Microorganisms, in particular Streptomyces species, produce a number of antibiotics including clavulanic acid and other clavams, cephalosporins, cephamycins, tunicamycin, holomycin, polyketides and penicillins. There is considerable interest in being able to manipulate the absolute and relative amounts of these antibiotics produced and accordingly there have been a large number of studies investigating the metabolic and genetic mechanisms of these pathways (Domain, A. L. (1990)"Biosynthesis and regulation of beta-lactam antibiotics"in"50 years of Penicillin applications, history and trends"). Many of the enzymes which carry out the various steps in the metabolic pathways and the genes which code for these enzymes are known.

In the cephalosporin metabolic pathway in, for example, Streptomyces clavuligerus three important antibiotics can be produced namely penicillin N, 0- carbamoyldeacetylcephalosporin C and cephamycin C. These antibiotics are synthesised from the tripeptide precursor d- (L-a-aminoadipyl)-L-cysteinyl-D-valine ; ACV (J. F.

Martin et al (1990),"Clusters of genes involved in Penicillin and cephalosporin biosynthesis"in"50 years of Penicillin applications, history and trends").

The recognised first dedicated step in the biosynthesis of both penicillins and cephalosporins in S. clavuligerus is the conversion of L-lysine to L-s-aminoadipic acid catalysed by the enzyme lysine-s-amino transferase (LAT) which is encoded by the lat gene. The nucleotide sequence of the S. clavuligerus lat gene, and consequently the derived amino acid sequence of the encoded protein, are known (Tobin, M. B et al., (1991) J. Bacteriol, 173,6223-6229). It has been shown that disrupting or deleting the lat gene can result in increased production of the 5R clavam clavulanic acid (W096/41886, SmithKline Beecham pic)

Clavams can be arbitrarily divided into two groups dependent on their ring stereochemistry (5S and 5R clavams). The biochemical pathways for the biosynthesis of 5R and 5S clavams have not yet been fully elucidated but it is believed that they are derived from the same starter units; a 3 carbon compound (probably D-glyceraldehyde-3- phosphate (Khaleeli et al. (1999) J. Amer. Chem. Soc. 121 (39); 9223-9224) and arginine (Valentine, B. P. et al (1993) J. Am Chem. Soc. 115,1210-1211) and share some common intermediates (Iwata-Reuyl, D. and C. A. Townsend (1992) J. Am. Chem. Soc.

114: 2762-63, and Janc, J. W. etal (1993) Bioorg. Med. Chem. Lett. 3: 2313-16.

Examples of 5S clavams include clavam-2-carboxylate (C2C), 2- hydroxymethylclavam (2HMC), 2- (3-alanyl) clavam, valclavam and clavaminic acid (GB 1585661, Rohl, F. et al. Arch. Microbiol. 147: 315-320, US 4,202,819). There are, however, few examples of 5R clavams and by far the most well known is the Slactamase inhibitor clavulanic acid which is produced by the fermentation of S. clavuligerus.

Clavulanic acid, in the form of potassium clavulanate is combined with the Slactam amoxycillin in the antibiotic AUGMENTIN (Trade Mark SmithKline Beecham).

Because of this commercial interest, investigations into the understanding of clavam biosynthesis have concentrated on the biosynthesis of the 5R clavam, clavulanic acid, by S. clavuligerus. A number of enzymes and their genes associated with the biosynthesis of clavulanic acid have been identified and published. Examples of such publications include Hodgson, J. E. et al., Gene 166,49-55 (1995), Aidoo, K : Å. et al., Gene 147,41-46 (1994), Paradkar, A. S. et al., J. Bact. 177 (5), 1307-14 (1995).

It is known that clavaminic acid, a 5S clavam which is also a clavulanic acid precursor, is produced by the action of clavaminic acid synthase in the clavulanic acid biosynthetic pathway in S. clavuligerus. Gene cloning experiments have identified that S. clavuligerus contains two clavaminic acid synthase isoenzymes, casl and cas2 (Marsh, E. N. et al Biochemistry 31,12648-657, (1992)) both of which can contribute to clavulanic acid production under certain nutritional conditions (Paradkar, A. S. et al., J.

