Journal of Biotechnology

Journal of Biotechnology Volu me 118, Issue 4, Pages 339-470 (10 September 2005)

Regular papers 1.

2.

3.

4.

5.

6.

7.

Molecular characterisation of a versatile peroxidase from a Bjerkandera strain • Pages 339-352 Patrícia R. Moreira, C. Duez, D. Dehareng, A. Antunes, E. Almeida-Vara, J.M. Frère, F. Xavier Malcata and J.C. Duarte Cloning and expression of the malolactic gene of Pediococcus damnosus NCFB1832 in Saccharomyces cerevisiae • Pages 353-362 Rolene Bauer, Heinrich Volschenk and Leon M.T. Dicks Genetically engineered horseradish peroxidase for facilitated purification from baculovirus cultures by cation-exchange chromatography • Pages 363-369 Gustavo Levin, Fernando Mendive, Héctor M. Targovnik, Osvaldo Cascone and María V. Miranda Development and characterization of a monoclonal antibody directed against human telomerase reverse transcriptase (hTERT) • Pages 370-378 Dario Soldateschi, Sara Bravaccini, Brunilde Berti, Alessandra Brogi, Tiziana Benicchi, Claudia Soldatini, Laura Medri, Francesco Fabbri, Franca De Paola, Dino Amadori and Daniele Calistri Functional periplasmic secretion of organophosphorous hydrolase using the twin-arginine translocation pathway in Escherichia coli • Pages 379-385 Dong Gyun Kang, Gio-Bin Lim and Hyung Joon Cha Surface character of pulp fibres studied using endoglucanases • Pages 386-397 Lars Hildén, Priit Väljamäe and Gunnar Johansson On-line monitoring of yeast cell growth by impedance spectroscopy • Pages 398-405 A. Soley, M. Lecina, X. Gámez, J.J. Cairó, P. Riu, X. Rosell, R. Bragós and F. Gòdia

8.

9.

10.

11.

12.

Bacterial inclusion bodies are cytotoxic in vivo in absence of functional chaperones DnaK or GroEL • Pages 406-412 Nuria González-Montalbán, M. Mar Carrió, Sergi Cuatrecasas, Anna Arís and Antonio Villaverde Microbial biomass production from rice straw hydrolysate in airlift bioreactors • Pages 413-420 Yu-Guo Zheng, Xiao-Long Chen and Zhao Wang Capture of bacterial endotoxins using a supermacroporous monolithic matrix with immobilized polyethyleneimine, lysozyme or polymyxin B • Pages 421-433 Amro Hanora, Fatima M. Plieva, Martin Hedström, Igor Yu. Galaev and Bo Mattiasson Engineering microenvironment for expansion of sensitive anchoragedependent mammalian cells • Pages 434-447 Ser-Mien Chia, Pao-Chun Lin, Chai-Hoon Quek, Chao Yin, Hai-Quan Mao, Kam W. Leong, Xi Xu, Cho-Hong Goh, Mah-Lee Ng and Hanry Yu Production of glucuronan oligosaccharides using a new glucuronan lyase activity from a Trichoderma sp. strain • Pages 448-457 C. Delattre, P. Michaud, J.M. Lion, B. Courtois and J. Courtois

13.

Author Index • Pages 458-461

14.

Subject Index • Pages 462-464

15.

Volume Contents • Pages 465-469

16.

Contents Cont'd • Page I

Journal of Biotechnology 118 (2005) 339–352

Molecular characterisation of a versatile peroxidase from a Bjerkandera strain Patr´ıcia R. Moreira a,b , C. Duez c , D. Dehareng c , A. Antunes d , E. Almeida-Vara a , J.M. Frère c , F. Xavier Malcata b , J.C. Duarte a,∗ a

c

Departamento de Biotecnologia, Instituto Nacional de Engenharia, Tecnologia e Inova¸ca˜ o (INETI), Est. Pa¸co do Lumiar, 22, P-1649-038 Lisboa, Portugal b Escola Superior de Biotecnologia, Universidade Cat´ olica Portuguesa, R. Dr. António Bernardino de Almeida, P-4200-072 Porto, Portugal Centre d’Ingénierie des Protéines, Université de Liège, Institut de Chimie, B6, Sart-Tilman, B-4000 Liège, Belgium d D´ epartement de Chimie Générale et Physique, Laboratoire de Spectrométrie de Masse, Université de Liège, B6a, Sart-Tilman, B-4000 Liège, Belgium Received 7 January 2005; received in revised form 2 May 2005; accepted 12 May 2005

Abstract The cloning and sequencing of the rbpa gene coding for a versatile peroxidase from a novel Bjerkandera strain is hereby reported. The 1777 bp isolated fragment contained a 1698 bp peroxidase-encoding gene, interrupted by 11 introns. The 367 amino acid-deduced sequence includes a 27 amino acid-signal peptide. The molecular model, built via homology modelling with crystal structures of four fungal peroxidases, highlighted the amino acid residues putatively involved in manganese binding and aromatic substrate oxidation. The potential heme pocket residues (R44, F47, H48, E79, N85, H177, F194 and D239) include both distal and proximal histidines (H48 and H177). RBP possesses potential calcium-binding residues (D49, G67, D69, S71, S178, D195, T197, I200 and D202) and eight cysteine residues (C3, C15, C16, C35, C121, C250, C286, C316). In addition, RBP includes residues involved in substrate oxidation: three acidic residues (E37, E41 and D183)—putatively involved in manganese binding and H83 and W172—potentially involved in oxidation of aromatic substrates. Characterisation of nucleotide and amino acid sequences include RBP in versatile peroxidase group sharing catalytic properties of both LiP and MnP. In addition, the RBP enzyme appears to be closely related with the ligninolytic peroxidases from the Trametes versicolor strain. © 2005 Published by Elsevier B.V. Keywords: Ligninolytic peroxidases; White-rot fungi; Cloning and sequencing; Tertiary structure; Introns; rbpa gene

1. Introduction ∗ Corresponding author. Tel.: +351 217 165 141; fax: +351 217 163 636. E-mail address: [email protected] (J.C. Duarte).

0168-1656/$ – see front matter © 2005 Published by Elsevier B.V. doi:10.1016/j.jbiotec.2005.05.014

During the latest decade, research on the lignindegradation ability of fungi has mainly focused on

340

P.R. Moreira et al. / Journal of Biotechnology 118 (2005) 339–352

a few basidiomycetes known as white-rot fungi. The complexity of the lignin-degradation mechanisms, which encompass different enzymes (depending on the fungus), has prompted isolation of novel fungal strains—in attempts to find and characterise new enzymes, regarding reaction mechanism and substrate preference. A novel class of ligninolytic peroxidases, named versatile peroxidases, with high affinity for manganese and dyes, has been described; these enzymes can also oxidise 2,6-dimethoxyphenol (DMP) and veratryl alcohol (VA) in a manganese-independent reaction (Camarero et al., 1996, 1999, 2000; Giardina et al., 2000; Heinfling et al., 1998a,b,c; Martinez, 2002; Mester and Field, 1998; Palma et al., 2000; Ruiz-Duenas et al., 1999b, 2001). Until now, however, those enzymes have only been isolated from Pleurotus ostreatus, Pleurotus eryngii, Pleurotus pulmonarius, Bjerkandera adusta and Bjerkandera sp. strain BOS55. The recently sequenced enzymes MnPL1 and MnPL2 from cultures of P. eryngii exhibit high sequence and structural similarities with LiP from Phanerochaete chrysosporium; however, molecular models show a putative manganese interaction site near the internal propionate moiety of heme. The presence of this binding site is essential for direct oxidation of Mn2+ (Ruiz-Duenas et al., 1999a). MnP isoenzymes, purified from P. ostreatus and duly characterised by Giardina et al. (2000), are able to oxidise phenolic substrates both in the presence and absence of manganese, as already reported (Giardina et al., 2000; Heinfling et al., 1998b; Mester and Field, 1998) for isozymes produced by other strains of P. ostreatus, P. pulmonarius, P. eryngii, B. adusta and Bjerkandera sp. The catalytic properties of versatile peroxidases from both Pleurotus and Bjerkandera spp. are similar to each other and differ from those of LiP and MnP; these differences in catalytic properties, as well as in structural characteristics (Camarero et al., 1999; RuizDuenas et al., 2001), justify the description of such versatile peroxidases as part of a new peroxidase family in class II (fungal) peroxidases (Martinez, 2002). A fungal strain, which exhibits high decolourisation activities on poly R-478 and Remazol Brilliant Blue R (RBBR) dyes, was collected from rotting lignocellulosic material; this novel fungal strain was iden-

tified in our laboratory as belonging to the Bjerkandera genus and was tentatively named Bjerkandera sp. strain B33/3. Analysis of peroxidase activities in the extracellular fluid of said strain demonstrated the existence of lignin peroxidase, as well as manganese-dependent and manganese-independent peroxidase activities (Moreira et al., 2001). Several peaks were obtained after MonoQ ion-exchange chromatography of the extracellular fluid, which were characterised by distinct enzyme activities. One of the main peaks associated with the dye decolourising activity exhibited also properties characteristic of the novel class of versatile peroxidases, with ability to oxidise manganese, as well as VA and DMP in a manganese-independent reaction. This new enzyme was named RBP. The objectives of the present study were to clone the RBP-encoding gene and to characterise both the enzyme and its gene-encoding structures.

2. Materials and methods 2.1. Performance of culturing Bjerkandera sp. strain B33/3 was grown in CDBYE medium, as described previously in detail (Moreira et al., 2001). 2.2. Determination of proteinaceous features From the extracellular fluid of the Bjerkandera sp. B33/3 strain, the main enzyme responsible for the dye decolourisation was isolated and purified following resolution by Mono-Q ion-exchange chromatography. This enzyme, named RBP is able to oxidise manganese, as well as veratryl alcohol and 2,6-dimethoxyphenol in a manganese-independent reaction; hence, it can be included in the new group of versatile ligninolytic peroxidases. Oxidation of DMP was estimated by measuring the absorbance at 469 nm (30 ◦ C), during oxidation of 1 mM DMP in 50 mM sodium malonate buffer (pH 3.0 and 5.0), in the presence of 1 mM ethylene diamine tetra-acetic acid (EDTA). Oxidation of veratryl alcohol was determined by measuring absorbance at 310 nm (30 ◦ C), during the oxidation of 4.0 mM veratryl alcohol in 100 mM sodium tartrate buffer (pH 3.0 and 5.0). The RBBR decolourising activity was assayed spectrophotometrically by measuring the


decrease in absorbance at 595 nm (30 ◦ C). The enzymatic standard reaction mixture consisted of 0.05 mM RBBR and 70 mM sodium tartrate buffer (pH 5.0). The manganese(II) activity was assayed spectrophotometrically by measuring the decrease in absorbance at 238 nm (30 ◦ C). The enzymatic standard reaction mixture consisted of 0.1 mM MnSO4 and 100 mM sodium tartrate buffer (pH 5.0). The 2,2 -azinobis(3ethylbenzothiazoline-6-sulphonic acid) diammonium salt (ABTS) oxidising activity was assayed spectrophotometrically by measuring the decrease in absorbance at 420 nm (25 ◦ C). The enzymatic standard reaction mixture consisted of 0.5 mM ABTS and 100 mM sodium tartrate buffer (pH 5.0). Kinetic constants were determined for several substrates. The RBP enzyme in stake presents high affinity for (the oxidising substrate) manganese at pH 5.0, with Km values (96 ␮M). Small values of Km for DMP oxidation in the absence of manganese were also obtained (66 and 99 ␮M) depending at the pH tested, 3.0 and 5.0, respectively. The smallest Km values were obtained were recorded for oxidation of RBBR and ABTS (2 ␮M for both); this realisation is consistent with the hypothesis that RBP is a versatile peroxidase with high affinity for dye substrates, e.g. anthraquinone-derived and high redox compounds, e.g. ABTS (which are substrates usually preferred by plant peroxidases). Oxidation of VA is favoured at pH 3.0, as happens with LiP peroxidases and has an apparent Km value of 3400 ␮M. Purified native protein (200 ␮g) was N-deglycosylated overnight at pH 5.5 with 80 mU of endoglycosidase H from Roche Molecular Biochemicals (Mannhein, Germany). SDS-PAGE of native and deglycosylated proteins was performed on PhastGel® high-density polyacrylamide gels (Amersham Pharmacia Biotech, Uppsala, Sweden). The gel was calibrated using the low molecular weight calibration kit (Amersham Pharmacia Biotech) as standard. Isoelectric focusing of purified native protein was performed in PhastGel® (Amersham Pharmacia Biotech) IEF polyacrylamide gels, with a pH range of 3.0–9.0. The gel was calibrated using the low pI calibration kit (Amersham Pharmacia Biotech) as standard. Protein bands were stained with Coomassie Brilliant Blue R-250. Protein mass was also estimated by ESI–MS/MS TM using a Q-Tof-2 mass spectrometer (Micromass,

341

Manchester, UK) using ca. 25 ␮g of purified native protein. 2.3. Sequencing of tryptic peptides The purified native protein was digested via addition of 1:20 (w/w) of a 0.1 ␮g ␮l−1 solution of trypsin (Roche Molecular Biochemicals), reconstituted in 1 mM HCl. In order to accelerate digestion, 1% (v/v) acetonitrile was added to a 500 mM ammonium acetate buffer (pH 7.4) containing 20 mM CaCl2 , which was incubated for 12 h at 37 ◦ C. The tryptic peptides were fractionated and desalted by elution on a ZipTipC18 pipette tip (Millipore, Billerica, USA). Elution was carried out with mixtures of water/acetonitrile/acetic acid: 93/5/2, 78/20/2 and 48/50/2 (v/v/v), respectively. The fractions obtained were analysed by nano-ESI–MS/MS TM using a Q-Tof-2 mass spectrometer (Micromass) as described by Shevchenko et al. (2000). Selection of the ions analysed and adjustment of the collision energy were made manually. Peptides were delivered to the mass spectrometer by silica capillaries obtained from Protana (MDS Proteomics, Odense, Denmark). The BioLynx software (Micromass) was used for the prediction of MS/MS fragmentation patterns from peptide sequences and comparison with existing mass spectra. The fragmentation data obtained were analysed using the PepSeq sequencing software (Micromass). Database searches using BLAST 2.1.3 (NCBI http server) programs (Altschul et al., 1997), were performed with the sequences obtained, in order to eliminate those resulting from trypsin autodigestion. 2.4. Extraction of DNA and RNA DNA and RNA were extracted from a 6-day-old mycelium, which was collected by filtration, frozen with liquid N2 and ground to powder form with a mortar-and-pestle. For the DNA extraction, the cell powder was suspended in 10 mM Tris–HCl buffer (pH 7.6) containing 1 mM EDTA (TE), homogenised in a guanidinium thiocyanate and N-lauroylsarcosine denaturing suspension according to Pitcher et al. (1989) and incubated for 2 h in ice; 10 M ammonium acetate was then added and after a further 10 min of incubation, the total DNA was extracted with 24:1 (v/v) chloroform/isoamyl alcohol. The mixture was centrifuged, the aqueous phase was transferred to another tube and

342


0.8 volume of isopropanol was added. The DNA was washed with 70% (v/v) ethanol and the resulting pellet was air-dried and redissolved in TE buffer. The mixture was treated with RNase (50 ␮g ml−1 ) for 30 min at 37 ◦ C and extracted once again with 24:1 (v/v) chloroform/isoamyl alcohol. The supernatant was recovered and supplemented with 1:10 (v/v) 3 M sodium acetate (pH 5.2) and 2.5 volume of cold (−20 ◦ C) absolute ethanol. After centrifugation for 5 min at 3000 × g, the pellet was rinsed with 70% (v/v) ethanol, air-dried and redissolved in TE. The quality and purity of DNA was evaluated spectrophotometrically at 260 and 280 nm and by electrophoresis in TAE buffer on 1.5% agarose gel stained with ethidium bromide. The RNA isolation from Bjerkandera sp. B33/3 was performed with the SV total RNA isolation system (Promega Corporation, Madison, WI, USA) according to the manufacturer’s instructions. The quality and purity of the total RNA was evaluated spectrophotometrically. 2.5. Performance of PCR Escherichia coli Top10F , XL1-Blue or DH5␣ were used for cloning recombinant plasmids. The polymerase chain reaction (PCR) products were cloned into the pGEM-T easy vector (Promega Corporation). The oligonucleotides were purchased from Eurogentec (Liège, Belgium) or from Amersham Pharmacia Biotech. The PCR amplifications were performed with biotools DNA polymerase (Biotechnological and Medical Laboratories, Madrid, Spain). For the cloning of fragments adjacent to a known sequence, the LA-PCR in vitro cloning kit was used according to the supplier’s instructions of the (Takara Shuzo, Shiga, Japan). 2.6. Synthesis of cDNA The cDNA was prepared from the total RNA with TM ImProm-II reverse transcription system (Promega Corporation), with small modifications of the manufacturer’s protocol. The reverse transcription step was initiated with primer NTAG1 (5 -GCAGTGGTAACAACTTTTTTTTTTTTMM-3 ) on 0.08 or 0.8 ␮g of total RNA in the buffer supplied, supplemented with 3 mM MgCl2 , in the presence or absence of SequencerX Enhancer Solution F (Invitrogen, Carlsbad, USA). The annealing temperature was 37 ◦ C.

PCR amplification of the cDNA was obtained using NTAG1 and F-peroxi (5 -TGCCCCGACGGCGTIAACACC-3 ) primers. F-Peroxi was synthesised on the basis of the N-terminal sequence of the purified peroxidase (VAXPDGVNTA) assuming that the nonidentified residue in the sequence determined was cysteine, a highly conserved residue at the third position of mature lignin peroxidases. SequencerX Enhancer Solution F was added to the reaction mixture. The annealing temperature was 39 ◦ C in the first 3 PCR cycles and 50 ◦ C in the next 22 cycles. The quality of the cDNA was evaluated by electrophoresis on a 1.0% agarose gel. The PCR products (without further purification) were shotgun cloned into the pGem-T easy vector according to the manufacturer’s instructions. In parallel, the PCR products of interesting sizes were purified from agarose gels with the geneclean spin kit (Qbiogene, Illkirch, France). A 3 -A overhang was added to the purified PCR fragments with the Biotools DNA polymerase, before cloning them into the pGemT easy plasmid. The ligation products were used to transform E. coli DH5␣ competent cells. Colonies were screened by PCR with M13 universal and reverse-40 primers. Clones with inserts of expected size were transferred to 2× YT liquid medium, supplemented with 50 ␮g ml−1 ampicillin and grown overnight at 37 ◦ C. Plasmids were isolated with the GFX Micro Plasmid Prep Kit (Amersham Pharmacia Biotech) according to the manufacturer’s instructions and digested with EcoRI to estimate the size of the inserts by electrophoresis. 2.7. Sequencing of DNA Sequencing reactions were carried out using the AutoRead sequencing kit (Amersham Pharmacia Biotech), with the fluorescent (cy5) M13 Universal and Reverse primers, SP6, T7 or specific internal oligonucleotides. Electrophoresis was done with an ALF Express II DNA sequencer (Amersham Pharmacia Biotech). The PSI-BLAST 2.2.8 at NCBI http server (Altschul et al., 1997) and the BioEdit Sequence Alignment Editor (Hall, 1999) programs were used for analysis, alignment and comparison of sequences of amino acids, DNA and cDNA sequences. A dendogram based on similarities between sequences of mature peroxidases was obtained using


the UPGMA method (after bootstrap of sequences), from PAM distance matrix of progressive pairwise multiple sequence alignment using the BioEdit Sequence Alignment Editor and phylogenetic inference package (PHYLIP) program (Felsenstein, 1993). LiP, MnP, VP and CIP, as well as other peroxidases from Bjerkandera sp. strain B33/3 (B33/3), B. adusta (BA), P. ostreatus (PO), P. eryngii (PE), Ceriporiopsis subvermispora (CS), Ganoderma applanatum (GA), Dichomitus squalens (DS), Coprinus cinereus (CC), Lepista irina (LI), Phlebia radiate (PR), Trametes versicolor (TV) and P. chrysosporium (PC) were compared to one another. The GenBank accession numbers (gene or cDNA sequences) are as follows: B33/3-RBP, AY217015; PE-MnPL1, AF007221; PE-PS1, AF175710; PO-MnP, POU21878; PO-MnP2, POS243977; PO-MnP3, AB011546; CSMnP1, AF013257; CS-MnP2A, AF161078; CSMnP3, AF161585; GA-MnP1, AB035734; DSMnP1, AF157474; DS-MnP2, AF157475; LI-VPS1, CAD56164; CC-CIP, X70789; PR-LigIII, P20010; PR-MnP2, CAC85963; PR-MnP3, CAC84573; TVMP2, Z30668; TV-PGVII, Z54279; TV-PGV, X77154; TV-CVMNP, D86493; TV-LP7, Z30667; TV-LPGII, X75655; TV-LiP12, M64993; TV-VLG1, M55294; TV-MRP, AF008585; PC-MnP1, M60672; PC-MnP2, L29039; PC-MnP3, U70998; PC-LiPA (H8), M27401; PC-LiPB (H8), M37701; PC-LiPC, M63496; PCLiPD (H2), X15599; PC-LiPE (H8), M92644; PCLiPF, M77508; PC-LiPH (H8), M24082; PC-LiPJ, AF140062. 2.8. Modelling of proteinaceous structure Molecular modelling of the Bjerkandera sp. RBP versatile peroxidase (devoid of signal peptide) was performed by sequence homology using the PROMODII program (Guex and Peitsch, 1997; Peitsch, 1995, 1996). It was based on alignment of four fungal peroxidases, for which crystal models are available that present the largest similarities at the sequence level: LiP–H2 (Broohaven PDB entries 1QPAA: 66.35% identity and 1QPAB: 66.35% identity); LiP–H8 (1LLP: 64.75% identity); MnP1 (1MNP: 53.58% identity) from P. chrysosporium. The complete enzyme, with its heme moiety, two calcium ions and surrounding water molecules, was built by superposition with the A chain of QPA peroxidase (1QPA) with the program

343

InsightII (Biosym Technologies, 1993a). The geometry of the whole system was optimised at the molecular mechanics (MM) level (Burkert and Allinger, 1982) with Amber force field (Weiner and Kollman, 1981; Weiner et al., 1986) using the program Discover (Biosym Technologies, 1993b) on a SGI Indy workstation. A dielectric constant equal to 2 was chosen and the threshold for convergence was selected as maxi˚ −1 . mum force fixed at 0.02 kcal mol−1 A The Swiss-PDBViewer v3.7 SP5 (Guex and Peitsch, 1997) and Rasmol v2.7.2.1 (Bernstein, 2003) programs were used for model comparison and presentation.

3. Results 3.1. Determination of proteinaceous features The mass of native and deglycosylated RBP, determined with SDS-PAGE on PhastGel® high-density polyacrilamide gels were ca. 45.7 and 43.2 kDa, respectively, indicating that the mass of non-proteinaceus residues amounts to 5.6% of the whole molecular mass. The mass of the RBP, which was also estimated by ESI–MS/MS, is ca. 37 kDa. The isoelectric point estimated by gradient PAGE is ca. 3.5. 3.2. Sequencing of tryptic peptides Sequencing of two internal peptides generated from RBP via tryptic digestion was possible using mass spectrometry, two positive ESI m/z spectrum were obtained. The sequences obtained were DSVTDILNR and LQSDADFAR. The first peptide presented a theoretical m/z of 516.7702 for a measured value of 516.7416 m/z, thus presenting a delta mass of −55 ppm. The second peptide presented a theoretical m/z of 511.7493 for a measured value of 511.7228 m/z, thus presenting a delta mass of −52 ppm. 3.3. Synthesis and cloning of rbpa cDNA The 1139 bp cDNA sequence (GenBank AF490538) corresponded to a single open reading frame. As expected, the deduced amino acid sequence of this fragment contained both the (previously identified) DSVTDILNR and LQSDADFAR internal tryptic peptides.

344


3.4. Cloning of rbpa gene With the purpose of cloning the gene encoding for RBP, the oligonucleotides F-peroxi (5 -TGCCCCGACGGCGTIAACACC-3 ) and Cterm6 (5 CACAATTCCTACGACGACGCCTTATTCCCTCC3 )—complementary to the 3 end of the cDNA sequence, were used in a PCR experiment pertaining to the total genomic DNA. A 1625 bp fragment was amplified and cloned. Six different templates were completely sequenced on both strands using internal specific oligonucleotides as primers. The Bestfit program from GCG Wisconsin package (Accelrys, San Diego CA, USA) clearly identified the correspondence between the amplified fragment and the cDNA and allowed accurate localisation of the introns: 10 short introns interrupted the ORF encoding the mature protein. Since the F-peroxi primer encodes seven amino acids beginning at the third residue of the mature protein N-terminus, the amplified fragment was devoid of the 5 end of the gene. In order to obtain it, the oligonucleotide Nterm2 (5 -GCACTTCTTCGCCGCACTCGCCGCCGTC-3 ), directed towards the 5 end of the gene, was combined with an hexamer (random priming) in a PCR reaction with the total genomic DNA; the PCR products were cloned into the pGEM-T easy vector and to improve the specificity of amplification, a second PCR was performed on the ligation mixtures—with the Nterm3 oligonucleotide (5 -GTTCTGTTGGATGTCATCACGGACGGCG-3 ) and the M13 Universal or Reverse primer. One clone contained an ORF, with a good score of similarity with the signal peptides of manganese or lignin peroxidases. The pair of oligonucleotides, 6 h complementary to the sequence encoding the peptide DDIEPNFHANN (5 -CGTTGTTGGCGTGGAAGTTGGGCTCGATGTCGTC-3 ) and PS1 (5 -ATGGCCTTCAAGCAACTCCTCACTG-3 ), allowed the expected signal peptide to be amplified. No promoter region was obtained for this gene. The sequence of the 1771 bp DNA fragment (Fig. 1) isolated during this research effort received the Accession no. AY217015 in the GenBank database. Referring to the GT–AG rule of the intron splice sequences and to the homology between this DNA fragment and the cDNA previously sequenced, it was possible to predict the intron/exon structure of the rbpa

gene. The 11 introns (for a total of 597 bp), with an average length of 54 bp and their processing sequences are highlighted in Fig. 1. The translation initiation and termination codons, as well as a polyadenylation signal on the flanking 3 region and the amplification primer locations are also indicated. The intron splice junction sequences strictly adhere to the GT–AG rule and all of the putative internal Lariat sites conform to the NNHTNAY rule. A schematic representation of the number, size and position of introns in the rbpa gene, as compared with those of related genes encoding fungal peroxidases from B. adusta, T. versicolor, P. ostreatus, P. eryngii and P. chrysosporium, is depicted in Fig. 2. The rbpa open reading frame has a 58% G + C content (62% for the first, 48% for the second and 65% for the third position, respectively); a much lower (42%) G + C content was found in the intron sequences. The 367 amino acid-deduced sequence (including a 27 amino acid-signal peptide) is provided in Fig. 3. The deduced molecular mass of the mature protein is 35.7 kDa and its (theoretical) isoelectric point is 4.3. RBP does not contain any tyrosine residue; this is characteristic of most ligninolytic peroxidases. Tyrosine residue presence would otherwise result in oxidation of the enzyme (Camarero et al., 2000). The sequence includes one potential N-glycosylation site (N103) according to the NXS/T consensus, which could account for the 5.6% N-linked carbohydrate. The deduced heme pocket residues (R44, F47, H48, E79, N85, H177, F194 and D239), that include both distal and proximal histidines (H48 and H177), are also highlighted in Fig. 3. RBP possesses potential calcium-binding residues (D49, G67, D69, S71, S178, D195, T197, I200 and D202) and eight cysteine residues (C3, C15, C16, C35, C121, C250, C286, C316), which are characteristics also shared by other fungal peroxidases (Camarero et al., 2000). In addition, RBP includes residues involved in substrate oxidation, three acidic residues (E37, E41 and D183)—putatively involved in manganese binding and H83 and W172—potentially involved in oxidation of aromatic substrates (Banci, 1997; Martinez, 2002; Ruiz-Duenas et al., 2001). The sequence relationship between 38 selected fungal (mature) peroxidases is illustrated in the dendogram


345

Fig. 1. Nucleotide sequence of the rbpa gene from Bjerkandera sp. strain B33/3. The polyadenylation signal AATAA in the 3 -flanking region, as well as the translation initiation ATG and termination TAA codons are indicated in underlined bold characters. The 11 deduced introns are indicated in lower case and the intron processing sequences (5 - and 3 -splicing and Lariat sites) are indicated in bold. The boxes denote the location of some important primers used for cloning the gene.

346


Fig. 1. (Continued ).

Fig. 2. Positions of exons (white) and introns (black) in genes encoding various LiP, MnP and other peroxidases. The second vertical line corresponds to the beginning of the mature protein. The distance between every line corresponds to 200 nucleotides. (1) C. cinereus peroxidase (GenBank no. X707892); (2) T. versicolor manganese peroxidase isoenzyme MP2 (GenBank no. Z30668); (3) T. versicolor lignin peroxidase isoenzyme LP7 (GenBank no. Z30667); (4) T. versicolor peroxidase PGV (GenBank no. X77154); (5) B. adusta lignin peroxidase (GenBank no. E03952); (6) P. ostreatus manganese peroxidase MnP2 (GenBank no. AJ243977); (7) P. eryngii polyvalent peroxidase MnPL1 (GenBank no. AF007224); (8) P. eryngii versatile peroxidase PS1 (GenBank no. AF175710); (9) P. chrysosporium manganese peroxidase MnP1 (GenBank no. M77513); (10) P. chrysosporium manganese peroxidase MnP2 (GenBank no. L29039); (11) P. chrysosporium manganese peroxidase MnP3 (GenBank no. U70998); (12) P. chrysosporium lignin peroxidase LiPDH2 (GenBank no. X15599); (13) P. chrysosporium lignin peroxidase LiPEH8 (GenBank no. M92644); (14) P. chrysosporium lignin peroxidase LiPH8 (GenBank no. M27884); (15) P. chrysosporium lignin peroxidase LiPCH10 (GenBank no. M63496); (16) Bjerkandera sp. strain B33/3 RBP peroxidase (GenBank no. AY217015).


347

Fig. 3. Amino acid sequence of the RBP versatile peroxidase from Bjerkandera sp. strain B33/3 deduced from the rbpa gene. The 367 amino acid-deduced sequence includes a 27 amino acid-signal peptide containing a protease-splicing sequence (shown in bold and italic). The underlined bold letters indicate a putative N-glycosylation residue (NMT), the eight cysteine residues, the heme pocket residues (R44, F47, H48, E79, N85, H177, F194 and D239) that include both distal and proximal histidines (H48 and H177, respectively), the potential calcium binding residues (D49, G67, D69, S71, S178, D195, I197, I200 and D202) and the residues involved in substrate oxidation (E37, E41 and D183 for manganese and H83, F149, W172 and A240 for aromatic substrates).

Fig. 4. Phylogenetic relationship between mature fungal peroxidases. The dendogram was built from PAM distances between mature proteins using the PHYLIP software package after multiple alignment of 38 fungal peroxidase sequences and using the UPGMA method. The presence of manganese binding site (*) and of exposed tryptophan involved in VA oxidation (䊉) are duly indicated. Nodes are represented as squares and bootstrap values of more than 50% (as percentage) are represented as (). Nodes with bootstrap values of less than 50% are represented as black squares.

348


Fig. 5. Schematic molecular representation of the RBP peroxidase from Bjerkandera strain B33/3, after homology modelling with PROMODII using LiP and MnP crystal models as templates and refinement by Discover. The positions of ␣-helices (A–J) and of the C-termini (Ct) and N-termini (Nt) are duly indicated.

three residues per strand (V1–C3 and N8–A10; F129–L131 and I284–C286; A180–A182 and S192–F194) complete the fold. Four disulphide bonds (C3:C16, C15:C286, C35:C121 and C250:C316) are possible, but are not shown for the sake of simplicity of representation. Two hypothetical, long-range electron transfer (LRET) pathways, presumably involved in aromatic substrate oxidation, would include exposed histidine H83 (proceeding via A84 and N85 and H-bonded to distal histidine, H48) and exposed tryptophan W172 (proceeding via L173 and H-bonded to the porphyrin ring) (Martinez, 2002; Ruiz-Duenas et al., 2001). When the predicted molecular model of Bjerkandera sp. RBP was superimposed on crystal models of P. chrysosporium peroxidases, more similar distances between backbone C␣ carbon atoms were obtained with LiP than with MnP. The root mean square (rms) ˚ (340 residues computed) after distances were ca. 1.43 A ˚ superimposition with LiP–H2 (PDB 1QPAA), 1.77 A (337 residues computed) with LiP–H8 (PDB 1LLP) ˚ (332 residues computed) with MnP1 (PDB and 2.13 A 1MNP). Low values were also obtained via superimposition of RBP and predicted (non-crystallographic) molecular models of versatile peroxidases from P. eryn˚ (328 gii. The RMS distances for RBP were 2.00 A residues computed) after superimposition with MnPL1 (sometimes referred to as VP–PL) (PDB 1A20) and ˚ (337 residues computed) with PS1 (PDB 1QJR). 1.81 A

4. Discussion in Fig. 4. Our dataset produced a single, fully resolved dendogram, but bootstrap support for some nodes is weak. 3.5. Modelling of proteinaceous structure A three-dimensional model for the mature RBP was obtained by homology modelling, as made available in Fig. 5; this molecular model includes 12 ␣-helices, named according to the cytochrome C peroxidase nomenclature (Finzel et al., 1984), viz.: helices A (A14–I25), B (V39–A50), B (P55–P59), B (S71–I74), C (E90–H102), D (A106–S119), E (V152–A162), F (I167–H177), G (Q204–T209), H (Q237–R244), I (A249–S254) and J (Q259–L273). Three short, antiparallel ␤-sheets, consisting of only

The molecular mass calculated from the deduced amino acid sequence and including the carbohydrate content, is ca. 37 kDa; this figure is in good agreement with that obtained via mass spectrometry. The discrepancy value obtained via SDS-PAGE could arise because of interference by the carbohydrate moiety; it has been suggested (Matagne et al., 1991) that a very large negative charge at neutral pH is responsible for anomalous behaviour of proteins in SDS/PAGE. The experimental isoelectric point (pI) of 3.5 is significantly different from the value 4.3 derived from the amino acid sequence; this realisation may be related to posttranslational processing, namely glycosylation, as described elsewhere (Gianazza, 1995). The abundance of acidic residues in the RBP amino acid sequence


(12.4% versus 4.4% of basic residues) is consistent with such a low pI. The rbpa nucleotide sequence revealed 98% identity with the cDNA over 1017 nucleotides. Two residues are different between the proteins translated from the cDNA and the rbpa gene; both changes, T167 in the cDNA to I167 in the rbpa gene and L301 to S301, respectively, are located in highly variable zones arising in fungal peroxidases. The primary structures of several fungal peroxidases contain either threonine or isoleucine at position 167, exactly at the beginning of ␣-helix F, without any effect on the secondary and tertiary structures of the protein. Furthermore, leucine or serine are both likely to be present at position 301. These results strongly suggest that both sequences might encode two very similar proteins (i.e. bearing more than 99% identity), most likely encoded by two allelic forms of the same gene. The work discussed hereafter relates only to the rbpa gene. The rbpa orf has a high codon usage bias, especially in the third codon base that is likely correlated to high gene expression, as detected during enzyme purification. RBP was the most abundant peroxidase produced and was secreted into the extracellular medium by Bjerkandera sp. B33/3. The mature RBP is preceded by a 27 amino acidleader sequence, characterised by a molecular mass of 2.9 kDa. Prediction of this signal peptidase cleavage site (Nielsen et al., 1999) suggests that RBP is synthesised as a preproenzyme, encompassing a 22 amino acid-prepeptide followed by a 5 amino acid-propeptide (Gold and Alic, 1993). The 22 amino acid-N-terminus of the RBP signal prepeptide consists of a positively charged region of 4 residues, following the initiation methionine residue (MAFKQ, with a net charge of +1) and of a hydrophobic region of 12 residues (LLTAALSIALAL), followed by a neutral but polar region of 5 residues (PFSQA). The five-residue propeptide (AITRR) possesses a cleavage site after a dibasic site consisting of two arginine residues; this indicates that its removal might involve an endoprotease belonging to the serine protease family. Many secreted proteins contain similar propeptides, which might be important for accurate cleavage by signal peptidases (Gold and Alic, 1993). Possible roles of propeptides involve protein targeting, polypeptide folding or enzyme maintaining in an inactive state during translocation. The RBP signal peptide might thus be cleaved during transit into the

349

endoplasmic reticulum, whereas its propeptide cleavage probably occurs in the trans-Golgi complex (Gold and Alic, 1993). However, the actual biological significance of the propeptide remains to be elucidated. Signal peptides are known to control the entry of eukaryotic proteins into the secretory pathway. RBP has all typical characteristics of a secreted protein, in agreement with the fact that it was first isolated in the extracellular fluid. One potential N-glycosylation residue (NMT), eight cysteine residues and heme pocket residues (R44, F47, H48, E79, N85, H177, F194 and D239) including both distal and proximal histidines H48 and H177, were detected. Potential calcium binding residues (D49, G67, D69, S71, S178, D195, I197, I200 and D202), as well as residues involved in substrate oxidation (E37, E41 and D183 for manganese and H83, F149, W172 and A240 for aromatic substrates) are also present (Banci, 1997; Martinez, 2002; Ruiz-Duenas et al., 2001). This further confirms that the previously isolated enzyme, as well as that encoded by the rbpa gene is a versatile peroxidase (Martinez, 2002; Ruiz-Duenas et al., 2001), which bears the ability to oxidise both manganese and aromatic substrates. On the basis of homology with cDNA and using the GT–AG rule of the intron splice sequences, it was possible to predict the intron/exon structure of the 1771 bp rbpa gene. The consensus sequences for fungal introns, 5 GTDNNN, YAG-3 , is highly conserved in the 1771 bp fragment, as already observed by Kimura et al. (1991) for the LiP gene of B. adusta. As happens in other filamentous fungi, all introns of the rbpa gene contain a NNHTNAY sequence close to the 3 -splicing site, which is related to the yeast TACTAAC sequence. Furthermore, the putative internal Lariat fully agrees with the NNHTNAY rule. Few introns of filamentous fungi follow exactly the yeast consensus, but most have CT at positions 4 and 5 of the element and A at position 6, with the consensus abiding to the aforementioned rule (Unkles, 1992). A previous study (Camarero et al., 2000) revealed that several groups of genes can be established (according to the intron positions) including genes coding for peroxidases of the same type (MnP, LiP or VP) isolated from the same genus. In fact, the MnP- and LiPencoding genes from P. chrysosporium exhibit very different intron positions and this is also the case in

350


Bjerkandera sp. genes (Fig. 2). The intron positions in fungal peroxidases presented in Fig. 2 emphasize that the rbpa gene is highly similar to the P. chrysosporium LiPEH8-, LiPH8- and LiPH10-encoding genes. In the dendogram that compares 38 sequences of mature fungal peroxidases (Fig. 4), the C. cinereus enzyme appeared unrelated to the white-rot fungi peroxidases. The T. versicolor manganese-repressed peroxidase (MRP) and the G. applanatum MnP1 form the well-separated (strongly supported by bootstrapping) Group V cluster. Analysis of the G. applanatum MnP1 sequence indicates that it might possess a long range hypothetical electron transfer pathway beginning at W172, thus revealing a new possible versatile peroxidase. Within the remainder of the white-rot peroxidases, the P. chrysosporium, C. subvermispora, P. radiate and D. squalens MnPs are clustered together in group I, which is separated from the other ligninolytic enzymes. The LiP clade (Group IV) is backed up by a very weak support, with a bootstrap value of 35%, although the internal clades that separate the Trametes and Phanerochaete LiPs are strongly supported by 97 and 100% bootstrap values, respectively. All Pleurotus sp. peroxidases including the versatile peroxidases, are clustered in Group II with L. irina VPS1, a putative versatile peroxidase (Zorn et al., 2003). Groups II and III are strongly supported in this dendogram, with 83 and 87% bootstrap values, respectively. Furthermore, it is possible to describe a versatile peroxidase broader clade (VP), that includes both Groups II and III, as already described by Martinez (2002) and which is supported by bootstrapping with a slightly lower score (70%). Well-characterised versatile peroxidases appear as two separate groups (Groups II and III); Pleurotus versatile peroxidases appear together, whereas RBP is associated with manganese peroxidases and other undefined peroxidases from Trametes sp. (not clustered in Group I), such as a MnP from P. radiate. At present, it is not possible to assess the presence of multiple genes coding for RBP in Bjerkandera, even though the differences between cDNA and DNA sequences might indicate so. The overall topology of such globular proteins as peroxidases is usually characterised by polypeptide segments (␣-helices and ␤-strands) linked by tight turns or bends, which are almost always on the molecule surface. A three-dimensional model for the

mature protein RBP (Fig. 5) was obtained by homology modelling. Its main characteristics are similar to those of fungal peroxidase models, especially those of versatile peroxidases. The most important feature of the molecular model is that it provides a structural basis to explain the catalytic properties of RBP. Its ability to oxidize manganese with high substrate affinity is related to the presence of a manganese-binding site, that enables oxidation of this cation by the internal heme moiety (Banci, 1997; Martinez, 2002; Ruiz-Duenas et al., 2001). The two residues, H83 and W172, located on the protein surface could be involved in oxidation of aromatic substrates. It is noteworthy that three aromatic residues (F47, F159 and F165) are close to one another and could also play a role in the LRET (Martinez, 2002; Ruiz-Duenas et al., 2001). When the predicted molecular model of RBP was superimposed onto crystal models of P. chrysosporium peroxidases, more similar distances between backbone C␣ carbons were obtained with LiP than with MnP. These results are also in agreement with those obtained from the intron and sequence analysis. Overall, our results indicate that the enzyme produced by the Bjerkandera sp. strain B33/3, of which some biochemical and molecular characteristics were conveyed in this communication, can be included, together with those produced by B. adusta and Pleurotus sp., in a new class of versatile peroxidases that share catalytic properties with both LiP and MnP.

Acknowledgements P. Moreira is a doctoral fellowship holder (PRAXIS XXI/BD/15825/98), granted by FCT (Portugal). C. Duez is a ‘Chercheur Qualifié’, funded by Fonds National de la Recherche Scientifique (Belgium). The work in Liège was specifically supported by the Belgian Programme on Interuniversity Poles of Attraction (PAI No. P05/33), funded by Services Fédéraux des Affaires Scientifiques, Techniques et Culturelles.

References Altschul, S.F., Madden, T.L., Schäffer, A.A., Zhang, J., Zhang, Z., Miller, W., Lipman, D.J., 1997. Gapped BLAST and PSI-

P.R. Moreira et al. / Journal of Biotechnology 118 (2005) 339–352 BLAST: a new generation of protein database search programs. Nucleic Acids Res. 25, 3389–3402. Banci, L., 1997. Structural properties of peroxidases. J. Biotechnol 53, 253–263. Bernstein, H., 2003. Program RasMol Molecular Rendering. Distributed by author. Biosym Technologies I., 1993a. Program InsightII. Biosym Technologies I., 1993b. Program Discover. Burkert, U., Allinger, N.L., 1982. Molecular Mechanics. American Chemical Society, Washington, DC, USA. Camarero, S., Bockle, B., Martinez, M.J., Martinez, A.T., 1996. Manganese-mediated lignin-degradation by Pleurotus pulmonarius. Appl. Environ. Microbiol. 62, 1070– 1072. Camarero, S., Sarkar, S., Ruiz-Duenas, F.J., Martinez, M.J., Martinez, A.T., 1999. Description of a versatile peroxidase involved in the natural degradation of lignin that has both manganese peroxidase and lignin peroxidase substrate interaction sites. J. Biol. Chem. 274, 10324–10330. Camarero, S., Ruiz-Duenas, F.J., Sarkar, S., Martinez, M.J., Martinez, A.T., 2000. The cloning of a new peroxidase found in lignocellulose cultures of Pleurotus eryngii and sequence comparison with other fungal peroxidases. FEMS Microbiol. Lett. 191, 37–43. Felsenstein, J., 1993. Phylogeny Inference Package (PHYLIP). Distributed by author. Finzel, B.C., Poulos, T.L., Kraut, J., 1984. Crystal structure of yeast ˚ resolution. J. Biol. cytochrome C peroxidase refined at 1.7 A Chem. 259, 3027–3036. Gianazza, E., 1995. Isoelectric focusing as a tool for the investigation of posttranslational processing and chemical modifications of proteins. J. Chromatogr. A 705, 67– 87. Giardina, P., Palmieri, G., Fontanella, B., Rivieccio, V., Sannia, G., 2000. Manganese peroxidase isoenzymes produced by Pleurotus ostreatus grown on wood sawdust. Arch. Biochem. Biophys. 376, 171–179. Gold, M.H., Alic, M., 1993. Molecular biology of the lignin degrading basidiomycete Phanerochaete chrysosporium. Microbiol. Rev. 57, 605–622. Guex, N., Peitsch, M.C., 1997. Swiss-model and the SwissPdbViewer: an environment for comparative protein modelling. Electrophoresis 18, 2714–2723. Hall, T.A., 1999. BioEdit: a user-friendly biological sequence alignment editor and analysis program for Windows 95/98/NT. Nucleic Acids Symp. Ser. 41, 95–98. Heinfling, A., Martinez, M.J., Martinez, A.T., Bergbauer, M., Szewzyk, U., 1998a. Purification and characterization of peroxidases from the dye-decolorizing fungus Bjerkandera adusta. FEMS Microbiol. Lett. 165, 43–50. Heinfling, A., Martinez, M.J., Martinez, A.T., Bergbauer, M., Szewzyk, U., 1998b. Transformation of industrial dyes by manganese peroxidases from Bjerkandera adusta and Pleurotus eryngii in a manganese-independent reaction. Appl. Environ. Microbiol. 64, 2788–2793. Heinfling, A., Ruiz-Duenas, F.J., Martinez, M.J., Bergbauer, M., Szewzyk, U., Martinez, A.T., 1998c. A study on reducing

351

substrates of manganese-oxidizing peroxidases from Pleurotus eryngii and Bjerkandera adusta. FEBS Lett. 428, 141– 146. Kimura, Y., Asada, Y., Oka, T., Kuwahara, M., 1991. Molecular analysis of a Bjerkandera adusta lignin peroxidase gene. Appl. Microbiol. Biotechnol. 35, 510–514. Martinez, A.T., 2002. Molecular biology and structure–function of lignin degrading heme peroxidases. Enzyme Microb. Technol. 30, 425–444. Matagne, A., Joris, B., Frere, J.M., 1991. Anomalous behaviour of a protein during SDS/PAGE corrected by chemical modification of carboxylic groups. Biochem. J. 280, 553– 556. Mester, T., Field, J.A., 1998. Characterization of a novel manganese peroxidase–lignin peroxidase hybrid isozyme produced by Bjerkandera species strain BOS55 in the absence of manganese. J. Biol. Chem. 273, 15412– 15417. Moreira, P.R., Almeida-Vara, E., Sena-Martins, G., Polonia, I., Malcata, F.X., Duarte, J.C., 2001. Decolourisation of Remazol Brilliant Blue R via a novel Bjerkandera sp. strain. J. Biotechnol. 89, 107–111. Nielsen, H., Brunak, S., von Heijne, G., 1999. Machine learning approaches for the prediction of signal peptides and other protein sorting signals. Protein Eng. 12, 3–9. Palma, C., Martinez, A.T., Lema, J.M., Martinez, M.J., 2000. Different fungal manganese-oxidizing peroxidases: a comparison between Bjerkandera sp. and Phanerochaete chrysosporium. J. Biotechnol. 77, 235–245. Peitsch, M.C., 1995. Protein modeling by e-mail. Biotechnology 13, 658–660. Peitsch, M.C., 1996. ProMod and Swiss-model: internet-based tools for automated comparative protein modelling. Biochem. Soc. Trans. 24, 274–279. Pitcher, D.G., Saunders, N.A., Owen, R.J., 1989. Rapid extraction of bacterial genomic DNA with guanidium thiocyanate. Lett. Appl. Microbiol. 8, 151–156. Ruiz-Duenas, F.J., Martinez, M.J., Martinez, A.T., 1999a. Heterologous expression of Pleurotus eryngii peroxidase confirms its ability to oxidize Mn2+ and different aromatic substrates. Appl. Environ. Microbiol. 65, 4705– 4707. Ruiz-Duenas, F.J., Martinez, M.J., Martinez, A.T., 1999b. Molecular characterization of a novel peroxidase isolated from the ligninolytic fungus Pleurotus eryngii. Mol. Microbiol. 31, 223– 235. Ruiz-Duenas, F.J., Camarero, S., Perez-Boada, M., Martinez, M.J., Martinez, A.T., 2001. A new versatile peroxidase from Pleurotus. Biochem. Soc. Trans. 29, 116–122. Shevchenko, A., Chernushevich, I., Wilm, M., Mann, M., 2000. De novo peptide sequencing by nanoelectrospray tandem mass spectrometry using triple quadrupole and quadrupole/time-of-flight instruments. In: Chapman, J.R. (Ed.), Protein and Peptide Analysis. Humana Press, Totowa, NJ, pp. 1–16. Unkles, S.E., 1992. Gene organization in industrial filamentous fungi. In: Kinghorn, J.R., Turner, G. (Eds.), Applied Molecular

352


Genetics of Filamentous Fungi. London University Press, Cambridge, pp. 28–53. Weiner, S., Kollman, P.A., 1981. AMBER: assisted model building with energy refinement. A general program for modeling molecules and their interactions. J. Comp. Chem. 2, 287– 303.

