Rhodococcus sp. Strain IGTS8 - NCBI

13 downloads 0 Views 2MB Size Report
culture supernatants were analyzed by high-performance liquid chromatography (HPLC). ... Midland Certified Reagent Co. (Midland, Tex.). The mutant.
Vol. 176, No. 21

JOURNAL OF BACTERIOLOGY, Nov. 1994, p. 6707-6716

0021-9193/94/$04.00+0 Copyright © 1994, American Society for Microbiology

Characterization of the Desulfurization Genes from Rhodococcus sp. Strain IGTS8 SYLVIA A. DENOME,' CHRISTOPHER OLDFIELD,2 LISA J. NASH,2 AND KEVIN D. YOUNG'* Department of Microbiology and Immunology, University of North Dakota School of Medicine, Grand Forks, North Dakota 58202,1 and Energy Biosystems Corporation, The Woodlands, Texas 773812 Received 13 May 1994/Accepted 29 August 1994

Rhodococcus sp. strain IGTS8 possesses an enzymatic pathway that can remove covalently bound sulfur from dibenzothiophene (DBT) without breaking carbon-carbon bonds. The DNA sequence of a 4.0-kb BstBI-BsVI fragment that carries the genes for this pathway was determined. Frameshift and deletion mutations established that three open reading frames were required for DBT desulfurization, and the genes were designated soxABC (for sulfur oxidation). Each sox gene was subcloned independently and expressed in Escherichia coli MZ1 under control of the inducible APL promoter with a XcII ribosomal binding site. SoxC is an -45-kDa protein that oxidizes DBT to DBT-5,5'-dioxide. SoxA is an -50-kDa protein responsible for metabolizing DBT-5,5'-dioxide to an unidentified intermediate. SoxB is an -40-kDa protein that, together with the SoxA protein, completes the desulfurization of DBT-5,5'-dioxide to 2-hydroxybiphenyl. Protein sequence comparisons revealed that the predicted SoxC protein is similar to members of the acyl coenzyme A dehydrogenase family but that the SoxA and SoxB proteins have no significant identities to other known proteins. The sox genes are plasmidborne and appear to be expressed as an operon in Rhodococcus sp. strain IGTS8 and in E. coli.

Burning sulfur-containing petroleum and coal contributes to environmental degradation in various ways (7). Although removal of inorganic sulfur from these fuels may be accomplished by physical, chemical, or biological means, organically bound sulfur is difficult to remove (1). One possible strategy for reducing the organic sulfur content is to expose these substrates to microorganisms or enzymes that can specifically break carbon-sulfur bonds (7, 18, 21), thereby releasing the sulfur in a water-soluble, inorganic form. The gram-positive bacterium Rhodococcus sp. strain IGTS8 is able to extract sulfur from a variety of organosulfur compounds, including thiophenes, sulfides, mercaptans, sulfoxides, and sulfones (15). Furthermore, the organism is able to remove organic sulfur from petroleum and soluble coal-derived materials without decreasing the caloric value of those substrates (21). Although the sulfur released becomes available for use in cell growth, the original substrates do not serve as carbon sources (15), indicating that a general degradation of the compounds does not occur. In particular, IGTS8 can remove sulfur from dibenzothiophene (DBT) to yield 2-hydroxybiphenyl (2HBP) as the final product, meaning that there is no degradation of the carbon-carbon bonds in the DBT molecule (3, 8, 21). Previously, we cloned a set of genes from IGTS8 that conferred the DBT desulfurization phenotype to a desulfurization-negative mutant of IGTS8 and to another species, Rhodococcus fascians (3). We now report the DNA sequences of the three genes that mediate this reaction and provide preliminary characterization of the enzymology by which the encoded proteins specifically break carbon-sulfur bonds.

contained (per liter) 0.2% glycerol, 4 g of NaH2PO4 * H20, 4 g of K2HP04 * 3H20, 2 g of NH4Cl, 0.2 g of MgCl2 6H20, 0.001 g of CaCl2 * 2H20, and 0.001 g of FeCl3 * 6H20. For assays involving Escherichia coli XL1-Blue or MZ1, BSM was supplemented with 0.01% Casamino Acids and thiamine (5 [ug/ml) to support growth. Luria broth (LB) contained 1% Bactotryptone, 0.5% yeast extract, and 1% NaCl. Ampicillin (100 ,ug/ml) and tetracycline (12.5 ,ug/ml) were included for plasmid selection as needed. DBT was purchased from Fluka. DBT-5,5'dioxide (DBTO2) was purchased from Pfaltz and Bauer. Unless otherwise specified, restriction enzymes, DNA-modifying enzymes, and linkers were purchased from New England Biolabs. Bacteria and plasmids. The bacterial strains used in this work are listed in Table 1. pLAFR5 was obtained from N. T. Keen (16). pRF29 (from J. Desomer) was the source of the Rhodococcus origin of replication (4). pBluescript (Stratagene) was used to generate nested deletion clones for DNA sequencing. pRE1, used to express the desulfurizing genes under the control of the X\PL promoter, was a gift of P. Reddy (30). Plasmid DNA isolation and Rhodococcus transformation. Plasmid DNA from E. coli was isolated by the method of Ish-Horowicz and Burke (14) or Holmes and Quigley (11). Plasmid DNA was isolated from Rhodococcus spp. as described by Singer and Finnerty (34). Transformation of Rhodococcus strain UV1 was performed via electroporation as described previously (3). Desulfurization assays. The plate assay for the identification of bacteria capable of modifying DBT was performed as described elsewhere (3, 23), with minor modifications; bacteria that metabolized DBTO2 were also detected with this assay. Briefly, cells were incubated on BSM-1% agarose plates sprayed with a 0.1% DBT (or DBTO2) solution in ethyl ether. The test substrates precipitate out as a white film immediately when sprayed onto the plate surface. Three distinct metabolic phenotypes are distinguishable when the plates are viewed under short-wave (254-nm) UV illumination. First, when 2HBP is produced from DBT (or DBTO2), a bright blue-white -

MATERIALS AND METHODS Media and reagents. Basal salts medium (BSM) was used to assay bacteria for their ability to desulfurize DBT. BSM * Corresponding author. Phone: (701) 777-2624. Fax: (701) 7773894. Electronic mail address: [email protected].

6707

6708

DENOME ET AL.

