Production of the Gram-positive Sarcina ventriculi pyruvate decarboxylase in Escherichia coli

Lee A. Talarico1, Lonnie O. Ingram1 and Julie A. Maupin-Furlow1

Department of Microbiology and Cell Science, University of Florida, Gainesville, FL 32611-0700, USA1

Author for correspondence: Julie A. Maupin-Furlow. Tel: +1 352 392 4095. Fax: +1 352 392 5922. e-mail: jmaupin{at}ufl.edu


   ABSTRACT
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS AND DISCUSSION
REFERENCES
 
Sarcina ventriculi grows in a remarkable range of mesophilic environments from pH 2 to pH 10. During growth in acidic environments, where acetate is toxic, expression of pyruvate decarboxylase (PDC) serves to direct the flow of pyruvate into ethanol during fermentation. PDC is rare in bacteria and absent in animals, although it is widely distributed in the plant kingdom. The pdc gene from S. ventriculi is the first to be cloned and characterized from a Gram-positive bacterium. In Escherichia coli, the recombinant pdc gene from S. ventriculi was poorly expressed due to differences in codon usage that are typical of low-G+C organisms. Expression was improved by the addition of supplemental codon genes and this facilitated the 136-fold purification of the recombinant enzyme as a homo-tetramer of 58 kDa subunits. Unlike Zymomonas mobilis PDC, which exhibits Michaelis–Menten kinetics, S. ventriculi PDC is activated by pyruvate and exhibits sigmoidal kinetics similar to fungal and higher plant PDCs. Amino acid residues involved in the allosteric site for pyruvate in fungal PDCs were conserved in S. ventriculi PDC, consistent with a conservation of mechanism. Cluster analysis of deduced amino acid sequences confirmed that S. ventriculi PDC is quite distant from Z. mobilis PDC and plant PDCs. S. ventriculi PDC appears to have diverged very early from a common ancestor which included most fungal PDCs and eubacterial indole-3-pyruvate decarboxylases. These results suggest that the S. ventriculi pdc gene is quite ancient in origin, in contrast to the Z. mobilis pdc, which may have originated by horizontal transfer from higher plants.

Keywords: pyruvate metabolism, thiamin pyrophosphate-dependent enzymes, alcohol dehydrogenase, ethanol fermentation, allosteric regulation

Abbreviations: ADH, alcohol dehydrogenase; E1, decarboxylase component of pyruvate dehydrogenase; IPD, indole-3-pyruvate decarboxylase; PDC, pyruvate decarboxylase; TPP, thiamin pyrophosphate; TK, transketolase.

The GenBank accession number for the sequence reported in this paper is AF354297.


   INTRODUCTION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS AND DISCUSSION
REFERENCES
 
Metabolic engineering of bacteria for high-level ethanol production incorporates the use of genes encoding pyruvate decarboxylase (PDC) and alcohol dehydrogenase (ADH) to channel the flow of pyruvate to ethanol (Ingram et al., 1999 ). There has been success in engineering Gram-negative bacteria for ethanol production using the pdc and adh genes from Zymomonas mobilis. However, similar approaches in engineering Gram-positive bacteria have been limited, possibly due to differences in transcription, translation or protein degradation between Gram-negative and Gram-positive bacteria (Barbosa & Ingram, 1994 ; Gold et al., 1996 ). Thus, the isolation of pdc operons from Gram-positive bacteria, such as Sarcina ventriculi, is likely to broaden the diversity of bacteria which can be modified for high-level ethanol production from biomass under conditions such as low pH, high temperature and high salt.

PDC (EC 4 . 1 . 1 . 1) serves as the key enzyme in all homo-ethanol fermentations. This enzyme catalyses the non-oxidative decarboxylation of pyruvate to acetaldehyde and carbon dioxide using Mg2+ and thiamin pyrophosphate (TPP) as cofactors. Acetaldehyde is subsequently reduced to ethanol by alcohol dehydrogenase (ADH, EC 1 . 1 . 1 . 1) during the regeneration of NAD+. PDC is widespread among plants, absent in animals, and rare in prokaryotes. This is in contrast to ADH, which is widespread among eukaryotes and prokaryotes. Prior to this study, the only bacterial pdc gene described was from the Gram-negative {alpha}-proteobacterium Zymomonas mobilis (Braü & Sahm, 1986 ; Conway et al., 1987 ; Neale et al., 1987a ; Reynen & Sahm, 1988 ). Z. mobilis PDC has been purified to homogeneity, crystallized, and extensively characterized (Candy & Duggleby, 1998 ). PDC has also been purified from one other bacterium, Sarcina ventriculi (Lowe & Zeikus, 1992 ), an unusual Gram-positive organism.

S. ventriculi is an obligate anaerobe that grows from pH 2 to pH 10, fermenting hexose and pentose sugars to produce acetate, ethanol, formate, CO2 and H2 (Goodwin & Zeikus, 1987 ; Stephenson & Dawes, 1971 ). In this organism, the relative production of ethanol and acetate varies with environmental pH. Under acidic conditions, where acetic acid is toxic to cells, ethanol is the primary product (Goodwin & Zeikus, 1987 ). At neutral pH and above, a near-equimolar mixture of ethanol and acetate is produced, with low levels of formate (Lowe & Zeikus, 1991 ). These changes in fermentation profiles have been attributed to changes in the levels of two enzymes that metabolize pyruvate: PDC and pyruvate dehydrogenase (Goodwin & Zeikus, 1987 ; Lowe & Zeikus, 1991 ).

The properties of the S. ventriculi PDC are very different from those of the Z. mobilis enzyme. Unlike the Michaelis–Menten kinetics of Z. mobilis PDC (Hoppner & Doelle, 1983 ; Neale et al., 1987a ), the S. ventriculi enzyme was activated by pyruvate (Lowe & Zeikus, 1992 ), similar to PDCs from yeast and higher plants. S. ventriculi PDC was reported to have an unusually high Km for pyruvate (13 mM), compared to Km values of 0·3–4·4 mM for other PDCs (Hoppner & Doelle, 1983 ; Neale et al., 1987a ; Ullrich & Donner, 1970 ). The phenylalanine content of purified S. ventriculi PDC was four- to fivefold greater than that of other PDCs, suggesting significant differences in primary structure (Lowe & Zeikus, 1992 ).

To further examine the unusual nature of the S. ventriculi PDC, the pdc gene encoding it was cloned, sequenced, and expressed in recombinant E. coli. This approach provided a genetic determination of primary amino acid sequence and facilitated PDC purification for further kinetic and biophysical characterization.


