ldentification of a Maior Soluble Protein in Mitochondria from ... - NCBI

Plant Physiol. (1993) 102: 1171-1177

ldentification of a Maior Soluble Protein in Mitochondria from Nonphotosynthetic Tissues as NAD-Dependent Formate Dehydrogenase' Catherine Colas des Francs-Small*, Françoise Ambard-Bretteville, lan David Small, and Rene Remy Laboratoire de Cénétique Moléculaire des Plantes, Centre National de Ia Recherche Scientifique-Unité de Recherche Associée 1 128, Université Paris-Sud, Bâtiment 360, F-91405 Orsay Cedex, France (C.C.d.F.-S., F A - B . , R.R.); and Laboratoire de Biologie Cellulaire, lnstitut National de Ia Recherche Agronomique Route de Saint Cyr, F-78026 Versailles, France (I.D.S.)

externa1 NAD(P)H, and rotenone-insensitive oxidation of interna1 NADH (Douce and Neuburger, 1989). With regard to matrix enzymes, although the tricarboxylic acid cycle functions similarly in a11 eukaryotes, plant mitochondria differ from their animal counterparts in oxidation (NADmalic enzyme) and levels of Gly oxidation (Gly decarboxylase) (Douce and Neuburger, 1989). Gly is oxidized in the Of Gly decarboxylase 'pace by the 'Ombined and Ser hYdroxYmethY1transferase. This is an important ste?' in the process of photorespiration. The Gly cleavage system purified from plant mitochondria (Neuburger et al., 1986; Walker and Oliver, 1986) is composed of four subunits, P (94 kD), L (60 kD), T (41 kD), and H (15 kD). This complex is very abundant in green tissues but Scarce in nonphotosynthetic tissues, and it increases dramaticallywhen exposed to the light. The L subunit (lipoamide dehydrogenase) has been found in root and other etiolated tissues, where it is thought to be Part Of two Other enzYme comP1exesfthe PYmvate and a-ketoglutarate dehydrogenases (Turner et al., 1992). It is thus clear that the physiological role and enzyme complement of mitochondria differ considerably among different tissues, as in fact does mitochondrial morphology (Bendich and Gauriloff, 1984). To study these differences in more detail, the variation in the protein composition of mitochondria from different plant tissues has been examined in our laboratory. In previous papers, we described the analysis of two-dimensional protein patterns of mitochondria isohted from differentPea (RémY et a1.r 1987; HumPherY-Smith et a1.r 1992) and Potato (Solanum tuberosum L.1 (Colas des Francs-Small et al., 1992) organs. In addition to the abovementioned abundance of the four subunits of the Gly decarboxYlase ComPlex in green tissues, we described a 40-kD PolYPeptide that iS VerY abundant in mitochondria from tubers, dark-grown shoots, and calli or cell suspensions but scarce in leaf mitochondria (Colas des Francs-Small et al., 1992). This differential distribution between green and nongreen tissues was observed in all plant species tested but shows noticeable quantitative variations between species. It

In many plant species, one of the most abundant soluble proteins (as judged by two-dimensional polyacrylamide gel electrophoresis) in mitochondria from nongreen tissues i s a 40-kD polypeptide that is relatively scarce in mitochondria from photosynthetic tissues. cDNA sequences encoding this po\ypeptide were iso\ated irem a X g t l l cDNA expression library from potato (Solanum tuberosum 1.) by screening with a specific antibody raised against the 40-kD polypeptide. The cDNA sequence contains an open reading frame of 1137 nucleotides whose predicted amino acid sequence shows strong homology to an NAD-dependent formate dehydrogenase (EC 1.2.1.2) from Pseudomonas sp. 101. Comparison of the cDNA sequence with the N-terminal amino acid sequence of the mature 4 0 - W PolYPePtide sW?Wts that the PolYPePtide i s made as a precursor with a 23-amino acid presequence that shows characteristics typical of mitochondrial targeting signals. The identity of the polypeptide was confirmed by assaying the formate dehydrogenase activity in plant mitochondria from various tissues and by activity staining of mitochondrial proteins run on native gels combined with antibody recognition. The abundance and distribution of this protein suggest that higher plant mitochondria from various nonphotosynthetic plant tissues (tubers, storage roots, seeds, darkgrown shoots, cauliflower heads, and tissues grown in vitro) might contain a formate-producing fermentation pathway similar to those described in bacteria and algae.

Mitochondria are one of the major sites of energy conversion in eukaryotic cells, providing ATP and various substrates for biosynthetic reactions that occur in the cytoplasm. It is thus understandable that many basic features of structure and function have been conserved between animal and plant mitochondria (e,g. general structure, cyanide-sensitive electron pathway, ATP synthase complex), despite the fact that they diverged a billion years ago. Nevertheless, plant mitochondria display particular characteristics due to the autotrophic metabolism of plant cells. These include the cyanideinsensitive electron pathway (Moore and Siedow, 1991; Rhoads and McIntosh, 1991), respiration-linked oxidation of

' This work was supported by the Centre National de Ia Recherche Scientifique. * Corresponding author; fax 33-1-69417492.

Abbreviations: FDH, formate dehydrogenase; IPTG, isopropyl-PD-thiogalactopyranoside; pI, isoelectric point. 1171

Colas des Francs-Small et al.