J. BACT1ERIOL.

TABLE 1. Bacterial strains Bacterium E. coli strains MZ1

NC3 (B/r) S17-1 XL1-Blue

Genotype or description his ilv rspL galK(Am) pglA8 (bio-uvrB)AH1 hsr rpsL (Strr) recA thi pro hsdR (r- m+) [RP4-2-Tc::Mu-Km::Tn71 recA1 lac thi endA1 gyrA96 hsdR7 supE44 relA1 (F' proAB lacdq lacZAM15 TnlO)

Source and/or reference

P. Reddy (30) C. Helmstetter A. Puhler (33)

Stratagene

(Tetr) Rhodococcus sp. strains IGTS8 UVi

DBT-desulfurizing environmental J. J. Kilbane (20) isolate ATCC 53968 3 Nondesulfurizing mutant of IGTS8

UV-fluorescent product that is very soluble is produced. This fluorescence can spread across the entire petri dish with prolonged incubations. The substrate sprayed on the plate completely disappears, beginning as a cleared zone around the colony and eventually spreading across the entire plate. A second fluorescent phenotype is observed when DBT is metabolized to DBTO2. This fluorescence is less bright than the fluorescence seen with 2HBP, and the DBTO2 fluorescence remains closely associated with the bacterial colony. Also, the production of DBTO2 from DBT is not associated with complete clearing of DBT from the assay plate as seen with the production of 2HBP. Unlike 2HBP and DBTO2, the third intermediate detected in the DBT plate assay is not UV fluorescent. Its presence is detected by the complete clearing of DBT or DBTO2 from the assay plate. The identity of this intermediate has not been determined. Desulfurization assays were also performed in liquid culture. For these assays, cells were incubated overnight at 30'C in LB medium, pelleted by centrifugation, and then washed three times in BSM. Washed cells were used to inoculate fresh BSM containing the test substrate so that the beginningA60 was 0.1. The test substrate (DBT or DBTO2) was supplied at a concentration of 0.1% or as specified in the figure legends. The assay cultures were shaken at 30'C for 24 to 48 h, and the culture supernatants were analyzed by high-performance liquid chromatography (HPLC). For desulfurization assays involving E. coli XL1-Blue and MZ1, 0.01% Casamino Acids and thiamine (5 pug/ml) were added to support growth. For desulfurization assays that involved an incomplete desulfurization pathway, 2 mM (NH4)2SO4 was included to support growth. DNA sequencing and analysis. Deletion clones for DNA sequencing were constructed in pBluescript by using exonuclease III and the method of Henikoff (9). The DNA nucleotide sequences of both strands were determined by the dideoxy chain termination method of Sanger et al. (32), using Sequenase 2.0 (U.S. Biochemical) and a-_3S-dATP (Amersham). Sequence assembly and analysis were performed with DNA Inspector II (Textco, Hanover, N.H.). The SwissProt protein database was searched for amino acid similarities with the BLITZ program of Sturrock and Collins (37), which implements the Smith and Waterman best local similarity algorithm (35). Predicted amino acid sequences were also compared with the Prosite motif compilation by using the BLOCKS electronic-mail server (10). Optimal alignments be-

tween two protein sequences were performed by using the FastP algorithm of Lipman and Pearson (24) at a ktuple of 1. Construction of mutations and subcloning of the sox genes. Frameshift mutations in individual sox open reading frames (ORFs) were constructed in pSAD69-37 at the specific sites indicated in Fig. 1 by using partial restriction enzyme digestion and blunting the ends with DNA polymerase I Klenow fragment. The blunt-ended ScaI restriction site was interrupted with HindIll linkers. By site-directed mutagenesis, the DNA sequence upstream of each sox ORF was altered to create an NdeI site that included the AUG initiation codon. This mutation facilitated the insertion of each sox gene downstream of the XcII ribosomal binding site in pSAD262-1 (see Fig. 3). Mutagenesis was performed by using oligonucleotide-directed in vitro mutagenesis kit version 2.1 (Amersham), with oligomers obtained from Midland Certified Reagent Co. (Midland, Tex.). The mutant oligomers (shown 5' to 3', with the NdeI site underlined) were as follows: soxA, GGACGCATACCATATGACTCAAC; soxB, AAGGACAACCCATATGACAAGCC; and soxC, AG GAACATCCATATGACACTGT. Induction of Sox proteins from the APL promoter. To induce expression of the Sox proteins from the APL promoter, we followed the protocol described by Reddy et al. (30). E. coli MZ1 containing pSAD267-1, pSAD277-7A, or pSAD269-2A was grown in LB-ampicillin overnight with shaking at 30'C. Cells were diluted 1:100 with fresh LB-ampicillin and shaken at 30'C until the A600 was between 0.4 and 0.6. An equal volume of LB-ampicillin (prewarmed to 50'C) was added, and the cells were shaken at the induction temperature (39 or 420C) for 15 min to 2 h. Cells were harvested by centrifugation at 3,000 X g for 15 min at 4°C and examined for protein expression by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) or analyzed for activity in the liquid desulfurization assay. For use in the desulfurization assay, pelleted cells were washed three times with BSM and used to inoculate fresh BSM-Casamino Acids-thiamine-ampicillin containing 0.05% DBT or DBTO2 to anA6. of 0.1. To assay more than one sox ORF together with the same substrate, the clones were induced separately under optimal induction conditions and cells from the separately induced cultures were combined for the assay. The assay cultures were shaken at 30°C for 24 to 48 h and analyzed by HPLC. HPLC analysis. An aliquot of the desulfurization assay cell cultures was diluted with an equal volume of acetonitrile and vortexed extensively to solubilize any water-insoluble intermediates. The cells were removed by centrifugation at 16,000 X g for 3 min at room temperature, and the supernatants were analyzed by reversed-phase HPLC with a Hewlett-Packard series 1050 system equipped with a diode array detector. The column was a Synchropak RP C18 column (4.6 by 100 mm), and elution was isocratic with nitrogen-purged acetonitrile-phosphate buffer (10 mM, pH 6.0) (50:50 [vol/vol]) pumped at 1.5 ml/min. The retention times of DBT, DBT-5-oxide (DBTO) DBTO2, and 2HBP were confirmed by internal and external standards. Values reported from HPLC analysis represent the area under the curve. SDS-PAGE. Proteins from whole cells (109 cells per lane) were separated on a 12.5% acrylamide resolving gel with a 4.0% acrylamide stacking gel by the methods of Dreyfuss et al.

