PREPARATION OF ENZYME WITH REDUCED -LACTAMASE ACTIVITY Field and background of the invention The present invention relates to the enzymatic bioconversion of P-lactam compounds. Among the antimicrobial agents the p-lactam antibiotics still stand out as an important class of compounds for their broad spectrum of activity and low toxicity to mammals. The typical feature of the antibiotic molecules of the -lactam type is the presence of a highly strained P-lactam ring. The P-lactam antibiotics are composed of two classes, the penicillins and cephalosporins. In penicillins the -lactam ring forms together with the pentagonal thiazolidine ring the penam nucleus, in cephalosporins the -lactam ring forms together with a six-membered dihydrothiazine ring the cephem nucleus.

In general the substituent at the 6-amino group of the penicillin nucleus or at the 7-amino group of the cephalosporin nucleus is referred to as the 'acyl side chain' or just as the 'side chain', the corresponding acid as the 'side chain acid'.

The side chains are of importance because of their ease of chemical modification and the profound effects which they exert on the properties of the molecule. Different side chains can render the -lactam nucleus more resistant to degradation by -lactamases, tolerant to gastric acidity, or enable the molecule to penetrate the outer envelope of gram-negative organisms.

-Lactam antibiotics are believed to kill bacteria by interference with the proper synthesis of the bacterial cell wall. Their mode of action is on certain essential hydrolytic enzymes which are involved in the cell wall synthesis. The hydrolysis of the p-lactam bond by these enzymes gives rise to a virtual irreversible inhibition by the hydrolysis product. In addition many of the target microorganisms may develop or have

already developed a defense mechanism against p-lactam antibiotics by producing hydrolytic enzymes which degrade the -lactam ring by hydrolysing the amide bond in the -lactam ring without subsequent inhibition. Enzymes which catalyze the hydrolysis of said amide bond are called -lactamases.

The basic antibiotics of the -lactam type are principally obtained by fermentation. Fungi of the genus Penicillium and Cephalosporium (Acremonium) are used for the production of raw material for P-lactam antibiotics such as penicillin G, penicillin V and cephalosporin C. These fermentation products, also referred to as PenG, PenV and CefC, respectively, are the starting materials for nearly all currently marketed penicillins and cephalosporins. The side chains of PenG, PenV and CefC are phenylacetyl, phenoxyacetyl and aminoadipyl, respectively. The side chains are removed by cleavage of an amide linkage (deacylation), resulting in 6-aminopenicillanic acid (6-APA) in case of the penicillin molecules and 7-aminocephalosporanic acid (7-ACA) in case of the cephalosporin molecule. In this respect also phenylacetyl-7-aminodesacetoxycephalosporanic acid (CefG) should be mentioned as a precursor of the cephalosporin 7-ADCA, although CefG is not a fermentation product. CefG is usually produced chemically from Penicillin G. In addition new routes have been disclosed which show a more efficient process for the production of 7-ADCA, 7-ADAC and 7-ACA (EPA 540 210 Al; WO 93/08287; WO 95/04148 and WO 95/04149).

-Lactam nuclei such as 6-APA, 7-ACA and 7-ADCA are used as starting points for synthetic manipulation leading to the so-called semi-synthetic penicillins and cephalosporins. At present these semisynthetic penicillins and cephalosporins form by far the most important market of P-lactam antibiotics (J.N.

Rolinson in J. Antimicrob. Chemother. 5 (1979) 7-14; J.

Antimicrob. Chemother. 22 (1988) 5-14) The production of semisynthetic -lactam products requires the deacylation of the penicillins and cephalosporins produced from fermentation. Although rather efficient chemical routes are available for the deacylation (J. Verweij & E. de Vroom, Recl.

Trav. Chim. Pays-Bas 112 (1993) 66-81), nowadays the enzymatic route is preferred in view of the high energy and solvents cost together with some environmental problems associated with the chemical route (Dunnill, P., Immobilised Cell and Enzyme Technol- ogy. Philos, Trans. R. Soc. London B290 (1980) 409-420). The enzymes which accomplish the deacylation of P-lactam compounds are classified as hydrolases based on the chemical reaction they catalyze. However, those hydrolases which are of particular interest in the deacylation of -lactam compounds are usually referred to in the art as 'acylases' or 'amidases', e.g.

'Penicillin G acylase', 'Penicillin V acylase', etc.. In general when a certain conversion is catalyzed by any biological derived activity, this activity is often referred to a biocatalyst. The biocatalyst may comprise whole cells, mixtures of enzymes or just one type of enzyme.

In addition it has been reported that under appropriate conditions many of these type of acylases are able to add alternative side chains to the -lactam nucleus which allows for enzymatic synthesis of semi-synthetic -lactam antibiotics (T.A.

Savidge in Biotechnology of Industrial Antibiotics (Ed E.J.Vandamme) Marcel Dekker, New York 1984; J.G.Shewale et al.