1172

was found to be extremely abundant in potato tubers and dark-grown shoots, accounting for up to 9% of the total mitochondrial protein revealed by Coomassie blue staining of two-dimensional gels. The abundance of this protein suggests that it plays an important role in plant metabolism, and its distribution (the converse to that observed for Gly decarboxylase) suggests its participation in a metabolic pathway largely confined to nonphotosynthetic tissues. It is encoded by a nuclear gene because it is not synthesized by isolated mitochondria. To determine the identity and role of this 40-kD polypeptide, we raised a polyclonal antibody against it, which was used to screen a cDNA expression library. Here we describe the isolation and characterization of the cDNA encoding this polypeptide and its identification as an NAD-dependent FDH. Although FDH has already been reported from green leaf mitochondria (Halliwell, 1974; Oliver, 1981), its abundance in mitochondria from nonphotosynthetic tissues has not been noted, and we suggest that the importance of this enzyme has been previously underestimated. The implications of the presence of this enzyme in such large amounts in certain tissues are discussed.

Plant Physiol. Vol. 102, 1993

Computer analysis of the sequence was carried out lusing the University of Wisconsin GCG package of programs (Devereux et al., 1984). Comparison of the sequence with the GenBank, EMBL, PIR, and SwissProt sequence data bases was performed using the BLAST (Altschul et al., 1990') network service at the National Center of Biotechnology lnformation (Bethesda, MD). RNA lsolation and Northern Analysis

Total RNA was isolated from tubers, dark-grown shoots, calli, cell suspensions, and green leaves according to the procedure of Logemann et al. (1987). mRNAs were purified from the total RNAs using mAP paper from Amersham according to the manufacturer's instructions. The nucleic acids (about 10 p g per lane) were separated on 3% formaldehyde, 1.5% agarose gels, transferred by blotting onto nitrocellulose, and probed with the 5' half of the cDNA (i1 870bp fragment generated by a XhoI digest of the plasmid pBluescript containing the cDNA 1440) labeled by relndom priming in the presence of [a-32P]dCTP.

In Vitro Translation MATERIALS A N D METHODS Plant Material

Potato (Solanum tuberosum L.) tubers cv BF15 were either kept in the dark at 16OC or planted in a greenhouse and produced either dark-grown shoots or green leaves. Calli and cell suspensions were obtained as described previously (Colas des Francs-Small et al., 1992). Xgtll cDNA Library Screening

A Xgtll cDNA library prepared from tuber poly(A)+ RNA (cv Désirée) was screened using a rabbit polyclonal antibody prepared against the denatured 40-kD polypeptide (Colas des Francs-Small et al., 1992). Before use, the antiserum was incubated with an Escherichia coli protein extract, to avoid any cross-reaction. About 2 X 105 phages were plated and overlaid with nitrocellulose filters saturated with IPTG. The antigen was detected using the ProtoBlot immunoscreening system from Promega. A11 positive plaques were replated and rescreened four times, until a11 plaques yielded only positive signals. The sizes of the cDNA inserts were determined by digestion of purified A-DNA with EcoRI and subsequent agarose gel electrophoresis. Initially, a partia1 cDNA clone named cDNA 707 (707 bp, discounting the poly[A] tail) and, subsequently, a second larger clone named cDNA 1440 (1440 bp, discounting the poly[A] tail) were isolated and used for further experiments. D N A Sequencing and Sequence Analysis

The cDNAs of interest were subcloned into the plasmid Bluescript (Stratagene). Single-stranded plasmid DNA was prepared according to the method of Ausubel et al. (1987). The DNA was sequenced according to the dideoxy-nucleotide chain termination method (Sanger et al., 1977) using an Applied Biosystems apparatus as recommended by the manuf acturer.

Total or poly(A)+ potato RNAs were translated in a rabbit reticulocyte lysate system from Promega. The polypeptide of interest was immunoprecipitated from the in vitro translation mixture by the addition of the antiserum raised against the denatured 40-kD polypeptide and protein A-Sepharose as described by Ausubel et al. (1987). Polypeptides synthesized in the presence of [35S]Metwere fractionated by electrophoresis through 12% polyacrylamide gels in the presence of SDS (Laemmli, 1970) and autoradiographed. The same experiment was done using poly(A)+ RNA that had previously been selected by hybridization with the cDNA 707 (hybrid selection). Expression of Potato FDH in E. coli

The cDNAs to be expressed in E. coli were generated by polymerase chain reactions using either of two 27" oligonucleotides carrying an NcoI site (underlined) and the beginning of the cDNA sequence coding for either the mature protein (GTACCCATGGCTCTTCAGGCTTCCCCT) or the precursor (GTACCCATGGCTAGTCGTGTAGCTTCT), together with a second oligonucleotide complementary to flanking vector sequences. The reaction mixture (100 pL) contained 10 ng of template DNA, 1 FM each oligonucleotide, 0.2 mM each deoxyribonucleotide triphosphate, 10 IT~MTrisHC1 (pH 8.3), 50 mM KCl, 2.5 mM MgCI2, 100 Fg mL-' of gelatin, and 2.5 units of Taq polymerase (Perkin-Elrner Cetus). Forty cycles were performed on a Perkin-Elmer apparatus as follows: 1 min at 94OC for denaturation, 2 min at 60°C for annealing, and 1 min (plus 10 s for each cycle) at 72OC for primer extension. The polymerase chain reaction products were digested with NcoI and XbaI and subsequently ligated into the strictly IPTG-inducible expression vector pTrc99A (Amman et al., 1988). The resulting plasmids thus expressed polypeptides that either contained or did riot contain the presequence.

