EP1027427A2 - Thermostable alcohol dehydrogenases - Google Patents

Abstract

The present invention provides proteins having alcohol dehydrogenase activity. In a particular embodiment, proteins having chiral alcohol selective alcohol dehydrogenase activity. In another embodiment, proteins having thermostable alcohol dehydrogenase activity are provided.

Description

THERMOSTABLE ALCOHOL DEHYDROGENASES Relation To Previous Applications

This application is a continuation of U S Provisional Application 60/063,517 filed October r 1998 (Allen, et al , THERMOSTABLE ALCOHOL DEHYDROGENASES) Statement of Government Rights

Some of the work m this specification has been performed under NSF Grant No 95 61842, therefore the United States Government may have certain rights in the invention Field of the Invention

The present invention relates to the filed of catalytic reagents for use in chemical synthesis In particular, the present invention relates to novel thermostable enzyme proteins with alcohol dehvdrogenase activity Background of the Invention

There is great need in the chemical industry to develop catalytic methods capable of efficienth converting molecules to products without generating excessive waste streams oi emissions Molecular conservation methods based on biocatalyt reagents, such as dehydrogenase enzymes, offer opportunities to solve the challenges found in attempting industrial scale synthesis Current dehydrogenase based technology has been limited by the lack of available enzymes The few enzymes which are available have narrow substrate specificities, poor stabilities or yield only a single type of stereocenter New dehydrogenase enzymes are needed in order to increase the chemo- and stereoselectivity for different substrates, thereby increasing the spectrum of useful chemical synthetic reactions over that which is currenth available

Dehydrogenase biocatalyst reagents efficiently reduce ketones to chiral alcohols The reduction ot the carbonyl group is one of the most versatile transformations in organic chemistry It can be used to form carbon-carbon bonds, introduce heteroatoms, serve as masked functionality and can be the site at which chirality is installed (March, 1985) Although many methods exist for operation on carbonyl groups, few combine the selectivity, mildness, cost effectiveness and minimal waste streams offered by enzyme catalyzed reactions Alcohol dehydrogenases are one of the principal enzymes that effectively operate on the carbonyl group In spite of industrial interest, \ery few alcohol dehydrogenases are commercially available that can produce molecules of high enantioselectivity under conditions encountered in industrial applications New dehydrogenase biocatalysts need to be developed that meet the requirements of the industrial synthetic chemist Biocatalyst reagents offer methods to install chiral centers in high value chemicals. Biocatalysts are increasingly being recognized as potential alternatives to traditional synthetic organic methods and are noted for their remarkable catalytic capacity (Faber, 1992). In spite of potential utility, biocatalysts and industrially important biotransformations remain to be discovered. Development of novel biocatalysts that are user friendly, economical and produce high yields of chiral chemicals are crucial to the future of industrial chemistry. Over 50% of existing therapeutic agents are chiral molecules and of those that are synthetic (528), 75% are prepared as racemic mixtures. Several companies have demonstrated that only one enantiomer of a racemate often produces the desired biological activity while the other antipode is ineffective or responsible for side effects. Drugs such as the recently redeveloped antiasthma agent (R)-albuterol, and the nonsteroidal antiinflammatory compounds ibuprofen and naprosyn represent chiral drugs originally manufactured as racemates in which one enantiomer is the active species (1994). Such agents have a huge market value each with sales in excess of one billion dollars worldwide(Stinson, 1994). In many such cases, biocatalysts offer the potential to dramatically streamline synthetic processes by minimizing the total number of steps and or complex purification schemes (Crout and Christen, 1989). Shorter synthetic routes to chiral compounds also reduce waste streams and minimize environmental pressures that are beginning to force chemical makers to search for alternative synthetic schemes. One emerging technology involves the use of dehydrogenase enzymes to effectively reduce carbonyl groups to the corresponding alcohols. Figure I highlights a few ketone substrates stereoselectively reduced to chiral alcohols that have been used as intermediates in chemical synthesis.

Commercially available dehydrogenases are limited in chemo- and stereoselectivity. Alcohol dehydrogenases are a family of enzymes capable of formal reversible two electron chemistry in which alcohols are oxidized to the corresponding ketones (Table I (Faber, 1992)).

Table I. Characteristics of commonly studied alcohol dehydrogenases.

a. The specificity is reversed for small ketone substrates. Depending on the conditions (Lemiere, 1986), ketones can be reduced to the respective alcohols via a stereospecific delivery of a hydπde equivalent catalyzed by the enzyme coupled to a bound cofactor (NADH or NADPH). This system represents a mild, extremely selective route to valuable chiral intermediates that can be used particularly by the pharmaceutical industry for the preparation of chiral therapeutics. Extensive investigations of alcohol dehydrogenases have shown that these proteins can be an efficient means for generanng a variety of compounds with capacity for industrial scale-up (See Drawing below) (Bradshaw, et al. 1992; Hummel, 1990; Seebach, et al., 1984)

Ketone substrates reduced by T. brocku alcohol dehydrogenase

Brief Summary of the Invention

The present invention provides for a protein with alcohol dehydrogenase activity selected from the group consisting of AD55.1, AD83.5, AD5.1, AD7.1, AD14.1, AD31.3, AD14. AD19. AD30, AD31, AD39, AD49.4, AD49.12, AD55, AD69, AD71, AD98, and XY (ADOl). In a particular embodiment, the present invention provides for a protein with chiral alcohol selective alcohol dehydrogenase activity selected from the group consisting of AD 19, AD30. AD39, AD49.4, AD49.12, AD55, AD69, AD71, AD98, and XY (ADOl) The present invention also provides for a protein with thermostable alcohol dehydrogenase activity selected from the group consisting of AD7 1, ADM 1. AD31 3. AD39 4. AD49 4. AD49.12. AD55.1, AD83.5, XY, and AD5.1 wherein said protein retains activity at temperatures above 32°C.

The present invention also provides for recombinant DNA constructs which encode for the proteins of the present invention. Such DNA constructs, which can be generated using recombinant DNA techniques, can be expression vector constructs whereby the nucleic acid sequence is transcribed and translated into the desired protein Said DNA constructs will be transformed into host cells which can express the proper protein. Thus the instant invention encompasses expressible recombinant DNA constructs which express the protein of the present invention, as well as host cells which are transformed with such constructs. The present invention also provides for using the DNA sequences and constructs of the present invention to hybridize with DNA or RNA in target gene banks, gene libraries, or expression libraries, under stringent hybridization conditions, to detect substantially related or related protein encoding nucleic acids from a pool of nucleic acids.

Thus the present invention provides for methods of producing recombinant proteins of the invention, as well as methods for detecting other related protein encoding nucleic acid sequences from pools of nucleic acid.