Process Biochemistry International June (1990) 97-103; E.J.

Vandamme in Enzyme Microb Technol. 5 (1983) 403-416).

The cephalosporin acylases are able to remove the polar side chains which occur in most cephalosporins. Cephalosporin acylases can be classified according to their substrate specificity in the true cephalosporin C acylases which show activity for both cephalosporin C and dicarboxyl N-acylated -lactam nuclei and in the dicarboxyl acylases which deacylate exclusively dicarboxyl N-acylated -lactam nuclei (K.K. Kumar et al. in Hindustan Antibiotics 35 (1993) 111-125; EP-A-322032)).

An example of a dicarboxyl acylase is glutarylacylase which deacylates glutaryl-7-ACA. Glutaryl-7-ACA is enzymatically prepared from CefC by enzymatic deamidation of the side chain with D-amino acid oxidase followed by chemical decarboxylation of the formed ketoadipoyl derivative with hydrogenperoxide, which

is produced in the first step. Such processes are performed at an industrial scale.

Useful enzymes for the conversion of p-lactam antibiotics can be identified by extensive screening programs. Once the enzyme has been identified the enzyme might be recovered from the original host. However this has several drawbacks. Usually, the fermentative manufacturing of large amounts of an enzyme catalyst in an economically feasible way is seriously hampered by low enzyme expression levels. Improvement of the expression level requires extensive random mutagenesis and screening for over-producing strains in order to improve the expression level of the enzyme of interest. In addition undesired side activities such as for example -lactamase activities are often produced in minor quantities which are very difficult to remove. Again, elimination of these undesired activities might be obtained by extensive random mutagenesis and screening for mutants which do not or do produce less P-lactamase. If such screening is unsuccessful then usually an elaborate purification is necessary to remove side activities such as for example p-lactamase activity.

Nowadays genetic engineering allows for the construction of very efficient production organisms in a very controlled manner. This implies homologous expression as well as heterologous expression of enzymes. If the identified enzyme is difficult to obtain from the original strain, recombinant-DNA techniques may be used to isolate the gene encoding the enzyme, express the gene in another organism, isolate and purify the expressed enzyme and test whether it is suitable for the intended application. Suitable hosts are for example E. coli, Bacilli, Aspergilli, and yeasts such as Kluyveromyces lactis. Nowadays several acylases are produced by genetically engineered strains (B.S. Deshpande et al. in World Journal of Microbiology & Biotechnology 10 (1994) 129-138; K.K.Kumar et al., v.s.). In addition some acylases have been expressed in B.subtilis, an organism which allows for excretion of the produced acylase (Aramori et al. in J.Bacteriology 173 (1991) 7848-7855, A. Olsson

et al. in Applied and Environmental Microbiology 49 (1985) 1084-1089, J.H. Kang et al. in J. Biotechnology 17 (1991) 99-108; H. Ohashi et al, Applied and Environmental Microbiology 55 (1989) 1351-1356).

The application of enzymes in the modification of -lactams is not limited to the acylases and the D-amino acid oxidases. Different esterases have been reported which allow for the enzymatic deacetylation of 7-ACA at the 3-position (B.J.

Abbot et al. in Applied Microbiol. 30 (1975) 413-419; K. Singh in Eur. J. Appl. Microbiol. Biotechnol. 9 (1980) 15-18, K.

Mitsushima et al. in Applied and Environmental Microbiol. 61 (1995) 2224-2229; EPA 0 454 478) . In large scale chemical synthesis of -lactam antibiotics enzymatic steps have been introduced in order to selectively convert one particular enantiomer when chemical steps lead to racemic mixtures (M.J.

Zmijewski et al. in Tetrahedron Letters 32 (1991) 1621-1622). In addition sensitive groups have to be protected during chemical synthesis. Afterwards this protection has to be removed under mild conditions which do not degrade the sensitive -lactam compound. Therefore these deprotection reactions are optimally suited for enzymatic reactions. An example of such enzymatic deprotection is the hydrolysis of the para-nitrobenzyl protected carboxyl function as a final step in the chemical synthesis of loracarbef (Brannon et al. in J. Antibiotics (1975) 29:121-124; US patent 3,725,359; J. Zock et al. in Gene 151 (1994) 37-43).

The enzymatic deprotection of a 7-ACA amino thiazolyl protected adduct has been described in EP 0 582 102. The introduction of a methoxygroup at the 7-position of the cephem nucleus (cephamycins) could be carried out enzymatically by a cell-free extract of Streptomvces Clavuliserus. In a similar way other enzymes from the biosynthetic -lactam synthesis pathway may be used in an isolated form for specific bioconversions. Examples are the use of desacetoxy-cephalosporin C synthetase (DAOCS) for the conversion of 3-exomethylenecephalosporin C to desacetylcephalosporin C (US 5,082,772) and the use of deacetoxycephalosporin C hydroxylase (DACS) in substantially pure

form for the hydroxylation of deacetoxycephalosporin C (DAOC) (EP 0 465 189). Although such enzymes may be isolated from their natural host, nowadays most of these enzymes are cloned and expressed in high productive hosts such as for example E. coli (Ghag et al. Biotechnol. Appl. Biochem 24 (1996) 109-119).