Weiner, S., Kollman, P.A., Nguyen, D.T., Case, D.A., 1986. An all atom force field for simulations of proteins and nucleic acids. J. Comp. Chem. 7, 230–252. Zorn, H., Langhoff, S., Scheibner, M., Nimtz, M., Berger, R.G., 2003. A peroxidase from Lepista irina cleaves beta-carotene to flavor compounds. Biol. Chem. 384, 1049–1056.


Cloning and expression of the malolactic gene of Pediococcus damnosus NCFB1832 in Saccharomyces cerevisiae Rolene Bauer a , Heinrich Volschenk b , Leon M.T. Dicks a,∗ b

a Department of Microbiology, Stellenbosch University, 7600 Stellenbosch, South Africa Department of Food and Agricultural Sciences, Cape Peninsula University of Technology, 8000 Cape Town, South Africa

Received 18 November 2004; received in revised form 29 March 2005; accepted 6 April 2005

Abstract Wine production is characterized by a primary alcoholic fermentation, conducted by Saccharomyces cerevisiae, followed by a secondary malolactic fermentation (MLF). Although most lactic acid bacteria (LAB) have the ability to metabolize l-malate, only a few species survive the high ethanol and SO2 levels in wine. Wines produced in colder viticultural regions have a lower pH than wines produced in warmer regions. The decarboxylation of l-malate in these wines leads to an increase in pH, more organoleptic complexity and microbiological stability. MLF is, however, difficult to control and problems often occur during filtering of such wines. Pediococcus spp. are known to occur in high pH wines and have strong malolactic activity. However, some pediococci synthesize exocellular polysaccharides, which may lead to abnormal viscosity in wine. In this study, the malolactic gene from Pediococcus damnosus NCFB1832 (mleD) was cloned into S. cerevisiae and co-expressed with the malate permease gene (mae1) of Schizosaccharomyces pombe. Expression of the mleD gene was compared to the expression of two other malolactic genes, mleS from Lactococcus lactis MG1363 and mleA from Oenococcus oeni Lal1. The genetically modified strain of S. cerevisiae decreased the level of l-malate in grape must to less than 0.3 g l−1 within 3 days. This is the first expression of a malolactic gene from Pediococcus in S. cerevisiae. © 2005 Elsevier B.V. All rights reserved. Keywords: Malolactic fermentation; Cloning; mleD gene; Saccharomyces cerevisiae

1. Introduction Winemaking normally involves two fermentation processes, i.e. an alcoholic fermentation conducted by Saccharomyces cerevisiae, followed by a malolactic ∗ Corresponding author. Tel.: +27 21 8085849; fax: +27 21 8085846. E-mail address: [email protected] (L.M.T. Dicks).

0168-1656/$ – see front matter © 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.jbiotec.2005.04.015

fermentation (MLF). This secondary fermentation is conducted by species of the genera Pediococcus, Lactobacillus, Leuconostoc and Oenococcus oeni, previously Leuconostoc oenos (Dicks et al., 1995a) and involves the conversion of l-malate to l-lactate. The direct consequence of this conversion is a decrease in total acidity, but MLF also plays a part in microbial stabilization, while the metabolic activity of the bacteria contributes to the organoleptic complexity of the wine.

354

R. Bauer et al. / Journal of Biotechnology 118 (2005) 353–362

The NAD+ /Mn2+ -dependent malolactic enzyme (MLE) transforms the C4 dicarboxylic acid l-malate to the C3 monocarboxylic acid l-lactate as a direct decarboxylation reaction without any free intermediates (Henick-Kling, 1995). This is contrary to the malic enzyme that converts l-malate to pyruvate, which can be further converted to l-lactate (Radler, 1986). The MLE has two or four identical subunits of 60–70 kDa (Caspritz and Radler, 1983; Spettoli et al., 1984; Naouri et al., 1990) and is homologous to the malic enzymes recorded for a number of other organisms. The amino acid sequences deduced from partially sequenced malolactic genes (mle) of Pediococcus, Leuconostoc, Oenococcus, Lactobacillus and Lactococcus spp. aligned with sequences recorded for 22 malic enzymes, but formed a distinct cluster with the malic enzymes of yeast and Escherichia coli (Groisillier and Lonvaud-Funel, 1999). Thus, far, only the mle genes from Lactococcus lactis subsp. lactis (mleS) and O. oeni (mleA) have been completely sequenced (Denayrolles et al., 1994; Labarre et al., 1996). Wild type S. cerevisiae metabolizes insignificant amounts of l-malate during alcoholic fermentation due to the absence of an active transport system for l-malate (Van Vuuren et al., 1995) and the low substrate affinity of its malic enzyme (Fuck et al., 1973). Schizosaccharomyces pombe, on the other hand, conducts an efficient malo-ethanolic fermentation under anaerobic conditions (Osothsilp and Subden, 1986) through the constitutive synthesis of malate permease, encoded by the mae1 gene (Grobler et al., 1995) and the malic enzyme, encoded by the mae2 gene (Viljoen et al., 1994). Volschenk et al. (1997b) increased the ability of S. cerevisiae to metabolize l-malate, by co-expressing the mae1 gene from S. pombe and the L. lactis malolactic gene (mleS). The recombinant yeast grew at pH values below that recorded for O. oeni (Kunkee, 1997) and metabolized l-malate to l-lactate within 3 days at 20 ◦ C in Cabernet Sauvignon and Shiraz grape musts (Volschenk et al., 1997a). In Chardonnay grape must, MLF was completed within 7 days at 15 ◦ C. In this paper, we describe the cloning, sequencing and heterologous expression of the mle gene from Pediococcus damnosus (mleD) in a laboratory strain of S. cerevisiae. The ability of the newly constructed strain to conduct MLF was compared with cells of the same yeast strain that have been transformed with the

mle genes from L. lactis subsp. lactis (mleS) and O. oeni (mleA), respectively.

2. Materials and methods 2.1. Bacterial strains, plasmids and culture conditions The strains and plasmids used in this study are listed in Table 1. E. coli DH5␣ was used as recipient strain for plasmid maintenance and library construction. The shuttle vector pTRKL2 was used for construction of the genome library. E. coli was cultured as described by Sambrook et al. (1989) and transformants were selected on either BHI agar (Difco, Difco Laboratories, Detroit, MI), supplemented with 150 ␮g ml−1 erythromycin (Ery), or LB agar (Difco) containing 100 ␮g ml−1 ampicillin. X-gal and IPTG were added at 70 and 40 ␮g ml−1 , respectively. P. damnosus NCFB1832 and L. lactis subsp. lactis MG1363 were grown at 30 ◦ C in MRS broth (Biolab, Biolab Diagnostics, Midrand, South Africa). O. oeni strain Lal1 was isolated from a commercial malolactic starter pack (Lallemand, Saint-Simon, France) and was cultured in acidic grape (AG) broth (Dicks et al., 1995b) at 30 ◦ C. S. cerevisiae was cultured at 30 ◦ C in YPD broth, as described by Sambrook et al. (1989). Yeast transformants were isolated on selective YNB agar plates, supplemented with the required amino acids (Volschenk et al., 1997a). 2.2. DNA isolation Genomic DNA of the LAB and S. pombe were isolated according to the methods described by Dellaglio et al. (1973) and Hoffman and Winston (1987), respectively. Plasmid DNA from E. coli was isolated by the method of Lee and Rasheed (1991) and purified through CsCl density gradient centrifugation (Sambrook et al., 1989). 2.3. Construction of the P. damnosus genomic library Genomic DNA was isolated from P. damnosus NCFB1832 as described before and partially digested with Sau3A (2.8 units to 125 ␮g DNA in a 500 ␮l reac-


355

Table 1 Strains and plasmids included in this study Description

Reference/source

Strain P. damnosus NCFB1832 L. lactis MG1363 O. oeni Lal1 E. coli DH5␣ S. cerevisiae YPH259

recA1 endA1 gyrA96 thi-1 relA1 hsdR17 supE44 80 lacZ∆M15 α ura3-52, lys2-801amber , ade2-101ochre , his3∆200, leu2-∆1

S. pombe IVPT 2010 Plasmids pTRKL2 pHVX2 YEP352-PGK1pt

Green et al. (1997) Department of Microbiology, Stellenbosch University, Stellenbosch, SA Lallemand, Saint-Simon, France Hanahan (1983) Sikorski and Hieter (1989) Institute for Wine Biotechnology, Department of Enology and Viticulture, University of Stellenbosch

Containing the lacZ-gene and Ery resistant marker YEplac181 (LEU2) containing the PGK1 promoter and terminator sequences YYCplac33 (URA3) containing the PGK1 promoter and terminator sequences

tion mixture) at 37 ◦ C for 2, 3, 5, 7 and 10 min, respectively. The samples were heated at 65 ◦ C for 15 min, followed by phenol–chloroform extraction and ethanol precipitation. Pooled samples of partially digested DNA were separated in a 10–40% linear sucrose gradient (Ausubel et al., 1994). Fragments between 4 and 10 kb were selected and ligated into plasmid pTRKL2 that has been linearized with BamHI. The constructs were transformed into E. coli DH5␣ that were made competent by using the RbCl-procedure, as adapted from a protocol obtained from the John Innes Institute, Norwich, England. The transformed cells were plated onto BHI Agar (Difco), supplemented with 150 ␮g ml−1 erythromycin (Ery), X-gal and IPTG as described before and incubated at 37 ◦ C. Transformants were handpicked with sterile toothpicks (50 white colonies per plate) and replica-plated on BHI agar supplemented with 150 ␮g ml−1 Ery, and incubated overnight at 37 ◦ C. Colonies were then washed from the agar surface and samples from each plate were stored in 40% (v/v) glycerol at −20 ◦ C. Cells collected from all plates were then pooled and incubated at 37 ◦ C for 3.5 h in BHI, supplemented with 150 ␮g ml−1 Ery. A glycerol freeze-culture prepared from this culture represented the genomic library. The number of transformants required to represent the entire genome (N) was calculated by using the equation N = ln(1 − p)/ln(1 − F/G), with p is the confidence

O’Sullivan and Klaenhammer (1993) Volschenk et al. (1997b) Institute for Wine Biotechnology, Department of Enology and Viticulture, University of Stellenbosch

limit set at 99.99%, G the genome size and F is the average size of DNA fragments ligated. The average size of the cloned fragments was determined by digesting the plasmid DNA, isolated from 50 transformants, with HindIII. Recombinant plasmid DNA was isolated and purified by CsCl density gradient centrifugation. 2.4. Screening of the P. damnosus genome library for a mleD gene All oligonucleotides were synthesized on a Beckman Oligo 1000 M DNA synthesizer (Beckman Instrument, California Avenue, CA). Primers for PCR were designed from two conserved sequence regions (boxes I and III) in the mle genes of L. lactis IL1441 (Denayrolles et al., 1994, 1995) and O. oeni IOEB 8413 (Labarre et al., 1996). The corresponding nucleotide sequences were 5 -ATC CAG TTG TTT ATG ATC (mle-oligo1F) (sense) and 5 -CAG TTC CTT GAA TRT CAT CRT T (R = A or G) (mle-oligo1R) (antisense). The DNA fragment amplified from the genome of P. damnosus NCFB1832 was labeled with [␣32 P]dATP (Amersham, Braunschweig, Germany) to be used as a probe for colony hybridization. Serial dilutions of the E. coli transformants containing the P. damnosus library were made and the cells plated on selective BHI agar, as described before. Colony hybridization was performed with the radiolabeled

356


probe under stringent conditions on MSI Nylon membranes (Osmonics, USA, Westborough) as recommended by the suppliers. 2.5. Amplification of the mae1 and mle genes Standard recombinant DNA techniques were used, as described by Ausubel et al. (1994). The mae1 gene was amplified from the genome of S. pombe by PCR, using oligonucleotides 5 -GAT CGA ATT CAT GGG TGA AAC TCA AGG AAA TC and 5 -GAT CAG ATC TTT AAA CGC TTT CAT GTT CAC T, based on the sequence of mae1 (Grobler et al., 1995). The primers introduced a EcoRI and BglII site at the 5 and 3 ends of the amplified mae1, respectively. PCR amplification of the L. lactis subsp. lactis MG1363 malolactic gene (mleS) was achieved by using the primers 5 -GAT CAG ATC TGA GGT TGT ACG ATG CGT GCA C and 5 GAT CCT CGA GCC CTT AGT ACT CTG GAT ACC AT, derived from sequences of mleS (Denayrolles et al., 1995). The primers introduced a BglII and XhoI site at the 5 and 3 ends, respectively. The malolactic gene of O. oeni Lal1 (mleA) was amplified by PCR using the oligonucleotide pairs 5 -GAT CGA ATT CGA GGA GAA AAT ATG ACA GAT CC and 5 -GAT CCT CGA GGC ATT CAT TAG TAT TTC GGA TC, derived from sequences of mleA (Labarre et al., 1996). These primers introduced a EcoRI and XhoI site at the 5 and 3 ends, respectively. The P. damnosus NCFB1832 malolactic gene (mleD) was amplified from an E. coli transformant, containing P. damnosus genomic DNA, that was identified through colony hybridization. The primers 5 -GAT CCT CGA GTG GAG GCT ATA AAT ATG GCA A and 5 -GAT CCT CGA GAG CTC ATG TAC TAT CTC TTA C were used, which introduced a XhoI site at the 5 and 3 of the mleD gene. 2.6. Cloning of the mae1 and mle genes into S. cerevisiae Standard recombinant DNA techniques were performed as described by Ausubel et al. (1994). The amplified mae1 fragment was digested with EcoRI and BglII, and cloned into the yeast plasmid YEP352PGK1pt to yield plasmid pRBMAE1. The PCR products representing the three malolactic genes mleD, mleA and mleS, were digested with corresponding restriction enzymes that were introduced at their 5 and

3 ends and cloned into the shuttle vector pHVX2 to yield plasmids pRBMleD, pRBMleA and pRBMleS, respectively. Orientation of the cloned genes was determined by sequencing. S. cerevisiae was co-transformed with pBRMAE1 and either pRBMleD, pRBMleA or pRBMleS using the lithium acetate procedure (Ausubel et al., 1994). 2.7. Sequencing of the cloned fragments DNA isolated with the miniprep method was sequenced by using the dideoxy chain termination method. An ABI Prism GigDye Terminator Cycle Sequencing Ready Reaction kit and an ABI Prism 377 DNA sequencer (PE Applied Biosystem, Foster City, CA) were used. Computer analysis of the sequences was performed by using DNA-MAN for Windows® (Lynnon Biosoft, Canada). Database searches were performed by using the BLASTN and BLASTX programs of the National Center for Biotechnology Information, Bethesda, MD (http://www.ncbi.nlm.nih.gov). Plasmid DNA from E. coli transformants containing P. damnosus genomic DNA that hybridized with the malolactic probes were sequenced with the PCR oligonucleotide mle-oligo1F. Plasmid pHVX2 containing the L. lactis MG1363 and O. oeni Lal1 mle enzymes were sequenced with oligonucleotides based on sequences of the PGK1 promoter (5 -GTT TAG TAG AAC CTC GT) and PGK1 terminator (5 AGC GTA AAG GAT GGG). Additional primers were synthesized based on new sequences generated. The nucleotide sequences of mleD and mleS have been submitted to GenBank with accession numbers AY450551 and AY450550. 2.8. Malolactic fermentation by S. cerevisiae Conversion of l-malate to l-lactate by recombinant strains of S. cerevisiae was determined by inoculating grape must (Denayrolles et al., 1995), supplemented with amino acids and adjusted to a pH of 3.3 (Volschenk et al., 1997a), with 106 cfu ml−1 . Cultures were incubated at 22 ◦ C until MLF was complete, with two shakings per day. The l-malate (initial concentration: 4.5 g l−1 ) and l-lactate concentrations during fermentation were measured enzymatically using the l-malic acid and l-lactic acid test kits (Roche Diagnostics, Mannheim, Germany), respectively.


357

3. Results

3.3. Nucleotide sequence analysis of mleD

3.1. Quality of the P. damnosus genomic library

The complete nucleotide sequence of the P. damnosus NCFB1832 mleD gene was established by sequencing both strands of the insert in pTRKL2 . A 1629-nucleotide open reading frame starting with an ATG initiation codon at position 334 and extending to the TAG stop codon at position 1960 was identified (Fig. 1). The open reading frame encoded a putative protein of 542 amino acids with a calculated molecular mass of 59,271 Da and a pI of 4.52. A potential ribosome binding site (RBS) at position 321 was identified. The mle genes of O. oeni Lal1 (mleA) and L. lactis MG 1363 (mleS) were also cloned into plasmid pTRKL2 and sequenced. The nucleotide sequence of mleA was identical to that of O. oeni IOEB 8413. A pairwise comparison of nucleotide sequences of mleS and the mle gene described for L. lactis IL1441 revealed 96.8% homology. Amino acid sequence comparison of the latter two genes revealed 99.1% homology. Pairwise comparison of the amino acid sequences revealed identity values of 72.5 and 64.4% for MleDp with MleAp and MleSp, respectively (Fig. 2). MleAp and MleSp are 67.1% identical (Fig. 2).

The size of the genome of Pediococcus spp. is approximately 1500 kb (Daniel, 1995). Since the average size of fragments cloned was 6 kb, and all the transformants contain inserts, the entire genome of P. damnosus NCFB1832 would be represented by 2303 transformants. A total of 3500 transformants were collected after E. coli transformation with the genomic library. Plasmid DNA representing the entire library was digested with HindIII. HindIII cuts pTRKL2 in the multiple cloning site (MCS) and twice at known positions outside the MCS. Two distinct bands were observed with a background smear between 1 and 10 kb (results not shown), suggesting that the library is representative. The plasmid pTRKL2 contains an EcoRI recognition site in the MCS and one site outside. Digestion of library DNA with EcoRI yielded one distinct band with a background smear, as expected (results not shown). 3.2. Cloning of the mleD gene of P. damnosus

3.4. Expression of mleD in yeast PCR amplification of the genomic DNA of P. damnosus with primers mle-oligo1F and mle-oligo1R, based on conserved sequences in other mle genes, generated fragments of approximately 580 bp, which corresponded to the expected size. To clone a copy of the gene, the PCR product was radiolabeled and used as a probe to screen the P. damnosus genomic library. Three E. coli transformants out of approx. 1500 screened, hybridized with this probe. Restriction enzyme digests of plasmid DNA isolated from these E. coli transformants differed from each other, but shared bands of similar sizes, suggesting that they contained partially overlapping genomic DNA inserts. The insert of one of the clones was sequenced by first using the two PCR primers (mle-oligo1F and mle-oligo1R) and then primers generated along the identified sequence. The entire mleD gene and flanking regions were sequenced. Amplification of the other two clones with primers mle-oligo1F and mle-oligo1R yielded DNA fragments with sequences identical to the first clone, which confirmed that the genomic library was representative.

The mleD ORF was cloned into the yeast expression vector pHVX2 described by Volschenk et al. (1997b), between the promoter and terminator sequences of the PGK1-gene to yield plasmid pRBMleD. The mle gene of O. oeni Lal1 (mleA) and L. lactis MG 1363 (mleS) were also cloned into the same vector leading to the plasmids pRBMleA and pRBMleS, respectively. Similarly, the S. pombe malate permease gene (mae1) was placed under control of the PGK1 promoter in the shuttle vector YEP352-PGK1pt. A recombinant strain of S. cerevisiae containing the malolactic gene of P. damnosus (mleD) and mae1 was constructed by co-transformation of plasmid pRBMleD and plasmid pRBMae1 into S. cerevisiae YPH259. Between 100 and 200 transformants were obtained per ␮g DNA. The presence of plasmids pRBM1eD and pRBMae1 in the recombinant strains was confirmed by restriction enzyme digests. Recombinant S. cerevisiae strains co-expressing the malolactic enzymes of L. lactis MG 1363 and O. oeni Lal1 and malate permease were constructed following the same strategy.

358


Fig. 1. Nucleotide and deduced amino acid sequences of mleD, and flanking DNA from Pediococcus damnosus NCFB 1832. The putative RBS is underlined. Numbers on the left represent the nucleotide positions.


359

Fig. 1. (Continued).

3.5. Malolactic fermentation by S. cerevisiae

4. Discussion

Malolactic fermentation in wine is regarded complete when the concentration of l-malate reaches 0.3 g l−1 (Martineau et al., 1995). S. cerevisiae transformants with the mleD gene converted l-malate to l-lactate within 3 days, reaching l-malate concentrations of below 0.3 g l−1 (Fig. 3). However, after 4 days of fermentation in grape must, only 2.8 g l−1 l-lactic was produced, compared to 3.3 g l−1 formed by transformants with mleA and mleS genes (Fig. 3). Yeast strains containing only plasmids pHVX2 or YEP352-PGK1pt (i.e. without mae1 or mle genes) converted 4.5 g l−1 malic acid to 3.0 g l−1 malic acid and produced only 1.5 g l−1 lactic acid after 4 days (Fig. 3).

This is the first report of a malolactic gene from P. damnosus cloned into and expressed by S. cerevisiae. Pairwise comparison of the amino acid sequences of mleDp with MleAp of O. oeni Lal1 and MleSp of L. lactis MG1363 revealed similarity values of 72 and 64%, respectively. The mleS nucleotide sequences of L. lactis MG1363 and L. lactis IL1441 revealed a high degree of homology (96.8%), while the nucleotide sequence recorded for the MLE of O. oeni Lal1 was identical to that recorded for the MLE of O. oeni IOEB 8413. A partial sequence covering 36% of the mle gene of O. oeni ATCC 23279T was also found to be identical to the corresponding nucleotide sequence of mleA (Groisillier and Lonvaud-Funel, 1999). The homology

360


Fig. 2. Sequence alignment of the deduced amino acid sequences of malolactic enzymes produced by Lactococcus lactis MG 1363 (MleSp), Oenococcus oeni Lal1 (MleAp) and Pediococcus damnosus NCFB 1832 (MleDp). Identical regions are enclosed in boxes.


Fig. 3. Degradation of l-malate (open symbols) and the production of l-lactate (closed symbols) in synthetic grape must fermented by recombinant strains of Saccharomyces cerevisiae co-expressing mae1 and the malolactic genes mleD, mleA and mleS, respectively. Controls were strains transformed with plasmids pHVX2 or YEP352-PGK1pt (i.e. without mae1 or mle genes). All data represent an average of three repeats. The values recorded in each experiment did not vary by more than 5%. Single data points are, therefore, presented in the figures without standard deviation bars.

361

lization of wine. Even with the use of bacterial starter cultures, stuck or sluggish MLF often cause delays in cellar operations. Furthermore, reduced sulfite levels in wine may lead to proliferation of spoilage organisms, producing off-odors and biogenic amines (LonvaudFunel, 1999). Malolactic strains of S. cerevisiae that degrade malate efficiently during alcoholic fermentation should prevent problems experienced with bacterial strains. We compared the kinetics of MLF between recombinant yeast strains expressing mleD, mleA and mleS. All three engineered S. cerevisiae strains efficiently converted l-malate to l-lactate. The malic acid conversion rate recorded for mleD and mleA was similar. However, transformants with the mleD gene converted less malic acid towards the end of the fermentation (last 48 h), which is also reflected in the lower lactic acid production at the end of fermentation.

Acknowledgement among the malolactic enzymes of O. oeni suggests that the species consist of a genetically homogeneous collection of strains. Analysis of a fragment representing 36% of the mle genes of 13 strains of lactic acid bacteria suggested that the gene evolved more rapidly than the 16S gene (Groisillier and Lonvaud-Funel, 1999). The 16S rRNA sequence tree revealed three distinct groups of lactic acid bacteria, viz. Lactococcus, Leuc. mesenteroides and O. oeni, and other species. Pediococcus and Lactobacillus spp. were intermixed in the 16S rRNA tree, whereas they were separated in the mle sequence phylogenetic tree. Although Leuc. mesenteroides and O. oeni were distinct from other species in the 16S rRNA tree, they were intermixed with Lactobacillus species and L. lactis in the mle tree. The mle of O. oeni is the closest related to that of L. lactis and the genus Pediococcus, with P. acidilactici and P. parvulus, was separated from the other LAB. After now having sequenced the mle gene of P. damnosus in its entirety, complete sequences are available for malolactic genes representing each of the three groups of LAB. Based on our results, the MLE of O. oeni is closeser related to P. damnosus. A complete and rapid malolactic conversion by S. cerevisiae can be achieved by co-expressing the P. damnosus mleD and S. pombe mae1 genes in this yeast. The removal of l-malate, one of the main organic acids of grape must, is essential for deacidification and stabi-

This research was funded by a grant from the National Research Foundation, South Africa.

References Ausubel, F.M., Brent, R., Kingston, R.E., Moore, D.D., Seidman, J.G., Smith, J.A., Struhl, K., 1994. Current Protocols in Molecular Biology. John Wiley and Sons, Inc., New York. Caspritz, G., Radler, F., 1983. Malolactic enzyme of Lactobacillus plantarum. Purification, properties and distribution among bacteria. J. Biol. Chem. 258, 4907–4910. Daniel, P., 1995. Sizing of the Lactobacillus plantarum genome and other lactic acid bacteria species by traverse alternating field electrophoresis. Curr. Microbiol. 30, 243–246. Dellaglio, F., Bottazzi, V., Trovatelli, L.D., 1973. Deoxyribonucleic acid homology and base composition in some thermophilic lactobacilli. J. Gen. Microbiol. 74, 289–297. Denayrolles, M., Aigle, M., Lonvaud-Funel, A., 1994. Cloning and sequence analysis of the gene encoding Lactococcus lactis malolactic enzyme: relationship with malic enzymes. FEMS Microbiol. Lett. 116, 79–86. Denayrolles, M., Aigle, M., Lonvaud-Funel, A., 1995. Functional expression in Saccharomyces cerevisiae of the Lactococcus lactis mleS gene encoding the malolactic enzyme. FEMS Microbiol. Lett. 125, 37–44. Dicks, L.M.T., Dellaglio, F., Colins, M.D., 1995a. Proposal to reclassify Leuconostoc oenos as Oenococcus oeni (corig.) gen. gov., comb. nov. Int. J. Syst. Bacteriol. 45, 395–397. Dicks, L.M.T., Van Vuuren, H.J.J., Dellaglio, F., 1995b. Taxonomy of Leuconostoc species particularly Leucocnostoc oenos,

362


as revealed by numerical analysis of total cell protein patterns, DNA base compositions, and DNA-DNA hybridisations. Int. J. Syst. Bacteriol. 40, 83–91. ¨ Fuck, E., Stärk, G., Radler, F., 1973. Apfels¨ auere-Stoffwechsel bei Saccharomyces. II. Anreicherung und Eigenschaften eines Malatenzymes. Arch. Mikrobiol. 89, 223–231. Green, G., Dicks, L.M.T., Bruggeman, G., Vandamme, E.J., Chikindas, M.L., 1997. Pediocin PD-1, a bactericidal antimicrobial peptide from Pediococcus damnosus NCFB1832. J. Appl. Microbiol. 83, 127–213. Grobler, J., Bauer, F., Subden, R.E., Van Vuuren, H.J.J., 1995. The mae1 gene of Schizosaccharomyces pombe encodes a permease for malate and other C4 dicarboxylic acids. Yeast 11, 1485–1491. Groisillier, A., Lonvaud-Funel, A., 1999. Comparison of partial malolactic enzyme gene sequences for phylogenetic analysis of some lactic acid bacteria species and relationships with the malic enzyme. Int. J. Syst. Bacteriol. 49, 1417–1428. Hanahan, D., 1983. Studies on transformation of Escherichia coli with plasmids. J. Mol. Biol. 166, 557–580. Henick-Kling, T., 1995. Control of malolactic fermentations in wine: energetics, flavour modification and methods of starter culture preparation. J. Appl. Bacteriol. Sympos. Suppl. 79, 29–37. Hoffman, C.S., Winston, F., 1987. A ten-minute DNA preparation from yeast efficiently releases autonomous plasmids for transformation of Escherichia coli. Gene 57, 267–272. Kunkee, R.E., 1997. A heady concoction of alcoholic and malolactic fermentations. Nat. Biotech. 15, 224–225. Labarre, C., Guzzo, J., Cavin, J.F., Davies, C., 1996. Cloning and characterization of the genes encoding the malolactic enzyme and malate permease of Leuconostoc oenos. Appl. Environ. Microbiol. 62, 1274–1282. Lee, S., Rasheed, S., 1991. A simple procedure for maximum yield of high quality plasmid DNA. Biotechniques 9, 676–679. Lonvaud-Funel, A., 1999. Lactic acid bacteria in the quality improvement and depreciation of wine. Antonie Van Leeuwenhoek 76, 317–331. Martineau, B., Acree, T.E., Henick-Kling, T., 1995. Effect of wine type on threshold for diacetyl. Food Res. Int. 28, 139–143.

Naouri, P., Chagnaud, P., Arnaud, A., Galzy, P., 1990. Purification and properties of a malolactic enzyme from Leuconostoc oenos ATCC 23278. J. Basic Microbiol. 30, 577–585. Osothsilp, C., Subden, R.E., 1986. Isolation and characterization of Schizosaccharomyces pombe mutants with defective NAD-dependant malic enzyme. Can. J. Microbiol. 32, 481– 486. O’Sullivan, D.J., Klaenhammer, T.R., 1993. High- and low-copynumber Lactococcus shuttle cloning vectors with features for clone screening. Gene 137, 227–231. Radler, F., 1986. Microbial biochemistry. Experientia 42, 884–893. Sambrook, J., Fritsch, E.F., Maniatis, T., 1989. Molecular Cloning: A Laboratory Manual, second ed. Cold Spring Harbour Laboratory Press, New York. Sikorski, R.S., Hieter, P., 1989. A system of shuttle vectors and yeast host strains designed for efficient manipulation of DNA in Saccharomyces cerevisiae. Genetics 122, 19–27. Spettoli, P., Nuti, M.P., Zamora, A., 1984. Properties of malolactic activity purified from Leucocnostoc oenos ML34 by affinity chromatography. Appl. Environ. Microbiol. 48, 900–903. Van Vuuren, H.J.J., Viljoen, M., Grobler, J., Volschenk, H., Bauer, F., Subden, R.E., 1995. Genetic analysis of the Schizosaccharomyces pombe malate permease (mae1) and malic enzyme (mae2) genes and their expression in Saccharomyces cerevisiae. In: Gousard, P.G. (Ed.), Proceedings of the first SASEV International Congress. Cape Town, South Africa, pp. 6–8. Viljoen, M., Subden, R.E., Krizus, A., Van Vuuren, H.J.J., 1994. Molecular analysis of the malic enzyme gene (mae2) of Schizosaccharomyces pombe. Yeast 10, 613–624. Volschenk, H., Viljoen, M., Grobler, J., Bauer, F.F., Lonvaud-Funel, A., Denayrolles, M., Subden, R.E., Van Vuuren, H.J.J., 1997a. Malolactic fermentation in grape musts by a genetically engineered strain of Saccharomyces cerevisiae. Am. J. Enol. Vitic. 48, 193–197. Volschenk, H., Viljoen, M., Grobler, J., Petzold, B., Bauer, F., Subden, R.E., Young, R.A., Lonvaud-Funel, A., Denayrolles, M., Van Vuuren, H.J.J., 1997b. Engineering pathways for malate degradation in Saccharomyces cerevisiae. Nature Biotech. 15, 253–257.


Genetically engineered horseradish peroxidase for facilitated purification from baculovirus cultures by cation-exchange chromatography Gustavo Levin a , Fernando Mendive b , Héctor M. Targovnik b , Osvaldo Cascone a,∗ , Mar´ıa V. Miranda a a

Cátedra de Microbiolog´ıa Industrial y Biotecnolog´ıa, Facultad de Farmacia y Bioqu´ımica (UBA), Jun´ın 956, 1113 Buenos Aires, Argentina b C´ atedra de Genética y Biolog´ıa Molecular, Facultad de Farmacia y Bioqu´ımica (UBA), Jun´ın 956, 1113 Buenos Aires, Argentina Received 14 December 2004; accepted 4 May 2005

Abstract An engineered horseradish peroxidase isozyme C (HRP C) gene was constructed by the addition of a 6xArg fusion tail to 6xHis–HRP C by the PCR strategy. The 6xHis–6xArg–HRP C cDNA was expressed in the Sf9 insect cell line from Spodoptera frugiperda infected with Autographa californica nuclear polyhedrosis virus. The recombinant peroxidase isoelectric point was 9.5 as judged by isoelectric focusing and was purified directly from the culture medium at day-6 post-infection by cation-exchange chromatography or immobilised metal ion-affinity chromatography. While the former technique gave a yield of 98.5% with a purification factor of 130, the latter gave only a 68% yield with a purification factor of 140. Results obtained provide evidence that the poly-Arg tag is more effective than the poly-His tag for peroxidase purification from a baculovirus expression system. © 2005 Elsevier B.V. All rights reserved. Keywords: Peroxidase; Purification; Fusion tail; Ion-exchange chromatography; Immobilised metal ion-affinity chromatography

1. Introduction

∗ Corresponding author. Tel.: +54 11 4901 6284; fax: +54 11 4901 6284. E-mail address: [email protected] (O. Cascone).


Horseradish peroxidase (HRP, EC 1.11.1.7) catalyses the oxidation of a broad variety of substrates by hydrogen peroxide. HRP isozyme C (HRP C) is the archetypal enzyme for the biochemical study of

364

G. Levin et al. / Journal of Biotechnology 118 (2005) 363–369

peroxidases and is probably, the most extensively studied member of the plant peroxidase superfamily. It consists of 308 amino acid residues, a ferric heme prosthetic group, two calcium ions per molecule and is glycosylated at eight sites of asparagine-linked glycans (Welinder, 1979; Smith et al., 1990; Dunford, 1991; Gray et al., 1998). HRP C complexity made difficult its expression in prokaryotic systems. On the other hand, its expression in Saccharomyces cerevisiae and Pichia pastoris rendered low yields (0.6 mg l−1 ) (Morawski et al., 2000). Baculovirus expression vectors are widely used to produce high levels of recombinant proteins during infection of insect larvae or established insect cell lines. One of the most appealing features of baculovirus–insect expression systems is the eukaryotic protein processing capability of the host. In a previous work, we expressed recombinant HRP C with a 6xHis tag (6xHis–HRP C) with high yield (41.3 mg l−1 ) in a baculovirus–insect cell system and purified it by immobilised metal ionaffinity chromatography (IMAC) (Segura et al., 2005). Probably, the 6xHis tail is the most commonly used tail to selectively purify recombinant proteins. However, at an industrial scale, this strategy becomes very expensive because of the matrix cost. Therefore, the growing demand for less expensive purification systems pushes the need to overcome the major drawbacks associated with this fusion tail, mainly the high purification costs and the difficulties in largescale processes (Kweon et al., 2002). Other fusion tags specifically designed to facilitate protein purification include FLAG peptide, streptavidin, polyaspartic acid, polyarginine, glutathione-S-transferase, polyphenylalanine, etc. (La Vallie and McCoy, 1995). Charged fusions are well-known by their small size and the availability of a variety of inexpensive separation methods based on charge. Brewer and Sassenfeld (1985) engineered the fusion of 6xArg tails to several proteins expressed in Escherichia coli, such as ␤-urogastrone and bacterial aspartate aminotransferase that were readily isolated from cell lysates by ionexchange chromatography. This purification concept based on polyarginine tail was assayed by Zhang and Glatz (1999) in the canola system but, in this case, proteases degraded the fusion tail. The aim of this work was to construct a fusion HRP C by adding a 6xArg tail to HRP C containing the 6xHis tail in order to increase its isoelectric point and there-

fore promote selective binding on cation exchangers directly from a baculovirus–insect cell culture medium. In addition, HRP C with both tails allowed us to compare the performance of IMAC and IEC for purification of said recombinant HRP C.

2. Materials and methods 2.1. Materials Horseradish peroxidase (type VI), hemin and 3,3 -diaminobenzidine (DAB reagent) were from Sigma–Aldrich (St. Louis, MO, USA). Grace’s insect tissue culture media and penicillin/streptomycin (ATB/ATM) were from Gibco–BRL (Gaithersburg, MD, USA). Fetal calf serum (FCS) was from Nutrientes Naturales S.A. (Buenos Aires, Argentina). E. coli strain DH5␣ was from Facultad de Farmacia y Bioqu´ımica (Buenos Aires, Argentina). Spodoptera frugiperda Sf9 cells were obtained from ABAC (Buenos Aires, Argentina). Ni(II)–NTA HiTrap prepacked columns and SP-Sepharose FF High Trap were from Amersham Biosciences (Uppsala, Sweden). 2.2. Strategy for fusion protein construction An engineered HRP C gene was constructed by addition of a poly-Arg fusion tail to 6xHis–HRP C gene by the PCR strategy. The template was plasmid pAcGP67HRP containing the gene synthesised by British Biotechnologies Ltd., generously provided by Dr. P.E. Ortiz de Montellano of the University of California. The synthetic gene does not include the endogenous 5 leader sequence found in the plant. The HRP C gene including a 6xHis tag at the 5 extreme was oriented, so that it could be expressed using the baculovirus polyhedrin promoter. The vector (pAcGP67B, Pharmingen, San Diego, CA, USA) encodes a sequence for the glycoprotein 67 leader peptide at the 5 end of the multiple cloning site which targets the protein for secretion. A 72 bp primer encoding the C-terminal region of HRP C was specially designed. This primer included five original codons of the native gene, a thrombin site and a tag of six arginine residues (AGG AGG AGA AGA AGG AGA). The 6xArg–HRP C amplified fragment was purified and cloned in the pGEM–T Easy


vector. JM109 competent cells were used for transformations and different clones were sequenced. The correct product and plasmid pAcGP67HRP were digested with SacI and EcoRI restriction endonucleases. The fragments were isolated on a 1% agarose gel and then recovered by column purification and ligated to obtain pAcGPHisArgHRP.

365

2.5. Cell culture S. frugiperda Sf9 cell cultures were conserved in monolayers in T-flasks at 27 ◦ C in a Grace’s medium containing 10% heat-inactivated fetal calf serum and routinely subcultured every 2–3 days. Cells were counted with a haemocytometer and cell viability was assessed by Trypan Blue staining.

2.3. PCR conditions The HRP C gene was 1.0 kb double-stranded DNA. PCR products amplified from plasmid pAcGP67HRP with the primer sequences AGT ATG GAT CCA TGC AGT TAA CCC CTA CAT T (forward) and CCG AAT TCA TCG CCG ACG TCG TCT CCT TGA TCC ACG GGG AAC CAG AGA GTT GCT GTT GAC CAC TCT GCA GTT (reverse). PCR conditions (100 ␮l final volume): 0.8 ␮M each primer, 1× Taq buffer, 0.4 mM each dNTP and 2 U Taq polymerase. PCR program: 95 ◦ C for 6 min and 40 cycles, 95 ◦ C for 30 s, 52 ◦ C for 30 s, 72 ◦ C for 1 min. An additional extension step of 10 min at 72 ◦ C was then applied. Free primers from the PCR product were removed by using the ConcertTM PCR purification system (Gibco, BRL, Gaithersburg, MD, USA). DNA concentration was 0.4 ␮g ␮l−1 . 2.4. Recombinant baculovirus construction and amplification Sf9 cells were co-transfected with 2 ␮g pAcGPHisTM ArgHRP and 0.5 ␮g wild linearised BaculoGold DNA (Pharmingen) in the presence of calcium phosphate according to Pharmingen (Gruenwald and Heitz, 1993). After a 5-day incubation period at 27 ◦ C, the cell culture supernatant was collected and centrifuged at 2500 × g for 5 min. Co-transfection efficiency was determined by the end-point dilution assay according to the supplier’s instructions (Pharmingen). After a round of plaque purification, a recombinant plaque was isolated and amplified to yield a high-titre virus stock. The purified virus was used to infect 1 × 107 Sf9 cells in monolayer in 15 ml Grace’s medium at a multiplicity of infection (MOI) of 0.1. Following two amplification steps, virus titre was determined by a plaque assay. Typically, the titre of this stock was 4.7 × 107 to 1 × 108 pfu ml−1 . This amplified virus stock was used at the production step.

2.6. Expression of 6xHis–6xArg–HRP C in Sf9 cells Infection was performed at a MOI 2 and, at the same time, heme was added at a 2.4 ␮M final concentration as was optimised as in a previous work (Segura et al., 2005). A negative control without virus was also performed. Plaques were incubated at 27 ◦ C under light protection and, after 6 days post-infection, the culture medium was harvested, cells were separated by centrifugation (1000 × g, 10 min) and peroxidase activity was measured in the supernatant. 2.7. Purification of 6xHis–6xArg–HRP C by cation-exchange chromatography The clarified supernatant brought to pH 8.5 was applied directly on a SP-Sepharose FF HiTrap column (bed volume 1.0 ml), equilibrated with 5 mM Tris–HCl buffer, pH 8.5. After a washing step with equilibration buffer, 1 ml fractions were collected at a linear flow rate of 0.4 cm min−1 and monitored by their absorbance at 280 nm and enzyme activity. Elution was performed with 1 M NaCl in the equilibration buffer. 2.8. Purification of 6xHis–6xArg–HRP C by IMAC The culture supernatant was diafiltered to change the buffer to 25 mM sodium phosphate, 300 mM NaCl, pH 8.0. A Ni(II)–NTA HiTrap column (bed volume 1.0 ml) equilibrated with the same buffer was loaded with 3 ml of the conditioned sample. Following a washing step with the same buffer containing 20 mM imidazole, the elution of the enzyme was performed by raising the imidazole concentration to 500 mM. Flow rate was 0.4 cm min−1 , and 1 ml fractions were collected and monitored by their absorbance at 280 nm and enzyme activity.

366


2.9. Total protein determination Protein concentration was determined by Bradford assay (1976) using bovine serum albumin as the standard. 2.10. Electrophoretic analysis SDS-PAGE 12.5% and three to nine range isoelectric focusing (IEF) analyses were carried out with a Phast System Equipment (Amersham Biosciences, Uppsala, Sweden). Gel staining was accomplished using the Coomassie Blue method or employing a selective staining with the DAB reagent to detect active HRP (Segura et al., 2005). 2.11. Enzyme activity measurement A reaction mixture containing 105 ␮M guaiacol and 250 ␮M hydrogen peroxide in 100 mM potassium phosphate buffer, pH 7.0, was prepared. Guaiacol oxidation was initiated by addition of 10 ␮l sample to 1 ml reaction mixture. The reaction was monitored by measuring its absorbance at 470 nm within 1 min. Activity calculations were performed as per Tjissen (1985). For enzyme mass quantitation, a specific activity of pure enzyme of 592.3 U mg−1 was assumed.

3. Results and discussion Recombinant DNA technology was utilised to produce horseradish peroxidase with a double tail. The PCR product was cloned in an appropriate transfer vector (pAcGP67B) under the control of the strong viral polyhedrin promoter. Fig. 1 shows the construction design of the protein. The tail 6xArg is present at the C-terminus and the tail 6xHis at the N-terminus of the original template. This type of genetic construction allows comparison of the performance of purification strategies based on IMAC or ion-exchange chromatography (IEC). As with 6xHis–HRP C (Segura et al., 2005), Sf9 cells showed a good performance in expressing recombinant 6xHis–6xArg–HRP C. At the day-6 postinfection, culture supernatant was collected and IEF analysis with DAB stain revealed a single discrete band showing peroxidase activity with pI 9.5 thus providing

Fig. 1. Construction design of the 6xHis–6xArg–HRP C gene. The PCR product was cloned into an intermediate vector pGEMT easy. The SacI–EcoRI restrict containing the gene of 6xArg–HRP C was then cloned into the transfer vector pAcGPHRP6xHis.

evidence on the integrity of the fusion tag as well as on the increased positive charge of the recombinant protein. Fig. 2 shows the isoelectric point increase in comparison with the 6xHis–HRP C molecule expressed in the same insect cell line. As judged by SDSPAGE, the molecular mass of HRP C with both tags is very close to that of the native enzyme (data not shown). The kinetics of 6xHis–6xArg–HRP C expression by Sf9 cells infected at different MOIs was very similar to that previously described for 6xHis–HRP C (Segura et al., 2005). Enzyme activity in the harvested supernatant increased continuously until day 7, but day-6 post-infection was chosen as the best


367

Fig. 3. Purification of 6xHis–6xArg–HRP C by immobilised metal ion-affinity chromatography. After a buffer shift to 25 mM sodium phosphate, 300 mM NaCl, pH 8.0, 3 ml of conditioned culture supernatant was loaded on a Ni(II)–NTA HiTrap column equilibrated with the same buffer. A washing step with the same buffer containing 20 mM imidazole was then performed and elution of the enzyme was carried out by raising the imidazole concentration to 500 mM; 1 ml fractions were collected at a linear flow rate of 0.4 cm min−1 and monitored by their absorbance at 280 nm () and enzyme activity (䊉). Fig. 2. Isoelectric focusing of 6xHis–6xArg–HRP C in comparison with 6xHis–HRP C. The 4 ␮l sample was loaded on a Phast System IEF 3–9 gel, and following separation, stained with DAB reagent as described in Section 2. Lane 1, HRP C standard; lane 2, 6xHis–6xArg–HRP C; lane 3, 6xHis–HRP C.

for product collection as it ensures a maximum yield of recombinant protein without significant intracellular protein contamination. The maximum level of active 6xHis–6xArg–HRP C produced in an FCSsupplemented Grace’s medium containing 2.4 ␮M hemin was 30.0 mg l−1 . In order to assess the usefulness of the tags through the performance of IMAC and IEC for enzyme purification from a culture supernatant, chromatographic runs were developed with 6xHis–6xArg–HRP C. 3.1. Purification of 6xHis–6xArg–HRP C by IMAC The culture supernatant was conditioned by diafiltration to remove amino acids present in high concentrations in the insect cell culture medium as they compete with 6xHis–6xArg–HRP C for the binding sites on the Ni–NTA matrix. On the other hand, the insect culture medium is acidic (pH 6.0–6.5) and the pH must be raised to 8.0 and NaCl concentration up to 300 mM to promote the protein binding to the Ni–NTA

matrix. Fig. 3 shows the chromatographic profile. This purification scheme yields 68% of active enzyme with a purification factor of 140. 3.2. Purification of 6xHis–6xArg–HRP C by IEC To assess the influence of the 6xArg tail on the chromatographic behaviour of the enzyme, ion-exchange purification was performed with 6xHis–6xArg–HRP C and 6xHis–HRP C. When the culture supernatant brought to pH 8.5 was loaded to an ion-exchange column, 6xHis–HRP C was found in the pass-through. In contrast, 6xHis–6xArg–HRP C was retained and only eluted by addition of 1 M NaCl to the adsorption buffer (Fig. 4). Recovery of 6xHis–6xArg–HRP C was 98.5% and the purification factor of 130. Both purification processes yielded electrophoretically homogeneous peroxidase. Different codons encoding the arginine amino acid are rare in E. coli; therefore, those grouped in clusters should be avoided. It has been found that a long stretch of similar codons decreases the expression level in E. coli. Furthermore, clusters of very rare codons can create translation errors and reduce the expression level. On the other hand, polyarginine fusions have proven to be prone to degradation

368


Acknowledgements This work was supported by grants from the Agencia Nacional de Promoción Cient´ıfica y Tecnológica de la República Argentina and the CONICET. MVM, HMT and OC are career researchers of the CONICET.

References

Fig. 4. Purification of 6xHis–6xArg–HRP C by ion-exchange chromatography. Clarified supernatant brought to pH 8.5 was loaded on an SP-Sepharose FF HiTrap column, equilibrated with 5 mM Tris–HCl buffer, pH 8.5. After a washing step with equilibration buffer, elution was performed with 1 M NaCl in the same buffer; 1 ml fractions were collected at a linear flow rate of 0.4 cm min−1 and monitored by their absorbance at 280 nm () and enzyme activity (䊉).

by E. coli proteases (Skerra et al., 1991; Niederauer et al., 1994). For these reasons, the strategy of the poly-Arg fusion tag for purification was not effective in prokaryotic systems unless a special E. coli strain was used (BL21 (DE3) codon plus, Stratagene). Zhang and Glatz (1999) described a similar method allowing selective recovery of the recombinant T4 lysozyme from canola. Unfortunately, the authors described the presence of proteases that degraded the fusion tag. In this work, we provide evidence that the 6xArg tag is stable in insect cells and allows us to purify recombinant peroxidase with a high yield by IEC in only one step. The enzyme purified by IMAC displays a similar purity degree but the yield is significantly lower. A culture medium supplemented with FCS was characterised by IEF and results indicate that other proteins of interest with a basic isoelectric point must be easily separated from contaminants by IEC.

4. Conclusion The addition of a 6xArg tag to 6xHis–HRP C results in an increased pI of the product expressed in the baculovirus system. This gives the opportunity for facilitated direct peroxidase recovery in high yield and purity employing inexpensive cation exchangers.