Brief Description of the Drawings

The enzymes of the present invention will be better understood in conjunction with the following figures in which: Figure 1 is a genetic map of the λpMYF construct;

Figure 2 shows the DNA fragment analysis of cloned dehydrogenases;

Figure 3 is a diagram of the SDS gel analysis of ADH candidates;

Figure 4 are graphs of Gel Filtration Profiles and Molecular Weight Determination for cloned enzymes; Figure 5 shows Native PAGE (Polyacrylamide Gel Electrophoresis) analysis of Crude Extracts;

Figure 6 are graphs of Optimal Temperature Charts for Several ADH Candidates;

Figure 7 are graphs of Residual Temperature Charts for Several ADH Candidates. Wherein the optimal temperature charts for enzymes produced from strains AD 19, AD30, AD39, AD49.4,

AD49.12, AD55, AD69, AD71, AD98, XY, and HLADH (containing the Horse Liver Alcohol Dehydrogenase control enzyme on plasmid pBPP), are shown. The data is presented as percent of original activity.

Detailed Description of the Invention

Dehydrogenases have been identified as useful biocatalysts for chemical synthesis applications, particularly in the reduction of carbonyl groups to alcohols. They allow simplification of reactions that are difficult by traditional synthetic methodology. T e highly stable biocatalysts of the present invention add a variety of new dehydrogenase specificities to the synthetic chemist's toolbox.

The present invention describes new enzymes from thermophilic organisms suitable for use as stable off-the-shelf reagents for selectively and mildly installing chiral centers from corresponding carbonyl groups. These enzymes have been characterized and show a variety of substrate specificities and enantioselectivities which indicates that they are useful biocatalysts which can be used economically to prepare fine chemicals and intermediates. The enzymes of the present invention are suitable for use in methods for large scale chemical reactions, and show an expanded range of substrate specficities.

Dehydrogenases from whole yeast cells are classic synthetic biocatalysts. Dehydrogenase catalyzed reactions have been performed by two principle methods. Most common to the organic chemist is the formation of chiral alcohols catalyzed by whole cell baker's yeast Baker's yeast has enjoyed reasonable notoriety as a biocatalyst because it can be used as an off the shelf reagent, requiring no special treatment such as refrigeration, purification or cofactors. The active component of yeast, is an alcohol dehydrogenase which has been extensively studied by X-ray crystallography, mechanistic biochemistry and organic chemistry. The broad range of molecules which serve as substrates has advanced baker's yeast as an important synthetic organic reagent and its use has been documented in several reviews (Csuk and Glanzer, 1991; Neuberg, 1949; Servi, 1990). In spite of its apparent general utility there are some limitations to intact Baker's yeast including the risk of secondary reactions resulting from other active proteins within the yeast. Other challenges associated with whole cell biocatalytic systems include limited substrate access to intracellular proteins and contaminating products of normal metabolic processes. Whole cells are often sensitive to immobilization and intolerant to temperature and inclusion of organic solvents.

Isolated dehydrogenase biocatalysts offer advantages over whole cell methods. Isolated enzymes offer an alternative method to whole cell systems to effect biotransformations. Chiral compounds with high enantiomeric excess have been prepared using purified dehydrogenase catalysts without contamination from undesired competing reactions. Dehydrogenases have shown promise for commercial application in the preparation of unusual amino acids (Benoiton, et al., 1957), b- hydroxyketones (Casy, et al., 1992) and resolution of racemic alcohols (Jones and Jakovic, 1982). Until recently, application of purified dehydrogenases has been limited by both the cost of cofactors (NAD(P)H) and the intolerance of cell free dehydrogenases to extreme environments often found in synthetic organic transformations. Biocatalysts developed from thermophilic organisms could increase environmental tolerance although thermophiles have yet to be widely investigated for useful dehydrogenase catalysts. The alcohol dehydrogenase isolated from Thermoanaerobium brockii is one of the few successfully isolated dehydrogenases from thermophilic organisms which show promise for industrial applications (Keinan, et al., 1986; Lamed and Ziekυs, 1981).

NAD(P)H Cofactors are Required for Dehydrogenase Based Synthetic Methods. The role of the cofactor in the catalytic system needs to be addressed since NAD(P)H are relatively costly reagents (as much as $250,000/mole) and are used stoichiometrically during the reaction. Cofactor cost currently limits the use of cofactor requiring enzymes to high value applications such as the preparation of pharmaceuticals. Generally, cofactors can not be simply disposed of at the conclusion of the reaction but rather need to be recycled. Ideally, cofactors should function at catalytic concentrations, necessitating in situ regeneration of the active species.

Current cofactor recycling techniques allow dehydrogenase biocatalysts to be used as valuable synthetic tools. The question of cofactor recycling has received considerable attention and several elegant means exist to regenerate reduced nicotinamide analogs for use in dehydrogenase catalyzed reactions (Jones, 1986). Promising methods for cofactor recycling include multiple dehydrogenase reactions in which the reduction by the dehydrogenase of interest is coupled to a second dehydrogenase such as formate dehydrogenase (Hummel, et al., 1987). Other methods involve using simple alcohols like isopropanol which have more favorable equilibria to regenerate cofactors without reoxidizing newly formed products (van Eys, 1961). Recycling systems offer the additional advantage of beginning the reaction with the oxidized form of the cofactor which is considerably less expensive then the reduced species. Recently NAD⁺ molecules have been treated with vanadate to produce a less expensive analog to the highly cosdy NADP⁺ molecule required for some dehydrogenase catalyzed reactions (Crans, et al., 1993).

Dehydrogenases reduce ketones stereoselectively following Prelog rules. The dehydrogenase chemical mechanism begins by binding the substrate either at the carbonyl or alcohol oxidation state. For the reduction of carbonyl groups, analysis of reaction kinetics and X-ray crystal structures of substrate bound dehydrogenases have shown that the pro-R hydrogen on the nicotinamide ring is positioned by the protein to efficiently transfer a formal hydride moiety to the carbonyl group. Several general observations have been made regarding the stereoselectivity of dehydrogenases which has led to some empirical rules for dehydrogenase mediated catalysis. In an early attempt to explain dehydrogenase stereoselectivity, Prelog developed a set of rules for reduction of carbonyl groups which predict that nucleophilic attack should occur at the face opposite the large substituent (shown in the drawing below)(Prelog, 1984). ^Φ Dehydrogenise The Prelog Rule for reduction of π carbonyl groups. NAD(P)H NAEKP

A more recent reinvestigation of dehydrogenase selectivity was carried out by Jones and coworkers who developed die cubic surface model for predicting stereoselectivity of dehydrogenase reactions shown in the drawing below (Dodds and Jones, 1988; Jacovac, et al., 1982; Jones and Jakovac, 1982).

The Cubic Space model for alcohol dehydrogenase catalyzed reactions (Jones and Jakovac, 1982). Cubic surface model of the active site of horse liver alcohol dehydrogenase. Cubes with solid lines are forbidden spaces occupied by active site amino acid residues. Dotted lines are cubes with disfavored spaces originating from potential interactions with charged groups on active site residues. Open spaces are available for substrate access to the active site, which allows positioning of the substrate in such a way that the active carbonyl functionality is located just above the hydride equivalent of the cofactor at the Cl-Dl boundary.