In the process of constructing recombinant strains for the production of enzymes for -lactamase sensitive bioconversions, it is common to use cloning vectors which are devoid of any p-lactamase activity instead of plasmids which contain the ampicillin resistance marker. However many microorganisms still show a basal level of P-lactamase activity which might be caused by chromosomal encoded P-lactamases. Also in this case one could try to inactivate the responsible gene(s) by genetic means, however this leads often to crippled organisms because such minor -lactamase activities are indispensable for cellular fitness and growth. Selection for hosts which do not produce detectable -lactamase activity for production of enzymes for -lactamase sensitive bioconversions is an approach to circumvent the problem. Unfortunately those hosts which are poor in -lactamase activity do not always coincide with the hosts which are optimal for production. In addition, under the stress conditions which may occur under large scale production conditions, expression of -lactamase activity might be induced to a significant level. So, the only way to be sure that no -lactamase activity will be introduced with the biocatalysts in a p-lactam bioconversion process is an extensive purification procedure for the biocatalyst.

In large scale bioconversion processes involving p-lactams, the operational efficiency as well as the economics of the process is improved by immobilisation of the enzymes on solid carriers (K. Matsumoto in Bioprocess Technol. 16 (1993) 67-88; J.G. Shewale & H. Sivaraman in Process Biochemistry August (1989) 146-154; T.A. Savidge in Biotechnology of Industrial Antibiotics (Ed E.J. Vandamme) Marcel Dekker, New York 1984; D.A.

Cole in Developments in Industrial Microbiology 30 (J. Ind.

Microbiol. suppl. 4) (1989) 121-130, WO 95/16773, USPatent

4,001,264, GB 2244711; European Patent Application 496993) Immobilisation usually leads to improved stability of the enzyme and allows for an easy separation from the liquid reactants. As a consequence immobilised enzymes can be reused many times which makes many enzymatic processes economically feasible. A large number of methods for biocatalyst immobilisation has been developed during the past three decades and still continues to expand ('Protein Immobilization Fundamentals and Applications', Ed R.F. Taylor, Marcel Dekker, Inc (1991)) . Immobilisation of enzymes can be obtained in many ways. These methods include cross-linking, physical adsorption, ionic binding, metal binding, covalent binding and entrapping. Coupling of enzyme molecules to solid supports involves the reactions between aminoacids of the enzyme and reactive groups on the carrier. As carriers for enzymatic activities, organic as well as inorganic, synthetic as well as natural materials have been used. The solid particles which result from the immobilisation of enzymatic activity are usually referred to as 'immob'. Immobs may contain one single type of enzyme or combinations of enzymes which work together.

In addition immobs may contain whole cells which in some cases are further coimmobilised with particular enzymes.

For carriers which contain amino groups as their functional group glutardialdehyde has been used to couple enzymes covalently to the carrier. The amino groups of the carrier are activated with a dialdehyde. After washing away the excess of dialdehyde the enzyme is added to the carrier and allowed to react with the aldehyde groups on the carrier. As an alternative the carrier, the enzymes and the glutardialdehyde may be mixed together. Entrapment of biological active material (cells or enzymes) by suitable matrix substances is usually carried out in the presence of cross-linking agents such as glutardialdehyde.

However until now it has not been reported that glutardialdehyde treatment of immobilised enzymes can reduce the undesired -lactamase activity which is co-immobilised in the immobilisation process.

Summary of the invention The present invention relates to an efficient method for the inactivation of undesired P-lactamase activity in the preparation of one or more enzymes (biocatalysts) which are used in the enzymatic conversion of P-lactamase sensitive compounds and to these preparations. The method comprises the treatment of the biocatalyst with an aldehyde, preferably a dialdehyde, compound. The method can be applied to liquid biocatalysts as well as immobilised biocatalysts, which also surprisingly results in reduction of leakage of protein and activity from the immobilised enzyme. In a preferred embodiment the dialdehyde is glutardialdehyde.

Detailed description of the invention The present invention relates to the use of dialdehydes to eliminate undesired P-lactamase activities in biocatalysts which are used in the enzymatic conversion of p-lactamase sensitive compounds. The present invention relates to the use of preparations of one or more enzymes as biocatalysts to convert -lactam compounds in large scale industrial processes. Such p-lactam compounds are known to be very sensitive to the degradation by -lactamases. The invention is based on the finding that the biocatalysts which are used often contain minor amounts of -lactamase activity. Whether such p-lactamase activity is a problem depends on the specificity of the p-lactamase and the P-lactam compounds which are present during the enzymatic conversion.