   METHODS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS AND DISCUSSION
REFERENCES
 
Materials.
Biochemicals were purchased from Sigma. Other organic and inorganic analytical-grade chemicals were from Fisher Scientific. Restriction endonucleases and DNA-modifying enzymes were from New England BioLabs. Oligonucleotides were from Sigma-Genosys. Digoxigenin-11-dUTP (2'-deoxyuridine-5'-triphosphate coupled by an 11-atom spacer to digoxigenin), alkaline-phosphatase-conjugated antibody raised against digoxigenin, and nylon membranes for colony and plaque hybridizations were from Roche Molecular Biochemicals. Positively charged membranes for Southern hybridization were from Ambion.

Bacterial strains and media.
Table 1 lists the E. coli strains used in this study, including strains TB-1 and DH5{alpha}, which were used for routine recombinant DNA experiments. E. coli strain SE2309 was used to create a subgenomic DNA library in plasmid pBR322. E. coli strains ER1647, LE392 and BM25.8 were used in conjunction with {lambda}BlueSTAR for a subgenomic DNA library. E. coli strains BL21(DE3), BL21-CodonPlus-RIL and BL21-CodonPlus-RIL/pSJS1240 were used to examine the expression of the S. ventriculi pdc gene from plasmid pJAM419. E. coli strains were grown in Luria–Bertani (LB) medium and supplemented with antibiotics as appropriate (30 mg chloramphenicol l-1, 100 mg carbenicillin l-1, 100 mg ampicillin l-1 and/or 50 mg spectinomycin l-1). S. ventriculi strain Goodsir was cultivated as described previously (Goodwin & Zeikus, 1987 ).


View this table:
[in this window]
[in a new window]
 
Table 1. Strains and plasmids used in this study

 
DNA isolation.
Plasmid DNA was isolated and purified using a Quantum Prep Plasmid Miniprep Kit from Bio-Rad. DNA fragments were eluted from 0·8% SeaKem GTG agarose (FMC Bioproducts) using either Ultrafree-DA filters from Millipore or the QIAquick gel extraction kit from Qiagen. S. ventriculi genomic DNA was isolated and purified as described by Harwood & Cutting (1990) .

Cloning of the S. ventriculi pdc gene.
A degenerate oligonucleotide, 5'-AARGARGTNAAYGTNGARCAYATGTTYGGNGT-3' (R is A or G; N is A, C, G or T; Y is C or T), was synthesized based on the N-terminal amino acid sequence of PDC purified from S. ventriculi (Lowe & Zeikus, 1992 ). This oligonucleotide was labelled at the 3'-end using terminal transferase with digoxigenin-11-dUTP and dATP as recommended by the supplier (Roche Molecular Biochemicals) and was used to screen genomic DNA from S. ventriculi.

For Southern analysis, genomic DNA was digested with BglI, EcoRI or HincII, separated by 0·8% agarose electrophoresis, and transferred to positively charged nylon membranes (Southern, 1975 ). Membranes were equilibrated at 58 °C for 2 h in 5x SSC (1x SSC is 0·15 M NaCl plus 0·015 M sodium citrate) containing 1% blocking reagent (Roche Molecular Biochemicals), 0·1% N-lauroylsarcosine, and 0·02% SDS. After the probe (0·2 pmol ml-1) and poly(A) (0·01 mg ml-1) were added, membranes were incubated at 58 °C for 18·5 h. Membranes were washed twice with 2x SSC containing 0·1% SDS (5 min per wash) at 25 °C and twice with 0·5x SSC containing 0·1% SDS (15 min per wash) at 58 °C. Signals were visualized using colorimetric detection according to the supplier (Roche Molecular Biochemicals).

For generation of a subgenomic library in plasmid pBR322, S. ventriculi chromosomal DNA was digested with HincII and fractionated by electrophoresis. The 2·5–3·5 kb HincII DNA fragments were ligated into the EcoRV site of pBR322 and transformed into E. coli SE2309. Colonies were screened with the degenerate oligonucleotide by colorimetric detection. By this method, plasmid pJAM400, which carries a HincII fragment containing 1350 bp of the pdc gene was isolated.

The {lambda}BlueSTAR Vector System (Novagen) was used to create an additional subgenomic library to facilitate isolation of the full-length pdc gene from S. ventriculi. Genomic DNA was digested with BclI, separated by electrophoresis in 0·8% agarose, and the 6·5–8·5 kb fragments were ligated with the {lambda}BlueSTAR BamHI arms. In vitro packaging and plating of phage was performed according to the supplier (Novagen). A DNA probe was generated using an 800 bp EcoRI fragment of the pdc gene from pJAM400 that was labelled with digoxigenin-11-dUTP using the random-primed method as recommended by the supplier (Roche Molecular Biochemicals). Plaques were screened using colorimetric detection. Cre-loxP-mediated subcloning was used to circularize the DNA of the positive plaques by plating {lambda}BlueSTAR phage with E. coli BM25.8, which expresses Cre recombinase (Novagen). The circularized plasmid pJAM410 was then purified and electroporated into E. coli DH5{alpha}.

For generation of a pdc expression vector, the promoterless pdc gene was subcloned into pET21d after amplification from pJAM413 (Table 1) by PCR. Primers were designed for directional insertion using BspHI (oligo 1) and XhoI (oligo 2) restriction sites. The resulting fragment was ligated into compatible NcoI and XhoI sites of pET21d (Novagen) to produce pJAM419 (Fig. 1). The fidelity of the pdc gene was confirmed by DNA sequencing.



View larger version (15K):
[in this window]
[in a new window]
 
Fig. 1. Partial map of restriction endonuclease sites for a 7 kb BclI genomic DNA fragment from S. ventriculi. Plasmids used in this study include pJAM410, which carries the complete 7 kb BclI fragment, and pJAM400, pJAM411 and pJAM413, which were used for DNA sequence analysis. Plasmid pJAM419 was used for expression of the S. ventriculi pdc gene in recombinant E. coli. The location of the pdc gene and ORF1* (a partial open reading frame of 177 amino acids with no apparent start codon) are shown directly below the physical map, with large arrows indicating the direction of transcription. The dashed line below the physical map indicates the 3886 bp HincII-to-HincII region sequenced.