This model more accurately describes the structure activity relationship between the active site binding surface and the structural geometry imposed by the substrate. The cubic surface model remains an effective stereoselectivity model since it evaluates catalysis based on boti a library of substrates with differing structure and high resolution X-ray crystal data instead of addressing specific interactions between substrate and protein. These investigations allow an understanding of dehydrogenase function that may be extended to newly developed catalysts. A newly discovered dehydrogenase can be assayed with a group of molecules to define the potential applications of the isolated catalyst

There is significant commercial utility for alcohol dehydrogenases. Development of reliable catalytic systems will provide an efficient means of generating chiral compounds without the need for alternative often tedious, costly resolution and purification techniques. Dehydrogenase catalysts will be synthetic reagents that are both mild and selective even in die presence of sensitive functionality to produce chiral building blocks for assembly of more complex compounds. Dehydrogenase catalyzed reactions are expected to be applied to the synthesis of complex therapeutic compounds such as antibiotics and antiviral agents. Few options are available for chemists who wish to develop biocatalytic methods using dehydrogenase enzymes. The two best studied commercially available biocatalytic methods are whole-cell yeast systems and horse liver alcohol dehydrogenase (HLADH). These methods represent relatively low cost biocatalytic options for the reduction of ketones to directly insert chiral centers into molecules that have demonstrated me potential of dehydrogenases for industrial use. Unfortunately, these two catalysts are limited in their substrate range and the HLADH enzyme in particular, is sensitive to inactivation. The structural specificity of these two enzymes has, however, laid the groundwork for development of other dehydrogenases with different cherao- and stereoselectivities in addition to providing opportunities for dehydrogenase catalysts that are more tolerant to conditions found in industrial applications. One technology which has been developed address the lack of stable enzymes is Altus Biologies' Cross-Linked Enyzme Crystal (CLEC) technology. Aldtough expensive, tiiis technology is useful to stabilize proteins for some applications and could possibly be used to stabilize thermophilic proteins even further if needed.

Thermophilic organisms provide stable biocatalyst reagents for organic synthesis. One of die greatest challenges still facing syndietic chemists is to develop inexpensive, efficient ways of producing compounds. In the field of biocatalysis, die problem is not only to find a system to effect a given transformation but also to implement the reaction on a large scale. Proteins derived from tiiermophilic sources provide an opportunity to satisfy the demands of the industrial process chemist. Thermophile derived biocatalysts maintain activity under relatively harsh conditions

(Daniel, et al., 1990; Fontana, 1984; Pham and Phillips, 1990) including: non-refrigerated storage, presence of organic cosolvents, and resistance to temperature inactivation.

Thermophile derived biocatalysts can reduce production and catalyst cost The ability to function at elevated temperatures should not be confused widi a need to operate under such conditions.

Tolerance to temperature reflects die inherent ability to survive as a functioning catalyst long after mesophile derived proteins have been inactivated, regardless of the means by which deactivation occurs. In addition, the costs associated widi cloned thermophilic enzymes (resulting from longer catalyst lifetime) should be lower man the corresponding mesophilic proteins, particularly for systems requiring purified catalyst preparations. Purification can be accomplished in fewer steps without the need for refrigeration equipment The effect of tiiermophilic biocatalysts is to lower catalyst cost by simplifying purification and increase catalyst half life to reduce the frequency of catalyst replacement The present invention and its various embodiments will be better understood in view of the following examples. The examples are presented only by way of illustration, and not limitation, of certain embodiments of die present invention. One of ordinary skill in die art would be able to understand and use die teachings of the present invention to develop additional equivalent emodimnents wid in die scope of die teaching and spirit of die disclosure.

Example 1. Isolation of Strains

Strains. Most isolated strains and cultures are grown on TT medium. This medium consists of (per liter): BBL Polypeptone (8 gm), Difco Yeast Extract (4 gm), and NaCl (2 gm). Small scale cultures for screening are grown at 55-65°C at 250-300 rpm with 1 liter of medium in a 2 liter flask.

Enrichment Procedures for Newly Isolated Thermophile s. Multiple stream sediments, composting organic materials, and soil samples are used to isolate new strains. These samples were collected from numerous geographic sites ranging from the Midwest to the Southeast United States.

Samples (~1 gm) are resuspended in 2 ml of TT broth and 50-100 μ\ of these samples were plated onto TT agar plates containing twice the usual amount of agar (3%). The increased agar concentration in the initial screening helped keep highly motile isolates from covering the entire plate. After initial purification, agar was usually added to a final concentration of 1.5% for solid media Plates are incubated at 55°C or 65° C for one to two days and isolates then purified by numerous restreaks onto fresh plates for single colony isolation. The initial basis for differentiation is color, colony morphology, microscopic examination, temperature of growth. Several hundred strains were initially isolated. Over 100 different microorganisms were chosen for further study.

Example 2. Construction of λpMYF

λpMYFl is a derivative of high capacity phagemid vector λpSL5, which was successfully applied for cloning and expression of a number of prokaryotic and eukaryotic genes in E. co/f.(Nick K.

Yankovsky, et al. 1989. Phasmids as effective tools for construction and analysis of gene libraries. Gene, 81, 203-210). The original λpSLS vector is a hybrid of phage λ vectors λgtWEC and 1 L47.1 and plasmid pUC19. λpSL5 DNA contains all the genes required for lytic λ development, and provides for effective amplification of the primary libraries in phage form. The vector itself cannot be effectively packaged into die λ capsid in monomeric or oligomeric forms due to its size of 35 kb. Thus, recombinant molecules of appropriate length have a selective advantage, which provides for very low background of non-recombinant molecules (<1% for λpSL5 vector). Recombinant libraries collected as phage particles can be transduced into host strain containing a helper phage and screened for dehydrogenase activity its bacterial colonies. The thermυsensitive λ repressor cI857 from resident λ prophage facilitates stability of λpSL5 based clones at 30°C.

Autonomous replication of phagemids in lysogens is mainlained by a pUC replication oπgin. The average size of cloned fragments is 12-15 kb, with maximum at about 20 kb, which is determined by die capacity of λ capsid (48 kb). E. coli LE392 (supE supF hsdR) was chosen as a host strain. LE392 is a standard host for phage and plasmid vectors bearing amber-mutations (such as λpSL5 and λpMYFl).

A map of λpMYFl is shown in Drawing 1. λpMYFl was constructed by elimination of one of two BamHI sites on λpSL5 (Fonstein, unpublished). This vector allows cloning of Sau3A generated fragments into its remaining BamHI site. λpMYFl was chosen as cloning vector because it provides d e highest level of foreign genes expression, presumably from P_ and PR promoters. Also, by unclear mechanism, copy number of λpMYFl is regulated dependent on toxicity of expressed foreign protein, i.e. λpMYFl allows cloning of unbearable for standard plasmid vectors proteins, such as proteases.

Example 3. Preparation of Clone Banks

Choice of the vectors and E. coli host strains. Several cloning systems were used to clone new dehydrogenases. The first was die λZAP/XLOLR system from Promega which we had previously used to clone many other types of enzymes. This system led to d e cloning of several new dehydrogenases but was not as effective as statistical analysis suggested it could be for the cloning of new genes. As a result, a second system, λpMYFl, which we have also used regularly in the past was employed (and led to most of the new clones described in tiiis work).

