Bacterial expression and refolding of single-chain Fv fragments with C-terminal cysteines

Two antibody single-chain Fv (scFv) fragments carrying five C-terminal histidine residues were expressed in Escherichia coli as periplasmic inclusion bodies. Their variable heavy (VH) and light (VL) domains are derived from the mouse monoclonal antibody 215 (MAb215), specific for the largest subunit of RNA polymerase II of Drosophila melanogaster and rat MAb Yol1/34, specific for pig brain alpha-tubulin. ScFv-215 contains an additional cysteine residue near to its C-terminus. After solubilization of inclusion bodies followed by immobilized metal affinity chromatography (IMAC) in 6M urea and a renaturation procedure, scFv monomers, noncovalent dimers, and aggregated antibody fragments were separated by size exclusion chromatography. In addition, a fraction of disulfide-bonded scFv-215 homodimers (scFv')2 was also isolated. The various antibody forms appear to be in equilibrium after renaturation since first peak composed mainly of aggregates could be resolved into a similar pattern of aggregates, dimers, and monomers after repeating the denaturation/renaturation procedure. All fractions of the recombinant scFv-215 demonstrated high antigen-binding activity and specificity as shown by enzyme-linked immunosorbent assay (ELISA) and Western blot analysis. Affinity measurements carried out by competitive immunoassays showed that covalently linked (scFv')2 have binding constants quite close to those of the parental MAbs and fourfold higher than scFv' monomers. ScFv derivatives, specifically biotinylated through the free sulfhydryl group, recognize the corresponding antigen in ELISA and Western blot analysis, thus demonstrating the possibility of using chemically modified scFv antibodies for immunodetection.     

Diffusing capacity of the lung for carbon monoxide

A phage-display technology was used to produce a single-chain Fv antibody fragment (scFv) from the 30AA5 hybridoma secreting anti-glycoprotein monoclonal antibody (MAb) that neutralizes rabies virus. ScFv was constructed and then cloned for expression as a protein fusion with the g3p minor coat protein of filamentous phage. The display of antibody fragment on the phage surface allows its selection by affinity using an enzyme-linked immunosorbent assay (ELISA); the selected scFv fragment was produced in a soluble form secreted by E. coli. The DNA fragment was sequenced to define the germline gene family and the amino-acid subgroups of the heavy (VH) and light (VL) chain variable regions. The specificity characteristics and neutralization capacity of phage-displayed and soluble scFv fragments were found to be identical to those of the parental 30AA5 MAb directed against antigenic site II of rabies glycoprotein. Phage-display technology allows the production of new antibody molecule forms able to neutralize the rabies virus specifically. The next step could be to engineer and produce multivalent and multispecific neutralizing antibody fragments. A cocktail of multispecific neutralizing antibodies could contain monovalent, bivalent or tetravalent scFv fragments, for passive immunoglobulin therapy.     

Construction of a functional single-chain variable fragment antibody against hemagglutinin from Porphyromonas gingivalis

Hemagglutinin is a major glycoprotein of Porphyromonas gingivalis vesicles and likely confers the ability to adsorb and penetrate into host tissue cells. To protect this bacterial invasion, murine monoclonal antibody (MAb) Pg-vc, which inhibited the hemagglutinating activity, was prepared by using P. gingivalis vesicles as an antigen. Western blot analysis revealed that when both MAb Pg-vc and anti-HA-Ag2 antibody raised against the P. gingivalis hemagglutinin adhesin (M. Deslauriers and C. Mouton, Infect. Immun. 60:2791-2799, 1992) were allowed to react with protein blots from P. gingivalis vesicles, a superimposable profile was observed. To obtain a recombinant antibody, cDNAs coding for the variable domains of the L and H chains of MAb Pg-vc were cloned by PCR, and a plasmid specifying a single-chain variable fragment (ScFv) was constructed. Following transformation of Escherichia coli cells, a recombinant ScFv protein was successfully expressed. The immunological properties of this protein were identical to those of the parental murine MAb, specifically recognizing the two proteins (43 and 49 kDa) originating from P. gingivalis vesicles. In addition, the ScFv antibody inhibited the P. gingivalis vesicle-associated hemagglutinating activity. The amino acid sequences deduced from nucleotide sequencing experiments confirmed that variable heavy-chain and variable light-chain regions belonged to VH1 and Vkappa12/13 families, respectively. Since the expression system used in this study can readily provide large quantities of single-chain recombinant antibody, it may be a useful in developing a therapeutic agent for passive immunization in humans.     

