Cloning of genes coding for L-sorbose and L-sorbosone dehydrogenases from Gluconobacter oxydans and microbial production of 2-keto-L-gulonate, a precursor of L-ascorbic acid, in a recombinant G. oxydans strain.

We have purified L-sorbose dehydrogenase (SDH) and L-sorbosone dehydrogenase (SNDH) from Gluconobacter oxydans T-100 that showed an ability to convert D-sorbitol to 2-keto-L-gulonate (2-KLGA). A genomic library of Gluconobacter oxydans T-100 was screened with a probe, a 180-bp PCR product which was obtained from degenerate oligodeoxyribonucleotides based on the elucidated sequence of the purified SDH (used as primers) and the genomic DNA of G. oxydans T-100 (used as a template). From sequencing of the DNA from a clone positive to the probe, the SNDH and the SDH were estimated to be coded in sequential open reading frames with 1,497 and 1,599 nucleotides, respectively, which was confirmed by expression of the DNA in Escherichia coli that showed both enzymatic activities. The DNA was introduced to a shuttle vector which was prepared from a plasmid of G. oxydans T-100 and pHSG298 to obtain an expression vector designated pSDH155. The production of 2-KLGA by pSDH155 in G. oxydans G624, an L-sorbose-accumulating strain, was improved to 230% compared to that of G. oxydans T-100. Chemical mutation of the host strain to suppress the L-idonate pathway and replacement of the original promoter with that of E. coli tufB resulted in improving the production of 2-KLGA. Consequently, high-level production from D-sorbitol to 2-KLGA (130 mg/ml) was achieved by simple fermentation of the recombinant Gluconobacter.     

Cloning and nucleotide sequencing of the membrane-bound L-sorbosone dehydrogenase gene of Acetobacter liquefaciens IFO 12258 and its expression in Gluconobacter oxydans.

Cloning and expression of the gene encoding Acetobacter liquefaciens IFO 12258 membrane-bound L-sorbosone dehydrogenase (SNDH) were studied. A genomic library of A. liquefaciens IFO 12258 was constructed with the mobilizable cosmid vector pVK102 (mob+) in Escherichia coli S17-1 (Tra+). The library was transferred by conjugal mating into Gluconobacter oxydans OX4, a mutant of G. oxydans IFO 3293 that accumulates L-sorbosone in the presence of L-sorbose. The transconjugants were screened for SNDH activity by performing a direct expression assay. One clone harboring plasmid p7A6 converted L-sorbosone to 2-keto-L-gulonic acid (2KGA) more rapidly than its host did and also converted L-sorbose to 2KGA with no accumulation of L-sorbosone. The insert (25 kb) of p7A6 was shortened to a 3.1-kb fragment, in which one open reading frame (1,347 bp) was found and was shown to encode a polypeptide with a molecular weight of 48,222. The SNDH gene was introduced into the 2KGA-producing strain G. oxydans IFO 3293 and its derivatives, which contained membrane-bound L-sorbose dehydrogenase. The cloned SNDH was correctly located in the membrane of the host. The membrane fraction of the clone exhibited almost stoichiometric formation of 2KGA from L-sorbosone and L-sorbose. Resting cells of the clones produced 2KGA very efficiently from L-sorbosone and L-sorbose, but not from D-sorbitol; the conversion yield from L-sorbosone was improved from approximately 25 to 83%, whereas the yield from L-sorbose was increased from 68 to 81%. Under fermentation conditions, cloning did not obviously improve the yield of 2KGA from L-sorbose.     

Pyrroloquinoline quinone-dependent dehydrogenases from Ketogulonicigenium vulgare catalyze the direct conversion of L-sorbosone to L-ascorbic acid

