IE950412A1 - Novel polypeptides - Google Patents

Abstract

The present invention relates to variants of Penicillin Binding Proteins (PBP), which proteins are involved in bacterial peptidoglycan biosynthesis. Disclosed are also DNA molecules coding for the said PBP variants, as well as vectors and cells harbouring such DNA molecules. The invention is also related to processes for assaying and designing therapeutically useful compounds which have high affinity to PBP, which processes utilize the said PBP variants.

Description

TECHNICAL FIELD The present invention relates to variants of Penicillin Binding Proteins (PBP), which proteins are involved in bacterial peptidoglycan biosynthesis. Disclosed are also DNA molecules coding for the said PBP variants, as well as vectors and cells harbouring such DNA molecules. The invention is also related to processes for assaying and designing therapeutically useful compounds which have high affinity to PBP, which processes utilize the said PBP variants.

BACKGROUND ART Bacteria and most other unicellular organisms possess a cell wall, which comprises a cross-linked polysaccharide-peptide complex called peptidoglycan. Peptidoglycan biosynthesis consists of three stages: (1) synthesis of precursors (sugar nucleotides) in the cytosol, (2) precursor transfer across the membrane and formation of the polysaccharide chain, and (3) cross-linking of individual peptidoglycan strands in the cell wall.

In the latter stage of peptidoglycan biosynthesis, new bonds must be made between nascent glycan strands and existing peptidoglycan. The newly synthesized chains are about 10 disaccharides long and are extended by transglycosylase enzymes to a final glycan strand of between 100 and 150 disaccharide units. The peptidoglycan is crosslinked by the action of transpeptidases which link the terminal D-ala of one glycan strand to a free ε-amino group on a diaminopimelic acid residue on an adjacent region. «τ q juA-s CtZN ts/5£. β/4τ xv/ce OPEN TO PUBLIC INSPECTION UNDER SECTION 28 AND RULE 23 JNL. No £444 OF A number of antibiotics inhibit bacterial growth by interfering with the formation of the peptidoglycan layer. The cross-linking reaction is the 0 4 1 2 --2target for action of two important classes of such antibiotics, the penicillins and the cephalosporins. Penicillin is thought to react irreversibly with the transpeptidase that catalyses cross-linking.

The penicillin interactive proteins fall into three groups: the β-lactamases, the Low Molecular Weight-Penicillin Binding Proteins (PBPs), which mainly include the carboxypeptidases, and the High Molecular WeightPenicillin Binding Proteins. Penicillin Binding Proteins are those enzymes which have been shown to bind radiolabelled penicillin G. In Escherichia coli such proteins are called e.g. PBP IA and PBP IB, both belonging to the class High Molecular Weight-PBPs. PBP IA and IB, which are known to be membrane bound proteins, maintain cell integrity and control peptidoglycan side wall extension during growth. Inactivation of either PBP IA or PBP IB can be tolerated by the bacteria while the deletion of both the genes, designated ponA and ponB, is lethal (Yousif et al., 1985).

PBP IB is known to be a bifunctional enzyme possessing both transpeptidase and transglycosylase activity (Ishino et al., 1980). PBP IA is believed to be bifunctional since it can substitute for PBP IB. The β-lactam antibiotics, such as penicillin, inhibit only the transpeptidase activity of these proteins.

The transglycosylase reaction is inhibited by e.g. moenomycin, which is a phosphoglycolipid used as a growth promoter in animal nutrition and which has been shown to possess broad spectrum bactericidal activity. The enzyme transglycosylase has been shown to be present in Escherichia coli, Staphylococcus aureus, Bacillus megaterium and Bacillus subtilis. This suggests that interference of peptidoglycan biosynthesis by inhibition of transglycosylase could be a lethal event in all clinically important pathogens. -3-950412 ι - - — ί* The putative transglycosylase domain of PBP IB has been assigned to the N-terminal 478 amino acids (Nakagawa et al.z 1987). This regions includes three conserved stretches of amino acids between the N-terminal half of both PBP 1A and IB and could represent residues involved in the transglycosylase activity.

Preparation of Penicillin Binding Protein 2A from Staphylococcus aureus is disclosed in EP-A-0505151.

DISCLOSURE OF THE INVENTION There is a growing number of reports of bacteria which are resistant to antibiotics. There is consequently a need for new compounds which inhibit bacterial growth by means of binding Penicillin Binding Proteins. The present invention provides PBP variants which facilitate processes for assaying and designing therapeutically useful compounds which have high affinity to PBPs.

Accordingly, it is an object of the invention to provide polypeptides which are water-soluble active derivatives of bacterial bifunctional Penicillin Binding Proteins, said Penicillin Binding Proteins being bound to the cell membrane when expressed in a bacterial cell and being capable of exhibiting both transglycosylase and transpeptidase activities and said derivatives lacking a membrane anchoring sequence but retaining the capability to exhibit one or both of said enzymic activities. The bacterial cell mentioned above is preferably an Escherichia coli cell or a Streptococcus pneumoniae cell.

The soluble PBP variants according to the invention retains transglycosylase activity, indicating that soluble variants of PBP, devoid of membrane anchoring sequences, can recognize lipid linked substrate and polymerise the disaccharide into repeating units. It can thus be assumed 9504121 -4— that other analogues of PBP lacking residues involved in membrane attachment would be enzymatically functional.

Molecules interacting with the penicillin interactive region of soluble PBP variants could be assumed to be capable of interacting identically with wild-type PBPs. Consequently the soluble PBP variants according to the invention can be used for identifying compounds which are interacting with wild-type Penicillin Binding Proteins.

It is furthermore well known that membrane-bound proteins are very difficult to crystallize. The soluble enzymatically active PBP variants can be used for crystallisation and will thereby facilitate a rational design, based on X-ray crystallography, of therapeutic compounds inhibiting High Molecular Weight-PBPs.

A further object of the invention is to provide polypeptides which are truncated water-soluble derivatives of bacterial bifunctional Penicillin Binding Proteins, said Penicillin Binding Proteins being bound to the cell membrane when expressed in a bacterial cell and being capable of exhibiting both transglycosylase and transpeptidase activities and said derivatives lacking the membrane anchoring sequence but retaining the capability to exhibit the transglycosylase activity. The bacterial cell mentioned above is preferably an Escherichia coli cell.

Alignment of amino acid sequences of High Molecular Weight-Penicillin Binding Proteins, and the compilation of the motifs involved in the penicillin binding of β-lactamases and carboxypeptidase, have suggested the C-terminal half of PBP 1A and IB to be the functional domain of the transpeptidase activity and includes the penicillin binding domain. In addition, Nakagawa et al. (1987) showed that a truncated ponB gene encoding the N-terminal 478 amino adds of PBP IB is capable of the transglycosylase reaction. -950412 ι -5On the basis of these findings, it has been suggested that the high molecular weight PBP IA and IB proteins are two domain-proteins, with the N-terminal half forming the transglycosylase domain and the C-terminal half the transpeptidase domain. The two domains have been predicted by computer analysis to be joined by a linker or hinge region which does not structurally or enzymatically contribute to the function of the protein. The linker region of E.coli PBP IB has been predicted to be from position 545-559 while that for E.coli PBP IA around position 501.

The monofunctional truncated variants of PBP according to the invention will, when used in x-ray crystallography, facilitate obtaining structural information of the transglycosylase domain of penicillin binding proteins. In addition, the reduced size of the monofunctional variant will facilitate crystallization.

In a preferred form, a water-soluble polypeptide according to the invention has an amino acid sequence which is identical to SEQ ID NO: 2, 4, 6, 12 or 13 in the Sequence Listing.

The observation that deletion of the ponA and ponB genes is lethal (Yousif et al., 1985) does not address the question of essentiality of the transglycosylase activity of the encoded PBP IA proteins, since the deletion results in the loss of both transglycosylation and transpeptidation activities. In addition, this experiment does not address the possibility that the transglycosylase enzyme activity can be contributed by a Penicillin Binding Protein other than PBP IA or PBP IB. It is also possible that hitherto undescribed Penicillin Binding Proteins and/or other proteins that contribute to the transglycosylase activity exist.

Alignment of the amino acids forming the putative transglycosylase domain of PBP IA and IB reveals three stretches of 9 out of 12 (Region 1), 9/10 (Region 2) and 8/10 (Region 3) amino acids identical within the 9504121 i N-terminal half of these two proteins (Broome-Smith et al., 1985) (Fig. 14). The same 3 regions are identically conserved among two other recently described protein sequences; Streptococcus pneumoniae PBP IA (Martin et al., 1992) and a 94 kDa protein from Haemophilus influenzae (Tomb et al., 1991). The conservation of these residues in such diverse species suggests their critical requirement in either maintaining structural aspects of the protein, or in the transglycosylation reaction itself.

The overlapping functional transglycosylase and transpeptidase activities of PBP IA and IB also suggests conservation of the catalytic centres and that molecules designed to interact with the catalytic centre of PBP IA would be reactive also with PBP IB.

The functional transglycosylase activity of the expressed protein can be studied either in a direct in vitro assay using appropriate substrates, or in an assay measuring the ability of the protein to complement the deletion of the corresponding genes in the chromosome. It has been shown that a plasmid with a gene encoding the wild type product (PBP IA or PBP IB) is capable of maintaining the viability of the E.coli cell (Yousif et al., 1985). This trans-complementation technique can be utilized to assess the functional nature of the mutant gene(s) encoding PBPs with mutations inactivating one of the enzymic (transglycosylation or transpeptidation) functions. The ability of such mutant products to complement in trans the deletion of the chromosomal ponA and ponB genes would define the essential requirement of the individual enzymic functions.

There is consequently a need for research tools which will make it possible to study the effects of specific inactivation of the transglycosylase activity of Penicillin Binding Proteins.

Consequently, a further aspect of the invention is a polypeptide which is a transglycosylase deficient derivative of a bacterial bifunctional penicillin 0 4 1 2^ —7— binding protein, said penicillin binding protein being bound to the cell membrane when expressed in a bacterial cell and being capable of exhibiting both transglycosylase and transpeptidase activities and said derivative lacking the capability to exhibit transglycosylase activity but retaining the capability to exhibit transpeptidase activity. The bacterial cell mentioned above is preferably an Escherichia coli cell.

The transglycosylase deficient PBP variants can advantageously be used in X-ray crystallography for the purpose of obtaining structural information of the activity sites of PBPs. Structural analysis of crystal form of soluble transglycosylase deficient PBP variants could allow delineation of the catalytic region and facilitate the design of molecules capable of specifically inhibiting the transglycosylase activity.