Bact. 177 (5), 1307-14 (1995)). Clavaminic acid synthase activity has also been detected in other clavulanic acid producing micro-organsims, ie. S. jumonjinensis (Vidal, C. M., ES 550549, (1987)) and S. katsurahamanus (Kitano, K. et al., JP 53-104796, (1978)) as well as S. antibioticos, a producer of the 5S clavam, valclavam (Baldwin, J. E. et al., Tetrahedron Letts. 35 (17), 2783-86, (1994)). The latter paper also reported S. antibioticos

to have proclavaminic acid amidino hydrolase activity, another enzyme known to be involved in clavulanic acid biosynthesis.

Recently genes which are specific for the biosynthesis of 5S clavams (ie. which are not essential for biosynthesis of 5R clavams) as exemplified by C2C and 2HMC in S. clavuligerus have been identified. For example Mosher, RH et al (1999) Antimicrobial Agents and Chemotherapy 43 (5) pl215-1224, disclose 3 open reading frames (ORFs) upstream of the casl gene (cvml, 2 and 3) and 3 ORFs downstream of casl (cvm4,5 and 6). Further it has been shown that by disrupting or deleting these 5S clavam specific genes it is possible to increase the production of 5R clavams, for example clavulanic acid, as well as reduce or eliminate the biosynthesis of 5S clavams such as C2C (W098/33896, SmithKline Beecham pic and the Governors of the University of Alberta).

There is a continuing need to improve organisms and processes for the production of 5R clavams, for example clavulanic acid. There is a particular need for organisms and processes that can provide quantitative and qualitative improvements in 5R clavam yield.

Quantitative means, essentially, a higher yield. Qualitative means, essentially, a 5R clavam product which has significantly reduced, or even undetectable, levels of unwanted 5S clavam or clavams.

Thus in one aspect the invention relates to a 5R clavam producing organism comprising: a) a cephalosporin biosynthetic pathway, or a part thereof, wherein the lat gene has been deleted, disrupted or otherwise made defective; and b) a clavam biosynthetic pathway, wherein one or more genes that are specific for 5S clavam biosynthesis have been deleted, disrupted or otherwise made defective.

In a preferred embodiment the organism is a Streptomycete, preferably S. clavuligerus. In a further preferred embodiment the gene specific for 5S clavam biosynthesis which is disrupted or otherwise made defective is selected from the group consisting of cvml, cvm2, cvm3, cvm4, cvm5 and cvm6, preferably cvml, cvm4 or cvm5, most preferably cvml. Preferably the 5R clavam has Flactamase inhibitory activity and most preferably the 5R clavam is clavulanic acid.

As used herein the term"deleted"means a total or partial deletion of the gene. A partial deletion can involve the removal of any amount of DNA from the target gene, from 1 base pair (bp) up to almost the entire polypeptide coding region of the gene. A

total deletion involves the removal of the entire coding region of the gene with or without flanking sequences, which may or may not include regulatory elements that are required for gene function, for example transcriptional promoters. Furthermore the deletion may result in the removal of just a regulatory region, such as a promoter, leaving the coding region intact. The result is that no mRNA can be produced and so the gene is rendered defective.

With regard to partial deletions the removal of just one or two base pairs from the coding region of the gene results in a frameshift mutation and consequently a truncated gene product. Preferably the truncated gene product will have no residual activity that may be possessed by the complete polypeptide. In this respect, deletions closer to the 5' end of the coding region are preferred.

As used herein the term"otherwise made defective"is understood to include, but is not limited to, such methods as inserting a strong transcriptional terminator downstream from the promoter to prevent transcription of the gene or part of the gene; the use of antisense (eg. O'Connor, J Neurochem (1991) 56: 560) or ribozymes (eg Usman, N, et al., Curr. Opin. Struct. Biol (1996) 6 (4), 527-33) to block expression at the post- transcriptional level by binding to or cleaving the mRNA respectively; or inhibiting transcription by using triplex-forming oligonucleotides to bind to the specific DNA sequence of the gene to be transcriptionally silence (eg. Dervan et al., Science (1991) 251 : 1360). The skilled man would be able perfom these aspects of the invention using methods that are well known in the art. Other methods for rendering the target gene defective are understood to fall within this definition.

In a further aspect the invention relates to a process for producing a 5R clavam, comprising growing the organism of the invention under fermentation conditions that facilitates the biosynthesis of the 5R clavam.

In one preferred aspect, the invention relates to the 5R clavam, preferably clavulanic acid, produced using the process of the invention.