Bradford, M., 1976. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein–dye binding. Anal. Biochem. 72, 248– 254. Brewer, S.J., Sassenfeld, J.M., 1985. The purification of recombinant proteins using C-terminal polyarginine fusions. Trends Biotechnol. 3, 119–122. Dunford, H.B., 1991. In: Everse, J., Everse, K.E., Grisham, M.B. (Eds.), Peroxidases in Chemistry and Biology, vol. 2. CRC Press, Boca Raton, pp. 1–24. Gray, J.S., Yun Yang, B., Montgomery, R., 1998. Heterogeneity of glycans at each N-glycosylation site of horseradish peroxidase. Carbohydr. Res. 311, 61–69. Gruenwald, S., Heitz, J., 1993. Generation of recombinant baculovirus by co-transfection. In: Baculovirus Expression Vector System: Procedures and Methods Manual, second ed. PharMingen, San Diego, pp. 48–49. Kweon, D.H., Lee, D.H., Han, N.S., Pha, C.S., Seo, J.H., 2002. Characterization of polycationic amino acids fusion systems for ion-exchange purification of cyclodextrin glycosyltransferase from recombinant Escherichia coli. Biotechnol. Prog. 18, 303– 308. La Vallie, E.R., McCoy, J.M., 1995. Gene fusion expression systems in Escherichia coli. Curr. Opin. Biotechnol. 6, 501– 506. Morawski, B., Lin, Z., Cirino, P., Joo, H., Arnold, F.H., 2000. Functional expression of horseradish peroxidase in Saccharomyces cerevisiae and Pichia pastoris. Protein Eng. 13, 377–384. Niederauer, M.Q., Suominen, I., Rougvie, M.A., Ford, C.F., Glatz, C.E., 1994. Characterization of polyelectrolyte precipitation of beta-galactosidase containing genetic fusions of charged polypeptides. Biotechnol. Prog. 10, 237–245. Segura, M.M., Levin, G.J., Miranda, M.V., Mendive, F.M., Targovnik, H.M., Cascone, O., 2005. High-level expression and purification of recombinant horseradish peroxidase isozyme C in SF-9 insect cell culture. Process Biochem. 40, 795–800. Skerra, A., Pfizinger, I., Pluckthun, A., 1991. The functional expression of antibody Fv fragments in Escherichia coli: improved vectors and a generally applicable purification technique. Bio/Technology 9, 273–278. Smith, A.T., Santama, N., Dacey, S., Edwards, M., Bray, R.C., Thorneley, R.N.F., Burke, J.F., 1990. Expression of a synthetic gene for horseradish peroxidase C in Escherichia coli and folding and activation of the recombinant enzyme with Ca2+ and heme. J. Biol. Chem. 265, 13335–13343.

G. Levin et al. / Journal of Biotechnology 118 (2005) 363–369 Tjissen, P., 1985. In: Burdon, R.H., van Knippenberg, P.H. (Eds.), Practice and Theory of Enzyme Immunoassays. Elsevier, NY, p. 173. Welinder, K.G., 1979. Amino acid sequence studies of horseradish peroxidase: amino and carboxyl termini, cyanogen bromide and tryptic fragments, the complete sequence, and some structural

369

characteristics of horseradish peroxidase C. Eur. J. Biochem. 96, 483–502. Zhang, C., Glatz, C.E., 1999. Process engineering strategy for recombinant protein recovery from canola by cation-exchange chromatography. Biotechnol. Prog. 15, 12–18.


Development and characterization of a monoclonal antibody directed against human telomerase reverse transcriptase (hTERT) Dario Soldateschi a , Sara Bravaccini b , Brunilde Berti a , Alessandra Brogi a , Tiziana Benicchi a , Claudia Soldatini a , Laura Medri c , Francesco Fabbri c , Franca De Paola c , Dino Amadori c , Daniele Calistri c,∗ a

Diesse Diagnostica Senese, Siena, Italy Istituto Oncologico Romagnolo, Forl`ı, Italy Division of Oncology and Diagnostics, Pierantoni Hospital, 47100 Forl`ı, Italy b

c

Received 2 December 2004; received in revised form 29 April 2005; accepted 4 May 2005

Abstract Telomerase activity plays an important role in the two complementary processes of cellular immortalization and senescence. This enzyme is active in almost all tumors, but also in inflammatory and many normal proliferating cells. Therefore, the main limits of molecular determinations, such as telomeric repeat amplification protocol assay is that they are not able to discriminate between the enzymatic activity of tumor and normal cells. The most appropriate technique for this would be immunohistochemical determination using monoclonal antibodies. Very few monoclonal antibodies (Mabs) directed against the human telomerase reverse transcriptase (hTERT) are commercially available and in the present study, we developed a new Mab directed against this protein (TERT-3 36-10) to investigate the possibility of detecting immunoreactivity to this Mab by immunohistochemical and flow cytometric approaches. Immunohistochemical determination showed a lack of reactivity to the Mab in highly differentiated striated muscle tissue, a variable reactivity in dysplastic cervical epithelial tissue and similar and widespread immunoreactivity in cell lines and clinical tumors. Furthermore, we demonstrated the ability of this Mab to inhibit enzyme activity in cell extract from MCR bladder tumor cell line. © 2005 Elsevier B.V. All rights reserved. Keywords: Telomerase; Monoclonal antibody; Immunohistochemistry; Flow cytometry; Inhibition assay

1. Introduction

∗ Corresponding author. Tel.: +39 0543 731623; fax: +39 0543 731736. E-mail address: [email protected] (D. Calistri).


Telomerase activity plays an important role in cell immortalization and transformation (Meyerson, 2000; Shay et al., 2001). The inability of the replication machinery to re-establish the telomeric sequences of

D. Soldateschi et al. / Journal of Biotechnology 118 (2005) 370–378

the 3 -ends of chromosomes after cell division causes a gradual decrease in DNA length, inducing apoptosis or causing chromosomal instability and consequently limiting cell division of somatic cells (Harley et al., 1990; Shay and Wright, 2000). The ribonucleoenzyme telomerase is capable of synthesizing these chromosomal ends by adding the hexameric repeats (TTAGGG) to maintain telomere length (Morin, 1989). This enzyme consists of two essential components: the RNA component (hTR) (Feng et al., 1995), which acts as a template for telomeric DNA synthesis and the catalytic subunit, human telomerase reverse transcriptase (hTERT) (Meyerson et al., 1997; Nakamura et al., 1997). The two components are not equally expressed and are probably regulated by different mechanisms. hTR is expressed in all tissues even in the absence of telomerase activity (Avilion et al., 1996). Conversely, the mRNA expression of the catalytic component hTERT is estimated at less than five copies per cell (Yi et al., 1999) and seems to be closely associated with telomerase activity. These data strongly suggest that hTERT is probably the rate-limiting component for enzyme activity. Recent studies have shown that telomerase activity, in association with the activation of oncogenes or the inactivation of tumor suppressor genes, is also capable of inducing tumorigenic conversion of normal human epithelial cells and fibroblasts (Hahn et al., 1999). Because of its role in cell immortalization and transformation, telomerase activity represents an important target for the development of new diagnostic tools and for therapeutic strategies. For this reason we aimed to develop a monoclonal antibody (Mab) directed against the essential telomerase subunit, hTERT. When the study was begun, no Mabs were available and to date only a few have been developed or commercialized (Yang et al., 2001; Kyo et al., 2003). In the present paper, we describe the procedure modalities, their application and the results obtained using different normal and tumor tissues. 2. Materials and methods

371

obtained from 1 ␮g of RNA by RT-PCR using the Gene AMP Gold RNA PCR Kit (Applied Biosystems, Foster City, CA) according to the manufacturer’s instructions. Synthetic oligonucleotides were used as primers for amplification of the sequence corresponding to the following amino acids: 1–291 (TERT-1), 144–249 (TERT-2), 227–338 (TERT-3), 325–465 (TERT-4), 458–559 (TERT-5), 540–852 (TERT-6) and 840–1132 (TERT-7). The primers used to amplify TERT-2, TERT3, TERT-4 and TERT-5 fragments were engineered to incorporate recognition sequences for EcoRI and BamHI restriction endonucleases, which permitted cloning in the bacterial expression vectors pGEX-2T (Amersham Biosciences, Milan, Italy), whilst the other primers contained the recognition sequences for EcoRI and HindIII restriction endonucleases, which permitted cloning in both the pET28 and pET32 expression vectors (Novagen, Madison, WI). TERT-1, TERT-3 and TERT-5 fragments were amplified with the GC-Rich PCR System (Roche Molecular Biochemicals, Milan, Italy) and TERT-2, TERT-4, TERT-6 and TERT-7 fragments were amplified with the Platinum Taq DNA Polymerase High Fidelity (Invitrogen, Milan, Italy). TERT-2, TERT-3, TERT-4 and TERT-5 fragments were cloned in the expression vector pGEX-2T in frame with the sequence coding for the glutathioneS-transferase (GST), originating plasmids pDST83/3, pDST78/2, pDST84/1 and pALT6/4, respectively. TERT-6 and TERT-7 fragments were cloned in the expression vector pET28 fused to the sequence coding for a 6 histidine tag (His6 -Tag), originating plasmids pABT62/1 and pALT4/3. TERT-1 fragment was cloned in the expression vector pET32 fused to the sequence coding for thioredoxin (TRX) and a His6 -Tag, originating plasmid pTBT5/1. TERT-4 and TERT-6 fragments were also cloned in the expression vector pET32 fused to the sequence coding for thioredoxin and the His6 Tag, originating plasmids pALT8/9 and pTBT11/3. The hTERT fragments were expressed by induction with 1 mM IPTG (Inalco, Milan, Italy) at 37 ◦ C for 1 h (TERT-1), 25 ◦ C for 4 h (TERT-2), 28 ◦ C for 4 h (TERT3), 37 ◦ C for 3 h (TERT-4, TERT-5) and 37 ◦ C for 2 h (TERT-6, TERT-7).

2.1. Cloning and expression of hTERT fragments 2.2. Preparation of inclusion bodies Total RNA was extracted from the LRWZ colorectal cancer cell line, established in our laboratory, with the Rneasy Mini Kit (Qiagen, Milan, Italy) and cDNA was

Bacteria from induced cultures were centrifuged and re-suspended to 1/40 of the original culture volume

372


in STE (0.1 M NaCl, 10 mM Tris–HCl, 1 mM EDTA pH 8.0) containing protease inhibitors (Roche Molecular Biochemicals). The bacterial suspension was lysed by sonication and the inclusion bodies of the TERT-7 fragment were purified by immobilized metal affinity chromatography (IMAC) on a HiTrap Chelating HP column (Amersham Biosciences) after solubilization with 6 M of guanidine hydrochloride, whilst the other fragments (TERT-4, TERT-5 and TERT-6) were used to immunize animals without any further purification. 2.3. Purification of hTERT fragments Bacteria from induced cultures were centrifuged and re-suspended in 1/40 of the original culture volume in STE for GST-fusion fragments and in lysis buffer (NaH2 PO4 0.02 mM, NaCl 1 M pH 7.2) for His6 -Tag fusion fragments. After the addition of protease inhibitors (Roche Molecular Biochemicals), the bacterial suspension was lysed by sonication and the soluble His6 -Tag fusion fragment (TERT-1) was purified with affinity columns HiTrap Chelating HP (Amersham Biosciences) whilst the soluble GST-fusion fragments (TERT-2, TERT-3) were purified with glutathione Sepharose 4 fast flow (Amersham Biosciences), according to the manufacturer’s instructions. 2.4. In vitro hTERT expression The open reading frame of the complete hTERT cDNA was cloned in the EcoRV site of pcDNA 3.1 (Invitrogen), originating the pcDNA3.1-hTERT plasmid. Cos-1 cells (American Type Culture Collection, Manassas, VA), seeded into 75 cm2 flasks 12–15 h before use at a density of 35,000/cm2 , were transiently transfected with this construct using Lipofectamine 2000 reagent (Invitrogen), according to the manufacturer’s instructions. Cells were harvested after 2 days, washed with phosphate-buffered saline containing protease inhibitors (Roche Molecular Biochemicals), plated onto microscope slides and fixed with ice-cold acetone for subsequent use in immunofluorescence (IF). hTERT expression was verified by IF with a polyclonal antibody directed against hTERT (Santa Cruz Biotechnology, Santa Cruz, CA).

2.5. Animal immunization and hybridoma cell generation Eight-week-old female Balb/c mice and 12week-old LEWIS/Crl Ico rats (Charles River) were immunized by intraperitoneal injection with 20–50 ␮g of the recombinant fragments, repeated eight to nine times at 2–3 week intervals. The recombinant fragment was mixed with complete Freund’s adjuvant (Sigma) for the first inoculation and with incomplete Freund’s adjuvant for the second one, whilst the others were performed without adjuvant. Before splenectomy, the animals were bled and serum was tested for the presence of antibody against hTERT by immunofluorescence with hTERT-expressing transfected COS-1 cells. Three days after the last intraperitoneal injection, splenocytes from immunized animals were fused at a 5:1 ratio with the myeloma cell line SP2/O-Ag14 using polyethylene glycol (PEG) 4000 (Serva, Heidelberg, Germany) as described by Kohler and Milstein (1975). The cells were suspended in RPMI 1640 medium (Invitrogen) containing 10 % FBS (Biochrom, Berlin, Germany), 2 mM glutamine (Invitrogen) supplemented with hypoxantine, aminopterine and thymidine (HAT) (Invitrogen) and then distributed in five 96-well culture plates. After 2 weeks’ cell culture the supernatants of the hybrid cell colonies were screened for antibodies by immunofluorescence on transfected Cos-1 cells and the HAT supplement was replaced by HT (Invitrogen) in the selection medium for hybridomas. Positive cell colonies were then subcloned in three 96-well culture plates at a density of 1 cell/well. 2.6. Preparation of Mabs Established clones were grown in 24-well plates in RPMI 1640 containing 10% FCS and 2 mM glutamine. Groups of three or four nude BALB/cByJIco-nu/nu (Charles River) mice were primed with an intraperitoneal injection of 0.5 ml of tetramethylpentadecane (Sigma) (Noeman et al., 1982). Two weeks after priming, each animal was injected with 3 × 106 hybridoma cells. Ascitic fluid was collected 10 days after injection. The IgG were purified by affinity chromatography with the protein G Sepharose 4 fast flow (Amersham Biosciences), according to the manufacturer’s instructions.


2.7. Immunofluorescence assay Cos-1 cells transfected with the pcDNA3.1-hTERT plasmid on microscope slides were fixed in ice-cold acetone and incubated with PBS containing 3% BSA for 1 h at room temperature. The hybridoma supernatants (30 ␮l) were added and the slides were incubated for 1 h at room temperature, after which the cells were stained with goat anti-rat IgG FITC conjugate secondary antibody (Sigma) and analyzed by fluorescence microscopy. The controls included immune and non-immune rat sera and non-specific hybridoma supernatants. 2.8. Western blotting analysis Cos-1 cells transfected with pcDNA3.1-hTERT or pcDNA3.1 (control) plasmids were washed once with PBS containing protease inhibitors (Roche Molecular Biochemical, Milan, Italy) and then re-suspended at a concentration of 107 cells/ml in Laemmli sample buffer (Laemmli, 1970). Samples were boiled for 5 min and 15 ␮l of the solution, corresponding to 1.5 × 104 cells, was loaded onto 7% polyacrylamide gel (National Diagnostics Ltd., Hessle, East Yorkshire, UK). After electrophoresis, the gel was transferred to nitrocellulose membrane (Schleicher and Schuell, Milan, Italy). The membrane was blocked for 1 h at room temperature with 5% non-fat dried milk in TBS (20 mM Tris–HCl, pH 7.5; 500 mM NaCl), washed with TTBS (TBS containing 0.05% Tween 20) and incubated for 1 h at room temperature with 30 ␮g/ml of MAb TERT3 36-10 in TTBS containing 5% non-fat dried milk (incubation buffer). After three washes with TTBS, the nitrocellulose membrane was incubated with alkaline phosphatase-conjugated goat anti-rat IgG antiserum (Sigma) in incubation buffer and then washed with TTBS. Recognized bands were detected by incubating the membrane with BCIP/NBT (Sigma) as substrate. 2.9. Immunohistochemical assay Commercial MCF-7 estrogen receptor-positive breast adenocarcinoma and HL60 myeloblastic leukemia cell lines were used. Cells were maintained as a monolayer at 37 ◦ C and subcultured weekly in DMEM/HAM’S F-12 (1:1) medium supplemented with fetal calf serum (FCS) (10%), l-glutamine (2 mM) and insulin (10 ␮g/ml).

373

The exponential growth phase was used for all experiments. Cells were detached and centrifuged at 1200 rpm for 10 min. The medium was removed and the cell pellet was suspended in human plasma. After the addition of thromboplastin (BioMerieux, Marcy l’Etoile, France), test tubes were agitated at 37 ◦ C until an agglomerate was obtained. Agglomerates were fixed in 10% neutral buffered formalin for 6 h and embedded in paraffin. Cell lines and histological paraffin blocks of normal striated muscle tissue, dysplastic cervical epithelial tissue and bladder tumors were cut into 4-␮m sections, mounted on positive-charged slides (BioOptica, Milan, Italy), deparaffinized with xylene and rehydrated. Endogenous peroxidase activity was blocked by 3% hydrogen peroxide solution for 10 min. hTERT antigen retrieval was performed by heating the slides in a water bath at 98.5 ◦ C for 45 min in 10 mM of citrate buffer (pH 6.0), followed by cooling at room temperature for at least 20 min. The sections were then treated for non-specific binding with 1% bovine serum albumin in PBS for 20 min, after which they were incubated for 1 h at room temperature with the monoclonal antibodies diluted with background reducing components (Dako Corporation, Carpinteria, CA). The sections were washed with PBS–Tween (0.05%), incubated with biotinylated anti-rat secondary antibody (Dako Corporation), rinsed again with PBS–Tween (0.05%) and incubated with streptavidin–peroxidase conjugate (Dako Corporation). The antibody binding was detected by staining with diaminobenzidine/hydrogen peroxidase chromogen solution (DAB+, liquid substrate–chromogen solution, Dako Corporation). Cell nuclei were counterstained blue by Mayer’s Hemalum. 2.10. Flow cytometry assay MCF7 cells were trypsinized and washed with PBS by centrifugation, re-suspended in 70% cold ethanol and fixed overnight at 4 ◦ C. Fixed cells were washed in PBS 1×, incubated 10 min at 25 ◦ C in solution A (PBS 1×, BSA 2%, Tween 20 0.1%) and then incubated with TERT-3-FITC Ab (Harlow and Lane, 1999) at different dilutions for 1 h at 4 ◦ C in the dark. MCF7 cells incubated with a FITC-conjugated (Dako Corporation) rat isotypic antibody were used as negative control.

374


Cells were then washed in PBS 1× and incubated with a DNA staining solution (PBS 1×, propidium iodide 10 ␮g/ml and RNase 50 ␮g/ml) for 2 h at 4 ◦ C in the dark, until determination. Flow cytometric analysis was performed with a FACSVantage flow cytometer (Becton Dickinson, San Jose, CA) equipped with a 488 nm argon laser. Data acquisition and analysis were performed using CELLQuest software (Becton Dickinson). About 15,000 events per sample were acquired. 2.11. Inhibition test Cells were isolated and washed with PBS and wash buffer. The cells were suspended in 200 ␮l of ice-cold TRAP lysis buffer (Tris–HCl pH 7.5 10 mM, MgCl2 1 mM, EGTA 1 mM, phenyl methylsulfonyl fluoride 0.1 mM, ␤-mercaptoethanol 5 mM, 3-[(3-cholamidopropyl) dimethylamino]-1-propanesulfonate (CHAPS) 0.5% and glycerol 10%), homogenized and incubated on ice for 1 h. The lysate was centrifuged for 20 min at 10,000 × g at 4 ◦ C. The supernatant was removed, snap frozen and stored at −70 ◦ C. The human bladder cancer cell line, MCR, established in our laboratory and expressing high telomerase activity, was used as positive control. Protein concentrations of each lysate were measured with Bio-Rad protein assay (Bio-Rad, Hercules, CA). Protein extract corresponding to 15,000 cells was mixed with scalar concentrations of Mab (0.02–0.4 mg/ml) in a final volume of 50 ␮l of cell lysis buffer and incubated overnight at 4 ◦ C. A Mab

not directed against hTERT was used as a negative control; 2 ␮l were used to determine the telomerase activity level by TRAP assay. Detection of telomerase activity was performed as described previously (Kim et al., 1994; Wright et al., 1995; Fedriga et al., 2001) using 2 ␮l of protein extract incubated with Mab. To verify the presence of RNAse activity in the buffer used for Mab preparation, we performed an assay on RNA obtained from MCF7 breast cancer cell line. Three dilutions of 25 ng, 250 ng and 1 ␮g of RNA were incubated overnight at 4 ◦ C with 0.4 mg/ml of monoclonal antibody (the highest concentration used in the inhibition test). Corresponding aliquots of RNA were incubated without monoclonal antibody as control. Two microliters of RNA were used for a reverse transcription assay and the cDNA obtained were quantified by real time PCR, amplifying the housekeeping gene GAPDH. 3. Results 3.1. Expression of recombinant hTERT fragments Recombinant fragments showed a somewhat different expression of hTERT but were always sufficient for animal immunization. The peptides showed a wide variability in terms of solubility also when derived from the same fusion protein. In particular, the peptides derived from TERT-2 and TERT-3, expressed as fusion proteins with GST, were soluble, whilst TERT-4 and TERT-5 were insoluble (Table 1).

Table 1 Generation of anti-telomerase Mabs by various hTERT recombinant fragments Fragment

TERT-1 TERT-2 TERT-3 TERT-4 TERT-4 TERT-5 TERT-6 TERT-6 TERT-7 a b c

Plasmid

pTBT5/1 pDST83/3 pDST78/2 pDST84/1 pALT8/9 pALT6/4 pABT62/1 pTBT11/3 pALT4/3

Immunofluorescence. Immunohistochemistry. Not done.

Fusion

TRX GST GST GST TRX GST His6 -Tag TRX His6 -Tag

Solubility

+ + + − + − − + −

Animals

No. of Mabs

Mice

Rats

IFa

IHCb

+ − + − n.d.c − − n.d. −

+ + + − + − − − +

7 (rat) 0 23 (rat) n.d. 0 n.d. n.d. 0 14 (rat)

0 n.d. 3 n.d. n.d n.d. n.d. n.d. 0


375

3.2. Generation of hTERT-reactive Mabs In mice, immunization was obtained only after injection with the TERT-1 and TERT-3 recombinant fragments, as evidenced by immunofluorescence reaction of serum antibodies with hTERT-expressing COS1 cells. In contrast, immunization in rats was obtained after injection with all but the TERT-5 and TERT-6 fragments (Table 1). Mabs were generated using spleen cells from all immunized animals. Out of more than 8500 hybridoma clones screened, 44 showed an immunofluorescence reaction with hTERT-expressing cells. Among these, 7, 23 and 14 Mabs were derived from animals immunized with TERT-1, TERT-3 and TERT-7, respectively (Table 1). 3.3. Western blot analysis Fig. 1. Western blot analysis of Cos-1 cells transfected and not transfected with hTERT gene. Only one band of the expected dimension (lane 1) can be seen in the Cos-1 trasfected cells. Cos-1 cells transfected with plasmid control do not show any aspecific reactivity (lane 2).

Cos-1 cells transfected with the pcDNA3.1-hTERT plasmid were used as positive control in Western blot analysis. This cell line showed the presence of a unique band consistent with the expected molecular weight of 127 kDa (Fig. 1). This band was absent in Cos-1 cells

Fig. 2. Reaction to Mab TERT-3 36-10 by immunohistochemistry: (A) normal striated muscle tissue; (B) cervical epithelial tissue; (C) MCF-7 cell line; (D) bladder tumor tissue (400×).

376


tranfected with plasmid control, confirming the ability of our antibody to react specifically with the hTERT protein.

No specific immunoreactivity was observed in MCF-7 cell line or tissue samples for any of the 12 TERT-7-derived Mabs tested and for any of the 6 TERT1-derived Mabs tested.

3.4. Immunohistochemistry 3.5. Flow cytometry Twenty-two of the 23 immunofluorescent reactive antibodies derived from the TERT-3 fragment were tested more exhaustively by immunohistochemistry in cell lines and tissue samples. Only three of these Mabs showed predominantly nuclear immunoreactivity. We chose the TERT-3 36-10 antibody, which showed the best signal localization in the absence of background for all subsequent determinations. A negative reaction to the antibody in highly differentiated striated muscle tissue, a variability in dysplastic cervical epithelium, with a constant positivity in the basal layer and similar and widespread immunoreactivity in tumor cell lines and clinical tumors were observed (Fig. 2A–D).

The potential of this assay to evaluate Mab immunoreactivity was tested on MCF-7 cell line. The 96% of immunoreactive cells were observed using a 15-␮g/ml concentration of TERT-3 36-10 fluorescence Mab (Fig. 3). 3.6. Telomerase activity inhibition The ability of TERT-3 36-10 Mab to inhibit telomerase activity was analyzed in protein extract from MCR cell line, which highly expresses this enzyme. The test was performed in triplicate at Mab

Fig. 3. Reaction to Mab TERT-3 36-10 by flow cytometry. Immunofluorescence plots of cells stained with isotypic control and TERT-3 36-10 Mab (A and B). Fluorescence histogram of isotypic control (grey line) and TERT-3 36-10 antibody (black line) (C).


Fig. 4. Telomerase inhibition assay with Mab TERT-3 36-10.

concentrations ranging from 0.02 to 0.4 mg/ml. A 50% inhibition of the enzymatic activity had already been obtained at a 0.08 mg/ml concentration and an 80% reduction was reached at a concentration of 0.4 mg/ml (Fig. 4). Real time analysis to verify the presence of RNAse in the buffers showed a similar RNA quantity in samples incubated with or without Mab hTERT, confirming the absence of aspecific RNA degradation (data not shown).

4. Discussion The role of telomerase enzyme in cell destiny is well known. By maintaining telomere length, it bestows cells with immortality (Kim et al., 1994; Shay and Bacchetti, 1997), whereas its inhibition, which leads to the progressive consumption of telomeres, inevitably results in senescence and death. Telomerase activity is a distinctive feature of almost all tumor cells, characterizes some types of normal cells, such as inflammatory elements, and is strongly repressed in most human somatic tissues (Holt and Shay, 1999). Telomerase activity determination by TRAP assay enables quantitative evaluations to be made, but requires pure, homogeneous cell populations to exclude contamination by different cell types. Therefore, one of the main aims of the present study was to develop a monoclonal antibody capable of discriminating between normal and tumor cells. The immunohistochemical results obtained in this study testify to the capacity of this monoclonal antibody to react specifically with dedifferentiated, but not

377

with some terminal differentiated tissues. TRAP assay data also bear witness to the ability of the antibody to strongly inhibit telomerase activity in homogeneous bladder cancer cell lines. Although some tumor cells use alternative pathways to maintain telomere integrity (alternative lengthening of telomeres: ALT pathway) (Bechter et al., 2003), this escape system to prevent shortening of chromosomal telomeric ends is present only in a low percentage of cases (Bryan et al., 1997). Telomerase reactivation is the preferential pathway in the vast majority of tumors. Therefore, our results open up important possibilities for non-invasive diagnosis aimed at the detection of telomerase activity in biological fluids, such as urine and sputum by discriminating between tumor and other contaminating normal cells. Moreover, the combination of immunomorphological and enzymatic analyses permits the cellular localization of the enzyme to be related to its activity and contributes to the generation of functional hypotheses and to a deeper understanding of metabolic and biomolecular mechanisms. Finally, the possibility of detecting reactivity by fluorescence cytometry opens up a whole new field of speculative research.

Acknowledgements The Authors wish to thank Professor Rosella Silvestrini for her invaluable scientific contribution and Gráinne Tierney for editing the manuscript. This work was supported by Consiglio Nazionale della Ricerca (CNR), Grant No. 0100216.ST97, Rome, by Istituto Oncologico Romagnolo, Forl`ı, by Ministero della Istruzione, Università e Ricerca (MIUR), P.N.R. Tecnologie in Oncologia, Tema 3, Progetto No. 7913, Rome, Italy.

References Avilion, A.A., Piatyszek, M.A., Gupta, J., Shay, J.W., Bacchetti, S., Greider, C.W., 1996. Human telomerase RNA and telomerase activity in immortal cell lines and tumor tissues. Cancer Res. 56, 645–650. Bechter, O.E., Zou, Y., Shay, J.W., Wright, W.E., 2003. Homologous recombination in human telomerase-positive and ALT cells occurs with the same frequency. EMBO Rep. 4, 1138–1143.

378


Bryan, T.M., Englezou, A., Dalla-Pozza, L., Dunham, M.A., Reddel, R.R., 1997. Evidence for an alternative mechanism for maintaining telomere length in human tumors and tumor-derived cell lines. Nat. Med. 3, 1271–1274. Fedriga, R., Gunelli, R., Nanni, O., Bacci, F., Amadori, D., Calistri, D., 2001. Telomerase activity detected by quantitative assay in bladder carcinoma and exfoliated cells in urine. Neoplasia 3, 446–450. Feng, J., Funk, W.D., Wang, S.-S., Weinrich, S.S., Avilion, A.A., Chiu, C.-P., Adams, R.R., Chang, E., Allsopp, R.C., Yu, J., Le, S., West, M.D., Harley, C.B., Andrews, W.H., Greider, C.W., Villeponteau, B., 1995. The RNA component of human telomerase. Science 269, 1236–1241. Hahn, W.C., Counter, C.M., Lundberg, A.S., Beijersbergen, R.L., Brooks, M.W., Weinberg, R.A., 1999. Creation of human tumour cells with defined genetic elements. Nature 400, 464–468. Harley, C., Futcher, A.B., Greider, C.W., 1990. Telomeres shorten during ageing of human fibroblasts. Nature 346, 866–868. Harlow, E., Lane, D., 1999. Using Antibodies. A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Holt, S.E., Shay, J.W., 1999. Role of telomerase in cellular proliferation and cancer. J. Cell. Physiol. 180, 10–18. Kim, N.W., Piatyszek, M.A., Prowse, K.R., Harley, C.B., West, M.D., Ho, P.L.C., Coviello, G.M., Wright, W.E., Weinrich, S.L., Shay, J.W., 1994. Specific association of human telomerase activity with immortal cells and cancer. Science 266, 2011–2015. Kohler, G., Milstein, C., 1975. Continuous cultures of fused cells secreting antibody of predefined specificity. Nature 256, 495–497. Kyo, S., Masutomi, K., Maida, Y., Kanaya, T., Yatabem, N., Nakamuram, M., Tanaka, M., Takarada, M., Sugawara, I., Murakami, S., Taira, T., Inoue, M., 2003. Significance of immunological detection of human telomerase reverse transcriptase. Am. J. Pathol. 163, 859–867. Laemmli, U.K., 1970. Cleavage of structural proteins during assembly of the head of bacteriophage T4. Nature 227, 680–685.

Meyerson, M., Counter, C.M., Eaton, E.N., Ellisen, L.W., Steiner, P., Caddle, S.D., Ziaugra, L., Beijersbergen, R.L., Davidoff, M.J., Liu, Q., Bacchetti, S., Haber, D.A., Weinberg, R.A., 1997. hEST2, the putative human telomerase catalytic subunit gene, is upregulated in tumor cells and during immortalization. Cell 90, 785–795. Meyerson, M., 2000. Role of telomerase in normal and cancer cells. J. Clin. Oncol. 18, 2626–2634. Morin, G.B., 1989. The human telomere transferase enzyme is a ribonucleoprotein that synthesizes TTAGGG repeats. Cell 59, 521–529. Nakamura, T.M., Morin, G.B., Chapman, K.B., Weinrich, S.L., Andrews, W.H., Lingner, J., Harley, C.B., Cech, T.R., 1997. Telomerase catalytic subunit homologs from fission yeast and human. Science 277, 955–959. Noeman, S.A., Misra, D.N., Yankes, R.J., Kunz, H.W., Gill, T.J., 1982. Growth of rat–mouse hybridomas in nude mice and nude rats. J. Immunol. Meth. 55, 319–326. Shay, J.W., Bacchetti, S., 1997. A survey of telomerase activity in human cancer. Eur. J. Cancer 33, 787–791. Shay, J.W., Wright, W.E., 2000. Hayflick, his limit, and cellular ageing. Nat. Rev. 1, 72–76. Shay, J.W., Zou, Y., Hiyama, E., Wright, W.E., 2001. Telomerase and cancer. Hum. Mol. Genet. 10, 677–685. Wright, W.E., Shay, J.W., Piatyszek, M.A., 1995. Modifications of a telomeric repeat amplification protocol (TRAP) result in increased reliability, linearity and sensitivity. Nucleic Acids Res. 23, 3794–3795. Yang, S., Zhang, B., Wang, J., Liao, S., Han, J., Wei, J., Hou, L., 2001. Monoclonal antibodies against human telomerase reverse transcriptase (hTERT): preparation, characterization, and application. Hybridoma 20, 249–255. Yi, X., Tesmer, V.M., Savre-Train, I., Shay, J.W., Wright, W.E., 1999. Both transcriptional and posttranscriptional mechanisms regulate human telomerase template RNA levels. Mol. Cell. Biol. 19, 3989–3997.


Functional periplasmic secretion of organophosphorous hydrolase using the twin-arginine translocation pathway in Escherichia coli Dong Gyun Kang a , Gio-Bin Lim b , Hyung Joon Cha a,∗ a

Department of Chemical Engineering, Division of Molecular and Life Sciences, Pohang University of Science and Technology, Pohang 790-784, Republic of Korea b Department of Chemical and Biochemical Engineering, University of Suwon, Hwasung 445-743, Republic of Korea Received 5 November 2004; received in revised form 3 May 2005; accepted 12 May 2005

Abstract Recombinant Escherichia coli systems expressing organophosphorous hydrolase (OPH) have been proposed for biotransformation of toxic organophosphate compounds. However, whole cell biocatalyst systems are critically disadvantaged due to substrate diffusion limitations. To enhance whole cell biocatalytic efficiency, we engineered E. coli, for the first time, to secrete metal ion cofactor-requiring OPH into the periplasmic space using the twin-arginine translocation (Tat) pathway. In particular, the twin-arginine signal sequence of E. coli trimethylamine N-oxide (TMAO) reductase (TorA) was employed. Even though total OPH activity in the cell lysate fraction was lower in the periplasmic-secreting strain than in the control cytosolic-expressing strain, whole cell OPH activity was approximately 2.8-fold higher due to successful translocation of OPH into the periplasmic space. In addition, whole cell OPH activity in the periplasmic-secreting strain was far more stable than that in the cytosolicexpressing strain. Therefore, Tat-driven periplasmic-secreting E. coli can be successfully employed as efficient whole cell biocatalysts. © 2005 Elsevier B.V. All rights reserved. Keywords: Periplasmic space; Secretion; Twin-arginine translocation pathway; TorA; Organophosphorous hydrolase; Escherichia coli; Whole cell biocatalyst

1. Introduction Organophosphate compounds are widely used in many pesticides (Paraoxon, Parathion, Coumaphos, ∗ Corresponding author. Tel.: +82 54 279 2280; fax: +82 54 279 5528. E-mail address: [email protected] (H.J. Cha).


and Diazinon) and chemical nerve agents (Sarin and Soman) (Donarski et al., 1989). Organophosphorous hydrolase (OPH) from Pseudomonas diminuta or Flavobacterium sp. is a homodimeric organophosphotriesterase that requires metal ion as a cofactor and can degrade a broad spectrum of toxic organophosphates (Mulbry and Karns, 1989; Grimsley et al., 1997). This enzyme can hydrolyze various phosphorus–ester bonds

380

D.G. Kang et al. / Journal of Biotechnology 118 (2005) 379–385

including P O, P F, P CN, and P S bonds (Lai et al., 1995). The application of OPH for bioremediation is of great interest due to its high turnover rate. Recombinant Escherichia coli expressing OPH can degrade a variety of organophosphates (Serdar and Gibson, 1985). The ability of E. coli to grow to much higher densities than P. diminuta or Flavobacterium enables the development of large-scale detoxification processes (Chen and Mulchandani, 1998). However, recombinant E. coli cells produce low yields of OPH due to the low solubility of this protein (Mulbry and Karns, 1989; Grimsley et al., 1997). In addition, the E. coli cell membrane can be a substrate diffusion barrier affecting whole cell biocatalytic efficiency (Rainina et al., 1996). Strategies attempted to enhance OPH production yield or biocatalytic efficiency include insertion of multiple gene fusions (Wu et al., 2001), fusion with a soluble partner to increase solubility (Cha et al., 2000), co-expression with Vitreoscilla hemoglobin (Kang et al., 2002), and display on the cell surface (Richins et al., 1997; Li et al., 2004). Periplasmic secretion of target proteins via translocation across the cytoplasmic membrane in Gramnegative bacteria such as E. coli can be a potential strategy for reducing the substrate diffusion barrier in whole cell biocatalyst systems (Kaderbhai et al., 2001). Periplasmic secretion of OPH in E. coli has not yet been reported. At least four distinct pathways have been reported to mediate periplasmic secretion, namely the general secretory (Sec) pathway (Pugsley, 1993), the signal recognition particle (SRP)-dependent pathway (Meyer et al., 1982), the YidC-dependent pathway (Samuelson et al., 1996), and the twin-arginine translocation (Tat) pathway (Berks, 1996). For the first three of these pathways, the polypeptide chain is translocated across the membrane in a largely unfolded state under requirement of ATP hydrolysis (Schatz and Dobberstein, 1996). However, in the Tat pathway, proteins are at least partially folded prior to export, and without the requirement of ATP (Cline et al., 1992). The Tat pathway has been successfully used for periplasmic secretion of several foreign proteins such as cofactorcontaining, multimeric, and disulfide-containing proteins (Halbig et al., 1999; Rodrigue et al., 1999; DeLisa et al., 2003). In the present work, we aimed to achieve functional secretion of OPH molecules into the periplasmic space using the Tat pathway. In particular, we

used the twin-arginine signal sequence of the E. coli enzyme trimethylamine N-oxide (TMAO) reductase (TorA) because OPH molecules require metal ion cofactors. TMAO reductase is a periplasmic enzyme which catalyzes reduction of TMAO to trimethylamine, and functions as a component of the anaerobic respiratory chain which provides energy for bacterial growth (Silvestro et al., 1989; Barrett and Kwan, 1985).

2. Materials and methods 2.1. Bacterial strains and plasmid construction E. coli strain TOP10 (Invitrogen, Carlsbad, CA, USA) was used for constructing recombinant plasmids. E. coli strain BL21(DE3) (Novagen, Madison, WI, USA) was used as a host for recombinant OPH expression. To make the cytosolic expression plasmid pEO (6434 bp), the OPH gene was polymerase chain reaction (PCR)-amplified from pTO (Kang et al., 2002) using OPH-N1 and OPH-C primers. The resulting 1.0 kb fragment was digested with NdeI and HindIII, and ligated into the NdeI/HindIII site of pET22b(+) (5493 bp; Novagen) (Fig. 1). To enable secretion of OPH, the twin-arginine signal sequence of E. coli TorA was used. The TorA signal sequence encoding the entire signal sequence and the first four amino acid residues of the mature protein (Thomas et al., 2001) was PCR-amplified from E. coli K-12 genomic DNA using TorA-SP1 and TorA-SP2 primers. This fragment was digested with NdeI and NcoI and cloned into the same sites of pET22b(+), resulting in pETat. To construct in-frame fusion of the TorA signal sequence and the OPH gene, the OPH gene was PCR-amplified using OPH-N2 and OPH-C primers. The PCR fragment was digested with NcoI and HindIII and ligated into the corresponding sites of pETat. The resulting plasmid pETO (6563 bp) contained an N-terminal TorA signal sequence-fused TorA SS::OPH construct (Fig. 1). The constructs were confirmed by DNA sequencing analysis. The strains, plasmids, and primers used in this study are summarized in Table 1. 2.2. Culture condition and cell fractionation Plasmids bearing strains were cultured in Luria broth (LB) or M9 minimal medium (12.8 g l−1


Fig. 1. Gene maps of recombinant plasmids pEO and pETO. Abbreviations: T7lac, T7 and lac hybrid promoter; Tat, twin-arginine TorA signal sequence of TMAO reductase; OPH, organophosphorous hydrolase gene; PelB, PelB signal sequence; MCS, multi-cloning site; term, terminator.

Na2 HPO4 ·7H2 O, 3 g l−1 KH2 PO4 , 0.5 g l−1 NaCl, 1 g l−1 NH4 Cl, 3 mg l−1 CaCl2 , and 1 mM MgSO4 ) containing 0.5% (w/v) glucose, 50 ␮g ml−1 of ampicillin, and 0.1 mM CoCl2 at the final concentration.

381

Sub-cultures were grown overnight in 6 ml LB and used to inoculate 50 ml M9 to a starting OD600 of 0.2. Cells were cultured in 250 ml Erlenmeyer flasks at 250 rpm and 37 ◦ C. When cultures reached a cell density (OD600 ) of 1.2, 0.5 mM IPTG was added to induce recombinant protein expression. After induction, cells were incubated at 250 rpm and 37 ◦ C for 24 h before fractionation. Total cell lysates, periplasmic, and cytoplasmic fractions were prepared according to the method described in the pET system manual (Novagen). One milliliter of cells was harvested, resuspended in 0.1 ml phosphate-buffered saline (PBS), and ultrasound sonicated for 2 × 30 s. This sample was saved as a total cell lysate. Ten milliliters culture was harvested and resuspended in 1 ml 30 mM Tris–HCl, pH 8.0, 20% sucrose, and then 2 ␮l 0.5 M EDTA, pH 8.0, was added. The cell suspension was incubated at room temperature for 10 min and pelleted by centrifugation at 10,000 × g at 4 ◦ C for 10 min. The pellet was resuspended in 1 ml ice-cold 5 mM MgSO4 and incubated on ice for 10 min. The osmotic shocked cells were pelleted by centrifugation as above, and the supernatant was collected as a periplasmic fraction and the pellet was saved as a cytoplasmic fraction. Each fractionated sample was analyzed using sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) or OPH activity assay. To check the fractionation was properly performed, each fraction was analyzed using Western blot for

Table 1 Strains, plasmids, and primers used in this study Description

References or sources

F− mcrA ∆(mrr-hsdRMS-mcrBC) φ80lacZ ∆M15 ∆lacX74 deoR recA1 araD139 ∆(ara-leu)7697 galU galK rpsL (StrR) endA1 nupG F− ompT hsdSB (rB− m− B ) gal dcm (DE3)

Invitrogen

Plasmids pET22b(+) pTO pEO pETat pETO

AmpR , T7lac promoter, pBR322 ori OPH gene source OPH gene from pTO cloned into pET22b(+) TorA signal sequence from K12 genomic DNA cloned into pET22b(+) TorA signal sequence and OPH gene cloned into pET22b(+)

Novagen Kang et al. (2002) This study This study This study

Primers OPH-N1 OPH-N2 OPH-C TorA-SP1 TorA-SP2

5 5 5 5 5

This study Kang et al. (2002) Kang et al. (2002) This study This study

Strains TOP10 BL21(DE3)

ggaattccatatgggatcgatcggcacagg 3 ggccatgggatcgatcggcacaggcg 3 ggaagctttcatgacgcccgcaaggtcg 3 ggaaatccatatgaacaataacgatctctttc 3 ctccatggccgcttgcgccgcagtcgc 3

Novagen

382


␤-lactamase as a periplasmic marker or GroEL as a cytoplasmic marker. 2.3. Analytical assays Cell density (OD600 ) was measured at 600 nm on a UV/VIS spectrophotometer (UV-1601PC; Shimadzu Corp., Kyoto, Japan). OPH activities of each samples were measured in 100 mM CHES [2-(N-cyclohexylamino)ethane-sulfonic acid] buffer, pH 9.0 (final concentration), at 400 nm (ε400 = 17,000 M−1 cm−1 ) using UV/VIS spectrophotometer by following the increase in absorbance of p-nitrophenol from the hydrolysis of 1 mM Paraoxon (Sigma, St. Louis, MO, USA) which was added to the sample immediately prior to the analysis. One unit of OPH activity was defined as micromoles Paraoxon hydrolyzed per minute (Caldwell et al., 1991).

For Western blot analysis, the gel was transferred onto a nitrocellulose membrane (Amersham Pharmacia Biotech, Buckinghamshire, England) with transfer buffer (48 mM Tris–HCl, 39 mM glycine, 20% methanol, pH 9.2) by using a Trans-Blot SD Cell (BioRad) at 15 V for 30 min. After blocking for 1 h in TBS buffer (20 mM Tris–HCl, 500 mM NaCl, pH 7.5) containing 5% (w/v) non-fat dry milk, the membrane was then incubated for 1.5 h at room temperature in antibody solution (1%, w/v, non-fat dry milk in TTBS (TBS with 0.05% Tween-20)) containing rabbit anti-␤lactamase antibody (1:1000, v/v) (Chemicon, Temecula, CA, USA) or rabbit anti-GroEL antibody (1:1000, v/v) (Sigma) and probed with secondary anti-rabbit IgG conjugated with alkaline phosphatase (1:5000, v/v) (Sigma). After successive washing with TTBS and TBS, BCIP/NBT color development solution (BioRad) was added to detect and the reaction was quenched with distilled water.

2.4. SDS-PAGE and Western blot analysis 3. Results and discussion Each sample was mixed with SDS sample buffer (10% SDS, 10% ␤-mercaptanol, 0.3 M Tris–HCl (pH 6.8), 0.05% bromophenol blue, 50% glycerol), boiled for 5 min, and resolved by 12.5% (w/v) SDS-PAGE. Each gel was detected by Coomassie blue staining (Sigma), silver staining (Bio-Rad, Hercules, CA, USA), or Western blot.

Using harvested 24 h post-infection samples, we compared the specific OPH activities in total cell lysates and intact cells for two cytosolic and periplasmic expression strains (Fig. 2). The cell lysates from the control strain expressing cytosolic OPH showed much higher (∼11.7-fold) OPH activity than intact

Fig. 2. Specific OPH activities of: (A) total cell lysate, (B) whole cell, and (C) periplasmic fractions. Abbreviations: CE, cytosolic-secreting cells; PE, periplasmic-expressing cells. Cells were grown in M9 media with supplement of 0.1 mM CoCl2 at 37 ◦ C for 24 h upon 0.5 mM IPTG induction. Activity analyses of each sample were carried out in 100 mM CHES buffer, pH 9.0, with 1 mM Paraoxon. Each value and error bar represents the mean of three independent experiments and its standard deviation.


cells. These data indicated there was a significant substrate diffusion problem due to the cell membrane. Even though the control strain expressing cytosolic OPH showed higher cell lysate activity (∼1.3-fold) than the periplasmic expression strain (Fig. 2A), the strain expressing periplasmic OPH exhibited enhanced (∼2.8-fold) whole cell activity compared to the cytosolic-expressing control strain (Fig. 2B). Note that the whole cell activity of the periplasmic expression strain was about 32% of the cell lysate activity. These results indicate that periplasmic expression reduced the substrate diffusion limitation experienced in whole cell biocatalyst systems. For successful use of the Tat pathway for periplasmic secretion of target OPH protein, it is essential to verify the location and to determine periplasmic translocation efficiency of expressed recombinant OPH. In order to achieve this, the periplasmic fraction was isolated and subjected to OPH activity analysis (Fig. 2C). Approximately 22% of OPH activity was found in the periplasmic fraction of cells expressing periplasmic OPH (Fig. 2C) compared to that in the total cell lysate fraction (Fig. 2A). However, the control strain expressing cytosolic OPH showed about 1.7% of activity in the periplasmic fraction (Fig. 2C) compared to that in the total cell lysate fraction (Fig. 2A). These differences between periplasmic-secreting and

383

cytosolic-expressing strains suggest correct localization of OPH proteins into the periplasmic space of host cells occurred. These fractions were also subjected to SDS-PAGE analyses (Fig. 3). In the case of periplasmic expression, OPH molecules were mainly found in a premature form (TorA SS::OPH) with noncleaved TorA signal sequences in the total cell lysate fraction (Fig. 3A). This high level of premature forms might be due to the high transcription rate driven by the strong T7lac hybrid promoter, or due to a low periplasmic translocation rate. However, we found that functional mature OPH was successfully translocated into the periplasmic fraction (Fig. 3B), and this successful periplasmic secretion enhanced whole cell OPH activity. To check the fractionation was properly performed, each fraction was analyzed using Western blot for ␤-lactamase as a periplasmic marker or GroEL as a cytoplasmic marker (Fig. 3C). While ␤-lactamase content in the periplasmic fraction was about 76% of the total (periplasmic and cytoplasmic) fraction, GroEL content was only about 7% in the periplasmic fraction. Therefore, we confirmed that fractionation was properly performed in this work. We assayed remaining whole cell OPH activities under resting cell conditions to investigate OPH stability in both strains (Fig. 4). Not surprisingly, the cytoplasmic expression strain showed steadily

Fig. 3. SDS-PAGE analyses of TorA SS::OPH proteins in: (A) total cell lysate and (B) periplasmic fractions. (C) Western blot analyses for ␤-lactamase as a periplasmic marker and GroEL as a cytoplasmic maker to evaluate fractionation. Abbreviations: CE, cytosolic-secreting cells; PE, periplasmic-expressing cells; M, protein molecular weight marker; CF, cytoplasmic fraction; PF, periplasmic fraction. Cells were grown in M9 media with supplement of 0.1 mM CoCl2 at 37 ◦ C for 24 h upon 0.5 mM IPTG induction. Each 12% SDS-PAGE gel was detected by: (A) Coomassie blue staining, (B) silver staining, or (C) Western blot.