(5).

Hexagonal pulsed-field gel electrophoresis. Rhodococcus sp. strains IGTS8 and UV1 were grown overnight with shaking at 30°C in LB. Cells were centrifuged at 3,000 X g for 15 min at 4°C and resuspended at 2 X 109/ml in 75 mM NaCl-25 mM EDTA. The cell suspension was warmed to 37°C and added to

CHARACTERIZATION OF DBT DESULFURIZATION GENES

VOL. 176X 1994

Sptd

BstZSclI

BstEII Xhdl

1 1 1 |[2

0

BME, BsIhe SPM

MItA

3

14 r

BstMl

I 5?9

soxA

soxC

6709

ECdR kb

-

soxB

mutant

Scal

MAA sB mutant

O mutant

mutant pSAD161-12

pSAD164-18 Apal

pSAD159-6

pSAD1 60-7

mutant

pSAD1 62-8 Sau3AI

EaI pSAD69-37

Scal

pSAD94-56

BsIeB 'tac BsI

BstBIsEII Flac *BslB l I *BslBI Bc BstBI

F/ac -0

Bsf1 = pSAD225-32 (pSAD1 67-32*) BsIEII

--44

I

NheI

BsNVI = pSAD231-4 BsNI ':j pSADI 74-4 I pSAD236-2

I

pSAD278-2 ADal 3 I ~~~~pSAD175-8 %lc FIG. 1. Partial restriction map of the 9.7-kb Sau3AI-EcoRI fragment that encodes the desulfurization genes from Rhodococcus sp. strain IGTS8. For clarity, only restriction sites discussed in the text are shown. ORFs >200 amino acids in length (horizontal arrows) and clones constructed from pSAD69-37 (vertical arrows) that contain frameshift mutations at the sites indicated are shown. Deletion plasmids are diagrammed below the restriction map; gene segments still present (open rectangles) and deleted DNA (lines) are indicated. pSAD167-32* contains the same DNA as indicated for pSAD225-32 but in the opposite orientation. 0. r-

equal volume of warm (37°C) 2% SeaPlaque agarose (FMC Corp.) in 10 mM Tris-HCl (pH 7.4)-10 mM MgCl2-0.1 mM EDTA. The cell-agarose mixture was poured into block molds (5 by 10 by 2 mm) and allowed to solidify at 4°C. The agarose blocks were equilibrated overnight at 4°C in 10 mM Tris-HCl (pH 7.5)-0.5 mM EDTA. Lysozyme (5 mg/ml) was added, and the blocks were incubated at 37°C for 4 h. Spheroplasts in the blocks were digested at 56°C for >15 h in 1% Sarkosyl-0.5 mM EDTA-0.5 mg of proteinase K per ml (pH 9.5), equilibrated extensively (>24 h) in 10 mM Tris-HCl (pH 7.4)-0.5 mM EDTA, and stored in this buffer at 4°C. Blocks were trimmed to 5 by 5 by 1 mm for electrophoresis in an LKB model 2015 Pulsaphor Unit (hexagonal electrode array). The separation gel (15 by 15 cm) was 1% agarose in 0.5X TBE (45 mM Tris base, 50 mM boric acid, and 0.5 mM EDTA). Samples were electrophoresed at 170 V with a 45-s pulse setting for 28 h at 15°C. DNA was visualized by ethidium bromide staining and transferred to Biotrans nylon membranes (ICN Biomedicals) by the method of Southern (36). The 4.0-kb BstBI-BsiWI DNA fragment from pSAD167-32 was labeled with [cx-32P]dCTP (Amersham) by the method of Feinberg and Vogelstein (6). Hybridizations were performed in 6X SSC (lx SSC is 0.15 M NaCl plus 0.015 M sodium citrate)-50% formamide at 42°C, and the membranes were washed under stringent conditions (0.1 X SSC at 50°C) as described by Sambrook et al. (31). Nucleotide sequence accession number. The nucleotide sequence of the 9.7-kb Sau3AI-EcoRI fragment from Rhodococcus sp. strain IGTS8 has been submitted to GenBank and assigned accession number U08850. an

RESULTS

The Rhodococcus desulfurization genes are expressed in E. coli. We previously cloned DNA fragments from Rhodococcus sp. strain IGTS8 and showed that the ability to desulfurize DBT to 2HBP could be transferred to a desulfurizationnegative mutant of IGTS8, strain UV1, and also to a separate species, R fascians (3). To make it easier to manipulate the genes, the desulfurization phenotype was also expressed in E. coli.

Two subclones that contained the 6.5-kb BstBI fragment from pSAD48-12 (3) in both orientations with respect to the lac promoter were constructed in pBluescript KS- and SK(pSAD225-32 and pSAD167-32) (Fig. 1). E. coli was transformed by the plasmids, and the ability of the E. coli cells to desulfurize DBT was measured in the desulfurization assay. HPLC analysis of whole cells cultured with DBT as the sole sulfur source confirmed that pSAD225-32 conferred the desulfurization phenotype to E. coli NC3 (Table 2). The production of 2HBP was increased by the addition of 0.01% Casamino Acids, but 0.1% Casamino Acids was inhibitory (Table 2). No 2HBP was detected in the negative control (NC3 containing pBluescript KS-). E. coli NC3 transformed with pSAD167-32 was incapable of desulfurizing DBT to 2HBP, as determined by the DBT plate assay. Since the same DNA fragment was present in opposite orientations in pSAD167-32 and pSAD225-32, this indicated that the desulfurization genes in pSAD225-32 were being expressed under control of the lac promoter and not from an endogenous Rhodococcus promoter. These results confirmed that the Rhodococcus gene

6710

DENOME ET AL.

J.

TABLE 2. Expression of desulfurization genes in E. coli

fere with the transcription of downstream ORFs, including the

Sox gene(s)

Additive

2HBP detected

NC3/pSAD225-32

soxABC

NC3/pBluescript KS-

soxABC soxABC None

None 0.01% Casamino Acids 0.1% Casamino Acids None

3.3 12.0 0.8 0

E. coli strain

BACTERIOL.