A novel enzyme, L-sorbosone dehydrogenase 1 (SNDH1), which directly converts L-sorbosone to L-ascorbic acid (L-AA), was isolated from Ketogulonicigenium vulgare DSM 4025 and characterized. This enzyme was a homooligomer of 75-kDa subunits containing pyrroloquinoline quinone (PQQ) and heme c as the prosthetic groups. Two isozymes of SNDH, SNDH2 consisting of 75-kDa and 55-kDa subunits and SNDH3 consisting of 55-kDa subunits, were also purified from the bacterium. All of the SNDHs produced L-AA, as well as 2-keto-L-gulonic acid (2KGA), from L-sorbosone, suggesting that tautomerization of L-sorbosone causes the dual conversion by SNDHs. The sndH gene coding for SNDH1 was isolated and analyzed. The N-terminal four-fifths of the SNDH amino acid sequence exhibited 40% identity to the sequence of a soluble quinoprotein glucose dehydrogenase from Acinetobacter calcoaceticus. The C-terminal one-fifth of the sequence exhibited similarity to a c-type cytochrome with a heme-binding motif. A lysate of Escherichia coli cells expressing sndH exhibited SNDH activity in the presence of PQQ and CaCl2. Gene disruption analysis of K. vulgare indicated that all of the SNDH proteins are encoded by the sndH gene. The 55-kDa subunit was derived from the 75-kDa subunit, as indicated by cleavage of the C-terminal domain in the bacterial cells.     

Sequence analysis of the Gluconobacter oxydans RecA protein and construction of a recA-deficient mutant.

The deduced amino acid sequence of Gluconobacter oxydans RecA protein shows 75.2, 69.4, and 66.2% homology with those from Aquaspirillum magnetotacticum, Escherichia coli, and Pseudomonas aeruginosa, respectively. The amino acid residues essential for function of the recombinase, protease, and ATPase in E. coli recA protein are conserved in G. oxydans. Of 24 amino acid residues believed to be the ATP binding domain of E. coli RecA, 17 are found to be identical in G. oxydans RecA. Interestingly, nucleotide sequence alignment between the SOS box of G. orphans recA gene and those from different microorganisms revealed that all the DNA sequences examined have dyad symmetry that can form a stem-loop structure. A G. oxydans recA-deficient mutant (LCC96) was created by allelic exchange using the cloned recA gene that had been insertionally inactivated by a kanamycin-resistance cassette. Such replacement of the wild-type recA with a kanamycin resistance gene in the chromosome was further verified by Southern hybridization. Phenotypically, the recA-deficient mutant is significantly more sensitive to UV irradiation than the wild-type strain, suggesting that the recA gene of G. oxydans ATCC9324 plays a role in repairing DNA damage caused by UV irradiation. Moreover, the mutant strain is much more plasmid transformable than its parent strain, illustrating that G. oxydans LCC96 could be used as a host to take up the recombinant plasmid for gene manipulation.     

Coenzyme specificity of enzymes in the oxidative pentose phosphate pathway of Gluconobacter oxydans

The coenzyme specificity of enzymes in the oxidative pentose phosphate pathway of Gluconobacter oxydans was investigated. By investigation of the activities of glucose-6-phosphate dehydrogenase (G6PDH) and 6-phosphogluconate dehydrogenase (6PGDH) in the soluble fraction of G. oxydans, and cloning and expression of genes in Escherichia coli, it was found that both G6PDH and 6PGDH have NAD/NADP dual coenzyme specificities. It was suggested that the pentose phosphate pathway is responsible for NADH regeneration in G. oxydans.     

Construction and co-expression of plasmid encoding xylitol dehydrogenase and a cofactor regeneration enzyme for the production of xylitol from D-arabitol

The biotransformation of D-arabitol into xylitol was investigated with focus on the conversion of D-xylulose into xylitol. This critical conversion was accomplished using Escherichia coli to co-express a xylitol dehydrogenase gene from Gluconobacter oxydans and a cofactor regeneration enzyme gene which was a glucose dehydrogenase gene from Bacillus subtilis for system 1 and an alcohol dehydrogenase gene from G. oxydans for system 2. Both systems efficiently converted D-xylulose into xylitol without the addition of expensive NADH. Approximately 26.91 g/L xylitol was obtained from around 30 g/L D-xylulose within system 1 (E. coli Rosetta/Duet-xdh-gdh), with a 92% conversion yield, somewhat higher than that of system 2 (E. coli Rosetta/Duet-xdh-adh, 24.9 g/L, 85.2%). The xylitol yields for both systems were more than 3-fold higher compared to that of the G. oxydans NH-10 cells (7.32 g/L). The total turnover number (TTN), defined as the number of moles of xylitol formed per mole of NAD(+), was 32,100 for system 1 and 17,600 for system 2. Compared with that of G. oxydans NH-10, the TTN increased by 21-fold for system 1 and 11-fold for system 2, hence, the co-expression systems greatly enhanced the NADH supply for the conversion, benefiting the practical synthesis of xylitol.     