In a preferred form, the transglycosylase deficient polypeptide according to invention is a polypeptide which is lacking transglycosylase activity because of a mutation or deletion in the second conserved region of the gene coding for said polypeptide.

In a further preferred form, the transglycosylase deficient polypeptide according to the invention has an amino acid sequence which is identical to SEQ ID NO: 7, 8, 9, or 10 in the Sequence Listing.

The conventional purification procedure employed for the enrichment of penicillin binding proteins has been the use of a penicillin affinity. The binding of the protein to penicillin is covalent and requires harsh conditions to elute the bound protein. This may lead to a certain degree of inactivation of the enzymic activity of the protein. There is consequently a need for alternate affinity matrices for the efficient purification of the proteins. 0 4 1 21 Included in the invention is consequently a polypeptide comprising (a) a first polypeptide which is a PBP variant according to the invention; and (b) an additional polypeptide which allows binding to an affinity matrix; there being a cleavage site between said polypeptides.

The additional polypeptide mentioned above can preferably be glutathione-S-transferase or a polypeptide substantially similar to glutathione-S-transferase. Such an additional polypeptide will enable rapid purification of the protein using Glutathione Sepharose affinity matrix. In another preferred form, the additional polypeptide is a polypeptide rich in histidine residues, which residues will confer on the protein the ability to bind to an Ni affinity column. The additional polypeptide can be fused either to the N-terminus or the C-terminus of the soluble/membrane bound PBP.

The ability of the fusion proteins to bind to an affinity matrix allows immobilisation of the protein. Such immobilised proteins can be used for analysis of competitive binding of different ligands to the bound active protein, and thus for screening of compounds binding to the enzymic domain of interest.

The polypeptides according to the invention are not to be limited strictly to any one of the sequences shown in the Sequence Listing. Rather the invention encompasses polypeptides carrying modifications like substitutions, small deletions, insertions or inversions, which polypeptides nevertheless have substantially the biochemical activities of the PBP variants which amino acid sequence is disclosed in the Sequence Listing. Included in the invention are consequently also polypeptides, the amino acid sequence of which is at least 90% homologous, preferably at least 95% homologous, with the amino add sequence of any of the PBP variants according to the invention. 950412; 9A further object of the invention is to provide isolated and purified DNA molecules which have nucleotide sequences coding for any one of the PBP variants according to the invention.

In a preferred form of the invention, the said DNA molecules have nucleotide sequences identical to SEQ ID NO: 1, 3 or 5 in the Sequence Listing. However, the DNA molecules according to the invention are not to be limited strictly to any of the sequences shown in the Sequence Listing. Rather the invention encompasses DNA molecules carrying modifications like substitutions, small deletions, insertions or inversions, which nevertheless encode proteins having substantially the biochemical activities of the PBP variants according to the invention.

Included in the invention is also a DNA molecule which nucleotide sequence is degenerate, because of the genetic code, to the said nucleotide sequence coding for a PBP variant according to the invention. The natural degeneracy of the genetic code is well known in the art. It will thus be appreciated that the DNA sequences shown in the Sequence Listing are only examples within a large but definite group of DNA sequences which will encode the PBP variants which amino acid sequences are shown in the Sequence Listing.

A further aspect of the invention is a replicable expression vector which carries and is capable of mediating the expression of a DNA molecule according to the invention. In the present context the term replicable means that the vector is able to replicate in a given type of host cell into which is has been introduced. Examples of vectors are viruses such as bacteriophages, cosmids, plasmids and other recombination vectors.

Nucleic add molecules are inserted into vector genomes by methods well known in the art. A vector according to the invention can preferably be one of the plasmids listed in Table 1 below. 950412; -10Included in the invention is also a host cell harbouring a vector according to the invention. Such a host cell can be a prokaryotic cell, a unicellular eukaryotic cell or a cell derived from a multicellular organism. The host cell can thus e.g. be a bacterial, yeast or mammalian cell. The methods employed to effect introduction of the vector into the host cell are wellknown to a person familiar with recombinant DNA methods.

A further aspect of the invention is a process for production of a polypeptide which is a derivative of penicillin binding protein, comprising growing a host cell according to the invention in or on a culture medium for expression of the polypeptide and optionally recovering the polypeptide. An appropriate host cell may be any of the cell types mentioned above, and the medium used to grow the cells may be any conventional medium suitable for the purpose.

The High Molecular Weight-Penicillin Binding Proteins have been shown to be anchored to the membrane, but the majority of the protein is within the periplasmic space of the cell (Edelman et al. 1987). Thus PBP derivatives, devoid of the membrane signal / anchoring sequences, are forced to fold into their native state in a heterologous environment, namely the cytosol. This often leads to misfolding, and the majority of the expressed protein aggregates into an inactive form referred to as inclusion bodies.

It has now surprisingly been found that high yields of an active watersoluble PBP variant can be obtained by regulated transcription of the gene encoding the said PBP variant. Such regulated transcription involves (i) using a suboptimal concentration of the inducer isopropyl thiogalactoside (IPTG); and (ii) culturing the cells expressing the PBP variant at reduced temperature. A cumulative effect of these factors contributes to the overall recovery of the active soluble protein. Consequently, lower rates of expression, achieved through the mentioned combination of (i) sub-optimal -119 5 0 4 1 2 j de-repression of promoter systems and (ii) increased generation time by lowering of the temperature of cultivation, will enhance the solubility of proteins lacking the membrane anchoring segment.

A further important aspect of the invention is a process for the production of a water soluble polypeptide according to the invention which comprises culturing Escherichia coli cells harbouring an expression vector wherein a DNA coding sequence for said polypeptide is under the control of an isopropyl thiogalactoside (IPTG) - inducible promoter, said culturing being carried out in the presence of a sub-optimal concentration of IPTG for induction of the said promoter and at a temperature in the range of 20 to 24°C, preferably 22°C. The concentration of IPTG can preferably be approximately 0.01 mM.

In the case of expression of ponAdel23, a gene encoding a PBP variant according to the invention, such regulated transcription by (i) controlled de-repression of the T7 promoter by using sub-optimal concentration of the inducer IPTG and (ii) reducing the growth rate by culturing at 22°C, resulted in yields of the active protein which reached nearly 50% of the total induced protein of interest. The growth and induction conditions were critical for the efficient recovery of the soluble protein, as growth at higher temperatures or induction with higher concentrations of IPTG resulted in the majority of the protein becoming inactive and forming inclusion bodies.

I will be appreciated that this method for controlled expression is applicable to other inducible promoter systems, e.g. the tac system, where the inducer is IPTG and the host is a lac Y negative host.

A route to obtain relevant structural information on the active site configuration of an enzyme is the production and characterisation of monoclonal antibodies capable of inhibiting the enzymic reaction. The - 95041 2 ,. -12antibodies inhibiting the activity represent molecules which block or compete with the substrate for entry into the active site pocket, or can represent molecules which can prevent structural transitions required for catalytic activity. In either case, these antibodies can be used as a tool to quantitate interaction of the target enzyme with binding of radiolabelled inhibitory compounds to judge the affinity of interaction provided the affinity of the inhibiting antibody is known. A further use of mapping the epitopes recognised by the inhibitory antibodies is the ability to delineate residues forming the active site.

Consequently, a further aspect of the invention is a method of identifying an antibody capable of binding a bacterial bifunctional penicillin binding protein which includes the step of employing a polypeptide according to the invention in an antibody binding assay and selecting antibodies that bind to the polypeptide.

Also included in the invention are monoclonal antibodies directed to a PBP variant according to the invention. Such a monoclonal antibody is prepared using known hybridoma technology by fusing antibody-producing B-cells from immunized animals with myeloma cells and selecting the resulting hybridoma cell line producing the desired antibody.

Another aspect of the invention is a method of assaying for compounds which bind to a penicillin binding protein, said method comprising (a) contacting a polypeptide which is a PBP variant according to the invention with a compound to be investigated; and (b) detecting whether said compound binds to the said PBP variant.

For example, a method of assaying for compounds which bind to a penicillin binding protein can comprise (a) culturing host cells according to the invention; (b) lysing the said cells and isolating the crude cell extract; (c) exposing the said cell extract to potential inhibitors of a penicillin 950412? -13binding protein; (d) introducing an agent, known to bind a penicillin binding protein, to the said cell extract; (e) removing the unbound fraction of said agent; and (f) assaying the presence of said agent remaining in the cell extract.

Another method of assaying for compounds which bind to a penicillin binding protein could comprise (a) exposing a polypeptide which is a PBP variant according to the invention, immobilised on a solid support, to a potential inhibitor of a penicillin binding protein; (b) exposing an agent, known to bind a penicillin binding protein, to the immobilised polypeptide; (c) removing the unbound fraction of said agent; and (d) assaying the presence of said agent bound to the immobilised polypeptide.

In a preferred form, the said method is a method of assaying for compounds which bind to the transglycosylase domain of a penicillin binding protein, said method comprising (a) exposing the transglycosylase domain of a polypeptide according to the invention, with the proviso that the polypeptide is not a transglycosylase deficient PBP variant, said polypeptide being immobilised on a solid support, to a potential inhibitor of the transglycosylase activity of a penicillin binding protein; (b) exposing an agent, known to bind the transglycosylase domain of a penicillin binding protein, to the immobilised polypeptide; (c) removing the unbound fraction of said agent; and (d) assaying the presence of said agent bound to the immobilised polypeptide.

Antibodies specific for transpeptidase can be immobilised on a BIAcore sensor chip surface. The BIAcore system, wherein BIA stands for Biospecific Interaction Analysis, is available from Pharmacia Biosensor, Sweden. Protein binding to the immobilised antibody is detected by the output RU-signal. Screening for TP inhibitors will be possible by a competitive assay wherein soluble protein is preincubated with test compounds. Binding of a test compound to the protein will result in a 0 4 1 2 j -14decrease in protein binding to TP specific antibody. In the same way, monoclonal antibodies specific for transglycosylase can be used in screening for TG inhibitors.

In a similar way, ampicillin or modified moenomycin can be coupled to the surface and used in an indirect competitive assay whereby protein is preincubated with test ligand prior to introduction in the BIAcore.

Consequently, yet another method of assaying for compounds which bind to a penicillin binding protein could comprise (a) exposing a polypeptide which is a PBP variant according to the invention to a potential inhibitor of a penicillin binding protein; (b) exposing the polypeptide to an agent, known to bind a penicillin binding protein, which agent is immobilised on a solid support; and (c) assaying the presence of polypeptide bound to the immobilised agent.