(AUC)b

a The substrate for the desulfurization assay was 0.1% DBT. b AUC, area under the curve (from HPLC analysis of desulfurization assays with whole cells). Values given are from the region of the curve matching the retention time of 2HBP.

products could be expressed in active form in E. coil. Furthermore, since the phenotype was dependent on only the activity of a single lac promoter, the results suggested that one gene mediated the desulfurization reaction or that multiple genes were being transcribed as an operon in E. col. A 4.0-kb BstBI-BsiWI fragment was subcloned from the 6.5-kb BstBI fragment to create pSAD231-4 (Fig. 1). This plasmid contained the smallest DNA fragment that could impart to E. coli the ability to desulfurize DBT to 2HBP (data not shown). DNA sequence and analysis. One hundred forty-three deletion subclones were generated by the exonuclease III deletion strategy and used to sequence both strands of the Rhodococcus 9.7-kb Sau3AI-EcoRI DNA fragment. Sequence analysis of the 4.0-kb BstBI-BsiWI fragment (which encodes the desulfurization genes) revealed that the DNA strand that was under control of the E. coli lac promoter in pSAD225-32 encoded three ORFs that began with AUG and were >200 amino acids in length (Fig. 1 and 2). These three ORFs were designated soxA, soxB, and soxC (for sulfur oxidation), reflecting the order in which they appeared in the sequence. The opposite DNA strand encoded two more ORFs within the 4.0-kb BstBI-BsiWI fragment, designated ORFi and ORF2 (Fig. 1). The ORFi sequence extended beyond the BsiWI site, making it unlikely to encode a desulfurization gene. ORF2, which spanned soxB and soxC, could be mutated without affecting desulfurization (see below), indicating that it was not involved in the desulfurization pathway. In addition, expression of these two ORFs under control of the lac promoter in pSAD167-32 did not result in DBT desulfurization (data not shown). We originally attempted to express the phenotype in E. coli by cloning the 15-kb EcoRI-HindIII DNA fragment from pSAD48-12 into pBluescript SK- and KS-. When cloned in either orientation with respect to the lac promoter, this DNA did not confer to E. coli the ability to desulfurize DBT (3). The DNA sequence in the region 5' to the sox genes was examined for possible E. coli transcription terminators. Such a sequence might explain why the 15-kb EcoRI-HindlIl fragment from pSAD48-12 did not confer the desulfurization phenotype to E. coli, even when cloned under the lac promoter (3). A short inverted repeat followed by four thymidines was located -300 nucleotides 5' to soxA (data not shown). If transcribed into RNA, this sequence (CGACGAUCUAACCAGAUCGACG GUUUU) could form a stem-loop structure that might inter-

sox genes. Thus, our previous inability to express these genes in E. coli (3) might be attributed to premature transcription

termination in the original subclone. Three gene products are involved in DBT desulfurization. To determine which genes were specifically involved in DBT desulfurization, deletions and point mutations were created in each of the sox genes. To avoid the possibility of premature transcription termination in E. coli (see above), the 1.5-kb Sau3AI-BstBI fragment was deleted by cloning the 6.5-kb BstBI fragment into pBluescript KS- (pSAD225-32; Fig. 1) under control of the lac promoter. Deletion of the 2.5-kb BsiWI-BstBI fragment created pSAD231-4 (Fig. 1), which still conferred to E. coli the ability to metabolize DBT to 2HBP. Other deletion mutants were constructed and examined for their ability to desulfurize DBT or DBTO2 in the plate or desulfurization assays. The cultures from the desulfurization assays were examined by HPLC for the accumulation of

desulfurization intermediates. The ABstEII-BstEII mutant (pSAD174-4; Fig. 1) contained only intact soxC and metabolized DBT to DBTO2 (Table 3). This result was confirmed in Rhodococcus organisms by using a frameshift mutation (see below) and indicated that the SoxC protein was sufficient to perform the first step in the desulfurization pathway, oxidation of DBT to DBTO2. The

ASphI-BsiWI

mutant

(pSAD236-2; Fig. 1)

contained

soxB, and ORF2 and metabolized DBTO2 to 2HBP (Table 3) but was incapable of metabolizing DBT to either DBTO2 or 2HBP (data not shown). Since the clone containing

soxA,

soxC alone (pSAD174-4) converted DBT to DBTO2, this implied that the SoxA, SoxB, and/or ORF2 protein carried out steps in the pathway subsequent to SoxC. This result confirmed that DBTO2 was an intermediate of DBT desulfurization and that the SoxA and/or SoxB protein could use DBTO2 as a substrate. A ANheI-BsiWI mutant (pSAD278-2; Fig. 1) with all of soxC and ORF1 removed, as well as the 5' end of ORF2, was constructed. This mutant, which retained soxA and soxB, converted DBTO2 to 2HBP, indicating that ORF2 was not essential for this step of the desulfurization pathway (Table 3). The role of soxA4 was tested by deleting the 1.7-kb ApaIBsiWI fragment (pSAD175-8; Fig. 1). This deletion inactivated soxB and soxC, leaving only soxA intact. E. coli NC3 containing pSAD175-8 metabolized DBTO2 to an unidentified intermediate. This intermediate was easily detected in the plate assay. When cells expressing soxA were incubated on agarose plates sprayed with DBTO2, the substrate was cleared from around the colony but no UV fluorescence was observed in the cleared area. This product was neither DBTO2 nor 2HBP, as both of those products were fluorescent in the plate assay. Subsequent experiments with soxA cloned under the control of the XpL promoter (pSAD267-1) revealed that the DBTO2 substrate in the desulfurization assay was completely consumed (see below), but we could not detect the appearance of a new peak on the HPLC chromatogram. It is possible that the new product peak was masked by the prominent peak at the solvent front. The fact that cells transformed with soxA metabolized DBTO2 but did not produce 2HBP suggested that the SoxA protein was responsible for the second step in the desulfurization

FIG. 2. DNA sequence of the 4.0-kb BstBI-BsiWI fragment containing the desulfurization genes from Rhodococcus sp. strain IGTS8. The sequence is shown 5' to 3', as determined from expression of the desulfurization phenotype from the lac promoter. Nucleotides representing possible ribosomal binding sites are marked above the sequence (>>>>). The predicted amino acids encoded by soxA, soxB, and soxC are shown below the nucleotide sequence. The Met amino acids in boldface type are the start codons for each of the three genes. The arrow at soxC indicates that the gene designation belongs to the open reading frame above soxC (compare with the locations of the soxA and soxB designations).

pn

p2.4a

a

00

C

0

0.4

9n10,

.4

0

Pt C-

0

"0

910

.4 c ..