Methods for preparing the organisms of the invention are well known in the art.

The starting point for preparing the organisms of the invention is a"parent"organism.

The parent organism may be a Streptomycete, for example a strain of S. clavuligerus (available for example from the American Type Culture Collection-ATCC 27064) or any modified strain derived therefrom. Such parent strains include those which have been

engineered, or selected by established strain improvement techniques, to produce higher levels of clavulanic acid.

Such parent strains comprise a cephalosporin biosynthetic pathway, or a part thereof. A part of a biosynthetic pathway is where one or more genes encoding one or more enzymes in that biosynthetic pathway are absent, for example by deletion, or have been otherwise made defective. The absence of a biosynthetic pathway gene could be a consequence of a natural event, or could be engineered using methods well known in the art (Sambrook, J., Fritsch, E. F. and Maniais, T. (1989) Molecular Cloning A Laboratory Manual (2nd Edition)).

The parent strain also comprises a pathway for the biosynthesis of a 5R clavam, most preferably clavulanic acid.

The lat gene and 5S clavam specific genes of the parent organism may be deleted, disrupted or otherwise made defective using well known methods (for example the methods disclosed in Sambrook, et al (1989) supra).

The genes may be disrupted using random mutagenesis methods, for example exposure to LEZ radiation or mutagenic chemicals (ethidium bromide or NTG), followed by selection of mutants disrupted in the lat or 5S clavam specific genes. Alternatively, since the DNA sequence of the lat and 5S clavam specific genes of S. clavuligerus are known (Tobin, M. B et al., (1991) supra ; Mosher, RH et al (1999) supra) the genes, or parts of the genes, may be deleted or otherwise mutated using standard in-vitro mutagenesis techniques. Thus, for example, the lat or 5S clavam specific gene or genes can be cloned, by PCR or library screening techniques, and mutated in-vitro (for example by single or multiple nucleotide deletion, substitution or insertion), and then reintroduced into a parent organism by replacing the existing wild-type gene. Methods of gene disruption are preferred (for example by inactivation by insertion of an apramycin resistance gene (Paradkar and Jensen (1995) J. Bacteriol. 177 (5) 1307-1314) and methods of gene deletion are particularly preferred. Gene replacement in vivo may be facilitated by, for example, double cross-over events using methods well known in the art (see for example Aidoo, KA et al (1994) Gene 147 p41-46). The cvml gene may be disrupted by insertional inactivation using an apramycin resistance gene, as described in W098/33896, or more preferably by deletion of the entire, or part of, the cvml gene. Most preferably a 654bp portion of the cvm 1 gene, between two AatII restriction sites, is deleted. The lat

gene may be disrupted by insertional mutagenesis, as described in W096/41886, wherein recombinant DNA techniques were used to introduce an apramycin resistance gene within the coding region of the lat gene. Preferably the lat gene, or a part thereof, is deleted. The organism of the invention can be constructed by deleting, disrupting or otherwise making defective the lat gene in a first step, followed by a second step wherby the cvml gene is deleted, disrupted or otherwise made defective. Alternatively the cvm I gene may be deleted or disrupted in the first step followed by the deletion or disruption of the lat gene in a second step. Preferably the cvm 1 gene is deleted or disrupted in the first step.

In the process of the invention, the organism of the invention is used to produce a 5R clavam with reduced amounts, or undetectable levels of 5S clavam. In a preferred embodiment the 5R clavam is clavulanic acid, preferably increased levels of 5R clavam are produced compared with the parent strain.

The processes of the invention may be carried out in flasks at the laboratory scale, in fermentors at pilot plant scale, or in fermentors at production scale.

Examples Example 1-Construction of a strain of S. clavuligerus containing gene disruptions of the lat and cvml genes To produce a strain of S. clavuligerus containing genes disruptions in both the lat and cvml genes a sequential strategy was adopted in which the parental S. clavuligerus strain was initially disrupted in the cvml gene then subsequently disrupted in the lat gene.

The basic method used for these gene disruptions are as described by Paradkar and Jensen (1995) using the plasmids pCEC061 (as described in W098/33896, SmithKline Beecham plc and the Governors of the University of Alberta) and p4861atap (as described in W096/41886, SmithKline Beecham pic and the Governors of the University of Alberta).

Disruption of the cvml gene.