In a preferred form, the said method is a method of assaying for compounds which bind to the transglycosylase domain of a penicillin binding protein, said method comprising (a) exposing the transglycosylase domain of a polypeptide according to the invention, with the proviso that the polypeptide is not a transglycosylase deficient PBP variant, to a potential inhibitor of a penicillin binding protein; (b) exposing the said polypeptide to an agent, known to bind to the transglycosylase domain of a penicillin binding protein, which agent is immobilised on a solid support; and (c) assaying the presence of polypeptide bound to the immobilised agent.

The agent known to bind a penicillin binding protein referred to above can e.g. be a monoclonal antibody or a labelled antibiotic compound such as pH]ampicillin. 950412. -15A further aspect of the invention is a method of determining the protein structure of a penicillin binding protein, characterized in that a polypeptide which is a PBP variant according to the invention is utilized in X-ray crystallography.

Some of the features of the preferred PBP variants according to the invention are summarised in Table 1 below. The plasmids listed in the Table have been deposited under the Budapest Treaty at the National Collection of Industrial and Marine Bacteria Limited (NCIMB), Aberdeen, Scotland, UK. The date of deposit is 28 June 1994. 0 4 1 2 j __ 1 -16TABLE 1 Example no. Features Plasmid (pARC) Deposit no. (NCIMB) Fig. SEQ ID NO: Soluble variants 1.1 E.coli PBP IA with aa 1-23 deleted 0558 40666 3 1,2 2.1 E.coli PBP IB with aa 65-87 deleted 0559 40667 9 3,4 3.1 S.pneumomae PBP IA with aa 1-38 deleted 0512 40665 12 5,6 Transglycosylase deficient variants 4.1 E.coli PBP IB with glutamines 270271 substituted to alanines 0438 40661 7 E.coli PBP IB with glutamines 270271 substituted to leucines 0468 40662 8 E.coli PBP IB with aa 264-271 deleted 0469 40663 9 4.2 E.coli PBP IA with glutamines 123124 substituted to alanines 0571 40668 19 10 Truncated variants 5.1 aa 1-553 of E.coli PBP IB 0592 40669 21 11 aa 1-553 of E.coli PBP IB, with aa 65-87 deleted 0593 40670 22 12 5.2 aa 210-368 of E.coli PBP IB 0392 40659 23 13 Fusion proteins 6.1 E.coli PBP IA with 23 aa deletion, ligated to glutathione-S-transferase 0499 40664 24 6.2 E.coli PBP IA with 23 aa deletion, ligated to histidine stretch 0400 40660 25 0 4 1 2 .1 -17EXAMPLES OF THE INVENTION In the following examples, the terms standard protocols and standard procedures are to be understood as protocols and procedures found in an ordinary laboratory manual such as the one by Sambrook, Fritsch and Maniatis (1989).

EXAMPLE 1 1.1. Construction of gene encoding soluble form of E.coli PBP IA The possible amino acid residues involved in the membrane anchoring region of PBP IA was deduced following the computer program described by Kyte & Dolittle (1982). The predicted hydrophobicity of the N-terminal 60 amino acid is shown in Fig. 1. Based on this hydrophobicity profile, it was predicted that the N-terminal 23 amino acids were strongly implicated to contribute to the membrane anchoring domain of the protein, but may not entirely encompass the membrane anchoring domain. This region was then putatively designated as the region involved in membrane anchoring.

The plasmid pBS98, harbouring the native ponA gene (encoding wild type PBP IA), was obtained from Prof. B.S. Spratt, Microbial Genetics Group, School of Biological Sciences, University of Sussex, Brighton, UK. The construction of pBS98 is described in Broome-Smith et al. (1985). Plasmid DNA from cells harbouring pBS98 was made following standard protocols.

Oligonucleotide primers for use in the polymerase chain reaction (PCR) were synthesized in Applied Biosystems Model 380 A. The 5'oligonucleotide primer used was TG-82: 0 4 1 2-1 # -18TAC ATC G-3' Y I 29 (Amino acid No.) Ncol '-TCG ACC ATG GGC CTA TAC M G L Y 23 24 25 26 CGC R TG-82 incorporates the following characteristics: (1) it allows construction of mutant ponA gene whose encoded product would have the 24th amino acid (glycine) of the wild type PBP IA as the second amino acid of the expressed mutant protein; and (2) it introduces DNA sequences recognized by the restriction enzyme Ncol. This introduces the codon ATG which corresponds to the first amino add of the mutant PBP IA when expressed in suitable systems.

The 3'-oligonucleotide primer used was TG-64: '-CGC GGA TCC GAA TCA CAA CAA TTC CTG TGC-3' T BamHI TG-64 has the following characteristics: (1) it introduces a termination codon following the 850th amino add of the structural protein of PBP IA; (2) it introduces a site for the restriction enzyme BamHI to facilitate cloning into suitable expression vectors.

Using these primers, PCR was carried out using pBS98 DNA as template following standard protocols. A DNA fragment of approximately 2.5 kb was amplified. The fragment was digested with the restriction enzyme Ncol followed by digestion with BamHI. This 2.5 kb Ncol - BamHI DNA fragment was then ligated to the vector pBR329 (Covarrubias et al., 1982) previously cut with Ncol and BamHI. Ligation of the two DNA fragments were carried out using standard protocols and the ligation mixture transformed into E.coli DH 5a. The transformed cells were plated on LB agar plates with 50 pg/ml ampidllin. Following overnight incubation at 37°C, individual ampidllin resistant colonies were tested for their 0 4 1 2,) -19tetracycline sensitivity as insertion into the Ncol - BamHI region renders the plasmid chloramphenicol and tetracycline sensitive. A recombinant plasmid bearing the 2.5 kb insert was designated pARC0488.

The Ncol - BamHI 2.5 kb DNA fragment was released from pARC0488 and ligated to Ncol - BamHI cleaved and purified pARC038 (Fig. 2). The plasmid pARC038 is a derivative of pETlld (Studier et al., 1990) in which the EcoRI and Pstl sites were made blunt ended with T4 exonuclease and the EcoRI - Pstl 0.75 kb DNA fragment replaced with a blunt ended kanamydn resistance cartridge (Pharmada Biochemicals). The ligation mixture was transformed into competent cells of E.coli BL 26 (DE3). The transformation mix was plated on LB agar with 50 pg/ml kanamydn. Mini-prep plasmid DNA was made from several kanamydn resistant colonies and screened by restriction endonuclease mapping using standard procedures.

One of the colonies harbouring plasmid with expected structure (Fig. 3) was labelled pARC0558 (NCIMB 40666). The DNA sequence of the mutant ponA gene labelled as ponAdel23 is shown as SEQ ID NO: 1. The amino add sequence of the soluble PBP lAdel23 is shown as SEQ ID NO: 2. 1.2. Expression of ponAdel23 E.coli BL 26 (DE3) cells (obtained from Dr. J.J. Dunn, Biology Dept., Brookhaven National Lab., Long Island, NY, USA) harbouring pARC0558 were grown in LB with 50 pg/ml kanamydn till an O.D. at 600 nm of 0.6 and induced with 0.01 mM isopropyl thiogalactoside (IPTG) for 6 hours.

Following 6 hours of induction, cells were harvested and broken by passing through a French press. After centrifugation at low speed to remove unbroken cells and debris, the cytosolic (soluble) fraction was obtained by either of the following two methods: (1) following a procedure 050412^ —20— described Page et al. (1982) in which the pellet, membrane and soluble proteins are separated by sucrose gradient centrifugation; or (2) by spinning the obtained supernatant at 200,000 x g for 90 minutes, whereafter the supernatant obtained is taken as the cytosolic / soluble protein fraction. 1.3. Penicillin binding of expressed PBP lAdel23 The obtained cytosolic fraction was tested for the presence of mutant PBP 1A by following the method of Rojo et al. (1984). This procedure involves using [ Ijcephradine as the labelled penicillin as it is specific for PBP 1A. Mutant PBP lAdel23 capable of binding the labelled cephradine could be demonstrated in the cytosolic fraction. Approximately 50% of the expressed mutant protein fractionated as a soluble protein, while the remaining 50% fractionated into the inclusion body and/or into the membrane associated fractions. Consequently, enhanced levels of active mutant PBP lAdel23 were obtained since the cells were induced with sub-optimal concentration of IPTG and the since cultures were grown at 22°C. The penicillin binding profile of the soluble PBP lAdel23 is shown in Fig. 4. 1.4. Purification of soluble PBP lAde!23 The cell pellet of E.coli BL26 (DE3) / pARC0558 obtained following 6 hours of induction at 22°C was washed twice with buffer A (30 mM Tris-Cl, pH 8.0; 10 mM EDTA; 10 pg/ml leupeptin; 10 pg/ml aprotinin; 5 mM DTT) and resuspended in the same buffer. The cell suspension was passed through a French press at 1200 psi. The lysate was spun at 10,000 rpm for 10 minutes and the obtained supernatant centrifuged at 200,000 x g for 45 minutes. The obtained supernatant was then adjusted to 30% saturation with ammonium sulphate. The mixture was centrifuged at 12,000 rpm for 10 min and the pellet resuspended in buffer A containing 1 M NaCl. The dissolved pellet was then treated with Cephradine-Affigel 10 matrix. 950412. -21Cephradine was conjugated to Affigel 10 following the instructions of the manufacturers (Biorad Laboratories, USA). The soluble PBP lAdel23 containing fraction, dissolved in buffer A containing 1 M NaCl, was incubated 16 hrs at 4°C. with cephradine-affigel 10 beads. The beads were then washed with Buffer A containing 1 M NaCl until the absorbance at 280 nm was nearly zero. Elution of PBP lAdel23 was monitored by assaying for penicillin binding activity in the wash. This activity was OR measured using cephradine prepared as described in Rojo et al. (1984). Bound PBP lAdel23 was eluted from the beads using 1 M hydroxylamine (pH 8.5) at 25°C for 120 minutes. This fraction was concentrated by ultrafiltration using YM 30 filters (Amicon, USA) in Buffer A with 0.25 M NaCl. The ultrafiltration also resulted in the removal of hydroxylamine. The purified fraction containing >85% of the protein species corresponding to PBP lAdel23 showed both penicillin binding and transglycosylase enzyme activities. The protein profile as seen by Coomassie Brilliant Blue staining and the [ IJcephradine / penicillin binding profile of the different fractions, obtained during the various stages of purification, are shown in Fig. 5. The N-terminal amino acid sequence of the soluble PBP lAdel23 was confirmed by sequencing the purified protein. 1.5. Transglycosylase activity of soluble PBP lAdel23 The transglycosylase activity of the soluble PBP lAdel23 protein was measured using essentially the method described by Ishino et al. (1980). The substrate for the detection of the enzymic activity were essentially prepared and purified following the protocols described by Heijenoort et al. (1992). The concentration dependent transglycosylase activity of PBP lAdel23 measured as the amount of peptidoglycan formed was compared to the amounts of peptidoglycan formed by different concentrations of the membrane bound form of native PBP 1A. As seen in Fig. 6, the peptidoglycan polymerizing efficiency of the mutant soluble PBP lAde!23 0 4 1 2 ·’ -22was nearly identical to the enzymic activity of the membrane bound form of the protein.