384


References

Fig. 4. Remaining relative whole cell OPH activities of the cells harboring cytosolic and periplasmic OPH proteins under resting cell incubations. Abbreviations: CE, cytosolic-secreting cells; PE, periplasmic-expressing cells. Cells were harvested and resuspended in an equal volume of 1× phosphate-buffered saline (PBS). The whole cell suspension was incubated at room temperature over time and analyzed by OPH activity assay.

decreasing OPH activity, indicating proteolysis of OPH proteins in the cytoplasm. However, importantly, whole cell OPH activity was stably maintained in the strain expressing periplasmic OPH. Interestingly, whole cell OPH activity increased during early incubation. This might be from time gap for periplasmic translocation of cytoplasmic premature OPH. This further translocated-functional OPH might cause increased apparent whole cell activity during early resting cell incubation. The present study showed that Tat-driven periplasmic secretion of OPH is a potential strategy to overcome traditional substrate diffusion limitations in whole cell biocatalyst systems. It appears this system can be successfully used as a whole cell biocatalyst for detoxification of organophosphate compounds.

Acknowledgment The authors would like to acknowledge support for fulfillment of this work by the Brain Korea 21 program issued from the Ministry of Education, Korea, and by the Korea Science and Engineering Foundation through the Advanced Environmental Biotechnology Research Center at Pohang University of Science and Technology.

Barrett, E.L., Kwan, H.S., 1985. Bacterial reduction of trimethylamine oxide. Annu. Rev. Microbiol. 39, 131–149. Berks, B.C., 1996. A common export pathway for proteins binding complex redox cofactors? Mol. Microbiol. 22, 393–404. Caldwell, S.R., Newcomb, J.R., Schlecht, K.A., Raushel, F.M., 1991. Limits of diffusion in the hydrolysis of substrates by the phophotriesterase from Pseudomonas diminuta. Biochemistry 30, 7438–7444. Cha, H.J., Wu, C.F., Valdes, J.J., Rao, G., Bentley, W.E., 2000. Observations of green fluorescent protein as a fusion partner in genetically engineered Escherichia coli: monitoring protein expression and solubility. Biotechnol. Bioeng. 67, 565–574. Chen, W., Mulchandani, A., 1998. The use of live biocatalysts for pesticide detoxification. Trends Biotechnol. 16, 71–76. Cline, K., Ettinger, W.F., Theg, S.M., 1992. Protein-specific energy requirements for protein transport across or into thylakoid membranes, two lumenal proteins are transported in the absence of ATP. J. Biol. Chem. 267, 2688–2696. DeLisa, M.P., Tullman, D., Georgiou, G., 2003. Folding quality control in the export of proteins by the bacterial twin-arginine translocation pathway. Proc. Natl. Acad. Sci. U.S.A. 100, 6115–6120. Donarski, W.J., Dumas, D.P., Heitmeyer, D.P., Lewis, V.E., Raushel, F.M., 1989. Structure–activity relationship in the hydrolysis of substrates by the phosphotriesterase from Pseudomonas diminuta. Biochemistry 28, 4650–4655. Grimsley, J.K., Scholtz, J.M., Pace, C.N., Wild, J.R., 1997. Organophosphorus hydrolase is a remarkably stable enzyme that unfolds through a homodimeric intermediate. Biochemistry 36, 14366–14374. Halbig, D., Wiegert, T., Blaudeck, N., Freudl, R., Sprenger, G.A., 1999. The efficient export of NADP-containing glucose–fructose oxidoreductase to the periplasm of Zymomonas mobilis depends both on an intact twin-arginine motif in the signal peptide and on the generation of a structural export signal induced by cofactor binding. Eur. J. Biochem. 263, 543–551. Kaderbhai, M.A., Ugochukwu, C.C., Kelly, S.L., Lamb, D.C., 2001. Export of cytochrome P450 105D1 to the periplasmic space of Escherichia coli. Appl. Environ. Microbiol. 67, 2136–2138. Kang, D.G., Kim, J.Y.H., Cha, H.J., 2002. Enhanced detoxification of organophosphates using recombinant Escherichia coli with co-expression of organophosphorus hydrolase and bacterial hemoglobin. Biotechnol. Lett. 24, 879–883. Lai, K., Stolowich, N.J., Wild, J.R., 1995. Characterization of P S bond hydrolysis in organophosphorothioate pesticides by organophosphorus hydrolase. Arch. Biochem. Biophys. 318, 59–64. Li, L., Kang, D.G., Cha, H.J., 2004. Functional display of foreign protein on cell surface of Escherichia coli using N-terminal domain of ice nucleation protein. Biotechnol. Bioeng. 85, 214–221. Meyer, D.I., Krause, E., Dobberstein, B., 1982. Secretory protein translocation across membranes—the role of the docking protein. Nature 297, 647–650. Mulbry, W.W., Karns, J.S., 1989. Parathion hydrolase specified by the Flavobacterium opd gene: relationship between the gene and protein. J. Bacteriol. 171, 6740–6746.

D.G. Kang et al. / Journal of Biotechnology 118 (2005) 379–385 Pugsley, A.P., 1993. The complete general secretory pathway in Gram-negative bacteria. Microbiol. Rev. 57, 50–108. Rainina, E., Efremenco, E., Varfolomeyev, S., Simonian, A.L., Wild, J.R., 1996. The development of a new biosensor based on recombinant E. coli for the detection of organophosphorous neurotoxins. Biosens. Bioelectron. 11, 991–1000. Richins, R.D., Kaneva, I., Mulchandani, A., Chen, W., 1997. Biodegradation of organophosphorus pesticides by surfaceexpressed organophosphorus hydrolase. Nat. Biotechnol. 15, 984–987. Rodrigue, A., Chanal, A., Beck, K., Muller, M., Wu, L., 1999. Cotranslocation of a periplasmic enzyme complex by a hitchhiker mechanism through the bacterial Tat pathway. J. Biol. Chem. 274, 13223–13228. Samuelson, J.C., Chen, M., Jiang, F., Moller, I., Wiedmann, M., Kuhn, A., Phillips, G.J., Dalbey, R.E., 1996. YidC mediates membrane protein insertion in bacteria. Nature 406, 637–641.

385

Schatz, G., Dobberstein, B., 1996. Common principles of protein translocation across membranes. Science 271, 1519–1526. Serdar, C.M., Gibson, D.T., 1985. Enzymatic hydrolysis of organophosphates: cloning and expression of a parathion hydrolase gene from Pseudomonas diminuta. Bio/Technology 3, 567–571. Silvestro, A., Pommier, J., Pascal, M.C., Giordano, G., 1989. The inducible trimethylamine N-oxide reductase of Escherichia coli K-12:its localization and inducers. Biochem. Biophys. Acta 994, 208–216. Thomas, J.D., Daniel, R.A., Errington, J., Robinson, C., 2001. Export of active green fluorescent protein to the periplasm by the twinarginine translocase (Tat) pathway in Escherichia coli. Mol. Microbiol. 39, 47–52. Wu, C.F., Valdes, J.J., Rao, G., Bentley, W.E., 2001. Enhancement of organophosphorus hydrolase yield in Escherichia coli using multiple gene fusions. Biotechnol. Bioeng. 75, 100–103.


Surface character of pulp fibres studied using endoglucanases Lars Hildén a,b,∗ , Priit Väljamäe c , Gunnar Johansson d a

WURC, Department of Wood Science, Swedish University of Agricultural Sciences, Box 7008, 750 07 Uppsala, Sweden b Holmen Paper Development Centre, Holmen Paper AB, SE-601 88 Norrk¨ oping, Sweden c Institute of Molecular and Cell Biology, University of Tartu, Vanemuise 46-138, Tartu 51010, Estonia d Department of Biochemistry, Uppsala University, BMC, Box 576, 751 23 Uppsala, Sweden Received 31 August 2004; received in revised form 2 May 2005; accepted 4 May 2005

Abstract The endoglucanase Cel5A from Trichoderma reesei and an endoglucanase from Aspergillus sp. (Novozym 476TM from Novozyme A/S) were evaluated as probes for the surface properties of soft- and hardwood chemical pulp fibres. The hydrolysis time curves were in accordance with a two-phase degradation model described by a biexponential function. The kinetic parameters corresponding to the amount of fast and slow degraded parts of the substrate correlated to tensile index, relative bonded area and z-strength of the paper. All paper properties showing a correlation with enzyme kinetic parameters were related to fibre–fibre interactions. Fluorescence labelling of the reducing end groups in pulp fibres followed by enzyme treatment indicated that the fast substrate class corresponds to the population of “loose” cellulose chain ends not tightly associated with the bulk cellulose. The correlation between the parameters of enzyme kinetics and mechanical properties of the paper produced from the corresponding pulp found in this study should allow a rapid evaluation of the raw fibre material used in paper making process. © 2005 Elsevier B.V. All rights reserved. Keywords: Cellulose; Cellulase; Pulp characterization; Endoglucanase; Fibre surface; Paper

1. Introduction Abbreviations: AA, anthranilic acid; AA-saccharide, anthranilic acid saccharide conjugate; ABTS, 2,2 -Azinobis(3-ethylbenzthiazoline-6-sulfonic acid); CMC, carboxy methyl cellulose; medium viscosity; DNS, dinitrosalicylic acid; Hp, hardwood pulp; N476, Novozym 476TM (an endoglucanase produced by a genetically modified Aspergillus sp.); RBA, relative bonded area; SEM, scanning electron microscopy; Sp, softwood pulp; TI, tensile index ∗ Corresponding author. Tel.: +46 18 67 24 75; fax: +46 18 67 34 89. E-mail address: [email protected] (L. Hildén). 0168-1656/$ – see front matter © 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.jbiotec.2005.05.001

Wood fibres form the main part of both softwoods and hardwoods. A typical mature wood fibre is a tubelike, lignified and dead cell. There are only a few living cells in a tree, the majority situated in the growth zone just under the bark. A typical spruce softwood fibre is 2–4 mm long and 20–40 ␮m wide and a typical birch hardwood fibre (libriform cell) is 1.1–1.2 mm long and 14–40 ␮m wide (Sjöström, 1981). Hardwood also has some shorter and wider cells known as vessel elements.

L. Hildén et al. / Journal of Biotechnology 118 (2005) 386–397

The fibres are composed mainly of cellulose but also of hemicellulose, lignin and some pectin. Pulp consists mainly of separate fibres from disintegrated wood. In the pulping-process the fibres are generally bleached which means that most of the hemicellulose and lignin is removed. Fibres from bleached (white) pulp thus consist mainly (about 80%) of cellulose. The cellulosic part in fibres has a heterogeneous structure and consists of crystalline and non-crystalline (amorphous) parts where the adjacent cellulose chains are not in the tight bonding with each other. In paper, fibres are connected to each other by surface–surface interactions so the ultrastructure of the fibre surface is, along with the properties of the fibre wall, important for paper properties (Duchesne and Daniel, 1999). The ultrastructure of wood fibre surfaces has been studied in numerous ways and aspects: by various microscopical techniques reviewed by Duchesne and Daniel (1999), by enzyme treatment followed by electron spectroscopy for chemical analysis (ESCA) (Buchert et al., 1996a) and by enzyme treatment alone (Yang et al., 1988; Buchert et al., 1996b; Laine et al., 1996; Oksanen et al., 1997). Dislocations (a fibre defect affecting both surface and interior of wood fibres) have also been studied both with enzymatic and chemical methods (Nyholm et al., 2001; Ander and Daniel, 2005). However, the number of studies on cross-sectioned wood cells and analysis of total chemical composition outnumber the studies on wood fibre surfaces. The exact distribution of constituents on the surface of a fibre is still unknown but most likely (at least on chemical pulp fibres) it is heterogeneous (Duchesne and Daniel, 2000). Remaining parts of the primary wall on a fibre surface most likely gives it a different surface character than a fully exposed outer secondary layer. The enzymatic hydrolysis rate of cellulose usually decreases far more rapidly than expected from the total degree of solubilization. This is sometimes referred to as non-linear kinetics. Resultant time curves of hydrolysis generally exhibit a bi-phasic character. The cause of this more or less gradual drop in reaction rate is not fully understood, but it has been postulated that both enzyme- and substrate-related properties contribute to this effect (for reviews see Mansfield et al., 1999; Lynd et al., 2002). In general, the proposed explanations fall into three distinct groups. (i) Substrate heterogeneity, meaning that the bulk cellulose contains several regions

387

(sub-substrates) that differ in their susceptibility to enzymatic attack. (ii) Strong product inhibition of cellulases by cellobiose and (iii) inactivation of cellulase. The real situation probably involves several factors in parallel and the relative contribution of any certain factor to a decrease in the rate may depend not only on the particular enzyme(s) or substrate used but also on the extent of hydrolysis and other experimental conditions like enzyme to substrate ratio, temperature or mixing regime. For endoglucanases the substrate heterogeneity seems to be the most relevant explanation for the decrease in rate. The opened active site topology of endoglucanases restricts them to the amorphous parts of cellulose, whereas the activity on highly crystalline cellulose is very low. Thus, it has been found by several authors that the pre-treatment of cellulose with endoglucanase leads to a strong decrease in the rate of subsequent hydrolysis with the same endoglucanase after washing of the pre-treated cellulose (Nidetzky et al., 1994; Zhang et al., 1999). Product inhibition of endoglucanases has been shown to be an order of magnitude weaker than for cellobiohydrolases and can be principally excluded as a cause for non-linear kinetics in cellulose hydrolysis (Zhang et al., 1999; Gruno et al., 2004). However, the inactivation of the enzyme (e.g., by entrapment of enzyme molecules into the pores of the substrate) should always be considered, especially on complex lignocellulosic substrates (Grethlein, 1985; Tanaka et al., 1988). In this study, we used the well characterized glycosyl hydrolase family 5 endoglucanase Cel5A from Trichoderma reesei and the commercially available endoglucanase Novozym 476TM from a genetically modified Aspergillus sp. (N476) for characterization of the softwood and hardwood derived pulps. Detailed information on the glycosyl hydrolase family nomenclature is to be found at the CAZy server (http://afmb.cnrs-mrs.fr/CAZY/) (Appel et al., 1994).

2. Materials and methods 2.1. Enzymes, substrates, chemicals and apparatus Cel5A (MW 50 kDa, ε280 = 78,000 M−1 cm−1 ) was purified from culture filtrate of T. reesei, as described

388


previously (Bhikhabhai et al., 1984; Saloheimo et al., 1988). N476 was a kind gift from Novozyme A/S, Bagsvaerd, Denmark. Typical fungal cellulase values, MW = 40 kDa, ε280 = 70 × 103 M−1 cm−1 were used in calculations. SDS-PAGE of crude N476 rendered one main band accompanied by a minor band in close proximity. The buffer of N476 was exchanged for the 50 mM sodium acetate pH 5, and this preparation was used without any further purification. CMC (carboxymethyl cellulose, medium viscosity) and ␤-glucosidase (from almonds) was from Fluka, Buchs, Switzerland. Birchwood xylan, glucose oxidase (from Aspergillus niger) and peroxidase (from horseradish) were from Sigma, St. Louis, Missouri, USA. All pulps (accompanied by mechanical and chemical data) were kindly provided by StoraEnso, Karlstad, Sweden in never dried state. All weight values given for pulps refer to calculated dry-weights. The softwood pulps (Sp) were prepared from spruce with small additions of pine and the hardwood pulps (Hp) were prepared from birch. All pulps were kraftprocessed (except for Hp4, see Table 1) and labbleached. All other chemicals were of highest grade available. Water purity was equal to distilled. The spectrophotometer was a UV-160A from Shimadzu, Kyoto, Japan. The fluorescence spectrophotometer was SPF-500 from Aminco, Silver Spring, MD, USA. The sputtering device was a Polaron E5000 from Quorum Technologies, UK and the scanning electron microscope (SEM) was an XL30 from Philips, The Netherlands.

2.2. Measurements of saccharide concentration The ABTS (2,2 -Azinobis(3-ethylbenzthiazoline-6sulfonic acid))-method was performed as described elsewhere (Gruno et al., 2004). ABTS was added to the saccharide containing solution to a final concentration of 0.2 mM. This solution was then incubated with a mixture of ␤-glucosidase, glucose oxidase and peroxidase at final concentrations of 0.5, 2.0 and 0.5 U ml−1 , respectively (units of enzyme activity were as provided by the manufacturers). The incubation was done at 25 ◦ C overnight. Cellobiose was used as a standard and the oxidation of ABTS was followed by the increase in absorbance at 420 nm. Since beside cellobiose, ␤-glucosidase also degrades higher cellooligosaccharides to glucose the assay gives us the total concentration of the soluble cello-oligosaccharide fraction in the equivalents of the standard saccharide used in calibration. The anthrone/sulphuric acid method was performed according to literature using cellobiose as a standard and absorbance measurements at 585 nm (Hörmann and Gollwitzer, 1962). In the anthrone method all saccharides are completely hydrolysed and measured as monosaccharide equivalents. The endoglucanases used release saccharides, mainly cellobiose and glucose, from the reducing end of cellulose chains but the anthrone method will detect all saccharides released into the solution. That means that any saccharides released from the non-reducing ends by other mechanisms, e.g., by pure mechanical wear, will also be

Table 1 Some technical and enzyme-kinetic parameters of the softwood (Sp) and hardwood (Hp) pulps used for paper properties correlation with N476 degradation in this study Pulp

TI (Nm g−1 )

z-Strength (kPa)

RBA (%)

Tear index (Nm2 kg−1 )

Re-wet zero-span (Nm g−1 )

aa (%)

ca (%)

kb (min−1 )

Sp4c Sp5d Sp6e Hp3f Hp4g

49.9 87 96.4 48.1 23.3

332 621 666 608 445

60.7 82.8 78.7 57.2 43

17.8 10 9.4 8.6 4.7

124.8 120.1 130.3 123.3 77.6

0.396 0.570 0.621 0.340 0.408

0.702 0.762 0.826 0.535 0.794

0.197 0.270 0.285 0.203 0.173

a b c d e f g

From the data in Fig. 4 according to Eq. (3), as a % of pulp dry weight. From the data in Fig. 4 according to Eq. (2). The un-milled raw material for Sp5 and Sp6. Laboratory milled Sp4. Factory milled Sp4. Sulphate processed. Sulphite processed.


registered as released saccharides. From samples run without cellulase present it was however evident that this contribution is negligible. The DNS (dinitrosalicylic acid) method is specific for the reducing end-groups in saccharides and was performed according to literature (Miller, 1959; Lawoko et al., 2000). In principle the cellulases are allowed to act on a substrate (e.g., CMC) for a certain time where after the reaction is stopped by addition of the DNSreagent. The substrate is spun down and amount of released reducing saccharides is measured as absorption at 575 nm in the supernatant. 2.3. Separation of fines The fines were separated from the fibres by washing 5 g of pulp with 4 × 250 ml water on a net with 140 ␮m wide openings (ca. 110 mesh). The fines were collected and concentrated with a 0.45 ␮m filter. The fines content was 2–3% for all pulps except for Sp4, which contained 1.5% fines. 2.4. Fluorescence labelling of reducing ends on cellulose with anthranilic acid The bleached softwood pulp Sp7 was labelled with anthranilic acid (AA) essentially as described in (Kipper et al., 2005): AA (40 mmol) was dissolved by stirring in 300 ml 50 mM sodium acetate buffer at 70 ◦ C followed by the addition of 2 g Sp7 and 200 mmol of sodium-cyanoborohydride and incubated for 8 h. AAlabelled pulp was extensively washed on a 0.22 ␮m filter, first twice with water/ethanol 1:1 (v:v), then twice with buffer and finally twice with water. The labelled pulp was stored in a moist/wet condition (ca. 60% water). 2.5. Enzymatic hydrolysis 2.5.1. Enzyme/substrate ratio In order to ensure a surplus of enzyme as well as easily detectable amount of released saccharides, a number of tests were performed. The concentration of pulp fibres was limited by the fact that over ca. 10 g l−1 , the pulp suspension becomes too viscous to allow sufficient mixing in syringes. Four different pulps Sp1, Sp2, Hp1 and Hp2 (1–8 g l−1 ) were incubated with 0.5 ␮M Cel5A at 40 ◦ C

389

in 50 mM sodium acetate buffer, pH 5.0 in a total volume of 1.0 ml for 1 h. The reaction was stopped by boiling for 5 min. The amount of released saccharides was determined with the ABTS-assay. For N476 the Sp7 (9.0 g l−1 ) was incubated with N476 (0.006–11.81 ␮M) in 50 mM sodium acetate buffer, pH 5.0 in a total volume of 10.0 ml for 1 h. The reaction was stopped by filtration through a 0.45 ␮m syringe filter. Triplicates from each sample were analyzed for released saccharides using the anthrone assay. In one experiment, an extra dose of N476 (equal to the initial dose) was added to the hydrolysis mixture of Sp4 (8.0 g l−1 ) in 50 mM sodium acetate buffer, pH 5.0 at 53 ◦ C after 1.5 h of hydrolysis. This addition raised the concentration of N476 from 3.34 to 7.0 ␮M. Released saccharides were analyzed before and after the addition of the new N476 dosage by the anthrone assay. 2.5.2. Enzymatic hydrolysis of pulps and fines Two softwood-pulps (Sp2 and Sp3) were incubated with 1 ␮M Cel5A in syringes on a vertical rotating disc at 40 ◦ C. Initial volume was 3.50 ml and initial concentration of pulp was 2.0 g l−1 . The buffer/pulp mixture was left to adjust to 40 ◦ C and the reaction was initiated by the addition of enzyme. Samples of ca. 250 ␮l (total 10 samples) were taken at selected times by pressing the solution through a 0.45 ␮m syringe filter. Later the samples were weighed to allow volume compensation in the calculations. The increased pulp concentration caused by the decreased volume during sampling was compensated in the subsequent analysis. The hydrolysis of fines was performed in 1.5 ml microcentrifuge tubes by incubating the suspension of fines (1.5 g l−1 ) with 1 ␮M Cel5A in 50 mM sodium acetate buffer, pH 5.0 at 40 ◦ C. The reaction was stopped by adding NaOH to pH 12.3. The tubes were thereafter centrifuged at 16,000 × g for 5 min and the supernatant was analyzed for the released saccharides using the anthrone assay. For N476 degradation of pulp three bleached softwood pulps of the same origin but slightly differently treated (Sp4, Sp5 and Sp6) and two bleached birch pulps of different origin (Hp3 and Hp4) (see Table 1) were incubated in syringes on a swaying table with 3.3 ␮M N476 in 50 mM sodium acetate buffer, pH 5.0 at 53 ◦ C. Initial volume of the reaction mixture and the concentration of pulp were 50.8 ml and 8.0 g l−1 ,

390


respectively. The buffer/pulp mixture was thermostated to 53 ◦ C and reaction was initiated by the addition of enzyme. Samples of ca. 500 ␮l (total 13 time points) were withdrawn at selected times by pressing the reaction mixture through a 0.45 ␮m filter and analysed for released saccharides in duplicate using the anthrone assay. Each pulp was run in duplicate with fines and without fines. Zero time points were made by pressing the pulp solution without the enzyme through the 0.45 ␮m filter under otherwise identical conditions. The background from the enzyme was also compensated using separate measurements. The amount of enzyme in samples was estimated using absorbance measurements at 280 nm and this figure was used further to compensate the enzyme background in the anthrone assay. 2.5.3. Hydrolysis of anthranilic acid labelled pulp with N476 Hydrolysis of AA-labelled Sp7 was performed by incubation with 3.3 ␮M N476 in 50 mM sodium acetate buffer, pH 5.0 at 53 ◦ C in a syringe. The initial concentration of labelled pulp was 5.0 g l−1 in 50.8 ml total volume. The buffer/pulp mixture was thermostated to 53 ◦ C and the reaction was initiated by the addition of enzyme. Samples of ca. 1.0 ml (total 13 time points) were withdrawn at selected intervals by pressing the reaction mixture through 0.45 ␮m filter and analyzed for released saccharides and fluorescence. Anthranilic acid cellobiose conjugate (AA-cellobiose) was used as a standard to calibrate the fluorescence signal. Excitation and emission wavelengths were set to 330 and 425 nm, respectively. AA-cellobiose was synthesised as described elsewhere (Sato et al., 1998; Huang et al., 2000; Kipper et al., 2005). 2.5.4. CMC-ase and xylanase activity of Cel5A and N476 Endoglucanase activity of the two endoglucanases was assessed on CMC as a substrate by following the production of reducing end-groups with the DNSmethod. CMC (18 mg ml−1 ) was incubated with 50 nM of either Cel5A or N476 in 50 mM NaAc buffer pH 5.0 at 40 ◦ C for 10 min in 0.2 ml total volume. The reaction was stopped by the addition of 0.2 ml DNS reagent. For xylanase activity birch wood xylan (10 mg ml−1 ) was incubated with 1.0 ␮M of either Cel5A or N476 in 50 mM NaAc buffer pH 5.0 at 40 ◦ C

for 30 min in 0.2 ml total volume. The reaction was stopped by the addition of 0.2 ml DNS reagent. In case of zero data points the enzyme was added after DNS reagent. In order to get into the linear region of the DNS assay 20 ␮l of 5 mM cellobiose was added (after DNS reagent) to all data points including blanks and calibration data points. Cellobiose was used as a standard in calibration. Standard errors are based on five parallel measurements. 2.6. Scanning electron microscopy of Cel5A treated pulps Cel5A-treated (1 ␮M Cel5A, 2.0 g l−1 pulp, 24 h at and untreated Sp2 and Sp3 fibres and the corresponding fines were washed repeatedly with water, freeze-dried and mounted on double bonding tape on metal discs. They were thereafter sputtered with gold and photographed with SEM. 40 ◦ C)

2.7. Data treatment The time curves of the enzymatic degradation of pulps were fitted to three different equations using nonlinear regression analysis. The following models were used: Pseudo-first order reaction: [P] = a[1 − exp(−bt)]

(1)

Fractal-like kinetics analogue of pseudo-first order reaction: [P] = a[1 − exp(−kt (1−h) )]

(2)

The sum of two pseudo first order reactions (biexponential fit), [P] = a[1 − exp(−bt)] + c[1 − exp(−dt)]

(3)

In all equations [P] stands for the extent of degradation (as a % of the total pulp) and t is the hydrolysis time. All other parameters are empirical. 2.8. The mechanical parameters of paper The values of the technical parameters were provided along with the pulps but for clarity a short introduction of the parameters is included here. z-Strength


(kPa) is the strength in the papers z-direction, i.e., perpendicular to the sheet. It is measured by placing the paper between two adhesive plates and pulling until it delaminates. Re-wet zero-span (Nm g−1 ) represents the interior strength of the fibres. It is measured by placing a piece of the re-wetted paper between two clamps, with a span as close to zero as possible between them, and then the paper is pulled apart. Tensile index (TI) (Nm g−1 ) is measured as zero-span but here the distance between the clamps is set to 100 mm. Thus, this parameter depends on the fibre–fibre interaction and, in turn, the strength of the bonds between the fibres as well as the relative bonded area. Relative bonded area (RBA) (%) is the amount of contact points between the fibres in a sheet of paper. Tear index (Nm2 kg−1 ) is the force required to make a certain tear out of the z-plane of the paper. It is partly dependant on fibre–fibre interactions. Wet fibre flexibility (N−1 m−2 ) is the ability of a single wet fibre to bend. Fibre length (mm) and width (␮m) simply describes the spatial extension of the fibre. Fibre shape factor (%) is the relation between the exposed length and the true length of the fibre. Coarseness (␮g m−1 ) is the relation between mass and total length of a portion of fibres.

391

Fig. 1. Extent of softwood pulp Sp7 degradation as a function of Aspergillus endoglucanase N476 concentration. Sp7 (9.0 g l−1 ) was incubated with N476 (0.006–11.81 ␮M) in 50 mM sodium acetate buffer, pH 5.0 at 53 ◦ C for 1 h. Extent of degradation represents a fraction (%) of released soluble saccharides to the initial dry-weight of pulp. Error bars are from the saccharide analysis.

3.1. Enzyme/substrate ratio and enzyme specificity

degradation had dropped significantly from its initial value. The extra addition of N476 after incubation of Sp7 (8.0 g l−1 ) for 90 min, increased its concentration from 3.3 to 7.0 ␮M but did not cause any enhancement in further degradation (Fig. 2). These results demonstrate that at 3.3 ␮M concentration the enzyme is in excess and that the possible inactivation of the enzyme during the degradation cannot be responsible for the observed non-linear kinetics under our experimental conditions. The incubation of four different pulps (one softwood and hardwood both bleached and

Contrary to the conventional Michaelis-Menten kinetics, the hydrolysis of the polymeric cellulose substrate exhibits a so-called dual saturation character, i.e., both enzyme and substrate can be saturated with each other (Bailey, 1989; Lynd et al., 2002). In order to characterize the substrate by means of enzyme kinetics, one must be sure that the kinetics is governed by the changes in the properties of the substrate and not those in the enzyme. For that it is better to use the enzyme in excess with the substrate. Incubation of a pulp Sp7 (9.0 g l−1 ) with varying concentrations of N476 demonstrates the saturation of the substrate with an enzyme (Fig. 1). The apparent half saturating concentration of N476 can be estimated around 0.1–0.2 ␮M. In order to ensure that the kinetics is not influenced by the changes in enzyme an experiment was made where a new amount of N476 was added at a point in the time curve where the rate of

Fig. 2. Time curve of the degradation of softwood pulp Sp7 by Aspergillus endoglucanase N476 with the addition of “fresh” enzyme. Sp7 (8.0 g l−1 ) was incubated with 3.3 ␮M N476 in 50 mM sodium acetate buffer, pH 5.0 at 53 ◦ C. After 1.5 h of hydrolysis (position indicated with arrow) a new portion of N476 was added to rise its concentration to 7.0 ␮M. Data points recorded after addition of “fresh” enzyme are represented by the filled labels. Error bars are from the saccharide analysis.

3. Results and discussion

392


unbleached) (1.0–8.0 g l−1 ) with 0.5 ␮M Cel5A for 1 h all rendered a linear correlation between the amount of the substrate and activity (data not shown). Thus, within this concentration range the amount of the substrate is rate limiting. Based on results just mentioned, in all further experiments, the concentration of the Cel5A and N476 were kept at 1.0 and 3.3 ␮M, respectively. As the product concentrations in our experiments (around 200 ␮M) remain far below the inhibition constants for endoglucanases (Gruno et al., 2004) the product inhibition can also be ruled out as a cause of the non-linear kinetics. Strong limitation of activity by the substrate has been reported for different endoglucanases acting on different substrates (Medve et al., 1998; Väljamäe et al., 1999; Zhang et al., 1999; Carrard et al., 2000). Both endoglucanases showed a high activity on CMC. At 40 ◦ C the catalytic activities of Cel5A and N476 were 3900 ± 70 and 1800 ± 40 min−1 , respectively. Considering the temperature difference of 10 ◦ C the activity of Cel5A found here correlates with the value of 2200 min−1 (30 ◦ C, pH 5) reported by Macarron et al., 1993. For both enzymes the activity on birch wood xylan was about three orders of magnitude lower than on CMC being 2.8 ± 0.4 and 1.5 ± 0.4 min−1 for Cel5A and N476, respectively. Low xylanase activity of T reesei Cel5A is supported also by the study of Lawoko et al. (2000) who found that the xylanase activity of Cel5A remained below the limit of detection (31 min−1 in their study). 3.2. Hydrolysis of pulps with Cel5A, the impact of fines All pulps have a broad distribution of fibres and particles with respect to length/size. The smallest particles, fragments of fibres, are generally referred to as fines. In this work, we have defined fines as all particles passing a net with 140 ␮m wide openings (ca. 110 mesh). The fraction of fines from the two-spruce softwood pulps Sp2 and Sp3 were separated by repeated washing on the net with the 140 ␮m wide openings. In both cases, the fines fraction corresponded to about 3% of the total weight of the pulp. Fig. 3 shows the degradation of the fibre fraction of the Sp2 and Sp3 pulps (2.0 g l−1 ) by 1.0 ␮M Cel5A. About 1.5 and 2.3% of the fibres from Sp2 and Sp3, respectively were degraded after 25 h of hydrolysis. The activity of Cel5A on both Sp2 and Sp3 fines was considerably higher, reaching 7% degrada-

Fig. 3. Hydrolysis of fibres and fines derived from two softwood pulps by T. reesei endoglucanase Cel5A. Fibres from Sp2 ( ) and Sp3 () (both at 2.0 g l−1 ) were incubated with 1 ␮M Cel5A in 50 mM sodium acetate buffer pH 5.0 at 40 ◦ C. The extent of degradation of corresponding fines (filled labels) are represented as their contribution to the total pulp degradation by considering their amount in pulp (3%). The concentration of the fines was 2.0 g l−1 and the actual extent of degradation was up to 7% after 21 h of hydrolysis. Solid lines are according to Eq. (3). Error bars are from the two parallel experiments.

tion after 20 h treatment under the same conditions (data not shown). Although the fines were degraded with about four times higher activity than the fibres, the contribution of the fines to the degradation of nonfractionated pulp is not high. The lower curves in Fig. 3 show the estimated contribution of the fines to the overall degradation of the non-fractionated pulp. 3.3. Detailed evaluation of five bleached pulps using endoglucanase N476 Three bleached spruce pulps (softwood) Sp4, Sp5, Sp6 and two bleached birch pulps (hardwood) Hp3 and Hp4 were treated with 3.3 ␮M Aspergillus endoglucanase N476. Although there were no apparent differences in using T. reesei endoglucanase Cel5A or Aspergillus sp. endoglucanase N476 in this study, the latter enzyme is commercially available in large quantities and was therefore preferred. Also, when the goal is surface characterization the specific activity of the enzyme is of minor importance provided that the same enzyme is used throughout. For each time curve, 13 samples between 1 and 90 min of hydrolysis were analyzed for released saccharides. Saccharide analysis was done by the anthrone assay and each data point was analyzed in duplicate. For each pulp the enzymatic degradation of fibres without fines was also followed.


393

of the pseudo-first order reaction (2) that was recently used to describe the synergistic cellulose degradation (Väljamäe et al., 2003) gave a good approximation to the hydrolysis data. For Eq. (2) the average R2 for all time curves (20 time curves with total 508 data points were analyzed) was 0.9926 ± 0.0043. Somewhat surprisingly, a slightly worse fit with an average R2 of 0.9904 ± 0.0055 was obtained using an equation for the sum of the two pseudo-first order reactions Eq. (3). The empirical parameters a and h obtained from Eq. (2) were scattered and did not correlate with the paper properties but there was a good correlation between the apparent first order rate constant k from Eq. (2) and the tensile index (see Fig. 7D) and RBA, however it is difficult to attach a clear-cut physical meaning to this parameter. In general, if Eq. (3) describes the reaction involving two distinct sub-substrates, then Eq. (2) accounts for the reaction with substrate that may be initially homogeneous but which gradually loses its reactivity as the reaction proceeds. An alternative to the latter proposal is the presence of numerous subsubstrates so that the change in net reactivity appears to be gradual. Fig. 4. Hydrolysis of the fibres from softwood (A) and hardwood (B) pulps by Aspergillus endoglucanase N476. Fibres (8.0 g l−1 ) were incubated with 3.3 ␮M N476 in 50 mM sodium acetate buffer, pH 5.0 at 53 ◦ C. Following pulps were examined A) Sp4 (♦), Sp5 () and Sp6 ( ) and B) Hp3 () and Hp4 (♦). All corresponding filled labels represent the results from parallel measurements under the same conditions. Solid lines are according to Eq. (3). Error bars are from the saccharide analysis.

The enzymatic hydrolysis of fibres was carried out in all respect similar to that of the non-fractionated pulps. For all pulps, the fines content was 2–3%. All hydrolysis experiments were run in two parallels. Fig. 4 shows the time curves for the release of soluble saccharides from the fibres of five pulps upon hydrolysis with N476. The general pattern of the degradation (curvature of the time curve) was similar for all of the pulps, both the softwood and hardwood, with or without fines. The only clear difference between different pulps was revealed in the absolute magnitude of the degradation. Three kinetic models (Eqs. (1)–(3)) were tested for the approximation of the hydrolysis data using non-linear regression analysis. The simple pseudo-first order reaction (1) was not satisfactory and gave a systematic deviation from the experimental data in all cases. The equation for fractal-like kinetics analogue

3.4. Identification of the cellulose sub-structures underlying the pattern of enzymatic degradation Non-linear kinetics observed for the degradation of pulps by N476 is typical for cellulose enzymatic hydrolysis. In the discussion above, we ruled out the enzyme-related factors as principal causes for the nonlinear kinetics under our experimental conditions. The multiple sub-substrate hypothesis seems to be most relevant here. According to this hypothesis the bulk cellulose contains several sub-substrates that differ in their susceptibility to enzymatic attack and the hydrolysis of each sub-substrate can be regarded to proceed according to a pseudo-first order reaction (Sattler et al., 1989; Nidetzky and Steiner, 1993). Eq. (3) that accounts for the two sub-substrates was sufficient to describe our experiments. The parameters a and c in Eq. (3) are amplitude factors and represents the total amount of given sub-substrate in cellulose (as a % of total mass of the pulp). Parameters b and d are pseudo-first order rate constants. It is well known that endoglucanases are most active against non-crystalline or amorphous cellulose and exhibit a very low activity on cellulose crystals. From the kinetics we can define that a and c represents

394


Fig. 5. Release of anthranilic acid saccharide conjugates (AAsaccharide) in hydrolysis of fluorescence end-labelled softwood pulp upon hydrolysis with Aspergillus endoglucanase N476. (A) Fluorescence labelled Sp7 (AA-Sp7, 5.0 g l−1 ) was incubated with 3.3 ␮M N476 in 50 mM sodium acetate buffer, pH 5.0 at 53 ◦ C. (B) The ratio of released saccharide (as cellobiose) to that of AA-saccharide (mol/mol) from the experiment in A. Error bars are from the two parallel experiments.

the amounts of the “fast” and “slow” substrate, respectively. This opens the question about the nature of the sub-structures that correspond to the parameters a and c. For this we used specific labelling of the reducing end groups on pulp with the fluorescent tag AA. Similar to observations for cellobiohydrolase Cel7A from T. reesei on AA labelled bacterial cellulose (Kipper et al., 2005), the AA-labelling of the pulp reduced the activity of the N476 by about 30% compared to the unlabelled substrate. Fig. 5A shows that there is a sudden release of AA-saccharide conjugates from the AA-labelled Sp7 softwood pulp upon hydrolysis with N476 at the initial stage of hydrolysis followed by a virtually linear stage. Thus, in the first rapid stage, endoglucanase releases mostly the end groups of the cellulose chain. This has been speculated previously that some of the end-groups of the cellulose chains may not always be tightly asso-

ciated with the bulk of the cellulose crystal and form a population of so-called “loose” ends (Medve et al., 1998). As our results suggest these “loose” ends is probably the most accessible substrate for endoglucanase. Fig. 5B shows the ratio of released cellobiose to that of the labelled end groups. This ratio increases sharply with time as the labelled-end groups of the “loose” chain-end fraction are depleted (note that in Fig. 5B the time scale is logarithmic). A possible contribution of the hydrolysis of xyloglucans to the values of released total saccharide and, hence to a and c from Eq. (3) can be excluded based on the very low xylanase activity of the enzymes used in this study (see Section 3.1). SEM observations of Cel5A treated pulps were also made to reveal any morphological change in the fibres upon enzymatic hydrolysis. The treatment of finesfree fibres with 1 ␮M Cel5A in 50 mM sodium acetate buffer, pH 5.0 at 40 ◦ C for 24 h. caused a slight, but clear increase in the number of cut-off fibres compared to the untreated samples (Fig. 6). This has been suggested by Oksanen et al. (1997) as one likely effect of Cel5A (formerly EGII) treatment. The fines fractions looked fuzzier after Cel5A treatment (data not shown). Larger magnification revealed a fine network covering larger particles. No peeling of the S2-wall and more extensive degradation, which is common with “complete” cellulase systems was observed. The “cuts” are

Fig. 6. SEM micrograph of T. reesei Cel5A treated softwood pulp. Sp2 (2.0 g l−1 ) was incubated with 1 ␮M Cel5A in 50 mM sodium acetate buffer, pH 5.0 at 40◦ for 24 h. An endoglucanase treatment caused an increase in the number of cut-off fibres (some of them are indicated by the arrows).


mainly rather sharp but also more blunt ones appear. The SEM observations suggest that the degradation with Cel5A takes place mainly at the outer fibre surface. However, the interior surface of the fibres is most likely also accessible to the enzymes via damages in the fibre wall and via hollow pits. Thus, from our results we suggest that the fast substrate for endoglucanase is the population of “loose” chain ends on the fibre surface. The length of these chain ends is apparently important for their accessibility to the enzyme. Furthermore, the length of “loose” chain ends is not likely to be even but can be rather expected to have a certain distribution. The enzymatic degradation of this heterogeneous population of “loose” chain ends may well be revealed as a gradual loss in substrate reactivity consistent with the apparent fractal-like kinetics described by Eq. (2).

395

3.5. Correlation between kinetic and paper/pulp mechanical parameters Kinetic parameters obtained from the non-linear regression analysis of the degradation of softwood pulps Sp4, Sp5 and Sp6 and hardwood pulps Hp3 and Hp4 with N476, according to Eq. (3) were compared to the mechanical parameters of the pulps and papers derived from corresponding pulps. The following pulp and paper parameters were assessed for correlation with the kinetic parameters: z-strength, zero-span, tensile index (TI), tear index, RBA (paper parameters) and wet fibre flexibility, fibre length, fibre width, fibre shape factor and fibre coarseness (pulp parameters). Some of the values of kinetic and mechanical parameters are listed in Table 1. The best correlation was found between the ratio of a/c (the ratio of the amount of fast and slow substrate) from the kinetics and the TI, RBA

Fig. 7. Correlation between the parameters from the enzymatic hydrolysis of pulp and the properties of the paper. (A–C) Correlation between the ratio of the amount of fast substrate to the amount of slow substrate, a/c from Eq. (3) and the tensile index, relative bonded area and z-strength of the paper derived from corresponding pulp. (D) Correlation between the tensile index and apparent first order rate constant from fractal-like kinetics, k in Eq. (2). Original data are represented in Fig. 4 and Table 1. Softwood pulp (♦), hardwood pulp ( ). Error bars are from the two parallel experiments.

396


and z-strength of the paper (Fig. 7A–C). Data in Fig. 7 are from the experiments with the fibre fractions. In the case of non-fractionated pulps (together with fines) the correlation trends were the same as in Fig. 7 but the results were more scattered. This may be attributed to the slight background scattering in the experiments with the non-fractionated pulps as some fines can occasionally pass through the 0.45 ␮m syringe filter. Results from correlating a, c and a + c to the above paper parameters followed the same trend as from correlating a/c and in no case were contradictory, although the results were also more scattered. We did not find any correlation between the pseudo-first order rate constants b and d from Eq. (3) and mechanical parameters. All of the mechanical parameters showing correlation with kinetic parameters were somehow related to fibre–fibre interactions. At the same time parameters describing intrinsic or geometrical properties of the fibre did not correlate with the kinetic parameters. TI, RBA and z-strength are all the parameters related to fibre–fibre interactions. The only at least partly fibre–fibre interaction dependent parameter that did not show a correlation with the kinetic parameters was the tear index. The kinetic parameters and their combinations that were found to be in correlation with paper parameters are all representative for the amount of non-crystalline cellulose. Thus, the fibre–fibre interaction can be suggested to be mediated mainly by the non-crystalline fraction of cellulose (“loose” chain ends and amorphous cellulose on the fibre surface). All these results are in accordance with the demonstrated coupling between Cel5A pre-treatment of pulp and negative impact on strength properties (Oksanen et al., 1997). Thus, the correlation between the parameters of enzyme kinetics and mechanical parameters of the paper derived from corresponding pulp should allow an easy raw material evaluation in paper making process.

Acknowledgements This work was carried out within the framework of the Wood Ultrastructure Research Centre (WURC), a VINNOVA competence centre based at the Swedish University of Agricultural Sciences (SLU), Uppsala, Sweden. P.V. was supported by the Estonian Science Foundation (grant no. 5848). Mao Bofeng, Anders

Moberg and Gunilla Söderstam at the Pulp Competence Centre, StoraEnso, Karlstad, Sweden is acknowledged for providing us with the pulps along with their technical parameters as well as for many valuable discussions on theoretical and practical aspects on the experiments. Novozyme A/S, Denmark is acknowledged for kindly providing us with N476. Dr. Göran Pettersson from Uppsala University and Dr. Geoffrey Daniel from Wood Science, SLU are acknowledged for valuable discussions and proof-reading of the manuscript. Magnus Gäredal is acknowledged for the laboratory work with the AA labelling.

References Ander, P., Daniel, G., 2005. Characterization of industrial and laboratory pulp fibres using HCl, cellulase and FibreMaster analysis. Nord. Pulp Paper Res. J. 20, 115–120. Appel, R.D., Bairoch, A., Hochstrasser, D.F., 1994. A newgeneration of information retrieval tools for biologists - the example of the Expasy www-server. Trends Biochem. Sci. 19, 258–260. Bailey, C.J., 1989. Enzyme kinetics of cellulose hydrolysis. Biochem. J. 262, 1001. Bhikhabhai, R., Johansson, G., Pettersson, G., 1984. Isolation of cellulolytic enzymes from Trichoderma reseei QM 9414. J. Appl. Biochem. 6, 336–345. Buchert, J., Carlsson, G., Viikari, L., Ström, G., 1996a. Surface characterisation of unbleached kraft pulps by enzymatic peeling and ESCA. Holzforschung 50, 69–74. Buchert, J., Suurnäkki, A., Tenkanen, M., Viikari, L., 1996b. Enzymatic characterisation of pulps. ACS Symp. Ser. 655, 38–43. Carrard, G., Koivula, A., Söderlund, H., Beguin, P., 2000. Cellulosebinding domains promote hydrolysis of different sites on crystalline cellulose. Proc. Natl. Acad. Sci. U.S.A. 97, 10342–10347. Duchesne, I., Daniel, G., 1999. The ultrastructure of wood fibre surfaces as shown by a variety of microscopical methods: a review. Nord. Pulp Paper Res. J. 14, 129–139. Duchesne, I., Daniel, G., 2000. Changes in surface ultrastructure of Norway spruce fibres during kraft pulping: visualisation by field emission-SEM. Nord. Pulp Paper Res. J. 15, 54–61. Grethlein, H.E., 1985. The effect of pore size distribution on the rate of enzymatic hydrolysis of cellulosic substrates. BioTechnology 2, 155–160. Gruno, M., Väljamäe, P., Pettersson, G., Johansson, G., 2004. Inhibition of the Trichoderma reesei cellulases by cellobiose is strongly dependent on the nature of the substrate. Biotechnol. Bioeng. 86, 503–511. Huang, Z., Prickett, T., Potts, M., Helm, R.F., 2000. The use of 2aminobenzoic acid tag for oligosaccharide gel electrophoresis. Carbohydr. Res. 328, 77–83. Hörmann, H., Gollwitzer, R., 1962. Bestimmung von hexosen in tryptophan-haltigen eiwisserkörpern. Ann. Chem. 655, 178–188.

L. Hildén et al. / Journal of Biotechnology 118 (2005) 386–397 Kipper, K., Väljamäe, P., Johansson, G., 2005. Processive action of cellobiohydrolase Cel7A from Trichoderma reesei is revealed as “burst” kinetics on fluorescent polymeric model substrates. Biochem. J. 385, 527–535. Laine, J., Buchert, J., Viikari, L., Stenius, P., 1996. Characterization of unbleached kraft pulps by enzymatic treatment, potentiometric titration, and polyelectrolyte adsorption. Holzforschung 50, 208–214. Lawoko, M., Nutt, A., Henriksson, H., Gellerstedt, G., Henriksson, G., 2000. Hemicellulase activity of aerobic fungal cellulases. Holzforschung 54, 497–500. Lynd, R.L., Weimer, P.J., van Zyl, W.H., Pretorius, I.S., 2002. Microbial cellulose utilization: fundamentals and biotechnology. Microbiol. Mol. Biol. Rev. 66, 506–577. Macarron, R., Acebal, C., Castillon, M.P., Dominguez, J.M., Mata, I., Pettersson, G., Tomme, P., Claeyssens, M., 1993. Mode of action of endoglucanase III from Trichoderma reesei. Biochem. J. 289, 867–873. Mansfield, S.D., Mooney, C., Saddler, J.N., 1999. Substrate and enzyme characteristics that limit cellulose hydrolysis. Biotechnol. Prog. 15, 804–816. Medve, J., Karlsson, J., Lee, D., Tjerneld, F., 1998. Hydrolysis of microcrystalline cellulose by cellobiohydrolase I and endoglucanase II from Trichoderma reesei: adsorption, sugar production pattern, and synergism of the enzymes. Biotechnol. Bioeng. 59, 621–634. Miller, G.L., 1959. Use of dinitrosalicylic acid reagent for determination of reducing sugar. Anal. Chem. 3 (31), 426–428. Nidetzky, B., Steiner, W., 1993. A new approach for modelling cellulase-cellulose adsorption and the kinetics of the enzymatic hydrolysis of microcrystalline cellulose. Biotechnol. Bioeng. 42, 469–479. Nidetzky, B., Steiner, W., Hayn, M., Claeyssens, M., 1994. Cellulose hydrolysis by the cellulases from Trichoderma reesei: a new model for synergistic interaction. Biochem. J. 298, 705–710. Nyholm, K., Ander, P., Bardage, S., Daniel, G., 2001. Dislocations in pulp fibres, their origin, characteristics and importance: a review. Nord. Pulp Paper Res. J. 16, 376–384.

