Cloning, sequencing and expression of the L-2-hydroxyisoicaproate dehydrogenase-encoding gene of Lactobacillus confusus in Escherichia coli

Cloning, sequencing and expression of the L-2-hydroxyisoicaproate dehydrogenase-encoding gene of Lactobacillus confusus in Escherichia coli

263 Gene, 83 (1989) 263-270 Elsevier GENE 03187 Cloning, sequencing and expression of the L-2-hydroxyisoicaproate dehydrogenase-encoding gene of Lac...

780KB Sizes 2 Downloads 20 Views


Gene, 83 (1989) 263-270 Elsevier GENE 03187

Cloning, sequencing and expression of the L-2-hydroxyisoicaproate dehydrogenase-encoding gene of Lactobacillus confusus in Escherichia coli (Recombinant DNA; promoter; terminator; codon usage; amino acid sequence similarity: L-lactate dehydrogenase)

Hans-Philipp Lerch, Ronald Frank and John Collins Gesellschaftfir Biotechnologische Forschung mbH, D-3300 Braunschweig (F.R. G.) Received by J.-P. Lecocq: 10 February 1989 Revised: 17 April 1989 Accepted: 27 April 1989


The gene (L-HicDH) encoding L-2-hydroxyisocaproate dehydrogenase (L-HicDH) from Lactobacillus confi~us was cloned in Escherichiu coli. A 69-mer oligodeoxyribonucleotide probe, derived to be complementary to the N-terminal amino acid (aa) coding sequence, was used for screening. The complete nucleotide (nt) sequence of the L-HicDH gene was determined. The 5’-end of the mRNA was mapped by primer extension and the promoter identified. Downstream from the L-HicDH gene is a typical Rho-independent terminator. The aa sequence of L-HicDH, deduced from the nt sequence, has an overall similarity of 30% to the aa sequence of L-lactate dehydrogenase (L-LDH) from Lactobucillus cusei. The aa residues involved in binding of coenzyme and substrate are highly conserved in L-HicDH with respect to prokaryotic and eukaryotic L-LDHs. The L-HicDH gene could be expressed under control of phage 1 ‘Leftward’ and ‘rightward’ promoters in E. coli up to 35% of total cell protein. The enzyme produced under these conditions exhibits full specific activity and is found exclusively in soluble form.


dehydrogenase (LL-2-hydroxyisocaproate HicDH; Schiltte et al. 1984) can be applied in an industrial process for the production of L-aa. The Correspondence to: Dr. J. Collins, Department of Genetics, GBF, Mascheroder Weg 1, D-3300 Braunschweig (F.R.G.) Tel. [0531]6181-200; Fax [0531]6181202. Abbreviations: aa, amino acid(s); bp, base pair(s); DSM, Deutsche Sammhmg von Mikroorganismen und Zellkulturen; ExoIII, bacteriophage 1 exonuclease III; FPLC, fast protein liquid chromatography; HicDH, 2-hydroxyisocaproate dehydro0378-l 119/89/$03.50


1989 Elsevier

Science Publishers

B.V. (Biomedical

tetrameric, NAD(H)-dependent enzyme catalyses the reversible and stereospecific interconversion between 2-ketocarboxylic acids and L-2-hydroxycarboxylic acids. Although L-2-hydroxyisocaproate is the best substrate for the enzyme, it accepts a genase; kb, kilobase or 1000 bp; LDH, lactate dehydrogenase; nt, nucleotide(s); oligo, oligodeoxyribonucleotide; ORF, open reading frame; PAGE, polyacrylamide-gel electrophoresis; 1 prpR, phage 3. ‘leftward’ and ‘rightward’ major early promoters; RBS, ribosome-binding site; RT, reverse transcriptase; SDS, sodium dodecyl sulfate; SSC, 0.15 M NaCl/0.015 M Na, . citrate pH 7.6; rsp, transcription start point; u, units; [ 1, denotes plasmid-carrier state. Division)