.



C)

*~P

~

0.In

'(

0

'0

04 .4 a~~

*-'0

a-'~ '0~ ~o

004

8.

Fo

Q0' bOO

'(03 91

'(0

'0$.44 n

:,8.r

.

pn o

.

¢MQ

9 00

Ca,0t* sO

*0

a

-a

c

00

P:.O

.w w o u

0

o4 0nb 0

M

p.4

o0

:00

2,*

a~

o

v.

910o F

'(

*05

0

~~

6711

91p I."0'9

~~~

o'0~~~.40

0

6712

DENOME ET AL.

J. BACTERIOL.

TABLE 3. Desulfurization activity of sox gene expression E. E. colicclistrain strain

Cloned gene(s)

Substrate

A

AUC of product

Ndl Sad Kpnl Apal Xhol Sall Clal Hindill

detecteda DBT02 2HBP

soxC 0.1% DBT soxAB, ORF2 0.05 mM DBT02 NC3/pSAD236-2 soxAB 0.05 mM DBT02 NC3/pSAD278-2 NC3/pBluescript KS- None 0.1% DBT NC3/pSAD174-4

None

0.05 mM DBT02

98.8 0 0 0 96.1

0 18.5 21.7 0 0

EowRV EcoRI Pstl Smal

a AUC, area under the curve (from HPLC analysis of desulfurization assays with whole cells). Values given are from the regions of the curve matching the retention times of DBT02 and 2HBP.

pathway. Since the SoxA and SoxB proteins together are capable of desulfurizing DBT02 to 2HBP (see discussion of pSAD236-2 above) but are incapable of converting DBT to 2HBP, this suggests that SoxB performs the third and final step in the desulfurization of DBT to 2HBP. The activity of each gene product was also tested in Rhodococcus sp. strain UV1. Mutations in the sox genes were created in pSAD69-37 (Fig. 1). The mutated DNA was subcloned into pLAFR5 on a 9.7-kb EcoRI fragment, and the Rhodococcus origin of replication from pRF29, contained on a 4.5-kb HindIII fragment, was added to each construct. Deletion of the -1.7-kb Sau3AI-ScaI fragment (pSAD94-56; Fig. 1), which removed the 5' end of soxA and the upstream region, completely blocked DBT desulfurization in Rhodococcus strain UV1. Since soxC was present but DBT was not metabolized to DBT02, it is likely that the 1.7-kb Sau3AI-ScaI fragment contains a Rhodococcus promoter and that the sox genes are transcribed as an operon in Rhodococcus sp. strain IGTS8. A frameshift mutation at one of the SphI sites (pSAD160-7; Fig. 1) did not affect the metabolism of DBT to 2HBP, as determined by the DBT plate assay. This result was consistent with our observations that the 4.0-kb BstBI-BsiWI fragment encoded the desulfurization pathway. An insertion mutation at the second SphI site (pSAD159-6; Fig. 1) interrupted soxC. Cells containing this plasmid had no detectable desulfurization activity when DBT was used as a substrate in the plate or desulfurization assays. This confirmed that the SoxC protein is involved in the early steps of DBT metabolism and is consistent with the results with E. coli which indicated that SoxC produces DBTO2 from DBT. A frameshift mutation at the ScaI site (pSAD161-12; Fig. 1) interrupted soxA. Results obtained with the DBT plate assay indicated that DBTO2 was being produced but not 2HBP. HPLC analysis of desulfurization assay cultures showed a DBTO2 peak, indicating that this mutation halts DBT metabolism at DBTO2. These data, combined with the results for E. coli (see above), make it likely that DBTO2 is the substrate for the SoxA protein in the desulfurization pathway. Both the ApaI mutation (pSAD162-8; Fig. 1), which interrupted soxB and ORF2, and the MluI mutation (pSAD164-18; Fig. 1), which interrupted only soxB, blocked the desulfurization of DBT to 2HBP. Since both SoxC (which oxidizes DBT to DBTO2) and SoxA (which metabolizes DBTO2 to an unidentified intermediate) were active in these clones, DBTO2 did not accumulate. Formation of the unidentified intermediate by SoxA was confirmed by the clearing of DBT in the plate assay without the appearance of a UV-fluorescent product. Overexpression of the Sox proteins in E. coli. In order to increase expression of the Sox proteins in E. coli, each of the sox genes was cloned separately into plasmid pSAD262-1 (Fig.

BamHI Spel Xbal EcoRV Sall

B Apal

Ndel

APL-

oxA

pSAD267-1 Xbal

Ndel . XPL-

sor Ndel

xPL

I

-I I

pSAD277-7A

Xhol

IoC

pSAD269-2A FIG. 3. (A) Schematic representation (not drawn to scale) of the expression vector pSAD262-1 constructed from pRE1 (30). The region between the HindIII and EcoRI sites was deleted, and the multiple cloning site was expanded by adding a KpnI-XbaI fragment from the multiple cloning site in pBluescript KS-. Transcription is initiated from the APL promoter. to and tR1, lambda transcription terminators; clI RBS, XcII ribosomal binding site; bla, gene encoding ampicillin resistance. (B) Maps of the individual desulfurization genes (soxA, soxB, and soxC) cloned under the control of the A'PL promoter in pSAD262-1. Site-directed mutagenesis was used to create an NdeI site that included the AUG initiation codon for each desulfurization gene. Each mutated gene was then cloned into pSAD262-1, as shown. To excise the soxB and soxC genes for cloning, restriction sites in pBluescript (XbaI and XhoI) were utilized. Rhodococcus and/or vector DNA downstream of the sox gene that was included during cloning (shaded boxes) is indicated.

3A and B). Expression of the sox genes was controlled in two ways when E. coli MZ1 was transformed with plasmids pSAD267-1, pSAD277-7A, and pSAD269-2A. First, MZ1 contains the temperature-sensitive XcI857 protein, which represses transcription from X\PL at 30°C. Second, a transcription terminator sequence exists between the X\PL promoter and the XcII ribosome binding site preceding the sox genes (Fig. 3A), which inhibits transcription unless the XN antitermination protein is present. E. coli MZ1 contains the XN gene under control of the NPL promoter. Thus, when E. coli MZ1 is shifted from 30 to 42°C, cloned genes are induced by removing these dual repression controls. Each sox gene, when cloned into pSAD262-1 and transformed into E. coli MZ1, expressed high levels of protein at 42°C (Fig. 4). The molecular masses were -49 kDa (sequence prediction, 49,579 Da) for SoxA, -41 kDa (sequence prediction, 39,001 Da) for SoxB, and -47 kDa (sequence prediction, 44,977 Da) for SoxC. Enzymatic activity of overexpressed Sox proteins in E. coli. The activity for each enzyme varied with the induction time and temperature (Table 4). For SoxA (pSAD267-1; Fig. 3B),