 
Nucleotide and protein sequence analyses.
DNA fragments of plasmids pJAM400 and pJAM410 (Fig. 1) were subcloned into plasmid vector pUC19 for determining the pdc sequence using the dideoxy termination method (Sanger et al., 1977 ) and a LI-COR automated DNA sequencer (DNA Sequencing Facility, Department of Microbiology and Cell Science, University of Florida). The nucleotide sequence of the S. ventriculi pdc gene and surrounding DNA is deposited in the GenBank database (accession number AF354297).

Genepro 5.0 (Riverside Scientific), CLUSTAL W version 1.81 (Thompson et al., 1994 ), Treeview version 1.5 (Page, 1996 ), and MultiAln (Corpet, 1988 ) were used for DNA and/or protein sequence alignments and comparisons. Deduced amino acid sequences were compared to protein sequences available in the GenBank, EMBL and SWISS-PROT databases at the National Center for Biotechnology Information (Bethesda, MD) using the BLAST network server (Altschul et al., 1990 ). The dense alignment surface (DAS) method was used for the prediction of transmembrane {alpha}-helices (Cserzo et al., 1997 ).

Production of S. ventriculi PDC in recombinant E. coli.
Plasmid pJAM419 was transformed into E. coli BL21-CodonPlus-RIL containing plasmid pSJS1240 (Table 1). Expression of the pdc gene in this plasmid is regulated by the bacteriophage T7 RNA polymerase promoter system (Novagen). Freshly transformed cells were inoculated into LB medium containing ampicillin, spectinomycin and chloramphenicol and grown at 37 °C (200 r.p.m.) until cells reached an OD600 of 0·6–0·8. Transcription was then induced for 2–3 h by the addition of 1 mM IPTG. Cells were harvested by centrifugation at 5000 g (10 min, 4 °C) and stored at -70 °C or in liquid nitrogen.

Purification of the S. ventriculi PDC protein.
All purification buffers contained 1 mM TPP and 1 mM MgSO4 unless indicated otherwise. Recombinant E. coli cells (14·8 g wet wt) were thawed in 6 vols 50 mM sodium phosphate buffer at pH 6·5 (buffer A) and passed through a French pressure cell at 20000 p.s.i. (138 MPa). Cell debris was removed by centrifugation at 16000 g (20 min, 4 °C). Supernatant was removed and filtered through a 45 µm filter membrane. Filtrate (692 mg protein) was applied to a Q Sepharose Fast Flow 26/10 column (Pharmacia) that was equilibrated at 2 ml min-1 with buffer A containing 300 mM NaCl. Fractions containing PDC activity (326 mg protein) were precipitated with 80% saturation (NH4)2SO4. Protein was resuspended in buffer A, dialysed against buffer A (4 °C, 16 h), and filtered (45 µm membrane). The filtrate (287 mg) was applied to a Q Sepharose column equilibrated with buffer A and developed with a linear NaCl gradient (0–400 mM NaCl in 220 ml buffer A) at 4 ml min-1. PDC active fractions eluted at 230–300 mM NaCl and were pooled. The pooled sample (23 mg) was applied to a 5 ml Bio-scale hydroxyapatite type I column (Bio-Rad) that was equilibrated at 1 ml min-1 with 5 mM sodium phosphate buffer at pH 6·5 (buffer B). The column was washed with 15 ml buffer B and developed with a linear sodium phosphate gradient (5–500 mM sodium phosphate at pH 6·5 in 75 ml). Protein fractions (11·4 mg) with PDC activity eluted at 200–300 mM sodium phosphate and were pooled. For further purification, portions of this material (0·25–0·5 mg protein per 0·25–0·5 ml) were applied to a Superdex 200 HR 10/30 column (Pharmacia) equilibrated at 0·25 ml min-1 in 50 mM sodium phosphate at pH 6·5 with 150 mM NaCl and 10% (v/v) glycerol in the presence and absence of 1 mM MgSO4 and 1 mM TPP.

Activity assays and protein electrophoretic techniques.
PDC activity was assayed by monitoring the pyruvic-acid-dependent reduction of NAD+ with baker’s yeast alcohol dehydrogenase (ADH) (Sigma) as a coupling enzyme at pH 6·5 as previously described (Conway et al., 1987 ). The assay also included the following modifications. Buffered enzyme (100 µl) was added to a final volume of 1 ml containing 0·15 mM NADH, 0·1 mM TPP, 0–25 mM pyruvate and 10 U ADH in 50 mM potassium MES or 100 mM sodium hydrogen maleate buffer with 5 mM MgCl2 at pH 6·5. Since this assay does not distinguish PDC from NADH-oxidizing enzymes such as lactate dehydrogenase, activity of cell lysate was estimated by correcting for control reactions performed in the absence of added ADH. One unit of enzyme activity is defined as amount of enzyme that generates 1 µmol acetaldehyde min-1. Thermostability was determined by incubating purified PDC in 50 mM sodium phosphate buffer at pH 6·5 with 1 mM TPP and 1 mM MgCl2 for 90 min and then assaying for activity with 10 mM pyruvate. Protein concentration was determined using Bradford protein reagent with bovine serum albumin as the standard (Bio-Rad).

Molecular mass and N-terminal amino acid sequence analyses.
Subunit molecular mass was estimated by reducing and denaturing SDS-PAGE using 12% polyacrylamide gels which were stained with Coomassie blue R-250. The molecular mass standards for SDS-PAGE were: phosphorylase b (97·4 kDa), bovine serum albumin (66·2 kDa), ovalbumin (45 kDa), carbonic anhydrase (31 kDa), trypsin inhibitor (21·5 kDa) and lysozyme (14·4 kDa). For determination of native molecular mass, samples were applied to a Superdex 200 HR 10/30 column equilibrated in 50 mM sodium phosphate buffer at pH 6·5 with 150 mM NaCl, 10% glycerol, and no added cofactors. Molecular mass standards included: bovine serum albumin (66 kDa), alcohol dehydrogenase (150 kDa), {alpha}-amylase (200 kDa), apoferritin (443 kDa), and thyroglobulin (669 kDa).

The N-terminal sequence was determined for PDC protein purified from recombinant E. coli. The protein was separated by SDS-PAGE and electroblotted onto a PVDF membrane (Immobilon-P). The sequence was determined by automated Edman degradation at the protein chemistry core facility of the University of Florida Interdisciplinary Center for Biotechnology Research.