broad range of 2-ketocarboxylic acids or corresponding L-2-hydroxycarboxylic acids as substrates (Schutte et al., 1984). L-lactate is a poor substrate for the enzyme. 2-Ketoacids with medium chain length (live to six C-atoms) are the best substrates and 2-ketoacids with a longer carbon backbone are not accepted. The enzyme can be used in combination with D-HicDH (Hummel et al., 1985) and L-leucine dehydrogenase (Schiitte et al. 1985a) in an enzyme membrane reactor, to convert a racemic mixture of 2-hydroxycarboxylic acids and ammonia to L-aa (Wandrey et al., 1984). The coenzyme NAD(H) is present in the reactor in a polyethyleneglycol-bound form and is continuously regenerated during the reaction cycle. In contrast to D-HicDH, which is found in a number of Lactobacillus strains (Schiitte et al., 1985b; Hummel et al., 1985), only one Lactobacillus strain is known to produce an L-HicDH, namely L. confisus DSM20196. Enzyme production in this strain is very low (about 0.5 u/mg protein in the crude extract). Our results in cloning, sequencing and expression of the L-HicDH gene are reported here. We have recently reported the cloning and expression in Escherichia coli of D-HicDH from L. casei ssp. the absence of pseudoplantarum in which AGG/AGA arginine codons in Lactobacillus genes was noted (Lerch et al., 1989).


(a) Materials,


strains and plasmids

Enzymes for molecular cloning and other materials were purchased from the same sources as described by Lerch et al. (1989) for the cloning of the D-HicDH gene of L. casei. Chromosomal DNA was isolated from L. confusus DSM20196. Other strains, plasmids and media were as described (Lerch et al., 1989). Plasmids pGEM3 and pGEM4 were obtained from Promega Biotec, Madison, WI. (b) Raising antiserum against L-HicDH and synthesis of an oligo probe for screening

L-HicDH was purified according to Schutte et al. (1984). After additional purification steps (interfacial

salting-out chromatography on Sepharose 4B and FPLC on MonoQ) homogeneous material was obtained. Intramuscular injections of 600 pg purified L-HicDH plus Freund’s complete adjuvants, followed two weeks later by 200 pg r_-HicDH plus incomplete adjuvants was used to raise specific anti-L -HicDH antibodies in a rabbit. The N-terminal aa sequence (Tsai et al., 1987) has been published. Based on the sequence of the first 25 aa, a 69-mer oligo probe was synthesised using the most frequent codon usage in the Staphylococcus hyicus lipaseencoding gene (G&z et al., 1985). The codon usage of this Gram-positive bacterial gene was used, since at that time no Lactobacillus gene sequence was known to us. (c) Restriction mapping of the L-HicDH gene on the Lactobacillus confusus genome

The 69-mer oligo was used as a probe in Southemblot experiments, to establish a physical map of the genomic region around the L-HicDH. Chromosomal DNA from L. confwus was cleaved with one or two restriction enzymes and subjected to 0.8% agarose gel electrophoresis. After Southern transfer, filters were hybridised with the 32P-labelled 69-mer oligo in 6 x SSC at 50°C. Each lane exhibited one stronglyhybridising band and four minor bands. It was assumed that the strong band represents L-HicDH. From the sizes of the hybridising fragments, a restriction map of the region containing L-HicDH was constructed and used for specilic fragment enrichment. This map is also correlated with the restriction map of the cloned gene (Fig. 1). (d) Cloning of the L-HicDH



in mind the difficulties in cloning D-HicDH from L. casei (Lerch et al., 1989), we chose the same strategy to clone L-HicDH by the fragment-enrichment approach. A number of L-LDH genes from Gram-positive bacteria had also been cloned by this method (Barstow et al., 1986; 1987; Ztilli et al., 1987). L. confusus chromosomal DNA (1 mg) was cleaved to completion with C/a1 and fragments of about 5.6 f 0.8 kb were isolated from a preparative agarose gel by electroelution. This ClaI-cleaved fraction was digested with EcoRI; fragments 4.6 kb




.. ..






Hind III Hmd III Sal1 BomHl EcoRV Clal EcoRV EcoRl Cl01 Salt Hpol EcoRl BamHl a... ... _.A_.. ..** / ..Q...._ / _...... /.’ / *... . .. ..__ ._.. .A._ ....

BstXlBstEll Dralll / Nrul Nael Htndlll Htndlll Nrul Sphl BomHl Sal1 Bell Pstl Cl0 I,,_..~~~~l~ A:. ............... ..