It has consequently been found that the elimination of the 23 amino acid residue stretch does not interfere with the ability of the protein to assume its native structure capable of both the enzymatic activities, i.e. the transglycosylase and the transpeptidase activities.

EXAMPLE 2 2.1. Construction of gene encoding soluble form of E.coli PBP IB The ponB gene encoding PBP IB was obtained on a plasmid pBS96 from Prof. B.S. Spratt, Microbial Genetics Group, School of Biological Sciences, University of Sussex, Brighton, UK. The construction of pBS96, as well as the nucleotide sequence of the wild-type ponB gene and the derived amino acid sequence, are described in Broome-Smith et al. (1985).

The hydropathy plot of the N-terminal approximately 150 amino acids as derived using the method of Kyte and Doolittle (1982) is shown in Fig. 7. Analysis of the hydropathicity plot indicated that the amino acids at positions 65 to 87 of the PBP IB sequence contributed largely to the hydrophobicity of the N-terminus and can be putatively assigned to be the membrane anchoring domain of the protein. In addition, β-lactamase studies of Edelman et al. (1987) had indicated that amino adds C-terminal to amino acid position 87 were present in the periplasmic space of the E.coli cell and that amino acids N-terminal to position 65 of PBP IB were within the cytoplasm of the cell.

The strategy employed to construct a mutant ponB gene encoding a soluble form of PBP IB is shown in Fig. 8. Initially a DNA fragment of approximately 200 bp of the 5'-end of the ponB gene was amplified by 9504^2 ;C · —23— PCR, from the ponB gene on the plasmid pBS96 (Broome-Smith et al., 1985). The oligonucleotide primers used were 5'-primer TG-77 (5'-GAA AAA CCA TGG CCG GGA ATG ACC-3') which includes a Ncol restriction enzyme site which also coincides with the start ATG codon of the sequence, and 3'-primer TG-84 (5'-AAG TCG CGA GCC GCG TTT GCC AC-3') which includes a site for the restriction enzyme Nrul and encodes for amino acids corresponding to position 64 of the PBP IB sequence.

Step 1: The PCR amplified fragment following restriction with the enzymes Ncol and Nrul was cloned into the Ncol - Nrul sites of the cloning vector pBR 329 (Covarrubias et al., 1982). Ligation, transformation and screening were carried out using standard protocols and the recombinant plasmid with the expected structure labelled pARC0547 (Fig. 8) was obtained.

Another DNA fragment of approximately 1.2 kb was amplified by PCR using primer sequences corresponding to amino acid 87 to 480. This DNA fragment encodes the C-terminal half of the TG domain of PBP IB. The primers used were 5'-primer TG-79 (5'-CGG ATA TCG ATC AAA AAA TTC GTA GCC G-3') which included the nucleotide sequence for the cleavage site for the restriction enzyme EcoRV, and 3'-primer TG-80 (5'-GCG GAT CCT TAG TCG ACG ACC ACA ATC GCA G-3') which included the sequence for BamHI cleavage.

Step 2: The PCR amplification of this fragment was done using the ponB gene on pBS96 (Broome-Smith et al., 1985) DNA as template. The amplified fragment was cloned into the EcoRV - BamHI sites of pBR 329 (Covarrubias et al., 1982) using standard protocols. The recombinant plasmid obtained was labelled pARC0534 (Fig. 8).

Step 3: The 200 bp Ncol - Nrul fragment cloned in pARC0547 was excised as a Ncol - Nrul fragment and cloned into Ncol - EcoRV cleaved pARC0534 to obtain pARC0551 (Fig. 8). —24— Λ The mutant ponB gene on pARC0551 has DNA sequences coding for the N-terminal 64 amino acids of PBP IB fused to the nucleotide sequences encoding the amino acids 88 to 480. A 1.3 kb Pstl - BamHI DNA fragment of pBS96 was then ligated to Pstl - BamHI cleaved pARC0551 and the ligation mixture transformed into E.coli DH5a using standard procedures. Individual transformants were then screened and colonies harbouring recombinant plasmid with the expected structure identified. The plasmid was labelled pARC0552. A Ncol - BamHI fragment from pARC0552 encompassing the entire mutant ponB gene was then excised and ligated to the T7 expression vector pARC038 to obtain pARC0559 (NCIMB 40667; Fig. 9).

The 3'-end of the cloned fragment of Step 1 has the nucleotide sequence TCG (partial Nrul site sequence) while the 5'-end of the fragment cloned in Step 2 has the sequence ATC (partial EcoRV cleavage sequence). The junction nucleotide sequence which is the outcome of the fusion of TCG and ATC results in the introduction of the codons for serine and isoleucine. Thus the mutant ponB gene encodes a PBP IB with the amino acid sequence 1 to 64 corresponding to the wild type PBP IB fused to the sequence 87 to 844. The two stretches are joined by the amino acids serine and isoleucine.

The nucleotide sequence of the mutant ponB gene is shown as SEQ ID NO: 3 and the derived amino add as SEQ ID NO: 4. 2.2. Expression of soluble PBP IB The plasmid DNA of pARC0559 was transformed into the T7 expression host E.coli BL 26 (DE3) and the restriction map profile of the transformed plasmid confirmed using standard procedures. E.coli BL26 (DE3)/pARC0559 were grown at 22°C and induced with 0.01 mM IPTG and the cells allowed to grow for 6 hours. Cells were then harvested and ο 0 4 1 2 J —25— broken by passage through a french press. The lysate was centrifuged at 10,000 rpm for 10 minutes and the supernatant obtained was centrifuged at 200,000 x g for 45 minutes in a Beckman ultracentrifuge. 2.3. Characterization of the expressed soluble PBP IB The obtained supernatant, i.e. the cytosolic / soluble fraction, was tested •pc for the presence of the mutant PBP IB using [ IJampicillin as the radio-ligand. The [^'^^Ilampicillin was prepared as described by Rojo et al. (1984) for the preparation of [^^Ijcephradine. The mutant PBP IB was detected in the soluble fraction and bound radioactive ampidllin.

Soluble PBP IB could also be purified using Ampidllin - Affigel beads by a procedure analogous to the one described in Section 1.4. The protein profile of the different fractions seen by Coomassie Blue staining and the -I JC binding of [ IJampicillin of the enriched PBP IB fraction is shown in Fig.

. The purified protein was enzymatically active in the peptidoglycan transglycosylase assay (Heijenoort et al., 1992) and bound penidllin with an affinity comparable to that of the membrane bound native PBP IB.

EXAMPLE 3 3.1. Construction of gene encoding soluble form of Streptococcus pneumoniae PBP IA The molecular architecture of the S.pneumoniae PBP IA is predicted to be similar to that of E.coli PBP IA and PBP IB protein in the fact that the protein is anchored to the membrane via a N-terminal membrane anchoring sequence. The nucleotide sequence of the gene encoding native membrane bound S.pneumoniae PBP IA and its derived amino acid 950412, —26— sequence are described in Martin et al., (1992). The hydropathicity profile of the N-terminal 100 amino acids as derived by the Kyte and Doolittle plot is shown in Fig. 11. A stretch of 38 amino acids contributed significantly to the hydrophobicity of this region and was assumed to be the membrane interacting domain. A mutant gene of S.pneumoniae PBP 1A was constructed by deleting the nucleotide sequence coding for the N-terminal 38 amino acids of S.pneumoniae PBP 1A.

Using standard PCR protocols, sequences encoding the wild type S.pneumoniae PBP 1A gene was amplified as a 2.5 kb DNA fragment from the chromosome of S.pneumoniae strain PM1 (obtained from S.A. Lacks, Biology Department, Brookhaven National Laboratory, Upton, New York, USA) (Lacks, 1968) using the primers designed based on the sequence reported by Martin et al. (1992) and the amplified fragment cloned into the pneumococcal vector pLS 101 (Balganesh and Lacks, 1984).

The mutant gene encoding a soluble form of S.pneumoniae PBP 1A was constructed by using of plasmid DNA harbouring the wild type gene as template and amplifying a 2.3 kb DNA fragment by using PCR following standard procedures. The sequence of the primers used were 5'-primer TG-24 (5'-TAC GTT ACC ATG GCT CCT AGC CTA TCC-3') and 3'-primer TG-25 (5'-GAC AGG ATC CTG AGA AGA TGT CTT CTC A-3').

The 5'-primer TG-24 includes the sequence for the restriction enzyme Ncol while the 3'-primer TG-25 includes the site for the restriction enzyme BamHI. The Ncol and BamHI digested PCR amplified DNA fragment was ligated to Ncol - BamHI cleaved pARC039. The plasmid pARC039 is a derivative of pET 8c (Studier et al., 1990) in which the gene coding for the β-lactamase has been replaced by a kanamycin resistance cartridge.

Following ligation and screening using standard protocols, the structure of the recombinant plasmid was confirmed by detailed restriction mapping 0 4 1 2 —27— and transformed into the T7 expression host E.coli BL 21 (DE3) (Studier et al., 1990). The recombinant plasmid was labelled pARC0512 (NCIMB 40665) and is schematically represented in Fig. 12.