   RESULTS AND DISCUSSION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS AND DISCUSSION
REFERENCES
 
S. ventriculi PDC operon
The N-terminal amino acid sequence of the PDC protein purified from S. ventriculi (Lowe & Zeikus, 1992 ) was used to generate a degenerate oligonucleotide for hybridization to genomic DNA. This approach facilitated the isolation of a 7·0 kb BclI genomic DNA fragment from S. ventriculi. The fragment was further subcloned in order to sequence both strands of a 3886 bp HincII-to-HincII region that hybridized to the oligonucleotide probe (Fig. 1). Analysis of the DNA sequence revealed an ORF of 1656 bp encoding a protein with an N-terminus identical to that of the previously purified S. ventriculi PDC. The ORF was therefore designated pdc. A canonical Shine–Dalgarno sequence is present 7 bp upstream of the pdc translation start codon. In addition, a region 82–110 bp upstream of pdc has limited identity to the eubacterial -35 and -10 promoter consensus sequence (aTaACA-N16-TATtAa, where N is any nucleotide and upper-case nucleotides match the consensus). Downstream (43 bp) of the pdc translation stop codon is a 40 bp region predicted to form a stem–loop structure followed by an AT-rich region, consistent with a {rho}-independent transcription terminator. Thus, the S. ventriculi pdc appears to be transcribed as a monocistronic operon like the Z. mobilis pdc gene (Conway et al., 1987 ).

A partial ORF was identified 722 bp upstream of pdc which encodes a 177 amino acid protein fragment (ORF1*) (Fig. 1). ORF1* has identity (28–29%) to several hypothetical membrane proteins (GenBank accession numbers CAC11620, CAC24018, CAA22902) and is predicted to form several membrane-spanning domains (data not shown).

S. ventriculi PDC protein sequence
The S. ventriculi pdc gene apparently encodes a protein of 552 amino acids (including the N-terminal methionine) with a calculated pI of 5·16 and anhydrous molecular mass of 61737 Da. Consistent with other Z. mobilis and fungal PDC proteins, the N-terminal extension of up to 47 amino acids that is common to plant PDC proteins is not conserved in the S. ventriculi PDC protein. Although the pI of the purified S. ventriculi PDC protein has not been experimentally determined, the calculated pI is consistent with the acidic pH optimum of 6·3–6·7 for stability and activity of this enzyme (Lowe & Zeikus, 1992 ). The amino acid composition of the protein deduced from the S. ventriculi pdc gene is similar to that determined for the Saccharomyces cerevisiae PDC1 (Kellermann et al., 1986 ) and Z. mobilis pdc genes (Conway et al., 1987 ; Neale et al., 1987b ). A notable exception is the alanine composition of Z. mobilis PDC, which is 1·8- to 2·2-fold higher than that of S. cerevisiae PDC1 and S. ventriculi PDC. Although the phenylalanine composition of the deduced S. ventriculi PDC protein is consistent with the other PDC proteins, it is almost 3·6-fold less than the composition previously reported for the purified S. ventriculi enzyme (Lowe & Zeikus, 1992 ). The reason for this discrepancy remains to be determined.

The amino acid sequence of S. ventriculi PDC was aligned with the sequences of the yeast (Sce) PDC1 and Z. mobilis (Zmo) PDC proteins, both of which have been analysed by X-ray crystallography (Arjunan et al., 1996 ; Dobritzsch et al., 1998 ; Dyda et al., 1993 ; König, 1998 ) (Fig. 2). The conserved motif of TPP-dependent enzymes identified by Hawkins et al. (1989) and known to be involved in Mg2+-TPP cofactor binding is highly conserved in all three PDC proteins. Amino acid residues located within 0·4 nm of the Mg2+- and TPP-binding site of the PDC proteins which have been crystallized from yeast and Z. mobilis are conserved in primary sequence with the S. ventriculi PDC protein. These include residues with similarity to the aspartate (SceD444, ZmoD440) and asparagine residues (SceN471, ZmoN467) that are involved in binding Mg2+. The S. ventriculi PDC appears to be more similar to the yeast PDC than to that of Z. mobilis in binding the diphosphates of TPP, where serine and threonine side chains (SceS446 and T390) as well as the main chain nitrogen of isoleucine (SceI476) are conserved. This contrasts with the Z. mobilis enzyme, which utilizes a main-chain nitrogen of aspartate (ZmoD390) instead of the threonine hydroxyl group (SceT390) for binding the ß-phosphate. S. ventriculi PDC residues are also similar to the aspartate, glutamate, threonine and histidine residues (SceD28, E477, T388, H114 and H115; ZmoD27, E473, T388, H113 and H114) which may potentially interact with intermediates during the decarboxylation reaction mediated by Z. mobilis and yeast PDCs. Furthermore, the isoleucine (Sce and Zmo I415) side chain, which appears to stabilize the V conformation of TPP through Van der Waals interactions, as well as the glutamate (SceE51, ZmoE50) which may donate a proton to the N1' atom of TPP, are conserved in the S. ventriculi PDC protein sequence.



View larger version (88K):
[in this window]
[in a new window]
 
Fig. 2. Multiple amino acid sequence alignment of S. ventriculi PDC with other PDC protein sequences. Abbreviations with GenBank or SWISS-PROT accession numbers: Sce, Saccharomyces cerevisiae P06169; Sve, Sarcina ventriculi; Zmo, Zymomonas mobilis P06672. Identical, functionally conserved and semi-conserved amino acid residues are shaded. Dashes indicate gaps introduced in protein sequence alignment. Indicated above the sequences are amino acid residues within 0·4 nm of the Mg2+- and TPP-binding site of yeast PDC1 (Dyda et al., 1993 ) ({blacktriangledown}), the Cys221 residue originally postulated to be required for pyruvate activation of yeast PDC1 ({bullet}), and the Tyr157 and Arg224 residues which form hydrogen bonds with allosteric activators such as pyruvamide ({blacksquare}). The underlined sequence is a conserved motif identified in TPP-dependent enzymes (Hawkins et al., 1989 ).