Biocatalysts may comprise whole cells, cell free extracts, or isolated enzymes in the free form or in any suitable immobilised solid form. The biocatalyst may comprise the natural host which expresses a certain biocatalytic activity or any recombinant organisms which expresses the desired biocatalytic activity. The biocatalytic activity may comprise one or more enzymes. General cloning techniques were performed as described

by Maniatis (Maniatis et al., "Molecular Cloning, A Laboratory Manual", Cold Spring Harbor Laboratory, 1982/1989) , Ausubel (Ausubel et al., "Current Protocols in Molecular Biology", John Wiley and Sons, Inc., New York, 1987 and Perbal (Perbal B., "A Practical Guide To Molecular Cloning", 2nd ed., John Wiley and Sons, Inc., New York, 1988). These handbooks describe in detail the protocols for construction and propagation of rDNA molecules, the procedures for making gene libraries and the protocols for mutating DNA in a site-directed or random fashion. Enzymes used for DNA manipulations were purchased from commercial suppliers and used according to their instructions. Plasmids and E. coli cloning hosts were obtained from public culture collections. In this way we were able to express the wild type genes coding for the acylases from Escherichia coli, Alcaligenes faecalis, Pseudomonas SY-77 and Pseudomonas SE-83 AcyII in E. coli (USP 5,457,032).

In addition a useful biocatalyst may comprise mutants of naturally occurring enzymes. E. coli strains JM101, WK6, HB101, PC2051, PC1243, NC1061, DH1 and RV308 have been used as expression hosts. Expression of acylase genes was obtained either from the homologous expression signals or from the E. coli lac, tac or trp promoter (De Boer et al., Proc. Natl. Acad. Sci. USA 80 (1983) 21-25) . However it will be understood that the described expression constructs and expression hosts are just given as examples and it will be apparent to those skilled in the art that certain changes and modifications may be made which result in alternative expression constructs and/or alternative expression hosts.

For the production of acylase by the described E. coli constructs, said E. coli constructs are grown in media which contain a carbon source (e.g. glucose, sucrose, lactose, etc.) a nitrogen source (yeast extract, soy meal, corn-steep liquor, bacto-peptone, etc.) and mineral salts. In the case of the autotrophic strains the essential amino acids are added to the fermentation. The temperature during the fermentation is between 150C and 370C. The pH set point is set to lower limit pH 6, the

upper limit is not controlled. Preferentially the pH is around pH 7. In case of intracellular production of the acylase, the acylase produced may be recovered by separating the cells and immobilise them by conventional procedures. Separation of the cells may be carried out by centrifugation or appropriate filtration methods such as a rotating vacuum filter. In case of any extracellular enzyme the enzyme activity can be recovered from the filtrate. Intracellular enzymes may be isolated from the cells by lysis which can be carried out by mechanical rupture (pressure, sonification, osmotic shock) or chemical methods (lysozyme/EDTA). The extract obtained in this way may be further purified when required by precipitation, chromatographic and membrane techniques. Typical operating conditions for a homogeniser to disrupt the cells may be 500-1000 bar, cooling in order to control the temperature below 150C, 1-3 passages.

Subsequently the suspension may be subjected to filtration in the presence of suitable flocculants such as C581 and/or dicalite 4108. Next the filtrate is subjected to germfiltration and ultrafiltration in order to wash and concentrate the enzyme activity. This enzyme liquid solution can be used as such for bioconversion but may also be immobilised on a suitable support.

Once the biocatalyst is contacted with compounds which comprise a P-lactam ring, it is essential that the biocatalyst is devoid of any -lactamase activity which can degrade any of the reactants. Therefore such a biocatalyst should be assayed for the presence P-lactamase activity by a relevant p-lactamase assay. The presence of P-lactamase activities is indicated by the presence of typical degradation products such as penicilloic or cephalosporoic acid, which can be measured by conventional analytical techniques such as for example HPLC or NMR. The destruction of the -lactam bond in cephalosporins and penicillins can also be followed spectrofotometrically. In addition to these direct methods there are also many indicator substrates which can reveal the presence of -lactamase such as for example the chromogenic cephalosporins 'Nitrocefin' or 'Oxoid' and (6R,7R)-3-[(E)-2-(2,4-dinitrofenyl)vinyl]-7-

fenylaceetamido-3-cefem-4-carboxylic acid, (6R,7R)-3-[(E)-2,4- dinitrostyryl] -7-fenylaceetamido-3-cefem-4-carboxylic acid <BR> <BR> <BR> ('Cefesone'), 3oe- (hydroxymethyl) -2, 2-dimethyl-6 - (phenoxyacetamido)penam or 6- (P-furylacryloylamido)penicillin (FAP).

