WO1998002556A2 - Screening for and use of an esterase for a stereospecific resolution - Google Patents

Abstract

A method of screening for microorganisms capable of the resolution of compounds of formula (I), which method comprises selecting for organisms capable of using compounds of formula (I) as a sole carbon source and assaying cells of the organism for specific esterase activity, can be used to identify enzymes useful in the manufacture of paroxetine.

Description

SCREENING FOR AND USE OF AN ESTERASE FOR A STEREOSPECIFIC RESOLUTION

The present invention relates to an enzyme useful in the resolution of isomers of a compound of formula (I), processes for preparing the enzyme, DNA coding for the enzyme and bioprocesses using the enzyme.

Me ( I ) where R is C 1-6 alkyl

The present invention also provides a novel screening method and a novel organism isolated from this screen containing the enzyme of the present invention.

International Application WO 93/22284 (SmithKline Beecham) the content of which is herein incorporated by reference, claims the resolution of compounds of formula (I) using an esterase enzyme and specifically discloses the use of pig liver or bovine liver esterase. However the use of esterases from animal livers is less ideal for large scale commercial operations which require an unlimited supply of enzyme. WO 94/03428 (Sepracor) describes a biocatalytic method for preparing optically pure precursors of paroxetine using an isolated Achromobacier sp. and an enzyme extract isolated therefrom having an optimum pH of between 7 and 8. WO 94/03428 (Sepracor) also outlines a method for identifying suitable esterases for the synthesis of paroxetine from isolated organisms belonging to the genus Achromobacier. In practice there are a number of problems with the latter identification method which is based on an 'overlay' approach. One problem is clumping of the substrate and the length of time (some several weeks) to produce a 'clearing zone' . We now provide a more specific screening procedure for organisms having the desired activity which overcomes these problems and is moreover particularly suitable for using with any source of organism whether from individual isolated organisms (eg. from a culture bank) or from any biological sample containing a mixture of microorganisms eg. soil. Accordingly the present invention provides a method of screening for microorganisms capable of the resolution of (I) by selecting for organisms capable of using paroxetine esters as a sole carbon source and then assaying cells of the microorganism, suitably using HPLC, to identify those having specific esterase activity. Other advantages of the method of the present invention are that it is more reliable and specific in that it does not rely on a clearing zone. It is a more sensitive and direct approach.

We have now found that other microbial enzymes suitable for use in the process of the invention can be derived from certain organisms by using the abovementioned screen. Particularly suitable microorganisms include organisms belonging to the genus Alcaligenes.

The strain NCIMB 40700 is a novel microorganism and as such forms a further aspect of the present invention. This strain has been deposited at the National Collection of Industrial and Marine Bacteria (NCIMB), Aberdeen. Scotland under accession number NCIMB 40700 on I December 1994. The characteristics of the organism are described in Example 5.

We have isolated an enzyme from NCIMB 40700 which is capable of carrying out the above reaction. In a further aspect of the invention there is provided a process for the resolution of (I) which comprises contacting (I) with the enzyme. Accordingly the present invention provides a process for stereospecifically hydrolysing a mixture of the ( + ) and (-) isomers of a compound of formula (I):

^

(ID

MR in which R is Cj_6 alkyl; using microorganism NCIMB 40700 or a microorganism identified using the screen of the invention or an enzyme isolated from said microorganisms (i) to form a compound of formula (II A), O N O

Me (+) isomer ( I IA ) and thereafter separating the resulting compound of formula (IIA) from the remaining (- ) isomer of formula (I); or ii) to form a compound of formula (IIB):

O O

Me (-) isomer ( I IB ) and thereafter separating the resulting compound of formula (IIB) from the remaining ( + ) isomer of formula (I). Process variant i) is preferred.

When using process variant (i) the stereoselectivity of the process is such that 0 from a racemic mixture of a compound of formula (I), after the action of the enzyme the ratio of (-) to ( +) isomer of formula (I), is greater than 60% , preferably greater than 70% , more preferably greater than 80% and most preferably greater than 85% .

When using process variant (ii) the stereoselectivity of the process is such that from a racemic mixture of a compound of formula (I), after the action of the enzyme, b the ratio of ( + ) to (-) isomer of a compound of formula (I), is greater than 60% , preferably greater than 70% , more preferably greater than 80% and most preferably greater than 85 % .

The process is suitably carried out by dissolving the ( ± ) unresolved compound of formula (I) into a suitable solvent such as an aqueous/organic solvent mixture and 0 adding the enzyme and stirring the resulting mixture until the reaction is completed.

Suitable temperatures for performing the reaction include 0-50°C more suitably 10-40°C and yet more suitably 25 to 35°C and most suitably at 30°C. Suitable aqueous/organic solvent mixtures include buffered aqueous solvents such as tris buffer which is mixed with DMSO. Suitable pH's for the reaction to be carried out include pH 4 to 8, more suitably pH 5 to 7 and preferably at pH 5.5. Suitable values for the variable R include methyl and ethyl. Preferably R is ethyl. In process variant i) where the carboxyl esterase enzyme stereospecifically hydrolyses the ( + ) form of a compound of formula (I), the remaining (-) form of a compound of formula (I) is separated by conventional techniques such as solvent extraction of the compound of formula (I) using a non-aqueous miscible solvent such as ethyl acetate and the resulting (-) compound of formula (I) may then be isolated using conventional techniques such as precipitation.

The present invention also extends to a process for subsequently converting the (-) compound of formula (I) prepared as described above to paroxetine or a pharmaceutically acceptable salt and/or solvate thereof such as the hydrochloride hemi- hydrate, for example, using the procedures outlined in US Patent No. 4,902,801 and US Patent 4,721 ,723 and the hydrochloride anhydrate using the procedures outlined in WO96/24595.

In process variant ii) where the carboxyl esterase enzyme stereospecifically hydrolyses the (-) form of a compound of formula (I) to yield the (-) form of a compound of formula (II), the remaining ( + ) form of a compound of formula (I) may be separated from the (-) form of a compound of formula (II) as mentioned above. The (-) compound of formula (II) may be converted to paroxetine by first converting it to a (-) compound of formula (I) using conventional esterification techniques. The ester may then be converted to paroxetine or a pharmaceutically acceptable salt and/or solvate thereof such as the hydrochloride hemi-hydrate, for example, using the procedures outlined in US Patent No. 4,902,801 and US Patent

4,721 ,723.

Alternatively, the (-) compound of formula (II) may be directly converted to paroxetine by reducing the carboxylic acid group to a hydroxymethyl group and reducing the two keto groups in the piperidine ring using conventional reducing agents such as lithium aluminium hydride. Subsequently, the resulting piperidine carbinol compound may be converted to paroxetine or a pharmaceutically acceptable salt and/or solvate thereof such as the hydrochloride hemi-hydrate, for example, using the procedures outlined in US Patent No. 4,902,801 and US Patent 4,721 ,723.

Compounds of formula (I) may be prepared according to the procedures outlined in US Patent 4,902,801

This transformation is particularly useful in the resolution of imide esters in the synthesis of paroxetine in which the novel esterase is capable of resolving paroxetine methyl ester and /or ethyl ester to give an enantiomer excess of the (-) isomer greater than 90% .

The medium for cultivating the microorganisms isolated according to the invention suitably contains sources of assimilable carbon and nitrogen together with inorganic salts and is suitably a rich medium eg. AJ3B (see example 4).

The process of the present invention can be carried out using whole cells, cell free extracts, permeabilised cells or the isolated enzyme from the microorganisms or any of these in immobilised form. Where the biotransformation is carried out using whole cells, the microorganism may be in the form of a growing culture, resting culture, washed mycelium, immobilised cells or protoplasts.

When cell-free extracts are used these are suitably produced by shear and/or chemical or enzymic lysis or other methods of disruption, preferably high pressure homogenisation and optionally thereafter removing cell debris, leaving the enzyme activity in solution. The enzyme is suitably prepared according to the examples below and forms a further aspect of the present invention. The enzyme may be prepared by culturing the microorganism in a conventional manner, especially under aerobic conditions in a suitable liquid or semi-solid medium. The culture conditions may be a temperature in the range from 15 - 35°C and pH in the range 6-8, most preferably 6.8. The enzyme may be isolated either from NCIMB 40700 or preferably from a suitable recombinant source, for example as described below, and used in purified form, partially purified form, as obtained in an impure state, as a filtrate from a disrupted cell preparation , or as a crude cell homogenate. Most suitably the enzyme is, for example, at least purified to remove other enzymes which might also catalyse the destruction of the starting material or the enzyme.

In a further aspect of the invention there is provided an enzyme capable of resolving compounds of the formula (I) obtainable by treating cells from NCIMB 40700, according to example 6 below, ie. by centrifugation and cell disruption (high pressure homogeniser) followed by fractionation and chromatography . The active enzyme is believed to exist as a dimer of two equal subunits each having a molecular weight of approximately 35,000 (by LCMS) and has other characteristics as indicated in Example 21 below. Preferably the enzyme has substantially the sequence shown SEQ ID NO:2. The term substantially means at least 85 % homologous, preferably 90 - 95% homologous, more preferably greater than 95% homologous. The present invention also provides for an enzyme having an αβ-hydrolase fold wherein the acidic residue of the catalytic triad appears at the end of strand 6, and which has two adjacent cysteine residues close to the active site that are linked together by a disulphide bond. In a further aspect the present invention also provides DNA encoding a protein capable of resolving compounds of formula (I). In particular the DNA comprises substantially the DNA sequence in SEQ ID NO: l .

The invention also encompasses DNA sequences which hybridise, preferably under conditions of high stringency, with the DNA of SEQ ID NO: 1 or fragments thereof. Suitable stringency conditions are, for example, 0.1 %w/v SDS, O. lxSSC at 60°C.

Preferably the DNA is comprised in a suitable vector. In yet a further aspect of the invention there is provided a host transformed by a vector according to the invention. Preferred hosts are E. coli and Agrobacterium, especially Agrobacterium.

Suitably the enzyme according to the invention is immobilised for example to an insoluble support material such as by the procedures discussed by Powell (1990) in Microbial Enzymes and Biotechnology ed. Fogarty and Kelly p369-394. This provides the advantage of increased yield and throughput. When the process is carried out using whole cells, the incubation medium comprises a medium such as medium A (see below). When the process is carried out using cell free extracts the incubation medium comprises a suitable buffer. The pH of the reaction is typically between 6 - 8. The temperature of the reaction should generally be in the range 22-48 °C preferably 37 - 47 °C, most preferably 47 °C. The reaction is typically carried out in aqueous medium , although alternatively the reaction can be carried out in the presence of organic solvents eg. acetone, toluene, methyl isobutyl ketone (MIBK). The reaction time depends on such factors as concentrations of reactants and cofactors, temperature and pH. After the reaction is complete the product can be isolated by conventional methods and as shown in the examples below. The initial purification conveniently involves solvent extraction or filtration.