The nucleotide sequence of the mutant S.pneumoniae PBP IA gene is shown as SEQ ID NO: 5 and the derived amino acid sequence is shown as SEQ ID NO: 6. 3.2. Expression and characterization of soluble form of Streptococcus pneumoniae PBP IA The gene coding for soluble S.pneumoniae PBP IA was expressed by a procedure analogous to the one described in Section 1.2.. The cytosolic fraction of E.coli BL 21 (DE3)/pARC0512 was isolated and tested for the presence of the soluble form of the S.pneumoniae PBP lAdel38. The radioactive ligand used for the binding studies was [ H]benzyl penicillin (Amersham) which was prepared as described earlier. Approximately 50% of the expressed protein from the mutant gene was found to be in the soluble fraction and bound P^^IJpenicillin (Rojo et al., 1984) or pHjpenicillin (Amersham) when the culture was grown and induced at 22°C with 0.01 mM IPTG. The growth and induction conditions were critical for the efficient recovery of the soluble protein, as growth at higher temperatures or induction with higher concentrations of IPTG resulted in the majority of the protein becoming inactive and forming inclusion bodies. Optimum levels of soluble active protein was found following induction for 6-8 h. (Fig. 13).

The soluble S.pneumoniae PBP lAdel38 protein could also be efficiently purified essentially following the protocol used for the purification of the soluble E.coli PBP IB protein. 95c The efficiency of penicillin binding of the soluble PBP lAdel38 was comparable to that of the native membrane bound S.pneutnoniae PBP IA.

EXAMPLE 4 4.1. Transglycosylase deficient E.coli PBP IB The conserved amino acids within Region 2 (Fig. 14) were chosen for sitedirected mutagenesis. Within this stretch of 10 amino acids three different mutations were constructed: (a) the glutamines at position 270 and 271 of the PBP IB sequence were changed to alanines; (b) the glutamines at position 270 and 271 of the PBP IB sequence were changed to leucines; and (c) a deletion of the nucleotide sequence encoding amino acids from position 264 to 271.

Mutants of the ponB gene were constructed essentially following the procedure of Kunkel et al. (1985). A 1.5 kb EcoRI - Sail fragment of the ponB gene of the plasmid pBS96 was excised and cloned into EcoRI - Sail cleaved M13mpl9 following standard protocols. (a) The primer used for mutating the nucleotide sequence coding for glutamine residues 270 and 271 into a sequence coding for alanine residues was TG-21: '-ACG CTG ACG GCC GCT CTG GTG AAA-3' TLTAALVK (b) The primer used for mutating the sequence coding for the glutamine residues 270 and 271 into leucine residues was TG-23: '-ACG CTG ACG CTA TTG CTG GTG AAA-3' —29— 950412, TLTLLLVK (c) The primer used for creating a deletion of the nucleotides encoding amino acids at position 264 to 271, all of which are within the conserved Region 2, was TG-22: '-CGC ACG GTA CAG CTG GTG AAA AAC-3' R Τ V Q L V K 260 261 262 263 272 273 274 (amino acid no.) Following mutagenesis, the nucleotide sequence of the mutagenized EcoRI - Sail fragment was determined following the protocol of Sanger et al. (1977). The sequencing confirmed the nucleotide changes and also ruled out any extraneous changes. This mutated 1.5 kb DNA fragment was ligated back to EcoRI - Sail cleaved pBS96 and the ligated DNA transformed in to E.coli DH5a cells following standard protocols. Kanamycin resistant transformants were analyzed for their plasmid profiles and the plasmid with the TG-21 mutation (a) was labelled pARC0438 (NCIMB 40661). The mutant protein is referred to as PBP IB QQ-AA (SEQ ID NO: 7).

The plasmid with the mutation (b) introduced by TG-23 was labelled pARC0468 (NCIMB 40662). The mutant protein is referred to as PBP IB QQ-LL (SEQ ID NO: 8) The plasmid with the deletion (c) obtained using TG-22 was labelled pARC0469 (NCIMB 40663). The mutant protein is referred to as PBP lBdel8 (SEQ ID NO: 9).

The four plasmid DNAs of pBS96, pARC0438, pARC0468 and pARC0469 were individually transformed into E.coli ponB.spF cells (Broome-Smith et al., 1985) in which a deleted ponB gene had been marked with spectinomycin resistance marker. 950412, -30E.coli ponB.spR cells having the individual plasmids pBS96, pARC0438 or pARC0469 were grown and membrane preparations made following the procedure described by Spratt (1977) and the profile of the penicillin binding proteins analyzed on a 8% SDS-PAGE following labelling with radioactive penicillin. The mutant proteins were initially analyzed for in vivo stability and localization into the membrane using anti-PBP IB sera raised against purified membrane bound native PBP IB (Fig. 15).

The mutant proteins were found to be localized to the membrane and no degraded protein fragments reacting with the antibody could be detected indicating no gross instability. In addition the mutant proteins bound penicillin with an affinity comparable to that of the wild type PBP IB (Fig. 15).

After assaying for transglycosylase activity as described in Heijenoort et al. (1978), no activity could be detected in the membranes expressing the mutant proteins, while the membrane with the wild type PBP IB showed transglycosylase activity. This defines the amino acids 263 to 271 as being critical for transglycosylase activity.

The ability of the mutant proteins to bind penicillin with an affinity comparable to that of the wild type suggests that the transpeptidase activity of the mutant proteins would also be comparable to that of the wild type. Knowing that the bifunctional protein PBP IB expressed on a plasmid can in trans complement the deletions of both ponA and ponB (Yousif et al., 1985) the ability of the transglycosylase negative / transpeptidase positive proteins PBP IB QQ-AA and PBP lBdel8 to complement the absence of chromosomally encoded PBP IA and IB was tested.

The wild type ponB and the mutant ponB genes were cloned into low copy vector pMAK 705 (Hamilton et al., 1989). The resulting plasmids were 950412. —31 designated pARC0462, (wild type ponB, Fig. 16), pARC0463, (ponBdel8, Fig. 17) and pARC0470 (ponB QQ-AA, Fig. 18). The plasmids were individually transformed into E.coli del ponA (E.coli with a deletion of the ponA gene).

E.coli del ponA /pARC0462, E.coli del ponA/pARC0463 and E.coli del ponA/pARC0470 were used as recipients of the Pl phage for the transduction of the ponBispc1 marker. The transduction was performed as described by Miller (1972) The phage Pl lysate was made on E.coli ponB.spc? strain (Yousif et al., 1985). Following infection, the infected cells were plated on spectinomycin. Integration of the DNA fragment harbouring and po«B:spcr transduced into any of the recipients results in the inactivation of the chromosomal ponB gene rendering the chromosome ponA' and ponB'. This genotype being lethal for the cell, the E.coli spectinomycin resistant transductants can remain viable only if the plasmid encoded ponB or the ponB mutant can functionally complement in trans.

The following E.coli strains were subject to phage Pl transduction analysis of trans-complementation: (1) E.coli AMA 1004 which has chromosomally coded wild type ponA and ponB; (2) E.coli AMA 1004 which has a chromosomally inactivated ponB and is the host for the plasmid coded mutant ponB genes; (3) E.coli AMA 1004 host bearing the plasmid pARC0462 encoding the wild type ponB gene; (4) E.coli AMA 1004 host bearing the plasmid pARC0463 encoding PBP lBdel8; and (5) E.coli AMA 1004 host bearing the plasmid pARC0470 encoding PBP IB QQ-AA. -32Results (Number of Kmr transductants / ml) (1) E.coli AMA 1004 (2) E.coli AMA 1004, ponB.spE (3) E.coli AMA 1004, ponB.spc1 (PBP IB wt) (4) E.coli AMA 1004, ponB.spc? (PBP lBdel8) (5) E.coli AMA 1004, ponB:sp 3.0 χ 104 < 1 1.1 χ 104 A comparable number of transductants were obtained for an internal marker : trp transduction using the same Pl phage lysate.

The above results show that viable transductants could be obtained only with wild type PBP IB, indicating that the TG’ TP+ product encoded by ponB QQ-AA or po«Bdel8 could not complement the loss of chromosomally encoded PBP 1A and IB. However, as these mutant proteins bind penicillin and thus can be assumed to have transpeptidase activity, the inability to complement must be the absence of the transglycosylase enzymic activity. These results confirm the essential nature of the transglycosylase activity of PBP 1A or IB for the viability of the E.coli cell.

The mutants described define the Region 2 to be involved in the transglycosylase activity of the protein. As this stretch of amino acids is conserved within the four high molecular weight penicillin binding proteins namely E.coli PBP 1A, IB and S.pneumoniae 1A and the 94 kDa protein of H.influenzae (Fig. 14) it is reasonable to assume similar catalytic or structural involvement of this region in all the transglycosylase enzymes utilizing substrates similar to that used by PBP 1A and IB of E.coli. 4.2. Transglycosylase deficient E.coli PBP 1A The conserved Region 2 was chosen for site-directed mutagenesis and the nucleotide sequence coding for glutamine at positions 123 and 124 of E.coli - 9 50 4 j 2 -33PBP IA was changed to a sequence coding for alanine by PCR mutagenesis as follows. The 5' half of the ponA gene was amplified as 2 fragments, the 5'-fragment corresponding to amino acid 1 to 123 (fragment A) and the 3'fragment corresponding to amino acid 124 to 434 (fragment B).

The sequence of the 5'-primer used for the amplification of fragment A was TG-93 (5'- GCG CGG ACC ATG GTG AAG TTC GTA AAG TAT-3') while the 3'-primer used for the amplification of fragment A was TG-106 (5'-CAG TGC TGC AGT AAT GGT ACT TGC CCC TTG-3').

The 3'-primer for fragment A amplification included the sequence for the restriction enzyme Psfl which allowed the conversion of the sequence encoding the glutamine residues in position 123 and 124 into a nucleotide sequence coding for alanine residues.

Fragment B was amplified with the 5'-primer TG-107 (5'-ATT ACT GCA GCA CTG GCG AGA AAC TTC TTC-3') and the 3'-primer TG-108 (5'-TCG CGA GAT ATC TGG CGG ATT GAT CGA CAC-3').

The 5'-primer for amplifying fragment B included the sequence for the restriction enzyme Psfl overlapping the sequence with that of 3'-primer for amplifying fragment A. Ligation of the 3'-end of fragment A to the 5'-end of fragment B recreated the site for Psfl and resulted in the change of the nucleotide sequence encoding glutamine 123 and 124 into alanine 123 and 124. The amplified fragments A and B were individually cloned into pBR 329, and corresponding clones pARC0565 and pARC0566 were obtained.

Fragment A and B obtained from pARC0565 and pARC0566 were ligated to obtain pARC0567. The ponA sequences were completed by introducing an Xhol - BamHI fragment of pARC0489 (which is identical to pARC0558 (Fig. 3) except for having additional Lad and Lac operator sequences) into pARC0567 to obtain pARC0568. The Mlul - Bglll fragment of pARC0568 - 95 0 4 1 2 ,1 -34which included the Qj23 * Qi24 t0 ^123 ' ^124 mutated region was then used to replace the otherwise identical Mini - Bglll fragment of pBS98 to obtain the plasmid pARC0571 (Fig. 19; NCIMB 40668). The mutant protein was labelled PBP IA QQ-AA (SEQ ID NO: 10).