CHARACTERIZATION OF DBT DESULFURIZATION GENES

VOL. 176, 1994 1

3

2

TABLE 5. Activities of Sox proteins expressed individually from the APL promoter in E. coli and then combineda

4

205,000 11 6,000 97,400 66,000

E.

None (no cells) MZ1/pSAD267-1 MZ1/pSAD277-7A MZ1/pSAD269-2A Mixedc

29 ,000 FIG. 4. Coomassie-stained SDS-polyacrylamnide gels showing overexpression of the desulfurization proteins expressed from the X{PL promoter in E. coli MZ1. Cells were induced at 420C for 60 min, with the equivalent of 109 cells loaded per lane. Lane 1, SoxA (-49,000 Da) expressed from pSAD267-1; lane 2, SoxB (-41,000 Da) expressed from pSAD277-7A; lane 3, SoxC (-47,000 Da) expressed from pSAD269-2A; lane 4, molecular weight markers.

enzyme activity was monitored by the decrease in the DBTO2 peak from the HPLC chromatogram. The greatest SoxA activity was observed with cells induced at 390C. Induction at 420C significantly reduced SoxA activity compared with that of inductions at 390C. SoxC (pSAD269-2A; Fig. 3B) oxidized DBT to DBTO2, and the greatest activity was observed when

cells were induced at 420C. Since we could not identify the product of DBTO2 metabolism by the SoxA protein, we were unable to assay for SoxB activity directly. Therefore, to measure SoxB activity, a mixed whole-cell assay was performed. Three cell cultures were grown, each containing one of the plasmids shown in Fig. 3B. Cultures containing the sox4 and soxC genes were induced at 39 and 420C, respectively, and soxB was induced under several conditions. The induced cultures were washed as described for the desulfurization assay and then mixed with one another and incubated overnight at 30'C with DBT as the test substrate. Thus, each of the Sox proteins was expressed in separate cells, and the intermediates produced by each were required to interact between cells. The SoxC culture by itself oxidized DBT TABLE 4. Activities of desulfurization gene products expressed from the A\PL promoter in E. coli sox gen gene

Substrate

Induction

potiol protocol

DBT02 detected

(AUC)Q

None (no cells)

None

20 ,uM DBTO2

No induction

54.4

MZ1/pSAD267-1

soxA soxA soxA soxA

20 20 20 20 20

39°C, 39°C, 42°C, 42°C,

No induction 15 min 60 min 15 min 60 min

14.2 4.9 4.1 29.1 32.7

No induction 39°C, 15 min 39°C, 60 min 42°C, 15 min 42°C, 60 min

12.6 44.5 77.2 115.9 165.6

soxA

MZ1/pSAD269-2A

soxC soxC soxC soxC soxC

p.M DBTO2

,uM DBTO2 ,uM DBTO2 p.M DBTO2

,uM DBTO2

0.1% 0.1% 0.1% 0.1% 0.1%

DBT DBT DBT DBT DBT

AUC, area under the curve (from HPLC analysis with whole cells). Values given are from the region of the curve matching the retention time of DBT02. a

No 2HBP was detected in these assays.

coli strain(s)

sox

s)

Induction protocol

AUC of products

detectedb DBTO2 2HBP

45,000

E. coli strain

6713

None soxA soxB soxC soxAC

No induction 390C, 15 min 39-420C, 15-60 min 420C, 30 mi As for soxA and soxC above For soxB, 390C, 15 min For soxB, 390C, 60 min For soxB, 42°C, 15 min For soxB, 42°C, 60 min For soxB, none

0 0 0 365.4 39.5

0

0 0 0 0

54.2 Mixed 1.5 soxABC soxABC 56.6 Mixed 2.1 Mixed soxABC 47.5 2.2 Mixed soxABC 35.7 2.3 30.9 3.7 Mixed soxABC a The substrate for all assays was 0.1% DBT. b AUC, area under the curve (from HPLC analysis of desulfurization assays with whole cells). Values given are from the regions of the curve matching the retention times of DBTO2 and 2HBP. C For assays involving mixed strains, clones of individual sox genes were induced in separate cells, mixed, and tested at 30°C in the desulfurization assay. For each mixed culture, MZ1/pSAD267-1 (soxA) was induced at 39°C for 15 min, MZ1/pSAD269-2A (soxC) was induced at 42°C for 30 min, and MZ1/ pSAD277-7A (soxB) was induced as indicated in the table.

to DBTO2. When the SoxA and SoxC cultures were mixed, the amount of DBTO2 detected decreased 10-fold, indicating that the SoxA protein was active. A small amount of 2HBP was produced when the three cultures were mixed together (Table 5). The activity of the SoxB protein (pSAD277-7A; Fig. 3B) was not affected by the time or temperature of soxB induction. Neither the SoxA nor the SoxB cultures alone produced 2HBP or any other detectable intermediate from DBT. In an attempt to improve the enzymatic activities of mixtures of individual Sox proteins, we lysed induced cultures of E. coli MZ1 containing pSAD267-1, pSAD277-7A, or pSAD269-2A. Cells were broken at 20,000 lb/in2 by passage through an Aminco French pressure cell. This treatment resulted in the total loss of detectable SoxA activity and greatly reduced SoxC activity (data not shown). Since the SoxA lysate lost all enzymatic activity, we could not determine if the SoxB lysate was active. SoxC is similar to members of the acyl-CoA dehydrogenase family. A search of the SwissProt protein sequence database using the FastA program of Pearson and Lipman (29) did not reveal any significant similarities for the Sox proteins. However, the BLOCKS search program identified a strong similarity between SoxC and several acyl coenzyme A (acyl-CoA) dehydrogenases (Table 6). acyl-CoA dehydrogenases contain six short, ungapped, highly conserved amino acid sequences (the blocks) (10). The SoxC protein contained sequences that matched the first four conserved blocks, and these were correctly spaced with respect to one another. The likelihood that such a combination would occur by chance was evaluated by the program as a P of 2HBP.