Description of the figures

Figure 1. Basic restriction map of the #713 genomic DNA fragment common to cosmids pROXl , pROX2 and pROX3. Figures 2 - 8 Plasmid maps (general features): the shaded area corresponds to the coding region of the esterase, Ptrc refers to the trc promoter, rrnBTιT2 to the rRNA transcriptional terminator, AmpR to ampicillin resistance, pBR322ori is the replication origin and LacN refers to an overexpressing lac repressor gene. Plasmid map of pROX6 Plasmιd map of pROX12 Plasmid map of pROX 18 Plasmid map of pROX29

Plasmid map of pROX24

Plasmid map of pROX26 (Cm^ refers to chloramphemcol resistance)

Plasmid map of pROX36 (Cm* refers to chloramphemcol resistance)

The following examples illustrate the invention

Example 1 - Primary Selection Screen

A selection screen was developed to isolate organisms capable of utilising paroxetine stage 1 esters (defined as a mixture of compounds with formula I where R = CH₂CII₃,85 %w/w, and R = CH₃, 15 %w/w) as sole carbon source Organisms were grown in minimal media with paroxetine stage 1 as the only carbon source in order to isolate those organisms having enzymes capable of modifying this substrate Sub- populations of these survivors were examined for those with esterases which would remove either the methyl and/or ethyl ester from the stage 1 paroxetine with chiral bias to the ( + )-ιsomer UK and foreign soil samples were obtained Small (micro-spatular) portions of these samples were added to 100ml shake-flask holding lOml of minimal medium A containing K₂HPO₄ lg/1, KH₂PO₄ 3H₂O, 3g/l, MgSO₄ 0 5g/l, (NH₃)₂SO₄ 7H₂0 2g/l, FeSO₄ 7H₂O lOmg/1, MnCl₂ 4H₂O lOmg/l, ZnSO₄ lOmg/l with Paroxetine Stage 1 at lOg/1 added as a solid after sterilisation These flasks were incubated at 26^υC on an orbital incubator for seven days after which time 1ml samples were removed and used as an inoculum for second set of identical flasks This subcultuπng was repeated for a second time and after 2 days incubation 10μl loop samples were streaked onto GM agar plates containing glycerol 7 5g/l, molasses 7 5g/l, yeast extract 5g/l, MgSO₄ 7H₂O 0 05g/l, KH₂PO4 6mg/l, NH₃FeSO₄ 12H₂O 16mg/l, CuSO₄ 7H₂O lOmg/l, CaSO4 0 25g/l and agar No 3 25g/l to give single colonies

After six days incubation at 26°C single morphologically different colonies ( 1 or 2 per plate) were picked and streaked out onto GM slopes for incubation at 26°C for seven days followed by storage at 4°C Each isolate was then prepared for the secondary screen by plating out as a lawn on GM agar induction plates containing 1 % paroxetine stage 1 (w/v) added as a suspended solid The substrate was added to this medium in case any esterases required induction A total of 1800 soil samples were processed in the primary screen which resulted in the isolation of 211 1 organisms which were analysed for esterase activity in the secondary screen.

Example 2 - Secondary Selection Screen

5 The assay for the secondary screen was designed to identify which isolated organisms could hydrolyse either the ethyl or methyl esters contained within paroxetine stage 1 to give the corresponding carboxyiic acid (Trans-3 carboxy-4-(4'-fluorophenyi)- N-methylpiperidine-2,6 dione) . Cells were scraped off GM agar induction plates and suspended by vortex mixing in 1ml of paroxetine stage 1 assay mixture containing 0 0.16M Tris.HCl, 20%v/v DMSO and 0.1 % w/v paroxetine stage 1 ester at pH 5.5. The reaction was incubated overnight at 30°C and the cells removed by centrifugation for 4min at 14,000 rpm (Eppendorf micro-centrifuge). The supernatant was filtered through a 0.45μm filter and 20μl injected onto a reversed phase column (Partisphere, C18, 5μm, 110mm x 4.7mm diameter) to measure paroxetine stage 1 acid. Separation

: 5 was achieved using gradient elution by mixing buffer A containing 0.05M ammonium acetate, 10% acetonitrile, pH 4.0 and buffer B containing 0.05M ammonium acetate, 60% acetonitrile, pH 4.0 by the following profile in Table 1.

Table 0

HPLC Equipment

Detector, Waters 490E

System Controller, Waters 600E

25 Injecter, Waters Autosampler 717 Approximately 50 isolates showed peaks which were consistent with carboxyiic acid product formation (retention time approximately 3.5min) but most of these were at extremely low levels. Effort was concentrated on the analysis of the top 15 most active isolates. The authenticity of the acid peak from these was verified by spiking and co- elution with the acid standard obtained from the PLE reaction (International Application WO 93/22284 SmithKline Beecham).

Example 3 - Tertiary Screen

The tertiary screen was designed to identify which of the carboxyiic acid producing organisms had chiral selectivity. The reactions were conducted in the same way as the secondary screen except that either the pure ethyl or methyl ester of compound I was used depending on the preference of the enzyme. Reaction supernatent (800μl) was vortex mixed for 10s with 500μl of ether. The phases were split by centrifugation at 14,000rpm for 10s and the top ether layer removed. The ether extract (lOOμl) was then separated using a Chiralcel OJ column based on cellulose 4-methyl benzoate on silica (Diacel Chemical Industries, 25cm x 4.6mm diameter) by elution with hexane containing 25 % v/v ethanol (lml/min). Resolved isomers were detected at 265nm. Chiral selectivity was expressed in terms of the ratio of the + and - esters.

Example 4 - Microbiological Analysis and Storage of active Isolates

Isolates were inspected for purity by plating out to single colonies on GM and TSA media and incubating at 26°C. Single colonies from pure plates were used to inoculate 250ml flasks containing 20ml of AJ3B (yeast extract 25g/l, NaH₂PO₄ 2g/l, MgSO₄ .711 ₂O lg/1, MnSO₄ .4H₂O O.Olg/l, FeSO₄ .7H₂O 0.01g/l, CaCl₂ .2H₂O 0.08 g/I, glycerol 15g/l, pH 6.8) and incubated overnight. 0.5ml of each culture was mixed with an equal volume of sterile 40% glycerol in cryotubes and frozen at -70°C. Stored cultures were analysed for purity and viability by plating out onto both GM and TSA.

Example 5 - Identification and characteristics of isolate NCIMB 40700 Appearance on GM plate; beige, circular, 1-3 mm, smooth, shiny, domed; in

AJ3B shakeflask medium; cloudy beige dispersed growth. Suitable growth conditions are AJ3B liquid media at 26°C for 1-2 days, inoculum is best from fresh agar plates or slopes using GM plate medium at 26 °C for 4 days. The organism is apparently motile at 37°C, yellow pigmented and oxidase negative and produces acid from few carbohydrates.

Partial 16S rDNA sequence analysis of isolate NCIMB 40700 was achieved by direct sequencing of PCR amplified 16S rDNA. The sequence data was compared with over 1500 sequences of members of the domain Bacteria. Similarity values showed that NCIMB 40700 is most closely related to Alcaligenes xylosoxidans (97.7% sequence similarity). The next highest similarity value was found to be for Alcaligenes faecalis (95.5 %). Although the strain was not identifiable to the species level this data places NCIMB 40700 in the Beta subclass of the Proteobacteria.

Example 6 Purification of esterase from the soil isolate NCIMB 40700 Cell disruption

2001 of NCIMB 40700 AJ3B fermentation broth (60h) was harvested using a desludging centrifuge (West Falia Separator Ltd. , Milton Keynes, UK.) to give 451 of cell suspension. The cells were then homogenised by four passes through a continuous homogeniser (A.P.V. Gaulin BV, S-Gravelandseweg, Hilversum, Holland) at 10,000 psi with temperature control below 20°C.

Enzyme capture and concentration

The homogenate was then diluted 1 : 1 with distilled water and stirred with 13.1kg of WK10 ion-exchange resin (Mitsubishi Kasei Corporation, Chiyoda-Ku, Tokyo, Japan). Enzyme uptake was completed in 3h with continuous control of pH at 6.5 using 5M NaOH. The loaded resin was removed by filtration and washed twice with 801 of distilled water. Esterase activity was then eluted from the resin by stirring with 801 of 1M NaCl for 90min with pH control at 7.5 using 5M NaOH. The enzyme was then concentrated to 61 by ultra-filtration (Spiral wound 10,000mw cut off, Amicon) and stirred for 16h with 515g of A568 resin (Rohm Haas, Chavny, France) with pH control at 7.0 using 5M NaOH. The resin, which had adsorbed unwanted protein, was removed by filtration to give 5.51 of clarified enzyme solution.

Phenyl Agarose chromatography.

Post A568 (41) extract was concentrated by ultra-filtration (Spiral wound 10,000mw cut off, Amicon) to 1.81. The concentrated extract was adjusted to pH 6.5 by adding 250ml of 0.39M sodium phosphate buffer to give a final buffer concentration of 50mM. This solution was stirred on ice and 396g of (NH₄)₂SO₄ (SigmaUltra) was added slowly to give a final (NH₄)₂SO₄ concentration of 1.3M. The solution was then re-adjusted to pH 6.5 with M NaOH and stood overnight at 8°C. Precipitated protein was then removed by centrifugation at 10,000rpm for 20min. Esterase activity was then adsorbed onto a Phenyl Agarose 6XL column (11.5cm x 5.0cm diameter, Affinity

Chromatography Ltd, Isle of Man, UK), at a flow rate of 20ml/min. This column had been previously equilibrated with 900ml of 50mM sodium phosphate buffer (pH 6.5) containing 1 5M (NH )₂SO₄ (buffer C) Un-bound protein was then removed by washing the loaded column with 1 21 of buffer C The esterase was then eluted using a linear gradient from 100% buffer C to 100% buffer D (50mM sodium phosphate, pH 6 5) over 21 80ml fractions were collected and assayed using the standard DMSO containing assay method given in example 2 Active fractions 24, 25 and 26 were pooled and dialysed against 51 of water overnight (room temp)

Q-Sepharose Chromatography

The dialysed extract from the phenyl agarose purification step was adjusted to pH 8 6 with a saturated solution of Na₂HPO After filtration (0 45μm) the solution was passed through a Q-sepharose ion exchange column at 3ml/mιn (Hiload HP, 10cm x 1 6cm diameter, Phamacia) which had been previously equilibrated with 25mM sodium phosphate buffer at pH 8 8 Esterase activity was not retained by this column but purification was achieved by binding of unwanted protein

S-Sepharose Chromatography

The active fraction from the Q-sepharose purification step (320ml) was adjusted to pH 7 0 using a saturated solution of NaH₂PO₄ After filtration (45μm), the esterase was bound to a S-sepharose ion exchange column at 3ml/mιn (Hiload HP, 10cm x 1 6cm diameter, Pharmacia) which had been previously equilibrated with buffer E containing 25mM sodium phosphate at pH 6 8 The loaded column was then washed with 100ml of buffer E The esterase was then eluted using a linear gradient from 100% buffer E to 100% buffer F (25mM sodium phosphate, pH 6 8, 1M NaCl) over 300ml at 3ml/mιn 7 8ml fractions were collected and assayed using the DMSO containing method Two overlapping peaks of activity were observed representing at least two differently charged active species Fractions 9, 10, 1 1 and 12 were pooled to give active species I and fractions 13 and 14 were pooled to give species II

Blue Sepharose Chromatography The active esterase species I was further purified use blue sepharose affinity chromatography 0 6ml of the pooled fraction from the S-sepharose purification step was diluted by adding 7 5ml of buffer E 7 5ml of this mixture was loaded onto a HiTrap blue affinity column at lml/min (5ml, Pharmacia) which had been previously equilibrated with buffer E The loaded column was then washed with 20ml of buffer E The esterase was then eluted using a linear gradient from 100% buffer E to 100% buffer F over 30ml at lml/min 1ml fractions were collected and assayed The UV elution profile showed resolution of two active species, with the first elut g peak, species I, at approximately twice the concentration of species II. Fractions 1 1 to 16 were pooled to give pure species I and fractions 17 to 21 were pooled to give species II. This process was repeated for the pooled fraction containing mainly species II from the S-sepharose column. This separation gave predominately species II in fractions 23 to 29 which were pooled to give pure species II.