When whole cells are used further purification in order to remove the p-lactamase activity before immobilisation is impossible as in general P-lactamase activity is connected to the cells. For isolated enzymes a further purification may be applied in order to remove the P-lactamase activity. Usually one or more chromatographic steps have to be performed. In general each additional purification step will also lead to losses of the target enzyme activity, reducing the overall yield of the purification process. In addition, each additional purification step requires additional handling and additional investments in down stream processing such as chromatography columns and resins.

This will increase the costs of the enzyme significantly.

The acylases from Escherichia coli, Alcaligenes faecalis, Pseudomonas SY-77 and Pseudomonas SE-83 AcyII and variants thereof were produced in E. coli strain HB101. As these enzymes would be contacted with cephalosporins, the enzyme liquid which resulted after recovery, was tested for the presence of -lactamase activity which would be able to degrade cephalosporins. The chromogenic substrate <BR> <BR> <BR> (6R, 7R) -3- [(E) -2- (2,4-Dinitrofenyl)vinyl] -7-fenylaceeta- mido-3-cefem-4-carboxylic-acid indicated that such activity was present. When contacting the enzyme liquid with CefG, CefG was degraded into non P-lactam products.

The enzyme liquids were immobilised on a carrier which was prepared as described in EPA 222462, using chitosan and gelatin as a gelling agent and glutardialdehyde as cross-linking agent.

The free aldehyde groups on the carrier are allowed to react with free amino groups of the enzyme. After 16-20 hours the immob is intensively washed with water. Upon said immobilisation the -lactamase activity was co-immobilised and could not be separated anymore from the desired enzyme activity. All the

benefits of immobilised enzymes with respect to repeated use and increased stability did hold also for the p-lactamase activity.

Although a chitosan-gelatin carrier was used in this case, it will be apparent to those skilled in the art that the principle of inactivation of -lactamase by glutardialdehyde treatment is not restricted to the type of carrier which is used.

Surprisingly we discovered that when the immob is treated with glutardialdehyde, even when the same does contain already glutardialdehyde, for instance as cross-linking agent, the -lactamase activity decreased enormously while the desired biocatalytic activity was maintained. Usually the final -lactamase activity became lower than the minimal levels which could be reliably detected with the used -lactamase assay. The glutardialdehyde concentration which is applied in said -lactamase inactivation process is preferably in the range of 0.05-10%, more preferably 1-5%. The pH of said process may vary from pH 4 up to pH 9, preferably around 7. The incubation temperature is usually between 150C and 400C. The incubation time of said process is usually between a few minutes and 14 hours, preferably 1-7 hours. In principle one could monitor the -lactamase activity during said inactivation process and stop when the -lactamase activity has become adequately low.

As an alternative to a batch wise treatment of the immob with glutardialdehyde which is removed after treatment, one might also add a low concentration of glutardialdehyde to the immob as a conservation agent. In addition to preserving the immob, the added glutardialdehyde will also reduce the p-lactamase activity effectively. The glutardialdehyde concentration which is applied under these circumstances is preferably in the range of 0.01-0.5% glutardialdehyde.

An extra advantage of the glutardialdehyde treatment of the immob is that the leakage of protein and activity from the immob was reduced considerably during the enzymatic conversion reaction. The reduced leakage will result in less protein impur- ities in the end product and will have a positive effect on the

total amount of product which can be produced with the immobilised catalyst.

Although said method works very efficient when treating immobs, also the liquid enzyme could be treated successfully with said procedure. When liquid enzyme was treated with glutardialdehyde, the P-lactamase activity could be decreased significantly. In comparison with the immobs the inactivation of the p-lactamase activity could be obtained at lower glutardialdehyde concentrations.

Glutardialdehyde concentration might be between 0.01% and 5%. The enzyme concentration should be such that aggregation and precipitation due to intermolecular crosslinking of the desired biocatalyst is prevented as much as possible. Suitable enzyme biocatalyst concentrations are in range of 0.001-8 mg/ml.

The invention will further be illustrated by the following non limiting examples.

Example 1 The inactivation of -lactamase activity in liquid enzyme preparations A liquid enzyme preparation of glutarylacylase which is applied as a biocatalyst in the conversion of glutaryl-7-ACA to 7-ACA is contacted with (6R,7R)-3- [(E)-2-(2,4-Dinitrofenyl)vi- nyl] -7-fenylaceetamido-3-cefem-4-carboxylic acid. This compound is sensitive to the presence of P-lactamase activity. After hydrolysis of the P-lactam ring in (6R,7R)-3-[(E)-2-(2,4-Dini- trofenyl)vinyl] -7-fenylaceetamido-3-cefem-4-carboxylic acid the yellow colour of the substrate turns into a bright red colour.