397

Oksanen, T., Pere, J., Buchert, J., Viikari, L., 1997. The effect of Trichoderma reesei cellulases and hemicellulases on the paper technical properties of never-dried bleached kraft pulp. Cellulose 4, 329–339. Saloheimo, M., Lehtovaara, P., Penttilä, M., Teeri, T.T., St˚ahlberg, J., Johansson, G., Pettersson, G., Claeyssens, M., Tomme, P., Knowles, J.K.C., 1988. EG III, a new endoglucanase from Trichoderma reesei: the characterization of both gene and enzyme. Gene 63, 11–21. Sato, K., Sato, K., Okubo, A., Yamazaki, S., 1998. Optimization of derivatization with 2-aminobenzoic acid for determination of monosaccharide composition by capillary electrophoresis. Anal. Biochem. 262, 195–197. Sattler, W., Esterbauer, H., Glatter, O., Steiner, W., 1989. The effect of enzyme concentration on the rate of the hydrolysis of cellulose. Biotechnol. Bioeng. 33, 1221–1234. Sjöström, E., 1981. Wood chemistry, fundamentals and applications. Academic Press, Inc., New York. Tanaka, M., Ikesaka, M., Matsuno, R., Converse, A.O., 1988. Effect of pore size in substrate and diffusion of enzyme on hydrolysis of cellulosic materials with cellulases. Biotechnol. Bioeng. 32, 698–706. Väljamäe, P., Kipper, K., Pettersson, G., Johansson, G., 2003. Synergistic cellulose hydrolysis can be described in terms of fractal-like kinetics. Biotechnol. Bioeng. 84, 254– 257. Väljamäe, P., Sild, V., Nutt, A., Pettersson, G., Johansson, G., 1999. Acid hydrolysis of bacterial cellulose reveals different modes of synergistic action between cellobiohydrolase I and endoglucanase I. Eur. J. Biochem. 266, 327–334. Yang, J.-L., Pettersson, B., Eriksson, K-E., 1988. Development of bioassays for the characterization of pulp fibre surfaces. I. Characterization of various mechanical pulp fibre surfaces by specific cellulolytic enzymes. Nord. Pulp Paper Res. J. 42, 19–25. Zhang, S., Wolfgang, D.E., Wilson, D.B., 1999. Substrate heterogeneity causes the non-linear kinetics of insoluble cellulose hydrolysis. Biotechnol. Bioeng. 66, 35–41.


On-line monitoring of yeast cell growth by impedance spectroscopy A. Soley a,∗ , M. Lecina a , X. Gámez a , J.J. Cairó a , P. Riu b , X. Rosell b , R. Bragós b , F. Gòdia a b

a Departament d’Enginyeria Qu´ımica, Escola T` ecnica Superior d’Enginyeria, Universitat Autònoma de Barcelona, 08193 Cerdanyola del Vallès, Barcelona, Spain Departament d’Enginyeria Electrònica, Universitat Politècnica de Catalunya, 08034 Barcelona, Spain

Received 2 December 2004; received in revised form 23 May 2005; accepted 30 May 2005

Abstract The application of impedance spectroscopy to estimate on-line cell concentration was studied. The estimation was based on the relative variation between electrical impedance measured at low (10 kHz) and high frequencies (10 MHz). Studies were carried out to characterise the influence of changes in physical and chemical parameters on the impedance measurement. Two different possibilities to perform on-line measurements were tested: a simple set-up, based on an in situ probe, gave good results but was not suitable for high agitation and aeration rates. An ex situ flow-through on-line measuring cell was used to overcome these problems, showing a better performance. The use of this set-up for the growth monitorisation of a Saccharomyces cerevisiae culture showed an efficient performance, having the correlation between estimated and measured S. cerevisiae a Pearson coefficient of 0.999. © 2005 Elsevier B.V. All rights reserved. Keywords: Impedance spectroscopy; Biomass estimator; On-line monitoring

1. Introduction In a great number of biotechnological processes the development of reliable, automated and optimised ∗

Corresponding author. Fax: +34 93 581 20 13. E-mail addresses: [email protected] (A. Soley), [email protected] (M. Lecina), [email protected] (J.J. Cairó), [email protected] (R. Bragós), [email protected] (F. Gòdia). 0168-1656/$ – see front matter © 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.jbiotec.2005.05.022

operation of bioreactors plays a very important role. In order to achieve this objective, an on-line measurement of the key variables of the process is required. In all the cases where microorganisms are used as biocatalysts, the cell concentration is clearly one of the parameters that should be determined, and particularly there’s a great interest in the determination of the living cells concentration. Without the establishment of reliable methods for real-time measurement of cell concentration, further

A. Soley et al. / Journal of Biotechnology 118 (2005) 398–405

advances in bioprocess monitoring and control will be limited. This situation has generated several studies that have tried different techniques to solve the problem. There are already some techniques available for online monitoring of biomass in bioreactors. However, it seems quite evident that a universally applicable and robust technique has not yet been developed as many of the studied methods show drawbacks and limitations. A review of them has been described, including for bacterial, yeast or mammalian cell culture processes in the literature (Sonnleitner et al., 1992; Konstantinov et al., 1994; Ducommun et al., 2000). Among the systems for biomass measurements, the dielectric analysis of cell suspensions is an attractive technique in order to estimate the cell concentration because it is a real-time and automated monitoring system, and applicable to complex media (Kell et al., 1990). The radio frequency electrical properties of cell suspensions are a direct and monotonic function of the radius and volume fraction of the particles constituting the suspended phase (Harris et al., 1987). The methods based on electrical measurements use the dependence of macroscopic variables of a material as conductance (G) and capacitance (C) or the intrinsic electrical properties as conductivity (σ) and permittivity (ε) with the cell volume fraction (P) at one or two frequency points. Some published results obtained with different types of biological cells have demonstrated a close relationship between the dielectric properties of suspended cells, like capacitance, and its biomass concentration (Fehrenbach et al., 1992; Davey et al., 1992; Matanguihan et al., 1994; November and Van Impe, 2000; Cannizzaro et al., 2003). This method gives a signal that is linear up to very high biomass levels (Harris and Kell, 1983) and provides systematic information on the conditions of yeast cells in culture (Asami and Yonezawa, 1995). Others authors (Mishima et al., 1991; Austin et al., 1994) have reported a relationship between the capacitance changes in the stationary growth phase and cell viability, but a completely quantitative description of it was not possible. The dependence of dielectric permittivity (ε) with the volume fraction (P) is described by the Maxwell model (Foster and Schwan, 1989). In practical measurements, ε is determined from the capacitance (C) through a geometrical factor called cell constant. However, neither ε nor C can be directly measured. They are determined from calculations after electrical voltage

399

Fig. 1. Effect of the yeast cell volume fraction on the impedance module. Cell volume fractions are expressed as cell concentrations (in g l−1 ). Impedance relaxations are shown between 10 kHz and 10 MHz.

and current measurements, and assuming an electrical model for the suspension that neglects the ionic conduction through the medium. This conduction mechanism becomes an interference in practical measurements. In order to avoid the inaccuracies derived from the cell constant measurements, the groups presenting the paper have worked in the development of biomass density estimators directly derived from electrical impedance measurements (Bragós, 1997; Gámez et al., 1996). The defined ratiometric estimator for biomass measures passive characteristics of cell suspension regions including all conduction mechanisms. The most direct measurement which can be calculated from voltage (V) and current (I) measurements is the electrical impedance (Z) that is the ratio between V and I. At low frequencies the impedance of a cell suspension is maximum because cell membranes isolate the intracellular space, thus decreasing the conductor volume. As frequency increases, the displacement current across membranes becomes higher, membranes no longer isolate intracellular space and the impedance of the suspension decreases towards its minimum value (Fig. 1). This behaviour, known as impedance relaxation, can be characterised with three parameters: the low frequency (Z0 ) and the high frequency (Z∞ ) impedance values, and the central relaxation frequency (fc ). fc is related to cell size; the smaller the cells, the higher fc . Impedance values at frequencies much lower or much higher than fc are related to conductivity of

400


intra and extracellular milieu and the ratio between intra and extracellular volumes. This behaviour can be modelled (Eq. (1)) adapting the Cole–Cole model for the complex permittivity of dielectric materials (Cole and Cole, 1941) to the electrical impedance of cell suspensions. The α parameter describes the dispersion of the cell sizes and shapes and is zero for equally sized spherical cells. Z(f ) = Z∞ +

Z0 − Z∞ 1−α 1 + j ffc

(1)

Apparently, a single frequency measurement made at low frequency could provide a good P estimation. However, this value is strongly influenced by conductivity changes due to temperature or ionic concentration variations. Because of this, estimators based on relative variations of impedance are expected to show a better performance. The biomass density estimator used in the present work, and which we call E2 , is defined as the relative variation between the low frequency (LF) and high frequency (HF) values of impedance magnitude with respect to the LF value (Eq. (2)). |Z(HF)| E2 = 100 1 − (%) (2) |Z(LF)| where E2 is the biomass estimator, |Z(LF)| the impedance magnitude at f fc and |Z(HF)| is the impedance magnitude at f fc . The biomass estimator, E2 , is a function of the cell size (R), membrane capacitance (Cm ), volume fraction (P) and intra and extracellular conductivities (σ i and σ e ). Assuming that P < 0.2 (v/v) and that the intra and extracellular conductivities are similar, it can be demonstrated that E2 is linearly dependant with P and that does not depend on the other parameters (Bragós, 1997). In practical conditions, R, Cm and σ i are constants for a given microorganism but σ e can present variations. To demonstrate the practical applicability of the impedance signal as an indicator of cell concentration in bioreactors, two different on-line measurement systems were developed, one in situ and another ex situ. In the present paper these systems are evaluated in terms of their applicability to yeast cells concentration measurement. The influence of temperature, ion concentration, aeration and agitation rates on the impedance measurement are discussed for both.

2. Materials and methods 2.1. Microorganism and culture conditions A wild type of the yeast Saccharomyces cerevisiae was used in the experiments to test the developed measurement systems. The yeast was grown on YPEG2% medium (w/v): yeast extract 1%, bacterial peptone 1% and glucose 2%. The pH was adjusted to 6.5 with HCl acid and the medium sterilised by autoclaving at 121 ◦ C for 20 min. A 100 ml culture incubated overnight at 37 ◦ C in YPEG2% medium on a shaker was used to inoculate the bioreactor. Bioreactor cultures were carried out in a Biostat B (Braun Biotech) equipped with a 3-l glass vessel. The bioreactor was filled with 2-l of sterile YPEG2% medium. Temperature was maintained at 37 ◦ C, agitation and aeration rates were set at 1000 rpm and 2 vvm, respectively. Neither pH nor pO2 control were performed. Samples for off-line analyses were taken aseptically every hour during the batch culture. 2.2. Off-line analyses Optical densities (OD) were measured at 550 nm using a cell with a path length of 10 mm in a Phillips spectrophotometer PU8620. Samples with optical density values greater than 0.7 were diluted. A calibration curve was used to determine the dry biomass concentration. To determine the biomass concentration by dry weight (DW), 10 ml samples were centrifuged (6000 rpm for 10 min at 4 ◦ C). The pellet was resuspended in a NaCl solution (0.9%, w/v), filtered trough a previously dried and weighted Whatman grade GF/F glass microfiber filter, washed and dried at 105 ◦ C to constant weight. In order to reduce the experimental error, duplicate measurements were performed. 2.3. Measurement system Impedance spectrum measurements in the 0.01–10 MHz range were performed using a HewlettPackard 4192A impedance analyser controlled by a Compaq Contura 410C M350 computer (486/16) via an interface bus conforming to the IEEE-488 standard. The data acquisition and processing program was


401

Fig. 2. Schematic diagram of an on-line monitored batch culture with the in situ measurement system: (1) personal computer, (2) HP4192A impedance analyser, (3) HPE3631A power supply, (4) front-end, (5) immersed probe, and (6) Biostat B bioreactor.

written by the authors in LabWindows CVI (National Instruments). The analyser was connected to the electrodes through a remote front-end to reduce the errors caused by the combined effect of the electrodes and the coaxial cables (Gersing, 1991). The four-electrode method was used to eliminate the artifacts due to the phenomenon of electrode polarisation (Harris et al., 1987; Rosell et al., 1988). Two electrodes were used as current injectors, whereas the other two were used to measure the voltage drop in the solution. On-line impedance measurements of S. cerevisiae cultures were performed using two alternative strategies: (a) The in situ sensor (Fig. 2) consisted of a tubular shaped probe made in Teflon with two stainless steel coaxial compound electrodes (10 and 4 mm in diameter, respectively), which was immersed

into the bioreactor liquid through a standard 25 mm (Ingold-type) port in the lid. (b) The ex situ sensor setup is depicted in Fig. 3. The culture broth was pumped from the bioreactor to a degassing bottle at flow of 108 ml min−1 using a peristaltic pump. After the broth degassing, the cell suspension enters a flow-through chamber made with a glass cylinder, whose working volume was 62 ml, with four stainless steel rod electrodes placed at the bottom. Each electrode measured 2 mm in diameter and 47 mm in length and was positioned parallel to each other. Temperature in the chamber was controlled by means of an external jacket with thermostatised water at 37.1 ◦ C. After each impedance measurement, the culture broth was pumped with the peristaltic pump from the flow-through chamber back to the bioreactor.

Fig. 3. Experimental set-up of a batch culture with the ex situ measurement system: (1) Masterflex peristaltic pump, (2) degassing bottle, (3) flow-through chamber, (4) thermostatised bath, (5) personal computer, (6) HP4194A impedance analyser, (7) front-end, (8) HPE3631A power supply, and (9) Biostat B bioreactor.

402


Both probes were steam sterilised together with the bioreactor. 2.4. Data analysis Cell suspensions were scanned at 31 frequency points in the frequency range, chosen with the aim of obtaining an even spacing on a log frequency scale. To reduce the effect of temporary variations of the impedance, ten successive frequency sweeps were averaged for each measure and recorded every 3 min. Biomass concentration for each yeast sample was obtained with estimator E2 using low and high frequencies: 10 kHz and 10 MHz, respectively. The low frequency was chosen to avoid electrode induced problems and the high frequency due to the limitations of the measurement system. Before addition of the inoculum, the impedance of the cell-free medium at culture conditions of temperature and pH was measured and the frequency sweep was used to perform a first-order on-line calibration of the growth relaxation curves. However, to be able to measure very low cell volume fractions, the standard gain and zero error correction method is not enough. We have adapted a method, based in a general transducer calibration formula (Bolk, 1985), which corrects non-linear load dependent systematic errors using three references measured at each frequency. The calibration formula is: Zc = C0 (f ) +

C1 (f )Zm (f ) 1 + C2 (f )Zm (f )

varying NaCl concentrations on the impedance were evaluated. Also, the influences of the agitation and aeration rate were performed with 2-l of 0.9% (w/v) NaCl solution while the temperature was maintained at 37 ◦ C.

3. Results 3.1. Results with the in situ probe 3.1.1. Effects of chemical and physical parameters on the measurement The temperature influence has been tested in the bioreactor, in a range from 33 to 39 ◦ C. The results show that at 10 kHz this variation is −0.29 ◦ C−1 (r = 0.999), meaning that 1 ◦ C variation would correspond to a 2% change on the impedance module. This effect is due to the increased mobility of the ions at higher temperature, which means a decrease of the resistance to current flow, that is, the impedance. Given the ratiometric nature of the E2 estimator, its temperature dependence is lower than 0.5% ◦ C−1 . The effect of aeration and agitation in the bioreactor is shown in Fig. 4. Agitation rate is varied from 0 to 600 rpm in two different situations: absence of aeration and 1 vvm aeration. In the first case, the influ-

(3)

where Zc is the corrected impedance (), Zm the measured impedance (), f the frequency (Hz) and C0 , C1 , and C2 are the correction coefficients. Correction coefficients C0 , C1 , and C2 were calculated at each frequency measuring three saline reference solutions of known conductivity. The measurement system together with the calibration method allowed measurement of 0.1% relaxation in the whole bandwidth. 2.5. Effects of chemical and physical parameters on the measurement The influence of temperature, ion concentrations, agitation and aeration rates were studied using the Biostat B bioreactor and the immersed probe. Effects of

Fig. 4. The influence of the agitation rate in a 0.9% (w/v) NaCl solution on the impedance values at 10 kHz with 1 vvm of aeration (triangles) and without air supply (squares).


ence of agitation is negligible up to 500 rpm. However, in the second case, the influence of agitation on the impedance measurements is notorious from 200 rpm. These results show that when the gas hold-up is increased in the system, as a consequence of aeration and agitation, the impedance at 10 kHz will also increase. A similar effect has been reported by other authors studying systems based on capacitance measurements (Mishima et al., 1991; Fehrenbach et al., 1992; Davey et al., 1993; November and Van Impe, 2000), in this case, observing a decrease of their values as a consequence of a higher gas hold-up. The variation of the value of the impedance module with gas hold-up can be explained by the fact that air bubbles are non-conductive of the electrical current due to their dielectric nature. For experimental conditions generating low air hold-up, such as for low aeration rates, this effect is negligible. However, when the gas hold-up is high, such as for high aeration rates, a negative effect is observed in the relaxation curve. This is the consequence of replacing a fraction of the volume previously occupied by a polarizable and conductive material (such as cells and culture medium) with a nonpolarizable and non-conductive material (such as air bubbles). The effect of this change is an increase of the impedance values. In broth with moderate cell concentrations, the effect of the changes of impedance due to the cell displacement by the gas is an important source of error for an accurate measurement of cell concentration. Moreover, the application of the impedance spectroscopy to biomass determination requires that the β-relaxation curves do not present an excessive noise for its correct treatment. Nevertheless, the noise generated by the gas bubbles on the measurement could be theoretically reduced by using very quick spike elimination algorithms (Sonnleitner et al., 1992). Effect of changes in ionic concentration: impedance measurement depends on the medium conductivity (σ e ) and intracellular conductivity (σ i ). The extracellular medium conductivity might change but the intracellular ionic environment remains almost constant (Asami et al., 1976). Ion concentration changes in the external medium affect not only to the absolute value of impedance but also to the central frequency and relaxation magnitude. However, the sensibility of estimator E2 to these changes is lower because it is a ratio between

403

impedances. Additionally, the changes of σ e in a culture of S. cerevisiae are low, since the concentration of salts remains almost invariable during the culture, and the main substrates and products of the metabolism (glucose, etanol) are not ionized. Therefore, the small changes in σ e observed can be neglected. In case bigger changes were observed, they could be corrected by measuring the impedance of yeast-free medium (Bragós, 1997). After these experiments characterising the behaviour of the impedance measurement as a function of the yeast culture main variables, it was possible to define the conditions that would be required to test its performance along a growth curve. These are: - Constant temperature. - pH control by acid/base addition was avoided to minimise changes in the medium conductivity. - As the culture conditions (1000 rpm, 2 vvm) generate a high hold-up in the reactor, the measurements are to be performed at an advantageous situation. Every 30 min the aeration is stopped, and the stirring reduced to 200 rpm. After 1 min, the measure is taken, and the normal conditions are restored immediately. This procedure was preferred to other possibilities, such as the correction of the measurement by the value of the gas hold-up (Mishima et al., 1991). 3.1.2. On-line measurements The results obtained according to the previous procedure, in a batch culture of S. cerevisiae cells using the in situ probe in the bioreactor, are presented in Fig. 5. The response follows the exponential growth phase, as well as the transition to the stationary phase. However, the results obtained for the lag phase are quite unstable (data not shown), due to the fact that cell concentration level is under the detection threshold for this measuring system, 0.5 g l−1 . As a general trend, these results show that, despite the need to stop aeration to proceed to the measurements, the probe shows a good profile, as it allows following the cell growth curve accurately. 3.2. Results with the ex situ system The in situ probe measurement would present important drawbacks to be applied to bioprocesses where maintaining constant aeration and agitation conditions is needed. To avoid these constraints, an ex situ

404


Fig. 5. On-line monitoring of a S. cerevisiae culture using the in situ probe. The estimation method based on impedance measurements (E2 estimator) (squares, dashed line) is compared to the measurements of dry weight (circles, solid line).

measurement system is proposed as an alternative. The measurement cell is non-agitated mechanically, and before the sample reaches the cell, it flows through a gas–liquid separation chamber, where the gas is eliminated. The system allows a continuous measurement of the cell suspension impedance, free of air bubbles and agitation effects, without any constraint for the agitation and aeration conditions in the bioreactor. In addition, the flow-through chamber for the ex situ measurement operate at constant temperature, and the liquid flow-rate allows an appropriate homogenisation in the chamber, thus avoiding any cell sedimentation, as it was confirmed by RTD experiments (data not shown). The impedance measurements obtained with this ex situ flow-through system during a batch growth of S. cerevisiae are presented in Fig. 6, together with the data obtained by optical density and dry weight measurements. More precisely, the data plotted in Fig. 6 corresponds to the E2 estimator obtained from the impedance measurements as described previously. Comparing the results obtained with the three techniques, it can be said that the use of the E2 estimator for on-line monitorisation of S. cerevisiae cultures shows an excellent performance, comparable to the obtained with other methods. The correlation observed between E2 and dry weight measurements (Fig. 7) has an excellent linearity (r = 0.999), with a slope of 0.33. The results in Fig. 6 show that in the lag phase the E2 value does not change significantly and the

Fig. 6. On-line monitoring of a S. cerevisiae using the ex situ set-up. In this case, the estimation method based on impedance measurements (E2 estimator) (solid line) is compared to the measurements of capacitance at 1 MHz (dashed line), dry weight (squares, solid line) and optical density at 550 nm (circles, solid line).

first detectable measurement over the base line noise (0.28%) correspond to a minimal detectable cell concentration of 0.3 g l−1 . During the exponential and stationary phase the results obtained show a high correlation coefficient versus dry weight measurements. As the stationary phase begins the estimator value decreases, probably due to the growth reduction observed after 6 h of culture. This situation is confirmed by the on-line pH measurements (data not shown) that showed no more hydrogen ion production after this

Fig. 7. Linear correlation between the impedance biomass estimator (E2 estimator) and the concentration of S. cerevisiae in dry weight (g l−1 ) during the batch culture performed.


period. This phenomenon has been well discussed for several authors (Austin et al., 1994; Matanguihan et al., 1994).

4. Discussion The impedance spectroscopy can be successfully applied to the measurement of cell concentration, particularly for the case of S. cerevisiae, and with the estimator E2 proposed in the present work. The application of this technique to on-line measurement of cell concentration in bioreactors has certain limitations when the measurement is done with an in situ probe. Indeed, in this case the technique could only be applied successfully in systems with low gas hold-up and agitation. For agitated and aerated systems, the system based on an external flow-through measurement cell proved better performance.

Acknowledgments The present work has been developed in the framework of the Centre de Referencia en Biotecnologia (Generalitat de Catalunya). Nuevas Tecnologias Espaciales, S.A. (Barcelona) provided the in situ probe used. This work has been supported by a PETRI type contract from Scientific Research of the Ministry of Education, Science and Culture, Spain.

References Asami, K., Hanai, T., Koizumi, N., 1976. Dielectric properties of yeast cells. J. Membr. Biol. 28 (2–3), 169–180. Asami, K., Yonezawa, T., 1995. Dielectric analysis of yeast cell growth. Biochim. Biophys. Acta 1245, 99–105. Austin, G.D., Watson, R.W.J., D’Amore, T., 1994. Studies of online viable yeast biomass with a capacitance. Biomass Monit. Biotechnol. Bioeng. 43, 337–341. Bolk, W.T., 1985. A general linearising method for transducers. J. Phys. E. Sci. Instrum. 18, 61–64. Bragós, R., 1997. Contribution to the characterization of biological systems using electrical impedance spectroscopy. Ph.D. Dissertation. Polithecnic University of Catalonia, Barcelona. Cannizzaro, C., Gügerli, R., Marison, I., von Stockar, U., 2003. Online biomass monitoring of CHO perfusion culture with scanning dielectric spectroscopy. Biotechnol. Bioeng. 84 (5), 597–610.

405

Cole, K.S., Cole, R.H., 1941. Dispersion and absorption in dielectrics. I. Alternating current characteristics. J. Chem. Phys. 9, 341–351. Davey, C.L., Davey, H.M., Kell, D.B., 1992. On the dielectric properties of cell suspensions at high volume fractions. Bioelectrochem. Bioenerg. 28, 319–340. Davey, C.L., Davey, H.M., Kell, D.B., Todd, R.W., 1993. Introduction to the dielectric estimation of cellular biomass in real time, with special emphasis on measurements at high volume fractions. Anal. Chim. Acta 279, 155–161. Ducommun, P., Bolzonella, I., Rhiel, M., Pujeaud, P., von Stockar, U., 2000. On-line determination of animal cell concentration. Biotechnol. Bioeng. 72 (5), 515–522. Fehrenbach, R., Comberbach, M., Pêtre, J.O., 1992. On-line biomass monitoring by capacitance measurement. J. Biotechnol. 23, 303–314. Foster, K.R., Schwan, H.P., 1989. Dielectric properties of tissues and biological materials: a critical review. CRC Crit. Rev. Biomed. Eng. 17, 25–104. Gámez, X., Sabès, M., Bragós, R., Riu, P., Cairó, J.J., Gòdia, F., 1996. Biomass monitoring using multifrequency impedance measurements: relationship between particle size and electrical impedance. In: Proceedings of the First European Symposium on Biochemical Engineering Science (ESBES), Dublin, Ireland (abstract 94). Gersing, E., 1991. Measurement of electrical impedance in organsmeasuring equipment for research and clinical applications. Biomed. Technik 36, 6–11. Harris, C.M., Kell, D.B., 1983. The radio-frequency dielectric properties of yeast cells measured with a rapid, automated, frequencydomain dielectric spectrometer. Biolectrochem. Bioenerg. 11, 15–28. Harris, C.M., Todd, R.W., Bungard, S.J., Lovitt, R.W., Morris, J.G., Kell, D.B., 1987. Dielectric permittivity of microbial suspensions at radio frequencies: a novel method for the real-time estimation of microbial biomass. Enzyme Microb. Technol. 9, 181–186. Kell, D.B., Markx, G.H., Davey, C.L., Todd, R.W., 1990. Real-time monitoring of cellular biomass: methods and applications. Trends Anal. Chem. 9, 190–194. Konstantinov, K.B., Chuppa, S., Sajan, E., Tsai, Y., Yoon, S., Golini, F., 1994. Real-time biomass-concentration monitoring in animalcell cultures. Trends Biotechnol. 12, 324–333. Matanguihan, R.M., Konstantinov, K.B., Yoshida, T., 1994. Dielectric measurement to monitor the growth and the physiological states of biological cells. Bioprocess Eng. 11, 213–222. Mishima, K., Mimura, A., Takahara, Y., 1991. On-line monitoring of cell concentrations during yeast cultivation by dielectric measurements. J. Ferment. Bioeng. 72, 296–299. November, E.J., Van Impe, J.F., 2000. Evaluation of on-line viable measurements during fermentations of Candida utilis. Bioprocess Eng. 23, 473–477. Rosell, J., Viscasillas, J., Riu, P.J., Pallàs, R., Murphy, D., Rolfe, P., 1988. Tomografia de impedancia eléctrica. Mundo Electrónico 187, 132–187. Sonnleitner, B., Locher, G., Fietcher, A., 1992. Biomass determination. J. Biotechnol. 25, 5–22.


Bacterial inclusion bodies are cytotoxic in vivo in absence of functional chaperones DnaK or GroEL Nuria González-Montalbán, M. Mar Carrió, Sergi Cuatrecasas, Anna Ar´ıs, Antonio Villaverde ∗ Institut de Biotecnologia i de Biomedicina, Departament de Genètica i de Microbiologia, Universitat Autònoma de Barcelona, Bellaterra, 08193 Barcelona, Spain Received 13 January 2005; received in revised form 6 May 2005; accepted 12 May 2005

Abstract Cytotoxicity of cytoplasmic bacterial inclusion bodies has been explored in vivo in cells producing a model, misfolding-prone ␤-galactosidase fusion protein. The formation of such aggregates does not result in detectable toxicity on Escherichia coli producing cells. However, a deficiency in the main chaperones DnaK or GroEL but not in other components of the heat shock system such as the chaperone ClpA or the protease Lon, promotes a dramatic inhibition of cell growth. The role of DnaK and GroEL in minimizing toxicity of in vivo protein aggregation is discussed in the context of the conformational stress and the protein quality control system. © 2005 Elsevier B.V. All rights reserved. Keywords: Aggregation; Chaperones; DnaK; E. coli; GroEL; Protein folding

1. Introduction Abnormal protein aggregation and its interference with cell physiology are matters of rising interest because of the increasing incidence of amyloid-linked pathologies. By experimentally approaching this issue, exogenously added aggregates have proved to be deleterious to cultured mammalian cells (Bucciantini et al., 2002), specially when occurring as fibril precur∗ Corresponding author. Tel.: +34 935812148; fax: +34 935812011. E-mail address: [email protected] (A. Villaverde).


sor, amorphous aggregates. The organized clustering of misfolded protein as less toxic, true fibrils has been then regarded as a cell protective mechanism. Organized protein packaging could prevent soluble cell proteins from harmful interaction with unusually exposed hydrophobic patches. Conformational inactivation of cell proteins by these contacts, or permeabilization of cell membranes by channel-like aggregates could be independent although not necessarily exclusive deleterious mechanisms of protein aggregation (Olofsson et al., 2002; Stefani and Dobson, 2003). However, the detailed biology of aggregate-promoted toxicity remains still unsolved.

N. González-Montalbán et al. / Journal of Biotechnology 118 (2005) 406–412

In bacteria, multiple arms of the heat shock system interact with cell polypeptides for their folding and protection from aggregation (Arsene et al., 2000; Baneyx and Mujacic, 2004). Heat shock proteins also modulate proteolysis and aggregation (Dougan et al., 2002; Hengge and Bukau, 2003), and actively assist the removal of aggregated polypeptides from protein deposits for their refolding or eventual degradation (Carrió and Villaverde, 2003; Weibezahn et al., 2004). The activities of the heat shock proteins (namely chaperones and proteases) in monitoring protein quality might be conditioned by their substrate load, and either folding, aggregation or degradation could be favoured depending on their availability (Villaverde and Carrió, 2003). In this context, high-rate production of misfolding prone proteins often results in the formation of large and refractile protein deposits known as inclusion bodies. Recent studies have contributed to understanding the formation of bacterial inclusion bodies as a dynamic result of protein aggregation and disaggregation, and to identify bacterial responses to conformational stress (Hoffmann and Rinas, 2000; Carrió and Villaverde, 2001; Hunke and Betton, 2003; Ignatova and Gierasch, 2004). However, the possible cytotoxicity associated to inclusion bodies remains essentially unexplored. In this work, we demonstrate that protein aggregation in bacteria is toxic in vivo when the main chaperones DnaK and GroEL are absent. The cytoprotector role shared by these chaperones is discussed in the context of the complex quality control network.

407

E. coli ␤-galactosidase, and its derivative pJVP1LAC a N-terminal ␤-galactosidase fusion containing the VP1 capsid protein of foot-and-mouth disease virus (Corchero et al., 1996). Its production results into cytoplasmic inclusion bodies containing about the half of the total VP1LAC in the cell (Corchero et al., 1996). The expression of both genes is under the control of both lambda lytic promoters and a constitutively expressed temperature-sensitive CI repressor. 2.2. Media and growth conditions LB rich medium (Sambrook et al., 1989) was used in all the experiments, with ampicillin at 100 ␮g/ml for plasmid maintenance. Cultures were performed in shaker flasks at 28 ◦ C at 250 rpm until the OD550 reached 0.3, and then transferred to a pre-warmed bath at 42 ◦ C to trigger recombinant gene expression. For analysis of cell viability in absence of protein synthesis, chroramphenicol was added at 200 ␮g/ml 5 h after induction of gene expression and the culture temperature shifted down to 28 ◦ C as described (Carrió and Villaverde, 2001). Since the GroEL44 phenotype is only observed at 42 ◦ C, the strain BB4565 was not included in this experiment that was partially performed at 28 ◦ C. Viable cell counts were obtained in triplicate from at least two independent experiments, by platting appropriate dilutions of culture samples in LB plates plus 100 ␮g/ml ampicillin, that were incubated at 28 ◦ C for about 20 h. 2.3. Analysis of protein fractions

2. Materials and methods 2.1. Bacterial strains and plasmids The Escherichia coli strains used in this work were MC1061 (hsdR mcrB araD139 ∆(araABC-leu)7679 ∆lacX74 galU galK rpsL thi) and MC4100 (araD139 ∆(argF-lac) U169 rpsL150 relA1 flbB5301 deoC1 ptsF25 rbsR). MC4100 derivatives GroEL44 BB4565 (groEL44 zdj::Tn10 zje::kan), DnaK− JTG20 (dnaK thr::Tn10), Lon− BB2395 (∆lon146::miniTn10) (Tomoyasu et al., 2001) and ClpA− JGT4 (clpA::kan) (Thomas and Baneyx, 1998) were also employed. Further details of all these strains can be found elsewhere (Sambrook et al., 1989; Carrió and Villaverde, 2003). Plasmid pJCO46 encodes a soluble, pseudo-wild type

The presence of model proteins in the soluble and insoluble cell fractions was determined by Western blot analysis. Briefly, culture samples of 5 ml were centrifuged for 15 min at 12,000 × g and resuspended in 500 ␮l of Z buffer without ␤-mercaptoethanol (Miller, 1972), with one tablet of protease inhibitor cocktail (Roche, ref. 1,836,170) per 10 ml buffer. Samples jacketed in ice were sonicated for 2 min at 50 W under 0.5 s cycles, and centrifuged for 15 min at 12,000 × g. The inclusion bodies in the insoluble fraction were washed three times in lysis buffer plus 0.5% Triton as previously described (Carrió et al., 2000) and both soluble and insoluble fractions were resuspended in denaturing buffer (Laemmli, 1970). After boiling for 15 min, defined sample volumes were loaded

408


onto SDS-polyacrilamide gels. Immunoreactive bands within the linear range, were visualised by using an anti-␤-galactosidase polyclonal antibody and quantified on digitalized blots by using the Quantity One software of Bio Rad. Images of intracellular protein aggregates were obtained by optical microscopy as described (Carrió and Villaverde, 2003).

3. Results 3.1. Strain dependent toxicity of cytoplasmic inclusion bodies The eventual citotoxicity of bacterial inclusion bodies was evaluated in E. coli producing a misfoldingprone polypeptide, namely the fusion protein VP1LAC (Corchero et al., 1996). More than 50% of the produced VP1LAC polypeptides aggregated as cytoplasmic inclusion bodies, while the parental ␤galactosidase encoded by pJCO46 remained in the soluble cell fraction, and no aggregates were then observed (Figs. 1 and 2). Cell viability was explored by analysing viable cell counts in two commonly used laboratory strains (wild-type for the heat-shock system) for the production of recombinant proteins, namely MC1061 and MC4100. Several MC4100 derivatives in which specific heat-shock genes had been either deleted

or inactivated were also included in the study, since the deficiencies in those genes, encoding either chaperones or proteases, modulate protein solubility and aggregation at different extend (Carrió and Villaverde, 2002, 2003). Despite this fact, in all these strains an important fraction of VP1LAC (but not ␤-galactosidase) protein was found in the insoluble cell fraction ranging from around 50 to almost 100% (Fig. 2 and Carrió and Villaverde, 2003). Also, inclusion bodies were clearly detectable by optical microscopy (data not shown). As shown in Fig. 3, protein aggregation was not associated to significant variations in cell growth in MC cells and also in BB2395 (Lon− ) and JGT4 (ClpA− ) strains, but instead it caused a dramatic reduction of cell viability in BB4565 (GroEL44) and in JGT20 (DnaK− ) strains when compared with the production of the soluble ␤-galactosidase. Kinetically, the lower concentration of viable cells was already observed 3 h after induction of recombinant gene expression, although differences were much more important after 5 h, representing around 30% for the GroEL deficient strain and around 50% for the DnaK-mutant (Fig. 4). This indicates that the occurrence of inclusion bodies, or eventually of misfolded protein chains involved in their construction or being removed from them, are deleterious for bacterial cells in absence of main, functional chaperones DnaK or GroEL.

Fig. 1. Optical micrographs of MC1061/pJCO46 (A) and MC1061/pJVP1LAC (B) cells, taken 5 h after induction of recombinant gene expression. As a reference, the arrow indicates one inclusion body in panel B, while these kind of aggregates are not observed in panel A. The synthesis rates of ␤-galactosidase and VP1LAC proteins are indistinguishable (Corchero et al., 1996).


409

Fig. 2. Western blot analysis of recombinant protein distribution (␤-galactosidase labelled as C and VP1LAC labelled as V) into soluble and insoluble cell fractions, in different E. coli strains. Samples were taken 5 h after induction of gene expression. For VP1LAC, percentages of soluble protein were calculated and matched those previously reported (Carrió and Villaverde, 2003).

3.2. Aggregation-dependent toxicity does not cause cell death but growth inhibition The kinetics shown in Fig. 4 reveal deleterious effects of the formation of cytoplasmic inclusion bodies in the absence of either functional chaperones DnaK or GroEL. In this context, we wanted to know if toxicity was caused by cell death or rather by inhibition of

cell growth. Therefore, we have explored cell viability when protein synthesis was arrested by the addition of chloramphenicol and the inducing temperature shifted down to 28 ◦ C. In absence of de novo protein synthesis, inclusion bodies slowly disintegrate at an important extend (Carrió and Villaverde, 2001), since as it has been proven, the volumetric growth of these aggregates is the result of an unbalanced equilibrium between pro-

Fig. 3. Viable cell counts taken 5 h after induction of recombinant gene expression as triggered in several strains. Cells carried either pJCO46 (black bars) or pJVP1LAC (grey bars). Only the significant differences between cells producing ␤-galactosidase and VP1LAC (as observed by an ANOVA comparison test) are indicated.

410


Fig. 4. Viable cell counts during induction of recombinant gene expression in MC4100 and two derivative strains, namely GroEL44 and DnaK− . Cells carried either pJCO46 (black symbols) or pJVP1LAC (white symbols).

Fig. 5. Viable cell counts and optical density in MC4100 cells and the DnaK− derivative carrying either pJCO46 (black symbols, continuous lines) or pJVP1LAC (white symbols, discontinuous lines). Five hours after induction of gene expression chloramphenicol was added to the cultures (time 0).

tein deposition and removal (Carrió et al., 1998, 1999; Carrió and Villaverde, 2002). As shown in Fig. 5, loss of viability was not observed in either parental cells and in the DnaK-deficient derivative.

4. Discussion Usually, foreign proteins overproduced in bacteria do not fold properly (Baneyx and Mujacic, 2004)


and trigger the expression of heat shock genes (Goff and Goldberg, 1985; Allen et al., 1992; Hoffmann and Rinas, 2000; Gill et al., 2000) as well as other cell responses directly or indirectly related to conformational stress (Aris et al., 1998; Harcum and Bentley, 1999; Jurgen et al., 2000). Despite the activation of these mechanisms, folding-defective polypeptides tend to aggregate as refractile particles known as inclusion bodies (see Fig. 1), in which minor amounts of cellular proteins are probably trapped during aggregation (Rinas and Bailey, 1992, 1993). The interferences of such protein deposits with cell functions have not been extensively explored. However, a recent study shows that the bacterial production of a MalE foldingdefective protein variant that forms periplasmic inclusion bodies is lethal for the producing cells (Hunke and Betton, 2003). It was not clear, however, if the aggregates themselves, their precursors or both of them were the deleterious agents. In the present work, we prove that in vivo protein deposition as cytoplasmic inclusion bodies is not toxic (or under detectable levels) in E. coli cells but clearly deleterious in strains devoid of functional GroEL or DnaK chaperones (Fig. 3). However, the absence of the chaperone ClpA or the protease Lon did not result in impaired cell growth, proving that the observed inclusion body toxicity in absence of the main GroEL or DnaK chaperones is not a general feature in heat-shock gene mutants. Interestingly, toxicity in Lon− was eventually expected since the periplasmic protease DegP prevents toxicity of the misfolded MalE31 variant in the periplasm (Hunke and Betton, 2003). DnaK and GroEL, together with their cochaperones DnaJ-GrpE and GroES respectively, are the main cytoplasmic folding systems in E. coli (Arsene et al., 2000). Also, DnaK, in cooperation with ClpB, efficiently disaggregates heat-induced protein deposits (Mogk et al., 2003a,b). Therefore, in absence of DnaK, inclusion bodies are larger than in parental cells (Carrió and Villaverde, 2003), probably as a combined consequence of less efficient folding and impaired removal from aggregates. Unexpectedly, the deficiency in GroEL results in very small and numerous inclusion bodies and more protein amount in the soluble protein fraction (Carrió and Villaverde, 2003). This fact suggests that GroEL could act also as a positive modulator of protein aggregation. The absence of both DnaK and GroEL activities, although rendering dis-

411

similar aggregate types, could increase the amount of toxic VP1LAC species in the cell. Since the amount of VP1LAC present in the soluble cell fraction of DnaK− cells is much lower than in the parental, in fact being hardly detectable (Fig. 2), these toxic species would probably be located in the inclusion bodies. This is in agreement with the fact that the protease Lon, removing misfolded protein in the soluble cell fraction is not defending from inclusion body-mediated toxicity. If the deleterious effects of aggregates depend on their interaction with cellular proteins, the surface of their interface would be a determinant of toxicity. In this regard, we have estimated such average inclusion body surface to be around 1.9, 8.8 and 3.5 ␮m2 per cell for MC4100, JGT20 and BB4565, respectively (according to data from (Carrió and Villaverde, 2003)), in the line of the previous hypothesis. Also in this line, recent studies about protein aggregates internalised by cultured mammalian cells point out the extent of contact surface as a main modulator of toxicity (Bucciantini et al., 2004). In this work, we show for a model, misfoldingprone fusion protein that cytoplasmic inclusion bodies are toxic to E. coli K12 cells, inhibiting cell growth, and that the main cytoplasmic chaperones DnaK and GroEL are cytoprotective. Although the conformational status of the recombinant protein in both soluble and insoluble cell fractions, as influenced by chaperone deficiencies, might be an important determinant of toxicity, data presented here also suggests that the extend of the inclusion body interface could be contributing to expose toxic protein patches to cellular targets.

Acknowledgements We are grateful to B. Bukau, A. Mogk and F. Baneyx for generously providing several bacterial strains. This work has been supported by grants BIO2004-00700 (MEC, Spain), 2002SGR-00099 (AGAUR, Spain) and by Maria Francesca de Roviralta Foundation.

References Allen, S.P., Polazzi, J.O., Gierse, J.K., Easton, A.M., 1992. Two novel heat shock genes encoding proteins produced in response to heterologous protein expression in Escherichia coli. J. Bacteriol. 174 (21), 6938–6947.

412


Aris, A., Corchero, J.L., Benito, A., Carbonell, X., Viaplana, E., Villaverde, A., 1998. The expression of recombinant genes from bacteriophage lambda strong promoters triggers the SOS response in Escherichia coli. Biotechnol. Bioeng. 60 (5), 551–559. Arsene, F., Tomoyasu, T., Bukau, B., 2000. The heat shock response of Escherichia coli. Int. J. Food Microbiol. 55 (1–3), 3–9. Baneyx, F., Mujacic, M., 2004. Recombinant protein folding and misfolding in Escherichia coli. Nat. Biotechnol. 22 (11), 1399–1408. Bucciantini, M., Calloni, G., Chiti, F., Formigli, L., Nosi, D., Dobson, C.M., Stefani, M., 2004. Prefibrillar amyloid protein aggregates share common features of cytotoxicity. J. Biol. Chem. 279 (30), 31374–31382. Bucciantini, M., Giannoni, E., Chiti, F., Baroni, F., Formigli, L., Zurdo, J., Taddei, N., Ramponi, G., Dobson, C.M., Stefani, M., 2002. Inherent toxicity of aggregates implies a common mechanism for protein misfolding diseases. Nature 416 (6880), 507–511. Carrió, M.M., Corchero, J.L., Villaverde, A., 1998. Dynamics of in vivo protein aggregation: building inclusion bodies in recombinant bacteria. FEMS Microbiol. Lett. 169 (1), 9–15. Carrió, M.M., Corchero, J.L., Villaverde, A., 1999. Proteolytic digestion of bacterial inclusion body proteins during dynamic transition between soluble and insoluble forms. Biochim. Biophys. Acta 1434 (1), 170–176. Carrió, M.M., Cubarsi, R., Villaverde, A., 2000. Fine architecture of bacterial inclusion bodies. FEBS Lett. 471 (1), 7–11. Carrió, M.M., Villaverde, A., 2001. Protein aggregation as bacterial inclusion bodies is reversible. FEBS Lett. 489 (1), 29–33. Carrió, M.M., Villaverde, A., 2002. Construction and deconstruction of bacterial inclusion bodies. J. Biotechnol. 96 (1), 3–12. Carrió, M.M., Villaverde, A., 2003. Role of molecular chaperones in inclusion body formation. FEBS Lett. 537 (1–3), 215–221. Corchero, J.L., Viaplana, E., Benito, A., Villaverde, A., 1996. The position of the heterologous domain can influence the solubility and proteolysis of beta-galactosidase fusion proteins in E. coli. J. Biotechnol. 48 (3), 191–200. Dougan, D.A., Mogk, A., Bukau, B., 2002. Protein folding and degradation in bacteria: to degrade or not to degrade? That is the question. Cell Mol. Life Sci. 59 (10), 1607–1616. Gill, R.T., Valdes, J.J., Bentley, W.E., 2000. A comparative study of global stress gene regulation in response to overexpression of recombinant proteins in Escherichia coli. Metab. Eng. 2 (3), 178–189. Goff, S.A., Goldberg, A.L., 1985. Production of abnormal proteins in E. coli stimulates transcription of lon and other heat shock genes. Cell 41 (2), 587–595. Harcum, S.W., Bentley, W.E., 1999. Heat-shock and stringent responses have overlapping protease activity in Escherichia coli. Implications for heterologous protein yield. Appl. Biochem. Biotechnol. 80 (1), 23–37. Hengge, R., Bukau, B., 2003. Proteolysis in prokaryotes: protein quality control and regulatory principles. Mol. Microbiol. 49 (6), 1451–1462. Hoffmann, F., Rinas, U., 2000. Kinetics of heat-shock response and inclusion body formation during temperature-induced produc-

tion of basic fibroblast growth factor in high-cell-density cultures of recombinant Escherichia coli. Biotechnol. Prog. 16 (6), 1000–1007. Hunke, S., Betton, J.M., 2003. Temperature effect on inclusion body formation and stress response in the periplasm of Escherichia coli. Mol. Microbiol. 50 (5), 1579–1589. Ignatova, Z., Gierasch, L.M., 2004. Monitoring protein stability and aggregation in vivo by real-time fluorescent labelling. Proc. Natl. Acad. Sci. U.S.A. 101 (2), 523–528. Jurgen, B., Lin, H.Y., Riemschneider, S., Scharf, C., Neubauer, P., Schmid, R., Hecker, M., Schweder, T., 2000. Monitoring of genes that respond to overproduction of an insoluble recombinant protein in Escherichia coli glucose-limited fed-batch fermentations. Biotechnol. Bioeng. 70 (2), 217–224. Laemmli, U.K., 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227 (259), 680–685. Miller, J.H., 1972. Experiments in Molecular Genetics. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Mogk, A., Deuerling, E., Vorderwulbecke, S., Vierling, E., Bukau, B., 2003a. Small heat shock proteins, ClpB and the DnaK system form a functional triade in reversing protein aggregation. Mol. Microbiol. 50 (2), 585–595. Mogk, A., Schlieker, C., Friedrich, K.L., Schonfeld, H.J., Vierling, E., Bukau, B., 2003b. Refolding of substrates bound to small Hsps relies on a disaggregation reaction mediated most efficiently by ClpB/DnaK. J. Biol. Chem. 278 (33), 31033–31042. Olofsson, A., Ostman, J., Lundgren, E., 2002. Amyloid: morphology and toxicity. Clin. Chem. Lab. Med. 40 (12), 1266–1270. Rinas, U., Bailey, J.E., 1992. Protein compositional analysis of inclusion bodies produced in recombinant Escherichia coli. Appl. Microbiol. Biotechnol. 37 (5), 609–614. Rinas, U., Bailey, J.E., 1993. Overexpression of bacterial hemoglobin causes incorporation of pre-beta-lactamase into cytoplasmic inclusion bodies. Appl. Environ. Microbiol. 59 (2), 561– 566. Sambrook, J., Fritsch, E., Maniatis, T., 1989. Molecular Cloning, A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Stefani, M., Dobson, C.M., 2003. Protein aggregation and aggregate toxicity: new insights into protein folding, misfolding diseases and biological evolution. J. Mol. Med. 81 (11), 678–699. Thomas, J.G., Baneyx, F., 1998. Roles of the Escherichia coli small heat shock proteins IbpA and IbpB in thermal stress management: comparison with ClpA, ClpB, and HtpG in vivo. J. Bacteriol. 180 (19), 5165–5172. Tomoyasu, T., Mogk, A., Langen, H., Goloubinoff, P., Bukau, B., 2001. Genetic dissection of the roles of chaperones and proteases in protein folding and degradation in the Escherichia coli cytosol. Mol. Microbiol. 40 (2), 397–413. Villaverde, A., Carrió, M.M., 2003. Protein aggregation in recombinant bacteria: biological role of inclusion bodies. Biotechnol. Lett. 25 (17), 1385–1395. Weibezahn, J., Bukau, B., Mogk, A., 2004. Unscrambling an egg: protein disaggregation by AAA+ proteins. Microb. Cell Fact 3 (1), 1.