Preparation ofgenomic DNA of thermostable microorganisms from ThermoGen collection. Strains were re-streaked from frozen stocks on TT agar plates, and 50 mi TT liquid cultures were inoculated from single colonies. Cultures grew overnight at 55^CC, and were washed twice in 1 x TE buffer. Pellets were resuspended in 2 ml of fresh buffer SI (50 mM Glucose, 50 mM Tris HC1 pH 8.0, 50 mM EDTA pH 8.0, 10 mg/ml lysozyme) and incubated at room temperature (RT) for 5 minutes. 5 ml of buffer S2 (50 mM Tris HCl pH 8.0, 50 mM Tris HCl pH 8.0, 1% SDS, 10 mg/ml Proteinase K)was added in each tubes, and samples were lysed at 65°C until solution became transparent Normally, this procedure required 3-4 hours, but time varied depending on strain. After completion of lysis several extractions with organic solvents were applied to get rid off cell proteins. The first two extractions were done with pure phenol saturated with Tris HC1, pH 7.5, followed by phenol/chloroform 1:1 and cWoroform/isoamyl alcohol 24:1 extractions. DNA was extracted from water phase by precipitation with 95% ethyl alcohol, and resuspended in 500 μl of 10 mM Tris HC1 pH 7.5 solution. Construction of recombinant DNA libraries on pMYFl. For the clone bank preparation we used the following microorganisms from die collection: # 2, 4, 7, 14, 16, 17, 19, 20, 22, 23, 24, 26, 27, 30, 31, 39, 45, 49, 51, 55, 57, 67, 69, 71, 75, 77, 83, 90, 98, 99, 116, 118, 122, 136, 146. In total, there were 55 new libraries constructed from 37 different strains (see Table 1 below).

Genomic DNA of each strain was partially cleaved with serial dilutions of which gave a gradient of restriction fragments lengths. The reactions continued for lh at 37°C followed by agarose electrophoresis to determine the sample widi optimal an average fragment size of 10-20 kb. λpMYFl DNA (0.5 μ Jμ\) was digested widi Bam HI to completion. Both genomic Sau 3A1 fragments and linearized λpMYFl DNA were precipitated wid etiiyl alcohol and resuspended in sterile distilled water at concentration 0.5 μ /μ\ and 0.8 g/ l. 2 μl of λpMYFl was ttien ligated with 3 μ\ of genomic Sau 3A1 fragments overnight at 16°C using 1 U of T4 DNA ligase (Stratagene) in a ligation mixture volume of 10 μ\. Upon of the completion of die reaction, 2 μl of die ligation mixture was incubated widi 12 μ\ of 1 packaging extract (Promega) for 90 minutes at room temperature. Extracts were plated on LB and covered widi top agar containing fresh E. coli LE392 cells. Plates were incubated for approximately 16 hours, and recombinant phages were collected. Libraries containing 10^ - 10^ independent plaques, were stored at 4^βC in SM buffer.

Table 2. Organisms and Clone Banks Screened in this Work Strains used from ThermoGen collection Organisms Screened Nearly 200 strains from die ThermoGen collection. pSL5 based clone banks 2; 10; 14; 16; 17; 19; 20; 23; 27; 30; 36; 51; 67; 69; 71; 75; 77; 83; 90;

98; 99; 116; 118; 122; 136; 146; xy (26total) pMYFl based clone banks 2; 4; 7; 14; 17; 19; 20; 22; 23; 24; 26b; 30; 31 ; 39; 45; 49; 51 ; 55; 57; 69; 71; 77; 83; 90; 98; 122; 136; xy (28 total) pZAP based clone banks 11; 13; 14; 15; 18; 21; 26b; 35; 36; 37; 49; 51; 59; 65; 83

Example 4. Screening for Novel Alcohol Dehydrogenases (ADH)

The following approach for the identification of new dehydrogenase catalysts from thermophilic organisms was used. Both native thermophilic organisms and clone banks from thermophilic organisms were screened for new dehydrogenase activities using a colorimetric para-rosaniline test (described below) since some enzymes can be produced more easily from die host organism, and others can be produced more easily in a cloned format since the expression control signals which may inhibit production of the gene in the native host are removed in the clone. As a result we screened botii native tiiermophilic organisms as well as clone banks from these organisms so we would have an optimal chance of finding new and unique enzyme activities. In general, clone banks were plated out at about 1,000 colonies per plate and native organisms were screened individually. In die first step, a para-rosaniline screen was used widi edianol as a substrate to identify colonies with alcohol dehydrogenase activity. This broad range of enzymes was then screened in a second step with a set of different alcohol substrates which allowed us to identify the key enzymes of interest for the project Once a set of potential candidates were identified, the enzymes were more carefully analyzed quantitatively.

Transduction and activity tests. For transduction, 50 μl of λpMYFl libraries were mixed widi 200 μl of fresh overnight LE392 (1) cells at room temperature. After 20 minutes, 600 μl of LB were added, and incubation continued for 60 minutes at 30°C. Upon transduction, cells were plated on LB agar, and incubated at 30° C for 24-36 hours. Grown colonies were screening for dehydrogenase activity according the following protocol. Clone banks were transduced into l_E392λ (λ lysogen) for MYF derivatives or XLOLR for pBK derivatives. The conditions for LE392λ and XLOLR transduction were identical, however die growth temperature was 30°C for LE392λ derivatives which contain a temperature inducible lambda instead of 37°C for XLOLR derivatives. Plates were grown with a colony density of approximately 500 colonies per plate.

Plate .Assay. Botii native strains and clone banks were screened for the presence of ADH activity using para-rosaniline method (modified from T. Conway et al., 1987, J. Bacteriol., 169:2591- 2597) as follows: Indicator plates were prepared by adding 8 ml of para-rosaniline (2.5 mg/ml in 96% edianol) and 100 mg of sodium bisulfite to 400 ml batches of precooied (45°C) Luria agar. Most of die dye was immediately converted to die leuco form by reaction with bisulfite to produce a rose-colored medium. Plates were stored away from fumes which contain aldehydes in die dark. A number of different primary and secondary alcohols such as ethanol, hexanol, isopropanol and cyclohexanol were included in agar containing LB media, 50 mg/ml p-rosaniline and 250 mg/ml sodium bisulfite (Conway, et al., 1987). Ethanol (or another substrate) diffuses into the bacterial cells to produce die acetaldehyde (or the appropriate product) by alcohol dehydrogenase. The leuco dye serves as a sink, reacting with the acetaldehyde to form a Schiff base which is intensely red. Utilizing this metiiod, we screened dirough several hundred new strain isolates as well as all of the clone banks described above. A number of different candidates were identified for further study. By far, die clone banks gave us the most colonies with the best overall enzyme production and yield. In fact, many native strains which did not show dehydrogenase activity on the screening plates, were actually good sources for cloned dehydrogenases from gene banks made from tiieir genome. About twenty positive clones per bank were picked on the para-rosaniline plates. Plasmid DNA was isolated from all die positive clones, and transductants containing the unique plasmids were used for die further analysis that included ADH activity assay with edianol as a substrate. Strains harboring plasmid pBPP (containing d e cloned Horse Liver Alcohol Dehydrogenase - HLADH) were used as a control. Upon restreaking, and testing for dehydrogenase activity, die most active isolates were identified for further study.