Mitochondrial Formate Dehydrogenase Synthesis of potato FDH was carried out using the E. coli strain x 53h (Tobey and Grant, 1986) transformed with the respective plasmids. After induction with IPTG according to the method of Ausubel et al. (1987), the proteins were separated by SDS-PAGE. The western transfer and subsequent antibody-staining reaction were as described by Colas des Francs-Small et al. (1992).

1173

GATGAGTCGTGTAGCTTCTACAGCAGCTCGTGCTATTACTTCACCTTCATCCTTAGTTTT M S R V A S T A A R A I T S P S S L V F f presequence

60

TACCAGAGAACTTCAGCCTTCCCCTGGCCCCAMAAGATTGTGGGTGTTTTCTACAAAGC T R E L Q A S P G P K K I V G V F Y K A N - t e r m i n a l sequence

120

--

-

AMTGAATATGCTGMATGAACCCCAACTTTTTGGGCTGTGCAGAMATGCCTTGGGGAT 180 N E Y A E M N P N F L G C A E N A L G I

+

ACGTGAATGGTTAGMTCTAMGGTCATCAGTACATAGTCACTCCTGACAAAGAAGGACC 240 R E W L E S K G H Q Y I V T P D K E G P

Enzyme Assays

FDH activity was measured as the increase in A340according to the method of Quayle (1966), using 0.025% Triton X100-treated tuber or leaf mitochondria or only the soluble mitochondrial proteins that were extracted as previously described (Colas des Francs-Small et al., 1992). Electrophoresis

Nondenaturing electrophoresis was performed on 7% acrylamide gels in the discontinuous buffer system of Laemmli (1970) without SDS. Soluble mitochondrial proteins (20 Wg) were loaded on each track, and the gels were run at 100 V for 2 h at 4OC in a Mini Protean I1 Bio-Rad apparatus. Subsequent FDH activity staining in the presence of nitroblue tetrazolium, phenazine methosulfate, formate, and NAD was performed according to the method of Uotila and Koivusalo (1979). For the experiment described in Figure 5 , after nondenaturing electrophoresis and Coomassie blue staining, a second dimension SDS-PAGE was performed, followed by transfer onto nitrocellulose membranes and immunoblotting according to the method of Towbin et al. (1979). Molecular mass estimations were done using the Pharmacia electrophoresis calibration kit for low molecular mass proteins. Quantitative estimations of the bands were performed by computerized image analysis using a MasterScan I (Computer Signal Processing, Inc., Billerica, MA).

RESULTS lsolation and Sequencing of cDNA Clones Reacting with the Antibody

Immunoscreening of the Xgtll cDNA library initially gave one positive clone, which was cloned into the plasmid Bluescript (to give cDNA 707) and fully sequenced. Subsequently, a longer clone was isolated, subcloned (to give cDNA 1440), and almost fully sequenced, to reconstitute as far as possible the complete cDNA sequence. No differences were observed in the approximately 300 bp sequenced from both clones; therefore, both cDNAs probably originate from the same gene. However, the two clones differ in their site of addition of the poly(A) tail; cDNA 1440 contains an extra 200 bp at the 3' end. Based on the size of the mRNA on northern blots (about 1.4 kb), we suspect that the poly(A) site in cDNA 707 is the major site used and that, in fact, although cDNA 1440 is at least as long as the major mRNA, it lacks 100 to 150 nucleotides at the 5' end. The combined cDNA sequence and the deduced amino acid sequence of the single, long, open reading frame are shown in Figure 1. The predicted amino acid sequence includes the 29 N-terminal amino acids previously identified

AGATTGTGAGCTTGAGAMCACATTCCTGATCTCCATGTGTTGATTTCGACCCCTTTTCA D C E L E K H I P D L H V L I S T P F H

300

TCCTGCCTATGTCACGGCTGMAGGATCAAGAAGGCMAMATCTACAACTACTGCTGAC P A Y V T A E R I K K A K N L Q L L L T

360

AGCTGGAATTGGATCAGATCATG~GATCTGAMGCTGCTGCTGCTGCTGGATTGACAGT 4 2 0 A G I G S D H V D L K A A A A A G L T V GGCAGAAGTCACTGGAAGCAATACTGTCTCAGTTGCGGAAGATGAACTGATGCGMTC~4 8 0 A E V T G S N T V S V A E D E L M R I L CDNA 7 0 7 s t a r t