Long-chain alkenes of the haptophytes Isochrysis galbana and Emiliania huxleyi

The gene families encoding the immunoglobulin variable regions of heavy (VH) and light (VL) chains in vertebrates are composed of many genes. However, the gene number and the extent of diversity among VH and VL gene copies vary with species. To examine the causes of this variation and the evolutionary forces for these multigene families, we conducted a phylogenetic analysis of VH and VL genes from the species of amniotes. The results of our analysis showed that for each species, VH and VL genes have the same pattern of clustering in the trees, and, according to this clustering pattern, the species can be divided into two groups. In the first group of species (humans and mice), VH and VL genes were extensively intermingled with genes from other organisms; in the second group of species (chickens, rabbits, cattle, sheep, swine, and horses), the genes tended to form clusters within the same group of organisms. These results suggest that the VH and VL multigene families have evolved in the same fashion: they have undergone coordinated contraction and expansion of gene repertoires such that each group of organisms is characterized by a certain level of diversity of VH and VL genes. The extent of diversity among copies of VH and VL genes in each species is related to the mechanism of generation of antibody variety. In humans and mice, DNA rearrangement of immunoglobulin variable, diversity, and joining-segment genes is a main source of antibody diversity, whereas in chickens, rabbits, cattle, sheep, swine, and horses, somatic hypermutation and somatic gene conversion play important roles. The evolutionary pattern of VH and VL multigene families is consistent with the birth-and-death model of evolution, yet different levels of diversifying selection seem to operate in the VH and VL genes of these two groups of species.     

Properties of a single-chain antibody containing different linker peptides

Single-chain antibodies were constructed using six different linker peptides to join the VH and VL domains of an anti-2-phenyloxazolone (Ox) antibody. Four of the linker peptides originated from the interdomain linker region of the fungal cellulase CBHI and consisted of 28, 11, six and two amino acid residues. The two other linker peptides used were the (GGGGS)3 linker with 15 amino acid residues and a modified IgG2b hinge peptide with 22 residues. Proteolytic stability and Ox binding properties of the six different scFv derivatives produced in Escherichia coli were investigated and compared with those of the corresponding Fv fragment containing no joining peptide between the V domains. The hapten binding properties of different antibody fragments were studied by ELISA and BIAcoreTM. The interdomain linker peptide improved the hapten binding properties of the antibody fragment when compared with Fv fragment, but slightly increased its susceptibility to proteases. Single-chain antibodies with short CBHI linkers of 11, six and two residues had a tendency to form multimers which led to a higher apparent affinity. The fragments with linkers longer than 11 residues remained monomeric.     

"Diabodies": small bivalent and bispecific antibody fragments

Bivalent and bispecific antibodies and their fragments have immense potential for practical application. Here we describe the design of small antibody fragments with two antigen-binding sites. The fragments comprise a heavy-chain variable domain (VH) connected to a light-chain variable domain (VL) on the same polypeptide chain (VH-VL). By using a linker that is too short to allow pairing between the two domains on the same chain, the domains are forced to pair with the complementary domains of another chain and create two antigen-binding sites. As indicated by a computer graphic model of the dimers, the two pairs of domains can pack together with the antigen-binding sites pointing in opposite directions. The dimeric antibody fragments, or "diabodies," can be designed for bivalent or bispecific interactions. Starting from the monoclonal antibodies NQ11.7.22 (NQ11) and D1.3 directed against the hapten phenyloxazolone and hen egg lysozyme, respectively, we built bivalent fragments (VHNQ11-VLNQ11)2 and (VHD1.3-VLD1.3)2 and bispecific fragments VHNQ11-VLD1.3 and VHD1.3-VLNQ11. The fragments were expressed by secretion from bacteria and shown to bind specifically to the hapten and/or antigen. Those with 5- and 15-residue linkers had similar binding affinities to the parent antibodies, but a fragment with the VH domain joined directly to the VL domain was found to have slower dissociation kinetics and an improved affinity for hapten. Diabodies offer a ready means of constructing small bivalent and bispecific antibody fragments in bacteria.     