Analysis of species I and II by LC-MS revealed more complex mixtures with both of these species containing more than one molecular mass. Some of these forms can be explained by differences in the N and C termini as given in Table 2.

Table 2. Modification of C- and N- termini of esterase

Preparation N-terminus (Determined by amino acid C-terminus Mass and sequencing) (from gene change in mass sequence)

0) E. coli pROXlό A P A A P P P V P K A H G R 35882 c

CD (fl -A(70)

H

H NCIMB 40700, iso-I, peak A P A A P P P V P K A H G R 35812 C H -R(156)

(0 NCIMB 40700, iso-I, peak B P A A P P P V P K A H G 35656 x m m -G(57)-H(137)

H u NCIMB 40700, iso-I, peak C P A A P P P V P K A 35562 r- m Agro pROX26, iso-Z A P P P V P K A H G R 35643

-R(156)-G(57)

Agro pROX26, iso-Y, peak A A P P P V P K A H 35430

Example 7 : Isolation of chromosomal DNA from soil isolate NCIMB 40700

Isolate NCIMB 40700 was cultured on AJ3B agar (as for AJ3B liquid media but containing Bacto agar, 415 g/1; pH 6.8) at 25°C for three days. After this time, a loopfull of cells was transferred to 200ml of AJ3B medium contained within a 11 conical flask, and the culture grown at 26°C with aeration (275 rpm) for three days. Cells were then harvested by centrifugation in a Sorval RC-5B Refrigerated Superspeed Centrifuge using a Sorval GSA rotor (7,500 rpm, 15min, 4°C). The supernatant was discarded and the cell pellet stored at -20°C for 16h. After this time, cells were thawed on ice and re-suspended in 50ml of 50mM Tris.HCl, 50mM EDTA (pH 8.0) solution. Equal portions of the suspension (5ml) were subsequently transferred to 30ml sterile glass universal bottles, to which was added 0.5ml of lysozyme solution (lOmg/ml in 0.25M Tris.HCl, pH 8.0). Samples were mixed by gentle inversion and then placed on ice for 30min. After this time, 6ml of lysis solution (SDS 0.5 % (w/v), 50mM Tris.HCl, 0.4M EDTA (pH 8.0) and proteinase K 1 mg/ml) was added. Following gentle mixing by inversion again, samples were heated to 50°C until lysis became evident, and then cooled to room temperature. At this point, 10ml of 50mM Tris.HCl, 50 mM EDTA (pH 8.0) was added to each cleared lysate, followed by 5ml of biphenol:CHCl₃ :ιso- amylalcohol (25:24: 1 mix, pH 8.0). Extraction was carried out by gentle inversion followed by separation of the phases in a Denley BR401 Refrigerated Centrifuge (5,000 rpm, 15min, 4°C). The aqueous phase from each sample was removed with a wide-bore pipette and extracted again with 5ml of biphenol:CHCl₃:iso-amylalcohol (25:24: 1 mix, pH 8.0). At this stage, all aqueous phases were pooled in a sterile 150ml plastic screw cap beaker using a wide-bore pipette. DNA was precipitated by adding 0.1 volume of 3M unbuffered sodium acetate (NaOAc) followed by one volume of cold (-20°C) absolute ethanol. The solution was mixed by gentle inversion until a clot of precipitated

DNA was evident. This was removed using a small sterile plastic loop and re-suspended in 1ml of TE solution (lOmM Tris.HCl, ImM EDTA, pH 8.0).

Example 8 : Construction of a Genomic Library of Soil Isolate NCIMB 40700. Plasmid pWE15 (Wahl, G. (1989) Strategies 2 : 17) was the cosmid vector used to construct a library of soil isolate NCIMB 40700 genomic DNA. In this instance, DNA (5μg) was digested with 40 units of restriction endonuclease BamW I (Gibco BRL, Paisley, PA3 4EF, Scotland) at 37°C for 5.5h in 50μl of the appropriate React™ Buffer (Gibco BRL). The reaction mixture was then extracted once with biphenol:CHCl₃:iso-amylalcohol (25:24: 1 mix, pH 8.0) and once with CHC1₃ . The recovered aqueous phase was treated with 0.1 volume 3M NaOac (pH 4.8) and 2.5 volumes of ethanol. Following mixing, the solution was incubated at -20°C for 16h to precipitate the DNA. After centrifugation in 1.5ml microcentrifuge tubes using a Jouan MRU. 11 Centrifuge (14,000 rpm, 25min, 4°C) the resultant DNA pellet was re- dissolved in 36μl of sterile double-distilled water. To this was added 4μl of shrimp alkaline phosphatase (SAP) reaction buffer (20mM Tris.HCl (pH 8.0) , lOmM MgCl₂ ) and 0.5 units of SAP enzyme (Amersham Life Science, Little Chalfont, HP7 9BR,

United Kingdom). De-phosphorylation of ends was allowed to proceed for 30min at 37° C. After this time, a further half unit of enzyme was added and incubation continued at 37°C for 30min. The reaction mixture was placed in a heated water bath held at 65°C for 20min, and then allowed to cool slowly to room temperature on the bench. De- phosphorylated pWE15 DNA was finally isolated from a 0.7% (w/v) Tris-borate-EDTA (TBE) agarose gel using the Sephaglas ™ Bandprep Kit (Pharmacia Biotech Ltd. , St. Albans, AL1 3AW, United Kingdom) according to the manufacturer's instructions.

In order to obtain partial NCIMB 40700 genomic DNA fragments of the desired size range (25 - 50 kb) for cosmid library construction, the following protocol was carried out. Intact genomic DNA (160μg) from isolate NCIMB 40700 (see Example 7) was digested with 0.156 units of restriction endonuclease Sau3A I (Gibco BRL) in 100 μl of the appropriate React™ Buffer (Gibco BRL) at 37°C for 25min. Twenty units of RNace-It™ Ribonuclease Cocktail (Stratagene Ltd. , Cambridge, CB4 4GF, United Kingdom) was then added and the reaction allowed to proceed for a further 5min. After this time, reactions were stopped by placing the sample in a heated water bath at 70°C for lOmin. Digested DNA (lOOμl) was loaded onto a 4.8ml sucrose gradient (10 - 40 % in lOmM Tris.HCl, ImM EDTA, 1M NaCl; pH 8.0) and fragments separated by centrifugation in a Beckman L5-65B Ultra-centrifuge (36,000 rpm, 16h, 4°C) using a Beckman SW50.1 swing-out rotor. Fractions (200μl) containing DNA fragments of the desired size range were added to lOOμl of TE buffer (lOmM Tris.HCl, ImM EDTA, pH 8.0) along with 1.25μl yeast transfer RNA (tRNA) in a 1.5ml microcentrifuge tube To this was added 300μl of propan-2-ol and precipitation of DNA allowed to proceed at -20°C for 16h. This was recovered by centrifugation using a Juan MR 14. 11 Centrifuge (12,000 rpm, 20min, 4°C) and the resulting DNA pellet washed with 70% (v/v) ethanol prior to air-drying for 30min at room temperature. DNA was subsequently re-suspended in 50μl TE buffer at 4°C for 16h.

Restricted pWE15 DNA and size-fractionated NCIMB 40700 genomic fragments were ligated and in vitro packaged using the Gigapack II Gold Packaging Kit (Stratagene Ltd.) according to the manufacturer's instructions (Catalog #251201 , 251202, November 20, 1991). Subsequent titration of the packaged library on E. coli strain XL-1 Blue MRF' (Stratagene Ltd.) indicated that a total of 4.6 X 10⁵ transductants had been obtained. Example 9 : Isolation of a Clone Specifying Esterase Activity.

Individual E. coli XL-1 Blue MRF' recombinant colonies (see Example 8) were inoculated into 2ml of 2YT medium (tryptonelό g/1, yeast extract lOg/1, NaCl 5g/l; pH 7.2) containing ampicillin at 50μg/ml final concentration, and grown at 25°C for 16h with aeration (298 rpm). Samples (0.5ml) were then removed from each culture and pooled in groups of three in 1.5ml microcentrifuge tubes. Cells were collected by centrifugation in an Eppendorf 5415 Centrifuge (14,000 rpm, 3min, room temperature) and re-suspended in 1.5ml of substrate solution containing compound I as the methyl ester dissolved in 10% (v/v) DMSO, 0.1M sodium phosphate buffer; pH 7.0. Samples were then mixed vigorously using a Vortex-Genie2™ shaker (CP Laboratories, Bishop's Stortford, CM23 3DX, United Kingdom) for 18h at room temperature. After this time, 700μl was removed from each reaction mixture and diluted 1 : 1 with methanol in a 1.5ml microcentrifuge tube. Samples were then clarified by centrifugation using an Eppendorf 5415 Centrifuge (14,000 rpm, 4min, room temperature). Clarified solution was then analysed by High Pressure Liquid Chromatography (HPLC) as given in example 2.

Of the 1 , 287 recombinants tested, 3 generated peaks which co-eluted with the carboxyiic acid standard and showed the same UV absorbance spectra as the standard, as determined by a photodiode array spectrometer.

Example 10 : Isolation and Characterisation of Cosmid pROX I, pROX 2 and pROX 3 from E. coli strain XL-1 Blue MRF'.

Cosmid DNA from the 3 positive clones which had been identified from the initial HPLC screen (see Example 9) was isolated using the large-scale alkaline lysis /

CsCl protocol as described in Sambrook et al. (1989) "Molecular Cloning, A Laboratory Manual " (Second Edition) . These were designated pROX 1 , pROX 2 and pROX 3, respectively. Restriction endonuclease analysis indicated that each clone contained common BamH I, EcoR I and Hind III restriction fragments covering a region of the genome approximately 12.0 kb in size (see Figure 1). The related-ness of the aforementioned clones was further verified by Southern analysis, whereby DNA (ca. 10 μg) from each clone was digested with 30 units of restriction endonuclease BamH I (Gibco BRL) in 40μl of the appropriate React™ Buffer (Gibco BRL) at 37°C for 5h. Intact NCIMB 40700 genomic DNA (32μg) was digested in a similar fashion under identical conditions. DNA fragments were size- fractionated by 0.7% (w/v) TBE agarose gel electrophoresis, and then transferred to Hybond -N nylon membrane (Amersham Life Science) as described in Sambrook et al. (1989) ibid. Following denaturation and renaturation, the membrane was probed with a 3 8 kb BamH I fragment (situated within the common region) purified from cosmid pROX 2, using the ECL ' Random Prime Labelling and Detection System (Amersham Life Science) according to the manufacturer's instructions Results indicated that the probe hybridised ^ to a 3 8 kb BamH I fragment from pROX 1 , pROX 2 and pROX 3, as well as a co- migrating fragment in the NCIMB 40700 genomic digest

Example 11 : Construction of plasmid pROX 6.