 
A notable exception in conservation is the yeast C221 residue (Fig. 2), which is highly conserved among the majority of fungal PDCs but is not conserved in either bacterial or plant PDC proteins. Based on site-directed mutagenesis, chemical modification and kinetic studies, this C221 residue has been proposed to be a primary binding site of the regulatory substrate molecule and the starting point of a signal transfer pathway to the active-site TPP in the yeast enzyme (Baburina et al., 1994 , 1998 ; Hübner et al., 1988 ). Consistent with these previous results the yeast C221 is positioned in a large cavity formed at the interface among all three PDC domains including the {alpha} or PYR (residues 1 to 189), ß or R (residues 190 to 356) and {gamma} or PP (residues 357 to 563) domains (Dyda et al., 1993 ). However, recent high-resolution structural analysis of the brewer’s yeast PDC crystallized in the presence of pyruvamide, a pyruvate analogue, enabled localization of the activator-binding site and revealed that cysteine does not play a direct role in this binding (Lu et al., 2000 ). Additionally, kinetic studies using stopped-flow techniques revealed that the C221A variant of yeast PDC was still substrate activated, and the lag phase of product formation did not disappear with progressive thiol oxidation (Lu et al., 2000 ). Instead, tyrosine (Y157) and arginine (R224) residues form hydrogen bonds with the amide group of pyruvamide. Both of these residues are conserved in the S. ventriculi PDC and not found in the Z. mobilis enzyme, which displays Michaelis–Menten kinetics. These results suggest that residues of the S. ventriculi PDC protein may allosterically bind the substrate activator with a mechanism common to the majority of fungal PDC proteins. Interestingly, these two residues that bind pyruvamide in the yeast enzyme are not universally distributed among the substrate-activated PDCs, most notably the plant PDCs.

Cluster analysis was performed to compare the PDC proteins and other TPP-dependent enzymes, including indole-3-pyruvate decarboxylase (IPD), the E1 component of pyruvate dehydrogenase (E1), acetolactate synthase (ALS), and transketolase (TK) (Fig. 3). The comparison reveals that all of these proteins are related in primary sequence and that there is a significant clustering of the sequences into families based on specific enzyme function. Of these, the S. ventriculi PDC appears most closely related to eubacterial IPD proteins as well as the majority of fungal PDC proteins. In contrast, the Z. mobilis PDC protein is most closely related to plant PDCs in addition to a couple of outgrouping fungal PDCs. Thus, it appears that the IPD protein family has close evolutionary roots with the PDC family and specifically the S. ventriculi PDC protein. In addition, the distant relationship of the S. ventriculi and Z. mobilis PDC proteins is consistent with the biochemical differences between these two enzymes (Hoppner & Doelle, 1983 ; Lowe & Zeikus, 1992 ; Neale et al., 1987a ; Ullrich & Donner, 1970 ). In contrast to Z. mobilis PDC, which may have originated by the horizontal transfer of a plant pdc gene, S. ventriculi PDC appears to have diverged quite early during evolution and last shared a common ancestor with most eubacterial IPD and fungal PDCs.



View larger version (35K):
[in this window]
[in a new window]
 
Fig. 3. Dendrogram summarizing the relationships between selected PDCs and other TPP-dependent enzymes. Deduced protein sequences were aligned using CLUSTAL X. Amino acid extensions at the N- or C-terminus, and apparent insertion sequences, were removed. Remaining regions containing approximately 520–540 amino acids were compared. Treeview was used to display these results as an unrooted dendrogram. Protein abbreviations: PDC, pyruvate decarboxylase; IPD, indole-3-pyruvate decarboxylase; ALS, acetolactate synthase; PDH E1 or E1, the E1 component of pyruvate dehydrogenase; TK, transketolase. Organism abbreviations and GenBank or SWISS-PROT accession numbers: Abr, Azospirillum brasilense P51852, Aor, Aspergillus oryzae AAD16178; Apa, Aspergillus parasiticus P51844; Asy, Ascidia sydneiensis samea BAA74730; Ath, Arabidopsis thaliana BAB08775; Bfl, Brevibacterium flavum A56684; Bsu, Bacillus subtilis P45694; Cgl, Corynebacterium glutamicum P42463; Cpn, Chlamydophila pneumoniae H72020; Dra, Deinococcus radiodurans A75387 (ALS), A75541 (E1); Ecl, Enterobacter cloacae P23234; Eco, Escherichia coli CAA24740; Ehe, Erwinia herbicola AAB06571; Eni, Aspergillus (Emericella) nidulans P87208; Fan, Fragaria x ananassa AAG13131; Ghi, Gossypium hirsutum S60056; Gth, Guillardia theta NP_050806; Huv, Hanseniaspora uvarum P34734; Kla, Kluyveromyces lactis Q12629 (PDC), Q12630 (TK); Kma, Kluyveromyces marxianus P33149; Mav, Mycobacterium avium Q59498; Mja, Methanococcus jannaschii Q57725; Mle, Mycobacterium leprae CAC31122 (ORF), 033112 (ALS), CAC30602 (E1); Mth, Methanobacterium thermoautotrophicum A69081 (ORF), C69059 (ALS); Mtu, Mycobacterium tuberculosis E70814 (IPD), 053250 (ALS); Ncr, Neurospora crassa P33287; Nta, Nicotiana tabacum P51846 (PDC), P09342 (ALS); Osa, Oryza sativa P51847 (PDC1), P51848 (PDC2), P51849 (PDC3); Pae, Pseudomonas aeruginosa G83123; Pmu, Pasteurella multocida AAK03712; Pop, Porphyra purpurea NP_053940; Ppu, Pseudomonas putida AAG00523; Psa, Pisum sativum P51850; Pst, Pichia stipitis AAC03164 (PDC1), AAC03165 (PDC2); Rca, Rhodobacter capsulatus JC4637; Reu, Ralstonia eutropha Q59097; Sav, Streptomyces avermitilis AAA93098; Sce, Saccharomyces cerevisiae P06169 (PDC1), P16467 (PDC5), P26263 (PDC6), Q07471 (ORF), NP_011105 (E1{alpha}), NP_009780 (E1ß); Sco, Streptomyces coelicolor T35828; Skl, Saccharomyces kluyveri AAF78895; Spl, Spirulina platensis P27868; Spo, Schizosaccharomyces pombe Q09737 (PDC1), Q92345 (PDC2); Sve, Sarcina ventriculi AF354297; Syn, Synechocystis sp. BAA17984; Vch, Vibrio cholerae A82375; Vvi, Vitis vinifera AAG22488; Zma, Zea mays P28516; Zbi, Zygosaccharomyces bisporus CAB65554; Zmo, Zymomonas mobilis P06672. Scale bar represents 0·1 nucleotide substitutions per site.

 
Production of S. ventriculi PDC protein
Significant differences in codon usage between the S. ventriculi and Z. mobilis pdc genes are evident and may impact translation in engineered host organisms (Table 2). In particular, the pdc gene of S. ventriculi encodes elevated use of tRNAAUA and tRNAAGA, both of which are relatively rare in E. coli. This is in contrast to the Z. mobilis pdc gene, which does not use the AUA codon and has only minimal use of the AGA codon. This suggests that production of the S. ventriculi PDC protein in recombinant E. coli may be limited by mRNA translation due to the requirement for elevated use of rare tRNAs.