Antibody VH domains as small recognition units

To develop immunoglobulin based recognition units of minimum size, a human heavy chain variable domain (VH) was designed for selection of phage displayed VH. Non-specific binding of the VH through its interface for the light chain variable domain (VL) was prevented through three mutations (G44E, L45R and W47G) in this interface. These mutations were introduced to mimic camelid antibody heavy chains naturally devoid of light chain partners. The third hypervariable loop of the modified VH was then randomised to yield a repertoire of 2 x 10(8) independent clones, which was displayed on phage and selected through antigen binding. VH clones specific for hapten and protein antigens were isolated. Soluble VH was expressed with an isoleucine residue at position 47 to improve expression and stability compared to VH containing a glycine residue at this position, which however was preferable for phage selection. Affinities of soluble VH for hapten were between 100 nM and 400 nM. The VH domains were highly specific, stable and well expressed in Escherichia coli. These positive biophysical properties and their small size make them attractive for biotechnological applications.     

Autoreactivity of human VH domains from cDNA libraries: analysis with a bacterial expression system

Previous studies showed that VH domains of several anti-DNA Abs can bind DNA in the absence of VL. In the current work, we tested the VH autoreactive potential more generally, examining VH domains that did not come from known autoantibodies. Using a bacterial expression system, we produced 11 fusion proteins, each containing a VH domain and a B domain of staphylococcal protein A. The VH domains were coded in cDNA libraries from circulating B cells of healthy young adult humans. Thus, binding properties of the Ig molecules from which they came were unknown. The B cells had not been stimulated in vitro. Seven cDNA clones combined the frequently expressed VH3-23 gene segment with varied DH and JH segments. The other clones contained unmutated VH3-7, VH3-9, VH3-53, and VH4-39 segments. We compared these bacterial expression products with single-chain Fv, VH and VL domains of IgM mAb 18/2, a VH3-23-encoded, DNA-binding autoantibody. Submicromolar concentrations of 5 of the 11 VH domains bound to ssDNA. Those and one more also bound to immobilized poly(dT), and two bound to circular plasmid dsDNA. Soluble poly(dT) was the most potent inhibitor in competitive ELISA. Seven of the VH domains also bound to immobilized nuclear ribonucleoprotein, four to histone and none to thyroglobulin. Two interacted with the matrix of a Sephacryl S-100 column. The polyreactive autoantigen-binding properties of these VH domains raise the question of whether these properties may play a role in the formation of the VH repertoire of circulating B cells.     

Antigen mimicry by an anti-idiotypic antibody single chain variable fragment

For the therapy of cancer patients whose disease is positive for Carcinoembryonic Antigen (CEA), we developed an active specific immunotherapy based on the idiotypic network. The anti-idiotype monoclonal antibody (mAb), 3H1 was generated by immunization of mice with the anti-CEA mAb, 8019. 3H1 mimics CEA both functionally and structurally and acts as a surrogate for CEA. To define the minimum structural requirements for antigen mimicry by 3H1, we constructed plasmid vectors for expression of single chain Fv (scFv) variants of 3H1 in Escherichia coli. Variable heavy (VH) and variable light (VL) chain domains of 3H1 were linked by a 15 amino acid linker (Ln), (Gly4Ser)3 in two constructs, VH-Ln-VL and VL-LnVH. Ln was omitted in two constructs, VH-VL and VL-VH. Each of the scFv constructs has a tag of six His [(His)6 tag] for purification by metal chelate affinity chromatography and detection by enzyme-linked immunoabsorbent assay (ELISA). Comparisons of the binding of 8019 to purified scFv proteins by ELISA and immunoblot experiments showed that only VH-Ln-VL had significant activity. VH-Ln-VL also showed maximum inhibition of binding of 8019 to CEA. Immunization of mice with naked VH-Ln-VL and VH-Ln-VL conjugated to keyhole limpet hemocyanin induced anti-CEA antibodies in mouse sera. Sera from immunized mice inhibited the binding of 8019 to 3H1 as well as CEA. Induction of anti-CEA antibodies in the immunized mice was confirmed by flow cytometric analysis using CEA positive MC-38cea cells. These results demonstrate that for antigen mimicry of 3H1 scFv, the presence of Ln is necessary and the domain order should be VH followed by VL.     