*.._.. *..... a... *..... a... Q.. .*-%._ . . . .._-... Hpal



L-Hi&H gene








Fig. 1. Restriction maps of the genomic region containing the L-HicDH gene (A and B) and sequencing strategy (C). Nested unidirectional deletions in pGEM subclones were introduced in both directions according to Her&off (1984). Alkali-denatured DNA (Chen and Seeburg, 1985) from minipreps was sequenced (Sanger et al., 1977) with [a-35S]dATP and Sequenase, a modified bacteriophage T7 DNA polymerase (Tabor and Richardson, 1987) and oligo primers. Oligos were synthesised as described by Frank et al. (1988). Samples ofthe sequencing reactions were run on a wedge-shaped (0.2-0.6 mm) 6% polyacrylamide gel. To make unidirectional deletions into L-ZficDH, the cloned DNA fragment was cut out from pHLS with CluI + PvaII (PvuII site is in the pBR322 vector) and inserted into AccI + PvuII-cleaved pGEM4, giving clone pHL7. For the deletions in the opposite direction, pHL5 was digested with PvuI + PvuII (both enzymes cutting in the vector at the left and the right of the cloned L. confusus DNA) and the fragment with L-HicDH was cloned into PvuI + HincII-cleaved pGEM3; the resulting clone is pHL24. pHL7 and pHL24 were both digested with XboI + SsrI in the polylinker region, treated with ExoIII and mung-bean nuclease and two sets of overlapping deletion clones were prepared according to Henikoff (1984). Then supercoiled DNA from these clones was sequenced. General cloning techniques used were the same as described for the cloning of the D-HicDH gene from L. cusei (Lerch et al. 1989).

in size were similarly purified from a gel. The final purified ClaI-EcoRI fraction, upon cleavage with HpaI, yielded five different fragments as seen. Attempts to clone the highly enriched CM-EcoRI fragment into pBR322 were unsuccessful. A smaller 2.0-kb BumHI-HpaI fragment could be cloned into BumHI + PvuII-cleaved pBR322, but contained only part of the L-HicDH gene as seen by sequencing from the BumHI site. The entire gene was obtained intact as follows: the fraction of 4.6-kb EcoRI-C/u1 fragments was digested with HpuI and a 2.9-kb CM-HpaI fragment cloned into ClaI + NruIcleaved pBR322. This cloned plasmid was designated pHL5.

(e) Gene sequence The complete nt sequence of L-HicDH and the sequencing strategy are outlined in Figs. 1 and 2. The L-HicDH structural gene consists of an ORF spanning 930 bp, starting with an ATG start codon and ending with a TAA stop codon. The 69-mer oligo matches the genomic sequence to 80 y0 (55 out of 69), with the longest continuous homologous stretch being 14 nt. Although the structural gene starts with ATG, no methionine was found as N-terminal aa during the aa sequencing. It may be possible that the protein is processed by a methionine-specific aminopeptidase. Besides this, no dis-















































Fig. 2. Complete nucleotide sequence of the L-HicDH gene, surrounding regions and deduced ORFs. The S’-end of the mRNA, -10 and -35 regions ofthe promoter, putative RBS, and inverted repeats are indicated. An asterisk indicates a stop codon. The end-points of the sequences in the deletion clones d-28 and 4-326 are shown. These deletions were derived from pHL7, as described in the legend to Fig. 1. A putative promoter sequence (nt 1260-1320) and Shine-Dalgarno sequence (nt 1319-1324) precede the lower ORF.

crepancies were found between the 25 sequenced N-terminal aa and the genomic sequence. The codon usage of L-HicDH is similar to other Lactobacillus genes, in that Arg codons AGG/A are not used (Lerch et al., 1989). (f) Promoter sequence The L-HicDH structural gene is preceded by a putative RBS containing GGGGG. The putative

Pig. 3. Mapping of the 5’-end of the L-HicDH mRNA. RNA was isolated from L. con&w as previously described for L. casei (Lerch et al., 1989).Total RNA from E. coli was prepared according to Maniatis et al. (1982). The 5’-end of the L-HicDH gene transcript was mapped by primer extension (Williams and Mason, 1985). A 22-mer oligo complements to aa positions 334-355 ofthe gene sequence (Fig. 2) was used as primer. Shown is an autoradiogram of primer extension mapping of the 5’-end of the L-HicDH mRNA (in E. coli). Lanes: 1, elongated oligo synthesised by RT after hybridisation of a 22-mer (32P-labelled) with RNA from E. coli DH 1[pHL5]. The dideoxy-sequencing reactions on denatured supercoiled pHL5 are run in lanes A, T, G and C. The same primer was used for both the RT ‘elongation’ and the sequencing reactions; 2, oligo control without elongation with RT. Arrowheads mark the positions of the elongation products in the autoradiogram.