Microbial biomass production from rice straw hydrolysate in airlift bioreactors Yu-Guo Zheng ∗ , Xiao-Long Chen, Zhao Wang Institute of Bioengineering, Zhejiang University of Technology, Hangzhou 310014, PR China Received 5 October 2004; received in revised form 29 March 2005; accepted 6 April 2005

Abstract Rice straw is a by-product of rice production, and a great bioresource as raw biomass material for manufacturing value-adding protein for animal feedstock, which has been paid more and more attention. In the present work, utilizing rice straw hydrolysate as a substrate for microbial biomass production in 11.5 L external-loop airlift bioreactors was investigated. Rice straw hydrolysate obtained through acid-hydrolyzing rice straw was used for the culture of yeast Candida arborea AS1.257. The influences of gas flow rate, initial liquid volume, hole diameter of gas sparger and numbers of sieve plates on microbial biomass production were examined. The best results in the external-loop airlift bioreactor were obtained under 9.0 L initial liquid volume, 1.1 (v/v)/min gas flow rate during culture time of 0–24 h and 1.4 (v/v)/min gas flow rate of 24–48 h at 29 ± 1 ◦ C. The addition of the sieve plates in the riser of the external-loop airlift bioreactor increased productivity. After 48 h, under optimized operation conditions, crude protein productivity with one sieve and two sieves were 13.6 mg/mL and 13.7 mg/mL, respectively, comparing 12.7 mg/mL without sieves in the airlift bioreactor and 11.7 mg/mL in the in the 10-L mechanically stirred tank bioreactor. It is feasible to operate the external-loop airlift bioreactors and possible to reduce the production cost for microbial biomass production from the rice straw hydrolysate. © 2005 Elsevier B.V. All rights reserved. Keywords: Agricultural residues; Bioreactor; Bioresource; Hydrolysis; Microbial biomass protein; Rice straw

1. Introduction Rice is one of the largest food crops in the world. In Asia alone, more than 2 billion people obtain

∗ Corresponding author. Tel.: +86 571 88320379; fax: +86 571 88320630. E-mail address: [email protected] (Y.-G. Zheng).


60–70% of their calories from rice (FAO, 2003). Thirty percent of the 135 million ha of total arable land in China is cultivated with rice. In the process of edible rice culture, a great amount of rice straw is generated as a byproduct. Because the rice straw is comprised mainly of lignin, cellulose, and hemicellulose, its direct utilization ratio as animal feedstock is very low. Most of the rice straw is used as a bedding material for animals or is burned. This

414

Y.-G. Zheng et al. / Journal of Biotechnology 118 (2005) 413–420

biomass resource is wasted and burnt which results in environmental pollution caused by global addition of carbon dioxide, a gas contribution to the greenhouse effect, and likely high health costs through increase in respiratory problems in the local population (Samar et al., 1999). There have been several reports on the use of the rice straw to produce chemicals and biochemicals (Roberto et al., 2003; Shambe and Kennedy, 1985; Sun and Cheng, 2002; Yang et al., 2003). Through appropriate hydrolysis of rice straw, this lignocellulosic biomass could be transferred into fermentable sugar as cultural substrate for growth of microorganisms. The effective utilization of the rice straw as a source of animal feedstock would not only help solve the problems of bioresource utilization, but also enable farmers to realize new income by selling value-added microbial biomass. Over the years efforts have been made to use rice straw for production of microbial biomass. Submerged fermentations have usually been used and mechanically stirred tank bioreactors have been the main type of bioreactor used for the submerged fermentation. However, it would be uneconomical to use this type of bioreactor to produce microbial biomass from the hydrolysate of rice straw because of relatively high agitation power requirements and capital cost (including impeller, motor, shaft seals, supports, and so on). To overcome these problems, alternative configurations, particularly airlift bioreactors, should be studied experimentally and developed. Airlift bioreactors have received much attention due to the simplicity of their design and construction, ease of operation, and lower energy consumption. There have been several studies about production of microbial biomass from industrial and agricultural wastes (Anupama and Ravindra, 2000; Choi and Park, 2003; Ferrer et al., 1996; Hussein et al., 1992; Kurbanoglu and Algur, 2002; Summers et al., 2003), but these studies did not deal with an external-loop airlift bioreactor with more than 10 L total volume. In order to find more efficient utilization of rice straw, we studied the possibility of using rice straw hydrolysate as a cultural substrate for yeast strains suitable for microbial biomass protein. External-loop bioreactors with a lower ratio 2.9 of height-to-diameter were used to produce microbial biomass from the rice straw hydrolysate to investigate its suitability for yeast culture.

2. Materials and methods 2.1. Microorganism and substrate preparation Candida arborea AS1.257 (from China General Microbiological Culture Collection Centre, CGMCCC) was used in present research. The rice straw was obtained from a local rice-milling factory. It contained 2.7% crude protein and 12.0% moisture. The rice straw had been ground into 80-mesh powder before being used in the experiments. A twostage procedure for acid hydrolysis of rice straw was employed. In the first-stage acid hydrolysis, 80-mesh rice straw was added into 2% H2 SO4 with solid/liquid (S/L) ratio of 0.2 (w/v), constantly stirred at a temperature of 100 ◦ C for 1 h in a big beaker. Through filtration, the first-stage hydrolysis liquor was obtained. The filtration residue was treated again in the second-stage procedure of the acid hydrolysis with the same conditions as that in the first stage. The filtrate from the two stages were joined and conserved in the fridge as the culture substrates for microbial biomass production through adjusting the pH, sugar concentration and adding nitrogen sources such as ammonia solution or urea. The filtration residue from the second-stage hydrolysis was washed using water with solid/liquid ratio of 1.0. The washed water was used as water for next hydrolysis and the residue was discarded. 2.2. Experimental apparatus The details of the 11.5 L total volume airlift bioreactor with external-loop used in this work are given in Table 1 and its structure was described by Zheng et al. (2000, 2001, 2004). The ratio of height (m) to diameter (m) of the riser of the external-loop airlift bioreactor (ELAB) was 2.9, and the ratio of riser to downcomer diameter was 6.6. A gas distributor was equipped on the bottom of the riser in ELAB. A horizontal pipe whose diameter was 0.025 m connected the riser and downcomer. For the study, the bioreacTable 1 Bioreactor dimensions Bioreactor

Dr (m)

Dd (m)

Ar /Ad

H (m)

VT (L)

ELAB MSTB

0.165 0.185

0.025 –a

43.56 –

0.479 0.370

11.5 10.0

a

In MSTB, there is no riser and downcomer.


tor was modified by fitting 5-mesh sieve plates made of stainless steel in the riser. The first sieve plate was fixed at 100 mm from the bottom of the bioreactor, and the second at 202 mm. The ELAB could be measured and controlled automatically by pumping temperaturecontrolled water through the jacket on the riser. The mechanically stirred tank bioreactor (MSTB) had a total volume of 10.0 L, a ratio of height to diameter of 2.0 and three flat-bladed disc turbine impellers stacked vertically. The rate of agitation and gas flow could be measured and controlled, as well as operation temperature by pumping temperature-controlled water through the jacket on the vessel. The above two bioreactors could be fed with the medium and added with the antifoam agents, together with the steam-generation system and the filtered gas system. 2.3. Media and cultivation The microorganism, C. arborea AS1.257, was first cultured at 29 ◦ C for 24 h on an agar slant medium with 10% (w/v) sugar degree of malt wort. Then it was transferred to a 500 mL shaking flask containing 50 mL of sterilized inoculum medium with 5% (w/v) sugar degree of malt wort. After the inoculation, the flasks were placed in an rotating shaker at 200 rpm for 24 h at 28–30 ◦ C. Fermentation medium for yeast biomass production contained the hydrolysate as carbon source and urea or ammonia solution as nitrogen source and pH 4.5–5.0 adjusted with 1.0 mol/L ammonia solution or NaOH solution. Reducing sugar concentration of fermentation medium was adjusted to about 4.5% (w/v). The ELAB and the MSTB were each filled with 8.0–9.5 L and 6.0–8.5 L fermentation media, respectively, to study the effects of liquid volume on microbial biomass production. The fermentation medium was sterilized at 121 ◦ C for 20 min, subsequently, cooled to 29 ± 1 ◦ C and inoculated with 5% (v/v) inoculum medium. The culture conditions in the ELAB were 0.08 MPa pressure, 29 ± 1 ◦ C temperature and 1.0–1.5 (v/v)/min gas flow rate for 48–50 h. The culture conditions in the MSTB were 0.08 MPa pressure, 29 ± 1 ◦ C temperature, 0.7–1.0 (v/v)/min gas flow rate and 150 rpm agitation rate for 48–50 h. Different gas flow rates were applied at different culture periods to investigate the effects of gas flow rate on microbial biomass production.

415

2.4. Analytical methods The yeast cell growth in culture broth was measured by determining optional density at 600 nm using a 722mode grating spectrophotometer (Shanghai Analytical Instrument Overall Factory, Shanghai, PR China). To determining dry yeast cell biomass content, the culture broth was sampled, and then centrifuged at 5000 × g for 15 min, and cell biomass was recovered after washed twice with distilled water. Dry microbial biomass weight was determined after the cells were dried at 105 ◦ C for 4 h. The crude protein content of cells was determined by the Kjeldahl method (Li et al., 2000), in which microbial cells were harvested by centrifugation and washed three times with distilled water to remove the dissolved nitrogen. The concentration of reducing sugar was assayed by a colorimetric method (Jia et al., 1983), where 3.5 mL of dinitrosalicylic acid (DNS) was used as color reagent and absorbance was measured using a 722-mode grating spectrophotometer at 550 nm.

3. Results and discussion 3.1. Influences of nitrogen source supplements on microbial biomass production To improve microbial biomass production and the protein content of C. arborea AS1.257 yeast cells produced from the rice straw hydrolysate, ammonia solution and urea as nitrogen source were added to fermentation medium. The ammonia solution was used as both a pH adjusting agent and an inorganic nitrogen source for yeast cells. The influences of varying concentration of ammonia solution and urea on crude protein content, sugar utilization ratio, and microbial biomass content are presented in Table 2. The crude protein content and sugar utilization ratio were the highest after a total fermentation time of 48 h when 2.7% ammonia solution and 0.8% urea were added. 3.2. Influences of gas flow rate on microbial biomass production in the ELAB The volumetric gas flow rate in a pneumatically agitated bioreactor has a large effect on oxygen transfer rate, degree of turbulence, and broth circulation

416


Table 2 Influences of nitrogen source supplements on microbial biomass production Ammonia solution (%, v/v)

Urea (%, w/v)

Crude protein content (%)

Cell biomass content (mg/mL)

Sugar utilization ratio (%)

0.9 1.8 2.7 3.6

1.4 1.1 0.8 0.5

56.2 57.8 60.8 59.0

18.7 20.0 20.8 20.5

77.7 77.8 81.9 80.9

velocity, which in turn affect microbial biomass concentration, substrate consumption rate and product accumulation (Sajc et al., 1995). The volumetric oxygen transfer of the airlift bioreactor increased as the gas flow rate increased. When dissolved oxygen could not be supplied sufficiently, the cell growth would become oxygen limited. It was necessary for the dissolved oxygen concentration in the fermentation broth to be higher than the critical oxygen level of the microorganisms. To optimize gas flow rate for the production of microbial biomass in the external-loop airlift bioreactor, the culture experiments were performed in a range of gas flow rate from 1.0 (v/v)/min to 1.5 (v/v)/min with 8.5 L initial liquid volume. The effects of varying gas flow rate are shown in Figs. 1–3. Yeast cell growth rate during the accelerating phase of cell growth increased with the increase of gas flow rate. Reducing sugar consumption rate and crude protein accumulation rate of yeast cell biomass also improved as gas flow rate increased. After a total fermentation time of 48 h, the highest value of 58.5% crude protein content was obtained at 1.5 (v/v)/min gas flow rate. When the fermentations were ended, the yeast cell biomass con-

Fig. 2. Effects of gas flow rate on reducing sugar consumption in the ELAB.

Fig. 1. Influences of gas flow rate on yeast cell growth in the ELAB.

Fig. 3. Effects of gas flow rate on crude protein accumulation in the ELAB.

centrations were not greatly different at the different gas flow rates. The maximum concentration of yeast cell biomass, 21.7 mg/mL, occurred at 1.2 (v/v)/min gas flow rate, and the minimum, 20.8 mg/mL, at 1.3 (v/v)/min after a total fermentation time of 48 h. Experiments on microbial biomass production in the external-loop airlift bioreactor with different aeration modes were also studied. One of aeration modes was to maintain a constant gas flow rate throughout the fermentation course, and another was to vary gas


417

Table 3 Effects of aeration modes in the ELAB on microbial biomass production Aeration modes

Cell biomass content (mg/mL)

Crude protein content (%)

Sugar utilization ratio (%)

Aeration mode I (1) Aeration mode I (2) Aeration mode II

21.6 20.9 20.9

51.7 58.5 60.8

76.4 78.7 81.5

Aeration mode I (1) was 1.1 (v/v)/min gas flow rate throughout fermentation course; aeration mode I (2) was 1.5 (v/v)/min throughout fermentation course; aeration mode II was 1.1 (v/v)/min in the earlyperiod and 1.4 (v/v)/min in the later-period.

flow rate at different fermentation periods. A comparison of experiment results between the two aeration modes is shown in Table 3 (with 8.5 L initial liquid volume). From Table 3, it was found that yeast cell biomass content with both two aeration modes after a total fermentation time of 48 h was not very different (21.6–20.9 mg/mL) but crude protein content with the second aeration mode of 1.1 (v/v)/min gas flow rate in the early period (before 24 h) and 1.5 (v/v)/min gas flow rate in the later period (after 24 h) was higher than that in the first aeration mode. Cutting down gas flow rate is advantageous to reduce the cost in microbial biomass production. Therefore, second mode is more suitable in the production. 3.3. Influences of liquid volume on microbial biomass production in the ELAB It has been proven that the liquid depth might influence mixing and particle circulation time in the airlift bioreactor (Snape et al., 1992). The depth of the liquid in the bioreactor also affected the gas holdup; the greater the depth of liquid in the bioreactor, the lower the gas holdup. As a result, liquid depth may affect the volumetric oxygen transfer, which then affects the microbial biomass production by fermentation. Because the depth of liquid was directly proportional to the liquid volume in the given bioreactors, fermentation trials were carried out in the external-loop bioreactor with initial liquid volume from 8.0 L up to 9.5 L, corresponding to liquid depth from 0.346 m up to 0.400 m, to investigate the impact of liquid height on the fermentation results. The results are shown in Fig. 4. Unfortunately, at the initial volume of 8.0 L, the circulation of bioreactor was worse and worse with the

Fig. 4. Effects of initial liquid volume on crude protein content.

fermentation time because of the vaporization of the broth. Therefore, the crude protein accumulation was low. In the experiments, the highest crude protein accumulation was obtained with a liquid volume of 9.0 L. Raising the liquid depth causes increase of hydrostatic pressure in bioreactor, which makes air bubbles compressed and become smaller, and in this case transfer of oxygen is decreased. But the residue time of air bubbles in the broth increases with increasing the depth of liquid, which improves the dissolution of oxygen. Our experimental results show that a liquid volume of 8.5–9.0 L results a hydrostatic pressure at which the concentration of crude protein in maximized; the compression of air bubbles in the liquid is offset by increased residue time and oxygen solubility. 3.4. Influences of sparger hole diameter in ELAB on microbial biomass production The oxygen transfer, hydrodynamic characteristics and fermentation characteristics affected by the gas sparger hole design. A smaller gas sparger hole diameter increased oxygen transfer and gas holdup because it reduced the gas bubble size and increased the characteristic gas velocity, increasing the fractional aerated area. The experimental external-loop airlift bioreactor was equipped with spargers possessing different hole diameters, including 1.0 mm, 1.5 mm, 2.0 mm, 3.0 mm, and 5.0 mm, and the influences on the microbial biomass production were examined under a optimized aeration mode (aeration mode II) and initial liquid volume of 8.5 L. The crude protein accumulation increased with decreasing gas sparger hole diameter (Fig. 5). The

418


Fig. 5. Effects of gas sparger diameter on crude protein content.

smaller the gas sparger hole diameter, the smaller the size of the gas bubbles. The highest value of crude protein content was obtained with 1.0 mm gas sparger diameter. Therefore, using smaller gas sparger hole diameter in the ELAB was advantageous to raise the productivity of crude protein of yeast cells. 3.5. Influences of equipping with sieve plates in the riser of the external-loop airlift bioreactor Equipping with inner devices inside the riser of an airlift bioreactor could promote bubble dispersion and gas–liquid mixing so as to enhance oxygen transfer. Different types of static mixers have been designed and developed by some researchers (Stejskal and Potucek, 1985; Gavrilescu et al., 1997). They were fixed in the airlift bioreactor to improve the reactor’s hydrodynamics and mass transfer characteristics. In our experiments, 5-mesh sieve plates were installed in the riser of the external-loop airlift bioreactor. The experiments to examine microbial biomass production response were then carried out in the external-loop airlift bioreactors with one stainless sieve plate and two stainless sieve plates, respectively. A comparison of microbial biomass culture was made in the external-loop airlift bioreactors among with one stainless sieve plate, two stainless sieve plates and without any inner-devices. The operation conditions of the airlift bioreactors were gas flow rate of 1.1 (v/v)/min in the early-period (before 24 h), 1.4 (v/v)/min in the later-period (after 24 h) and initial liquid volume of 8.5 L. Time courses of microbial

Fig. 6. The time courses of yeast cell biomass content in the ELAB and the MSTB. The ELAB (1), without sieve plates; the ELAB (2), with one sieve plate; the ELAB (3), with two sieve plates.

biomass content, crude protein content and reducing sugar content in the external-loop airlift bioreactors are shown in Figs. 6–8, respectively. It was found that yeast cell biomass growth rate, reducing sugar consumption rate and crude protein accumulation rate in the external-loop airlift bioreactor with two stainless sieve plates were the highest among three modifications of the ELAB. After 48 h under abovementioned optimum operating conditions, the averages of 21.8 mg/mL microbial biomass content, 62.9% crude protein content, 13.7 mg/mL crude protein production and 82.0% sugar utilization ratio were obtained in the ELAB with two stainless sieve plates; 21.9 mg/mL microbial biomass content, 62.1% crude protein content, 13.6 mg/mL crude protein production

Fig. 7. The time courses of reducing sugar content in the ELAB and the MSTB. The ELAB (1), without sieve plates; the ELAB (2), with one sieve plate; the ELAB (3), with two sieve plates.


419

sugar change and cell growth during microbial biomass fermentation are shown in Figs. 6–8. After a total fermentation time of 48 h, the averages of 19.8 mg/mL microbial biomass content, 59.0% crude protein content, 11.7 mg/mL crude protein production and 76.1% sugar utilization ratio were obtained in the MSTB under above-mentioned optimum operating conditions. From Figs. 6–8, it was observed that crude protein content and yeast cell biomass content in the ELAB were higher than these in the MSTB. Fig. 8. The time courses of crude protein content in the ELAB and the MSTB. The ELAB (1), without sieve plates; the ELAB (2), with one sieve plate; the ELAB (3), with two sieve plates.

and 82.9% sugar utilization ratio in the ELAB with one stainless sieve plate. Comparing with them, 20.9 mg/mL microbial biomass content, 60.8% crude protein content, 12.7 mg/mL crude protein production and 77.8% sugar utilization ratio in the ELAB without stainless sieve plates. So the highest crude protein content, 62.9%, was obtained in the external-loop airlift bioreactors with two stainless sieve plates, which was higher than that in other type of bioreactors. The results illustrated that the addition of sieve plates to the riser of the airlift bioreactor could accelerate microbial cell growth and increase protein content. The reason was that the volumetric oxygen transfer coefficient, KL a, of the airlift bioreactor increased. 3.6. Comparison of microbial biomass production between in the ELAB and the MSTB Mechanically stirred tank bioreactors have widely used for the production of different chemicals and biochemicals by the submerged fermentation. To investigate the possibility of substituting the ELAB for the MSTB, a comparison was made of microbial biomass fermentation both two types of the bioreactors. The basis for comparison was established according to the highest crude protein accumulation on the yeast cells under the optimized operation conditions in both ELAB and MSTB. The effects of gas flow rate, agitation rate and initial liquid volume were examined in the MSTB. The optimum operating conditions of the MSTB were gas flow rate of 1.0 (v/v)/min, agitation rat of 150 rpm and initial liquid volume of 7.5 L. Time courses of crude protein accumulation in yeast cells, reducing

4. Conclusions The microbial biomass production from the hydrolysate of rice straw by C. arborea using the external-loop airlift bioreactors with a lower heightto-diameter ratio of 2.9 and a riser-to-diameter ratio of 6.6 was studied in order to reduce the production cost of microbial biomass. An optimum crude protein content was obtained by growing C. arborea under defined cultural operation conditions. The experimental results obtained in the ELAB were better than these in the MTSB. Because the total consumed power of the ELAB was that of the MSTB, the ELAB was beneficial to the production of microbial biomass. Because of the ELAB with a low ratio of height-to-diameter, which is similar to that of most MSTB, it is feasible to modify the MSTB, which has been widely used in microbial submerge fermentation, particularly in China. To increase the productivity of microbial biomass could be realized by optimizing the different operation conditions and refitting the structure of the ELAB such as equipping with sieve plates in the riser.

Acknowledgement This work was supported by the National Fund of the Major Basic Research Development Program 973 of China (2003CB716005).

References Anupama, Ravindra, P., 2000. Value-added food: single cell protein. Biotechnol. Adv. 18, 459–479.

420


Choi, M.H., Park, Y.H., 2003. Production of yeast biomass using waste Chinese cabbage. Biomass Bioenergy 25, 221–116. FAO, 2003. http://www.fao.org/rice2004/en/world.htm. Ferrer, J., Paez, G., Marmol, Z., Ramones, E., Garcia, H., Forster, C.F., 1996. Acid hydrolysis of shrimp-shell wastes and the production of single cell protein from the hydrolysate. Bioresour. Technol. 57, 55–60. Gavrilescu, M., Roman, R.V., Tudose, R.Z., 1997. Hydrodynamics in the external-loop airlift bioreactors with static mixers. Bioprocess. Eng. 16, 93–99. Hussein, A.M., El-Saied, H., Yasin, M.H., 1992. Bioconversion of hemicelluloses of hull black liquor into single-cell protein. J. Chem. Technol. Biotechnol. 53, 147–152. Kurbanoglu, E.B., Algur, O.F., 2002. Single-cell protein production from ram horn hydrolysate by bacteria. Bioresour. Technol. 85, 125–129. Jia, S., Mu, G., Ren, S., 1983. A rapid method for determination of reducing sugar, glucose and starch in fruits and vegetables with 3,5-diatrosalycylic acid. Food Ferment. Ind. 10, 30–34. Li, J.W., Xiao, N., Yu, R.Y., Yuan, M.X., Chen, L.R., Chen, Y.H., Chen, L.T., 2000. Principles and Methods of Biochemical Experiments. Peking University Press, Beijing, pp. 160–164. Roberto, I.C., Mussatto, S.I., Rodrigues, R.C.L.B., 2003. Dilute-acid hydrolysis for optimization of xylose recovery from rice straws in a semi-pilot reactor. Ind. Crops Prod. 17, 171–176. Sajc, L., Oberadovic, B., Vakovic, D., Bugarki, B., 1995. Hydrodynamics and mass transfer in a four-phase external-loop airlift bioreactor. Biotechnol. Prog. 11, 420–428. Samar, S., Malik, R.K., Mangat, R., Singh, S., Ram, M., 1999. Effect of rice straw burning on the efficacy of the herbicides in wheat (Triticum aestivum). Indian J. Agron. 44, 361–366.

Shambe, T., Kennedy, J.F., 1985. Acid and enzymatic hydrolysis of chotropically pretreated millet stalk, acha and rice straws and conversion of the products to ethanol. Enzyme Microb. Technol. 7, 115–120. Snape, J.B., Fialova, M., Zahradnik, J., 1992. Hydrodynamic studies in an external-loop airlift reactor containing aqueous electrolyte and sugar solutions. Chem. Eng. Sci. 47, 3387–3394. Stejskal, J., Potucek, F., 1985. Oxygen transfer in liquids. Biotechnol. Bioeng. 27, 503–508. Summers, M.D., Jenkins, B.M., Hyde, P.R., Williams, J.F., Mutters, R.G., Scardacci, S.C., Hair, M.W., 2003. Biomass production and allocation in rice with implications for straw harvesting and utilization. Biomass Bioenergy 24, 163–173. Sun, Y., Cheng, J., 2002. Hydrolysis of lignocellulosic materials for ethanol production: a review. Bioresour. Technol. 83, 1– 11. Yang, H.S., Kim, D.J., Kim, H.J., 2003. Rice straw-wood particle composite for sound absorbing wooden construction materials. Bioresour. Technol. 86, 117–121. Zheng, Y.G., Chen, X.L., Wang, Z., Shen, Y.C., 2004. Production of validamycins from crude substrates by Streptomyces hygroscopicus in an external-loop airlift bioreactor with a low height-todiameter ratio. Chin. J. Chem. Eng. 12, 102–107. Zheng, Y.G., Wang, Z., Chen, X.L., 2000. ␣-Amylase production by Bacillus subtilis with dregs in an external-loop airlift bioreactor. Biochem. Eng. J. 5, 115–121. Zheng, Y.G., Wang, Z., Chen, X.L., Zhang, C.H., 2001. Production of extracellular protease from crude substrates with dregs in an external-loop airlift bioreactor with lower ratio of height to diameter. Biotechnol. Prog. 17, 273–277.


Capture of bacterial endotoxins using a supermacroporous monolithic matrix with immobilized polyethyleneimine, lysozyme or polymyxin B Amro Hanora a,c , Fatima M. Plieva b , Martin Hedström a , Igor Yu. Galaev a,b , Bo Mattiasson a,b,∗ a

Department of Biotechnology, Center for Chemistry and Chemical Engineering, Lund University, P.O. Box 124, SE-22100 Lund, Sweden b Protista Biotechnology AB, P.O. Box 86, SE-26722 Lund, Sweden c Department of Microbiology and Immunology, Faculty of Pharmacy, Suez Canal University, Ismailia, Egypt Received 17 December 2004; received in revised form 25 April 2005; accepted 2 May 2005

Abstract Bacterial endotoxins (BEs) are integrated part of Escherichia coli, a microorganism widely used for the production of recombinant proteins. BEs should be eliminated in the course of down stream processing of target protein produced by these bacteria. Supermacroporous monolith (continuous bed) columns, so called cryogel columns, with immobilized polyethyleneimine (PEI), polymyxin B (PMB) and lysozyme were employed for BEs capture. Due to the large interconnected pores it was possible to use cryogel columns at flow rates as high as 10 ml/min. The columns packed with Sepharose CL-4B with immobilized PEI, PMB and lysozyme were impossible to use at these high flow rates due to the collapse of the bed. The dynamic capacities of the cryogel columns were nearly independent of the flow rate. In the presence of EDTA, BEs were quantitatively captured from mixtures with a model protein, bovine serum albumin (BSA) at pH 7.2 with practically no protein losses. At pH 3.6 BEs were captured directly from non-clarified E. coli cell lysate resulting in more than 104 times BEs clearance. © 2005 Elsevier B.V. All rights reserved. Keywords: Supermacroporous gel; Monolith column; Endotoxin; Affinity chromatography

1. Introduction Bacterial Endotoxins (BEs), also known as lipopolysaccharides (LPS), are an integrated part of ∗ Corresponding author. Tel.: +46 46 2228264; fax: +46 46 2224713. E-mail address: [email protected] (B. Mattiasson).


outer membrane of Gram negative bacteria. BEs are composed of three parts. The first part is a lipid part called lipid A, which has two 1,6-linked phosphoryl glucosamines acylated asymmetrically with four hydroxylated fatty acids (Seydel et al., 2003). The second part is a core part consisting of an inner core region with three 2-keto-3-deoxyoctonic acid residues linked to heptose residues and an outer core hexose

422

A. Hanora et al. / Journal of Biotechnology 118 (2005) 421–433

region (Petsch and Anspach, 2000). The third part is O-specific part consisting of repeating (4–40) oligosaccharides units with 3–8 monosaccharides in each unit (Petsch and Anspach, 2000). Lipid A is the most conserved part of BEs and it is responsible for the toxicity, however the O-specific part is responsible for the antigenicity of BEs. BEs are heterogeneous both in size and composition and are amphiphilic in nature having both a hydrophilic part (the phosphoric acid groups and sugar part) and a hydrophobic part (the fatty acids part in lipid A) (Hirayama and Sakata, 2002). Due to their amphiphilic character, BEs are prone to form micelle-like structures in aqueous media with molecular weights up to 1000 kDa (Santos et al., 2003). Upon administration in blood stream, BEs cause inflammation and septic chock (Liu et al., 1997). The general threshold level of endotoxin set by pharmacopoeias world-wide for intravenous applications is 5 endotoxin units (EU) per kg body weight per an hour with 1 EU being equal to 100 pg BEs (Hirayama and Sakata, 2002). The amount of BEs allowed in pharmacological protein preparations depends on the intended use. For the proteins that are administrated in relatively low doses, the regulatory demands are modest, for example insulin and ␣-interferon could contain up to 10 and 100 EU/mg, respectively (Petsch and Anspach, 2000). However, for the preparations of antibodies and human serum albumin that are administered in large amounts, the allowed amount of BEs per mg of protein is extremely low. As BEs are integral part of Escherichia coli, a microorganism widely used for the production of recombinant proteins, there is always a risk of BEs copurification with the target product. BEs concentrations in starting material varies from less than 100 EU/ml in cell culture supernatants to more than 2,000,000 EU/ml in supernatants after homogenization of high cell density bacterial cultivations (Petsch and Anspach, 2000). Various methods have been used for BE removal from non-protein preparations including ion-exchange chromatography, ultra-filtration and sucrose gradient centrifugation (Hirayama and Sakata, 2002). Affinity adsorption methods have been proven to be efficient for BE removal from protein solutions (Anspach, 2001). Selective endotoxin adsorption implies the use of negative chromatography. In this technique, the target product (protein of interest) passes through the column, whereas the contaminants are bound.

LPS-binding proteins like endotoxin-binding protein, bactericidal/permeability-increasing protein and palate lung nasal epithelial clone protein have a high specificity towards BEs (Ghafouri et al., 2004). Alternatively, less specific, but more cheap and robust ligands like polyethyleneimine (PEI), lysozyme, Polymyxin B (PMB), polyhistidine, polyhistamine, polylysine and histidine have been used for BEs removal (Anspach, 2001; Matsumae et al., 1990). Traditionally these ligands were immobilized on matrices like Sepharose used for protein chromatography. These matrices have some inherent disadvantages when being used for negative chromatography: the columns have high flow resistance resulting in pronounced pressure drops when pumping feed through the column; slow mass transfer especially of large entities like BEs micelles due to the diffusion in and out of the pores; impossibility to process turbid feeds due to the column blocking by particulate material. The development of supermacroporous monolith (continuous bed) columns has opened new perspectives for selective capture of biomolecules at high flow rates from non-pretreated feeds like cell homogenates (Arvidsson et al., 2003) or cell suspensions (Dainiak et al., 2004). Monoliths have many advantages over beds of packed beads, as they are inexpensive, the production can be performed in situ in different formats from 96 well plate format (Hanora et al., in press) to large column for industrial application with the bed volume up to 8000 ml (Podgornik et al., 2004). The mass transfer in monoliths is due to convection rather than diffusion as in case of packed bed columns allowing for operation at high flow rates without loosing resolution (Bedair and Rassi, 2004) or capacity (Podgornik et al., 2004). Indeed, fast processing is highly advantageous for decontamination using negative chromatography. In this study, supermacroporous monolith columns with immobilized ligands, polyethyleneimine, lysozyme, and polymyxin B were evaluated for BEs removal from model BEs–protein mixtures and crude cell lysate.

2. Material and methods 2.1. Chemicals and reagents Polymyxin B (MW 1450 Da), polyethyleneimine (MW 60,000 Da), lysozyme (MW 14,300 Da), pic-


rylsulfonic acid (2,4,6-trinitrobenzene sulfonic acid hydrate; TNBS) and bovine serum albumin (BSA) were purchased from Sigma (St. Louis, USA). Sepharose CL-4B was obtained from Amersham Biosciences (Sweden). All other chemicals were of analytical grade and commercially available. The supermacroporous monolith columns with bed volume of 2 ml (2.5 cm × 1 cm i.d.) produced from polyacrylamide and carrying reactive epoxy groups were provided by Protista Biotechnology AB (Lund, Sweden) in dry state. 2.2. BEs preparation BEs were prepared from E. coli cells according to Westphal and Jann (1965). Briefly, cells were grown in Luria Bertani media (LB) until the OD600 reached 0.6. The cells were centrifuged and the cell pellet was suspended in water. The cell suspension was treated with preheated 90% phenol solution. The mixture was vigorously shaken at a temperature of 65–70 ◦ C. The emulsion was then centrifuged for 30 min at 1000 × g and the aqueous layer containing endotoxin was separated. The aqueous extract was dialyzed and lyophilized. The lyophilized preparation contained about 50–60% bacterial RNA. To purify BEs further, the lyophilized preparation was dissolved in 5 ml 0.5 M NaCl. A 2% solution of hexadecyl-trimethylammonium bromide (CTAB) in 0.5 M NaCl was added until the amounts in weight between the CTAB and crude extract was about 1.5:1 and the precipitate was removed by centrifugation for 10 min at 11000 × g. The supernatant was poured into a 10-fold volume of ethanol. The precipitate formed was removed by centrifugation, dissolved in water, dialyzed against deioinized water, and lyophilized. 2.3. BEs assay The endotoxin content was assayed using a quantitative, chromogenic LAL assay (endpoint method) following the instructions of the supplier. The released amount of p-nitroaniline was measured at λmax = 405 nm and endotoxin from E. coli O111:B4 was used as standard. Endotoxin (LPS) was hydrolyzed by acetic acid hydrolysis. A 200 ␮l of purified extract was added to 800 ␮l of 1% SDS dissolved in 20 mM sodium acetate buffer pH 4.5 and the mixture was

423

heated at 100 ◦ C for 2 h. The mixture was lyophilized. The lyophilized substance was dissolved in 100 ␮l water. To remove SDS, 200 ␮l acidified ethanol (20 ml ethanol acidified with 100 ␮l of 4 M HCl) was added and centrifuged. The pellet was washed three times with 200 ␮l ethanol and was extracted with chloroform:methanol:water (12:6:1) mixture. The mass spectrometrical analysis of the endotoxin preparation was performed using a QSTAR® hybrid pulsar-i instrument (Applied Biosystems, Foster City, CA) equipped with a nano spray ion source. The endotoxin containing sample (10 ␮l) was desalted using a C18 zip tip (Millipore, Billerica USA) and was applied to a metal-coated fused silica capillary needle (Proxeon, Denmark). Data collection was performed with connected software (Analyte QS) from Applied Biosystems. Tandem mass spectrometry was performed on selected parent ions. 2.4. Ligand immobilization Sepharose 4B was activated with bisoxirane according to Hermanson et al. (1992) for the introduction of reactive epoxy groups. Briefly, 100 ml of Sepharose CL-4B was washed with water, suspended in 75 ml of 0.6 M NaOH containing 150 mg sodium borohydride with stirring. Seventy-five milliliters of 1,4-butandiol diglycidyl ether was added slowly with continuous stirring and the mixture was stirred overnight at room temperature. Finally, the reaction mixture was washed with water to remove non-reacted reagents. Cryogels were wetted with five column volumes (CVs) of deionized water and sodium phosphate buffer pH 7, respectively. Sodium phosphate buffer pH 7.2 containing either 10 mg/ml of PEI or lysozyme or 2 mg/ml polymyxin B was used for coupling ligands to epoxyactivated monolith columns and activated Sepharose 4B as described by Kumar et al. (2003). 2.5. Chromatographic experiments The chromatographic experiments were performed using a Biologic DuoFlow Chromatographic System (Bio-Rad, Hercules, CA, USA). The supermacroporous monoliths with immobilized ligands were put into FPLC columns (Pharmacia, Uppsala, Sweden) with 10 mm inner diameter. Alternatively the column (2.5 cm × 1 cm i.d.) was packed with Sepharose CL-4B

424


with immobilized ligands. The columns were equilibrated with 10 column volumes using various buffers ranging from pH 2 to 11 (100 mM sodium phosphate, 100 mM glycine and 100 mM sodium acetate buffers). The samples were dissolved in the corresponding buffer and loaded to the column at various flow rates 0.1, 1 and 10 ml/min. Washing was performed with at least 10 CVs of the corresponding equilibrating buffer. Elution was carried out with 2 M NaCl at a flow rate of 1 ml/min. The absorbance was monitored at 280 nm and the fractions corresponding to the eluted peaks were collected. 2.6. BEs capture from BEs—bovine serum albumin solutions A solution of BE and bovine serum albumin was prepared in 0.1 M sodium phosphate buffer with and without 0.01 M EDTA at different pH values. The BEs concentration was adjusted to be equal to the dynamic capacities at 5% breakthrough of the corresponding columns. The solution was applied to monolith columns with immobilized ligands as mentioned above at a flow rate of 1 ml/min. Protein concentration was determined using BCA protein assay reagent according to the established method (Smith et al., 1985). BEs concentrations were assayed by measuring the absorbance at 280 nm taking into account the contribution of BSA into absorbance at 280 nm, which was calculated from the results of BCA protein assay. 2.7. BEs capture from E. coli cell lysate E. coli cells were cultivated as described above until the OD600 reached 4. Twenty-five milliliters portions of cultural liquid were centrifuged and the pellets were suspended each in 1 ml of the running buffer (0.1 M sodium phosphate pH 7.2 alternatively 0.1 M glycine buffer pH 3.6). The suspensions were sonicated and the cell homogenates were applied directly to monolith columns with immobilized ligands at a flow rate of 1 ml/min. The columns were washed with the 10 CVs of the corresponding running buffer. Elution was performed with 2 M NaCl in the corresponding running buffer. Absorbance at 280 nm was measured and BEs were assayed using LAL test.

2.8. Cleaning in place and ligand leakage Ten column volumes of 0.5 M NaOH were applied to supermacroporous monolith with immobilized PEI at flow rate of 4 ml/min, followed by deionized water. Sodium hydroxide eluate was collected and neutralized. Part of the neutralized eluate was dialyzed, freeze dried, resuspended in 1 ml of 0.1 M sodium carbonate buffer pH 9 and finally filtered using a filter disk with pores of 0.2 ␮m. Half milliliter of neutralized or concentrated eluate was added to 0.5 ml of 170 ␮M TNBS dissolved in 0.1 M sodium carbonate buffer pH 9, incubated for 5 min at room temperature. The presence of PEI was determined by measuring absorbance at 350 nm. 3. Results and discussion 3.1. BEs preparation and characterization BE has been purified from E. coli cells and LAL test confirmed the endotoxin activity in purified preparation to be 5 × 107 endotoxin units/ml. The structure of purified endotoxin was confirmed by mass spectrometry. The mass spectrum had a signal peak at m/z 1797 (Fig. 1 and structure I Fig. 2B), which was corresponding to lipid A moiety (Karibian et al., 1999). The negative-ion mode of mass spectrometry promoted the fragmentational cleavage between and within the glucosamine residues resulting in the signal peak at m/z 710 (structure IV Fig. 2B) which represented the glucosamine I moiety (reducing) (Karibian et al., 1999). Signal peak at m/z 1244 (structure II Fig. 2B) corresponded to lipid A with the following modifications: the glucosamine I (reducing) loses phosphate group at position 1 and the double bond is formed between C1 and C2, glucosamine II (distal) loses the primary fatty acid 3-hydroxy fatty acid (3-OH C14 ) in position 3 and the double bond is formed between C3 and C4 . The signal peak at m/z 1226 (structure III Fig. 2B) corresponded to the following: the loss of primary 3-hydroxy fatty acid at position 3 and phosphate group at position 1 in glucosamine I (reducing), the loss of the secondary fatty acid (C14 ) attached to primary fatty acid in position 3 of glucosamine II (distal) and the formation of double bonds between C1–C2, C3–C4 and C3–C4 in primary fatty acid at position C3 (Lee et al., 2004) (Fig. 2).


425

Fig. 1. Mass spectrum of purified endotoxin showing lipid A peak at m/z 1797 and the structure of lipid A.

3.2. Chromatographic behavior of BE on supermacroporous monolith columns with different immobilized ligands The chromatographic behavior of purified BE has been studied on affinity columns produced by immobilization of polymyxin B, lysozyme or polyethyleneimine to supermacroporous polyacrylamide monoliths, so called cryogels. Cryogels (from the Greek ␬␳␫о␴ (kryos) meaning frost or ice) are produced by polymerization of monomers and crosslinker at subzero temperatures when most of the solvent, water, is frozen while the dissolved substances (monomers or polymers) are concentrated in small non-frozen regions. The reaction proceeds in these non-frozen regions while the crystals of frozen solvent perform like porogen. After melting the ice crystals, a system of large interconnected pores is formed. The

size and shape of the pore formed depends on the initial concentration of reagents in solution and the freezing conditions. The cryogel have large interconnected pores of 5–100 ␮m (supermacroporous), nonporous and dense walls that were clearly shown by SEM picture (Fig. 3). The large interconnected pores endow the monoliths cryogels with unique elastic and spongy morphology that make these monoliths fundamentally different from the well known rigid poly (glycidyl methacrylate) monoliths (Svec and Frechet, 1999; Peters et al., 1997, 1999) or superporous molded monoliths agarose (Gustavsson and Larsson, 1999; Gustavsson et al., 1998). The most part of water (70%) is squeezed mechanically from the pAAm monoliths. The functional epoxy groups on the surface of the monoliths allowed their modification with required ligands at high pH. The cryogel allows direct processing of cell suspensions with no need in pretreatment with

426


flow rates up to 2000 cm/h at normal pressure chromatography mode (Arvidsson et al., 2003; Dainiak et al., 2004; Plieva et al., 2004a,b). Thus supermacroporous monoliths are promising as matrices for negative chromatography aimed at capturing BEs. Three ligands selected were polymyxin B, a cyclic cationic decapeptide antibiotic with pK >9 for the ␣-amino group and pK = 10 for ␦-amino group of diaminobutyric acid, respectively (Petsch et al., 1997); PEI is a highly branched polymer (the ratio of primary, secondary and tertiary amines is 1:2:1) of molecular weight 60 kDa with pK >9 for primary amino groups and pK >10.5 for secondary amines (Petsch et al., 1997, 1998b); and lysozyme from hen egg white is a small and relatively stable protein with MW 14.4 KDa and pI of 11.2. Lysozyme is capable of binding oligosaccharides and hydrolyzing preferentially the ␤-1,4 glucosidic linkages between N-acetylmuramic acid and

N-acetylglucoseamine which occur in the mucopeptide cell wall structure of certain microorganisms. There are 39 amino groups (with equal or more than 30% accessibility) at the protein surface available for binding negative moieties. Thus, ligands are positively charged under neutral conditions. However, BEs are negatively charged under neutral conditions due to phosphate groups with pK1 1.3 and pK2 8.2 (Hou and Zanieweki, 1990; Petsch et al., 1998a). Apart of electrostatic interactions, one could expect some BEs binding to PEI or polymyxin B via hydrophobic interactions (Hirayama and Sakata, 2002) and some BEs binding to lysozyme due to the protein affinity towards oligosaccharides. All three ligands bind BEs efficiently at pH 7.2, the bound BEs were quantitatively eluted with 2 M NaCl (Fig. 4). As the columns were saturated with BEs (BEs were applied until nearly the complete breakthrough) the eluted peaks represented the static binding capacity,

Fig. 2. (A) Negative-ion mode nanospray mass spectrum of lipid A; (B) the fragmentational structures corresponding to the peaks.


427

Fig. 2. (Continued ).

which was much higher in the case of PEI as compared to the other two ligands. There was no BEs binding to PEI-ligands or lysozyme-ligands at high pH values probably due to the deprotonation of ligands and their loss of positive charge. Contrary, below pH 7, the efficiency of BEs capture improved with decreasing pH as judged by decreasing area of small breakthrough peaks (Fig. 5A and B). The BEs aggregation in aqueous media was presumed to be the reason for the observed breakthrough peaks. The large micelle-like structures were formed by BEs due to non-polar interactions between neighboring alkyl chains as well as due to the bridging between phosphate groups by bivalent cations like

Ca(II) (Petsch and Anspach, 2000; Santos et al., 2003). The micelle could pass non-hindered through the pores of superporous monolith columns without being bound by the ligands. To investigate the effect of micelle formation on chromatographic profile, the eluted fractions from the column with immobilized lysozyme were combined, dialyzed, lyophilized, dissolved in sodium phosphate buffer pH 7.2 and re-loaded on the column. Despite that all the material applied on the column has been already once bound to the column, the breakthrough peak was still observed, indicating that whatever was in the peak was aggregated in situ. Mass spectrometry confirmed the presence of BEs in all three fractions—eluted after the first chromatographic run, breakthrough and eluted fractions after

428


Fig. 3. Scanning electron microphotograph of the radial cross-sections of cryogel. The sample was fixed in 2.5% glutaraldehyde in 0.12 M sodium phosphate buffer, pH 7.2 overnight, post-fixed in 1% osmium tetroxide for 1 h, dehydrated in ethanol and critical point dried. The dried sample was coated with gold/palladium (40/60) and examined using a JEOLJSM-5600LV scanning electron microscope.

re-chromatography (data not shown). This observation agreed well with the data on the formation of BEs aggregates with a size of 119 nm at concentrations below 14 ␮g/ml and larger micelle-like aggregates

Fig. 4. Chromatographic profiles of endotoxin breakthrough and elution on superporous monolith columns with immobilized PEI (solid line), lysozyme (dash line) and polymyxin B (dotted line); arrow indicates the start of elution. Experimental conditions: 0.1 M sodium phosphate buffer pH 7.2 was used as running buffer and elution was carried out with 2 M NaCl in the running buffer at a flow rate of 1 ml/min.

of 190 nm in size at concentrations above 40 ␮g/ml (Santos et al., 2003). The dissociation constant for BEs aggregation was estimated to be 34 nM at pH 7.4 and was relatively independent of temperature and ionic strength (Takayama, 1996). The dissociation rate of micelles is very low (Petsch and Anspach, 2000). The BEs micelles even at low BEs concentrations have higher stability than micelles formed by simple detergents. The addition of a surfactant, Tween 80 in concentrations 0.5 or 5 of its critical micelle concentration (CMC) had no effect on the breakthrough portion of BEs when the sample was applied to a lysozyme–ligand column (data not shown). Thus, acidic conditions happened to be the most appropriate for BEs capture probably due to the decreased tendency of micelle formation. However, the feasibility of using this technique for decontamination of target protein from BEs depends strongly on the stability of the target proteins. Two pH values were selected for further studies, pH 7.2 as the medium which is appropriate for the decontamination of target protein as most of the recombinant proteins are stable under neutral conditions, and pH 3.6 as the medium which is appropriate for the decontamination


429

Fig. 6. The pressure drop for superporous monolith column (closed square) and a column of similar size (2.5 cm × 1 cm i.d.) packed with Sepharose CL-4B (open circle). The experiments were carried out on Bio-Rad FPLC system, where the flow rate was increased stepwise every one minute. The backpressure was recorded and the experiment was stopped when the Sepharose CL-4B bed was compressed.