Open sandwich ELISA: a novel immunoassay based on the interchain interaction of antibody variable region

We describe an immunoassay that is based on the interchain interaction of separated VL and VH chains from a single chain antibody variable region. In the presence of antigen, the chains reassociate. VL fragments of anti-hen egg lysozyme (HEL) antibody HyHEL-10 were immobilized on microtiter plates. Samples were coincubated with an M13-displayed VH chain, and assayed with peroxidase-labeled anti-M13 antibody. Signal was detected in direct proportion to the amount of HEL in the sample. Wide dynamic range with < 15 ng/ml sensitivity was attained.     

Expression of monovalent fragments derived from a human IgM autoantibody in E. coli. The input of the somatically mutated CDR1/CDR2 and of the CDR3 into antigen binding specificity

A hybridoma producing a polyspecific human monoclonal IgM antibody (named CB03) has been derived from a fusion of mouse myeloma cells with human spleen lymphocytes obtained from an autoimmune patient suffering from chronic idiopathic thrombocytopenia. The antibody was found to be encoded by somatically mutated VHI and VlambdaIII genes. To study the input of mutated complementarity regions (CDRs) into antibody specificity, the antigen binding features of the purified complete IgM antibody were compared with (i) a Fab fragment by hot tryptic digestion and (ii) recombinant monovalent fragments expressed in E. coli. In detail, vectors were constructed encoding for (i) rFab03 and single chain Fv03 fragments containing the VH and VL genes connected by a linker sequence, (ii) scFc1.1. fragments containing the VH germline equivalent and the CB03 wild-type CDR3 region, and (iii) scFv fragments containing the CDR1 and CDR2 in germline configuration and the CDR3 expressed in the CB253 human fetal B cell hybridoma producing a polyspecific IgM antibody. The expression vectors contained at the 3' end either a (His)6 motif allowing purification on Ni(2+)-agarose or a c-myc tag for specifically detecting the expression products by a murine monoclonal antibody. Western blotting and ELISA analyses of the expression products indicate: (i) recombinant Fab fragments were found in the bacterial periplasm in extremely low amounts (1-10 micrograms from 1 litre bacterial culture), (ii) scFv fragments were obtained in suitable amounts from bacterial periplasm (800-1000 micrograms/l), (iii) the monovalent recombinant fragments as well as the Fab obtained by tryptic digestion reflected the polyspecific antigen binding features of the complete IgM antibody, but did bind to the antigens with much lower affinity, and (iv) the CDR3 was found to be of critical importance for the antigen binding pattern of this particular IgM. We discuss the expression of recombinant scFv fragments in E. coli as a suitable method in studying the role of the somatic mutation in autoantibody generation.     

Hepatic leukostasis and hypoxic stress in adhesion molecule-deficient mice after gut ischemia/reperfusion

An anti-p185HER2/anti-CD3 humanized bispecific diabody was previously constructed from two cross-over single-chain Fv in which YH and VL domains of the parent antibodies are present on different polypeptides. Here this diabody is used to evaluate domain interface engineering strategies for enhancing the formation of functional heterodimers over inactive homodimers. A disulfide-stabilized diabody was obtained by introducing two cysteine mutations, VL L46C and VH D101C, at the anti-p185HER2.VL/VH interface. The fraction of recovered diabody that was functional following expression in Escherichia coli was improved for the disulfide-stabilized compared to the parent diabody (> 96% versus 72%), whereas the overall yield was > 60-fold lower. Eleven "knob-into-hole" diabodies were designed by molecular modeling of sterically complementary mutations at the two VL/VH interfaces. Replacements at either interface are sufficient to improve the fraction of functional heterodimer, while maintaining overall recoverable yields and affinity for both antigens close to that of the parent diabody. For example, diabody variant v5 containing the mutations VL Y87A:F98M and VH V37F:L45W at the anti-p185HER2 VL/VH interface was recovered as 92% functional heterodimer while maintaining overall recovered yield within twofold of the parent diabody. The binding affinity of v5 for p185HER2 extracellular domain and T cells is eightfold weaker and twofold stronger than for the parent diabody, respectively. Domain interface remodeling based upon either sterically complementary mutations or interchain disulfide bonds can facilitate the production of a functional diabody heterodimer. This study expands the scope of domain interface engineering by demonstrating the enhanced assembly of proteins interacting via two domain interfaces.     