View this table:
[in this window]
[in a new window]
 
Table 2. Codon usage of S. ventriculi (Sve) and Z. mobilis (Zmo) pdc genes

 
To further investigate this, the levels of the tRNA genes that are rare in E. coli were modified during pdc expression by including multiple copies of these genes on the chromosome (E. coli strain BL21-CodonPlus-RIL) and/or on a complementary plasmid (pSJS1240) (Table 1). These modified E. coli strains were transformed with plasmid pJAM419, which carries the S. ventriculi pdc gene positioned 8 bp downstream of an optimized Shine–Dalgarno consensus sequence and controlled at the transcriptional level by T7 RNA polymerase promoter and terminator sequences from plasmid vector pET21d. Detectable levels of PDC activity (0·16 U per mg protein at 5 mM pyruvate) were observed after induction of pdc transcription in E. coli host strains with additional chromosomal and/or plasmid copies of the ileU/X, argU and leuW genes encoding the rare tRNAAUA, tRNAAGG/AGA and tRNACUA. A 5- to 10-fold increase in the levels of a 58 kDa protein with a molecular mass comparable to the S. ventriculi PDC (Fig. 4) were produced in these strains compared to a similar E. coli strain without added tRNA genes [BL21(DE3)] (data not shown). These results suggest that the S. ventriculi PDC protein is synthesized in recombinant E. coli and that the high percentage of AUA and AGA codons of the pdc gene limits translation.



View larger version (78K):
[in this window]
[in a new window]
 
Fig. 4. S. ventriculi PDC protein synthesized in recombinant E. coli. Proteins were analysed by reducing SDS-PAGE using 12% polyacrylamide gels and stained with Coomassie blue R-250. Lanes 1 and 4, molecular mass standards (5 µg). Lanes 2 and 3, cell lysate (20 µg) of IPTG-induced E. coli BL21-CodonPlus-RIL/pSJS1240 transformed with pET21d or pJAM419, respectively. Lane 5. S. ventriculi PDC protein (2 µg) purified from recombinant E. coli.

 
Interestingly, additional proteins of 43 and 27 kDa were also observed when the PDC protein was synthesized in E. coli compared to control strains (Fig. 4). The origin of these proteins remains to be determined. They may be fragments of PDC generated by proteolysis or truncated PDC produced from errors in translation/transcription. Alternatively, increased production of PDC may increase the levels of acetaldehyde, which can be toxic to the cell, and may subsequently induce the levels of proteins in response to this stress.

Biochemical and biophysical properties of the S. ventriculi PDC protein from recombinant E. coli
The S. ventriculi PDC protein was purified over 136-fold from recombinant E. coli. The N-terminal amino acid sequence of this protein (MKITIAEYLLXR, where X is an unidentified amino acid) was identical to the sequence of PDC purified from S. ventriculi (Lowe & Zeikus, 1992 ). Both PDC proteins have an N-terminal methionine residue, which suggests that this residue is not accessible for cleavage by either the S. ventriculi or E. coli aminopeptidases.

The thermostability of the purified S. ventriculi PDC was examined in the presence of 1 mM cofactors TPP and Mg2+ at pH 6·5. Enzyme activity was stable up to 42 °C but was abolished after incubation for 60–90 min at temperatures of 50 °C and above. This is consistent with the significant loss of PDC activity observed when a thermal treatment step (60 °C for 30 min) was included in the purification (data not shown). In contrast, methods used to purify Z. mobilis and other PDC proteins (König, 1998 ) typically incorporate thermal treatment to remove unwanted proteins. These results suggest that the recombinant S. ventriculi PDC protein is not as thermostable as other PDC proteins, including that of Z. mobilis.

PDC proteins have been shown to bind TPP and Mg2+ cofactors with high affinity at slightly acidic pH (König, 1998 ). Consistent with this, the recombinant S. ventriculi PDC retains full activity after incubation at 37 °C for 90 min in the presence of 25 mM EGTA or EDTA in pH 6·5 buffer without cofactors. This is similar to the PDC protein purified directly from S. ventriculi, which is fully active after similar treatment with metal chelators.

The recombinant S. ventriculi enzyme displays sigmoidal kinetics (Fig. 5), suggesting that the enzyme is substrate activated, similar to the fungal and plant PDC proteins (König, 1998 ). This contrasts with the Z. mobilis PDC, which is the only PDC known to display non-sigmoidal, Michaelis–Menten kinetics. Based on the kinetic data, it appears that the S. ventriculi PDC has two affinities for pyruvate (Km values of 2·8 and 10 mM) (Fig. 5). The Km value for the low-affinity site is comparable to the Km value (13 mM) reported by Lowe & Zeikus (1992) . The fivefold lower Km observed in this present study is apparently due to the difference in the buffer, since maleate buffer significantly inhibited the PDC activity at low pyruvate concentrations (Fig. 5). This difference in activity of S. ventriculi in maleate buffer is similar to that observed with Z. mobilis PDC, with a Km of 4·4 mM for pyruvate in Tris/maleate buffer at pH 6 (Hoppner & Doelle, 1983 ) compared to Km values of 0·3–0·4 mM in potassium MES buffer at pH 6 (Bringer-Meyer et al., 1986 ) and sodium citrate buffer at pH 6·5 (Neale et al., 1987b ).



View larger version (10K):
[in this window]
[in a new window]
 
Fig. 5. Kinetics for pyruvate of the S. ventriculi PDC purified from recombinant E. coli. The data represent mean results from triplicate determinations of PDC activity by the ADH coupled assay using 1 µg purified enzyme in 1 ml final assay volume with either potassium MES ({bullet}) or sodium hydrogen maleate ({circ}) buffer as described in Methods.

 
The high-affinity Km value of 2·8 mM obtained for the S. ventriculi PDC is similar to the Km values obtained with many fungal and plant PDCs, including those purified from Saccharomyces cerevisiae (1–3 mM) (Boiteux & Hess, 1970 ; Hübner et al., 1978 ), Zygosaccharomyces bisporus (1·73 mM) (Neuser et al., 2000 ), orange (0·8–3·2 mM) (Raymond et al., 1979 ), and wheat germ (3 mM) (Zehender et al., 1987 ). However, these Km values are several-fold higher than those reported for the PDCs of Z. mobilis (0·3–0·4 mM) (Bringer-Meyer et al., 1986 ; Neale et al., 1987b ) and rice (0·25 mM) (Rivoal et al., 1990 ).