Fig. 5. Chromatographic profiles of endotoxin breakthrough and elution on superporous monolith columns with immobilized PEI (A) and lysozyme (B) at different pH values. Experimental conditions: BEs were applied in 1 ml of the running buffer followed by washing and elution with 2 M NaCl in the running buffer at a flow rate of 1 ml/min. Sodium phosphate buffers are as follows: pH 11 (open circle with dotted line), pH 7.2 (open rhomb with dotted line), pH 4.6 (open square with dotted line), and pH 3.8 (open triangle with dashed line). Glycine buffers are as follows: pH 3 (closed square with straight line) and pH 2 (closed triangle with dotted line).

of wastewater, where the stability of other components is irrelevant. The most important characteristic of the chromatographic decontamination process is the dynamic capacity of the column which indicates how fast and how efficient BEs could be captured. The dynamic capacities at 5% BEs breakthrough level of the columns with different ligands are presented at two flow rates 1 and 10 ml/min, respectively (Table 1). For comparison, the capacities for the column packed with Sepharose CL-4B with the same ligands immobilized are presented at flow rate of 1 ml/min. It was impossible to use Sepharose column at flow rates above 10 ml/min due to the progressively increased pressure drop and finally collapse of the bed (Fig. 6). At low flow rate, the dynamic capacity of the column with immobilized polymyxin B was higher at pH 7.2

Table 1 Relative dynamic capacities at 5% BEs breakthrough for cryogel and Sepharose columns with immobilized polyethyleneimine, lysozyme and polymyxin B at different pH values and different flow rates Ligands

Polyethyleneimine

Polymyxin B

Matrix

Cryogel

Sepharose

Cryogel

pH

3.6

7.2

7.2

3.6

7.2

Flow rate ml/min 1 5.5 10 5.5

1.7 3.3

6 –

8.4 5.8

3.5 2.9

Lysozyme Sepharose

Cryogel

7.2

3.6

7.2

7.2

2.3 3.5

2 3

6 –

12 –

Relative dynamic capacities are presented as absorbance units of BE at 280 nm per 1 ml of column volume.

Sepharose

430


than that of the columns with immobilized PEI and lysozyme, respectively, probably due to some specific interactions of polymyxin B with the lipid A part of BEs on top of ion-exchange interactions characteristic for all three ligands (Petsch et al., 1998c). However at pH 3.6, the differences in capacities were less pronounced. At low flow rate, the dynamic capacities of Sepharose-based absorbents were higher than that of supermacroporous monolith columns for all three ligands. It is not surprising since Sepharose CL-4B has much smaller pores than supermacroporous monoliths and hence larger surface available for ligand immobilization. On the other side, smaller pores in Sepharose CL-4B entail higher pressure drops in Sepharosepacked columns as compared to supermacroporous monolith columns and impede using Sepharose-packed columns at flow rate as high as 10 ml/min. However, the supermacroporous columns perform at high flow rate nearly as well as at low flow rate for polymyxin B-ligand. It is interesting to note, that for PEI- and lysozyme-ligands, the capacity even increased with the flow rate. A similar behavior has been observed earlier for supermacroporous monoliths with ion-exchange or iminodiacetate ligands (Plieva et al., 2004b). The rationale behind could be as follows. As the supermacroporous monoliths have very large pores and present a very low flow resistance, at low flow rates the liquid could pass through the column using only some of the pores, probably the largest ones, so some of the smaller pores are not exposed to the mobile phase resulting in a decrease in the apparent capacity. With increasing flow rate, the mobile phase starts passing through smaller pores as well as through the larger ones. Hence, more ligands are exposed to the mobile phase and more BEs could be bound, the capacity increases. The even further decrease in capacity

for supermacroporous monolith column with immobilized PEI from 1.7 to 0.5 absorbance units, respectively, when the flow rate was decreased to an extremely small value of 0.1 ml/min indicates in favor of this assumption. 3.3. BEs capture from protein solutions A prerequisite for decontamination of protein solutions from BEs is a fast and selective binding of endotoxin to the adsorbent without affecting the recovery of the protein. BSA, a slightly acidic protein with pI of about 4.6 (Chaiyasut and Tsuda, 2001), was used as a model protein for optimizing the conditions for BEs decontamination (Table 2). Lysozyme happened to be the least suitable ligand for BEs capture in the presence of protein at pH 7.2, allowing only for around 60% of BEs to be captured. The incomplete BEs capture could be attributed to BEs binding to BSA with a dissociation constant KD ∼4.3 × 10−6 M under neutral conditions (Petsch et al., 1998c). It was reported that, the binding due to the electrostatic interaction between lysozyme and lipid A-phosphate groups with a [LPS]:[lysozyme] molar ratio 3:1 (Brandenburg et al., 1998). However, the association constant for BEs to bind to PEI immobilized on Nylon membrane is at least two order of magnitude higher than that for binding to immobilized lysozyme (Petsch et al., 1998a). The efficient (around 95%) capture of BEs along with quantitative protein recovery in the eluate was achieved either at pH 4.7 and 6 for lysozyme, or pH 4.7 and 7.2 for PEI, respectively. While the dissociation constant of polymyxin B–BEs is 2.46 × 10−9 M (Minobe et al., 1988), quantitative recovery of bound BEs were obtained at pH 6 and 7.2. In these cases probably, the stripping off BEs from the complex with BSA takes

Table 2 Recoveries of BSA and BEs in different fractions after passing through cryogel columns with immobilized polyethyleneimine, polymyxin B and lysozyme at a flow rate of 1 ml/min pH

7.2 7.2 + 0.01 M EDTA 6 4.7

Polyethyleneimine (PEI)

Polymyxin B

Lysozyme

BSA (breakthrough)

BSA eluted

BEs eluted

BSA (breakthrough)

BSA eluted

BEs eluted

BSA (breakthrough)

BSA eluted

BEs eluted

100 100 100 ∼100

0 0 0 5

100 100 85 100

100 100 ∼100 92

0 0 8 6

100 100 100 91

100 100 97 ∼100

0 5 0 14

63 90 94 100

Recoveries are presented in percentage taking the applied amounts of BSA and BEs as 100%, respectively.


431

Table 3 BEs capture from E. coli cell lysate BEs, EU/ml Polyethyleneimine pH Applied Breakthrough

3.6 4,100,000 101; (++) Nf /Ni of 11–100; (+) Nf /Ni of 2–10; (−) no proliferation. The amplification scores of the different cells in both configurations are similar.

442

S.-M. Chia et al. / Journal of Biotechnology 118 (2005) 434–447 Table 2 Collagen concentrations that support cell proliferation in 3D cultures Collagen concentration (mg/mL)

Amplification score (3D)

0.5 1.0 1.5

+ ++ +++

With the optimal cell seeding density of 1 × 106 cells/mL, the effects of different collagen concentrations on the rate of amplification of the cells in the microcapsules are tabulated. Cells could proliferate in the collagen concentration range of 0.5–1.5 mg/mL, with the optimal concentration at 1.5 mg/mL. Amplification score is defined as the ratio of the final cell number/initial cell number (Nf /Ni ). (+++) Nf /Ni of >101; (++) Nf /Ni of 11–100; (+) Nf /Ni of 2–10.

Fig. 4. Cells were greatly amplified in the microcapsules without splitting. (A) PC12, HepG2 and BMMNCs cells were cultured for 7 days in microcapsules and 2D culture without splitting. The amplification of: () 2D cultured cells; ( ) cells cultured in microcapsules with the slightly methylated collagen; ( ) cells cultured in microcapsules with highly methylated collagen. All three types of cells were amplified greatly in microcapsules but not in 2D culture. The highly methylated collagen could double the amplification of cells cultured in microcapsules. (B) Cell re-proliferation in the microcapsules. PC12, HepG2 and BMMNCs cells harvested from the microcapsules were microencapsulated again, and the cell amplification was quantified. The amplification of cells cultured in microcapsules with: ( and ) slightly methylated collagen in the completely dissociated and incompletely dissociated states, respectively, and ( and ) highly methylated collagen in the completely dissociated and incompletely dissociated states, respectively. Harvested cells in single cell state re-proliferated to the full potential while the cell aggregates exhibited a reduced level of re-proliferation.

163 ± 12-, 295 ± 19- and 170 ± 9-fold, respectively (Fig. 4A). The higher level of collagen methylation could increase the cell amplification in the microenvironment. The highly methylated collagen exhibits

more net positive charges, which might modify the ionic equilibrium of the cells that in turn stimulates the cell proliferation. Alternatively, the highly methylated collagen can affect the cell–collagen interactions that link to the signalling pathways regulating cell proliferation. The exact mechanisms of the cell proliferation stimulation by collagen methylation are being studied. To investigate whether the 3D cultured cells harvested from the microenvironment can re-proliferate in further passages, we re-encapsulated the harvested 3D cultured cells, and allowed them to proliferate again in the microenvironment for 7 days. The BMMNCs, PC12 and HepG2 cells were amplified 89 ± 6-, 150 ± 12- and 100 ± 8-fold, respectively, in the microcapsules with slightly methylated collagen and 150 ± 12-, 295 ± 24- and 220 ± 19-fold, respectively, with highly methylated collagen (Fig. 4B). When the harvested cells were not completely dissociated and many remain as cell aggregates, the BMMNCs, PC12 and HepG2 cells were amplified 50 ± 4-, 55 ± 6- and 50 ± 5-fold, respectively, in the microcapsules with slightly methylated collagen and 100 ± 8, 250 ± 10 and 150 ± 12 times, respectively, with highly methylated collagen. The completely dissociated cells could re-proliferate to the full potential (as in Fig. 4A and B) while the incompletely dissociated cells exhibited a reduced level of re-proliferation (Fig. 4B). It was interesting to note that the highly methylated collagen could enhance the cell amplification with no significant differences in cell amplification in completely or incompletely dissociated states (Fig. 4).

S.-M. Chia et al. / Journal of Biotechnology 118 (2005) 434–447

3.3. Exhibition of extended proliferation lifespan with higher viability of 3D cultured cells Like most adult stem cells, BMMNCs in 2D culture have limited proliferation lifespan of approximately 10–13 passages (Pittenger et al., 1999). We have observed a decreased in the viability of the cells when repeated cultured in 2D as compared to cells cultured in 3D (Fig. 5A). We hypothesized that the limited proliferation lifespan of adult stem cells in vitro might partially be imposed by the repeated disruptive detachment procedures in the conventional 2D culture.

443

We would like to investigate whether the 3D cultured BMMNCs harvested from the microenvironment would exhibit extended proliferation lifespan. Since the typical proliferation rate for BMMNCs in 2D culture is 48 h per cell doubling generation, each passage in 2D culture is equivalent to one cell doubling generation. We have compared the proliferation rates of the 3D cultured BMMNCs and the 2D cultured BMMNCs as they were plated onto the cell culture polystyrene dish. Indeed, the 2D cultured BMMNCs showed a reduced rate of cell proliferation 6 generations after the onset of the comparative experiment (10 generations after

Fig. 5. 3D cultured BMMNCs harvested from the microenvironment exhibited extended proliferation lifespan with higher viability. (A) (Panel a) The 3D cultured cells stained with cell tracker green and propidium iodide and (Panel b) 2D cultured cells. Scale bars represent 5 ␮m. (B) () The rate of proliferation per 24 h of the 3D cultured cells after the onset of the comparative experiment and () the rate of proliferation per 24 h of the 2D cultured cells after the onset of the comparative experiment. The 3D cultured BMMNCs cells maintained the same proliferation rate as when they were first isolated from bone marrow, while the 2D cultured cells stopped proliferation after seven generations from the onset of the comparative experiment.

444


isolation from bone marrow) and almost completely stopped proliferation after 7 generations (11 generations after isolation from bone marrow) (Fig. 5B). For 3D cultured BMMNCs, the proliferation rate is 24 h per cell doubling generation (Fig. 5B). Thus, each passage every 7 days is equivalent to seven cell doubling generations. After a month of continuous culture (30 generations) in the microenvironment, the 3D cultured BMMNCs were harvested and plated onto the cell culture polystyrene dish. Seven generations after the onset of the comparative experiment (37 generations after isolation from bone marrow), the 3D cultured BMMNCs maintained the same proliferation rate (Fig. 5B) as when they were first isolated from bone marrow. Therefore, we can greatly extend the proliferation lifespan of BMMNCs, to amplify them and potentially other anchorage-dependent mammalian cells to sufficient numbers for therapeutic applications. The potential to greatly expand rare adult stem cells from various tissue sources will be important for stem cell research. 3.4. Exhibition of superior attachment kinetics, cell morphology and specific functions of 3D cultured cells For tissue engineering applications, it is highly desirable to transplant harvested cells that mimic the characteristics in vivo. We hypothesized that the 3D cultured cells would exhibit better attachment kinetics, cell morphology and functions than the conventionally cultured cells that were harvested with disruptive detachment procedures. Therefore, we compared the characteristics of PC12 cells and rat hepatocytes cultured on specific ligand-conjugated polymer surfaces. Rat hepatocytes were used here in place of HepG2 because rat hepatocytes have well-characterized attachment characteristics to specific ligands (Kobayashi et al., 1992). Rat hepatocytes can also be characterized for the cytochrome P450-dependent mono-oxygenase activity that is a good indicator of the detoxifying functions of hepatocytes (Selden et al., 1999). Rat hepatocytes specifically bind to motifs on certain sugars such as lactose and galactose (Bennatt et al., 1997). PC12 cells interact with the Tyr–Ile–Gly– Ser–Arg (YIGSR) peptide-binding domain of the laminin. Thus, we chemically conjugated the lactose and YIGSR fragment of the laminin onto polyester

Fig. 6. 3D cultured cells harvested from the microenvironment exhibited superior attachment kinetics, cell morphology and specific functions than the 2D cultured cells. (A) Attachment of PC12 cells onto YIGSR conjugated to polyester coverslip: ( and ) the attachment of the 3D cultured PC12 cells onto YIGSR-conjugated and plain PET coverslips, respectively, and ( and ) the attachment of the 2D cultured PC12 cells onto YIGSR-conjugated and plain PET coverslips, respectively. (B) Attachment of rat hepatocytes on lactose-conjugated PET coverslips: ( and ) the attachment of 3D cultured rat hepatocytes onto lactose-conjugated and plain PET coverslips, respectively, and (♦ and ×) the attachment of the 2D cultured rat hepatocytes onto lactose-conjugated and plain PET coverslips, respectively. The 3D cultured PC12 cells and rat hepatocytes exhibit better attachment to ligand-conjugated PET coverslip than the 2D cultured cells.

(PET) membranes and measured the kinetics of the cell attachment to these membranes. Indeed, the 3D cultured cells exhibited superior attachment kinetics than the 2D cultured cells (Fig. 6A and B) on binding to these ligand-conjugated polymer surfaces.


445

Fig. 7. Morphology and neurite extension of PC12 cells. (Panel a) 3D cultured PC12 cells 3 h after attached to YIGSR-conjugated PET coverslip. (Panel b) 2D cultured PC12 cells. The 3D cultured PC12 cells have secreted significant amount of the extra-cellular matrices than the 2D cultured factor induction. (Panel d) 2D cultured PC12 cells. The maximum cells. (Panel c) 3D cultured PC12 cells 48 h after nerve growth length = (length of the longest neurite on each cell)/n. The mean length = (mean length of all the neurites on each cell)/n, n = 50 cells. The 3D cultured PC12 cells show longer neurite extension than the 2D cultured cells.

The morphology of PC12 cells changes in response to extra-cellular microenvironment (Berthiaume et al., 1996). We have examined the cell morphology of PC12 cells as they attached to the ligand-conjugated PET membranes. Three hours after the cells were incubated with the membranes, the cells harvested from 2D culture started to attach with spread morphology (Fig. 7). In contrast, the 3D cultured PC12 cells exhibited the healthy round morphology by binding to significant amount of extra-cellular matrices that might have been secreted by the cells themselves (Fig. 7). Un-differentiated PC12 cells normally exhibit round morphology until after many passages (Kozlowski et al., 1989). Therefore, the 3D cultured cells seem to preserve better cell morphology and matrix-secretion than the 2D cultured cells. The PC12 cells can be induced to form neurites by nerve growth factor (Baldwin et al., 1996). The max-

imum and mean lengths of the neurite extension are good measures of the PC12 cell functions (Leoni et al., 1999). We have measured the lengths of the neurites extended from PC12 cells 48 h after induction by nerve growth factor. The neurites extended from the 3D cultured PC12 cells were two-fold longer, in terms of both maximum and mean lengths, than the neurites extended from the 2D cultured PC12 cells (Fig. 7). We have also measured the cytochrome P450dependent mono-oxygenase activity of the rat hepatocytes (Wang et al., 1997). The 3D cultured rat hepatocytes exhibited a two-fold higher detoxification function than the 2D cultured cells 3 h after attaching to the lactose-conjugated PET membranes (data not shown). Therefore, the 3D cultured cells have demonstrated improved substrate attachment kinetics, better preserved cell morphology and other specific functions than the conventionally cultured cells.

446


3.5. Conclusion The three-dimensional culture system presented here allows cell growth and proliferation in three dimensions, allowing cell splitting without subjecting cells to disruptive conditions that potentially affects cell structures and functions. The matrix used in this system can provide a good microenvironment for culturing of anchorage-dependent mammalian cells (Toh et al., 2005). Numerous efforts have been undertaken to develop novel biomaterials and substrate surfaces for improving cell attachment and functions (Folch and Toner, 2000; Kam et al., 2002). The 3D cultured cells exhibiting the improved attachment kinetics can complement these materials science and engineering efforts. Multiple types of cells can be seeded sequentially within a short period of time to better control the formation of complex tissue constructs. These 3D cultured cells should have intact cell surface receptors and other membrane components that can enhance the interaction between these 3D cultured cells and other cells or extra-cellular matrices to generate functioning tissues when transplanted in vivo for tissue repair and regeneration. Furthermore, the microencapsulated cells can be cultured in well-established macro-environments such as the packed-bed and fluidized-bed bioreactors for large-scale production of highly functional anchorage-dependent cells for various applications. We conclude that a microcapsule-based threedimensional microenvironment has been engineered for expansion of sensitive anchorage-dependent mammalian cells. These cells exhibit superior characteristics that would be useful in cell-based therapeutics, tissue engineering and stem cell research.

Acknowledgements The authors thank members of the Yu Laboratory for technical assistance and Dr. Jan-Thorsten Schantz for providing the BMMNCs. This study is supported, in part, by research grants from the National Medical Research Council and the Biomedical Research Council, and institutional budgets from the Institute of Bioengineering and Nanotechnology (IBN) of the Agency for Science, Technology and Research.

Contract grant sponsors are the National Medical Research Council and Biomedical Research Council of A*STAR, Republic of Singapore, Singapore. References Ali, S.A., Pappas, I.S., Parnavelas, J.G., 1998. Collagen type IV promotes the differentiation of neuronal progenitors and inhibits astroglial differentiation in cortical cell cultures. Brain Res. Dev. 110 (1), 31–38. Avital, I., Inderbitzin, D., Aoki, T., Tyan, D.B., Cohen, A.H., Ferraresso, C., Rozga, J., Arnaout, W.S., Demetriou, A.A., 2001. Isolation, characterization, and transplantation of bone marrowderived hepatocyte stem cells. Biochem. Biophys. Res. Commun. 288 (1), 156–164. Baba, M., Hirai, S., Kawakami, S., Kishida, T., Sakai, N., Kaneko, S., Yao, M., Shuin, T., Kubota, Y., Hosaka, M., Ohno, S., 2001. Tumor suppressor protein VHL is induced at high cell density and mediates contact inhibition of cell growth. Oncogene 20 (22), 2727–2736. Baldwin, S.P., Krewson, C.E., Saltzman, W.M., 1996. PC12 cell aggregation and neurite growths in gels of collagen, laminin and fibronectin. Int. J. Dev. Neurosci. 14, 351–364. Baldwin, S.P., Saltzman, W.M., 2001. Aggregation enhances catecholamine secretion in cultured cells. Tissue Eng. 7 (2), 179–190. Barhoumi, R., Bowen, J.A., Stein, L.S., Echols, J., Burghardt, R.C., 1993. Concurrent analysis of intracellular glutathione content and gap junctional intercellular communication. Cytometry 14 (7), 747–756. Bennatt, D.J., Ling, Y.Y., McAbee, D.D., 1997. Isolated rat hepatocytesbind lactoferrins by the RHL-1 subunit of the asialoglycoprotein receptor in a galactose-independent manner. Biochemistry 36, 8367–8376. Berthiaume, F., Moghe, P.V., Toner, M., Yarmush, M.L., 1996. Effect of extracellular matrix topology on cell structure, function, and physiological responsiveness: hepatocytes cultured in a sandwich configuration. FASEB J. 10, 1471–1484. Bissell, M.J., Barcellos-Hoff, M.H., 1987. The influence of extracellular matrix on gene expression: is structure the message? J. Cell. Sci. Suppl. 8, 327–343. Burdsal, C.A., Damsky, C.H., Pedersen, R.A., 1993. The role of E-cadherin and integrins in mesoderm differentiation and migration at the mammalian primitive streak. Development 118 (3), 829–844. Burghardt, R.C., Barhoumi, R., Lewis, E.H., Bailey, R.H., Pyle, K.A., Clement, B.A., Phillips, T.D., 1992. Patulin-induced cellular toxicity: a vital fluorescence study. Toxicol. Appl. Pharmacol. 112 (2), 235–244. Campenot, R.B., Eng, H., 2000. Protein synthesis in axons and its possible functions. J. Neurocytol. 29 (11–12), 793–798. Chia, S.M., Leong, K.W., Li, J., Xu, X., Zeng, K.Y., Er, P.N., Gao, S.J., Yu, H., 2000. Hepatocyte encapsulation for enhanced cellular functions. Tissue Eng. 5 (6), 481–495. Dai, W., Belt, J., Saltzman, W.M., 1994. Cell-binding peptides conjugated to poly(ethylene glycol) promote neural cell aggregation. Biotechnol. (N. Y.) 12, 792–801.

S.-M. Chia et al. / Journal of Biotechnology 118 (2005) 434–447 Folch, A., Toner, M., 2000. Microengineering of cellular interactions. Annu. Rev. Biomed. Eng. 2, 227–256. Gumbleton, M., Audus, K.L., 2001. Progress and limitations in the use of in vitro cell cultures to serve as permeability screen for the blood–brain barrier. J. Pharm. Sci. 90 (11), 681–698. Gustavson, K.H., 1956. The Chemistry and Reactivity of Collagen. Academic Press Inc., New York. Hartmann-Petersen, R., Walmod, P.S., Berezin, A., Berezin, V., Bock, E., 2000. Individual cell motility studied by time-lapsed video recording: influence of experimental conditions. Cytometry 40, 260–270. Ishaug-Riley, S.L., Crane-Kruger, G.M., Yaszemski, M.J., Mikos, A.G., 1998. Three-dimensional culture of rat calvarial osteoblasts in porous biodegradable polymers. Biomaterials 19 (15), 1405–1412. Jaroszeski, M.J., Gilbert, R., Fallon, P.G., Heller, R., 1994a. Mechanically facilitated cell–cell electrofusion. Biophys. J. 67 (4), 1574–1581. Jaroszeski, M.J., Gilbert, R., Heller, R., 1994b. Detection and quantitation of cell–cell electrofusion products by flow cytometry. Anal. Biochem. 216 (2), 271–275. Jirosva, K., Sodaar, P., Mandys, V., Bar, P.R., 1997. Cold jet: a method to obtain pure Schwann cell cultures without the need for cytotoxic, apoptosis-inducing drug treatment. J. Neurosci. Methods 78, 133–137. Kam, L., Shain, W., Turner, J.N., Bizios, R., 2002. Selective adhesion of astrocytes to surfaces modified with immobilized peptides. Biomaterials 23 (2), 511–515. Karkkainen, M.J., Makinen, T., Alitalo, K., 2002. Lymphatic endothelium: a new frontier of metastasis research. Nat. Cell Biol. 4 (1), E2–E5. Kicic, A., Shen, W., Rakoczy, P.E., 2001. The potential of marrow stromal cells in stem cell therapy. Eye 15 (Pt 6), 695–707. Kobayashi, A., Kobayashi, K., Akaike, T., 1992. Control of adhesion and detachment of parenchymal liver cells using lactosecarrying polystyrene as substratum. J. Biomater. Sci. Polym. Ed. 3, 499–508. Kozlowski, M.R., Rosser, M.P., Hall, E., Longden, A., 1989. Effects of bradykinin on PC12 cell differentiation. Peptides 10, 1121–1126. Lapinsky, S.E., Glencross, D., Car, N.G., Kallenbach, J.M., Zwi, S., 1991. Quantification and assessment of viability of Pneumocystis carinii organisms by flow cytometry. J. Clin. Microbiol. 29 (5), 911–915. Leonard, D.W., Meek, K.M., 1997. Refractive indices of the collagen fibrils and extrafibrillar materials of the corneal stroma. Biophys. J. 72, 1382–1387. Leoni, C., Menegon, A., Benfenati, F., Toniolo, D., Pennuto, M., Valtorta, F., 1999. Neurite extension occurs in the absence of regulated exocytosis in PC12 subclones. Mol. Cell Biol. 10 (9), 2919–2931. Maria-Engler, S.S., Mares-Guia, M., Correa, M.L., Oliveira, E.M., Aita, C.A., Krogh, K., Genzini, T., Miranda, M.P., Ribeiro, M., Vilela, L., Noronha, I.L., Eliaschewitz, F.G., Sogayar, M.C., 2001. Microencapsulation and tissue engineering as an alternative treatment of diabetes. Braz. J. Med. Biol. Res. 34 (6), 691–697.

447

Melkonian, A.D., Gaylo, J.D., Cousins, R.B., Grant, M.H., 1994. Culture of a differentiated liver cell line, Hep G2, in serum with application to a bioartificial liver: effect of supplementation of serum with amino acids. Artif. Organs 18, 611– 617. Ng, S., Wu, Y.N., Zhou, Y., Toh, Y.E., Ho, Z.Z., Chia, S.M., Zhu, J.H., Mao, H.Q., Yu, H., 2005. Optimization of 3-D hepatocyte culture by controlling the physical and chemical properties of the extra-cellular matrices. Biomaterials 26 (16), 3153– 3163. Nyberg, S.L., Remmel, R.P., Mann, H.J., Peshwa, M.V., Hu, W.S., Cerra, F.B., 1994. Primary hepatocytes outperform Hep G2 cells as the source of biotransformation functions in a bioartificial liver. Ann. Surg. 220 (1), 59–67. Pittenger, M.F., Mackay, A.M., Beck, S.C., Jaiswal, R.K., Douglas, R., Mosca, J.D., Moorman, M.A., Simonetti, D.W., Craig, S., Marshak, D.R., 1999. Multilineage potential of adult human mesenchymal stem cells. Science 284 (5411), 143–147. Selden, C., Shariat, A., McCloskey, P., Ryder, T., Roberts, E., 1999. Three-dimensional in vitro cell culture leads to a marked upregulation of cell function in human hepatocyte cell-lines—an important tool for the development of a bioartificial liver machine. Ann. N. Y. Acad. Sci. 875, 353–363. Semler, E.J., Moghe, P.V., 2001. Engineering hepatocyte functional fate through growth factor dynamics: the role of cell morphologic priming. Biotechnol. Bioeng. 75 (5), 510–520. Shimizu, T., Yamato, M., Kikuchi, A., Okano, T., 2001. Twodimensional manipulation of cardiac myocyte sheets utilizing temperature-responsive culture dishes augments the pulsatile amplitude. Tissue Eng. 7, 141–151. Shofuda, K.I., Hasenstab, D., Shofuda, T., Kenagy, R.D., Clowes, A.W., 2001. Control of smooth muscle cell function by membrane-type matrix metalloproteinases. Ann. N. Y. Acad. Sci. 947, 337–340. Toh, Y.C., Ho, S.T., Zhou, Y., Hutmacher, D.W., Yu, H., 2005. Application of a polyelectrolyte complex coacervation method to improve seeding efficiency of bone marrow stromal cells in a 3D culture system. Biomaterials 26 (19), 4149–4160. Toma, J.G., Akhavan, M., Fernandes, K.J., Barnabe-Heider, F., Sadikot, A., Kaplan, D.R., Miller, F.D., 2001. Isolation of multipotent adult stem cells from the dermis of mammalian skin. Nat. Cell Biol. 3 (9), 778–784. Wang, L., Sun, J., Li, L., Koutalistras, N., Wang, C., Sheil, R., 1997. Evaluation of the methods for assessment of hepatocytes P450IA1, P450IIB1, viability and proliferation in bioartificial liver support system matrices. Transplant. Proc. 29 (1–2), 395–396. Ward, M.D., Hammer, D.A., 1993. A theoretical analysis for the effect of focal contact formation on cell–substrate attachment strength. Biophys. J. 64, 936–959. Waring, M.J., 1965. Complex formation between ethidium bromide and nucleic acids. J. Mol. Biol. 13 (1), 269–282. Yamato, M., Okuhara, M., Karikusa, F., Kikuchi, A., Sakurai, Y., Okano, T., 1999. Signal transduction and cytoskeletal reorganization are required for cell detachment from cell culture surfaces grafted with a temperature-responsive polymer. J. Biomed. Mater. Res. 44 (1), 44–52.


Production of glucuronan oligosaccharides using a new glucuronan lyase activity from a Trichoderma sp. strain C. Delattre, P. Michaud ∗ , J.M. Lion, B. Courtois, J. Courtois Laboratoire des Glucides-EPMV CNRS FRE 2779, IUT/GB, UPJV, Avenue des Facultés, Le Bailly, 80025 Amiens Cedex, France Received 10 August 2004; received in revised form 1 April 2005; accepted 6 April 2005

Abstract Sinorhizobium meliloti M5N1CS synthesizes a homopolymer of glucuronic acids ␤-(1,4) linked and variably C2 and/or C3 O-acetylated. To obtain ␤--(4,5)-unsaturated oligoglucuronans, various acetylated forms of this bacterial polymer were cleaved by a Trichoderma sp. GL2 glucuronan lyase. Oligomers with polymerization degrees up to 8 were then produced, purified by liquid chromatography (size exclusion and anions exchange) and characterized using 1 H NMR and ESI-Q/TOF-MS. Finally, the production (in gram quantity) of pure unsaturated oligoglucuronans non-acetylated (di- and trisaccharide) was investigated thanks to the complete depolymerization of deacetylated glucuronan. © 2005 Elsevier B.V. All rights reserved. Keywords: Glucuronan; Oligosaccharides; Polysaccharide lyases

1. Introduction Oligosaccharides from plants, algae, fungi and animals have been widely studied these last years for their potentialities as signaling molecules. For plants, oligosaccharides as ␤-oligoglucans, oligochitosans and oligogalacturonates are called oligosaccharins (Darvill Abbreviations: DEAE, di-ethyl-amino-ethyl; ESI-Q/TOF-MS, electrospray ionization quadrupole/time of flight mass spectrometry; HMW, high molecular weight; 1 H NMR, nuclear magnetic resonance of proton; LMW, low molecular weight; TLC, thin-layer chromatography ∗ Corresponding author. Tel.: +33 3 22 53 40 98; fax: +33 3 22 95 71 17. E-mail address: [email protected] (P. Michaud). 0168-1656/$ – see front matter © 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.jbiotec.2005.04.014

et al., 1992; Aldington and Fry, 1993; Ozeretskovskaya and Romenskaya, 1996) and are well known as elicitors of phytoalexins synthesis and reactive oxygen species (ROS) production (Darvill et al., 1992; Ebel and Corsio, 1994; Bolwell, 1999; Lee et al., 1999). More, oligosaccharins affect plant cell growth and development as, for example: flow formation, root organogenesis, auxin induced cell elongation and differenciation (Darvill et al., 1992; Côté and Hahn, 1994; Creelman and Mullet, 1997; Willats et al., 1999). For algae and seaweeds, a similar concept can be applied with analogous oligosaccharides but also with more specific structures as oligoalginates, oligocarrageenans and oligolaminarins (Potin et al., 1999). We noted that these “marine oligomers” are also elicitors

C. Delattre et al. / Journal of Biotechnology 118 (2005) 448–457

of terrestrial plants and used as seaweed fertilizers (Painter, 1993; Patier et al., 1993). Concerning Fungi, a notable role of several oligosaccharides (oligoalginates) was clearly identified in secondary metabolites production (Ariyo et al., 1998). Among animals, therapeutic actions as antitumoral agent, keratinocyte growth stimulator, antithrombotic and anticoagulant were, respectively, found with oligochitins (Suzuki et al., 1986), oligoalginates (Kawada et al., 1997), oligoglycosaminoglycans (Pineo and Hull, 1997), and oligofucans (Nardella et al., 1996). For all anionic oligosaccharides, we pointed out that only little information is available (excepted for oligoglycosaminoglycans, oligogalacturonates and oligoalginates) for two reasons. Firstly, the acidic polysaccharides and oligosaccharides are poorly represented in plants and are in majority microbial compounds. Secondly, as for microbial alginates (Gacesa, 1988), microbial polysaccharides possess a high degree of structural variability and complexity, notably with O-acetylation that decrease the cleavage by enzymes. These last are in majority polysaccharide lyases (PL) (Michaud et al., 2003). They depolymerize anionic polymers, by a ␤-elimination mechanism, leading to -(4,5)-unsaturated oligouronic acids. This reaction consists of a general base-catalyzed abstraction of the proton at C-5 of an uronic acid. An electron transfers from the carboxyl group to form a double bond between C-4 and C-5 results in the elimination of the 4-O-glycosidic bond and in the formation of 4-deoxy-l-erythro-hex-4-enopyranosyluronic acid. This reaction conducts to the formation of an unsaturated uronate at the newly generated non-reducing end (Sutherland, 1995; Michaud et al., 2003). Viewing all examples of biological activities, it appears necessary to produce and isolate rapidly new unsaturated anionic oligosaccharides, notably after cleavage by PL of anionic microbial polysaccharides. The Sinorhizobium meliloti M5N1CS (NCIMB 40472) strain, which induces the formation of effective nodules on alfalfa roots (Gonzales et al., 1996), produces the glucuronan as exopolysaccharide (Courtois et al., 1993). This anionic homopolymer is a (1 → 4)-␤-d-polyglucuropyranosyluronic acid variably O-acetylated at C-3 and/or C-2 position depending to the Mg2+ concentration in the culture medium (Michaud et al., 1995). In order to obtain oligoglucuronans, a bacterial polysaccharide lyase degrading

449

various acetylated (but not 2,3-di-O-acetylated) and deacetylated glucuronan was purified (Da costa et al., 2001). Because of the low activity of this enzyme, we have researched another way for the production of unsaturated oligoglucuronans. We now propose, in this paper, to study the enzymatic depolymerization of all glucuronans by a Trichoderma sp. glucuronan lyase (GL) in order to generate unsaturated oligoglucuronans.

2. Materials and methods 2.1. Glucuronan production, purification and characterization The S. meliloti M5N1CS mutant strain (NCIMB 40472) was cultivated at 30 ◦ C in a 20-l bioreactor (SGI, Toulouse, France) containing 15 l of Rhizobium complete (RC) medium (Courtois et al., 1983) supplemented with sucrose 1% (w/v) (RCS medium). MgSO4 ·7H2 O was added (0.15%, w/v) or not for the production of mainly 2,3-di-O-acetylated glucuronan (highly acetylated glucuronan) or 3-O-acetylated glucuronan (standard), respectively (Michaud et al., 1995). The inoculum was a 1.5 l of RCS medium inoculated with S. meliloti M5N1CS and incubated 20 h at 30 ◦ C on a rotary shaker (100 rpm). After 72–96 h of cultivation, the broths were centrifuged at 33,900 × g for 40 min at 20 ◦ C. The polysaccharides in the cell-free broths were then precipitated by addition of 3 isopropanol volumes and collected by centrifugation (33,900 × g for 40 min at 20 ◦ C). After freeze-dried, the glucuronan pellets were dissolved (4 g/l) in H2 O and the isopropanol precipitation was repeated twice. In order to produce deacetylated glucuronan, a standard glucuronan was treated overnight by KOH (2 M) at 50 ◦ C (pH 12) and purified as described above. 2.2. Preparation of crude enzyme fraction Trichoderma sp. GL2 strain was isolated from compost and grown on potato agar (PDA). To recover glucuronan cleavage activity, the strain was grown on 100 ml MM medium (Pentillä et al., 1987) supplemented with 1% (w/v) glucuronan as sole carbon source. After 48 h of incubation at 25 ◦ C, on a rotary

450


shaker (200 rpm), the extracellular medium was collected by centrifugation (10,000 × g, 20 min and 4 ◦ C) and concentrated to 10 ml by ultrafiltration in a stirred Amicon cell (Beverly, MA) through a 104 normalmolecular weight cutoff (NMWCO) polyethersulfone membrane from Amicon (Sortorius, Göttingen, Germany). After a wash step with H2 O, as described above, the enzymatic fraction was stored at −80 ◦ C. 2.3. Enzyme assay The GL activity of the enzyme was measured by monitoring the increase of absorbance at 235 nm using an Uvikon 930 spectrophotometer (Kontron, Montigny Lebretonneux, France). The reaction mixture was composed of 1 ml of 0.2% (w/v) glucuronan solution in 50 mM potassium acetate buffer pH 5.5 and an appropriate volume (5–10 ␮l) of enzyme preparation. According to literature (Hashimoto et al., 2003), 1 unit (U) of the enzyme activity was defined as the enzyme amount required to cause an increase of 1 in absorbance at 235 nm/min. Specific activity was expressed as units per milligram of protein. 2.4. Production of unsaturated oligoglucuronans Solutions of standard, highly acetylated and deacetylated glucuronans (3%, w/v) in 50 mM potassium acetate buffer pH 5.5 were incubated at 20 ◦ C with enzymatic preparation during various times (between 4 and 96 h). After incubation, the enzymatic ␤-elimination was stopped by dipping the reaction medium into a 95 ◦ C water bath. The mixture of oligomers was then centrifuged (15,000 × g for 20 min at 20 ◦ C) and the supernatant was recovered. 2.5. Purification of unsaturated oligoglucuronates All the purifications were performed at room temperature using low-pressure liquid chromatography system (Proteam LC system 210, Lincoln, NE). Unsaturated oligoglucuronates were size-fractionated by low-pressure gel-permeation chromatography on a Biogel P6 fine (Biorad) column (2.6 cm × 100 cm, Amersham Bioscience). The oligosaccharide mixture was loaded (100–500 mg in 10 ml) and eluted with a 50 mM ammonium formate solution at a flow rate of 0.8 ml/min. Detection was achieved

with a UV detector (UA-6 from ISCO) at 254 nm and with RI detector (Melz). Fractions (5 ml) were collected with a Foxy 200 (ISCO) collector. Fractions belonging to a same peak were pooled and freeze-dried. Unsaturated oligoglucuronates with low (8. The most representative dp was the dp 3 (26.9%). The DS and O-acetylation distributions of standard oligosaccharides (Table 2) and polysaccharide (Table 1) were compared. All standard unsaturated oligoglucuronans were predominantly 3-O-acetylated as the standard glucuronan and no significant differences were detected excepted a substitution reduced of around 10% for the dp 3–7. Depolymerization of the highly acetylated glucuronan was carried out as the standard one. In these conditions, the GL affinity with 2,3-di-O-acetylated

Table 2 DS, molar ratio of O-acetylated residues species and massic ratio of oligoglucuronans from standard glucuronan degradation dpa >8

8

7

6

5

4

3

2

DS (%) 3-O-Acb (%) 2-O-Ac (%) 2,3-di-O-Ac (%)

63.9 48.0 8.9 6.9

65.0 53.4 8.6 3.0

50.9 41.6 2.8 6.4

41.2 36.9 2.2 2.1

46.6 36.8 5.2 4.7

51.4 37.6 6.3 7.5

42.9 38.0 2.3 2.4

60.0 48.0 3.7 8.3

Massic ratio (%)

25.1

10.7

13.2

4.1

8.8

8.5

26.9

2.7

a b

Degree of polymerisation. Acetylated.


substrate was revealed by presence of H-4 signal characteristic of ␤-elimination (unpublished data). This result established the degradation of highly substituted glucuronans. Nevertheless, its cleavage generated an heterogeneous mixture of LMW with a very large distribution in high dp (>10). As for the standard polysaccharide, the fungal lyase preserved the distribution of the polysaccharide O-acetyl substituents. The high level of GL activity on deacetylated glucuronan authorized us to envisage a larger production of unsaturated deacetylated oligoglucuronans by comparison with acetylated ones. Consequently, 6 g of deacetylated polymer (real mass = 3.6 g depending on purity%) were incubated with 100 U of enzyme activity during 4, 8, 24 and 96 h. The oligomers formed were fractionated by size exclusion chromatography and revealed (Table 3) a large repartition of dp (dp 1 to >8) for short incubations (4 h). After 8 h, the dp 4, 3, 2 and 1 were the sole species identified in the degradation medium. The mixture of unsaturated oligoglucuronans was analysed by using of ESI-Q/TOF-MS. We noted for

453

Table 3 Massic ratio of oligoglucuronans from deacetylated glucuronan degradation after different incubation times Time (h)

4 8 24 96 a

Massic ratio (%) >8a

8a

7a

5a

4a

3a

2a

1a

23.4 – – –

1.8 – – –

2.5 – – –

7.2 – – –

13.6 12.0 – –

3.8 35.0 59.6 26.0

36.8 50.0 35.1 66.0

10.9 2.0 5.3 8.0

Degree of polymerisation.

long incubation times (24 h) the presence of three major molecular weights: [M − H]− : 175.40, 351.23, and 527.03. They were identified as: 4-deoxy-␤-d-hex4-enopyranosylglucuronate (G); 4-deoxy-␤-d-hex4-enopyranosylglucuronate-(1 → 4)-O-␤-d-glucuropyranosyluronate (GG) and; 4-deoxy-␤-d-hex-4-enopyranosylglucuronate-(1 → 4)-O-␤-d-glucuropyranosyluronate-(1 → 4)-O-␤-d-glucuropyranosyluronate (GGG), respectively (Fig. 2). For GG and GGG, others molecular ions as [M − 2H + Na]− ,

Fig. 2. ESI-Q/TOF-mass spectra of an oligosaccharides mixture (96 h of incubation) containing: unsaturated monomer (G), dimer (GG) and trimer (GGG).

454


[M − 2H + K]− , [M − 3H + Na + K]− and [M − 2H + Na]− were distinguished with molecular weights of 373.22, 389.18, 586.88 and 548.99, respectively. 3.3. Large scale-production of unsaturated oligoglucuronans Thanks to our previous results, the production and the purification of unsaturated glucuronic acid (G), unsaturated dimer (GG) and unsaturated trimer (GGG) was investigated. For that purpose, 12 g of deacetylated glucuronan (real mass = 7.2 g depending on purity%) were incubated during 96 h with 200 U of crude enzyme fraction. After 24 h, 50% of the sample was loaded on a DEAE-sepharose column. Fractions (5 ml) were collected and uronic acids were assayed (Fig. 3). Three families of oligoglucuronans were recognized and analysed as G, GG and GGG with mass spectrometry (data not shown) and 1 H NMR spectroscopy (Fig. 4). The differences between protons of unsaturated monomer and oligomers were the lack of H-1 (5.64 ppm), H-1␤ (5.16 ppm), H-1␣ (5.74 ppm) and H-4/H-5 (4.32 ppm) signals for unsaturated glucuronic acid. Additionally, new signals appeared for the anomeric protons (H-1␣ at 5.68 ppm and H-1␤ at 5.20 ppm) of this compound. All the average dp

Fig. 4. 1 H NMR (80 ◦ C) spectra of (a) G; (b) GG; (c) GGG (20 g/l in D2 O).

were then estimated by comparaison between H-1 signal integration of ␤--(4,5)-glucuronic acid (G) and all H-1 signals integrations (H-1, H-1␤, H-1␣ and H-1). The results corroborated the mass spectrometry analysis. During the steps of anionic exchange chromatography, the unsaturated glucuronic acid (G) appeared unstable when it was quantify by A235 nm measurement or analysed by 1 H NMR or mass spectrometry. We noted that this instability increased notably after several elutions on anions exchange chromatography. Table 4 shows the different mass of each oligosaccharide purified on DEAE-sepharose. At 24 h of incuTable 4 Mass (g) and percentage of polyglucuronate recoveries for each oligoglucuronate purified by anion exchange chromatography on DEAE-sepharose after T = 24 and 96 h of deacetylated glucuronan depolymerization Mass (g)

Fig. 3. DEAE-Sepharose 6B-CL anion-exchange liquid chromatography of an oligomers mixture from a 24 h depolymerization of deacetylated glucuronan. Each fraction (5 ml) was collected and 40 ␮l was analysed by uronic acids assay. G, GG and GGG refer, respectively, to the unsaturated monomer, the unsaturated dimer and the unsaturated trimer.

Ga GGb GGGc a b c

Monomer. Dimer. Trimer.

T (24 h)

T (96 h)

0.18 (5.6%) 0.30 (9.4%) 1.06 (33.0%)

0.27 (7.5%) 0.92 (25.5%) 0.31 (9.7%)


Fig. 5. TLC analysis of oligoglucuronans and GGG degradation by Trichoderma sp. GL2. TLC analysis of oligoglucuronans. Lane 1, ␣-keto-acid (DKI) from G (monomer) conversion; lane 2, GG (dimer); lane 3, GGG (trimer); lane 4, GGG (1 mg) incubated with 0.5 U of glucuronan lyase.

bation, the unsaturated trimer (1.06 g) was the most representative oligosaccharide of the mixture compared with dimer (0.30 g) and monomer (0.18 g). On the other hand, after 96 h of incubation, the unsaturated dimer was predominantly present (0.92 g) by comparison with the trimer (0.31 g) and monomer (0.27 g). This result suggested that the unsaturated trimer (GGG) was cleaved into GG and G seeing that di- and monomer increased whereas trimer decreased. This was confirmed by TLC analysis of GL products (G and GG) using GGG (1 mg) as substrate (Fig. 5). The lyase activity on unsaturated trisaccharide oligoglucuronan (GGG) was observed by formation of GG and G after 30 min of incubation with enzymatic preparation (0.5 U). The conversion of unsaturated monosaccharide (G) established above was confirmed in regard to the differences between migration of G in GGG degradation products mix and purified G.

4. Discussion In this paper, we report the isolation of a novel glucuronan lyase (GL) activity from Trichoderma sp. GL2 and its partial characterization. Contrary to our previous studies about GlyA, a bacterial glucuronan lyase (Da costa et al., 2003), we noted that this fungal polysaccharide lyase activity was able to cleave highly acetylated glucuronans and had higher activities for the others (deacetylated and standard). It was clearly established that the GL affinity toward glucuronan increased

455

when acetyl esterification and more particularly 2,3di-O-acetyl esterification decreased. So, the specific activity against standard and highly acetylated glucuronan was reduced to 2.5 and 4.0, respectively, compared to deacetylated one. This inhibitor effect of O-acetylation on the Trichoderma sp. GL2 glucuronan lyase activity was correlated to literature, where it was described that polysaccharide lyase activities were inhibited by the presence of O-acetyl substituents on alginate (Skjak braek et al., 1986; Kennedy et al., 1992), gellan (Giavasis et al., 2000), pectin (Pagel and Heitifuss, 1990; Oosterveld et al., 2000) or xanthan (Shatwell and Sutherland, 1991). Studies about deacetylated glucuronan degradations by this new glucuronan lyase showed that GG and GGG were the most abundant products. After 96 h of deacetylated glucuronan degradation, the proportion of GG increased compared with the GGG one. We observed that the -(4,5) unsaturated glucuronic acid was unstable notably during purification steps by anion exchange chromatography (DEAE-Sepharose). Indeed, after several minutes, the detection of 4,5-unsaturated monomer (A235 nm ) decreased. The non-enzymatic transformation of this unsaturated monomer in α-keto-glucuronic acid has been considered in agree with literature data (Preiss and Ashwell, 1962; Shevchik et al., 1999; Hashimoto et al., 2000). In the last, an -(4,5) unsaturated pyranose monomer may spontaneously transform into 4,5-enolic chain which was isomerized into the most stable 5-keto structure: the 4-deoxy-l-threo-5-hexosulose uronic acid (DKI). This conversion of G was confirmed by TLC analysis. Thus, this glucuronan lyase activity from Trichoderma sp. GL2 allowed to completely depolymerize glucuronan up to obtention of ultimate products which were unsaturated monomer (G) and unsaturated dimer (GG). This conducted us to suggest that unsaturated trimer (GGG) was the limit substrate for the GL activity. Standard and highly acetylated glucuronans were depolymerized by the enzymatic preparation with the aim to produce variably acetylated oligosaccharides. With the mainly 3-O-acetylated glucuronan degradation we obtained unsaturated acetylated oligoglucuronans with various dp (2 to >8) whereas 2,3di-O-acetylated glucuronan cleavage led us to envisage the recovery of highly acetylated LMW polysaccha-

456


rides. Contrary to our previous works about glucuronan lyase from S. meliloti M5N1CS (Da costa et al., 2003), the Trichoderma sp. GL2 glucuronan lyase authorized to produce low dp with O-acetylation degrees similar to the native polymer (standard glucuronan). In regards to the implication of O-acetyl distribution on polysaccharide and oligosaccharide biological activities (Komalavilas and Mort, 1989; Dinand et al., 1997), these studies opened the way to a large-scale production of variably O-acetylated oligoglucuronans. Then we envisage new applications as it is actually investigated with other bacterial (Potin et al., 1999; Iwasaki and Matsubara, 2000) and vegetal (Spiro et al., 1998; Lee et al., 1999) oligosaccharides. Actually, the works for characterize and identify proteins present in the enzymatic preparation are in progress. The large production of unsaturated oligoglucuronates directly in bioreactor will be also envisaged by adding Trichoderma sp. GL2 glucuronan lyase after or during the glucuronan biosynthesis.