-35 region of the promoter (TGGTGA) is situated 17 bp upstream from a typical TATA-box, which is itself located at -10 relative to the actual tsp measured in L. confusus and E. coli. Fig. 3 shows the results of primer extension on L-HicDH mRNA from E. eoli. The tsp was detected at the same position in L. confusu (not shown). The L-HicDH promoter lacks a poly(A) region at -45, which is present in most promoters from Gram-positive bacteria (Graves and Rabinowitz, 1986). The -35 module of the promoter is located in a region of inverted repeats, which may form a secondary structure similar to a Rho-independent terminator (Fig. 2), perhaps terminating the transcript of the ORF located upstream from L-HicDH.


(h) Primary sequence comparison of L-HicDH to other dehydrogenases

(g) Terminator sequence Downstream from the L-HicDH structural gene there is a region of inverted repeats, 37 bp in length, which may form a secondary structure typical for Rho-independent terminators of E. coli (Friedman et al., 1987). The content (Fig. 3) of GC pairs of the stem (four out of 15) is low, as is the free energy of formation (dG = -5.8 kcal). Although the 3’ end of the L-HicDH transcript has not been mapped, the 5’ end is known from the primer extension experiment (Fig. 3) and the total length (about 1 kb) of the mRNA from clone pHL5 could be deduced from a Northern blot (not shown). It is, therefore, probable that this terminator is used at least in E. coli.

1 : :

Dogflsh H4 t-tDH T. caldophllus t-LDH E. stearathermophilus L. casei t-LDH L. confusus L-HicDH


%I lQhGsLfLhtakiVs&jVsv IlmpfAh-pVwVrAgsVgd fNhGkVfApkpVdIwhgDVdd 1s IpftspkkiysA-EVsd fQ&AnLeahGnIvIn~aa

: :


1 : :

: 3 4 5

+ + 209




: 3 4 5


: : 5 110

I : 4 5


tLwdiqkDLk-f iLkeaafaLg-f tLkrv1arAftr qtkkv1tDAfakndietrq yIqqrfdEIvdt1

Fig. 4. Alignment of the aa sequences of L-LDH from dogfish muscle (Taylor, 1977), T. caldophilus (Kunai et al., 1986), B. stearothemophilus ([email protected] et al., 1987) and L. casei (Hensel et al., 1983) and L-HicDH from L. conjuus according to Eventoff et al. (1977), Stangl et al. (1987) and Kunai et al. (1986). Numhering of aa residues follows the N-system of Eventoff et al. (1977). Identical aa are boxed, homologous aa are in upper-case letters. Regions binding NAD are indicated by ’ + ‘. Substrate-binding regions are designated by small circles.

A computer search in the EMBL/SWISSPROT data base (version 14) for similarities of the aa sequence of L-HicDH to other known sequences, showed that the L-HicDH sequence resembles the sequences of both eukaryotic and prokaryotic L-LDHs. An alignment of the L-HicDH and L-LDH sequences is presented in Fig. 4. The sequence of the L-HicDH is 30% identical to the sequence of the L-LDH from L. casei. Sequence similarity with the eukaryotic L-LDH from spiny dogfish is 24 2. Like all the prokaryotic L-LDHs, L-HicDH lacks the N-terminal 12-15 aa present in the eukaryotic enzymes, which are in part responsible for the subunit-to-subunit interaction at the R-axes of the tetrameric enzymes from eukaryotes (Holbrook et al., 1975). It seems likely that L-HicDH (L. confusus) and L-LDH (L. casei) genes have evolved from a common precursor. Since, however, strong conservation of structural features is found amongst the dehydrogenases in general (cf., spiny dogfish and Lactobacillus) no preference can be made for a decision between divergence or convergence at the evolutionary level. We note that L. confusus has no L-LDH (it has an unrelated D-LDH; Hummel et al., 1983) and may have thus been free to evolve a new, and extremely rare function. Starting with the sequence alignment in Fig. 4, we compared the aa residues which are involved in the binding of the coenzyme and the substrate (Fig. 4). Most of the coenzyme-binding residues of r_-HicDH are in good agreement with those of the L-LDHs, although there are some conservative exchanges, e.g., Ile-Val transitions. The three substrate-binding aa, Arg-109, Arg-171 and His-195, are identical in L-HicDH and the L-LDHs. One may conclude from all these data that the polypeptide backbone of L-HicDH is very similar to that of the L-LDHs and that the chemical reaction mechanism of the dehydrogenation may be the same. His-188, which is highly conserved in both the bacterial and vertebrate L-LDHs, is replaced by Ser in L-HicDH. It was shown that His-188 is at the binding site of the allosteric effector fructose 1,6-bisphosphate for the bacterial enzymes (Hensel et al., 1983; Clarke et al., 1985). In the case of the verte-