Cloning and functional characterization of the 5''-flanking region of human methionine adenosyltransferase 2A gene

Electron microscopy of dimeric and trimeric single chain antibody Fv fragments (scFvs) complexed with anti-idiotype Fab fragments was used to reveal the orientation of antigen binding sites. This is the first structural analysis that discloses the multivalent binding orientation of scFv trimers (triabodies). Three different scFv molecules were used for the imaging analysis; NC10 scFv-5 and scFv-0, with five- and zero-residue linkers respectively between the VH and VL domains, were complexed with 3-2G12 anti-idiotype Fab fragments and 11-1G10 scFv-0 was complexed with NC41 anti-idiotype Fab fragments. The scFv-5 molecules formed bivalent dimers (diabodies) and the zero-linker scFv-0 molecules formed trivalent trimers (triabodies). The images of the NC10 diabody-Fab complex appear as boomerangs, not as a linear molecule, with a variable angle between the two Fab arms and the triabody-Fab complexes appear as tripods.     

Folding and assembly of an antibody Fv fragment, a heterodimer stabilized by antigen

The folding and assembly of the Fv fragment of the phosphorylcholine binding antibody McPC603, a non-covalent heterodimer of the variable domains VH and VL, was investigated. Since both domains, each engineered for stability and folding efficiency, could now be obtained in native and soluble form by themselves, fluorescence spectra of VH and VL in unfolded, folded and associated states can be reported. VH and VL only associate when they are native, and the stability of the heterodimer is strongly increased in the presence of antigen. VH rapidly folds into an hyperfluorescent intermediate, and the native state is reached in two parallel, proline-independent reactions. VL displays two fast refolding reactions, which are followed by two slower phases, limited by proline cis/trans-isomerization. The rate-limiting step for both the Fv and the scFv (single-chain Fv) fragment is the formation of the native VH-VL interface, which depends on ProL95 being in cis. The folding of the Fv fragment is fast after short-term denaturation or in the presence of proline cis/trans-isomerase catalysis, but the scFv fragment falls into a kinetic trap, observed by the persistence of the slow phases under all conditions. Furthermore, the scFv fragment, but not the Fv fragment, gives rise to premature interface formation, indicated by the fluorescence spectra and a much higher transient binding of 8-anilino-1-naphthalene sulfonate. The analysis of the folding pathway of the domains VH and VL in isolation and in non-covalent and covalent assemblies should provide helpful insights into the folding of multimeric proteins in general, and for the further engineering of stable and well-folding antibody fragments in particular.     

Evidence for immunoglobulin heavy chain variable region gene replacement in a patient with B cell chronic lymphocytic leukemia

Immunoglobulin heavy chain variable (V) gene replacement is an unusual recombinatorial event characterized by rearrangement of a germline V gene to a preformed VDJ gene complex. This phenomenon has occasionally been implicated in the emergence of clonal subpopulations during the course of acute lymphoblastic leukemia; it has also been found in murine precursor B cell lines. V gene replacement has never been described in lymphoproliferative disorders corresponding to more differentiated stages of B cell ontogeny. The present communication provides evidence for the operation of the same mechanism in B cell chronic lymphocytic leukemia (B-CLL). Genomic DNA and total cellular RNA extracted from peripheral blood mononuclear cells of a 48-year-old female patient diagnosed as having typical B-CLL were subjected to polymerase chain reaction (PCR) amplification aiming to detect rearranged clonal heavy and light chain variable genes (VH and VL, respectively). PCR consistently gave two VH amplification products, both at the DNA and the RNA level; similar analysis for the VL region revealed the presence of a single rearranged VK gene. Direct sequence analysis of the PCR products revealed that, except for a number of silent mutations, the single rearranged VK gene was identical to the germline A1-A17 VK gene. The two rearranged VH gene segments belong to the VHl and VHIII gene families and are closely homologous, respectively, to the germline gene segments V1-18 and V3-30, which have been shown to be used by autoantibodies. Both rearranged VH genes showed identical in-frame D-N-JH junctions and JH gene usage (JH5b), whereas the VH-N-D junctions were different. The above findings indicate that, during the course of the disease of our patient, VH gene replacement took place giving rise to two different clonally related subpopulations. This raises the intriguing possibility that the recombinase machinery, which governs Ig recombinatorial processes, might be operative even at more advanced stages in B cell ontogeny.     