r,

AATTCTTGTCCGAAATTTCTTGCCTGGACATCATCAGGTTATCAATGGGGAGTGGAATGT I L V R N F L P G H H Q V I N G E W N V

540

TGCTGCAATTGCGCACAGAGCTTATGATTTGGMGGCAAGACTGTTGGMCTGTTGGTGC A A I A H R A Y D L E G K T V G T V G A

600

AGGACGTATTGGCAGACTTTTACTCCMCGATTGAAGCCATTTAACTGTAATCTACTTTA 660 G R I G R L L L Q R L K P F N C N L L Y

CCATGACCGTCTTAAGATGGATTCTGAGCTAGAGAATCAMTTGGAGCMAGTTCGAGGA H D R L K M D S E L E N Q I G A K F E E

720

BglII

AGATCTCGATMAATGCTCTCGAMTGTGACATAGTGGTCATCAATACGCCTCTTACAGA D

L

D

K

M

L

S

K

C

D

I

V

V

I

N

T

P

L

T

780

E

GMAACAMAGGAATGTTTGACAAGGMAGAATTGCMAGTTGAAGAAGGGAGTTCTAAT 8 4 0 K T K G M F D K E R I A K L K K G V L I

XhoI TGTAMCMTGC-GAGCAATTATGGACACTCAAGCAGTTGTTGATGCTTGTAACAG V N N A R G A I M D T Q A V V D A C N

900 S

CGGTCATATTGCAGGATACAGTGGAGATGTTTGGTACCCACAACCAGCTCCCAAGGACCA 960 G H I A G Y S G D V W Y P Q P A P K D H TCCGTGGCGCTACATGCCAAACCMGCTATGACTCCTCACATTTCTGGGACTACCATTGA P W R Y M P N Q A M T P H I S G T T I D

1033

TGCACAG~GAGGTATGCAGCTGGAACAAAGGATATGTTGGACAGGTACTTCAAGGGTGA 1080 A Q L R Y A A G T K D M L D R Y F K G E GGATTTCCCTGCTGMAATTACATTGTCAAGGATGGGGAATTGGCACCTCAGTATCGCTA 1140 D F P A E N Y I V K D G E L A P Q Y R AATCATTMTGTACCTTAGTTGTGCCTAGATTTCTGTGTCAAATAMCTGGCAAGCTGCG 1200

TCTTCAATAAGCTGATAMATTGGCTACTTATGTTCTTTTAGM

1245

Figure 1. Nucleotide sequence of the cDNA encoding potato FDH

together with the deduced amino acid sequence of the open reading frame. The 23-amino acid presequence and the 29-amino acid N-terminal sequence of the mature protein (Colas des FrancsSmall et al., 1992)are indicated by arrows. The start of cDNA 707 is indicated; cDNA 1440 was sequenced u p to the Bglll site at nucleotide 721. The Xhol (used to generate the 870-bp probe for northern blots) and BgllI sites are underlined.

by protein sequencing (Colas des Francs-Small et al., 1992), confirming the correspondence between the cDNA sequence and the 40-kD polypeptide. The size of the protein expected from the 1137-nucleotide open reading frame is 379 amino acids (predicted molecular mass 41.9 kD and pl 7.08), i.e. significantly larger than the polypeptide seen on two-dimen-


1174

Mr

A

B

C

D

x10'3

43 _ 40-

'1

•m

30-

20-

—

Figure 2. SDS-PACE analysis of translation products. Total potato RNAs were translated in a rabbit reticulocyte lysate system, and the translation product of interest was immunoprecipitated using the antibody directed against the 40-kD polypeptide. Lane A, Coomassie blue-stained tuber mitochondrial proteins; lane B, corresponding immunoblot showing the 40-kD polypeptide; lane C, fluorograph of an SDS-gel of the immunoprecipitate showing the 43-kD labeled polypeptide (the extra band that can be seen above the 43-kD product corresponds to rabbit immunoglobulin C band slightly labeled by the unremoved 43-kD product); lane D, fluorograph of the supernatant left after immunoprecipitation. Lanes A and B are from one gel, and lanes C and D are from another gel, which was autoradiographed. Molecular mass markers (Pharmacia kit) are indicated on the left; 40 kD is an estimate.

sional gels. However, there are 23 amino acids before the Nterminal sequence of the mature polypeptide, suggesting that it is synthesized as a longer precursor that is subsequently cleaved. This was confirmed by immunoprecipitation (with the antibody against the 40-kD polypeptide) of in vitro translation products of potato callus RNA (Fig. 2) and by the hybrid selection experiment (not shown), which revealed a single product of 43 kD (as estimated by SDS-PAGE). This 43-kD product was cleaved to give a 40-kD polypeptide in the presence of isolated mitochondria (not shown). The predicted molecular mass of the mature polypeptide as deduced from the cDNA sequence is 39.44 kD, and the predicted pi is 6.74, both values corresponding closely to those observed on two-dimensional gels (Colas des Francs-Small et al., 1992). As discussed above, it is likely that cDNA 1440 lacks sequence at the 5' end, and we cannot be certain that the first ATC in the sequence is the true initiation codon; however, the 23-amino acid presequence has all the features expected of a mitochondrial targeting sequence, i.e. it is rich in hydroxylated and positively charged amino acids, and secondary structure predictions for the first 14 amino acids suggest that it should form an a-helix. A helical wheel plot of this region suggests that the helix that would be formed would be highly amphiphilic (not shown). Such amphiphilicity has been suggested to be necessary for efficient mitochondrial targeting. Amino Acid Sequence Comparisons