brate enzymes, citrate and other anions can bind at this position (Adams et al., 1973). Indeed, L-HicDH activity is not regulated by fructose 1,6-bisphosphate (H. SchUtte, W. Hummel and M.-R. Kula, pers. communication). (i) Expression of the L-HicDH gene in Escherichia

coli The expression of L-HicDH in several clones was investigated by SDS-PAGE and Western blot (Fig. 5), as well as by assaying the enzyme activity in the soluble cell extract. The initial plasmid construct, pHL5, contained L-HicDH on a 2.9-kb ClaI-HpaI fragment in pBR322 and directed the synthesis of L-HicDH up to 6% of soluble cell protein in E. coli. The insert bearing L-HicDH was cut out from A-28 with EcoRI and cloned into the EcoRI-cut expression vector pJLA501 (Schauder et al., 1987), so that L-HicDH is under control of the 1 pgR (clone pHL6). Surprisingly, L-HicDH expression is rather high in pHL5, whereas A-326 and A-28 ExoIII-deletion clones from pHL7 (see legend to Fig. 1) produce



Fig. 5. Expression of the L-Hi&H gene. (A) 0.1% SDS-12.5% PAGE (stained with Coomassie blue) of crude extract; (B) Western blot. Specific L-HicDH activity in the crude extract is given below (in parentheses). Pure L-HicDH has a specific activity of 650 u/mg. Lanes: 1, L. conjiuus (0.88 u/mg); 2, clone pHL6 after heat induction (244 u/mg); 3, clone pHL6 before heat induction (0.0 u/mg); 4, pHL5 (50 u/mg); 5, clone A-326 (0.5 u/mg); 6, clone A-28 (1.3 u/mg). The arrows indicate the position of L-HicDH. L-HicDH activity was measured according to SchUtte et al. (1984).

much less enzyme. Since A-28 lacks the promoter of L-HicDH, a tsp may be present somewhere in the pGEM3 vector. The amount of mRNA present in the cells is in agreement with the enzyme levels. Under control of I pgR L-HicDH can be produced up to more than 35 % of total cell protein. The enzyme is soluble and activity correlates with the amount of enzyme observed assuming that it has normal full specific activity. Formation of inclusion bodies was not observed. (j) Conclusions

(1) L-HicDH was successfully cloned by employment of the fragment-enrichment strategy, or ‘forced cloning, which has been useful in the cloning of D-HicDH and other genes from Gram-positive organisms. The codon usage of L-HicDH corresponded to that of other Lactobacillus genes, i.e., the Arg codons AGA/AGG were not used. (2) L-HicDH is expressed at a very low level in its natural host L. confi~.~. In E. coli, 50-fold higher expression of L-HicDH was found in the original clone, in which more than 1 kb of the original upstream sequence is also present. Since removal of this upstream region gave a 50-fold loss of expression, it may be possible that an activator protein is encoded in the ORF upstream of L-HicDH. Under control of lprpR in the expression vector pJLA501, L-HicDH could be overproduced despite the fact that its RBS is not optimal. In contrast to D-HicDH, L-HicDH remains soluble and fully active upon overexpression to 35% of total cell protein. (3) By comparison of the aa sequence of r_-HicDH and other dehydrogenase sequences, it became obvious that L-HicDH is closely related to the well known L-LDHs. The highest overall sequence similarity (30%) was found with the L-LDH from L. casei. There is no similarity in sequence between L-HicDH from L. confusus and D-HicDH from L. casei; these two genes may therefore have evolved from different ancestors. The aa residues involved in the binding of coenzyme and substrate are highly conserved. It is, therefore, very likely that the polypeptide backbone and the dehydrogenation reaction mechanism of L-HicDH are the same as for the L-LDHs. These findings provide a good basis for further studies of the structure-function relationship of L-HicDH as part of a protein design project.