Example 5. Production of Novel ADH Candidates

Production of ADH enzymes in shake flash for analysis. Strains LE392 harboring MYF-ADH plasmids were grown as 1 1 cultures in 21 shake flasks for 2 days in LB medium widi 100 g/ml ampicillin (Amp) at 30°C. Strains XLOLR/49.12 and ADOl were grown in LB-Ka_4o medium at 37°C overnight Cells were harvested by centrifugation, lysed using a Sonics & Materials homogenizer, incubated for 5 mi n. at 65°C in die 50 ml tubes (without temperature control inside of die solution). After centrigugation to remove the denatured protein debris die supernatant containing the purified protein was lyophilized.

Production of ADH enzymes infermenters. Strains LE392λ harboring MYF derivatives widi cloned adh genes were grown in a 17 liter fermenter (LH Fermentation, Model 2000 series 1) in 15 liters of LB medium widi 100 g/ml Ampicilin at 30°C overnight Strains XLOLR/49.12 and ADOl were grown in the LB medium widi 40 g/ml Kan at 37°C. Native strains were grown in 15 liters of TT brotii at 55-65°C. All cells were grown with approximate stirring at 250 rpm and 03 to 0.5 wm (volumes air/volume media per minute). Temperature is maintained by circulating water from a 28 liter water reservoir dirough hollow baffles within die stirred jars. The cells were disrupted by using two passes of an Avestin Emulsiflex C5 homogenizer between 10,000-15,000 PSI. The cells were tiien spun down and purified by heat treatment at 65°C for 5 min. widi temperature control inside of die solution. The sample was then spun down again in a centrifuge to remove die cell debris. The protein was lyophilized and characterized by substrate specificity assay. Example 6. Assay for ADH in cell extracts

Quantitative Assay. A standard method for die quantification of alcohol dehydrogenase based on NAD(P) utilization was used. Overnight cultures of cells to be assayed are grown in rich media. The cells are washed, resuspended in 600 μl of assay buffer (83 mM KH2PO4 [pH 7.3], 40 mM KCl, 0.25 mM EDTA), sonicated, and centrifuged. The assay mixture typically contained 50 μl of cell extract, 100 μl EtOH, 20 μl 100 mM NAD and/or NADP, 830 μl buffer and is carried out at room temperature. The assay measures a substrate dependent reduction of NAD(P)⁺ monitored by the change in absorbence at 340 nm in order to confirm dehydrogenase activity. Since crude lysates were expected to have background activity towards NAD(P)⁺, control experiments were performed to correct for spontaneous reaction with botii reduced and oxidized cofactors in the absence of added substrate. The reactions were run for approximately 3 minutes while continually measuring absorbence at 340 nM. This mediod produced a reliable quantitative determination of ADH activity present in the cell. Units of activity were calculated as μm A per minute of product formed. Specific activity is calculated as units per mg of protein used.

Example 7. DNA Analysis.

For the cloned candidate dehydrogenases, die DNA insert size was analyzed for comparison to each other. To analyze plasmids which were cloned using the pMYF system, an EcoRV digestion was used and to analyze plasmids which were cloned from die XLOLR/λZAP system, an EcoRI + Psti restriction digestion was used. Insert sizes for die key new dehydrogenase activities we have discovered during die course of this work are compiled in Table 4 below. It is interesting to note tiiat clone numbers 19 and 39 show nearly identical restriction patterns, yet they were obtained from colonies with si nificantiy different morphologies, and tiieir substrate specificities are different (see substrate characterization data below). This may indicate tiiat tiiey are highly related, but not identical enzymes.

Isolated plasmid clones which express the desired proteins of die present invention can be sequenced to determine die nucleic acid sequence for the expressed protein. By direct analysis of the DNA sequence it is possible to determine the start-codon for die initiation of transcription of die full-length transcript from which the amino acid sequence of die protein can be determined. Where tiiere may be more titan one possible start-codon, performing N-terminal amino acid sequence determination on isolated protein will allow for the identification of the proper start-codon. Methods for nucleic acid sequnencing, amino acid sequencing, protein isolation and die procedures for growing and/or manipulating cells, proteins and/or nucleic acids can be found in die general literature, for example see Sambrook et al., Molecular Cloning 2nd edition, 1989, Cold Spring Harbor Press.

Example 8. SDS-PAGE Gel Analysis.

In order to prepare samples for protein gel analysis, 50 ml of LB containing 100 l/ml ampicillin were inoculated widi different E. coli LE3921 cells containing ADH plasmids derived from pSL5 vector. The cultures were grown overnight at 30°C. The strains were named according to the native organism's DNA library number. 5, 7, 14, 19, 30, 39, 55, 69, 71, 98, 136 and 49.4 AD49.12 were clones derived from the pBK vector and transformed into E. coli XLOLR cells instead of LE392λ. XY, CA, CB, and CC are native cells that were found as new organisms on para-rosaniline selection plates (which may be mesophiles). The turbid cultures were spun down in 50 ml conical tubes at 3,000 rpm for 10 minutes. The clear supernatant was discarded and die cell pellets were washed by resuspending in ~1 ml of TE buffer and transferring die volume into 1.5 ml eppendorf tubes. The cells were pelleted by centrifuging at high speed for 2-3 minutes. The supernatant was tiien aspirated and die cell pellets were resuspended in 0.5-1.0 ml of ADH buffer (80 mM KH₂PO₄, 40 mM KC1, 0.25 mM EDTA, pH 7.3) and put on ice. The cell solution was sonicated twice for 20-30 seconds (output control at 4, percent duty cycle at 50, with pulsing). After sonication, die cell debris was spun down for 5 minutes. The clear supernatant was transferred to fresh eppendorf tubes and kept on ice.

Figure 3 depicts analysis of crude extracts from the ADH candidates. Crude extracts from ADH positive colonies were run on SDS gels. Fifteen microliters of each extract was mixed widi 15 μl of 2X SDS loading dye (100 mM Tris pH 6.8, 200 mM DTT, 4% SDS, 0.2% Bromophenol Blue, 20% glycerol). The mixtures were heated at 100°C for 5 minutes. Thirty microliters of each mixture was loaded onto an 8% SDS Gel (8% acrylamide, 375 mM Tris pH 8.8, 0.1% SDS, 0.1% ammonium persulfate, 10-12 μl TEMED/20 ml gel solution) widi 10-15 μl of molecular weight standard (Kaleidoscope Prestained Standards, Catalog number 161-0324, Bio-Rad). The gel was run for ~2 hours at 80-100 V and tiien stained with Coomassie Brilliant Blue (0.25g Coomassie Brilliant Blue R250 in 90 ml methanol.water (1: 1 v/v) and 10 ml glacial acetic acid) for 30-45 minutes and destained widi destaining solution (90 ml metiianol: water (1:1 v/v) and 10 ml glacial acetic acid). Labels 5, 7, 14, 19, 30, 39, 49.4, 49.12, 55, 69, 71, 98, 136, XY, CA, CB, CC refer to extracts made from strains AD5, AD7, AD 14, AD 19, AD30, AD39, AD49.4, AD49.12, AD55, AD69, AD71, AD98, AD136, XY (ADOl), CA, CB, and CC respectively. As can be seen, they did not show any outstanding protein bands which corresponded to the ADH gene. In order to further characterize the ADH content of die cells, Native protein gels and semipurified extracts were tested as described in die following section.