The amino acid sequence derived from the open reading frame was compared with sequences in the GenBank/EMBL,


PIR, and SwissProt data bases. The comparisons revealed similarity between the derived amino acid sequence from cDNA 1440 and several NAD-specific, D-isomer 2-hydroxy acid dehydrogenases, but the most significant homology is with Pseudomonas sp. 101 FDH (Popov et al., 1990) (51% identity in a 325-amino acid overlap, i.e. from amino acids 51-376). The alignment of the two sequences can be seen in Figure 3, where the sequences corresponding to the NADbinding domain are boxed. Confirmation of the 40-kD Polypeptide as FDH

The sequence alignments strongly suggested that the 40kD polypeptide was an NAD-dependent FDH, and therefore, a number of enzyme activity assays were run on soluble mitochondrial proteins from potato tubers or green leaves. Initially, samples were run on nondenaturing gels (Fig. 4), and then one-half of each track was stained for FDH activity, and the other half was run on a second-dimension SDS gel (Fig. 5). Figure 4 shows the FDH activity revealed on nondenaturing gels for green leaf- and tuber-soluble mitochondrial proteins. The intensity of band revealed on the tuber mitochondria gel is about 5-fold greater than that observed for leaf mitochondrial proteins (Fig. 4). This result was confirmed by spectrophotometric enzyme assays performed on tuber and leaf mitochondria. The FDH activities found were (an average of three separate experiments) 0.27 ^mol min"1

Potato FDH Pseudomonas FDH

LQASPGPKKIVGVFYKANEYAEMN 47 :: II...:... II ::.. IDHYPGGQTLPTP..KAIDFTP.. 46

48 PNFLGCAENALGIREWLESKGHQYIVTPDKEGPDCELEKHIPDLHVLIST 97 . . : I I : . . . . 1 1 : I . : I I 1 . 11 . : I I . I I . I I I : : ! : . : . ! . 1 : 1 1 47 GQLLGSVSGELGLRKYLESNGHTLVVTSDKSGPDSVFERELVDADVVISQ % 98 PFHPAYVTAERIKKAKNLQLLLTAGIGSDHVDLKAAAAAGLTVAEVTGSN 147 II 1 1 1 : 1 : 1 1 1 I I I I I . I I I I I I I I I I I I I.. I . .:. . I I I I I 97 PFWPAYLTPERIAKAKNLKLALTAGIGSDHVDLQSAIDRNVVTAEVTYVN 146 148 TVSVAEDELMRILILVRNFLPGHHQVINGEWNVAAIAHRAYDLEGKTVGT 197 . : ! I I I . :l .1 I 1 1 1 1 : 1 1 : 1 . . . 1 : 1 1 : 1 . . . :lI I I I:. Ill 147 SISVAEHVVMMILSLVRNYLPSHEWARKGGWNIADCVSHAYDLEAMHVGT 196

198 V IAGRIGRLLLQRLKPFNCNLLYHD iLKMDSELENQIGAKFEEDLDKMLS 247 1 : 1 1 1 1 1 : 1 . 1 1 1 1 : III 1 ::...:!.::. .: . . . :. 1.. 197 V IAGRIGLAVLRRLAPFDHVLHYTD (HRLPESVEKELNLTWHATREDMYP 246 248 KCDIVVINTPLTEKTKGMFDKERIAKLKKGVLIVNNARGAIMDTQAVVDA 297 I I:I.:I.II ..I. I::.I : :I:I..III.III : I :II. I 247 VCDVVTLNCPLHPETEHMINDETLKLFKRGAYIVNTARGKLCDRDAVARA 296 298 CNSGHIAGYSGDVWYPQPAPKDHPWRYMPNQAMTPHISGTTIDAQLRYAA 347 :l I : :l I I . I I I I:I I I I I I I I I I I II :: I I I I I I I I I : . I I I II I 297 LESGRLAGYAGDVWFPQPAPKDHPWRTMPYDGMTPHISGTTLTAQARYAA 346 348 GTKDMLDRYFKGEDFPAENYIVKDGELAP 376 I I : : : ! : : l . I . : . . ! . II . : I . I I . 347 GTREILECFFEGRPIRDEYLIVQGGALAG 375

Figure 3. Alignment of potato FDH and Pseudomonas FDH amino acid sequences. Caps inserted for optimal alignment are represented by dots in the sequence. Identical amino acid residue matches between the two sequences are connected by solid vertical lines. Conservative and semiconservative changes are represented by double vertical dots and single dots, respectively. The sequences corresponding to the NAD-binding domain are boxed.

Mitochondrial Formate Dehydrogenase

1175

43kO. -40kD

Figure 4. Nondenaturing acrylamide gel electrophoresis of potato soluble mitochondrial proteins. Coomassie blue-stained gels of green leaf proteins (lane A) and tuber proteins (lane B). Lanes C and D show similar gels stained for FDH activity. Each lane was loaded with 20 Mg of soluble mitochondrial proteins.