The plasmid pCEC061, was prepared from a wild type strain of S. clavuligerus and used to transform S. clavuligerus strain SB 203 (a high titre variant of the S. clavuligerus strain ATCC 27064) using methods described in WO 94/18326 (except that in addition to thiostrepton the overpour used to select the transformants also contained apramycin (20ug/ml final concentration). The transformants which were found to be resistant to both thiostrepton and apramycin were then subclutured through two successive rounds of sporulation on solid agar media which had not been supplemented with either thiostrepton or apramycin (as described in W098/33896). Individual colonies were then replica plated onto antibiotic containing media to identify apramycin-resistant and thiostrepton-sensitive transformants. From this process 84 putative mutants were chosen for further analysis.

Because these isolates were apramycin resistant and thiostrepton sensitive it was assumed that the plasmid had been lost and a double crossover had occurred between the pCEC061 and the S. clavuligerus chromosome, creating a copy of the cvml gene in the chromosome in which the apramycin resistance gene had been inserted. To confirm that the cvm 1 gene had been inactivated the strains were also fermented and analysed for the production of clavam-2-carboxylate and 2 hydroxymethyl clavam. This was achieved by growing these strains in shake flask fermentations (as described in WO 94/18326) and using HPLC analysis (as as described in W098/33896) to determine the presence or absence of detectable levels of clavam-2-carboxylate and 2 hydroxymethyl clavam. The

HPLC anaysis failed to detect either clavam-2-carboxylate or 2 hydroxymethyl clavam.

One of these strains GEM1801 was then selected to for the next step; the inactivation of the LAT gene by gene disruption.

Disruption of lat in GEM 1801.

The plasmid p4861atap, was prepared from a wild type strain ofS. clavuligerus and used to transform S. clavuligerus strain GEM1801 as described in WO 94/18326.

Since GEM 1801 already contained the apramycin gene inserted into its chromosome, transformants were selected for thiostrepton resistance only. The transformants which were found to be resistant to thiostrepton and were then subclutured through four successive rounds of sporulation on solid agar media which had not been supplemented with either thiostrepton or apramycin (as described in the previous example). Individual colonies were then replica plated to antibiotic containing media to identify strains which were apramycin-resistant and thiostrepton-sensitive. From this process 9 putative mutants were chosen for further analysis.

Because these isolates were apramycin resistant and thiostrepton sensitive it was assumed that the plasmid had been lost. Since GEM1801 was already apramycin resistance, in order to identify isolates where a double crossover event had occurred between the p4861atap and the S. clavuligerus chromosome, thereby creating a copy of the lat gene in the chromosome in which the apramycin resistance gene had been inserted, it was necessary to test thiostrepton sensitive strains for the production of products of the cephamycin C/cephalosporin biosynthetic pathway. This was achieved by fermenting the isolates in shake flask fermentations as described in the previous example and analysing the resultant broth by HPLC as described in WO 94/18326. From this analysis several strains (GEM 1894,1895,1896,1897,1898 18100, and 18101) were identified for further analysis which were unable to produce any products of the cephamycin C/cephalosporin biosynthetic pathway. This indicated that that the lat gene had been inactivated.

Example 2-Clavulanic acid productivity of the cvml/lat double disrupted strains Seven strains identified from example 1 as being inactivated in both the lat and cvml genes (strains GEM 1894, 1895,1896,1897,1898 18100, and 18101) were analysed in shake flask fermentations for clavulanic acid productivity. The shake flask

fermentations and assays for clavulanic acid productivity were carried out as in WO 94/18326. From the results shown in table 1 it is clear that the strains containing both of these mutations have increased clavulanic acid productivity compared to parental control.

The improvements range from 6% to 16% above the productivity of the parental strain.

Table 1 Culture Clavulanic acid productivity (% of control-SB 203) GEM1894109 GEM1895111 GEM 1896 106 GEM1897115 GEM 1898 111 GEM18100 114 GEM18101 116 Example 3-Construction of a strain of S. clavuligerus containing gene deletions of the cvml and lat genes To produce a strain of S. clavuligerus containing gene deletions in both the lat and cvm I genes a sequential strategy was adopted in which the parental S. clavuligerus strain (SB203) was initially deleted in the cvml gene then subsequently deleted in the lat gene.

Deletion of the cvml gene.

S. clavuligerus genomic PCR products were generated using the oligonucleotide primers listed below and pCEC061 (as described in W098/33896 supra) as the template.