Example 9. Determination of Molecular Size of the Proteins by Gel Filtration Chromatography

The molecular size of the active form of die enzymes was determined by separation by size on a gel filtration column containing Sephadex S-200 (Pharmacia ) agarose gel. The column was equilibrated by the dehydrogenase enzyme assay buffer (83 mM potassium phosphate buffer, 43 mM Potassium chloride, 1 mM EDTA). The column conditions were, flow rate, 14.4 ml per hour. Sample volume, 200 μL and die column size 50 cm by 1 cm (BioRad Labs.), and bed volume,

-37.5 ml. The proteins were detected by absorption at 280 nm and for the enzymes the activity was monitored and, the fraction with highest activity was taken as the elution volume. Molecular sizes were calculated from the plot of Log of die known protein standards against the ratio of elution volume (V_e) to void volume (V_D) of the column. The ratio of Ve/Vo values obtained for die unknown proteins (dehydrogenases) was used to determine die molecular size from the standard plot developed for the proteins of known molecular size. The standard proteins used were, yeast alcohol dehydrogenase (150 kDa), bovine serum albumin (67 kDa), ovalbu in (49 kDa) and blue dextran (2200 kDa). When possible, subunit composition was determined by comparing active protein sizes of the fractions from die chromatography widi SDS Gel electrophoresis results of the semipurified protein. The results of this analysis is presented in die Summary Table 4 below.

Example 10. Method of Protein Characterization by migration on Native PAGE.

We were able to identify die active protein band in a crude extract by running d e sample on a native (nondenaturing) protein gel and staining with a phenazine-based method. The cell extracts were run on an 8% Native PAGE, which (which were identical to our 8% SDS PAGE gels described above widiout the SDS). Fifteen microliters of cell extract was mixed with 15 μ\ of 2X Native gel loading dye (100 mM Tris pH 6.8, 200 mM DTT, 0.27c Bromophenol Blue, 20% glycerol). The mixture was loaded onto the gel without boiling. The gels were run at 80-100 V for 2-3 hours. Labels 5, 7, 14, 19, 30, 39, 49.4, 49.12, 55, 69, 71, 98, 136, XY, CA, CB, CC refer to extracts made from strains AD5, AD7, ADM, AD19, AD30, AD39, AD49.4, AD49.12, AD55, AD69, AD71, AD98, AD136, XY (ADOl), CA, CB, and CC respectively. The activity stain was made either in 0.4-0.7% agarose or in liquid form (1 mg/ml nitro blue tetrazolium, 0.1 mg/ml phenazine metiiylsulfale (dissolved in 1.0-1.2 ml ethanol before adding to mixture), 0.25- 0.35 mM NAD, 0.25-0.35 mM NADP) and added to the top of native PAGE gels as shown in Figure 4. Approximately 30 ml was adequate for staining each gel. The activity staining reaction was run at 37°C in the dark. In just a few minutes dark bands began to appear for highly active enzymes.

Example 11. Optimum Temperature Analysis.

Lyophilized ADH samples were dissolved in ADH buffer (80 mM KJL-PO^ 40 mM KC1, 0.25 mM EDTA, pH 7.3) at 20 mg/ml and put on ice. The reaction mixture was made up of an appropriate dilution of die extract 30 l 0.1 M NAD or NADP, 100 μl ethanol, and measured up to 1 ml widi ADH buffer. The activities were measured on die spectrophotometer at 340 nm for 80 seconds at room temperature. The enzymes were assayed for activity at 30°C, 40°C, 50°C, 60°C, and 70°C. At each temperature, d e spectrophotometer chamber was heated by a circulating water temperature bath. The ADH buffer is also heated to die respective temperature, but the NAD or NADP and ethanol were at room temperature. The amount of tiiose reagents in the total volume is not enough to lower die temperature significantly. The 80 second activity reading is run 2-3 times consecutively to get an average. At 40°C, 50°C, and 60^βC, activity increases over several minutes. At 70°C, the activity is die highest for nearly all the enzymes, but die duration of the activity lasts only up to 5-7 minutes before die enzymes denature. The optimal temperature charts for enzymes produced from strains AD19, AD30, AD39, AD49.4, AD49.12, AD55, AD69, AD71, AD98, and strain XY are depicted in Drawing 6. The data is presented as specific activity. Reactions were run for 80 seconds at the various temperature, and generally reflect die initial rates of the reaction. Most enzymes begin to denature at around 60°C and have very short half lives at 70°C.

Example 12. Residual Activity Analysis.

Lyophilized samples were dissolved in ADH buffer (80 mM KH,PO₄, 40 mM KC1, 0.25 mM EDTA, pH 7.3) at 20 mg/ml. The enzymes were initially assayed at room temperature as a standard measurement. The samples were tiien incubated at 40°C, 50°C, and 60^βC for varying lengths of time. Plots for many of the enzymes we analyzed during die course of this work are depicted in Figure 7. As can be seen, most enzymes retain a significant amount of tiieir activity with very little loss in activity at 40°C. Example 13. Specific Activity and Optimal Cofactor Analysis.

The specific activity of each enzyme identified in a crude extract from shake flasks or from fermenters was analyzed using ethanol as the main substrate to determine relative reaction rates. In addition, botii NAD and NADP were tested as cofactors in order to determine die ideal cofactor. For all of the enzymes discussed in this report NAD was die optimal cofactor . Other enzymes preferred NADP, however activities were generally low and made it difficult to develop ti ese enzymes further. Since NAD is a more cost effective cofactor than NADP, we decided to preferentially focus on the NAD utilizing enzymes during die initial characterization. The optimized results from cultures which were grown in 15 L fermenters are listed in Table 4 at the end of tiiis section. In many instances these number are several orders of magnitude higher tiian die original specific activities identified in die crude extracts before growth optimization.

Example 14. General Substrate Specificity Analysis

Further characterization was done to assess die initial structure/function parameters of the active site of the newly discovered dehydrogenases and to investigate die stereochemical preference of die catalysts. This would help develop a reasonable initial set of enzymes widi a different range substrate specificities. The compounds and methods we have chosen to study in the feasibility portion of this project were chosen for tiieir ease of analysis. The general type of alcohol resolution reaction we tested is shown in die scheme below.