The SoxC protein mediates an initial sulfur oxidation reaction that produces DBTO2 from DBT. This reaction is observed when the cloned soxC gene is expressed in the absence of other sox genes either in Rhodococcus sp. strain UV1 or in whole cells or cell lysates of E. coli. The SoxA protein acts after SoxC in the desulfurization sequence, because DBTO2 is degraded by SoxA alone but not by SoxB. SoxA converts DBTO2 to an intermediate which was not identified by HPLC, but the DBT plate assay indicates that the compound is more water soluble

6715

than DBT. The fact that the HPLC analysis we employed could not detect the appearance of a new peak when SoxA was given DBT02 as a substrate, even though DBT02 was completely consumed, suggests that the intermediate produced by SoxA may be a very polar compound, one that is not distinguished from the other compounds found at the solvent front of the chromatogram. 2-HBP sulfonic acid and 2-HBP sulfinic acid have been identified as desulfurization intermediates of DBT metabolism by Rhodococcus sp. strain IGTS8 (3, 8), and either of these compounds may be the intermediate that SoxA produces from DBT02. The sulfonic and sulfinic acid intermediates may be involved in a second desulfurization pathway of DBT, one that does not produce 2HBP (28). Nevertheless, we have established that the SoxB protein completes the conversion of the unidentified intermediate to 2HBP. Establishment of the exact nature of the chemical intermediates will require efficient reconstitution of these enzyme reactions in vitro. To date, we have successfully overexpressed each of the Sox proteins in E. coli under control of the APL promoter, but most of the protein remains in an inactive form, presumably as inclusion bodies. We have been unable to determine a definitive role for DBTO in the pathway, since little or no DBTO is observed in the desulfurization reactions. In the absence of SoxC, only DBT02 was desulfurized to 2HBP by cells expressing the remaining two enzymes in the pathway, namely, SoxA and SoxB. Thus, DBTO does not accumulate as the product of a single enzyme, nor is it metabolized further if SoxC is absent. Experiments with DBT as a substrate revealed that DBTO can appear in cultures of bacteria that are desulfurization negative-for example, in the supernatants of wild-type E. coli (data not shown). There is some evidence that higher levels of DBTO may actually inhibit the metabolism of DBT02 to 2HBP in E. coli (2). This could, in part, explain why Rhodococcus sp. strain IGTS8 grows very poorly (8) or not at all (3) when supplied with DBTO as the sole source of sulfur. Prolonged storage of DBTO can result in the accumulation of DBT02 in the stock bottle, and DBT02 was found to be a contaminant in commercially available preparations of DBTO. These factors have made it difficult to determine an exact role for DBTO in the desulfurization pathway. Although our results, for reactions catalyzed by single gene products, suggest that DBTO is not a stable, individual enzymatic intermediate product of a unique enzyme in the pathway of desulfurization of DBT to 2HBP, we cannot rule out the possibility that it may be a transient step in the formation of DBT02 during the reaction catalyzed by SoxC. It is unexpected that the SoxC protein should be so closely related to the family of acyl-CoA dehydrogenases. The similarity of SoxC to this family is distributed throughout the first four-fifths of the protein and includes a large proportion of amino acids that are strictly conserved in the consensus sequence of these enzymes. Members of this family are oxidoreductases involved in the first or third steps of fatty acid P-oxidation in mitochondria (25, 38). The acyl-CoA dehydrogenases bind FAD (38) and, consistent with its relationship to this family, the activity of SoxC was increased by the addition of FAD to cell lysates. The eukaryotic acyl-CoA dehydrogenases transfer reducing equivalents from FAD to a flavoprotein and from there to the rest of the mitochondrial electron transport chain (38). If SoxC has a similar requirement for FAD, this could explain in part why the pellet and supernatant fractions of lysates containing SoxC were less active than were the original unfractionated lysates. Finally, the location of the desulfurization genes on an endogenous plasmid suggests that the pathway may also be

6716

DENOME ET AL.

distributed among other soil microorganisms. The fact that the sox genes can be expressed in E. coli argues that with the appropriate regulatory signals the enzymes could be active in a broad spectrum of genera beyond the rhodococci. ACKNOWLEDGMENTS This work was supported by Department of Energy contract DEAC22-89PC89901 and by Energy Biosystems Corporation. REFERENCES 1. Bos, P., and J. G. Kuenen. 1990. Microbial treatment of coal, p. 343-377. In H. L. Ehrlich and C. L. Brierly (ed.), Microbial mineral recovery. McGraw-Hill, New York. 2. Denome, S. A. Unpublished results. 3. Denome, S. A., E. S. Olson, and K. D. Young. 1993. Identification and cloning of genes involved in specific desulfurization of dibenzothiophene by Rhodococcus sp. strain IGTS8. Appl. Environ. Microbiol. 59:2837-2843. 4. Desomer, J., P. Dhaese, and M. van Montagu. 1990. Transformation of Rhodococcus fascians by high-voltage electroporation and development of R fascians cloning vectors. Appl. Environ. Microbiol. 56:2818-2825. 5. Dreyfuss, B., S. A. Adams, and Y. D. Choi. 1984. Physical change in cytoplasmic messenger ribonucleoproteins in cells treated with inhibitors of mRNA transcription. Mol. Cell. Biol. 4:415-423. 6. Feinberg, A. P., and B. Vogelstein. 1984. Addendum: a technique for radiolabelling DNA restriction endonuclease fragments to high specific activity. Anal. Biochem. 137:266-267. 7. Foght, J. M., P. M. Fedorak, M. R Gray, and D. W. S. Westlake.