Expression studies on the mutant indicated that the mutant protein was localised to the membrane (as detected by anti PBP IA antibodies) and bound penicillin with an affinity comparable to that of the native PBP IA (Fig. 20).

An in vivo complementation assay, similar to that described in the previous section, was performed by checking the ability of mutant PBP IA protein to complement in trans. The in vivo complementation was performed using phage Pl transduction and transducing ponBispc1 into the host E.coli (recipient) del ponA harbouring the plasmid encoding the mutant protein PBP IA QQ-AA.

In order to carry out the complementation analysis the wild type ponA gene was cloned into the low copy vector pMAK 705 (Hamilton et al, 1989) to obtain pARC0583 and the mutant ponA gene encoding PBP IA QQ-AA cloned into pMAK 705 to obtain pARC0582.

The following E.coli strains were subject to phage Pl transduction analysis of trans-complementation: (1) E.coli AMA 1004 which has chromosomally coded ponA and ponB; (2) E.coli AMA 1004 ponA which has a chromosomally inactivated ponA and is the host for the plasmid coded mutant ponA genes; (3) host bearing the plasmid pARC0583 encoding the wild type ponA gene; (4) host bearing the plasmid pARC0582 encoding PBP IA QQ-AA. -35950412.

Results (Number of SpZ transductants / ml) (1) E.coli AMA 1004 (2) E.coli AMA 1004, ponA (3) E.coli AMA 1004, ponA (PBP IA wt) (4) E.coli AMA 1004, ponA (PBP IA QQ-AA) 2.1 x 103 < 1 1.64 x 103 A comparable number of transductants were obtained for internal marker : trp transduction using the same Pl phage lysate.

As shown above, no viable transductants could be obtained with E.coli del ponA / pARC0582 as recipient indicating that the mutant PBP IA QQ-AA could not complement the absence of chromosomally encoded PBP 1A/1B. This indicates that the Qj23 and Qi24 regi°n 2 of PBP IA also affects transglycosylase activity of the protein as the loss of the complementing function must be a reflection of the loss of transglycosylase activity. The transpeptidase activity of the protein is unaffected as tested by its affinity to bind penicillin.

These results argue in favour of the region 2 as a critical stretch of amino acids involved in the transglycosylase enzymic function and may be the explanation for the strong evolutionary conservation of this stretch of amino acids.

EXAMPLE 5 .1. Truncated E.coli PBP IB A mutant gene encoding the truncated PBP IB consisting of the N-terminal 553 amino acids was constructed by PCR amplification using the 5'-primer TG-77 (5'-GAA AAA CCA TGG CCG GGA ATG ACC-3') and the 3'primer TG-116 (5'- ATG GGA TCC TTA ATC ATT CTG CGG TGA-3'). —36— The 5' end of the primer corresponded to the amino acid 553 in the wild type followed by the stop codon and a site for the restriction enzyme BflznHI. A fragment of 1.7 kb was amplified using pBS96 DNA as template. The PCR amplified fragment was cut with Pstl and BamHl and cloned into Pstl-BamHl restricted pARC0555 (pARC0555 has the full length ponB gene cloned as Ncol-BamHl fragment into the expression vector pETlld. The Ncol site includes the initiation codon ATG) to obtain pARC0592 (NCIMB 40669; Fig. 21) The expressed protein (SEQ ID NO: 11) was shown to have transglycosylase activity, thus confirming the functional independence of this domain.

The soluble truncated PBP IB, i.e. PBP IB with N-terminal 553 amino acids but lacking the membrane anchoring hydrophobic domain from 65-87, was constructed by replacing the Pstl-BamHl fragment of pARC0559 (Fig. 9) with the Pstl-BamHl fragment of pARC0592 to obtain pARC0593 (NCIMB 40670; Fig. 22). The mutant ponB gene encodes the soluble form of PBP IB and the expressed protein (SEQ ID NO: 12) was found to have transglycosylase activity. .2. Minimum substrate binding domain of truncated E.coli PBP IB Detailed computer analysis of the anatomy of the presumptive TG domain (aa 1-553) of PBP IB indicated that aa 210-368 were probably sufficient for the binding of the lipid linked substrate and the transglycosylase reaction. This stretch of amino acids includes the 3 conserved domains Region I, II and ΠΙ. The mutant gene encoding the truncated protein stretch 210-368 was constructed as follows.

A fragment of approx size 480 bp was amplified from pBS96 as substrate with the 5'-primer having the sequence TG-154 (5'-CAA TCC ATG GGT GAG CAG CGT CTG TTT G-3') were the initiation ATG codon is 950412, -37immediately followed by the sequence encoding the 210th amino acid of PBP IB.

The 3'-primer corresponded to the sequence TG-155 (5'-T CCA GAA TTC CAG TTT TGG GTT ACG-3') were the sequence encoded the amino acid 368 of PBP IB followed by the nucleotide sequence that provides the restriction site for EcoRI, enabling fusion to sequences encoding an enterokinase site and a histidine stretch, which allows rapid purification of the protein on an Ni affinity column (cf. section 6.2 below).

A NcoI-EcoRI fragment was cloned into the plasmid pARC0400 that was restricted with NcoI-EcoRI to obtain the recombinant plasmid pARC0392 (NCIMB 40659; Fig. 23). The recombinant plasmid was transformed into E.coli BL26 (DE3) and a protein of approximately 17 kDa was detected largely in the soluble fraction after induction with IPTG.

Along similar lines the minimum substrate binding region of PBP IA could be predicted to involve the stretch 62-220 in the wild type protein. Production of this protein as a fusion with a histidine stretch allows high efficiency affinity purification of the expressed product using the Ni^+ column. That the results will be similar to that obtained with truncated PBP IB can be anticipated.

EXAMPLE 6 6.1. N-terminal fusion of soluble E.coli PBP IA to glutathione-S-transferase Fusion of the ponAdel23 gene at its 5'-end in frame to sequences coding for glutathione-S-transferase was made as described in the following section.

The vector chosen for the fusion gene construction was pGEX-3X obtained from Pharmacia Biochemicals. In order to fuse the 5'-initiation ATG of -9504ι2, -38ponAdel23 in frame with the gene encoding glutathione-S-transferase a BamHI site was introduced using a PCR primer whose sequence included the sequence for the restriction enzyme EcoRI. The 5'-primer used was TG115: '-TCG AGG ATC CCC ATG GGC CTA TAC CGC TAC ATC G-3' BcoRI BamHI The 3'-primer used was TG-106, described in Section 4.2. The PCR amplified DNA Fragment A was digested with BamHI and Pstl and cloned into the BamHI - Pstl sites of the standard cloning vector pUC8 to obtain pARC0496. This Fragment A includes the N-terminal 102 amino acids of the PBP lAdel23 protein. A BnmHI - Mini (site present within the fragment A) 270 bp fragment obtained from Fragment A, a 2.2 kb M/uI - EcoRI fragment which includes the rest of the portion of the ponA gene obtained from pARC0490 (pARC0490 has the wild type ponA gene cloned into the Xbal - BamHI sites of the low copy vector pWKS29 (Fu Wang et al., 1991) facilitating the 3'-end of the ponA del 23 gene to be excised as an EcoRI fragment) and a EcoRI - BamHI cleaved pGEX-3X were ligated together and transformed into competent E.coli cells. Individual transformants were screened for recombinant plasmid and the plasmid with the expected structure was designated pARC0499 (NCIMB 40664; Fig. 24). The encoded fusion product on pARC0499 has the glutathione-S-transferase sequences at its C-terminus linked to PBP lAdel23 sequences via a Factor Xa cleavage recognition sequence.

Following induction with 1 mM IPTG, a fusion protein of expected size was found to be induced. The protein bound penicillin and was active in the transglycosylase assay. Following cell lysis by passing the suspension through a French press, the cell free supernatant fraction was prepared as detailed in Section 1.4. for the purification of PBP lAdel23. The supernatant fraction was passed through a Glutathione Sepharose® matrix 950412. 39— (Pharmacia Biochemicals) and the bound GST-PBP lAdell23 was eluted with glutathione. The eluted protein was found to be 80% homogeneous. Free glutathione was removed by dialysis and the GST-PBP lAdel 23 was cleaved with factor Xa.

PBP lAdel23 thus purified was found to be active in both penicillin binding and the transglycosylase reactions. 6.2. C-terminal fusion of soluble E.coli PBP 1A to histidine stretch Fusion of the ponAdel23 gene at its 3'-end in frame to sequences encoding a stretch of 6 histidines was made as described below.

In the first step the ponAdel23 gene was amplified using pBS98 DNA as template using the 5'-primer TG-115 (5'-TCG AGG ATC CCC ATG GGC CTA TAC CGC TAC ATC G-3') and the 3'-primer TTG-121 (5'-GTT AGA ATT CGA ACA ATT CCT GTG-3').

The 3'-primer introduced an EcoRI site at the 3' end of the ponAdel23 gene while eliminating the translation stop codon. The PCR amplified modified poziAdel23 gene fragment was digested with Pstl and EcoRI to release a 930 bp 5'-end fragment and ligated to Pstl-EcoRI digested pBR 329 to obtain the recombinant plasmid pARC0467.

In the next step, a double stranded synthetic oligonucleotide with sequences encoding the six histidines and the DNA sequence coding for amino acids recognised as the enterokinase cleavage site was synthesised and ligated to the newly created EcoRI site at the 3'-end of the ponAdel23 gene on pARC0467. The synthetic oligonucleotides used were TG-122: 850412. -40EcoRI '-AA TTC GAC GAC GAC GAC AAG CAC CAC CAC CAC CAC CAC TGA TAA G-3 ' ENTEROKINASE HISTIDINES and TG 123 (5'-GAT CCT TAT CAG TGG TGG TGG TGG TGG TGC TTG TCG TCG TCG TCG-3').