Five micrograms of plasmid pTrc99A (Amann, E et al (1988) Gene 69 301- 0 315) DNA was digested with 40 units of either restriction endonuclease EcoR I, BamH I or Hind III (Gibco BRL) at 37°C for 4h in 60μl of the appropriate React™ Buffer (Gibco BRL) After this time, each reaction mixture was extracted once with biphenol CHC1₃ iso-amylalcohol (25 24 1 mix, pH 8 0) and once with CHC1₃ DNA was then ethanol precipitated, de-phosphorylated and gel-purified as described in J Example 8 Cosmid pROX 1 DNA (lOμg) was digested with 40 units of restriction endonuclease EcoR I (Gibco BRL) at 37°C for 5h in 50μl of the appropriate React^{1 M} Buffer (Gibco BRL) In addition, cosmid pROX 2 DNA (lOμg) was digested in a similar fashion with either restriction endonuclease BamH I or Hind III (Gibco BRL) All three reaction mixtures were then size-fractionated by 0 7% (w/v) TBΕ agarose gel 0 electrophoresis The portion of the gel containing either the 5 7 kb EcoR I, 3 8 kb

BamH I or 5 0 kb Hind III fragment, was excised and the DNA recovered using the Sephaglas™ Bandprep Kit (Pharmacia Biotech Ltd ) as described previously (see Example 8)

Approximately 500ng of purified cosmid DNA was mixed with ca 500ng of 5 purified de-phosphorylated vector DNA (which had been digested in a similar fashion) in a 1 5ml microcentrifuge tube, to give a total volume of 18μl Tubes were placed in a water bath held at 65°C for 7mιn, then removed to the bench for lOmin to allow the DNA mixture to cool slowly to room temperature After this time, 2μl of T4 DNA ligase buffer (Boehringer Mannheim, Lewis, BN7 1LG, United Kingdom) and 1 unit of 0 T4 DNA ligase (Boehringer Mannheim) was added, and the reaction allowed to proceed at 16°C for 16h The hgation mix was then used to transform competent E coli strain XL-1 Blue MRF' cells as described in Sambrook et al (1989) ibid The cells were spread on L-agar plates (tryptone 10 g/1, yeast extract 5g/l, NaCl 5g/l agar 15g/l, pH 7 2) supplemented with ampicilhn (50μg/ml) and growth allowed to proceed at 37°C 5 tor 16h After this time, 12 ampicillin-resistant recombinant colonies from each transformation were then individually inoculated into 5ml of 2YT medium and grown for 16h at 37°C with aeration (275rpm) After this time, small-scale isolation of plasmid DNA was performed using the alkaline-lysis procedure, as described in Sambrook et al. (1989) ibid. A portion of the DNA preparation (lOμl) of each clone was digested with the appropriate restriction endonuclease (s) to verify sub-cloning of the desired NCIMB 40700 genomic DNA fragment, and to determine the orientation of the aforementioned fragment in relation to the trc promoter region of the vector. Subsequent 0.7% (w/v) TBE agarose gel electrophoresis indicated that each NCIMB 40700 genomic fragment had been successfully cloned in both orientations.

Recombinant cells (representing each type of clone) were inoculated into 10ml of 2 YT medium supplemented with 50μg/ml ampicillin, as well as 10ml of 2 YT medium containing 50μg/ml ampicillin and 0.5mM isopropyl β-D- thiogalactopyranoside (IPTG). Cultures were grown at 25°C for 16h with aeration (275 rpm), and then harvested by centrifugation in a Denley BR401 Refrigerated Centrifuge (4,000rpm, lOmin, 4°C). Resultant cell pellets were re-suspended in 1.5ml of compound I, methyl ester substrate solution and assayed for esterase activity as described in Example 9. No esterase activity could be detected in cells harbouring the 3.8 kb BamH I or 5.0 kb Hind III fragment clones, in either orientation. However, in contrast cells harbouring the 5.7 kb EcoR I fragment clones displayed esterase activity, the level of which depended on orientation. One of the isolates from the class giving the higher level of enzyme activity was retained and designated pROX 6 (see Figure 2).

Example 12 : Construction of Plasmid pROX 12.

Mini-preparation plasmid pROX 6 DNA (10 μl) was digested with 20 units of restriction endonuclease BamH I (Gibco BRL) in 30μl of the appropriate React™ Buffer (Gibco BRL) for 4h at 37°C. Twenty units of RNace-It ™ Ribonuclease Cocktail (Stratagene Ltd.) were then added and incubation allowed to continue at 37°C for a further 20min. After this time, DNA fragments were loaded onto a 0.7 % (w/v) TBE agarose gel and size-fractionated. The portion of the gel containing the 9.2 kb BamH I fragment was excised and DNA purified using the Sephaglas ™ Bandprep Kit (Pharmacia Biotech Ltd.) as described in Example 8. Re-circularisation of purified DNA using T4 DNA ligase (Boehringer

Mannheim) was carried out as described in Example 1 1. The ligation mix was then used to transform competent E. coli XL-1 Blue MRF' cells, which were subsequently plated on L-agar plates supplemented with ampicillin (50μg/ml ) and grown at 37°C for 16h. DNA was then prepared on a small-scale from 8 ampicillin-resistant clones, as outlined in Example 11. A portion of the DNA preparation (lOμl) was digested with 20 units of restriction endonuclease EcoR I in 30μl of the appropriate React™ Buffer (Gibco BRL) at 37°C for 3h. RNA was removed and DNA fragments size-fractionated by agarose gel electrophoresis, as described above. All the clones analysed gave a 9.2 kb DNA fragment, indicating a loss of the 700 bp BamH I region from pROX 6. HPLC analysis of recombinant cells was performed as described in Example 3, and results indicated that esterase activity was retained. One representative clone was kept, and - designated pROX 12 (see Figure 3).

Example 13 : Construction of Plasmid pROX 18.

Mini-preparation pROX 12 DNA (20μl) was digested with 20 units of restriction endonuclease H d III and BamH I (Gibco BRL) in 40μl of the appropriate React ™ Buffer (Gibco BRL) at 37°C for 2h. Following further treatment of the reaction mixture with RNace-It ™ Ribonuclease Cocktail (Stratagene Ltd.) at 37°C for lh, DNA fragments were size-fractionated by 0.7% (w/v) TBE agarose gel electrophoresis. The portion of the gel containing the 3.0 kb DNA fragment was excised and DNA purified using the Sephaglas ™ Bandprep Kit (Pharmacia Biotech Ltd.) as described in Example 8.

A ligation reaction containing the purified 3.0 kb BamH I / Hind III NCIMB 40700 fragment (ca. 500ng) and purified plasmid pTrc99A DNA (ca. 500ng, previously digested with the same restriction endonucleases and de-phosphorylated) was set up as described in Example 1 1. The ligation mix was then used to transform competent E. coli XL-1 Blue MRF' cells, which were subsequently spread onto L-agar plates supplemented with ampicillin (50μg/ml) and grown at 37°C for 16h. After this time, 6 ampicillin-resistant colonies were individually inoculated into 10ml of 2YT medium containing ampicillin (50μg/ml) and grown at 25°C for a further 16h. Plasmid DNA was prepared on a small-scale from 3ml of culture as described in Example 11 , and a portion of this (lOμl) digested with restriction endonuclease BamH I and Hind III

(Gibco BRL). DNA fragments were size-fractionated by 0.7% (w/v) TBE agarose gel electrophoresis, and results indicated that the 3.0 kb NCIMB 40700 genomic fragment had been successfully sub-cloned in all cases. One such clone was retained and designated pROX 18 (see Figure 4).

Example 14 : Determination of the Nucleotide Sequence of the NCIMB 40700 DNA in Plasmid pROX 18.

Plasmid pROX18 DNA was prepared on a large-scale using the alkaline-lysis/ CsCl purification procedure described in Sambrook et al. (1989) ibid. This was 5 digested with restriction endonuclease BamH I and Hind III (Gibco BRL), to release the

3.0 kb NCIMB 40700 genomic fragment, which was subsequently sub-cloned into the corresponding sites within plasmid pBluescript II SK^" (Stratagene Ltd). The resulting construct was then used to transform competent E. coli strain XL-1 Blue cells. Plasmid DNA was then purified using the alkaline-lysis procedure as described in Sambrook et al. (1989) ibid. Nested deletions were generated from both the T7 and T3 ends using the Εrase-a-Base™ System Kit (Promega limited, Southampton, United Kingdom) following the procedure of Henikoff, S. (1984) Gene 28 : 351-359. Deletion clones were size-selected by agarose gel electrophoresis, and DNA sequencing performed using standard dideoxynucleotide termination reactions containing 7-deaza dGTP and [³⁵ S] dATP as label. In order to resolve severe GC band compressions, 7-deaza dITP was added. All sequencing reactions were analysed on 6% (v/v) polyacrylamide wedge gels containing 8M urea. Internal oiigonucleotide primers were synthesised as necessary in order to fill in sequencing gaps.

In total, 3,032 bp of sequence was obtained on both strands (see ID SΕQ 1). Translation of this region in all six possible reading frames, reveals the presence of three open reading frames (ORF's). Of these, the first ORF (nucleotide position 337 to 1404) is entire and encodes for a predicted 355 amino acid polypeptide, 38.3 kiloDaltons (kDa) in size . This has identical homology (residues 26 to 54, inclusive) to the amino acid region determined by N-terminal sequencing of the purified NCIMB 40700 esterase enzyme. It is therefore proposed that this represents the structural gene encoding the aforementioned enzyme. Within the same reading frame, the second entire ORF lies downstream of the first (nucleotide position 1558 to 2466) and encodes for a predicted 302 amino acid polypeptide, 32.6 kDa in size. Finally, a third incomplete ORF lies downstream of the second (nucleotide position 2641 to 3032) again in the same reading frame, which encodes for a predicted 130 amino acid truncated polypeptide, 13.5 kDa in size.

Example 15 : Construction of Plasmid pROX 29.