B ,,.,»

A. Mr ,10'

-94 -67 -43

-30

-20

Figure 6. Immunoblot of the polypeptide expressed in E. coli. Part a of the gel shows the kinetics of the IPTC induction of the expression of the 43-kD precursor protein (plasmid containing the putative full-length cDNA), and part b shows that of the 40-kD mature protein (plasmid containing the cDNA encoding the mature protein). The samples were taken after 0 min (lane 0), 10 min (1), 20 min (2), and 60 min (3) of induction for both experiments. About 5 Mg protein were loaded on part a and 50 Mg on part b of the gel such that the induced mature protein could be detected. Lane M shows an immunoblot of tuber mitochondrial proteins. These experiments were done using the polyclonal antibody raised against the 40-kD polypeptide. An unidentified 70-kD E. coli polypeptide slightly cross-reacts with the antibody but is clearly not IPTG induced. The sizes indicated are estimated from molecular mass markers.

mg ' of soluble protein for tuber mitochondria and 0.035 jjmol min"1 mg"1 of soluble protein for leaf mitochondria. No FDH activity was found outside mitochondria. Figure 5 shows that the single band that had FDH activity ran as a 40-kD polypeptide on an SDS gel and was recognized by our antibody after western blotting. This experiment was performed because this antibody (raised against the denatured protein) does not recognize the native protein very efficiently. To further confirm the identity of the 40-kD polypeptide, cDNAs corresponding to the precursor protein or to the mature protein were expressed in £. coli using an IPTGinducible expression vector. Both cDNAs gave rise to polypeptides of the expected sizes (43 or 40 kD, respectively, on the SDS-gels that were specifically recognized by our anti-

Lo

L«

Li«

Lao

C

S

T

I Figure 5. Identification of the major 40-kD polypeptide of soluble mitochondrial proteins from potato tubers as FDH. (FDH activity co-migrates on nondenaturing gels with the 40-kD polypeptide recognized by the antibody.) A nondenaturing gel of soluble mitochondrial proteins (A) similar to that presented in Figure 4B was subjected to a second-dimensional separation by SDS-PACE (C). On the left side is a one-dimensional gel of the same denatured sample (B). The position of the major 40-kD polypeptide is indicated by an arrow. On the right side a bracket shows the part of the gel represented in D as an immunoblot. The major polypeptide corresponding to the FDH activity shown in Figure 4 clearly is specifically recognized by the 40-kD polypeptide polyclonal antibody. The sizes of molecular mass markers are indicated on the right.

Figure 7. Northern blot of total potato RNA hybridized with a labeled 870-bp Xriol fragment of the cDNA 1440. Lanes LO to L20 show the accumulation of the transcript in the leaf after 0 h (LO), 8 h(L8), 16 h (L16), and 20 h (L20) of darkness. The other lanes show hybridization to callus RNA (C), dark-grown shoot RNA (S), and tuber RNA (T). About 10 Mg of total RNA (as estimated by absorbance measurements and ethidium bromide staining of the gel) were loaded per lane.


1176

body, but the precursor form was much more highly expressed than the mature form, which was barely detectable (Fig. 6). Regulation of Expression of FDH in Plant Tissues

Figure 7 shows a northern blot hybridized with a 870-bp XhoI fragment of the cDNA 1440. The size of the transcript was estimated to be approximately 1.4 kb using brome mosaic virus RNA and rRNAs as markers. The difference in transcript abundance in various organs is striking: it is abundant in dark-grown shoots and calli (and cell suspensions), but scarce or undetectable in tubers. In leaves, the transcript level is low and does not seem to vary during the day but starts increasing by 8 h of darkness and has increased dramatically after 16 h in the dark. DlSCUSSlON

The sequence comparisons, activity tests, and antibody recognition data presented in this paper strongly suggest that the cDNA we have sequenced encodes a precursor for a mitochondrial NAD-dependent FDH corresponding to the abundant 40-kD polypeptide previously observed on twodimensional gels. FDH is a dimeric enzyme that catalyzes the oxidation of formate to COz in the presence of NAD+. Although the reverse reaction is theoretically possible, the thermodynamic equilibrium is strongly in favor of C 0 2 production, and attempts to demonstrate the reverse reaction have been unsuccessful (Mathews and Vennesland, 1950; Davison, 1951). The enzyme has been found previously in plants (Davison, 1951), animals (Mathews and Vennesland, 1950), and microorganisms (Quayle, 1966; Kato et al., 1974). As opposed to the FDH found in E. coli and severa1other bacteria (Ferry, 1990), which is Cyt b,-linked and membrane bound, the plant, yeast, and Pseudomonas enzymes are soluble. FDH has been purified from the yeast species Kloeckera (Kato et al., 1974) and Candida boidinii (Schütte et al., 1976) and from pea (Nason and Little, 1955; Uotila and Koivusalo, 1979). The molecular mass and pI for the potato enzyme are close to those found for the pea (42.4 kD for the monomer and a pI of 6.2), C. boidinii (36 kD) (Schütte et al., 1976), and Caizdida niethanolica (43 kD) (Izumi et al., 1989) enzymes. In methanol-using microorganisms, FDH is cytosolic and catalyzes the last step of methanol oxidation, the true substrate being the S-formylglutathione produced by formaldehyde dehydrogenase (Van Dijken et al., 1976). Plants do contain a cytosolic formaldehyde dehydrogenase, which has been purified from pea (Uotila and Koivusalo, 1979), but the lower K, of the plant FDH for formate (about 2 mM, as opposed to 13 mM [Schütte et al., 19761 or 22 mM [Kato et al., 19741 for yeast FDHs) suggests that the plant enzyme may use formate directly. In higher plants, in which FDH has previously been shown to be mitochondrial (Halliwell, 1974; Oliver, 1981), its function is not clear. Formate can be oxidized to COz by FDH or by the peroxidative action of catalase but can also be reduced to hydroxymethyl tetrahydrofolate and participate in Ser synthesis by the Ser hydroxymethyl transferase reaction. What remains obscure is the source of this formate (Cossins, 1980). In leaves, it is