The change in optical density can be measured spectrophoto- metrically at 510 nm. One unit (U) is defined as the increase of hOD/min x 100 under the following reaction conditions. The reaction is carried out in a cuvette in a spectrophotometer at 370C. The cuvette which contains 980 41 0.2 mM <BR> <BR> <BR> <BR> (6R,7R)-3-[(E)-2-(2,4-Dinitrofenyl)vinyl]-7-fenylaceeta- mido-3-cefem-4-carboxylic acid in 0.1 M TRIS pH 8.0 is preheated

during 5 minutes. Subsequently 10-20y1 of the glutarylacylase preparation which may contain enzyme impurities with -lactamase activity is added to the cuvette, the solution is thoroughly mixed and the reaction is followed at 370C by measuring the OD during 8 minutes at 1 minute intervals. A linear reaction rate is obtained over 8 minutes between 0.19 and 2.16 U/ml mixture in the cuvette. The increase of the OD is corrected for of a blank solution of (6R,7R)-3-[(E)-2-(2,4-dinitrofenyl)vinyl]- 7-fenylaceetamido-3-cefem-4-carboxylic acid to which water instead of a sample was added.

Table 1, incubation A, shows that the glutarylacylase preparation contains a significant amount of -lactamase activity.

The glutarylacylase preparation was treated with glutardialdehyde in the following way: 0.1 ml of the glutarylacylase preparation, 50 mg/ml, was mixed with 0.2 ml sodiumphosphate buffer 0.1 M pH 7.5 and 0.7 ml water.

Subsequently 0 41 (A), 10 41 (B) and 20 41 (C) of a 25% glutardialdehyde solution was added. After 3.75 and 5 hours reaction at 200C the P-lactamase activity was measured in the 3 samples by addition of 10-20 41 of the undiluted samples into the -lactamase activity test.

The results of the P-lactamase inactivation experiment are shown in table 1.

Table 1 Inactivation of -lactamase in glutaryl acylase solution 3.75 hrs 5 hrs µl sam- U/ml in de- # U/ml in U/ml % residual U/ml in # U/ml in U/ml % resi- ple termination determi- sample activity determi- determi- sample dual nation nation nation activity Blank 10 0.23 0.17 A 10 1.31 1.08 108 (100) 1.29 1.12 112 (100) B 10 0.86 0.63 31 29 0.74 0.57 28.5 25 C 20 0.58 0.35 18 17 0.44 0.27 13.5 12

As shown in table 1 the -lactamase activity has been reduced in the samples which have been treated with glutardialdehyde. With twice the amount of glutardialdehyde the p-lactamase activity is reduced more. The reduction is a time dependant process which still proceeds after 3.75 hours reaction time. The best result obtained in this experiment is a reduction of the p-lactamase activity to 12 W of the original activity.

The glutarylacylase activity was measured after 3 and 5 hrs reaction time in the samples A, B and C. The glutaryl- acylase activity can be determined with the chromogenic substrate glutaryl-para-NitroAnilide (glu-pNA) (Franzosi et al. in Appl. Microbiol. Biotechnol. (1995), 43, 508). One glu-pNA unit (U) of enzyme activity will release 1 ymol of the yellow coloured p-nitroanilide in one minute under the given reaction conditions. A 10 mM solution of glu-pNA in 100 mM glycylgly- cine buffer at pH 8.0 is used. The formation of reaction product is measured spectrofotometrically at 405 nm (E = 9.90 x 103 W1 cam~1). In a 1 ml cuvette 990 41 of a preincubated glu- pNA solution is mixed with 10 41 of the enzyme preparation.

The reaction is followed during 10 minutes at 370C with one reading per minute. The activity of the sample is calculated from the change in OD/min by using the linear regression method. A linear reaction is obtained at an enzyme dosage of 200 - 1500 U/l.

In table 2 it is shown that the glutarylacylase activity is reduced to maximal 72 W of the original activity, while the p-lactamase activity in the same sample is reduced to 12 W.

However it should be evident for a person skilled in the art that further elimination P-lactamase activity at the expense of less glutarylacylase activity loss may be obtained by optimisation of the reaction parameters such as for example the glutardialdehyde dosage, the total reaction time, and the reaction circumstances such as pH, temperature, ionic strength etc.

Table 2 Glutarylacylase activity after glutardialdehyde treatment 3 3 hrs reaction time 5 hrs reaction time Sample U/l |t U/l U/l A 1217 (100) 1235 (100) B 1026 85 1010 83 C 881 75 867 72 Example 2 Variations in inactivation method of the -lactamase activity in immobilised glutarylacylase applied on adipyl-7-ADCA 1 ml of settled volume of immobilised glutarylacylase was incubated with 5 ml sodium acetate buffer pH 6.0 (samples 1 to 3) with different amounts of glutardialdehyde (glut) as shown in table 3. In addition 1 ml of settled volume of immobilised glutarylacylase was incubated with 5 ml sodium acetate buffer pH 5.0 (sample 4) with 2% glutardialdehyde end concentration. After 2 hours of incubation at 250C the immobilised enzyme was washed several times with water in order to wash out the unbound glutardialdehyde.