Structure and binding properties of monoclonal antibodies to core histones from autoimmune mice

Histones are frequent targets of self-reactive antibodies during autoimmune syndromes. We report the specificities and V region genes of three IgG anti-histone MAbs obtained from autoimmune mice. Each of the MAbs, named LG2-1, LG2-2 and BWA3, is directed against a different determinant located in the basic amino-terminal domain of core histones. LG2-1 reacts with a peptide from histone H3 (residues 30-45), LG2-2 recognizes the amino-terminus of H2B (residues 1-13) and BWA3 binds an epitope corresponding to a region of high sequence similarity between H2A and H4 (residues 1-20 and 1-29, respectively). The analysis of their V region sequences indicates that the H chain CDRs of these MAbs are remarkable for the presence of negatively charged amino acid residues that may play a role in the binding to cationic histones. The H chain importance in conferring reactivity to histones is corroborated by the observation that each of the VH gene segments of these MAbs is very similar to VH genes of previously described murine anti-histone antibodies.     

Structures of SAP-1 bound to DNA targets from the E74 and c-fos promoters: insights into DNA sequence discrimination by Ets proteins

Rabbit monoclonal antibodies (RmAb) are not routinely obtained by eukaryotic cell fusion techniques. Therefore, we have applied phage display technology to produce a recombinant rabbit Fab molecule directed against the KLH model antigen. The Fab fragments selected from the rabbit phage display library were subcloned in an expression vector to permit the production of a fusion protein comprising a dimer of bacterial alkaline phosphatase (phoA). This fusion protein was directly produced into the periplasmic space of Escherichia coli. We show that a crude extract containing these conjugates can be used in a direct enzyme immunoassay, as exemplified in the case of the KLH antigen.     

Role of mouse VH10 and VL gene segments in the specific binding of antibody to Z-DNA, analyzed with recombinant single chain Fv molecules

A plasmid vector was constructed for the expression of a single chain Fv domain of mouse mAb to Z-DNA (antibody Z22), which is encoded by VH10 and V kappa 10 gene family members along with Dsp2, JH4, and J kappa 4 segments. The vector coded for a PhoA secretion signal, VH segment, flexible peptide linker, VL segment, (His)5, and a protein A domain. Unique restriction sites allowed exchange of the segments as cassettes. Bacteria transformed with the vector secreted soluble recombinant Fv with specific Z-DNA-binding activity. When the L chain of Z22 was replaced with a library of splenic VL cDNA from a mouse immunized with Z-DNA, only a light chain closely resembling that of the original Z22 (differing at six amino acid positions) yielded Fv with Z-DNA-binding activity. The Fv with this L chain replacement had a lowered affinity, but remained selective for Z-DNA. Replacement of the Z22 H chain with a mixture of 11 VH10-encoded H chains yielded two Z-DNA binding clones, but they bound B-DNA and denatured DNA as well as Z-DNA. The replacement clones indicate the importance of the H chain CDR3 and particular VH-VL combinations in formation of specific antibodies to Z-DNA.     

Molecular cloning, sequencing, and expression of the genes encoding adenosylcobalamin-dependent diol dehydrase of Klebsiella oxytoca

The pdd genes encoding adenosylcobalamin-dependent diol dehydrase of Klebsiella oxytoca were cloned by using a synthetic oligodeoxyribonucleotide as a hybridization probe followed by measuring the enzyme activity of each clone. Five clones of Escherichia coli exhibited diol dehydrase activity. At least one of them was shown to express diol dehydrase genes under control of their own promoter. Sequence analysis of the DNA fragments found in common in the inserts of these five clones and the flanking regions revealed four open reading frames separated by 10-18 base pairs. The sequential three open reading frames from the second to the fourth (pddA, pddB, and pddC genes) encoded polypeptides of 554, 224, and 173 amino acid residues with predicted molecular weights of 60,348 (alpha), 24,113 (beta), and 19,173 (gamma), respectively. Overexpression of these three genes in E. coli produced more than 50-fold higher level of functional apodiol dehydrase than that in K. oxytoca. The recombinant enzyme was indistinguishable from the wild-type one of K. oxytoca by the criteria of polyacrylamide gel electrophoretic and immunochemical properties. It was thus concluded that these three gene products are the subunits of functional diol dehydrase. Comparisons of the deduced amino acid sequences of the three subunits with other proteins failed to reveal any apparent homology.