Scheme 1. Resolution of alcohols by dehydrogenase

During the course of this work we screened nearly 200 native organisms and 57 clone banks (approximately 5-10,000 individual clones from each bank on plate assay). Of these, we detected about 100 positive dehydrogenase activities from the native organisms and clone banks combined. After performing preliminary activity analysis similar to that which was described above for these positive activities, about 16 were chosen for a more complete study based on their overall performance and production level consistency. These enzymes and tiieir origin are listed in Table 3. Of these about a dozen showed unique dehydrogenase activity and were followed up in more detail - and a summary of their physical properties are shown below in Table 4. Table 3. Key to Main Strains and Initially Chosen for Follow-up

Strain Name Source Vector Resistance Comments

Strains*

#l-#246 Strains in the ThermoGen collection. Used for screening and construction of clone banks AA-XY Strains in the ThermoGen collection. Used primarily for screening

Naήve ADH candidates XY (ADOl) native Km^r, Apr ThermoGen strain collection

CA native Km^r. Ar ThermoGen strain collection

CB native Km^r, Ap¹ ThermoGen strain collection

CC native Km^r, Apr ThermoGen strain collection

Cloned ADH candidates

AD7 clone from strain #7 pTGAD14 Ap* λpMYF clone

ADM clone from strain #14 pTGAD14 Apr λpMYF clone

AD19 clone from strain #19 pTGAD19 Apr λpMYF clone

AD30 clone from strain #30 pTGAD31 Apr λpMYF clone

AD31 clone from strain #31 pTGAD14 Ar/ λpMYF clone

AD39 clone from strain #39 ρTGAD14 Apr λpMYF clone

AD49.4 clone from strain #49 pTGAD14 Apr λpMYF clone

AD49.12 clone from strain #49 pTGAD14 Km¹ pBK clone

AD55 clone from strain #55 pTGAD14 Ap¹ λpMYF clone

AD69 clone from strain #69 pTGAD14 Ap λpMYF clone

AD71 clone from strain #71 pTGAD14 Apr λpMYF clone

AD98 clone from strain #98 pTGAD14 λpMYF clone

^♦approximately 200 ThermoGen strains total were screened

Table 4. ADH Physical Properties Summary Table

Enzyme mw DNA useful Temp Sp. Act*, tl/2 (KDal) insert Range (u/mg) 40^βC (hrs)

(kb)

ADM αl 16.8 RT- 0^βC 28 txl

AD19 105 6 RT-40°C Dd 20

AD30 121 16.4 RT 0°C 4 24

AD31 c 11.8 RT- 0°C 1.45 nd

AD39 73.6 6 R 40^βC 60 15

AD49.4 78.2 8.4 RT-40°C 16 15-20

AD49.12 121.3 3.3 RT-50°C 306 >50

AD55 97.3 9.3 RT^0^βC 30 20

AD69 121.3 12.8 RT 0^βC 30 20

AD71 78.1 7.4 RT-40^QC 136 12

AD98 78.1 12 RT^0^βC 80 28

XY (ADOl) 78-131 NA RT 0^βC 460 >50

^♦Specific Activity is calculated from one run in the 15 Uter fermenter using ethanol as a substrate. Reactions were run at room temperature. Extracts from the clones were heat purified crude extracts, nd - not determined We tested die substrate specificity of the newly discovered catalysts on a variety of alcohols which could easily be analyzed spectrophotometrically. Table 5 lists relative activities of the ten enzymes we chose from die previous section of the work to study against a series of alcohols. The activities presented are relative to tiieir activity on edianol . The data which is presented has been compiled from several independent runs of enzyme in order to verify repeatability of the ratios which have been presented here.

Table 5. General Alcohol Preference

As can be seen, several of die enzymes prefer alternative alcohols compared to edianol. Very few have a high degree of activity on cyclohexanol relative to edianol, but enough activity is present to detect in the short reaction.

Example 15. Enantioselectivity of the Dehydrogenases

To obtain data on chiral compounds, we utilized chiral alcohols and measured relative rates on the two different enantiomers of substrates of interest This type of data is presented in Table 6. Again, for comparison, the data is presented as activity relative to using ethanol as d e substrate. Two pairs of chiral alcohols were tested: R- and S-butanol; and R- and S-pentanol.

The data clearly indicates different preferences for the enzymes. Most of the enzymes prefer the S- alcohols, however the XY dehydrogenase prefers R-pentanol over S-pentanol neariy six times more. Interestingly enough, it does not have the same degree of selectivity for die butanol compound. Several of the otiier enzymes prefer the S-enantiomer up to twenty times higher than die R-enantiomer. Most enzymes are eitiier as selective or more selective using pentanol as a substrate versus butanol, however, die AD49.4 is more selective with the butanol. It is comforting to see tiiat only one enzyme in this test, AD49-12, is not highly selective on at least one of die two compounds. This underscores die fact tiiat we can find selective catalysts using die methods employed during tiiis project Another interesting note is that AD49-4 and AD49-12 both were isolated from the same clone bank (created from strain #49) yet are clearly different dehydrogenase activities.

Table 6. Chiral Alcohol Selectivit

Summary. After having screened nearly 200 new organisms and over 50 clone banks we have isolated and performed characterization of about a dozen new dehydrogenases. The enzymes of the present invention have demonstrated synthetic utility as dehydrogenases by identifying R- and S- selective catalysts.

"y. Example 16. Heat Denaturation Purification for Production of Enzyme

Tables 7A and 7B below list die results for characterizations of die Alcohol dehydrogenase specific activities (u= mol/mg of the total protein) in die ADH producing strains after heat treatment at 55°C.

Table 7 Heat ; purified I ADH activity

Specific activities (u=umol/m_ proU % Activity

Strain 55°C heat inactivation

0 min. 15 min. 30min. 60 min. 90 min. 0 min. 15 min. 30 min. 60 min. 90 min.

AD7-1 9.0 3.5 2.0 6.8 0.0 100% 39% 22% 76% 0%

AD14-1 14.0 60.0 60.0 0.0 0.0 100% 429% 429% 0% 0%

AD31-3 4.7 14.0 10.5 0.0 0.0 100% 298% 223% 0% 0%

AD39-4 67.0 149.0 127.0 39.0 24.0 100% 222% 190% 58% 36%

AI549-4 36.5 22.0 15.0 8.1 0.0 100% 60% 41% 22% 0%

AD49-12 2170.0 1830.0 1060.0 597.0 450.0 100% 84% 49% 28% 21%

AD55-1 3.0 5.0 4.8 0.0 0.0 100% 167% 160% 0% 0%

AD83-5 21.0 28.0 23.0 6.5 0.0 100% 133% 110% 31% 0%

XY 1680.0 4450.0 7822.0 1048.0 822.0 100% 265% 466% 62% 49%

AD5-1 35 75 50 180 100% 214% 143% 514% 0%

Strains were grown in 1 liter of the LB _Amp (°^r LBκ_n for die 49-12 and XY) at 30°C overnight, washed in TE buffer, concentrated x20 times in assay buffer and homogenized in EMulsiFlex homogenizer. Activities were assayed with ethanol as a substrate; cofactors tiiat have been used were: NAD for the strains #14-1, 39-4, 49-4, 49-12 and XY, and NADP for the #7-1, 7-2, 31-3, 55-1, 83-5; Protein concentrations are given in mg/ml.

Table 8. Total ADH activities (u/1 liter) in the grown cultures without and with heat treatment.