The plasmid pARC0467 was linearised with EcoRI and the synthetic double stranded oligonucleotide ligated. Following ligation a Psfl - BamHI (Fragment A) was released from the ligation mixture and cloned into the Pstl - BamHI sites of pARC0558 (Fig. 3), to obtain pARC0400 (NCIMB 40660; Fig. 25). The mutant ponAdel23 fusion gene thus encoded a protein with the PBP lAdel23 sequence fused to the amino acid sequence AspAsp-Asp-Asp-Lys fused to His-His-His-His-His-His at its C-terminus. The Asp-Asp-Asp-Asp-Lys sequence is recognised by the protease enterokinase and cleaves following the lysine residue. The six histidine residues confer on the protein the ability to bind to the metal nickel.

The recombinant plasmid pARC0400 was transformed in E.coli BL26(DE3) cells and induced under culture and temperature conditions identical to those used for the purification of PBP lAdel23. The cells were lysed by passing through a French press. The lysate was centrifuged at 10,000 rpm for 10 min. The supernatant obtained after low speed centrifugation was then spun at 200,000 x g for 45 min and the supernatant obtained represented the cytosolic fraction. This fraction contained the protein encoded by the fusion gene and the recombinant fusion protein was labelled PBP lAdel23EH. This protein PBP lAdel23EH bound [lz^I]cephradine and was also active in transglycosylase assay. The soluble fraction was passed through a Ni affinity column and bound protein eluted in batches with increasing concentrations of imidazole essentially following the procedure described in 'The Qia Expressionist obtained from QIAGEN Inc. 9259 Eton Avenue, Chateworth, CA 91311 USA. The majority of PBP -41950412 lAdel23EH eluted with 250 mM imidazole and was approximately 85% homogenous. It was the only cephradine binding protein eluted from the column. Thus the ability of fusion protein to bind to the Ni column can be easily exploited both for efficient purification and immobilisation of the active protein.

EXAMPLE 7 7.1. Use of cell extracts for enzyme assays and in screening The crude cell extract made according to Example 6 can be analyzed for the ability to bind penicillin by reacting with pHjampidllin prepared according to Hackenbeck (1983). To adapt the procedure to large-scale screening, a 96 well microtitre plate is used to contain the reactions and the assay is performed using a Beckman Biomek robot. Crude cell extract is a mixed with [ HJampicillin for 15 min at 37°C. The proteins in the reactionare are precipitated with TCA and collected on a glass filter, unbound ampidllin is washed off and filters counted in a sdntillation counter. Alternatively, autoradiography can be used to assay the degree of binding of ampidllin.

Based on the above method, a competitive assay can be used to assess the ability of test compounds to bind to the transpeptidase site of a PBP variant. In this assay, the test compound is exposed to the crude cell extract for 15 min prior to the addition of ampidllin. A positive result is indicated by a reduction in the amount of radioactivity present on the glass filter. 7.2. Use of soluble immobilised protein in screening Protein containing a histidine peptide which has been purified as described can be used for screening for compounds which inhibit transpeptidase b.850 412 -42activity or transglycosylase activity. The purified full length or truncated protein is immobilised onto agarose gel to which Ni(II) has been coupled. Aliquots of the beads containing immobilised protein are then transferred to the wells of a microtitre plate, test compounds are added to the plate and incubated before unbound test substance is washed free. Compounds which bind to the transpeptidase site of the bifunctional protein can be detected by adding [^HJampidllin to the reaction vessel and continuing essentially as described above. Alternatively monoclonal antibodies known to bind to the transpeptidase region can be used. Compounds which bind to the transglycosylase site can be assessed in a competitive assay by the use of monoclonal antibodies which bind to the transglycosylase region of the protein.

EXAMPLE 8 8.1. Production of monoclonal antibodies to PBP IA The protocol for the production of monoclonal antibodies (mAbs) was essentially that described in Antibodies - a laboratory manual (ed. Harlow David Lane, Cold Spring Harbor, USA). Purified membrane bound PBP IA was used as the immunogen. Balb-C mice, 6-8 weeks old were immunised with 50 pg of purified native PBP IA in Freunds Complete Adjuvant. A booster injection of 20 pg PBP IA in incomplete Freunds adjuvant was given intraperitonially. Two weeks later the presence of serum antibodies was checked by ELISA using PBP 1A as the coated antigen. Mice with circulating antibodies were immunised intraperitonially daily for 4 days with 20 pg of PBP IA in saline and the mice sacrificed for isolating splenocytes for generating fusions.

The myeloma cell line used in fusion experiments was Sp 2/0-Ag 14 and these cells were fused with splenocytes from immunised mice at an ratio of : 1. Fusion was carried out using standard protocols and antibody r -43production from the clones was monitored by ELISA against PBP IA when the cells were > 90% confluent. high producing clones were expanded to 24 well plates and the secreted 5 antibody characterised using the following screens: (1) ELISA against membrane bound form of PBP IA; (2) ELISA against soluble form of PBP lAdel 23; (3) Dot blot analysis against membrane bound PBP IA to eliminate monoclonals reacting with the detergent solubilised purified PBP IA protein only due to changes in the configuration during purification; and (4) ELISA against membrane bound form of PBP IB.

Based on these screens, a panel of 5 secreting clones were selected and subcloned twice to ensure monoclonality. Ascites with these hybridoma clones were raised following standard procedures and IgG was purified from these ascites fluids, using Protein G-Sepharose® affinity chromatography as recommended by the manufacturers of Protein G-Sepharose (Pharmacia Biochemicals).

These purified antibodies react specifically with PBP IA in both the membrane bound and the soluble forms in ELISA, Dot blots and in Western blotting. Clones were obtained by a cloning procedure employing 3 cells / well. To ensure the monoclonality these clones were subcloned into 96 well microtitre plates by limiting dilution at 1 cell / well. The wells receiving one cell were carefully confirmed under the microscope and allowed to grow with macrophage feeder layers so as to obtain progeny from a single hybrid cell. Following sub-cloning the secretion of mAbs to PBP IA was again assayed in ELISA using full length PBP IA. Finally two clones from each parent hybridoma were selected and one of them was expanded as ascites in pristine primed Balb/c mice. All the five clones adapted to grow in peritoneal cavities and produced ascitic mAbs. .950412 -44The ascitic mAbs were titrated against purified PBP IA in ELISA. All the ascitic mAbs had a titre of > 5 x 10^ in ELISA and recognised full length protein in western immunoblots. The ascitic mAbs were purified by protein-G affinity columns.

The immunoglobulin isotype of mAbs was determined by mouse Ig isotype by ELISA using a kit obtained from Sigma chemicals USA. Four of the monoclonals belonged to IgGl and one belonged to IgG2a immunoglobulin isotype.

Further characterization of mAbs was done by using full length membrane bound PBP 1A/1B in western blots. In addition the transglycosylase (TG) and transpeptidase (TP) domain specificity of mAbs was determined by using various truncated forms of the membrane-bound N-terminal of PBP IA, N-terminal of PBP IB and C-terminal of PBP IB in Western immunoblots. Various full length and truncated membrane bound PBPs were expressed and the prepared membrane fractions were resolved on a SDS-PAGE. The proteins were transferred onto nitrocellulose membranes and subjected to western blot analysis using polyclonal E.coli PBP IA antibodies and monoclonal antibodies.

Assessment of the penicillin binding inhibitory potential of the mAbs was determined essentially following the protocol described by den Blaauwen et al. (1990). The protein-G affinity purified mAbs was preincubated with PBP IA followed by addition of pH] benzyl penicillin or [^^^Ilcephradine. Two of the mAbs competitively inhibited binding of the radiolabelled penicillin to PBP IA.

Monoclonal antibodies specific for the TG domain of PBP IA have been obtained by screening the secreted antibody of the original hybridoma clones to react with the protein representing the N-terminal 434 amino adds of PBP IA in western blots. Antibody from clone TG-2 reacted with •«-95 04 12 -45the N-terminal truncated 434 amino acid analogue of PBP 1A but also inhibited (>80% inhibition) the transglycosylase activity of PBP 1A. This indicates that the antibody recognises sequences in the protein which are involved in (a) binding of the substrate; (b) catalysing the enzymic action; or (c) altering conformation of the protein allosterically. In either of the three possibilities, identification of compounds competing for the binding of TG-2 to PBP 1A would represent molecules interacting with identical sequences on PBP 1A. Thus the competitive binding assay could be used as a screening assay for the identification of the TG inhibitory compound.

BRIEF DESCRIPTION OF THE DRAWINGS Figure 1 Hydropathicity profile of E.coli PBP 1A. The figure shows in expanded form the hydropathicity pattern of the N-terminal 55 amino acids of PBP 1A.

Figure 2 Schematic representation of the T7 translation fusion expression vector pARC038. - vector sequences ------ gene conferring kanamycin resistance Kmr, gene encoding the lactose repressor (lac I ), origin of replication (ori), T7 lac operator promoter, T7 phage terminator.

The direction of transcription of the different genes are shown by arrows. Relevant restriction enzyme sites are shown. Numbers next to the restriction site represent the nucleotide position taking the nucleotide at the upper twelve o'clock-position as zero. -46Figure 3 Schematic representation of the vector pARC0558 encoding soluble PBP lAdel 23 of E.coli. - vector sequences = mutant gene encoding PBP lAdel23, kanamycin resistance Kmr, lactose repressor (lac Iq) and the origin of replication ori.

Figure 4 Expression of soluble PBP lAdel23. Panel A represents the autoradiogram of the [^^^I]cephra.dine binding profile of the uninduced and induced cultures of E.coli BL 26 (DE3) harbouring pARC0558. Panel B represents the Coomassie Brilliant Blue staining protein profile of the same uninduced and induced cells. Lane (1): uninduced cytosol fraction; (2): uninduced membrane fraction; (3): induced cytosol fraction; (4): induced membrane fraction; (M): molecular weight markers.

Figure 5 SDS-PAGE pattern of purified PBP lAdel23. Panel A: Coomassie blue staining. Panel B: [ I]cephradine binding protein profile. Lanes (1): E.coli BL 26(DE3)/pARC0558 cytosolic fraction (200,000g supernatant); (2): 30% Ammonium sulphate supernatant fraction; (3): 30% Ammonium sulphate pellet fraction; (4): Cephradine affigel breakthrough fraction; (5): Molecular weight markers; (6-8): Cephradine affigel eluate.

Figure 6 Transglycosylase activity profile of wild type PBP 1A and mutant PBP lAdel23 using purified proteins. (a—A) represents activity of soluble PBP lAdel23; (·-·) represents activity of membrane bound PBP 1A solubilised with octyl-3-glucoside. X-axis represents the concentration of the proteins used in pg. Y-axis represents the quantities of peptidoglycan formed. 0 411 —47— Figure 7 Hydropathic!ty profile of E.coli PBP IB. The figure represents the expanded hydropathicity profile of the N-terminal 150 amino acids of E.coli PBP IB.