A region (nucleotides 325 to 1490, comprising the esterase structural gene and flanking regions) within the sequenced 3.0 kb NCIMB 40700 genomic fragment, was amplified from CsCl-purified plasmid pROX18 template using the GeneAmp^R PCR Reagent Kit (Perkin Elmer Cetus, 761 Main Avenue, Norwalk, CT 06859, USA) essentially according to the manufacturer's instructions. In brief, a series of modified lOOμl reaction mixtures (1 X Reaction Buffer / 10 % (v/v) dimethyl sulfoxide (DMSO) containing 50-400ng pROX 18 DNA; lμM of oiigonucleotide dTGAGGAGACACCATGGCTATGCACCGTAC (Cruachem Ltd. , Glasgow, G20 0UA, United Kingdom); lμM of oiigonucleotide dTATTAGCGCTAAGCTTAGTGCCAACAG (Cruachem Ltd.) and 200μM dNTP's) were set up in capped 0.5ml polypropylene microcentrifuge tubes. These were then placed in a heated water bath maintained at 98°C for 5min, and then allowed to cool slowly on the bench for a further lOmin. After this time, tubes were placed in a Hybaid ™ Thermal Cycler (Hybaid Ltd., Teddington, Middlesex, TW1 1 8LL, United Kingdom.) and 5 units of Amplitaq DNA polymerase (Perkin Elmer Cetus) were added to each reaction mixture. Amplification of DNA was allowed to proceed over 30 cycles, each comprising of: a denaturing step of 94°C for lmin; an annealing step of 65°C for 2min, and an elongation step of 72°C for 2min. The last cycle was slightly modified to include an additional lOmin elongation step. Following this, the reaction mixtures were cooled to 30°C and subsequently placed on ice. Samples were then pooled and extracted once with biphenol:CHCl₃:iso-amylalcohol (25:24: 1 mix, pH 8.0), once with CHC1₃ and finally ethanol precipitated as described in Example 8.

PCR-amplified DNA was digested with 20 units of restriction endonuclease Nco I and Hind III (Gibco BRL) in 40μl of the appropriate React™ Buffer (Gibco BRL) at 37°C for 4h; after which time, DNA fragments were size-fractionated by 0.7% (w/v) TBE agarose gel electrophoresis. The portion of the gel containing the 1. 1 kb Nco I / Hind III DNA fragment was excised and DNA purified using the Sephaglas ™ Bandprep Kit (Pharmacia Biotech Ltd.) as described in Example 8.

A ligation reaction containing ca. 500ng of purified 1.1 kb Nco I / Hind III PCR product and ca. 500ng of plasmid pKK233-2 DNA (Amann, E. and Brosius, J. (1985) Gene 40 : 183-190) , previously digested with the same restriction endonucleases and de-phosphorylated; was set up as described in Example 11. The ligation mix was then used to transform competent E. coli JM109 cells, which were subsequently spread onto L-agar plates supplemented with ampicillin (50μg/ml) and 1 % (w/v) glucose. Growth was allowed to proceed at 37°C for 16h, after which time 12 ampicillin-resistant colonies were individually inoculated into 3ml of 2YT medium containing ampicillin (50μg/ml) and 1 % (w/v) glucose. Cultures were grown with aeration (275rpm) at 25°C for a further 16h, and plasmid DNA was prepared on a small-scale as described in Example 11. A portion of this (lOμl) was digested with Nco I and Hind III restriction endonucleases, and DNA fragments size-fractionated by 0.7% (w/v) TBE agarose gel electrophoresis. Results indicated that the 1.1 kb NCIMB 40700 genomic fragment had been successfully sub-cloned in all cases.

HPLC analysis of E. coli JM109 recombinant cultures was performed as described in Example 2 (whole broth and washed cells). Results indicated that esterase activity was retained. One clone was kept and designated pROX 29 (see Figure 5).

Example 16 : Construction of plasmid pROX 26 Approximately 5μg of plasmid pKK233-2 DNA was digested to completion with 10 units of restriction endonuclease Hind III (Gibco BRL) in the appropriate React™ Buffer (Gibco BRL) at 37°C for 3h. The reaction mixture was then extracted once with biphenol:CHCl₃:iso-amylalcohol (25:24: 1 mix, pH 8.0) and once with CHC1 ₃. Vector DNA was finally ethanol precipitated, de-phosphorylated using Calf Intestinal Alkaline Phosphatase (CIAP; Gibco BRL) according to the manufacturer's instructions, and gel- purified as described in Example 8.

Plasmid pROX 12 DNA (ca. 5μg) was digested in a similar fashion to that described above, and the reaction mixture size-fractionated by 0.7% (w/v) TBE agarose gel electrophoresis. The portion of the gel containing the 3.0 kb H d III fragment was excised, and the DNA recovered using the Sephaglas™ Bandprep Kit (Pharmacia Biotech Ltd.) as described previously (see Example 8).

A ligation reaction (15μl of IX ligase buffer (Gibco BRL) containing ca. 500ng of purified pROX12 DNA, ca. 500ng of purified de-phosphorylated pKK233-2 DNA and 1 unit of T4 DNA ligase; Gibco BRL), was set up in a 1.5ml microcentrifuge tube. Incubation was allowed to take place at 16°C for 18h, after which the reaction mixture was used to transform competent E. coli JM109 cells. These were plated onto L-agar containing ampicillin (50μg/ml) and 1 % (w/v) glucose, and allowed to grow at 37°C for 16h. Ten antibiotic-resistant colonies were individually inoculated into 3ml of 2YT medium containing ampicillin (50μg/ml) and cultivated for a further 16h at the same temperature. DNA was then prepared on a small-scale and re-suspended in 20μl TE buffer. A portion of this (5μl) was digested with restriction endonuclease BamH I (Gibco BRL) and DNA fragments size-fractionated by 0.7% (w/v) TBE agarose gel electrophoresis, as described in Example 8. One clone gave DNA fragments of 3.3 kb and 4.3 kb in size, indicating successful cloning of the pROX 12 Hind III fragment.

This was retained and designated pROX 24 (see Figure 6).

Approximately 3μg of CsCl-prepared plasmid pROX24 and pWOR904 DNA (cited in International Patent Publication WO 94/00577) was digested to completion with restriction endonuclease BamH I (Gibco BRL) in the appropriate React™ Buffer (Gibco BRL) in a similar fashion to that described in previous Examples. Each plasmid digest was then size-fractionated by 0.7% (w/v) TBE agarose gel electrophoresis. The portion of the gel containing the 3.3 kb pROX 24 BamH I fragment and the 10.4 kb pWOR904 BamH I fragment was excised and DNA purified using the Sephaglas ™ Bandprep Kit (Pharmacia Biotech Ltd.) as described in Example 8. A ligation reaction containing ca. lOOng of the aforementioned purified DNA fragments was set up and allowed to proceed, as described in Example 11. The mixture was then used to transform competent E. coli JM109 cells, which were subsequently

-22 -

ULE 26 plated onto L-agar containing chloramphenicol (25μg/ml) and incubated overnight at 37 °C. After this time, 24 antibiotic-resistant colonies were used to inoculate 2ml of L- broth containing chloramphenicol (25μg/ml) and grown for a further 18h at 37°C. Plasmid DNA was then isolated on a small-scale using the alkaline lysis procedure described by Sambrook et al. (1989) ibid, and re-suspended in 20μl TE buffer. A portion of this (5μl) was digested with restriction endonuclease Bgl II (Gibco BRL) and size-fractionated by 0.7% (w/v) TBE agarose gel electrophoresis. One clone gave DNA bands which were 10.8 kb and 2.9 kb in size, indicating that the pROX 24 fragment had been successfully sub-cloned into pWOR904, with the esterase structural gene running towards the chloramphenicol-resistance marker. This construct was retained and designated plasmid pROX 26 (see Figure 7).

Example 17 : Construction of plasmid pROX 36

Approximately 2μl of CsCl-prepared pWOR904 DNA was digested with 50 units of restriction endonuclease Kpn I (Gibco BRL) in 60μl of the appropriate React™ Buffer (Gibco BRL) at 37°C for 4h. Likewise, 5μg of CsCl-prepared pROX 29 DNA was digested with 50 units of restriction endonuclease Hind HI (Gibco BRL) in a similar fashion. Each reaction mixture was then extracted once with biphenol:CHCl₃:iso-amylalcohol (25:24: 1 mix; pH 8.0); once with CHC1₃ and ethanol precipitated (see Example 8). Recovered DNA was then blunted using T4 DNA polymerase (Promega Ltd., Southampton, S01 7NS, United Kingdom) as described by Sambrook et al. (1989) ibid. , and further digested to completion with restriction endonuclease BamH I (Gibco BRL). Reactions were size-fractionated by 0.7% (w/v) TBE agarose electrophoresis and the portion of the gel containing the 10.4 kb pWOR904 fragment and the 1.4 kb pROX 29 fragment were excised and DNA purified, as described in Example 8.

A ligation reaction containing ca. 500ng of purified pWOR904 and pROX 29 DNA was set up and allowed to proceed as described in Example 1 1. The reaction mixture was then used to transform competent E. coli JM109 cells, which were subsequently spread onto L-agar plates containing chloramphenicol (25μg/ml) and 1 % (w/v) glucose, and grown at 37°C for a further 16h. Plasmid DNA was then prepared on a small-scale as described in Example 11 , and a portion of this ( lOμl) digested with EcoR I restriction endonuclease (Gibco BRL). The resulting DNA fragments were size- fractionated by 0.7% (w/v) TBΕ agarose gel electrophoresis, and in all cases bands of 10.4 kb and 1.4 kb in size were detected. This indicated the successful sub-cloning of the pROX 29 fragment into pWOR904, with the esterase structural gene running

-23-

SUBSTITUTΕ SHEET (RULE 26) towards the chloramphenicol-resistance marker. One clone was retained and designated pROX 36 (see Figure 8).

Example 18 : Introduction of plasmid pROX 26 and pROX 36 into Agrobacterium

Approximately 400ng of "midi-preparation" plasmid DNA (previously prepared from 10ml of recombinant E. coli culture using a scaled-up version of the alkaline lysis procedure described by Sambrook et al. (1989) ibid.) was used to transform electro-competent Agrobacterium sp. " 15-10" cells (cited in International Patent Publication WO 94/00577) as described by Wen-jun, S. and Forde, B.G. (1989) Nucleic Acids Research 17(20) : 8385. Following expression, recombinant bacteria were spread onto L-agar supplemented with chloramphenicol (lOμg/ml) and incubation allowed to proceed at 30°C for 3 days. After this time, antibiotic-resistant colonies were patched onto L-agar containing chloramphenicol (lOμg/ml) and growth allowed to proceed at 30°C overnight. Plasmid DNA was then purified on a small-scale using a modified version of the Rapid Boiling Method (Holmes, D.S. and Quigley, M. (1981 ) Analytical Biochemistry 114 : 193) as described in International Patent Publication WO 94/00577; and digested with the appropriate restriction endonucleases in order to verify both presence and structure.

Example 19: Demonstration of esterase activity in Agrobacterium

Recombinant Agrobacterium cells containing either plasmid pROX 26 or plasmid pROX 36 were plated onto L-agar containing chloramphenicol (lOμg/ml) and incubated at 30°C for 2 days. A loopfull of cells were then used to inoculate 20 ml of AJ3B medium containing chloramphenicol (lOμg/ml) in a 250ml springed conical flask.