利用山梨糖脱氢酶活性筛选氧化葡糖杆菌PQQ合成基因簇

【目的】利用山梨糖脱氢酶醌酶活性从氧化葡糖杆菌H24中分离PQQ生物合成基因簇。【方法】利用ptsG位点整合sdh基因的大肠杆菌JM109作为宿主菌构建了氧化葡糖杆菌H24的基因组DNA文库。通过山梨糖脱氢酶活性检测,从文库中筛选具有PQQ合成能力的单菌落并进行亚克隆。【结果】从氧化葡糖杆菌H24的基因组文库中筛选得到一株具有山梨糖脱氢酶活性的单菌落,亚克隆后序列分析显示插入片段全长5400bp,对应5个编码框(pqqABCDE),与其他细菌PQQ生物合成基因簇有很高的序列同源性。【结论】利用山梨糖脱氢酶醌酶活性成功从氧化葡糖杆菌H24中分离克隆得到了PQQ生物合成基因簇pqqABCDE。     

Cloning of the xylitol dehydrogenase gene from Gluconobacter oxydans and improved production of xylitol from D-arabitol

Xylitol dehydrogenase (XDH) was purified from the cytoplasmic fraction of Gluconobacter oxydans ATCC 621. The purified enzyme reduced D-xylulose to xylitol in the presence of NADH with an optimum pH of around 5.0. Based on the determined NH2-terminal amino acid sequence, the gene encoding xdh was cloned, and its identity was confirmed by expression in Escherichia coli. The xdh gene encodes a polypeptide composed of 262 amino acid residues, with an estimated molecular mass of 27.8 kDa. The deduced amino acid sequence suggested that the enzyme belongs to the short-chain dehydrogenase/reductase family. Expression plasmids for the xdh gene were constructed and used to produce recombinant strains of G. oxydans that had up to 11-fold greater XDH activity than the wild-type strain. When used in the production of xylitol from D-arabitol under controlled aeration and pH conditions, the strain harboring the xdh expression plasmids produced 57 g/l xylitol from 225 g/l D-arabitol, whereas the control strain produced 27 g/l xylitol. These results demonstrated that increasing XDH activity in G. oxydans improved xylitol productivity.     

Transaldolase/glucose-6-phosphate isomerase bifunctional enzyme and ribulokinase as factors to increase xylitol production from D-arabitol in Gluconobacter oxydans

Xylitol production from D-arabitol by the membrane and soluble fractions of Gluconobacter oxydans was investigated. Two proteins in the soluble fraction were found to have the ability to increase xylitol production. Both of these xylitol-increasing factors were purified, and on the basis of their NH(2)-terminal amino acid sequences the genes encoding both of the factors were cloned. Expression of the cloned genes in Escherichia coli showed that one of the xylitol-increasing factors is the bifunctional enzyme transaldolase/glucose-6-phosphate isomerase, and the other is ribulokinase. Using membrane and soluble fractions of G. oxydans, 3.8 g/l of xylitol were produced from 10 g/l D-arabitol after incubation for 40 h, and addition of purified recombinant transaldolase/glucose-6-phosphate isomerase or ribulokinase increased xylitol to 5.4 g/l respectively, confirming the identity of the xylitol-increasing factors.     

氧化葡糖杆菌sndh-sdh基因簇的克隆表达

目的:从氧化葡糖杆菌H763中克隆sndh-sdh基因簇,在大肠杆菌和氧化葡糖杆菌621H中分别表达山梨酮脱氢酶-山梨糖脱氢酶(SNDH-SDH),并检测其活性。方法与结果:以氧化葡糖杆菌H763基因组DNA为模板,PCR扩增包括启动子、结构基因及终止序列在内的sndh-sdh基因簇,回收3533 bp的扩增产物,连入pMD18T载体,转化至大肠杆菌DH5α中表达;以山梨糖或木糖为底物,DCIP法检测菌体裂解液,DCIP检测液颜色由蓝绿色变为黄色,表明大肠杆菌表达产物具有脱氢酶活性。构建pBBR1MCS2-sndh-sdh载体,通过接合转移导入氧化葡糖杆菌621H,重组葡糖杆菌在以山梨醇或山梨糖为底物的培养基中培养,采用薄层层析检测法检测其培养上清中的代谢产物,层析板上显示了2-酮基-L-古龙酸斑点。结论:重组大肠杆菌DH5α和氧化葡糖杆菌621H中均表达了有脱氢酶活性的SNDH-SDH。     

Nucleotide sequence encoding the flavoprotein and iron-sulfur protein subunits of the Bacillus subtilis PY79 succinate dehydrogenase complex.