Total ADH activities (u=/πnol, 1 Uter) with and without heat treatment strain 55oC heat inactivation

0 min. 15 min. 30 min. 60 min. 90 min. 0 min. 15 min. 30 min. 60 min. 90 min.

AD7-1 730 141 72 170 0 100% 19% 10% 23% 0%

AD14-1 3,400 1935 1250 850 100% 57% 37% 25% 0%

AD31-3 1050 504 314 0 0 100% 48% 30% 0% 0%

AD39^ 5700 4500 3200 770 360 100% 79% 56% 14% 6%

AD49^ 2400 968 532 242 0 100% 40% 22% 10% 0%

AD49-12 500000 110000 58000 30000 20000 100% 22% 12% 6% 4%

AD55-1 300 101 72 0 0 100% 34% 24% 0% 0%

AD83-5 2100 706 580 97 0 100% 34% 28% 5% 0%

XY 286000 111300 78200 52400 0 100% 39% 27% 18% 0%

AD5-1 3500 1130 900 850 100% 32% 26% 24% 0%

Literature Cited

Reagent cost estimated from prices found in the Sigma catalog.

( 1994) Single-enantiomer version of antiasthma drug patented. Chem. and Eng. News. 72:45. Benoiton, L., M. Winitz, S. M. Bimbaum and J. P. Greenstein. (1957) Studies on diastereomreic a-amino acids and corresponding a-hydroxy acids. IX. Configuration of the isomeπc g-hydroxyglutamic acids. J. Am. Chem. Soc. 79:6192. Bradshaw, C. W., W. Hummel and C.-H. Wong. (1992) Lactobacillus kefir alcohol dehydrogenase: a useful catalyst for syndiesis. J. Org. Chem. 57:1532. Casy, G., T. V. Lee and H. Lovell. (1992) Tetrahedron tt. 33:817.

Conway, T., G. W. Sewell, Y. A. Osman and L. O. Ingram. (1987) Cloning and Seqiencing of the Alcohol Dehydrogenase 11 Gene from Zymomonas mobilis. J. Bact. 169:2591-2597. Crans, D. C, R. W. Marshman, R. Nielsen and I. Felty. (1993) NADV: A New Cofactor for Alcohol Dehydrogenase from Thermoanaerobium brockii. J. Org. Chem. 58:2244. Crout, D. H. and M. Christen. (1989) Modern Synthetic Methods. 5:1-114.

Csuk, R. and B. 1. Glanzer. (1991) Baker's Yeast Mediated Transformations in Organic

Chemistry. Chem. Rev. 91:49-97. Daniel, R. M., J. Bragger and H. W. Morgan. Enzymes from Extreme Environments. In Biocatalvsis. Ed. by D. A. Abramowicz, Van Nostrand Reinhold, New York. p. 243. 1990. Dodds, D. R. and J. B. Jones. (1988) Enzymes in organic syndiesis. 38. Preparations of enantiomerically pure chiral hydroxydecalones via stereospecific horse liver alcohol dehydrogenase catalyzed reductions of decalindiones. J. Am. Chem. Soc. 1 10:577-583. Faber, K. Biotransformations in Organic Chemistry. Springer- Verlag, Berlin. 1992. Fontana, A. Thermophilic Enzvmes and tiieir Potential Use in Biotechnology.. Dechema, Weinheim. 1984.

Hummel, W. (1990) Biotechnol. Lett. 12:403.

Hummel, W., H. Schutte, E. Schmidt, C. Wandrey and M. R. Kula. (1987) Appl. Microbiol.

Biotechnol. 26:49. Jacovac, 1. J., H. B. Goodbrand, K. P. Lok and J. B. Jones. (1982) Enzymes in organic synthesis. 24. Preparations of enantiomerically pure chiral lactones via stereospecific horse liver alcohol dehydrogenase catalyzed oxidations of monocyclic meso diols. J. Am. Chem. Soc. 104:4659-4665. Jones, J. B. (1986) Enzymes in organic Synthesis. Tetrahedron. 42:3351-3403. Jones, J. B. and I. J. Jakovac. (1982) Can. J. Chem. 60: 19. Jones, J. B. and I. J. Jakovic. (1982) A new cubic space section model for predicting the specificity of horse liver alcohol dehydrogenase-catalyzed oxidoreductions. Can. J. Chem. 60: 19. Keinan, E., E. K. Hafeli, K. K Seth and R. Lamed. (1986) Thermostable Enzymes in Organic Synthesis. 2. Asymmetric reduction of ketones widi alcohol dehydrogenase from diermoanaerobium brockii. /. Am. Chem. Soc. 108: 162-169. Lamed, R. and J. G. Ziekus. (1981) Novel NADP-linked alcohol aldehyde/ketone oxidoreductase in tiiermophilic etiianologenic bacteria Biochem. J. 195: 183- 190. Lemiere, G. L. Alcohol Dehydrogenase Catalyzed Oxidoreduction Reactions in Organic Chemistry. In Enzymes as Catalysts in Organic Synthesis. Ed. by M. P. Schneider, Reidel,

Dordrecht p. 17. 1986. March, J. Advanced Organic Chemistry. John Wiley and Sons, New York. 1985. Neuberg, C. (1949) Biochemical reductions at the expense of sugars. Adv. Carbohydr. Res. 4:75. Pham, V. T. and R. S. Phillips. (1990) /. Am. Chem. Soc. 1 12:3629. Prelog, V. (1984) Pure Appl. Chem. 9: 119.

Seebach, D., M. A. Sutter and R. H. Weber. (1984) Org. Synth. 63: 1-. Servi, S. (1990) Baker's Yeast as a Reagent in Organic Syndiesis. Synthesis. 1-25. Stinson, S. C. (1994) Chiral Drugs. Chemical and Engineering News. 72:38-72. van Eys, J. (1961) Aldehyde-ketone isomerization activity of liver alcohol dehydrogenase. J. Biol. Chem. 236:1531.

Claims

We Claim:

1. A protein with alcohol dehydrogenase activity selected from the group consisting of AD55.1, AD83.5, AD5.1, AD7.1, AD14.1, AD31.3, ADR AD19, AD30, AD31, AD39, AD49.4, AD49.12, AD55, AD69, AD71, AD98, and XY (ADOl).

2. A protein widi diermostable alcohol dehydrogenase activity selected from the group consisting of AD7.1, AD14.1, AD31.3, AD39.4, AD49.4, AD49.12, AD55.1, AD83.5, XY, and AD5.1.

3. A protein with chiral alcohol selective alcohol dehydrogenase activity selected from the group consisting of AD19, AD30, AD39, AD49.4, AD49.12, AD55, AD69, AD71, AD98, and XY(ADOl).

4. An expressible recombinant DNA construct which contains a DNA sequence which when expressed encodes for a protein of claim 1.

5. A recombinant DNA which hybridizes widi a recombinant DNA of claim 4.

6. A recombinant DNA which encodes for the amino acid sequence of die protein of claim 1.

7. A DNA sequence which hybridizes widi die DNA of claim 6 under high stringency conditions.

8. A cell transformed with the construct of claim 4.