Acknowledgements We thank Serge Pilard for mass spectrometry analysis. This work was supported by the European social found and by Conseil Régional de Picardie (France).

References Aldington, S., Fry, S., 1993. Oligosaccharins. Adv. Bot. Res. 19, 1–101. Ariyo, B., Tamerler, C., Bucke, C., Keshavarz, T., 1998. Enhanced penicillin production by oligosaccharides from batch cultures of Penicillium chrysogenum in stirred-tank reactors. FEMS Microbiol. Lett. 166, 165–170. Bolwell, P.G., 1999. Role of active oxygen species and NO in plant defence responses. Plant Biol. 2, 287–294. Bradford, H.M., 1976. A rapid and sensitive method for the quantification of microgramm quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72, 248–254. Côté, F., Hahn, M.G., 1994. Oligosaccharins: structure and signal transduction. Plant Mol. Biol. 12, 1379–1411. Courtois, B., Hornez, J.P., Courtois, J., Derieux, J.C., 1983. Mise en e´ vidence d’une propriété métabolique de Rhizobium meliloti utilisable pour sa classification. Anal. Microbiol. 182, 141–147. Courtois, J., Séguin, J.P., Declomesnil, S., Heyraud, A., Colin-Morel, P., Dantas, L., Barbotin, J.N., Courtois, B., 1993. A (1 → 4)-␤d-glucuronan excreted by a mutant of Rhizobium meliloti M5N1 strain. J. Carbohydr. Chem. 12, 441–448.

Creelman, R.A., Mullet, J.E., 1997. Oligosaccharins, brassinolides and jasmonates: non traditional regulators of plant growth, development and gene expression. Plant Cell. 9, 1211–1223. Da costa, A., Michaud, P., Petit, E., Heyraud, A., Colin-Morel, P., Courtois, B., Courtois, J., 2001. Purification and properties of a glucuronan lyase from Sinorhizobium meliloti M5N1CS (NCIMB 40472). Appl. Environ. Microbiol. 67, 5197–5203. Da costa, A., Michaud, P., Petit, E., Heyraud, A., Colin-Morel, P., Courtois, B., Courtois, J., 2003. Acetyl substitution of glucuronan influences glucuronan cleavage by GlyA from Sinorhizobium meliloti M5N1CS (NCIMB 40472). Carbohydr. Polym. 51, 223–228. Dantas, L., Courtois, J., Courtois, B., Séguin, J.P., Gey, C., Heyraud, A., 1994. NMR spectroscopic investigation of oligoglucuronates prepared by enzymatic hydrolysis of a (1 → 4)-␤-d-glucuronan. Carbohydr. Res. 265, 303–310. Darvill, A., Augur, C., Bergmann, C., Carlson, R.W., Cheong, J.J., Eberhard, S., Hahn, M.G., Lo, V.M., Marfa, V., Meyer, B., Mohnen, D., O’Neill, M.A., Spiro, M.D., Van Halbeek, H., York, W.S., Albersheim, P., 1992. Oligosaccharins—oligosaccharides that regulate growth, development and defence responses in plants. Glycobiology 2, 181–198. Dinand, E., Excoffier, G., Lienart, Y., Vignon, M.R., 1997. Two rhamnogalacturonide tetrasaccharides isolated from semi-retted flax fibers are signaling molecules in Rubus fructicosus L. cell. Plant Physiol. 115, 793–801. Ebel, J., Corsio, E., 1994. Elicitors of plant defense responses. Int. Rev. Cytol. 148, 1–36. Gacesa, P., 1988. Alginate. Carbohydr. Polym. 8, 161–182. Giavasis, I., Harvey, L.M., McNeil, B., 2000. Gellan gum. Crit. Rev. Biotech. 20, 177–211. Gonzales, M.L., Courtois, J., Heyraud, A., Colin-Morel, P., Séguin, J.P., Barbotin, J.N., Courtois, B., 1996. Selection of a succinoglycan deficient Rhizobium meliloti mutant producing (1 → 4)-␤-dglucuronan partially acetylated: symbiotic properties analysis. Ann. N. Y. Acad. Sci. 782, 53–60. Hashimoto, W., Miyake, O., Momma, K., Kawai, S., Murata, K., 2000. Molecular identification of oligoalginate lyase of Sphingomonas sp. strain A1 as one of the enzymes required for complete depolymerization of alginate. J. Bacteriol. 182, 4572– 4577. Hashimoto, W., Nankai, H., Mikami, B., Murata, K., 2003. Crystal structure of Bacillus sp. GL1 xanthan lyase, which acts on the side chains of xanthan. J. Biol. Chem. 278 (9), 7663–7673. Iwasaki, K., Matsubara, Y., 2000. Purification of alginate oligosaccharides with root growth-promoting activity toward lettuce. Biosci. Biotechnol. Biochem. 64, 1067–1070. Kawada, A., Hiura, N., Shiraiwa, M., Tajima, S., Hiruma, M., Hara, K., Ishibashi, A., Takahara, H., 1997. Stimulation of human keratinocyte growth by alginate oligosaccharides, a possible cofactor for epidermal growth factor in cell culture. FEBS Lett. 408, 43–46. Kennedy, L., McDowell, K., Sutherland, I.W., 1992. Alginases from Azotobacter species. J. Gen. Microbiol. 138, 3007–3013. Komalavilas, P., Mort, A.J., 1989. The acetylation at O-3 of galacturonic acid in the rhamnose-rich region of pectins. Carbohydr. Res. 189, 261–272.

C. Delattre et al. / Journal of Biotechnology 118 (2005) 448–457 Lee, S., Choi, H., Suh, S.S., Doo, I.S., Oh, K.Y., Choi, E., SchroederTaylor, A.T., Low, P.S., Lee, Y., 1999. Oligogalacturonic acid and chitosan reduce stomatal aperture by inducing the evolution of reactive oxygen species from guard cells of tomato and Commolina communis. Plant Physiol. 121, 147–152. Michaud, P., Roblot, C., Courtois, J., Séguin, J.P., Heyraud, A., Barbotin, J.N., Courtois, B., 1995. Effect of Mg2+ on production and O-acetylation of glucuronan excreted by the Rhizobium meliloti M5N1CS strain during fermentation. Lett. Appl. Microbiol. 20, 110–112. Michaud, P., Da Costa, A., Courtois, B., Courtois, J., 2003. Polysaccharides lyases—recent developments as biotechnological tools. Crit. Rev. Biotechnol. 23 (4), 233–266. Nardella, A., Chaubet, F., Boisson-Vidal, C., Blondin, C., Durand, P., Jozefonvicz, 1996. Anticoagulant low molecular weight fucans produced by radical process and ion exchange chromatography of high molecular weight fucans extracted from the brown seaweed Ascophyllum nodosum. Carbohydr. Res. 289, 201–208. Oosterveld, A., Beldman, G., Searle-van Leeuwen, M.J.F., Voragen, A.G.J., 2000. Effect of enzymatic deacetylation on gelation of sugar beet pectin in the presence of calcium. Carbohydr. Polym. 43, 249–256. Ozeretskovskaya, O.L., Romenskaya, I.G., 1996. Oligosaccharins as regulatory molecules of plants. J. Plant Physiol. 43, 648–655. Pagel, W., Heitifuss, R., 1990. Enzymatic activities in soft rot pathogenesis of potato tubers: effect of calcium, pH, and degree of pectin esterification on the activities of polygalacturonase and pectate lyase. Physiol. Mol. Plant Pathol. 37, 9–25. Painter, T.J., 1993. Carbohydrate polymers in desert reclamation: the potential of microalgal biofertilizers. Carbohydr. Res. 20, 77–86. Patier, P., Yvin, J.C., Kloareg, B., Liénart, Y., Rochas, C., 1993. Seaweed liquid fertilizer from Ascophyllum nodosum contains elicitors of plant d-glycanase. J. Appl. Phycol. 5, 343–349. Pentillä, M., Nevalainen, H., Rätto, M., Salminen, E., Knowles, J., 1987. A versatile transformation system for the cellulolytic filamentous fungus Trichoderma reesei. Gene 61, 155–164. Pineo, G.F., Hull, R.D., 1997. Low-molecular-weight heparin: prophylaxis and treatment of venous thromboembolism. Annu. Rev. Med. 48, 79–91.

457

Potin, P., Bouarad, K., Küpper, F., Kloareg, B., 1999. Oligosaccharide recognition signals and defence reactions in marine plant–microbe interactions. Curr. Opin. Microbiol. 2, 276– 283. Preiss, J., Ashwell, G., 1962. Alginic acid metabolism in bacteria. J. Biol. Chem. 237, 309–316. Seguin, J.P., Senechal, L., Roblot, C., Pheulpin, P., Courtois, J., Courtois, B., Barbotin, J.N., 1996. Structural studies of a new extracellular polysaccharide excreted by Rhizobium meliloti M5N1 mutant strain. Carbohydr. Lett. 2, 83–90. Shatwell, K.P., Sutherland, I.W., 1991. Influence of the acetyl substituent on the interaction of xanthan with plant polysaccharidesIII. Xanthan–Konjac Mannan Systems. Carbohydr. Polym. 14, 131–147. Shevchik, V.E., Condemine, G., Robert-Baudouy, J., HugouvieuxCotte-Pattat, N., 1999. The exopolygalacturonate lyase PelW and the oligogalacturonate lyase OgL, two cytoplasmic enzymes of pectin catabolism in Erwinia chrysanthemi 3937. J. Bacteriol. 181, 3912–3919. Skjak braek, G., Grasdalen, H., Larsen, B., 1986. Monomer sequence and acetylation pattern in some bacterial alginates. Carbohydr. Res. 154, 239–250. Spiro, M.D., Ridley, B.L., Eberhard, S., Kates, K.A., Mathieu, Y., O’Neill, M.A., Mohnen, D., Guern, J., Darvill, A., Albersheim, P., 1998. Biological activity of reducing-end-derivatized oligogalacturonides in tobacco tissue cultures. Plant Physiol. 116, 1289–1298. Sutherland, I.W., 1995. Polysaccharide lyases. FEMS Microbiol. Rev. 16, 323–347. Suzuki, K., Mikami, T., Okaya, Y., Tokoro, A., Suzuki, S., Suzuki, M., 1986. Antitumour effect of hexa-N-acetychitohexose and chitohexaose. Carbohydr. Res. 151, 403. Van Den Hoogen, B.M., Van Weeren, P.R., Lopes-Cardozo, M., Van Golde, L.M.G., Barnveld, A., Van de Lest, C.H.A., 1998. A microtiter assay for the determination of uronic acids. Anal. Biochem. 257, 107–111. Willats, W.G.T., Steele-King, C.G., Marcus, S.E., Knox, J.P., 1999. Side chains of pectic polysaccharides are regulated in relation to cell proliferation and cell differenciation. Plant J. 20, 619–628.


Author Index Adam, A.C., see Latorre-García, L. (118) 167 Almeida-Vara, E., see Moreira, P.R. (118) 339 Álvaro, G., see Vidal, L. (118) 75 Alves, P.M., see Teixeira, A. (118) 290 Amadori, D., see Soldateschi, D. (118) 370 Antunes, A., see Moreira, P.R. (118) 339 Arís, A., see González-Montalbán, N. (118) 406 Bauer, R., Volschenk, H., Dicks, L.M.T., Cloning and expression of the malolactic gene of Pediococcus damnosus NCFB1832 in Saccharomyces cerevisiae (118) 353 Benaiges, M.D., see Vidal, L. (118) 75 Benicchi, T., see Soldateschi, D. (118) 370 Berti, B., see Soldateschi, D. (118) 370 Bi, J., see Yun, Q. (118) 67 Bi, J., see Zhi, W. (118) 157 Boonstra, J., see Thomassen, Y.E. (118) 270 Bragós, R., see Soley, A. (118) 398 Bravaccini, S., see Soldateschi, D. (118) 370 Brogi, A., see Soldateschi, D. (118) 370 Cairó, J.J., see Soley, A. (118) 398 Calistri, D., see Soldateschi, D. (118) 370 Caminal, G., see Vidal, L. (118) 75 Cardoso, A.G., see Tsao, Y.-S. (118) 316 Carrió, M.M., see González-Montalbán, N. (118) 406 Carrondo, M.J.T., see Teixeira, A. (118) 290 Cascone, O., see Levin, G. (118) 363 Cha, H.J., see Kang, D.G. (118) 379 Chen, Q.-X., see Zhou, J.-Y. (118) 201 Chen, T., see Yun, Q. (118) 67 Chen, T.-F., see Zhou, J.-Y. (118) 201 Chen, X.-L., see Zheng, Y.-G. (118) 413 Chen, Y.-H., Wu, J.-C., Wang, K.-C., Chiang, Y.-W., Lai, C.-W., Chung, Y.-C., Hu, Y.-C., Baculovirus-mediated production of HDV-like particles in BHK cells using a novel oscillating bioreactor (118) 135 Chesnut, J.D., see Yahata, K. (118) 123 Chia, S.-M., Lin, P.-C., Quek, C.-H., Yin, C., Mao, H.-Q., Leong, K.W., Xu, X., Goh, C.-H., Ng, M.-L., Yu, H., Engineering microenvironment for expansion of sensitive anchorage-dependent mammalian cells (118) 434 Chiang, Y.-W., see Chen, Y.-H. (118) 135

doi:10.1016/S0168-1656(05)00417-7

Christakopoulos, P., see Panagiotou, G. (118) 304 Chung, Y.-C., see Chen, Y.-H. (118) 135 Clemente, J.J., see Teixeira, A. (118) 290 Collet, C., Gaudard, O., Péringer, P., Schwitzguébel, J.-P., Acetate production from lactose by Clostridium thermolacticum and hydrogen-scavenging microorganisms in continuous culture— Effect of hydrogen partial pressure (118) 328 Condon, R.G.G., see Tsao, Y.-S. (118) 316 Courtois, B., see Delattre, C. (118) 448 Courtois, J., see Delattre, C. (118) 448 Cruz, H.J., see Teixeira, A. (118) 290 Cuatrecasas, S., see González-Montalbán, N. (118) 406 Cullen, D., see Wymelenberg, A.V. (118) 17 Cunha, A.E., see Teixeira, A. (118) 290 De Paola, F., see Soldateschi, D. (118) 370 Dehareng, D., see Moreira, P.R. (118) 339 Delattre, C., Michaud, P., Lion, J.M., Courtois, B., Courtois, J., Production of glucuronan oligosaccharides using a new glucuronan lyase activity from a Trichoderma sp. strain (118) 448 Diaz Sanchez-Bustamante, C., see Kelm, J.M. (118) 213 Dicks, L.M.T., see Bauer, R. (118) 353 Djonov, V., see Kelm, J.M. (118) 213 Duarte, J.C., see Moreira, P.R. (118) 339 Duez, C., see Moreira, P.R. (118) 339 Ehler, E., see Kelm, J.M. (118) 213 Fabbri, F., see Soldateschi, D. (118) 370 Ferrer, P., see Vidal, L. (118) 75 Fournier, S.M., see Taylor, R.H. (118) 265 Freire-Picos, M.A., see Seoane, S. (118) 149 Frère, J.M., see Moreira, P.R. (118) 339 Fussenegger, M., see Kelm, J.M. (118) 213 Galaev, I.Y., see Hanora, A. (118) 421 Gámez, X., see Soley, A. (118) 398 Gaskell, J., see Wymelenberg, A.V. (118) 17 Gaudard, O., see Collet, C. (118) 328 Georgiev, M., see Pavlov, A. (118) 89 Gòdia, F., see Soley, A. (118) 398 Goh, C.-H., see Chia, S.-M. (118) 434 Gong, H., see Zhou, J.-Y. (118) 201

Author Index González-Montalbán, N., Carrió, M.M., Cuatrecasas, S., Arís, A., Villaverde, A., Bacterial inclusion bodies are cytotoxic in vivo in absence of functional chaperones DnaK or GroEL (118) 406 Guadalupe, A.R., see Muñoz-Serrano, L. (118) 233 Guiard, B., see Seoane, S. (118) 149 Guo, J.-Q., see Zhou, J.-Y. (118) 201 Hall, L.D., see Macaskie, L.E. (118) 187 Hanora, A., Plieva, F.M., Hedström, M., Galaev, I.Y., Mattiasson, B., Capture of bacterial endotoxins using a supermacroporous monolithic matrix with immobilized polyethyleneimine, lysozyme or polymyxin B (118) 421 He, Z., see Zhu, K. (118) 257 Hedström, M., see Hanora, A. (118) 421 Hefford, M.A., see Taylor, R.H. (118) 265 Hildén, L., Väljamäe, P., Johansson, G., Surface character of pulp fibres studied using endoglucanases (118) 386 Hoerstrup, S.P., see Kelm, J.M. (118) 213 Honda, K., see Sakamoto, K. (118) 99 Horgan, G.W., Optimising two-dye microarray designs for estimating associations with a quantitative trait (118) 1 Hotta, J., see Yahata, K. (118) 123 Hu, Y.-C., see Chen, Y.-H. (118) 135 Huhtala, A., Linko, P., Mutharasan, R., Protein response of insect cells to bioreactor environmental stresses (118) 278 Ilieva, M., see Pavlov, A. (118) 89 Imamoto, F., see Yahata, K. (118) 123 Ittner, L., see Kelm, J.M. (118) 213 Jin, H., see Zhu, K. (118) 257 Johansson, G., see Hildén, L. (118) 386 Kang, D.G., Lim, G.-B., Cha, H.J., Functional periplasmic secretion of organophosphorous hydrolase using the twin-arginine translocation pathway in Escherichia coli (118) 379 Kaplan, H., see Taylor, R.H. (118) 265 Kataoka, M., see Sakamoto, K. (118) 99 Kearns, B.G., see Tsao, Y.-S. (118) 316 Kelm, J.M., Diaz Sanchez-Bustamante, C., Ehler, E., Hoerstrup, S.P., Djonov, V., Ittner, L., Fussenegger, M., VEGF profiling and angiogenesis in human microtissues (118) 213 Kersten, P.J., see Wymelenberg, A.V. (118) 17 Kishine, H., see Yahata, K. (118) 123 Kita, S., see Sakamoto, K. (118) 99 Kleerebezem, R., see Paulo, P.L. (118) 107 Kovacheva, E., see Pavlov, A. (118) 89 Lagos, J.C., see Tsao, Y.-S. (118) 316 Lai, C.-W., see Chen, Y.-H. (118) 135 Latorre-García, L., Adam, A.C., Manzanares, P., Polaina, J., Improving the amylolytic activity of Saccharomyces cerevisiae glucoamylase by the addition of a starch binding domain (118) 167 Lecina, M., see Soley, A. (118) 398 Leisola, M., see Nyyssölä, A. (118) 55 , Guide for Authors (118) 117

459

Leong, K.W., see Chia, S.-M. (118) 434 Lettinga, G., see Paulo, P.L. (118) 107 Levin, G., Mendive, F., Targovnik, H.M., Cascone, O., Miranda, M.V., Genetically engineered horseradish peroxidase for facilitated purification from baculovirus cultures by cation-exchange chromatography (118) 363 Lim, G.-B., see Kang, D.G. (118) 379 Lin, J.-M., see Zhao, L. (118) 177 Lin, P.-C., see Chia, S.-M. (118) 434 Linko, P., see Huhtala, A. (118) 278 Lio, P., see Tsao, Y.-S. (118) 316 Lion, J.M., see Delattre, C. (118) 448 Liu, T., Zhang, Y.-Z., Wu, X.-F., High level expression of functionally active human lactoferrin in silkworm larvae (118) 246 Liu, Z., see Tsao, Y.-S. (118) 316 Ma, G., see Yun, Q. (118) 67 Ma, Y., see Zhu, K. (118) 257 Macaskie, L.E., Yong, P., Paterson-Beedle, M., Thackray, A.C., Marquis, P.M., Sammons, R.L., Nott, K.P., Hall, L.D., A novel non line-of-sight method for coating hydroxyapatite onto the surfaces of support materials by biomineralization (118) 187 Malcata, F.X., see Moreira, P.R. (118) 339 Manzanares, P., see Latorre-García, L. (118) 167 Mao, H.-Q., see Chia, S.-M. (118) 434 Marquis, P.M., see Macaskie, L.E. (118) 187 Martinez, D., see Wymelenberg, A.V. (118) 17 Mattiasson, B., see Hanora, A. (118) 421 Medri, L., see Soldateschi, D. (118) 370 Mendive, F., see Levin, G. (118) 363 Michaud, P., see Delattre, C. (118) 448 Miranda, M.V., see Levin, G. (118) 363 Moreira, J.L., see Teixeira, A. (118) 290 Moreira, P.R., Duez, C., Dehareng, D., Antunes, A., Almeida-Vara, E., Frère, J.M., Malcata, F.X., Duarte, J.C., Molecular characterisation of a versatile peroxidase from a Bjerkandera strain (118) 339 Muñoz-Serrano, L., Guadalupe, A.R., Vega-Bermudez, E., Morphological studies of oligodeoxyribonucleotides probes covalently immobilized at polystyrene modified surfaces (118) 233 Mutharasan, R., see Huhtala, A. (118) 278 Ng, M.-L., see Chia, S.-M. (118) 434 Nose, H., see Sakamoto, K. (118) 99 Nott, K.P., see Macaskie, L.E. (118) 187 Nyyssölä, A., Pihlajaniemi, A., Palva, A., von Weymarn, N., Leisola, M., Production of xylitol from D-xylose by recombinant Lactococcus lactis (118) 55 Okabe, M., see Yahata, K. (118) 123 Okamoto, K., see Yanase, H. (118) 35 Oliveira, R., see Teixeira, A. (118) 290 Olsson, L., see Panagiotou, G. (118) 304 Ouyang, F., see Zhi, W. (118) 157 Palva, A., see Nyyssölä, A. (118) 55

460

Author Index

Panagiotou, G., Christakopoulos, P., Olsson, L., The influence of different cultivation conditions on the metabolome of Fusarium oxysporum (118) 304 Paterson-Beedle, M., see Macaskie, L.E. (118) 187 Paulo, P.L., Kleerebezem, R., Lettinga, G., Lens, P.N.L., Cultivation of high-rate sulfate reducing sludge by pH-based electron donor dosage (118) 107 Pavlov, A., Popov, S., Kovacheva, E., Georgiev, M., Ilieva, M., Volatile and polar compounds in Rosa damascena Mill 1803 cell suspension (118) 89 Pedrosa, F.O., see Ramos, H.J.O. (118) 9 Péringer, P., see Collet, C. (118) 328 Pihlajaniemi, A., see Nyyssölä, A. (118) 55 Plieva, F.M., see Hanora, A. (118) 421 Polaina, J., see Latorre-García, L. (118) 167 Popov, S., see Pavlov, A. (118) 89 Quek, C.-H., see Chia, S.-M. (118) 434 Rajangam, A.S., see Wymelenberg, A.V. (118) 17 Ramos, H.J.O., Souza, E.M., Soares-Ramos, J.R.L., Pedrosa, F.O., A new system to control the barnase expression by a NifAdependent promoter (118) 9 Ren, Z., see Zhu, K. (118) 257 Riu, P., see Soley, A. (118) 398 Rodríguez-Torres, A.-M., see Seoane, S. (118) 149 Rosell, X., see Soley, A. (118) 398 Sabat, G., see Wymelenberg, A.V. (118) 17 Sakamoto, K., Honda, K., Wada, K., Kita, S., Tsuzaki, K., Nose, H., Kataoka, M., Shimizu, S., Practical resolution system for DLpantoyl lactone using the lactonase from Fusarium oxysporum (118) 99 Sammons, R.L., see Macaskie, L.E. (118) 187 Sasaki, Y., see Yahata, K. (118) 123 Sato, D., see Yanase, H. (118) 35 Schwitzguébel, J.-P., see Collet, C. (118) 328 Seoane, S., Guiard, B., Rodríguez-Torres, A.-M., Freire-Picos, M.A., Effects of splitting alternative KlCYC1 3-UTR regions on processing: Metabolic consequences and biotechnological applications (118) 149 Shang, S.-B., see Zhou, J.-Y. (118) 201 Shen, H.-G., see Zhou, J.-Y. (118) 201 Shimizu, S., see Sakamoto, K. (118) 99 Simons, B.L., see Taylor, R.H. (118) 265 Soares-Ramos, J.R.L., see Ramos, H.J.O. (118) 9 Soldateschi, D., Bravaccini, S., Berti, B., Brogi, A., Benicchi, T., Soldatini, C., Medri, L., Fabbri, F., De Paola, F., Amadori, D., Calistri, D., Development and characterization of a monoclonal antibody directed against human telomerase reverse transcriptase (hTERT) (118) 370 Soldatini, C., see Soldateschi, D. (118) 370 Soley, A., Lecina, M., Gámez, X., Cairó, J.J., Riu, P., Rosell, X., Bragós, R., Gòdia, F., On-line monitoring of yeast cell growth by impedance spectroscopy (118) 398 Sone, T., see Yahata, K. (118) 123 Song, J., see Zhi, W. (118) 157

Souza, E.M., see Ramos, H.J.O. (118) 9 Su, Z., see Yun, Q. (118) 67 Targovnik, H.M., see Levin, G. (118) 363 Tay, J.-H., see Zhuang, W.-Q. (118) 45 Tay, S.T.-L., see Zhuang, W.-Q. (118) 45 Taylor, R.H., Fournier, S.M., Simons, B.L., Kaplan, H., Hefford, M.A., Covalent protein immobilization on glass surfaces: Application to alkaline phosphatase (118) 265 Teeri, T.T., see Wymelenberg, A.V. (118) 17 Teixeira, A., Cunha, A.E., Clemente, J.J., Moreira, J.L., Cruz, H.J., Alves, P.M., Carrondo, M.J.T., Oliveira, R., Modelling and optimization of a recombinant BHK-21 cultivation process using hybrid grey-box systems (118) 290 Thackray, A.C., see Macaskie, L.E. (118) 187 Thomassen, Y.E., Verkleij, A.J., Boonstra, J., Verrips, C.T., Specific production rate of VHH antibody fragments by Saccharomyces cerevisiae is correlated with growth rate, independent of nutrient limitation (118) 270 Tsao, Y.-S., Cardoso, A.G., Condon, R.G.G., Voloch, M., Lio, P., Lagos, J.C., Kearns, B.G., Liu, Z., Monitoring Chinese hamster ovary cell culture by the analysis of glucose and lactate metabolism (118) 316 Tsuzaki, K., see Sakamoto, K. (118) 99 Väljamäe, P., see Hildén, L. (118) 386 Vega-Bermudez, E., see Muñoz-Serrano, L. (118) 233 Verkleij, A.J., see Thomassen, Y.E. (118) 270 Verrips, C.T., see Thomassen, Y.E. (118) 270 Vidal, L., Ferrer, P., Álvaro, G., Benaiges, M.D., Caminal, G., Influence of induction and operation mode on recombinant rhamnulose 1-phosphate aldolase production by Escherichia coli using the T5 promoter (118) 75 Villaverde, A., see González-Montalbán, N. (118) 406 Voloch, M., see Tsao, Y.-S. (118) 316 Volschenk, H., see Bauer, R. (118) 353 von Weymarn, N., see Nyyssölä, A. (118) 55 Wada, K., see Sakamoto, K. (118) 99 Wang, B., see Zhu, K. (118) 257 Wang, K.-C., see Chen, Y.-H. (118) 135 Wang, Z., see Zheng, Y.-G. (118) 413 Wu, J.-C., see Chen, Y.-H. (118) 135 Wu, J.-X., see Zhou, J.-Y. (118) 201 Wu, X.-F., see Liu, T. (118) 246 Wymelenberg, A.V., Sabat, G., Martinez, D., Rajangam, A.S., Teeri, T.T., Gaskell, J., Kersten, P.J., Cullen, D., The Phanerochaete chrysosporium secretome: Database predictions and initial mass spectrometry peptide identifications in cellulosegrown medium (118) 17 Xiao, C., see Zhu, K. (118) 257 Xu, X., see Chia, S.-M. (118) 434 Yahata, K., Kishine, H., Sone, T., Sasaki, Y., Hotta, J., Chesnut, J.D., Okabe, M., Imamoto, F., Multi-gene Gateway clone design for expression of multiple heterologous genes in living cells:

Author Index Conditional gene expression at near physiological levels (118) 123 Yamamoto, K., see Yanase, H. (118) 35 Yanase, H., Yamamoto, K., Sato, D., Okamoto, K., Ethanol production from cellobiose by Zymobacter palmae carrying the Ruminocuccus albus -glucosidase gene (118) 35 Yang, R.E., see Yun, Q. (118) 67 Yi, S., see Zhuang, W.-Q. (118) 45 Yin, C., see Chia, S.-M. (118) 434 Yong, P., see Macaskie, L.E. (118) 187 Yu, H., see Chia, S.-M. (118) 434 Yun, Q., Yang, R.E., Chen, T., Bi, J., Ma, G., Su, Z., Reproducible preparation and effective separation of PEGylated recombinant human granulocyte colony-stimulating factor with novel “PEG-pellet” PEGylation mode and ion-exchange chromatography (118) 67 Zhang, F., see Zhu, K. (118) 257 Zhang, Y.-Z., see Liu, T. (118) 246 Zhao, L., Lin, J.-M., Development of a micro-plate magnetic chemiluminescence enzyme immunoassay (MMCLEIA) for

461

rapid- and high-throughput analysis of 17-estradiol in water samples (118) 177 Zheng, Y.-G., Chen, X.-L., Wang, Z., Microbial biomass production from rice straw hydrolysate in airlift bioreactors (118) 413 Zhi, W., Song, J., Ouyang, F., Bi, J., Application of response surface methodology to the modeling of -amylase purification by aqueous two-phase systems (118) 157 Zhou, J.-Y., Shang, S.-B., Gong, H., Chen, Q.-X., Wu, J.-X., Shen, H.-G., Chen, T.-F., Guo, J.-Q., In vitro expression, monoclonal antibody and bioactivity for capsid protein of porcine circovirus type II without nuclear localization signal (118) 201 Zhu, K., Jin, H., Ma, Y., Ren, Z., Xiao, C., He, Z., Zhang, F., Zhu, Q., Wang, B., A continuous thermal lysis procedure for the large-scale preparation of plasmid DNA (118) 257 Zhu, Q., see Zhu, K. (118) 257 Zhuang, W.-Q., Tay, J.-H., Yi, S., Tay, S.T.-L., Microbial adaptation to biodegradation of tert-butyl alcohol in a sequencing batch reactor (118) 45


Subject Index -Amylase, (118) 157 Acetate, (118) 328 Aerobic granules, (118) 45 Affinity chromatography, (118) 421 AFM, (118) 233 Aggregation, (118) 406 Agricultural residues, (118) 413 Alkaline phosphatase, (118) 265 Amide bond, (118) 265 Anaerobic, (118) 107 Anchorage-dependent mammalian cells, (118) 434 Anoxia, (118) 278 Aqueous two-phase systems, (118) 157 Baculovirus, (118) 135 Barnase, (118) 9 Barstar, (118) 9 BelloCell, (118) 135 BHK-21 cell lines, (118) 290 Biodegradation, (118) 45 Biofilm, (118) 187 Biointerfaces, (118) 233 Biological containment system, (118) 9 Biomass estimator, (118) 398 Biomass, (118) 149, 35 Bioprocess design, (118) 75 Bioreactor cultivation, (118) 89 Bioreactor, (118) 278, 413 Bioresource, (118) 413 Calcium phosphate, (118) 187 Carbon limited, (118) 270 Cell density, (118) 316 Cell expansion, (118) 434 Cell suspension, (118) 107 Cell-based therapies, (118) 213 Cellobiose, (118) 35 Cellulase, (118) 17, 386 Cellulose, (118) 386 Chaperones, (118) 406 Chinese hamster ovary cells, (118) 316 Chromatography, (118) 67 Cloning and sequencing, (118) 339

doi:10.1016/S0168-1656(05)00418-9

Cloning, (118) 353 Clostridium, (118) 328 Colitis, (118) 246 Competition, (118) 107 Continuous boiling lysis, (118) 257 Continuous culture, (118) 270, 328 Controllable expression of transgenes, (118) 123 Cross-linking, (118) 265 Cytochrome c, (118) 149 Different carbon sources, (118) 304 DNA sensors, (118) 233 DnaK, (118) 406 E. coli, (118) 406 Endoglucanase, (118) 386 Endotoxin, (118) 421 Environmental water sample, (118) 177 Enzymatic activity, (118) 265 Enzymatic resolution, (118) 99 Enzyme engineering, (118) 167 17-Estradiol (E2), (118) 177 Escherichia coli, (118) 75, 379 Ethanol, (118) 167, 278, 35 Family 15 glycosyl hydrolases, (118) 167 Fed-batch culture, (118) 75 Fermentation, (118) 149, 167 Fibre surface, (118) 386 Flow cytometry, (118) 370 Formic acid, (118) 107 Fusarium oxysporum, (118) 304, 99 Fusion tail, (118) 363 Galactose limited, (118) 270 Gene expression, (118) 149 Gene regulation, (118) 9 Genetic engineering, (118) 35 Genome, (118) 17 Glass, (118) 265 Glucose, (118) 316 Glucuronan, (118) 448

Subject Index Glycosyl hydrolase, (118) 17 GroEL, (118) 406 Haldane kinetics, (118) 45 Heat shock protein, (118) 278 Hepatitis delta virus, (118) 135 Heterologous protein production, (118) 270 Human lactoferrin, (118) 246 Hybrid grey-box modelling, (118) 290 Hybrid protein, (118) 167 Hydrogen, (118) 328 Hydrolysis, (118) 413 Hydroxyapatite, (118) 187 IgG1–IL2 fusion protein, (118) 290 Immobilised metal ion-affinity chromatography, (118) 363 Immobilization, (118) 99 Immunohistochemistry, (118) 370 Immunorelevant epitope, (118) 201 Impedance spectroscopy, (118) 398 Inhibition assay, (118) 370 Introns, (118) 339 Ion-exchange chromatography, (118) 363 Kluyveromyces, (118) 149 LacI, (118) 9 Lactate, (118) 316 Lactococcus lactis, (118) 55 Lactonase, (118) 99 Lactose, (118) 328 Large-scale, (118) 257 Ligninolytic peroxidases, (118) 339 Loop, (118) 1 Low expression promoters, (118) 123 Malolactic fermentation, (118) 353 Mammalian cell, (118) 135 Metabolic engineering, (118) 55 Metabolism, (118) 316 Metabolite profile, (118) 304 Methyl tert-butyl ether, (118) 45 Microarray, (118) 1 Microbial biomass protein, (118) 413 Microencapsulation, (118) 434 Micro-plate magnetic chemiluminescence enzyme immunoassay (MMCLEIA), (118) 177 Micro-plate magnetic separator, (118) 177 mleD gene, (118) 353 Monoclonal antibody, (118) 201, 370 Monolith column, (118) 421 mRNA processing, (118) 149 Multi-gene expression clone, (118) 123 Multisite Gateway cloning, (118) 123 Nanostructures, (118) 233 Nif promoter, (118) 9

463

Nitrogen limited, (118) 270 Non line-of-sight coatings, (118) 187 Non-disruptive harvesting, (118) 434 Oligosaccharides, (118) 448 On-line monitoring, (118) 398 Organophosphorous hydrolase, (118) 379 Paper, (118) 386 PEGylation, (118) 67 Periplasmic space, (118) 379 Peroxidase, (118) 363 pH change, (118) 278 Phanerochaete chrysosporium, (118) 17 Plasmid, (118) 257 Poly(ethylene glycol), (118) 67 Polymers, (118) 233 Polysaccharide lyases, (118) 448 Polyurethane foam, (118) 187 Porcine circovirus type II, (118) 201 Principal component analysis, (118) 304 Process monitoring, (118) 316 Process optimization, (118) 290 Protein folding, (118) 406 Protein immobilization, (118) 265 Protein, (118) 67 Proteome, (118) 17 Pulp characterization, (118) 386 Purification, (118) 157, 363 Quantitative trait, (118) 1 rbpa gene, (118) 339 Recombinant baculovirus, (118) 246 Recombinant Capsid protein defecting nuclear localization signal, (118) 201 Redox balance, (118) 304 Response surface methodology, (118) 157 Rhamnulose 1-phosphate aldolase, (118) 75 rhG-CSF, (118) 67 Rice straw, (118) 413 Rosa damascena Mill 1803 cell suspension culture, (118) 89 Saccharomyces cerevisiae, (118) 353 Salinity, (118) 278 Secretion, (118) 379 Secretome, (118) 17 Self-assembly, (118) 213 Sequencing batch reactor, (118) 45 Serratia, (118) 187 Shear stress, (118) 278 Silkworm larvae, (118) 246 Simultaneous introduction of multiple cDNAs, (118) 123 Stem cell research, (118) 434 Stress protein, (118) 278 Substrate inhibition, (118) 45

464 Sulfate reduction, (118) 107 Supermacroporous gel, (118) 421 T5 promoter, (118) 75 Telomerase, (118) 370 Tert-butyl alcohol, (118) 45 Tertiary structure, (118) 339 Thermophilic conditions, (118) 328 Tissue Engineering, (118) 213 Tissue engineering, (118) 434 Titanium disc, (118) 187 TorA, (118) 379 Twin-arginine translocation pathway, (118) 379 Two-phase cultivation, (118) 89

Subject Index Vascularization, (118) 213 Virus transduction, (118) 135 Virus-like particle, (118) 135 Volatile and polar metabolites, (118) 89 White-rot fungi, (118) 339 Whole cell biocatalyst, (118) 379 Xylitol, (118) 55 Xylose reductase, (118) 55 Xylose transport, (118) 55 Xylose, (118) 55 Zymobacter palmae, (118) 35 Zymomonas mobilis, (118) 35


Contents of Volume 118 Nucleic Acids/Molecular Biology Regular papers G.W. Horgan (UK) Optimising two-dye microarray designs for estimating associations with a quantitative trait

1

H.J.O. Ramos, E.M. Souza, J.R.L. Soares-Ramos and F.O. Pedrosa (Brazil) A new system to control the barnase expression by a NifA-dependent promoter

9

A.V. Wymelenberg, G. Sabat, D. Martinez, A.S. Rajangam, T.T. Teeri, J. Gaskell, P.J. Kersten and D. Cullen (USA, Sweden) The Phanerochaete chrysosporium secretome: Database predictions and initial mass spectrometry peptide identifications in cellulose-grown medium

17

Physiology/Biochemistry Regular paper H. Yanase, K. Yamamoto, D. Sato and K. Okamoto (Japan) Ethanol production from cellobiose by Zymobacter palmae carrying the Ruminocuccus albus -glucosidase gene

35

Biochemical Engineering/Bioprocess Engineering Regular papers W.-Q. Zhuang, J.-H. Tay, S. Yi and S.T.-L. Tay (Singapore) Microbial adaptation to biodegradation of tert-butyl alcohol in a sequencing batch reactor

45

A. Nyyssölä, A. Pihlajaniemi, A. Palva, N. von Weymarn and M. Leisola (Finland) Production of xylitol from D-xylose by recombinant Lactococcus lactis

55

Industrial Processes/New Products Regular papers Q. Yun, R.E. Yang, T. Chen, J. Bi, G. Ma and Z. Su (PR China) Reproducible preparation and effective separation of PEGylated recombinant human granulocyte colonystimulating factor with novel “PEG-pellet” PEGylation mode and ion-exchange chromatography

67

L. Vidal, P. Ferrer, G. Álvaro, M.D. Benaiges and G. Caminal (Spain) Influence of induction and operation mode on recombinant rhamnulose 1-phosphate aldolase production by Escherichia coli using the T5 promoter

75

doi:10.1016/S0168-1656(05)00419-0

466

Contents of Volume 118

A. Pavlov, S. Popov, E. Kovacheva, M. Georgiev and M. Ilieva (Bulgaria) Volatile and polar compounds in Rosa damascena Mill 1803 cell suspension K. Sakamoto, K. Honda, K. Wada, S. Kita, K. Tsuzaki, H. Nose, M. Kataoka and S. Shimizu (Japan) Practical resolution system for DL-pantoyl lactone using the lactonase from Fusarium oxysporum

89 99

P.L. Paulo, R. Kleerebezem, G. Lettinga and P.N.L. Lens (The Netherlands, Brazil) Cultivation of high-rate sulfate reducing sludge by pH-based electron donor dosage

107

Guide for Authors

117

Nucleic Acids/Molecular Biology Regular papers K. Yahata, H. Kishine, T. Sone, Y. Sasaki, J. Hotta, J.D. Chesnut, M. Okabe and F. Imamoto (Japan, USA) Multi-gene Gateway clone design for expression of multiple heterologous genes in living cells: Conditional gene expression at near physiological levels

123

Y.-H. Chen, J.-C. Wu, K.-C. Wang, Y.-W. Chiang, C.-W. Lai, Y.-C. Chung and Y.-C. Hu (Taiwan) Baculovirus-mediated production of HDV-like particles in BHK cells using a novel oscillating bioreactor

135

S. Seoane, B. Guiard, A.-M. Rodríguez-Torres and M.A. Freire-Picos (Spain, France) Effects of splitting alternative KlCYC1 3-UTR regions on processing: Metabolic consequences and biotechnological applications

149

Physiology/Biochemistry Regular papers W. Zhi, J. Song, F. Ouyang and J. Bi (PR China) Application of response surface methodology to the modeling of -amylase purification by aqueous two-phase systems

157

L. Latorre-García, A.C. Adam, P. Manzanares and J. Polaina (Spain) Improving the amylolytic activity of Saccharomyces cerevisiae glucoamylase by the addition of a starch binding domain

167

Biochemical Engineering/Bioprocess Engineering Regular papers L. Zhao and J.-M. Lin (China) Development of a micro-plate magnetic chemiluminescence enzyme immunoassay (MMCLEIA) for rapid- and high-throughput analysis of 17-estradiol in water samples

177

L.E. Macaskie, P. Yong, M. Paterson-Beedle, A.C. Thackray, P.M. Marquis, R.L. Sammons, K.P. Nott and L.D. Hall (UK) A novel non line-of-sight method for coating hydroxyapatite onto the surfaces of support materials by biomineralization

187

Industrial Processes/New Products Regular papers J.-Y. Zhou, S.-B. Shang, H. Gong, Q.-X. Chen, J.-X. Wu, H.-G. Shen, T.-F. Chen and J.-Q. Guo (China) In vitro expression, monoclonal antibody and bioactivity for capsid protein of porcine circovirus type II without nuclear localization signal

201


467

Medical Biotechnology Regular papers J.M. Kelm, C. Diaz Sanchez-Bustamante, E. Ehler, S.P. Hoerstrup, V. Djonov, L. Ittner and M. Fussenegger (Switzerland, UK) VEGF profiling and angiogenesis in human microtissues

213

Nucleic Acids/Molecular Biology Regular papers L. Muñoz-Serrano, A.R. Guadalupe and E. Vega-Bermudez (USA) Morphological studies of oligodeoxyribonucleotides probes covalently immobilized at polystyrene modified surfaces

233

T. Liu, Y.-Z. Zhang and X.-F. Wu (China) High level expression of functionally active human lactoferrin in silkworm larvae

246

K. Zhu, H. Jin, Y. Ma, Z. Ren, C. Xiao, Z. He, F. Zhang, Q. Zhu and B. Wang (China) A continuous thermal lysis procedure for the large-scale preparation of plasmid DNA

257

Physiology/Biochemistry Regular paper R.H. Taylor, S.M. Fournier, B.L. Simons, H. Kaplan and M.A. Hefford (Canada) Covalent protein immobilization on glass surfaces: Application to alkaline phosphatase

265

Biochemical Engineering/Bioprocess Engineering Regular papers Y.E. Thomassen, A.J. Verkleij, J. Boonstra and C.T. Verrips (The Netherlands) Specific production rate of VHH antibody fragments by Saccharomyces cerevisiae is correlated with growth rate, independent of nutrient limitation

270

A. Huhtala, P. Linko and R. Mutharasan (Finland, USA) Protein response of insect cells to bioreactor environmental stresses

278

A. Teixeira, A.E. Cunha, J.J. Clemente, J.L. Moreira, H.J. Cruz, P.M. Alves, M.J.T. Carrondo and R. Oliveira (Portugal) Modelling and optimization of a recombinant BHK-21 cultivation process using hybrid grey-box systems

290

G. Panagiotou, P. Christakopoulos and L. Olsson (Denmark, Greece) The influence of different cultivation conditions on the metabolome of Fusarium oxysporum

304

Industrial Processes/New Products Regular papers Y.-S. Tsao, A.G. Cardoso, R.G.G. Condon, M. Voloch, P. Lio, J.C. Lagos, B.G. Kearns and Z. Liu (USA) Monitoring Chinese hamster ovary cell culture by the analysis of glucose and lactate metabolism

316

C. Collet, O. Gaudard, P. Péringer and J.-P. Schwitzguébel (Switzerland) Acetate production from lactose by Clostridium thermolacticum and hydrogen-scavenging microorganisms in continuous culture—Effect of hydrogen partial pressure

328

468


Nucleic Acids/Molecular Biology Regular papers P.R. Moreira, C. Duez, D. Dehareng, A. Antunes, E. Almeida-Vara, J.M. Frère, F.X. Malcata and J.C. Duarte (Portugal, Belgium) Molecular characterisation of a versatile peroxidase from a Bjerkandera strain

339

R. Bauer, H. Volschenk and L.M.T. Dicks (South Africa) Cloning and expression of the malolactic gene of Pediococcus damnosus NCFB1832 in Saccharomyces cerevisiae

353

Physiology/Biochemistry Regular papers G. Levin, F. Mendive, H.M. Targovnik, O. Cascone and M.V. Miranda (Argentina) Genetically engineered horseradish peroxidase for facilitated purification from baculovirus cultures by cation-exchange chromatography

363

D. Soldateschi, S. Bravaccini, B. Berti, A. Brogi, T. Benicchi, C. Soldatini, L. Medri, F. Fabbri, F. De Paola, D. Amadori and D. Calistri (Italy) Development and characterization of a monoclonal antibody directed against human telomerase reverse transcriptase (hTERT)

370

D.G. Kang, G.-B. Lim and H.J. Cha (Republic of Korea) Functional periplasmic secretion of organophosphorous hydrolase using the twin-arginine translocation pathway in Escherichia coli

379

L. Hildén, P. Väljamäe and G. Johansson (Sweden, Estonia) Surface character of pulp fibres studied using endoglucanases

386

Biochemical Engineering/Bioprocess Engineering Regular papers A. Soley, M. Lecina, X. Gámez, J.J. Cairó, P. Riu, X. Rosell, R. Bragós and F. Gòdia (Spain) On-line monitoring of yeast cell growth by impedance spectroscopy

398

N. González-Montalbán, M.M. Carrió, S. Cuatrecasas, A. Arís and A. Villaverde (Spain) Bacterial inclusion bodies are cytotoxic in vivo in absence of functional chaperones DnaK or GroEL

406

Y.-G. Zheng, X.-L. Chen and Z. Wang (PR China) Microbial biomass production from rice straw hydrolysate in airlift bioreactors

413

Industrial Processes/New Products Regular papers A. Hanora, F.M. Plieva, M. Hedström, I.Yu. Galaev and B. Mattiasson (Sweden, Egypt) Capture of bacterial endotoxins using a supermacroporous monolithic matrix with immobilized polyethyleneimine, lysozyme or polymyxin B

421

S.-M. Chia, P.-C. Lin, C.-H. Quek, C. Yin, H.-Q. Mao, K.W. Leong, X. Xu, C.-H. Goh, M.-L. Ng and H. Yu (Singapore, USA) Engineering microenvironment for expansion of sensitive anchorage-dependent mammalian cells

434


469

C. Delattre, P. Michaud, J.M. Lion, B. Courtois and J. Courtois (France) Production of glucuronan oligosaccharides using a new glucuronan lyase activity from a Trichoderma sp. strain 448 Author Index

458

Subject Index

462


465

I (Contents continued from outside back cover) Industrial Processes/New Products Regular papers A. Hanora, F.M. Plieva, M. Hedström, I.Yu. Galaev and B. Mattiasson (Sweden, Egypt) Capture of bacterial endotoxins using a supermacroporous monolithic matrix with immobilized polyethyleneimine, lysozyme or polymyxin B

421

S.-M. Chia, P.-C. Lin, C.-H. Quek, C. Yin, H.-Q. Mao, K.W. Leong, X. Xu, C.-H. Goh, M.-L. Ng and H. Yu (Singapore, USA) Engineering microenvironment for expansion of sensitive anchorage-dependent mammalian cells

434

C. Delattre, P. Michaud, J.M. Lion, B. Courtois and J. Courtois (France) Production of glucuronan oligosaccharides using a new glucuronan lyase activity from a Trichoderma sp. strain 448 Author Index

458

Subject Index

462


465

doi:10.1016/S0168-1656(05)00424-4