The nucleotide sequence of a 2.7-kilobase segment of DNA containing the sdhA and sdhB genes encoding the flavoprotein (Fp, sdhA) and iron-sulfur protein (Ip, sdhB) subunits of the succinate dehydrogenase of Bacillus subtilis was determined. This sequence extends the previously reported sequence encoding the cytochrome b558 subunit (sdhC) and completes the sequence of the sdh operon, sdhCAB. The predicted molecular weights for the Fp and Ip subunits, 65,186 (585 amino acids) and 28,285 (252 amino acids), agreed with the values determined independently for the labeled Fp and Ip antigens, although it appeared that the B. subtilis Fp was not functional after expression of the sdhA gene in Escherichia coli. Both subunits closely resembled the corresponding Fp and Ip subunits of the succinate dehydrogenase (SDH) and fumarate reductase of E. coli in size, composition, and amino acid sequence. The sequence homologies further indicated that the B. subtilis SDH subunits are equally related to the SDH and fumarate reductase subunits of E. coli but are less closely related than are the corresponding pairs of E. coli subunits. The regions of highest sequence conservation were identifiable as the catalytically significant flavin adenine dinucleotide-binding sites and cysteine clusters of the iron-sulfur centers.     

葡萄糖酸氧化杆菌木糖醇脱氢酶基因的克隆与表达

【目的】获得葡萄糖酸氧化杆菌(Gluconobacter oxydans CGMCC 1.637)的木糖醇脱氢酶基因,研究其酶学性质及碳源特别是D-阿拉伯醇和木糖醇对该酶活性的影响。【方法】通过已报道序列的木糖醇脱氢酶的保守区设计引物,用聚合酶链式反应(polymerase chain reaction,PCR)扩增获得目的基因片段。根据获得的片段序列设计引物克隆目的基因的5’和3’片段,将所获得的片段拼接,获得完整的木糖醇脱氢酶基因。通过构建工程菌获得重组蛋白,并利用氧化还原反应测定重组酶的活性。用含不同碳源的培养基培养G.oxydans CGMCC 1.637,并测定其破胞上清液木糖醇脱氢酶氧化木糖醇的活性;用不同碳源培养的G.oxydans CGMCC 1.637转化木酮糖,用高效液相色谱法测定木糖醇的产量。【结果】获得一个新的798bp的木糖醇脱氢酶基因,所编码的木糖醇脱氢酶含265个氨基酸,属于短链脱氢酶家族。酶学性质研究发现,该木糖醇脱氢酶催化木糖醇氧化的最适合条件为35℃、pH 10.0,最高活性为23.27 U/mg,催化木酮糖还原为木糖醇的最适条件为30℃、pH 6.0。最高活性为255.55 U/mg;该木糖醇脱氢酶的对木糖醇的Km和Vmax分别为78.97 mmol/L和40.17 U/mg。碳源诱导实验表明,d-山梨醇对G.oxydans CGMCC 1.637木糖醇脱氢酶的活性有明显的促进作用,而葡萄糖、果糖、木糖、木糖醇、D-阿拉伯醇对木糖醇脱氢酶活性有明显的抑制作用。而在转化实验中,用d-甘露糖培养的G.oxydans CGMCC 1.637的转化能力明显高于其他碳源培养的G.oxydans CGMCC 1.637的转化能力,其中,用阿拉伯醇培养的G.oxydans CGMCC 1.637的转化能力最低,仅为对照的35%。【结论】克隆自G.oxydans CGMCC 1.637的木糖醇脱氢酶基因是一个新的基因,用阿拉伯醇培养的G.oxydans CGMCC 1.637破胞液木糖醇脱氢酶活性低;且阿拉伯醇对G.oxydans CGMCC 1.637木酮糖的还原能力具有抑制作用。     

Synthesis of optically active amino acids from alpha-keto acids with Escherichia coli cells expressing heterologous genes.