At pH 6·5, the recombinant S. ventriculi PDC forms a 235 kDa homotetramer consisting of a 58 kDa protein, as determined by Superdex 200 gel filtration chromatography and SDS-12% PAGE (see Methods). Exclusion of the cofactors from the buffer during gel filtration chromatography at pH 6·5 did not alter the tetramer configuration or enzyme activity, suggesting that the cofactors are tightly bound. The configuration of the PDC complex is consistent with that purified from S. ventriculi as well as the majority of those isolated from fungi, plants and Z. mobilis. There are, however, plant PDCs which have been reported to form larger complexes, including the PDC from Neurospora crassa, which forms aggregated filaments of 8–10 nm (Alvarez et al., 1993 ) as well as the PDC from Pisum sativum, which forms up to 960 kDa complexes (Mücke et al., 1995 ).

Conclusions
Based on this study, the S. ventriculi PDC protein appears to share similar primary sequence structure with TPP-dependent enzymes and is highly related to the fungal PDC and eubacterial IPD enzymes. The close relationship of the S. ventriculi and fungal PDC structures is consistent with the similar biochemical and biophysical properties of these enzymes. Both types of enzymes display substrate cooperativity with similar affinities for pyruvate. The structure and biochemistry of the S. ventriculi PDC, however, dramatically contrast with the only other bacterial PDC (Z. mobilis) which has been characterized. The Z. mobilis PDC is closely related to plants in primary structure and it is the only PDC known to display Michaelis–Menten kinetics.

This study also demonstrates the synthesis of active, soluble S. ventriculi PDC protein in recombinant E. coli. Only two other genes, the Z. mobilis pdc and Saccharomyces cerevisiae PDC1 genes, have been reported to synthesize PDC protein in recombinant bacteria (Braü & Sahm, 1986 ; Candy et al., 1991 ; Conway et al., 1987 ). Of these, at least 50% of the S. cerevisiae PDC1 forms insoluble inclusions in E. coli and thus has not been useful in engineering bacteria for high-level ethanol production (Candy et al., 1991 ). Due to codon bias, accessory tRNA is essential for efficient production of S. ventriculi PDC in recombinant E. coli. However, the low-G+C codon bias of the S. ventriculi pdc gene should broaden the spectrum of bacteria that can be engineered as hosts for high-level production of PDC protein and the engineering of homo-ethanol pathways (Ingram et al., 1999 ).


   ACKNOWLEDGEMENTS
 
We thank Francis Davis and Jack Shelton for DNA sequencing. Thanks also to K. T. Shanmugam for generously providing E. coli strain SE2309, and Sung-Hou and Rosalind Kim for providing plasmid pSJS1240. This work was supported in part by the Department of Energy (ZDH-9-29009-04 and DE-FG02-96ER20222) and the Florida Agricultural Experiment Station (Journal Series R-08274).


   REFERENCES
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS AND DISCUSSION
REFERENCES
 
Altschul, S. F., Gish, W., Miller, W., Myers, E. W. & Lipman, D. J. (1990). Basic local alignment search tool. J Mol Biol 215, 403-410.[Medline]

Alvarez, M. E., Rosa, A. L., Temporini, E. D., Wolstenholme, A., Panzetta, G., Patrito, L. & Maccioni, H. J. (1993). The 59-kDa polypeptide constituent of 8–10-nm cytoplasmic filaments in Neurospora crassa is a pyruvate decarboxylase. Gene 130, 253-258.[Medline]

Arjunan, P., Umland, T., Dyda, F. & 7 other authors (1996). Crystal structure of the thiamin diphosphate-dependent enzyme pyruvate decarboxylase from the yeast Saccharomyces cerevisiae at 2·3 resolution. J Mol Biol 256, 590–600.[Medline]

Baburina, I., Gao, Y., Hu, Z., Jordan, F., Hohmann, S. & Furey, W. (1994). Substrate activation of brewers’ yeast pyruvate decarboxylase is abolished by mutation of cysteine 221 to serine. Biochemistry 33, 5630-5635.[Medline]

Baburina, I., Li, H., Bennion, B., Furey, W. & Jordan, F. (1998). Interdomain information transfer during substrate activation of yeast pyruvate decarboxylase: the interaction between cysteine 221 and histidine 92. Biochemistry 37, 1235-1244.[Medline]

Barbosa, M. F. S. & Ingram, L. O. (1994). Expression of the Zymomonas mobilis alcohol dehydrogenase II (adhB) and pyruvate decarboxylase (pdc) genes in Bacillus. Curr Microbiol 28, 279-282.

Boiteux, A. & Hess, B. (1970). Allosteric properties of yeast pyruvate decarboxylase. FEBS Lett 9, 293-296.[Medline]

Braü, B. & Sahm, H. (1986). Cloning and expression of the structural gene for pyruvate decarboxylase of Zymomonas mobilis in Escherichia coli. Arch Microbiol 144, 296-301.

Bringer-Meyer, S., Schimz, K.-L. & Sahm, H. (1986). Pyruvate decarboxylase from Zymomonas mobilis. Isolation and partial characterization. Arch Microbiol 146, 105-110.

Candy, J. M. & Duggleby, R. G. (1998). Structure and properties of pyruvate decarboxylase and site-directed mutagenesis of the Zymomonas mobilis enzyme. Biochim Biophys Acta 1385, 323-338.[Medline]

Candy, J. M., Duggleby, R. G. & Mattick, J. S. (1991). Expression of active yeast pyruvate decarboxylase in Escherichia coli. J Gen Microbiol 137, 2811-2815.[Medline]

Conway, T., Osman, Y. A., Konnan, J. I., Hoffmann, E. M. & Ingram, L. O. (1987). Promoter and nucleotide sequences of the Zymomonas mobilis pyruvate decarboxylase. J Bacteriol 169, 949-954.[Medline]

Corpet, F. (1988). Multiple sequence alignment with hierarchical clustering. Nucleic Acids Res 16, 10881-10890.[Abstract]

Cserzo, M., Wallin, E., Simon, I., von Heijne, G. & Elofsson, A. (1997). Prediction of transmembrane {alpha}-helices in prokaryotic membrane proteins: the dense alignment surface method. Protein Eng 10, 673-676.[Abstract]

Dobritzsch, D., König, S., Schneider, G. & Lu, G. (1998). High resolution crystal structure of pyruvate decarboxylase from Zymomonas mobilis. Implications for substrate activation in pyruvate decarboxylases. J Biol Chem 273, 20196-20204.[Abstract/Free Full Text]