thought to arise from the glyoxylate produced by photorespiration, and FDH has been shown to be active in spinach and tobacco leaves (Oliver, 1981). However, our experinients show that the protein is much more abundant in nongreen tissues (i.e. potato tubers, dark-grown shoots, cell suspensions, cauliflower heads, and storage roots such as sweet potato and carrot [Colas des Francs-Small et al., 19921) than in photosynthetic tissues. The activity we found in tuber mitochondria is 5 - to 8-fold greater than the activity ir1 leaf mitochondria. High activities have also been observcrd in mature seeds (Davison, 1949). At least part of the tissue-specific distribution of FDH is regulated by mRNA levels. Northern blot experiments show that the transcript is abundant in dark-grown shoots, calli, and cell suspensions, which are actively dividing tissues. In dormant tubers, the protein is abundant but the transuipt is scarce, suggesting that protein turnover is low inside the mitochondrion. In leaves, the transcript level seem5 to t ' e low and constant during the day but accumulates after 16 h in the dark. The slow response suggests a control by changing metabolite levels rather than directly by light levels.. This type of regulation is interesting, and the study of the FDH gene promoter is about to start in our laboratory. In conclusion, the abundance and tissue distribution of FDH suggest that a major uncharacterized metabolic pathway exists in mitochondria from various nonphotosynthetic plant tissues that produces large quantities of formate. In these tissues, the origin of the formate is likely to be different from that in leaves, and it is apparently produced in much higher amounts. In general, high FDH activity seems to be found in nonphotosynthetic tissues that have low surface area to volume ratios (tubers, storage roots, etc.) or that are growing in confined conditions, e.g. in vitro cultures, developing seeds. This suggests to us a correlation with anaerobic metabolism. Formate has long been known to be produced by potato tubers and by roots under anaerobic stress (Davison, 1951), although in recent years much more emphasis has been placed on ethanol production. Formate is also produced in large quantities during anaerobic fermentation in bacteria (Ferry, 1990) and unicellular algae (Kreuzberg, 1984). Ir1 these organisms, formate is produced by thioclastic cleavage of pyruvate by the enzyme pyruvate-formate lyase, which is activated under strictly anaerobic conditions by a specific "activase."The enzymes of the formate fermentation pathway found in Chlnmydomonas are mostly mitochondrial (.Kreuzberg et al., 1987). The possible occurrence of such a pathway in higher plants should stimulate interest in the clistinct physiological roles of mitochondria in different tissues,. ACKNOWLEDCMENTS

The authors wish to thank Dr. M. Bevan (Cambridge Laboratory, Norwich, UK) for providing the potato tuber cDNA library ,md Dr. G.A. Grant (Washington University School of Medicine, St. Louis, MO) for the E. coli strain we used for the expression. We are grateful

to A. Lepingle and D. Lancelin for help with the sequencing, to A. Darpas for the tissue culture, and to D. Froger for the photographs. We thank F. Vede1 for his continuous interest in this work. Received January 26, 1993; accepted April 15, 1993. Copyright Clearance Center: 0032-0889/93/102/1171/07.

Mitochondrial Formate Dehydrogenase The GenBank/EMBL accession number for the sequence reported in this article is 221493.