We thank Dr. H.L. Paul (BBA, Braunschweig) for his help in immunising the rabbits. We thank Dr. H. Tsai for his interest in our work, Dr. J.E.G. McCarthy for the pJLA501 vector and B&be1 Seeger-Kunth and Sabine Peters for secretarial assistance.

REFERENCES Adams, M.J., Liljas, A. and Rossmann, M.G.: Functional anion binding sites in dogfish M4 lactate dehydrogenase. J. Mol. Biol. 76 (1973) 519-531. Barstow, D.A., Clarke, A.R., Chia, W.N., Wigley, D., Sharman, A.F., Holbrook, J.J., Atkinson, T. and Minton, N.P.: Cloning, expression and complete nucleotide sequence of the BacilZus stearothemophilus L-lactate dehydrogenase gene. Gene 46 (1986) 47-55. Barstow, D.A., Murphy, J.P., Sharman, A.F., Clarke, A.R., Holbrook, J.J. and Atkinson, T.: Amino acid sequence of the L-lactate dehydrogenase of Bacillw caldotenax deduced from the nucleotide sequence of the cloned gene. Eur. J. Biochem. 165 (1987) 581-586. Chen, E.Y. and Seeburg, P.H.: Supercoil sequencing: a fast and simple method for sequencing plasmid DNA. DNA 4 (1985) 165-170. Clarke, A.R., Atkinson, T., Campbell, J.W. and Holbrook, J.J.: The assembly mechanism of the lactate dehydrogenase tetramer from Bacillus steurothemophilus; the equilibrium relationship between quatemary structure and the binding of fructose 1,6-bisphosphate NADH and oxamat. Biochim. Biophys. Acta 829 (1985) 387-396. Eventoff, W., Rossmann, M.G., Taylor, S.S., Torff, H.-J., Meyer, H., Keil, W. and Kiltz, H.-H.: Structural adaptions of lactate dehydrogenase isoenzymes. Proc. Natl. Acad. Sci. USA 74 (1977) 2677-2681. Frank, R., Meyerhans, A., Schwellnus, K. and Blocker, H.: Simultaneous synthesis and biological applications of DNA fragments: an efficient and complete methodology. Methods Enzymol. 154 (1988) 201-249. Friedman, P.J., Imperiale, M.J. and Adhya, S.L.: RNA 3’-end formation in the control of gene expression. AMU. Rev. Genet. 21 (1987) 453-488. G&z, F., Popp, F., Kom, E. and Schleifer, K.H.: Complete nucleotide sequence of the lipase gene from Staphylococcus hyicus cloned in Staphylococcus carnosus. Nucleic Acids Res. 13 (1985) 5895-5906. Graves, M.C. and Rabinowitz, J.C.: In vivo and in vitro transcription of the Clostridium pasteurianum ferredoxin gene. J. Biol. Chem. 261 (1986) 11409-11415. Henikoff, S.: Unidirectional digestion with exonuclease III creates targeted breakpoints for DNA sequencing. Gene 28 (1984) 351-359.