Growth was allowed to proceed at 30°C for 20h with aeration (275rpm). After this time, the culture was assayed for esterase activity using a slightly modified procedure to that outlined in Example 9. In this instance, whole-broth or whole-cell activities were measured by diluting either a sample of culture or washed-cells, 10-fold in sodium phosphate buffer (100 mM, pH 6.5). An aliquot of this (0.2ml) was mixed with 1.6ml of sodium phosphate buffer (lOOmM, pH 6.5) in a glass universal bottle, and the contents pre-heated to 40°C in a water bath. To this was added 0.2ml of DMSO containing 5% (w/v) compound I as the methyl ester substrate, and incubation allowed to continue at 40°C for a further 15min. After this time, 1ml of methanol was added to terminate the reaction. A sample was then transferred to an 1.5ml Eppendorf tube and spun in an Eppendorf 5415 Centrifuge (14,000rpm, 5min, room temperature). The supernatant was finally analysed by HPLC as described in Example 9. Assay of soluble cell-extracts was carried out using an identical procedure to that described above, except for several modifications. In this instance, a sample (0.1ml) of cell-free extract was mixed with 0.35ml of sodium phosphate buffer ( lOOmM, pH 6.5) in a 1.5 ml Eppendorf tube. DMSO (0.05ml) containing 5 % (w/v) compound I as the methyl ester substrate was added to start the reaction, and 0.25ml of methanol added to stop the reaction.

Overall, results indicated that recombinant Agrobacterium " 15-10" cells containing plasmid pROX 26 produced 700-fold higher levels of esterase activity compared with Agrobacterium " 15-10" cells harbouring pWOR904. Moreover, Agrobacterium " 15-10" cells containing pROX 36 displayed slightly lower activity levels compared with the above.

Example 20 Purification of Esterase from Agrobacterium clones

Material from the genetically engineered Agrobacterium producing organism was processed to the WK10 step (see Example 6) and then concentrated and used for crystallographic studies (see Example 22) and immobilisation studies (see Example 24). This over-producer gives approximately 6% of its cell protein as the esterase. Unlike the material produced by genetically engineered E.coli, Agrobacterium clones produce at least two iso-enzymes. In addition the leader sequence is cleaved off at a different point to that observed for NCIMB 40700 and the E.coli clones.

Example 21 General enzyme properties

The enzyme is encoded from a gene of 1 ,065 base pairs and is expressed as a polypeptide which is minus a leader sequence of approximately 25 amino acids (see Table 2). This suggests that the enzyme is exported to the periplasm. The cleavage site for this leader sequence varies by a few amino acids according to the expression host. The native protein has a subunit size of approximately 35,500 and the active protein consists of two of these sub-units giving an enzyme with a mw of around 71,000. There are 4 cysteine residues per sub-unit indicating that there could be up to 4 disulphide bridges. These observation indicate that the enzyme has evolved to work outside the cytoplasm and that it is designed to withstand fairly harsh environments. It is not known if the disulphide bridges hold the sub-units together or are just cross-links within each sub-unit. The enzyme is inhibited by DTTA suggesting that di-sulphide bridges are critical to conformation and activity. It has been demonstrated using liquid chromatography-mass spectroscopy (see

Table 2) that the enzyme can lose 1 , 2 or 3 amino acids from its carboxyl terminus. This leads to a family of iso-enzymes, which have been isolated, and shown to have the same chiral selectivity. These modifications are thought to occur after translation.

The enzyme has an iso-electric point average of 8.9, with the iso-enzymes grouped very closely together. This high pi value explains the very good selectivity of the WK10 purification step allowing enzyme preparations at approximately 90% purity even at pilot plant scale.

The post WK10 extract shows the following hydrolysis rates (Table 3):

Table 3

Example 22 Crystallisation and x-ray diffraction analysis of the esterase The procedures used were standard crystallisation and crystallographic techniques using CCP4 software, Daresbury laboratory (Collaborative Computational Project Number 4 (1994) Acta CrystD50, 760-763 "The CCP4 Suite: Programs for Protein Crystallography").

Crystallisation trials

Crystallisation trials were first carried out using the active protein fraction after a gel filtration step on Superose 12™ to remove any aggregated material and trace contaminants. The protein was concentrated to 14 mg/ml (estimated using A260/A280), before crystallisation using vapour phase diffusion. Good quality crystals were obtained after one week using 30-35% ammonium sulphate as the precipitant and either 1 OmM

PIPES at pH6.5 or l OmM Tris buffer at pH 7.0. The initial crystals obtained ranged from clusters of fine needles to large thin plates. The plate like crystals grown in 10 mM Tris pH 7.0 diffracted to 1.7A resolution with cryo-cooling and x-ray data were collected to 97.7% completeness. The crystal symmetry is in the space group monoclinic C2 with unit cell parameters a = 134.7A, b = 55.8A, c = 1 10.3A, α = γ = 90°, b = 125°.

Seeding Fragments of crystals obtained from the above expeπments were used as seeds in further crystallisation tnals This produced larger crystals (with different, orthorhombic morphology) The crystals were exposed to a Synchrotron radiation source and native diffraction data were collected to 1 5 A resolution with 88 1 % completeness The crystals have symmetry in the space group P2 ] 2 j 2 j , with unit cell parameters a = 56 7A, b =

1 15 2A, c = 131 5 A, a = b = g = 90° The individual crystals were non-isomorphous, but by cutting a crystal into several fragments it was possible to collect native and deπvative data on different pieces of one crystal

Structure

The structure was solved using the techniques of multiple isomorphous replacement and multi-crystal averaging Phases for the model were generated using data collected from three deπvatised crystals soaked in a solution of either uranium acetate (5mM), mercury acetate (15 mM) or gold chloride (2mM) The structure has been refined to 1 6A resolution with an overall R factor = 20 87 % and free R factor= 23 10%

The overall structure of the enzyme is an αβ-hydrolase fold. By analogy with other hydrolase enzyme structures it has been possible to identify a catalytic tπad formed by seπne (233), glutamic acid (257) and hιstιdme(325) All hydrolase enzymes studied to date have the acidic residue of the catalytic tπad in a loop at the end of strand 7 The enzyme of the invention has the acidic residue at the end of strand 6 There are two disulphide bπdges within each subunit of the dimer, these are formed between Cys(261 )- Cys(296), and Cys(98)-Cys(99) The disulphide bπdge formed between Cys(98) and Cys(99) is only the second reported example of adjacent cysteine residues forming a disulphide (the other being mcthanol dehydrogenase from Methvlobactenum extorque ) There are two proline residues in a cw-confirmation, these are Pro(l 79) and Pro(210) (All amino acid residues are numbered according to SEQ ID NO 2)

Example 23 Modification of esterase enzyme kinetics Initial experiments using enzyme isolated directly from NCIMB 40700 showed that the methyl ester was the preferred substrate and that chiral selectivity of the enzyme was good (90 to 95 % yield with a 95 5 chiral ratio) On cloning of the encoding gene into E coli, it was found that the enzyme was less specific The enzyme showed poor chiral selectivity and increased hydrolysis rate for the ethyl ester Part of this difference was due to the host, but the effect was enhanced by the purification process This modification in kinetic behaviour was also seen when the same purification methods were applied to the enzyme from isolate NCIMB 40700 The change in kinetics was associated with an ion exchange chromatography step which required high pH condition. In addition the enzyme was also exposed to high ammonium sulphate levels. The effects of these factors on chiral selectivity were therefore studied and are shown below. Assays were carried out using NCIMB 40700 enzyme and the standard DMSO reaction mixture (see Example 2).

Table 5

These results showed that both high pH and ammonium sulphate treatment 0 reduce the chiral selectivity of the esterase. The effects of ammonium sulphate are the most dramatic and appear to be reversible by dialysis. The pH is irreversible. Reduction in chiral selectivity caused by ammonium sulphate is associated with an increase in reaction rate which can be seen from the time taken to reach a chiral ratio of 95:5. These treatments in some way must effect the enzymes tertiary structure. 5 Expression of the enzyme in Agrobacterium, gave enzyme with better chiral selectivity. In addition, all fermentations and down stream processing are controlled below pH 7.0 to avoid loss of selectivity.

Example 24 Immobilisation of esterase onto OC1050 resin

3.61 of WK10 purified esterase enzyme from Agrobacterium" \5-\0 pROX 36 (8mg/ml, 0.18mM/mg/h ethyl substrate) was dialysed overnight against running tap water using cellulose dialysis tubing (mol. wt. cut-off 12,000 to 14,000). The conductivity of the recovered enzyme was than increased to 4.5mS with the addition of solid NaCl. The enzyme solution was then stirred with 400g. of OC1050 resin (damp

? as supplied by the manufacturer, Bayer) and the pH controlled to 6.5 using 1M H SO₄. The adsorption was completed in 20h at room temperature with an activity uptake of 99% . The enzyme-resin was allowed to settle and the supernatant decanted and discarded. The enzyme-resin was then washed by resuspending in 3.61 of distilled water and recovered by decantation. Covalent coupling of the enzyme to the resin was achieved by resuspended the enzyme-resin in 71 of 0 5 % glutaraldehyde containing 0 1 M sodium phosphate buffer pH7 0 After stirring for lh at room temperature excess glutaraldehyde was removed by washing the resin with 3x51 of distilled water, 51 of 0 IM Tris buffer pH7 0 (30mιn) and 51 of 0 I M sodium phosphate buffer pH7.0 ( lh), between each wash enzyme-resin was recovered by decantation Finally, immobilised enzyme (397g) was recovered using vacuum filtration on a Buchner funnel and stored damp at 4°C in a sealed plastic container

1 Enzyme solution used

Methyl substrate activity 5.22mM/ml/h Ethyl substrate activity 0 18mM/ml/h Methyl ethyl activity ratio 29.1 Protein 7 97mg/ml

2 Enzyme challenge

Methyl activity challenge 47 0 mM/g/h Ethyl activity challenge 1 62mM/g/h Protein challenge 71 7mg protein/g resin

3 Post-ads supnt

Methyl substrate activity 88,500 nM/ml/hr (98 % ads ) Ethyl substrate activity 2,400 nM/ml/hr (99% ads ) Protein 0 28 mg/ml (96% ads )

4 Post-ads enzyme/resin ethyl activity 0 87 mM/g/hr

Example 25 Resolution of Stage 1 Paroxetine Ester using a Typical Batched Bio- transformation Reactor 120g of unresolved Stage 1 Paroxetine ester was pre-ground in a mortar and pestle, then homogenised into 750ml of distilled water using a Silverson Laboratory Mixer-Emulsifier for 30min. The substrate suspension was then degassed under vacuum for 30min. The degassed substrate was added to a thermostatically controlled water jacketed glass reactor vessel (21, 47°C) and the pH of the suspension adjusted to 6.5, with the dropwise addition of 13ml 1.0M NaOH (Fisons analytical grade) care was taken not to exceed pH 8.0. Prior to use, 200g of immobilised OC1050 esterase resin was washed with 41 of distilled water over lOOmicron wire mesh filter and dried over a Whatman 54 filter. 120g of resin was weighed out. The reaction was started by the addition of 120g resin to the substrate in the reactor and washed in with 250ml of distilled water, the total reactor volume was 11. The reaction was stirred (PTFE 76 x 19mm flat blade) and the pH was maintained at 6.5 using a pH titrator, with 1.0M NaOH. The reaction was monitored directly for acid production and indirectly using caustic addition and conductivity. Chiral measurements were determined throughout the run, and the reaction was stopped when the -/ + ethyl chiral ester ratio was > 95:5.