The original nucleotide sequence was altered to incorporate a PstI site into oligonucleotide 1 and a SphI site in oligonucleotide 4.

Oligonucleotide Pair 1 used to generate PCR product 1: Primer 1 5'dCTGACGCTGCAGGAGGAAGTCCCGC 3'-SEQ ID NO: 1 Primer 2 5'dCGGGGCGAGGACGTCGTCCCGATCC 3'-SEQ ID NO : 2

Oligonucleotide Pair 2 used to generate PCR product 2: Primer 3 5'dGAGCCCCTGGACGTCGGCGGTGTCC 3'-SEQ ID N0 : 3 Primer 4 5'dGACGGTGCATGCTCAGCAGGGAGCG 3'-SEQ ID NO : 4 Standard PCR reactions were carried out using the PTC-200 Peltier thermal cycler from GRI (Felsted, Dunmow, Essex, CM6 3LD). PCR product 1 was generated using primers I and 2. This product is approximately I Kb and contains the 3'region of cvml from the second AatII site and downstream regions. PCR product 2 was generated using primers 3 and 4. This product is approximately 1. lKb and contains the 5'region of cvml from the first AatII site and upstream regions. PCR product 2 was ligated to SrfI digested pCR- Script Amp SK (+) as per instruction manual (Strategene Ltd, Cambridge Science Park, Milton Road, Cambridge CB4 4GF). The ligation mixture was used to transform Epicurian E. coli XL1-Blue MRF'Kan supercompetent cells (available from Strategene)- to ampicillin resistance (as per manufacturers instructions). Plasmid DNA was isolated from the resulting transformants and DNA restriction analysis using Salll HindIII and BamHIlNotI revealed that 7 clones contained plasmid into which PCR product 2 had been ligated. Isolate 8 contained the insert in the correct orientation and was designated pSB 1.

PCR product 1 was digested with PstI and AatII and the resultant DNA fractionated by agarose gel electrophoresis. The 1Kb fragment was excised and eluted from the gel using the Sephaglas band prep kit. (Pharmacia, St. Albans, Herts, ALI 3AW). The isolated fragment was then ligated to AatII and PstI digested pSB 1. The ligation mixture was used to transform competent cells ofE. coli XLl-Blue (available from Strategene) to ampicillin resistance (as per manufacturers instructions). Plasmid DNA was isolated from the resulting transformants and restriction analysis using NotIlEcoRI, SrfI, PstI and SalI revealed that 1 clone contained plasmid into which PCR product 1 had been ligated. This plasmid which now contained a cloned copy of the cvml gene with a 654 nucleotide deletion was designated pSB2.

To introduce pSB2 into S. clavuligerus the plasmid was further modified to contain an origin of replication that could function in streptomyces. To achieve this the plasmid DNA of pSB2 was digested with EcoRI and HindIII and ligated into the Streptomyces vector pIJ486 which encodes a thiostrepton resistance gene (6.2Kb: Ward et al., (1986) Mol. Gen. Genet. 203: 468-478) and was also digested with EcoRI and HindIII.

The ligation mixture was used to transform E. coli (JM109) competent cells (Strategene Ltd) to ampicillin resistance. Plasmid DNA was isolated from the resulting transformants and restriction analysis using EcoRI, SalI and BglII revealed 6 clones possessed pSB2 containing pIJ486. One of these plasmids was designated pSB3.

The plasmid pSB3 was used to transform a derivative of the S. clavuligerus strain SB203 in which the cvml gene had already been disrupted by insertion of the apramycin resistance gene (as described in W098/33896 supra). Thiostrepton resistant transformants were selected and these transformants were then put through 3 rounds of sporulation on non-selective media and screened for the loss of apramycin and thiostrepton resistance.

This was carried out to identify those transformants where homologous recombination had taken place and an"exchange"had occurred between the apramycin disrupted chromosomal copy of cvml and the deleted copy of cvm I on the plasmid. From this process 45 isolates were identified which had lost apramycin resistance. These isolates were then subjected to further biochemical and genetic tests to confirm that the apramycin disrupted cvml gene had been deleted as expected. From this analysis several strains were identified which now contained a simple deletion in the clavam gene cvml and one of these strains, SB82, was chosen as a parent strain to delete the lat gene.

Deletion of the lat gene in SB82.