We describe a simple method for enzymatic synthesis of L and D amino acids from alpha-keto acids with Escherichia coli cells which express heterologous genes. L-amino acids were produced with thermostable L-amino acid dehydrogenase and formate dehydrogenase (FDH) from alpha-keto acids and ammonium formate with only an intracellular pool of NAD+ for the regeneration of NADH. We constructed plasmids containing, in addition to the FDH gene, the genes for amino acid dehydrogenases, including i.e., leucine dehydrogenase, alanine dehydrogenase, and phenylalanine dehydrogenase. L-Leucine, L-valine, L-norvaline, L-methionine, L-phenylalanine, and L-tyrosine were synthesized with the recombinant E. coli cells with high chemical yields (> 80%) and high optical yields (up to 100% enantiomeric excess). Stereospecific conversion of various alpha-keto acids to D amino acids was also examined with recombinant E. coli cells containing a plasmid coding for the four heterologous genes of the thermostable enzymes D-amino acid aminotransferase, alanine racemase, L-alanine dehydrogenase, and FDH. Optically pure D enantiomers of glutamate and leucine were obtained.     

Molecular cloning and sequence determination of the lpd gene encoding lipoamide dehydrogenase from Pseudomonas fluorescens

The lpd gene encoding lipoamide dehydrogenase (dihydrolipoamide dehydrogenase; EC 1.8.1.4) was isolated from a library of Pseudomonas fluorescens DNA cloned in Escherichia coli TG2 by use of serum raised against lipoamide dehydrogenase from Azotobacter vinelandii. Large amounts (up to 15% of total cellular protein) of the P. fluorescens lipoamide dehydrogenase were produced by the E. coli clone harbouring plasmid pCJB94 with the lipoamide dehydrogenase gene. The enzyme was purified to homogeneity by a three-step procedure. The gene was subcloned from plasmid pCJB94 and the complete nucleotide sequence of the subcloned fragment (3610 bp) was determined. The derived amino acid sequence of P. fluorescens lipoamide dehydrogenase showed 84% and 42% homology when compared to the amino acid sequences of lipoamide dehydrogenase from A. vinelandii and E. coli, respectively. The lpd gene of P. fluorescens is clustered in the genome with genes for the other components of the 2-oxoglutarate dehydrogenase complex.     

Stepwise metabolic engineering of Gluconobacter oxydans WSH-003 for the direct production of 2-keto-l-gulonic acid from d-sorbitol

2-Keto-l-gulonic acid (2-KLG), the direct precursor of vitamin C, is currently produced by a two-step fermentation route from d-sorbitol. However, this route involves three bacteria, making the mix-culture system complicated and redundant. Thus, replacement of the conventional two-step fermentation process with a one-step process could be revolutionary in vitamin C industry. In this study, different combinations of five l-sorbose dehydrogenases (SDH) and two l-sorbosone dehydrogenases (SNDH) from Ketogulonicigenium vulgare WSH-001 were introduced into Gluconobacter oxydans WSH-003, an industrial strain used for the conversion of d-sorbitol to l-sorbose. The optimum combination produced 4.9 g/L of 2-KLG. In addition, 10 different linker peptides were used for the fusion expression of SDH and SNDH in G. oxydans. The best recombinant strain (G. oxydans/pGUC-k0203-GS-k0095) produced 32.4 g/L of 2-KLG after 168 h. Furthermore, biosynthesis of pyrroloquinoline quinine (PQQ), a cofactor of those dehydrogenases, was enhanced to improve 2-KLG production. With the stepwise metabolic engineering of G. oxydans, the final 2-KLG production was improved to 39.2 g/L, which was 8.0-fold higher than that obtained using independent expression of the dehydrogenases. These results bring us closer to the final one-step industrial-scale production of vitamin C.     

Cloning of Escherichia coli lacZ and lacY genes and their expression in Gluconobacter oxydans and Acetobacter liquefaciens

An efficient transformation protocol for Gluconobacter oxydans and Acetobacter liquefaciens strains was developed by preparation of electrocompetent cells grown on yeast extract-ethanol medium. Plasmid pBBR122 was used as broad-host-range vector to clone the Escherichia coli lacZY genes in G. oxydans and A. liquefaciens. Although both lac genes were functionally expressed in both acetic acid bacteria, only a few transformants were able to grow on lactose. However, this ability strictly depended on the presence of a plasmid expressing both lac genes. Mutations in the plasmids and/or in the chromosome were excluded as the cause of growth ability on lactose.     