At the end of the reaction (ca. 20h) the resolved Paroxetine ester was recovered from the reactor, by gently stirring the reactor, allowing the resin to settle and filtering off the suspended ester via a lOOmicron wire mesh filter. This process of stirring and filtering was repeated 4 times by the additions of 11 of fresh distilled water to the reactor. 38g of solid (-) isomer was recovered from the combined washing by filtration on a No.54 Whatman filter paper. The ester was then dissolved in 500ml of toluene, filtered through No.54 Whatman filter paper and rotary evaporated to dryness. The product was dried over silica gel in a desiccator under vacuum for 24h. The resolved product could also be extracted directly from the resin by extraction into toluene.

Example 26 Increasing reaction rate by continuous acid product removal and choice of immobilisation conditions and support The carboxyiic acid product generated by the action of the esterase has been shown to act as a inhibitor of the enzyme catalysed reaction. This has the effect of reducing reaction rate as the reaction proceeds and as acid accumulates. Therefore, rate enhancement has been achieved by interfacing the slurry reactor system with a recalculation system containing an electrodialysis unit to continuously remove the acid from the system. Similar success has been achieved by using a variety of ion-exchange resins in the form of columns. Both these methods have the advantages of reducing reaction times. In addition, careful choice of the immobilisation conditions and resin supports can tailor the kinetic properties of the enzyme with respect to chiral selectivity, rate of reaction and acid inhibition Immobilisation onto OC1050 resin as described in Example 24 is one of the best options in terms of rate, chiral selectivity and resistance to product inhibition

Example 27 Preparation of (-) trans-3-Ethoxycarbonyl-4-(4'fluorophenyl)-N- methyl piperidine-2,6-dione

A solution of (±) trans-3-Ethoxycarbonyl-4-(4'-fluorophenyl)-N-methyl pιpeπdιne-2,6-dιone (1 51g 5 15mmol) in DMSO (100ml) is added to Tris buffer (900ml, 0 2m, pH 5 5) The pH readjusted to 5 5, 35mg of enzyme, added and the reaction stirred for twenty hours pH is maintained as required by addition of aqueous sodium hydroxide (0 105m, 25ml, 2.63mmol)

The reaction mixture is extracted with ether (3 x 300ml), the combined organic extracts are washed with Tris buffer (0 IM, pH 8 5, 2 x 250ml), the Tris buffer extracts washed with ether ( 1 x 200ml) and the combined organic extracts dried over anhydrous magnesium sulphate The mixture is assayed (up to 20 hours) for enantiomeπc ratio

The solution evaporated to an oil and replaced with toluene/THF This solution was then reduced with lithium aluminium hydride

SEQUENCE LISTING

(1) GENERAL INFORMATION (l) APPLICANT

(A) NAME: SmithKline Beecham pic

(B) STREET: New Horizons Court

(C) CITY: Brentford

(D) STATE OR PROVINCE: Middlesex

(E) COUNTRY: Enqland

(!) POSTAL CODE: TW8 9EP (11) TITLE OF THE INVENTION: Novel Bioprocess

(in) NUMBER OF SEQUENCES: 4 (lv) COMPUTER-READABLE FORM:

(A) MEDIUM TYPE: Diskette

(B) COMPUTER: IBM Compatible

(C) OPERATING SYSTEM: DOS

(D) SOFTWARE: FastSEQ for Wmαows Version 0

(v) CURRENT APPLICATION DATA: (A) APPLICATION NUMBER:

(?) INFORMATION FOR SEQ ID NO : I :

U) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 3032 base pairs

(B) TYPE: nucleic acid

(C) STRANDEDNESS: double

(D) TOPOLOGY: unknown

(n) MOLECULE TYPE: Genomic DNA (vi ) ORIGINAL SOURCE:

(B) STRAIN: NCIMB 40700

(xi) SEQUENCE DESCRIPTION: SEQ ID NO:l: AAGCTTCCAG GCCGCATACG CTGACGGAAA CAGCGGCATA CCGGAACTGG ATGCTTTGGC 60

CAAGACCGAG GGACAATAAT CCACCGCACT GAGTAAAGTA CAAGCACAGC GTCATAACTG 120

AATTGAGCAG GCTCATTCAA ACGCGGTGAA CAACAAAAAT GGCCGGTAGC GCAAATGCTG 180

CGGGCCGTTT TCGCATCGAT GGTGATGTGA TTCTAAAGGG ATTGGGTCCC CCATCAGTTA 24 f

ATCACCGCAG AGAGCCGCGC ACACTGGTTT TCACTCAAGG ATTTGCAACA TAAAAAGTTT 300

CATGTGAAAG ACGCCCCCCC TAΛTTGAGGΛ GACATCATGC CTATGCACCG TACCCTGGCC 360

GCGCTTTGTG TCGCGGCAGC CAGCCTGCTG CTTGACTGCG CGCATGCCGC GCCCGCGGCG 420

CCACCTCCCG TACCGAAGAC ACCTGCTGGT CCCTTGACGC TGAGCGGGCA AGGCAGCTTT 480

TTCGTCGGCG GACGTGATGT CACGTCGGAA ACGCTGTCGC TGTCGCCCAA ATACGATGCG 540

CATGGCACGG TCACGGTGGA CCAGATGTAT GTGCGTTACC AGATTCCACA ACGCGCCAAG 600

CGTTATCCCA TCACCTTGAT CCACGGCTGC TGCCTGACCG GCATGACCTG GGAAACCΛCG 660

CCCGATGGCC GCATGGGCTG GGACGAGTAC TTCCTGCGCA AGGGCTACTC GACCTACGTC 720

ΛTCGATCAAT CGGGACGCGG CCGCTCCGCC ACGGACATCT CCGCCATCAA CGCGGTGAAG 78C

CTGGGCAACC CGCCCGCCΛG CTCGCTGCCC GACCTGTTTG CCGCCGGACA CGAAGCGGCC 84'

IGGGCGATAT TCCGCTTCGG TCCGCGCTAC CCGGACGCGT TCAAGGACAC CCAGTTTCCC 900

GTGCAGGCCC AGGCCGAACT GTGGCAACAG ATGGTGCCCG ACTGGCTCGG TTCCATGCCC 60

ACACCCAATC CGACCGTGGC CAACCTGTCC AAGCTGGCCA TCAAGCTGGA CGGCACCGTG 102C

CTGCTTAGCC ATTCACAGTC AGGCATCTAT CCCTTCCAGA CCGCGGCCAT GAACCCCAAG 1080

GCCATTΛCCG CCATCGTTTC GGTGGAGCCG GGCGAATGCC CGAAGCCGGA GGACCTCAAG 114C

CCGTTGACCA GCATTCCCGT CCTGGTCGTC TTCGGCGATC ACATCGAGGA ATTTCCACGC 1200

TGGGCGCCGC GCCTGAAGGC CTGCCATGCC TTCATCGACG CCCTGAACGC GGCGGGCGGA 126C

AAAGGACAGC TCATGAGCCT GCCGGCATTG GGCGTGCACG GCAATTCGCA CATGATGATG 1320

CAGGACAGGA ACAACCTGCA GGTGGCGGAT CTGATCCTGG ATTGGATCGG CCGCAACACC 1380

GCCAAACCGG CGCATGGCCG CTGAGCGGCG CGGCGGCCGA CGATGCGTAC CGTGGTCAGG 1440

CCGAGΛCTGG TGCCGGCCGT CCGCTGTTGG CACTGCGCTT AGCGCTAATA TTGCCCGATC 1500

CCCATCCCAA GCGACCGCAA GATCGCACGC CAAGCACATA GAGGAAACAC CGGCATCATG 1560

ATGACGCAGA TGAAAGGACT GTATGTAGTG GCGCAGACGC CGTTCGCGGC CGATGGCGCA 1620

CTGGACCTGG ACAGCATCGA CACCCTCACC GACTTCTACT TCAAGCACGG CGCCGCCGGC 1680

CTGACGGTGC TGGGCGTGGC GGGCGAAGCG GCCAAGCTGT CGGAAGCTGA ATCCGTCAGC 1740

GCGGTCAAGC GTTTCGTCGC GCGTGCCCAG GGCAAGCCCG TCATCGCCGG CGTCAGCAAT 1800

CCCAGCGTCG CGCAACTGGC CACGCTGACC CAGGCCGTGA CCGAAGTCGG CGCCAGCGGC 1860

GTCATGΛTCG CCCCGCCGCC CGGCCTGCGC ACCGAGGAAG ACGTCATCΛG CTACTACGGC 1920

GCCGTGTTCG ATCGTATCGG CGACGTCCCT GCCGTGCTGC AGGACTTTCC TTTTTCCACC 1980

GGCGTCTGGA TGTCGGTGCC CACCATGACC CGGCTGATCG AGCGCTTCCC GCAGATCCAG 2040

GCCATCAAGG AAGAAGACAT CCCCAGCGCC GCCAAGATTA CGCGCCTGCG TGAGGCGATG 2100

CCGCGCCACG TGCCCATCCT CACCGGCAAC AACGCCGTGT TCCTGCCACA GGAACTGGGC 2160

CGCGGCATCG AGGGCCCCAT GGCGGGTTTT TCCCATCCGG AGATGCTGTC CGGCGTGTAT 2220 GCGCTGTACA CCGAAGGCAA TCGCGACGCC GCCTTCGACC TGTTCAATCT GTACCTGCCG 2280

CTGCTGAACT TCGAGAACCA GGCGCAATGG GGCGTCGCCG TGCGCAAGGA AATCCTGCGC 2340

CGCCGCGGCG CCATCGCGTC CGCGGTCATG CGCGCGCCTG GCCCGCGCCT GAGCCCGCTG 2400

GACCTCGCGG ATATCGACTA CCTGCTTGAA CGCCTGGGGC AGGCGCTCGC GCAACGGCGC 2460

GTTTGATGTC CGACCATTCC TGCTACGCCA CCGCCGGGCG TCAGCCCCGG CACGCCTGGC 2520

GCTGACGATC CGACCTCCGC TCAGGTACAA GACAACCGCC CTGTGGCAGT ATCGCGGGTA 2580

TCGTGGACAG ACGCCGCGGC GGCTACCGCG CACCACGCTΛ TCGGGATTCC AAGGGGAAAC 2640

ATGAAGAAGA TCTGTATTTT CGGCGCCGGT TCGGTCGGCG GCCATATGGC GGCGCGCCTG 2700

GCGCAAGCCG GCCTGGACGT TTCGGTGGTG GTGCGCGGCG CGCATCTCGC GGCCATCCGC 2760

AAGAACGGCC TGCGCTTCAC CGGACCGGAC GCCGACTTCA CCGTGCCGGT GCACGCCAGC 2820

GACAATCCGG CCGACCTGGG CCCGCAGGAC CTGGTCATCT CCACCCTGAA AGCCCAGTCG 2880

CTGGGCGCCG CGGCCGCCGG CATGGCCACG CTGCTCAAGC CGGAAACCCC GGTGGTATTC 2940

GCGGTCAACG GCATTCCCTG GTGGTACTTC CATGGCATGC CCGACGACGG CAGCGACAAG 3000

CCCGCCGCCA CTCGCCTGCC CCGCCTGGAT CC 3032

(2) INFORMATION FOR SEQ ID NO : 2 :