Hensel, R., Mayr, U. and Woenckhaus, C.: Affinity labelling of the allosteric site of the L-lactate dehydrogenase of Lactobucillus casei. Eur. J. B&hem. 135 (1983) 359-365. Hensel, R., Mayr, U. and Yang, C.-Y.: The complete primary structure of the allosteric L-lactate dehydrogenase from Lactobacillur casei. Eur. J. Biochem. 134 (1983) 503-511. Holbrook, J.J., Liljas, A., Steindel, J.S. and Rossmann, M.G.: Lactate dehydrogenase. In Boyer, P.D. (Ed.), The Enzymes, 3rd ed., Vol. 11. Academic Press, New York, 1975, pp. 191-292. Hummel, W., Schiitte, H. and Kula, M.-R.: Large scale production of D-hCtatC dehydrogenase for the stereospecific reduction of pyruvate and phenylpyruvate. Eur. J. Appl. Microbial. Biotechnol. 18 (1983) 75-85. Hummel, W., Schtitte, H. and Kula, M.-R.: D-2-Hydroxyisocaproate dehydrogenase from Lactobacilius casei - a new enzyme suitable for stereospecific reduction of 2-ketocarboxylic acids. Appl. Microbial. Biotechnol. 21 (1985) 7-15. Kunai, K., Machida, M., Matsuzawa, H. and Ohta, T.: Nucleotide sequence and characteristics of the gene for L-lactate dehydrogenase of Thermus caldophilus GK24 and the deduced amino acid sequence of the enzyme. Eur. J. Biochem. 160 (1986) 433-440. Lerch, H.-P., Blocker, H., Kalwass, H., Hoppe, J., Tsai, H. and Collins, J.: Cloning, sequencing and expression in Escherichiu coli of the D-2-hydroxyisocaproate dehydrogenase gene of Lactobacillw casei. Gene 78 (1989) 47-57. Maniatis, T., Fritsch, E.F. and Sambrook, J.: Molecular Cloning. A Laboratory Manual Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, 1982. Sanger, F., Nicklen, S. and Coulson, A.R.: DNA sequencing with chain-terminating inhibitors. Proc. Natl. Acad. Sci. USA 74 (1977) 5463-5467. Schauder, B., Blocker, H., Frank, R. and McCarthy, J.E.G.: Inducible expression vectors incorporating the Escherichia coli atpE translation initiation region. Gene 52 (1987) 279-283. Schtltte, H., Hummel, W. and Kula, M.-R.: L-2-Hydroxyisocaproate dehydrogenase - a new enzyme from Lactobacillus conjkus for the stereospecilic reduction of 2-ketocarboxylic acids. Appl. Microbial. Biotechnol. 19 (1984) 167-176. Schtitte, H., Hummel, W. and Kula, M.-R.: Improved enzyme screening by automated fast protein liquid chromatography. Anal. B&hem. 151 (1985a) 547-553. Schtitte, H., Hummel, W., Tsai, H. and Kula, M.-R.: L-Leucine dehydrogenase from Bacillus cereus - production, large scale purification and protein characterization. Appl. Microbial. Biotechnol. 22 (1985b) 306-317. Stangl, D., Wiederkehr, F., Suter, F. and Zuber, H.: The complete amino acid sequence of the mesophilic L-lactate dehydrogenase from Bacillus megaterium. Biol. Chem. Hoppe-Seyler 368 (1987) 1157-1166. Tabor, S. and Richardson, C.C.: DNA sequence analysis with a modified bacteriophage T7 DNA polymerase. Proc. Natl. Acad. Sci. USA 84 (1987) 4767-4771. Taylor, S.S.: Amino acid sequence of dog&h muscle lactate dehydrogenase. J. Biol. Chem. 252 (1977) 1799-1806.

270 Tsai, H., Lerch, H.-P., Kalwass, H., Schiitte, H., Hoppe, J. and Collins, J.: The D- and L-2-hydroxyisocaproate dehydrogenases from Lactobacilli are structurally and immunologically unrelated enzymes. In Neijssel, O.M., Van der Meer, R.R. and Luyben K.C.A.M. (Eds.), Proc. 4th Eur. Congr. Biotechnol., Vol. 2. Elsevier, Amsterdam, 1987, pp. 228-231. Wandrey, C., Fiolitakis, E., Wichmann, U. and Kula, M.-R.: L-Amino acids from a racemic mixture of a-hydroxy acids. Ann. N.Y. Acad. Sci. 434 (1984) 91-94.

Williams, J.G. and Mason, P.J.: Hybridisation in the analysis of RNA. In Hames, B.D. and Higgins, S.J. (Eds.), Nucleic Acid Hybridisation: A Practical Approach. IRL Press, Oxford, 1985, pp. 139-160. Zlilli, F., Weber, H. and Zuber, H.: Nucleotide sequences of lactate dehydrogenase genes from the thermophilic bacteria Bacillus stearothermophilus, B. caldolyticus and B. caldotenax. Hoppe-Seyler J. Biol. Chem. 368 (1987) 1167-l 177.