The P-lactamase activity was determined in the following way. An amount of immob was sucked on a glassfilter under re- duced vacuum for several minutes. The resulting wet immob is quantified by its weight, usually indicated as wet weight.

40 mg of wet weight immobilised glutarylacylase were incubated at room temperature with 2 ml 0.1 mM (6R,7R)-3-[(E)-2-(2,4- dinitrofenyl)vinyl]-7- fenylaceetamido-3-cefem-4-carboxylic acid in 0.1 M Tris buffer pH 8.0 in a tube on a tube shaker.

In the first instance the dark reddish colour developed in the immobilised particles, subsequently also the supernatant slowly took up a reddish colour. The colour development was followed by eye and compared to a reference immobilised product which did not contain any P-lactamase activity. The velocity at which the reddish colour develops is a measure for

the amount of -lactamase activity present in the immobilised enzyme product.

The glutarylacylase activity was measured with adipyl-7- ADCA as substrate. 40 mg of wet weight immob were incubated with 20 ml of 100 mM adipyl-7-ADCA in 0.1 M Tris buffer at pH 8.0 in a shaking waterbath at 370C. After 10 and 30 minutes of incubation, a sample was withdrawn from the supernatant. The reaction was stopped by dilution with the eluent which is used in the RP-HPLC for the determination of the amount of 7-ADCA produced. A RP-HPLC C18 column was used. Detection of p-lactams occurred at 254 nm. The composition of the eluent used was 5.5 g NaH2PO4.H2O, 0.3 g sodium-dodecylsulphate and 180 ml acetonitrile per litre. The pH was corrected to pH 3.0 with 4M H3PO4. A standard with known 7-ADCA content was used for the calculation of the amount produced 7-ADCA. The activity was expressed as the amount of ymol 7-ADCA produced per minute, between 10 and 30 minutes incubation at 370C, per gram wet weight immobilised enzyme (AU/g ww).

The results are shown in table 3. It was observed that the reddish colour in sample 1 develops only slightly slower than in the untreated reference sample 5. In sample 5 the immobilised enzyme develops a dark red colour within only a few minutes after the start of the reaction. The velocity at which the colour develops in sample 2 and 3 was 2-3, respectively 10-20 times slower than in the untreated reference sample. At pH 5 the reduction of the rate at which the colour develops is very similar to that observed at pH 6 with 2 W glutardialdehyde, about 10-20 times. The glutardialdehyde treatment hardly had any effect on the glutarylacylase activity.

The conclusion is that after treatment with 0.5 - 2 W glutardialdehyde a considerable P-lactamase reduction has been obtained while the glutarylacylase activity has been maintained.

Table 3 Reduction of -lactamase in immobilised glutarylacylase Sample pH b glut pH after AU/g ww Reduction reaction I of - lactamase 1 6 0.1 6.8 93 slight 2 6 0.5 6.5 99 ~ 2 - 3 x 3 6 2 6.1 98 10 - 20 x 4 5 2 5.0 97 10 - 20 x 5 - 0 - 101 Reference Example 3 Inactivation of the P-lactamase activity in immobilised glutarylacylase and determination on CefG.

To 50 ml of settled washed immobilised glutarylacylase 100 ml water and 12 ml glutardialdehyde (25%) was added, resulting in + 2 W glutardialdehyde concentration. The mixture was incubated for 2 hours with stirring at room temperature.

Unbound glutardialdehyde was washed out and the glutarylacylase activity and the -lactamase activity were determined.

Immobilised glutarylacylase was contacted with CefG before and after treatment with glutardialdehyde. The degradation of the amount of CefG was determined with the same RPHPLC determination as used for the determination of the glutarylacylase activity on adipyl-7-ADCA.

0.5 gram of wet weight immob was contacted with 2 ml 0.23 M CefG in 0.1 M TRIS buffer pH 8.0 at 370C in a shaking waterbath. At several intervals between 10 and 360 minutes incubation samples of the supernatant were withdrawn and diluted in the HPLC eluent. Appropriate standards were used for the calculation of the concentration of CefG and 7-ADCA.

The -lactamase activity was calculated from the reduction of the total amount of intact -lactam in the assay, corrected for the degradation in absence of enzyme. The -lactamase activity is expressed in ymol CefG degradation between 10 and

360 minutes incubation, per minute and per gram of wet weight immobilised product.

The lactamase reduction was expressed as a percentage of the original activity (see table 4).

Table 4 Acylase act. P-lactamase % AU/g ww ymol/min/g -lactamase ww Blank substrate 0 0 untreated immob 75.3 0.24 (100) treated immob 75.5 0.020 8 After treatment with glutardialdehyde the lactamase activity has been reduced to 8 %, while no loss of glutarylacylase activity was observed.