Dyda, F., Furey, W., Swaminathan, S., Sax, M., Farrenkopf, B. & Jordan, F. (1993). Catalytic centers in the thiamin diphosphate dependent enzyme pyruvate decarboxylase at 2·4- resolution. Biochemistry 32, 6165-6170.[Medline]

Gold, R. S., Meagher, M. M., Tong, S., Hutkins, R. W. & Conway, T. (1996). Cloning and expression of the Zymomonas mobilis ‘‘production of ethanol’’ genes in Lactobacillus casei. Curr Microbiol 33, 256-260.[Medline]

Goodwin, S. & Zeikus, J. G. (1987). Physiological adaptations of anaerobic bacteria to low pH: metabolic control of proton motive force in Sarcina ventriculi. J Bacteriol 169, 2150-2157.[Medline]

Harwood, C. R. & Cutting, S. M. (1990). Molecular Biological Methods for Bacillus. New York: Wiley.

Hawkins, C. F., Borges, A. & Perham, R. N. (1989). A common structural motif in thiamin pyrophosphate-binding enzymes. FEBS Lett 255, 77-82.[Medline]

Hoppner, T. C. & Doelle, H. W. (1983). Purification and kinetic characteristics of pyruvate decarboxylase and ethanol dehydrogenase from Zymomonas mobilis in relation to ethanol production. Eur J Appl Microbiol Biotechnol 17, 152-157.

Hübner, G., Weidhase, R. & Schellenberger, A. (1978). The mechanism of substrate activation of pyruvate decarboxylase: a first approach. Eur J Biochem 92, 175-181.[Abstract]

Hübner, G., König, S. & Schellenberger, A. (1988). The functional role of thiol groups of pyruvate decarboxylase from brewer’s yeast. Biomed Biochim Acta 47, 9-18.[Medline]

Ingram, L. O., Aldrich, H. C., Borges, A. C. & 9 other authors (1999). Enteric bacterial catalysts for fuel ethanol production. Biotechnol Prog 15, 855–866.[Medline]

Kellermann, E., Seeboth, P. G. & Hollenberg, C. P. (1986). Analysis of the primary structure and promoter function of a pyruvate decarboxylase gene (PDC1) from Saccharomyces cerevisiae. Nucleic Acids Res 14, 8963-8977.[Abstract]

Kim, R., Sandler, S. J., Goldman, S., Yokota, H., Clark, A. J. & Kim, S.-H. (1998). Overexpression of archaeal proteins in Escherichia coli. Biotechnol Lett 20, 207-210.

König, S. (1998). Subunit structure, function and organisation of pyruvate decarboxylases from various organisms. Biochim Biophys Acta 1385, 271-286.[Medline]

Lowe, S. E. & Zeikus, J. G. (1991). Metabolic regulation of carbon and electron flow as a function of pH during growth of Sarcina ventriculi. Arch Microbiol 155, 325-329.

Lowe, S. E. & Zeikus, J. G. (1992). Purification and characterization of pyruvate decarboxylase from Sarcina ventriculi. J Gen Microbiol 138, 803-807.[Medline]

Lu, G., Dobritzsch, D., Baumann, S., Schneider, G. & König, S. (2000). The structural basis of substrate activation in yeast pyruvate decarboxylase. A crystallographic and kinetic study. Eur J Biochem 267, 861-868.[Abstract/Free Full Text]

Mücke, U., König, S. & Hübner, G. (1995). Purification and characterisation of pyruvate decarboxylase from pea seeds (Pisum sativum cv. Miko). Biol Chem Hoppe-Seyler 376, 111-117.[Medline]

Neale, A. D., Scopes, R. K., Wettenhall, R. E. & Hoogenraad, N. J. (1987a). Nucleotide sequence of the pyruvate decarboxylase gene from Zymomonas mobilis. Nucleic Acids Res 15, 1753-1761.[Abstract]

Neale, A. D., Scopes, R. K., Wettenhall, R. E. & Hoogenraad, N. J. (1987b). Pyruvate decarboxylase of Zymomonas mobilis: isolation, properties, and genetic expression in Escherichia coli. J Bacteriol 169, 1024-1028.[Medline]

Neuser, F., Zorn, H., Richter, U. & Berger, R. G. (2000). Purification, characterisation and cDNA sequencing of pyruvate decarboxylase from Zygosaccharomyces bisporus. Biol Chem 381, 349-353.[Medline]

Page, R. D. (1996). TreeView: an application to display phylogenetic trees on personal computers. Comput Appl Biosci 12, 357-358.[Medline]

Raymond, W. R., Hostettler, J. B., Assar, K. & Varsel, C. (1979). Orange pyruvate decarboxylase: isolation and mechanistic studies. J Food Sci 44, 777-781.

Reynen, M. & Sahm, H. (1988). Comparison of the structural genes for pyruvate decarboxylase in different Zymomonas mobilis strains. J Bacteriol 170, 3310-3313.[Medline]

Rivoal, J., Ricard, B. & Pradet, A. (1990). Purification and partial characterization of pyruvate decarboxylase from Oryza sativa L. Eur J Biochem 194, 791-797.[Abstract]

Sanger, F., Nicklen, S. & Coulson, A. R. (1977). DNA sequencing with chain-terminating inhibitors. Proc Natl Acad Sci USA 74, 5463-5467.[Abstract]

Southern, E. M. (1975). Detection of specific sequences among DNA fragments separated by gel electrophoresis. J Mol Biol 98, 503-517.[Medline]

Stephenson, M. P. & Dawes, E. A. (1971). Pyruvic acid and formic acid metabolism in Sarcina ventriculi and the role of ferredoxin. J Gen Microbiol 69, 331-343.[Medline]

Thompson, J. D., Higgins, D. G. & Gibson, T. J. (1994). CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res 22, 4673-4680.[Abstract]

Ullrich, J. & Donner, I. (1970). Kinetic evidence for two active sites in cytoplasmic yeast pyruvate decarboxylase. Hoppe Seylers Z Physiol Chem 351, 1026-1029.[Medline]

Zehender, H., Trescher, D. & Ullrich, J. (1987). Improved purification of pyruvate decarboxylase from wheat germ. Its partial characterisation and comparison with the yeast enzyme. Eur J Biochem 167, 149-154.[Abstract]

Received 8 March 2001; revised 26 May 2001; accepted 4 June 2001.