山梨酮脱氢酶模块与酮古龙酸杆菌底盘细胞的适配分析

酮古龙酸杆菌Ketogulonigenium vulgare是维生素C二步混菌发酵过程中的产酸菌。山梨酮脱氢酶(L-sorbosone dehydrogenase,缩写为SNDH)作为维生素C直接前体2-酮基-L-古龙酸(2-KGA)合成的关键酶,其作用机制并不十分清楚。借助全基因组测序抽提2个山梨酮脱氢酶基因,分别位于基因组(缩写为sndhg)和质粒(缩写为sndhp)上。通过工程化改造技术在工业产酸菌中构建山梨酮脱氢酶功能模块,比较其对2-KGA产量的影响。研究发现sndhg过表达对菌株产酸影响不明显,sndhp过表达使菌株明显产生副产物。将sndhg和sndhp分别配合辅因子PQQ合成基因pqq A,分别构建sndhg-pqq A和sndhp-pqq A模块,得到的工程菌株产酸情况与之前的结果大致相同。将4株K.vulgare工程菌株分别与内生芽孢杆菌Bacillus endophyticus混合培养传代50 d后,分离菌株进行混菌发酵,其2-KGA的转化率分别提高了15.4%、179%、0.65%和125%。表明混菌适应性进化策略是一种增加功能模块与底盘细胞适配性,进而快速获得优良性状菌种的有效方法。     

Cloning and sequence analysis of the gene for glucodextranase from Arthrobacter globiformis T-3044 and expression in Escherichia coli cells

The gld gene for glucodextranase from Arthrobacter globiformis T-3044 was cloned by using a combination of gene walking and probe methods and expressed on the recombinant plasmid pGD8, which was constructed with pUC118, in Escherichia coli cells. The enzyme gene consisted of a unique open reading frame of 3,153 bp. The comparison of the DNA sequence data with the N-terminal and 6 internal amino acid sequences of the purified enzyme secreted from A. globiformis T-3044 suggested the enzyme was translated from mRNA as a secretory precursor with a signal peptide of 28 amino acids residues. The deduced amino acids sequence of the mature enzyme contained 1,023 residues, resulting in a polypeptide with a molecular mass of 107,475 daltons. The deduced sequence showed about 38% identity to that of the glucoamylase from Clostridium sp. G0005. The glucodextranase activity of transformant harboring pGD8 was about 40 mU/ml at 30 degrees C for a 16-h culture. Although the GDase that was produced from the transformant was shorter than authentic GDase by 2 amino acid residues at the N-terminal end side, its enzymatic properties were almost same as the authentic one. Two kinds of genes, dex1 and dex2, for endo-dextranases from A. globiformis T-3044 were also cloned into Escherichia coli cells. The N-terminal of the purified endo-dextranase from A. globiformis T-3044 agreed with the deduced amino acid sequence, after the 33rd alanine residue, of only the dex1 gene for edo-dextranase. This result suggests that the endo-dextranase is translated from mRNA as a secretory precursor with a signal peptide of 32 amino acids residues. The deduced sequence of endo-dextranase 1 and endo-dextranase 2 showed about 93% and 65% identity with that of known endo-dextranase from Arthrobacter sp. CB-8, respectively.     

Cloning, nucleotide sequence, and overexpression of smoS, a component of a novel operon encoding an ABC transporter and polyol dehydrogenases of Rhodobacter sphaeroides Si4.

The gene coding for sorbitol dehydrogenase (SDH) of Rhodobacter sphaeroides Si4 was located 55 nucleotides upstream of the mannitol dehydrogenase gene (mtlK) within a previously unrecognized polyol operon. This operon probably consists of all the proteins necessary for transport and metabolization of various polyols. The gene encoding SDH (smoS) was cloned and sequenced. Analysis of the deduced amino acid sequence revealed homology to enzymes of the short-chain dehydrogenase/reductase protein family. For structure analysis of this unique bacterial enzyme, smoS was subcloned into the overexpression vector pET-24a(+) and then overproduced in Escherichia coli BL21(DE3), which yielded a specific activity of 24.8 U/mg of protein and a volumetric yield of 38,000 U/liter. Compared to values derived with the native host, R. sphaeroides, these values reflected a 270-fold increase in expression of SDH and a 971-fold increase in the volumetric yield. SDH was purified to homogeneity, with a recovery of 49%, on the basis of a three-step procedure. Upstream from smoS, another gene (smoK), which encoded a putative ATP-binding protein of an ABC transporter, was identified.