In different experiments a maximal reduction in p-lactamase activity has been obtained of 20 - 30 times the original activity after treatment with 2 - 5 W glutardialdehyde and 2-6 hours reaction time.

Example 4 Reduction of the enzyme leakage from immob by glutardialdehyde treatment Enzyme leakage of an immobilised glutarylacylase product without glutardialdehyde treatment (glut -) was compared to the leakage of an immob treated with 3 W glutardialdehyde during 5.5 hours (glut +) as has been described in example 2.

Enzyme leakage was measured by incubation of 400 mg wet weight immob in 10 ml buffer, containing 0.05 M TRIS, 0.2 M NaCl pH 8.5, 0.04 W sodium azide as a preservative and 0.5 mg/ml bovine serum albumin as stabilizer for the enzyme, in a shaking water bath at 370C. The released glutarylacylase was measured in the supernatant with the glu-pNA method described in example 1 after various time intervals.

The amount of leakage of enzyme was expressed as the percentage of activity in the supernatant of the total measured activity in the immob added to the test.

The results shown in table 5 indicate that the leakage of activity has strongly been reduced by the glutardialdehyde treatment.

Table 5 Activity leakage from immob before and after treatment with glutardialdehyde Hours in test % leakage of activity - - glut + glut 1 6.8 0.24 5.5 8.7 0.5 22 10.9 0.9 30 12.2 1.1 46 13.2 1.3 53.5 13.5 1.4 70 14.3 1.4 In a comparable test the total amount of released protein, determined with the BCA method, executed according to the instructions of the supplier (Pierce), could be reduced from 48 Hg/ml (- glut) to 10 yg/ml or even lower (+ glut) after 4 hours incubation.

Example 5 Reduction of p-lactamase with formaldehyde To 8 gram wet weight of immobilised glutarylacylase, water was added to a total volume of 30 ml and 2.4 ml 37 W formaldehyde solution was added while shaking in a shaking water bath at room temperature. During incubation the pH was 5.6. After 4.5 hours incubation in a shaking water bath at room temperature the immob was washed several times with water

in order to wash out the unbound formaldehyde. The glutaryl- acylase activity was determined with adipyl-7-ADCA substrate as described in example 2. The P-lactamase assay was carried out with CefG substrate: 2 gram of wet weight immob, or a lower amount in case of high -lactamase activity, was incubated with 4 ml 40 mM CefG in 0.2 M tris buffer pH 8.0 at 370C in a shaking water bath. Detection of the reaction products was performed as described in example 3. -lactamase activity was expressed as the reduction of CefG due to -lactamase activity in ymol/min/g wet weight, under the reaction conditions used. The lactamase activity was compared to an untreated immobilised glutaryl acylase and expressed in the reduction factor (table 6).

Table 6 Effect of formaldehyde treatment on P-lactamase reduction Acylase P-lactamase Factor activity activity -lactamase U/g ww I ymol/min/g ww reduction Untreated 135 1.5 1 immob Formaldehyde 53 0.22 7 treated After treatment of the immob with formaldehyde the p-lactamase activity was reduced with a factor 7.

Example 6 Effect of increased pH during treatment of immob with glutardialdehyde In a slightly modified glutardialdehyde treatment procedure the pH was controlled during treatment. During glutardialdehyde treatment of the immob in an experiment without pH correction the pH was 5.5. In a parallel experiment the pH during glutardialdehyde treatment was increased by

adding sodium hydroxide 4 M, resulting in pH 7.8 measured after 15 minutes of incubation.

The incubation procedure and the determination of the glutaryl acylase and -lactamase activity were the same as described in example 5. The leakage of the acylase activity was determined in the treated and untreated immobs. 0.4 gram wet weight of immob was, after several washes with water, incubated with 10 ml of 50 mM tris and 200 mM NaCl pH 8.5 at 300C in a shaking water bath during 4 hours, in order to simulate bioconversion conditions. The released acylase activity was determined in the supernatant with Glu-pNA substrate as in example 1. The amount of leaked activity was expressed as the percentage of the total activity determined in the immob added to the test.

Table 7 Effect of pH during glutardialdehyde treatment at different pH Acylase -lactamase Factor Activity activity activity -lactamase leakage U/g ww ymol/min/g reduction ww Untreated 135 1.5 1 18 immob Glut ald. 133 0.08 19 0.47 treated pH 5.5 Glut ald. 121 0.05 30 0.51 treated pH 7.8 In the experiment at pH 7.8 the -lactamase activity was reduced with a factor 30 (table 7). It can be concluded that at increased pH an increased P-lactamase reduction can be obtained. At pH 5.5 as well as pH 7.8 a considerable reduction of activity leakage was obtained.