The DNA sequence of the lat gene and its surrounding DNA is known (Tobin et al J. Bact 173,6223-9 (1991) and Perez-llarena et al J. Bact 2053-9 1997). A PCR strategy was used to create an approximately lkb deletion of the lat gene as discribed below.

PCR was used to amplify the 3'region of the open reading frame (orfl2) adjacent to the lat gene using the primers: FZ1 [5'-dGTTAAAGCTTCCGATACAGACAAGGGGTTCCTTCAC-3']-SEQ ID NO : 5 ; and RZ I [5'-dATTAGGATCCCGAGCTGTTGCTCAGCTTCGTGTTCA-3']-SEQ ID NO : 6 PCR was further used to amplify the 3'region of the lat gene from S. clavuligerus chromosomal DNA using the following oligonucleotide primers

FLI [5'-dATATGAATTCTGCAGCGGCTCTGCCACGAGAACG-3']-SEQ ID NO : 7 and; RL1 [5'-dATATGGATCCATCGCAACGCTCGACACTCCTTGGAATG-3']-SEQ ID NO : 8.

The purified PCR products of the orfl2 and lat gene fragments were then individually cloned into pCRscript using the Stratagene protocol before transforming into competent XLlBlue MRF cells. Transformants were screened on plates containing Luria agar containing 100 ug/ml ampicillin, 0.5 mM ITPG and 40 ug/ml Xgal.

Plasmids were prepared from each transformation, pCR-lat for the lat fragment and pCR-Z for the orfl2 fragment were digested with EcoRI. This linearised the plasmid pCR-Z but in the case of pCR-lat two DNA fragments were produced one of which contained the 3'portion of the lat gene. The digests were electrophoresed in low melting point agarose and DNA fragments recovered as Sambrook et al. (1989) supra.

The gel purified pCR-Z EcoRI cut vector was ligated to the gel purified EcoRl DNA fragment from pCR-lat containing the 3'portion of the lat gene. The ligation mix was then used to transform competent JM109 cells using the CaClz method as Sambrook et al.

1989. Transformants were selected on Luria agar containing 100 ug/ml ampicillin.

The resultant plasmid, pSB13, was then further modified to remove a small amount of sequence (111 nucleotides) upstream of the deleted copy of the lat gene. This was achieved by digesting pSB13 with the DNA restriction enzyme NotI and religating the plasmid together. The ligation mix was used to transform competent JM109 cells using the CaC12 method as Sambrook et al. (1989) supra. Transformants were selected on Luria agar plus 100 ug/ml ampicillin.

A Streptomyces replicon and selectable marker was then introduced into the vector. This plasmid was named pSB486. This plasmid pSB486 was then introduced into SB82 by transformation. From the transformation, 88 thiostrepton resistant transformants were selected and tested for cephalosporoate production. Several of these primary transformants were found to have reduced cephalosporoate productivity compared to SB82. These primary transformants were then put through a further round of sporulation on non-selective media and screened for the loss of thiostrepton resistance. This was carried out to identify those transformants where the plasmid had been lost. In some cases homologous recombination had taken place and an"exchange"had occurred between the

chromosomal copy of the lat gene and the deleted copy of lat gene on the plasmid. These thiostrepton sensitive strains were then tested for cephalosporoate productivity to identify those where the"exchange"had taken place. From this process primary transformant produced 90 thiostrepton sensitive isolates which on further testing were found to be unable to produce cephalosporoate. These isolates were then subjected to further biochemical and genetic tests to confirm that the lat gene had been deleted as expected.

These isolates were subsequently confirmed to have both the cvml and lat genes deleted.

Example 4-Clavulanic acid productivity of the cvml/LAT double deleted strains Five strains identified from example 3 as being inactivated in both the lat and cvml genes as described above (strains GEM1850, 1851,1855,1856, and 1860) were analysed in shake flask fermentations for clavulanic acid productivity. The shake flask fermentations and assays for clavulanic acid productivity were carried out as in WO 94/18326. From the results shown in table 1 it is clear that the strains containing both of these mutations have increased clavulanic acid productivity compared to parental control. The improvements range from 9% to 19% above the productivity of the parental strain.

Table 2 Culture Clavulanic acid productivity (% of control-SB 203) GEM1850 110 GEM1851 109 GEM1855 117 GEM1856 119 GEM1860 109