(I) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 355 amino acids

(B) TYPE: amino acid

(C) STRANDEDNESS: single

(D) TOPOLOGY: linear

(ii) MOLECULE TYPE: protein (v) FRAGMENT TYPE: N-terminal (vi) ORIGINAL SOURCE:

(B) STRAIN: NCIMB 40700

(xi) SEQUENCE DESCRIPTION: SEQ ID NO: 2:

Met Pro Met His Arg Thr Leu Ala Ala Leu Cys Val Ala Ala Ala Ser

1 5 10 15

Leu Leu Leu Asp Cys Ala His Ala Ala Pro Ala Ala Pro Pro Pro Val

20 25 30

Pro Lys Thr Pro Ala Gly Pro Leu Thr Leu Ser Gly Gin Gly Ser Phe

35 40 45

Phe V L Gly Gly Arg Asp Val Thr Ser GIu Thr Leu Ser Leu Ser Pro 50 55 60

Lys Tyr Asp Ala His Gly Thr Val Thr Val Asp Gin Met Tyr Val Arg 65 70 75 80

Tyr Gin lie Pro Gin Arg Ala Lys Arg Tyr Pro lie Thr Leu lie His

85 90 95

Gly Cys Cys Leu Thr Gly Met Thr Trp Glu Thr Thr Pro Asp Gly Arg

100 105 110

Met Gly Trp Asp Glu Tyr Phe Leu Arg Lys Gly Tyr Ser Thr Tyr Val

115 120 125

He Asp Gin Ser Gly Arg Gly Arg Ser Ala Thr Asp He Ser Ala He

130 135 140

Asn Ala Val Lys Leu Gly Lys Ala Pro Ala Ser Ser Leu Pro Asp Leu 145 150 155 160

Phe Ala Ala Gly His Glu Ala Ala Trp Ala He Phe Arg Phe Gly Pro

165 170 _75

Arg Tyr Pro Asp Ala Phe Lys Asp Thr Gin Phe Pro Val Gin Ala Gin

180 185 190

Ala Glu Leu Trp Gin Gin Met Val Pro Asp Trp Leu Gly Ser Met Pro

195 200 205

Thr Pro Asn Pro Thr Val Ala Asn Leu Ser Lys Leu Ala He Lys Leu

210 215 220

Asp Gly Thr Val Leu Leu Ser His Ser Gin Ser Gly He Tyr Pro Phe 225 230 235 240

Gin Thr Ala Ala Met Asn Pro Lys Gly He Thr Ala He Val Ser Val

245 250 255

Glu Pro Gly Glu Cys Pro Lys Pro Glu Asp Val Lys Pro Leu Thr Ser

260 265 270

He Pro Val Leu Val Val Phe Gly Asp His He Glu Glu Phe Pro Arg

275 280 285

Trp Ala Pro Arg Leu Lys Ala Cys His Ala Phe He Asp Ala Leu Asn

290 295 300

Ala Ala GJ y Gly Lys Gly Gin Leu Met Ser Leu Pro Ala Leu Gly Val 305 310 315 320

His Gly Asn Ser His Met Met Met Gin Asp Arg Asn Asn Leu Gin Val

325 330 335

Ala Asp Leu He Leu Asp Trp He Gly Arg Asn Thr Ala Lys Pro Ala 340 345 350 H s G l y Arg

(2) INFORMATION FOR SEQ ID NO : 3 :

(l) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 29 base pairs

(B) TYPE: nucleic acid

(C) STRANDEDNESS: unknown

(D) TOPOLOGY: unknown

(n) MOLECULE TYPE: Other

!xι) SEQUENCE DESCRIPTION: SEQ ID NO : :

TGAGGAGACA CCATGGCTAT GCACCGTAC 29

(2) INFORMATION FOR SEQ ID NO: :

( ) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 27 base pairs

(B) TYPE: nucleic acid

(C) STRANDEDNESS: unknown ([-) TOPOLOGY: unknown

(xi) SEQUENCE DESCRIPTION: SEQ ID NO : 4 :

TATTAGCGCT AAGCTTAGTG CCAACAG ? /

ΛppliLΛnt s or

s file International application . rclcrcncc num 3 1575

INDICATIONS RELATING TO A DEPOSITED MICROORGANISM

(PCT Rule I3bιs)

A. The indications made below relate to the microorganism relerred to in the description on page , line 1 3

II IDENTIFICATION OF DEPOSIT Further deposits are identified on an additional sheet | |

Name ot depositary institution

National Collection of Industrial and Marine Bacteria Ltd (NCIMB)

Address ot depositary institution (including postal code and country)

Aberdeen Scotland United Kingdom

Date ol deposit Accession Number

1 Decemoer 199^ NCIMB 40700

C. ADDITIONAL INDICATIONS (leave blank if not applicable) This information is continued on an additional sheet Q

In respect of those designations m which a European Patent is sought, a sample of the deposited micro-organism will be made available until the publication of the mention of grant of the European Patent or until the date on which the application has been refused or withdrawn, only by the issue of such a sample to an expert nominated by the person requesting the sample.

D. DESIGNATED STATES FOR WHICH INDICATIONS ARE MADE (if the indications are not for all designated States)

E. SEPARATE FURNISHING OF INDICATIONS (leave blank if not applicable)

The indications lisied belowwill be submitted to the International Bureau later (specify the general nature of the indications e , 'Accession

Number of Deposit'')

For receiving Office use onlv For International Bureau use only

I ) This sheet was received with the international application [ I This sheet was received by the International Bureau on.

Authorized officer Authorized officer

Form PCT RO/134 (July 19.₂)

Claims

1. A method of screening for microorganisms capable of the resolution of compounds of formula (I), which method comprises selecting for organisms capable of using compounds of formula (I) as a sole carbon source and assaying cells of the organism for specific esterase activity.

where R is c 1-6 alkyl

? A microorganism identified using the method of claim 1.

3. Microorganism NCIMB 40700.

4. An enzyme capable of resolving compounds of the formula (I) obtainable from NCIMB 40700.

5. An enzyme according to claim 4 having an apparent molecular weight of 35 kD by LCMS.

6. An enzyme according to claim 4 or 5 in the form of a dimer.

7. An enzyme according to any one of claims 4 to 6 which comprises substantially the sequence of amino acids shown in SEQ ID NO:2.

8. A process for the production of an enzyme according to any one of the claims 4 to

7 comprising stirring disrupted cells of NCIMB 40700 with WK10 ion-exchange resin, eluting the enzyme with NaCl at pH 7.5, concentrating the enzyme solution by ultrafiltration, mixing with A568 resin then subjecting the extract to phenylagarose, Q- sepharose, S-sepharose and blue sepharose chromatography.

9. An isolated polynucleotide comprising a nucleotide sequence that has at least 80% identity over its entire length to a nucleotide sequence encoding an enzyme as defined in any one of claims 4 to 7; or a nucleotide sequence complementary to said isolated polynucleotide.

10. The polynucleotide of claim 9 wherein said polynucleotide comprises the nucleotide sequence contained in SEQ ID NO: 1 encoding the polypeptide of SEQ ID

NO:2.

1 1. The polynucleotide of claim 9 wherein said polynucleotide comprises a nucleotide sequence that is at least 80% identical to that of SEQ ID NO: 1 over its entire length.

12. The polynucleotide of claim 1 1 which is polynucleotide of SEQ ID NO: 1.

13. The polynucleotide of any one of claims 9 to 12, which is DNA or RNA.

14. DNA which hybridises under conditions of high stringency to the DNA of any one of claims 9 to 13 or fragments thereof.

15. A vector comprising DNA according to any one of claims 9 to 14 for expressing an enzyme having esterase activity in a suitable host organism.

16. A host cell transformed with a vector according to claim 15.

17. A host cell of claim 16, wherein host cell is an Agrobacterium species.

18. A process for producing recombinant enzyme having esterase activity in host cell of claims 16 or 17, comprising culturing the transformed host cell under conditions permitting expression of the said enzyme and recovering the said protein.

19. A process for stereospecifically hydrolysing a mixture of the (+) and (-) isomers of a compound of formula (I) using an enzyme as claimed in any one of claims 4 to 7 or a microorganism of claim 2, 3, 16 or 17:

(i) to form a compound of formula (IIA),

O N O

Me (+) isomer (HA)

and thereafter separating the resulting compound of formula (IIA) from the remaining (-) isomer of formula (I); or ii) to form a compound of formula (IIB):

COV-/₂,H

O N O

Me .-) isomer (IIB)

and thereafter separating the resulting compound of formula (IIB) from the remaining (+) isomer of formula (I).

20. A process according to variant (i) in claim 19 in which the stereoselectivity of the process is such that from a racemic mixture of a compound of formula (I), after the action of the enzyme, the ratio of (-) to (+) isomer of formula (I), is greater than 60%.

21. A process according to variant (ii) in claim 19 in which the stereoselectivity of the process is such that from a racemic mixture of a compound of formula (I), after the action of the enzyme or microorganism , the ratio of (+) to (-) isomer of a compound of formula (I), is greater than 60%.

22. A process according to any one of claims 19 to 21 which is carried out by dissolving (±) unresolved compound of formula (I) into an aqueous/organic solvent mixture and adding the enzyme or microorganism and stirring the resulting mixture until the reaction is completed.

23. A process according to any one of claims 19 to 22 in which the variable R in formula (I) is methyl or ethyl.

24. A process as claimed in any one of claims 19 to 23 wherein the microorganism belongs to the genus Alcaligenes.

25. A process as claimed in any one of claims 19 to 23 wherein the microorganism is NCIMB 40700.

26. A process as claimed in any one of claims 19 to 25 wherein cell free extracts of the microorganism are used.

27. A process as claimed in any one of claims 19 to 26 wherein the enzyme, extract or microorganism is immobilised.

28. A process for subsequently converting a compound of formula (IIB), prepared as described in any one of claims 19 to 23, to paroxetine or a pharmaceutically acceptable salt and/or solvate thereof.

29. A process according to claim 28, comprising first converting the compound of formula (IIB), as defined in claim 19, using conventional esterification techniques followed by subsequent convertion to paroxetine or a pharmaceutically acceptable salt and/or solvate thereof.

30. A process according to claim 28, comprising reducing the carboxyiic acid group of the compound of formula (IIB) to a hydroxymethyl group, and reducing the two keto groups in the piperidine ring and subsequently, converting the resulting piperidine carbinol to paroxetine or a pharmaceutically acceptable salt and/or solvate thereof. 31 A process according to any one of claims 28 to 30 wherein the pharmaceutically acceptable salt is paroxetine hydrochloπde

32 A process according to claim 31 wherein the pharmaceutically acceptable salt is paroxetine hydrochloπde hemihydrate

33 A process according to claim 31 wherein the pharmaceutically acceptable salt is paroxetine hydrochloπde anhydrate

34 A protein having an αβ-hydrolase fold wherein the acidic residue of the catalytic tπad appears in the end of strand 6

35 A hydrolase enzyme having two adjacent cysteme residues close to its active site that are linked together by a disulphide bond