LITERATURE ClTED

Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ (1990) Basic local alignment search tool. J Mo1 Biol 215: 403-410 Amman E, Ochs B, Abel KJ (1988) Tightly regulated tac promoter vectors useful for the expression of unfused and fused proteins in Escherichia coli. Gene 6 9 301-315 Ausubel FM, Brent R, Kingston RE, Moore DD, Seidmann JG, Smith JA, Struhl K (1987) Current Protocols in Molecular Biology. Wiley-Interscience, New York Bendich AJ, Gauriloff LP (1984) Morphometric analysis of cucurbit mitochondria: the relationship between chondriome volume and DNA content. Protoplasma 1 1 9 1-7 Colas des Francs-Small C, Ambard-Bretteville F, Darpas A, Sallantin M, Huet JC, Pernollet JC, Remy R (1992) Variation of the polypeptide composition of mitochondria isolated from different potato tissues. Plant Physiol 98: 273-278 Cossins EA (1980) One-carbon metabolism. In DD Davies, ed, The Biochemistry of Plants, Vol 11. Academic Press, New York, pp 365-418 Davison DC (1949) The distribution of formic and alcohol dehydrogenases in the higher plants, with particular reference to their variation in the pea plant during its life cycle. Proc Linn Soc NSW 74: 26-36 Davison DC (1951) Studies on plant formic dehydrogenase. Biochem J 4 9 520-526 Devereux J, Haeberli P, Smithies O (1984) A comprehensive set of sequence analysis programs for the VAX. Nucleic Acids Res 12: 387-395 Douce R, Neuburger M (1989) The uniqueness of plant mitochondria. Annu Rev Plant Physiol Plant Mo1 Biol 40: 371-414 Ferry JG (1990) Formate dehydrogenase. FEMS Microbiol Rev 87: 377-382 Halliwell B (1974) Oxidation of formate by peroxisomes and mitochondria from spinach leaves. Biochem J 138: 77-85 Humphery-Smith I, Colas des Francs-Small C, Ambard-Bretteville F, Rémy R (1992) Tissue-specific variations of pea mitochondrial polypeptides detected by computerized image analysis of two-dimensional electrophoresis gels. Electrophoresis 13: 168-1 72 Izumi Y, Kanzaki H, Morita S, Futazuka H, Yamada H (1989) Characterization of crystalline formate dehydrogenase from Candida methanolica. Eur J Biochem 1 8 2 333-341 Kato N, Kano M, Tani Y, Ogata K (1974) Purification and characterisation of formate dehydrogenase in a methanol-utilizing yeast, Kloeckera sp. No. 2201. Agric Biol Chem 38: 111-1 16 Kreuzberg K (1984) Starch fermentation via formate producing pathway in Chlamydomonas reinhardii, Chlorogonium elongatum, and Chlorella fusca. Physiol Plant 61: 87-94 Kreuzberg K, Klock G, Grobheiser D (1987) Subcellular distribution of pyruvate-degrading enzymes in Chlamydomonas reinhardii studied by an improved protoplast fractionation procedure. Physiol Plant 6 9 481-488 Laemmli UK (1970) Cleavage of structural proteins during the

1177

assembly of the head of bacteriophage T4. Nature (Lond.) 227: 68O -685 Logemann J, Schell J, Willmitzer L (1987) Improved method for the isolation of RNA from plant tissues. Ana1 Biochem 163: 16-20 Mathews MB, Vennesland B (1950) Enzymic oxidation of formic acid. J Biol Chem 186: 667-682 Moore AL, Siedow JN (1991) The regulation and nature of the cyanide-resistant alternative oxidase of plant mitochondria. Biochim Biophys Acta 1059 121-140 Nason A, Little HN (1955) Formic dehydrogenase from peas. Methods Enzymol 1: 536-539 Neuburger M, Bourguignon J, Douce R (1986) Isolation of a large complex from the matrix of pea leaf mitochondria involved in the rapid transformation of glycine into serine. FEBS Lett 207: 18-22 Oliver DJ (1981) Formate oxidation and oxygen reduction by leaf mitochondria. Plant Physiol 68: 703-705 Popov VO, Shumilin IA, Ustinnikova TB, Lamzin VS, Egorov TA (1990) NAD-dependant formate dehydrogenase from methylotrophic bacterium Pseudomonas sp, 101. I. Amino acid sequence. Bioorg Khim 16: 324-335 Quayle JR (1966) Formate dehydrogenase. Methods Enzymol 9: 360-364 Remy R, Ambard-Bretteville F, Colas des Francs C (1987) Analysis by two-dimensional gel electrophoresis of the polypeptide composition of pea mitochondria isolated from different tissues. Electrophoresis 8: 528-532 Rhoads DM, McIntosh L (1991) Isolation and characterization of a cDNA clone encoding an alternative oxidase protein of Sauromatum guttatum (Schott). Proc Natl Acad Sci USA 88: 2122-2126 Sanger F, Nicklen S, Coulson AR (1977) DNA sequencing with chain-terminating inhibitors. Proc Natl Acad Sci USA 7 4 5463-5467 Schiitte H, Flossdorf J, Sahm H, Kula MR (1976) Purification and properties of formaldehyde dehydrogenase and formate dehydrogenase from Candida boidinii. Eur J Biochem 62: 151-160 Tobey KL, Grant GA (1986) The nudeotide sequence of the serA gene of Escherichia coli and the amino acid sequence of the encoded protein, D-3-phosphoglycerate dehydrogenase. J Biol Chem 261: 12179-12183 Towbin H, Steahelin T, Gordon J (1979) Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some applications. Proc Natl Acad Sci USA 76: 4350-4354 Turner SR, Ireland R, Rawthorne S (1992) Purification and primary amino acid sequence of the L subunit of glycine decarboxylase. Evidence for a single lipoamide dehydrogenase in plant mitochondria. J Biol Chem 267: 7745-7750 Uotila L, Koivusalo M (1979) Purification of formaldehyde and formate dehydrogenases from pea seeds by affinity chromatography and S-formylglutathione as the intermediate of formaldehyde metabolism. Arch Biochem Biophys 196: 33-45 Van Dijken JP, Oostra-Demkes GJ, Otto R, Harder W (1976) Sformylglutathione: the substrate for formate dehydrogenase in methanol-utilizing yeasts. Arch Microbiol 111: 77-83 Walker JL, Oliver DJ (1986) Glycine decarboxylase multienzyme complex. Purification and partia1 characterization from pea leaf mitochondria. J Biol Chem 261: 2214-2221