8.

9. 10.

11.

12. 13.

14.

15.

16.

17.

1990. Microbial desulfurization of petroleum, p. 379-407. In H. L. Ehrlich and C. L. Brierly (ed.), Microbial mineral recovery. McGraw-Hill, New York. Gallagher, J. R., E. S. Olson, and D. C. Stanley. 1993. Microbial desulfurization of dibenzothiophene: a sulfur specific pathway. FEMS Microbiol. Lett. 107:31-36. Henikoff, S. 1984. Unidirectional digestion with exonuclease III creates targeted breakpoints for DNA sequencing. Gene 28:351359. Henikof, S., and J. G. Henikoff. 1991. Automated assembly of protein blocks for database searching. Nucleic Acids Res. 19:65656572. Holmes, D. S., and M. Quigley. 1981. A rapid boiling method for the preparation of bacterial plasmids. Anal. Biochem. 114:193197. Ikeda, Y., and K. Tanaka. 1983. Purification and characterization of isovaleryl CoenzymeA dehydrogenase from rat liver mitochondria. J. Biol. Chem. 258:1077-1805. Indo, Y., T. Yang-Feng, R. Glassberg, and K. Tanaka. 1992. Molecular cloning and nucleotide sequence of cDNAs encoding human long-chain acyl-CoA dehydrogenase and assignment of the location of its gene (ACADL) to chromosome 2. Genomics 12:626. Ish-Horowicz, D., and J. F. Burke. 1981. Rapid and efficient cosmid cloning. Nucleic Acids Res. 9:2989-2998. Kayser, K. J., B. A. Bielaga-Jones, K. Jackowski, 0. Odusan, J. J. Kilbane II. 1993. Utilization of organosulphur compounds by axenic and mixed cultures of Rhodococcus rhodochrous IGTS8. J. Gen. Microbiol. 139:3123-3129. Keen, N. T., S. Tamaki, D. Kobayashi, and D. Trollinger. 1988. Improved broad-host-range plasmids for DNA cloning in Gramnegative bacteria. Gene 70:191-197. Kelly, D. P., J.-J. Kim, J. J. Billadello, B. E. Hainline, T. W. Chu, and A. W. Strauss. 1987. Nucleotide sequence of medium-chain acyl-CoA dehydrogenase mRNA and its expression in enzymedeficient human tissue. Proc. Natl. Acad. Sci. USA 84:4068-4072.

J. BAc-rERIOL.

18. Kilbane, J. J. 1989. Desulfurization of coal: the microbial solution. Trends Biotechnol. 7:97-101. 19. Kilbane, J. J. 1990. Microbial removal of organic sulfur from coal: current status and research needs, p. 487-506. In D. L. Wise (ed.), Bioprocessing and biotreatment of coal. Marcel-Dekker, New York. 20. Kilbane, J. J. April 1992. Mutant microorganisms useful for cleavage of organic C-S bonds. U.S. patent 5,104,801. 21. Kilbane, J. J., I, and K. Jackowski. 1992. Biodesulfurization of water-soluble coal-derived material by Rhodococcus rhodochrous IGTS8. Biotechnol. Bioeng. 40:1107-1114. 22. Kim, J.-J. P., M. Wang, and R Paschke. 1993. Crystal structures of medium-chain acyl CoA dehydrogenase from pig liver mitochondria with and without substrate. Proc. Natl. Acad. Sci. USA 90: 7523-7527. 23. Kiyohara, H., K. Nagao, and K. Yana. 1982. Rapid screen for bacteria degrading water-insoluble, solid hydrocarbons on agar plates. Appl. Environ. Microbiol. 43:454-457. 24. Lipman, D. J., and W. R Pearson. 1985. Rapid and sensitive protein similarity searches. Science 227:1435-1441. 25. Matsubara, Y., Y. Indo, E. Naito, H. Ozasa, R Glassberg, J. Vockley, Y. Ikeda, J. Kraus, and K. Tanaka. 1989. Molecular cloning and nucleotide sequence of cDNAs encoding the precursors of rat long chain acyl-Coenzyme A, short chain acyl-Coenzyme A, and isovaleryl-Coenzyme A dehydrogenases. J. Biol. Chem. 264:16321-16331. 26. Matsubara, Y., M. Ito, R Glassberg, S. Satyabhama, Y. Ikeda, and K. Tanaka. 1990. Nucleotide sequence of messenger RNA encoding human isovaleryl-Coenzyme A dehydrogenase and its expression in isovaleric acidemia fibroblasts. J. Clin. Invest. 85:10581064. 27. Naito, E., H. Ozasa, Y. Ikeda, and K. Tanaka. 1989. Molecular cloning and nucleotide sequence of complementary DNAs encoding human short chain acyl-coenzyme A dehydrogenase and the study of the molecular basis of human short chain acyl-coenzyme A dehydrogenase deficiency. J. Clin. Invest. 83:1605-1613. 28. Oldfield, C. Unpublished results. 29. Pearson, W. R., and D. J. Lipman. 1988. Improved tools for biological sequence comparison. Proc. Natl. Acad. Sci. USA 85: 2444-2448. 30. Reddy, P., A. Peterofsky, and K. McKenney. 1989. Hyperexpression and purification of Escherichia coli adenylate cyclase using a vector designed for expression of lethal gene products. Nucleic Acids Res. 17:10473-10488. 31. Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecular cloning: a laboratory manual, 2nd ed. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y. 32. Sanger, F., S. Nicklen, and A. R Coulson. 1977. DNA sequencing with chain-terminating inhibitors. Proc. Natl. Acad. Sci. USA 74: 5463-5467. 33. Simon, R., U. Priefer, and A. Puhler. 1983. A broad host range mobilization system for in vivo genetic engineering: transposon mutagenesis in gram negative bacteria. Biotechnology 1:784-791. 34. Singer, M. E., and W. R Finnerty. 1988. Construction of an Escherichia coli-Rhodococcus shuttle vector and plasmid transformation in Rhodococcus spp. J. Bacteriol. 170:638-645. 35. Smith, T. F., and M. S. Waterman. 1981. Identification of common molecular subsequences. J. Mol. Biol. 147:195-197. 36. Southern, E. M. 1975. Detection of specific sequences among DNA fragments separated by gel electrophoresis. J. Mol. Biol. 98: 503-517. 37. Sturrock, S. S., and J. F. Collins. 1993. MPsrch version 1.3. Biocomputing Research Unit, University of Edinburgh, Edinburgh, United Kingdom. 38. Tanaka, K., and Y. Indo. 1992. Evolution of the acyl-CoA dehydrogenase/oxidase superfamily. Prog. Clin. Biol. Res. 375:95-110.