Figure 8 Schematic representation of the cloning of the soluble transglycosylase domain of E.coli PBP IB. - vector sequences == sequences encoding ponB gene fragments and β-lactamase The Ncol-Nrul fragment encoding the N-terminal 64 amino acids of PBP IB was cloned into the Ncol-EcoRV sites of pARC0534 to obtain the plasmid pARC0551. This recombinant plasmid harbours the gene encoding amino acid 1 to 480 of PBP IB with internal deletion of amino acid 65 to 87.

Figure 9 Schematic representation of pARC0559 encoding soluble PBP IB. - vector sequences = sequences of the mutant ponB gene encoding the soluble form of PBP IB (solPBP IB), lactose repressor (lac I ), kanamycin resistance 4 (Kmr) and the origin of replication (ori).

Arrows represent direction of transcription of the genes.

Figure 10 Purification of soluble PBP IB. Panel A: SDS-PAGE, Coomassie blue staining of the different fractions. Panel B: [^^IJampicillin binding profile of the same fractions. Lanes (1) and (2): Cytosolic fraction of E.coli BL 26(DE3)/pARC0559 induced cells; (3): Breakthrough fraction of AmpicillinAffigel column; (4): Molecular weight markers; (5) and (6): Eluted fraction from the Ampicillin-Affigel column.

Figure 11 ^950412 -48Hydropathicity profile of S.pneumoniae PBP IA. The figure shows the expanded profile of the hydropathicity profile of the N-terminal 100 amino acids of S.pneumoniae PBP IA.

Figure 12 Schematic representation of the plasmid pARC0512 encoding soluble form of S.pneumoniae PBP IA. - represents vector sequences = represents sequences of the gene encoding soluble PBP IA of S.pneumoniae (sPBP IA), kanamycin resistance Kmr and the origin of replication (ori).

Figure 13 Penicillin binding profile of soluble S.pneumoniae PBP IA. Host: E.coli BL 21(DE3)/pARC0512. Panel A: Coomassie Blue staining. Panel B: In vivo labelling with [3H]benzyl penicillin followed by SDS-PAGE. Lanes (1) and (2) : Cytosolic fraction of cells induced at 22°C for 2 h and 20 h respectively; (3) : Cytosolic fraction of cells induced at 30°C for 2 h; (4): Cytosolic fraction of cells induced at 37°C for 2h; (5): Molecular weight markers.

Figure 14 Amino acid alignment of conserved regions of the transglycosylase domain of high molecular weight penicillin binding proteins. The figure compares the conserved residues of the Regions 1, 2 and 3 among E.1A (E.coli PBP ΙΑ), E.1B (E.coli PBP IB), S.1A (S.pneumoniae PBP IA), and H.inf (Haemophilus influenzae PBP IA). (*) indicates identical amino acid residues.

Figure 15 Analysis of membrane protein of E.coli cells harbouring plasmids with genes encoding mutant PBP IB. Panel A: [3H]benzyl penicillin binding profile. Panel B: Western blotting with anti-PBP IB sera. Lanes (1): Molecular weight markers; (2): Membrane fraction of E.coli JM 101/pBS96 (950412 —49— cells; (3): Membrane fraction of E.coli 900521 ponBiSpc1 cells (This host lacks chromosomal encoded PBP IB); (4): Membrane fraction of E.coli 900521 ponB:spc./pARC0438 cells; (5): Membrane fraction of E.coli 900521 ponB:spc/pARC0469; (6): Membrane fraction of E.coli 900521 po«B:spc/pARC0468.

Figure 16 Schematic representation of plasmid pARC0462 encoding wild type PBP IB: - vector sequences == sequences of the ponB gene, replication origin (ori), chloramphenicol acetyl transferase (cmr) and portions of the lac Z multiple cloning site.

Figure 17 Schematic representation of plasmid pARC0463 encoding mutant ponB gene. - vector sequences = sequences of mutant the ponB gene encoding PBP lBdel8 amino acids, replication origin (ori), chloramphenicol acetyl transferase (cmr) and portions of the lac Z multiple cloning site.

Figure 18 Schematic representation of plasmid pARC0470 encoding mutant ponB gene. - vector sequences = sequences of mutant the ponB gene encoding PBP IB Q271-272 ' ^271-272' rephcation origin (ori), chloramphenicol acetyl transferase (cmr) and portions of the lac Z multiple cloning site.

Figure 19 Schematic representation of pARC0571 harbouring mutant ponA gene. -50- vector sequences ·· = sequences of mutant ponA gene (PBP 1A QQ-AA), kanamycin resistance Kmr origin of replication (ori).

Figure 20 1 OR [ IJPenicillin binding protein profile of wild type and mutant E.coli PBP 1A. Lane (1): E.coli AMA 1004 ponB.spc//pBS 98 (w.t. ponA); (2): E.coli BL21 (DE3) porc B:spcr/pARC0570 (w.t. ponA); (3): E.coli AMA 1004 del ponA/pARC0571 (QQ-AA ponA); (4): E.coli AMA 1004 del ponA/pBS 98 (w.t. ponA); (5): Molecular weight markers.

Figure 21 Schematic representation of plasmid pARC0592. - vector sequences —-= sequences of truncated ponB gene encoding for the N-terminal 553 amino acids of PBP IB (hinge IB), kanamycin resistance (Kmr) and origin of replication (ori) Figure 22 Schematic representation of plasmid pARC0593. - vector sequences = - = sequences of mutant truncated ponB gene encoding a soluble form of the truncated N-terminal 553 amino acids of PBP IB (soluble hinge IB), kanamycin resistance Kmr and origin of replication (ori).

Figure 23 Schematic representation of plasmid pARC0392. - vector sequences =..... sequences of mutant gene encoding truncated fragment of PBP IB protein, representing amino adds 210-368 sequences fused in frame at its 3'-end to sequences encoding a enterokinase site followed by a —51 stretch of 6 histidines, kanamycin resistance Kmr and origin of replication (ori).

Figure 24 Schematic representation of plasmid pARC0499. - vector sequences • sequences of mutant ponAdel23 gene fused at its 5'-end in frame to sequences encoding Glutathione-S-transferase encoding sequences, β-lactamase ampr and origin of replication (ori).

Figure 25 Schematic representation of plasmid pARC0400. - vector sequences — sequences of mutant ponAdel23 sequences fused in frame at its 3'end to sequences encoding a enterokinase site followed by a stretch of 6 histidines, kanamycin resistance Kmr and origin of replication (ori).

REFERENCES Balganesh, T.S. and Lacks, S. (1984): Gene 29, 221-230 den Blaauwen, T. et al. (1990): J. Bact. 172, 63-70 Broome-Smith, J.K. et al. (1985): Eur. J. Biochem. 147, 437-446 Covarrubias, L. & Bolivar, F. (1982): Gene 17, 79-89 Edelman, A. et al. (1987): Molecular Microbiology 1, 101-106 Fu Wang, R. and Kushner, S.R. (1991): Gene 100, 195-199 Hackenbeck et al. (eds.) The target of penicillin. W. de Gruyter publications, Berlin/New York 1983.

Hamilton, C.A. et al. (1989): J. Bacteriol. 171, 4617-4622 Heijenoort, Y. van, et al. (1978): FEBS Letters 89, 141-144 -52Heijenoort, Y. van, et al. (1992): J. Bacteriol. 174, 3549-3557 Ishino, F. et al. (1980): Biochem. Biophys. Res. Comm. 97, 287- 293 Kunkel, T.A. (1985): Proc. Natl. Acad. Sci. U.S.A. 82, 488-492 Kyte and Doolittle (1982): J. Mol. Biol. 157, 105-132 Lacks, S.A. (1968): Genetics 60, 685-706 Martin, C. et al. (1992): J. Bacteriol. 174, 4517-4523 Miller, J.H. (ed.) (1972): Experiments in Molecular Genetics. Cold Spring Harbor Publications.

Nakagawa, J.S. et al. (1984): J. Biol. Chem. 259, 13937-13946 Page, W.J. et al. (1982): J. Bacteriol. 151, 237-242 Rojo et.al. (1984): J. Antibiotics. 37, 389-393 Sambrook, J., Fritsch, E.F. and Maniatis, T. (1989): Molecular Cloning: A laboratory manual, 2nd Ed., Cold Spring Harbor Laboratory, Cold Spring Harbor, NY Sanger, et al. (1977): Proc. Natl. Acad. Sci. U.S.A. 74, 5463-5467 Spratt, B.S. (1977): Eur. J. Biochem. 72, 341-352 Studier, F.W. et al. (1990): Methods in Enzymology 185, 61-89 Tomb, E et al. (1991): Gene 104, 1-10 Yousif, S.Y. et al. (1985): J. Gen. Microbiol. 131, 2839-2845

Claims (30)

1. A polypeptide which is a water-soluble active derivative of a bacterial bifunctional penicillin binding protein, said penicillin binding protein being bound to the cell membrane when expressed in a bacterial cell and being capable of exhibiting both transglycosylase and transpeptidase activities and said derivative lacking a membrane anchoring sequence but retaining the capability to exhibit one or both of said enzymic activities.

2. A polypeptide according to claim 1 which amino acid sequence is identical to, or substantially similar to, SEQ ID NO: 2, 4, 6, 12 or 13 in the Sequence Listing.

3. A polypeptide which is a transglycosylase deficient derivative of a bacterial bifunctional penicillin binding protein, said penicillin binding protein being bound to the cell membrane when expressed in a bacterial cell and being capable of exhibiting both transglycosylase and transpeptidase activities and said derivative lacking the capability to exhibit transglycosylase activity but retaining the capability to exhibit transpeptidase activity.

4. A polypeptide according to claim 3 wherein the said derivative is lacking transglycosylase activity because of a mutation or deletion in the second conserved region of the gene coding for said polypeptide.

5. A polypeptide according to claim 3 which amino acid sequence is identical to, or substantially similar to, SEQ ID NO: 7, 8, 9, or 10 in the Sequence Listing. -876. A polypeptide according to claim 1 or 3 wherein the bacterial cell is an Escherichia coli cell or a Streptococcus pneumoniae cell.

6. 7. A polypeptide comprising (a) a first polypeptide according to claim 1 or 3; and (b) an additional polypeptide which allows binding to an affinity matrix; there being a cleavage site between said polypeptides.

7. 8. A polypeptide according to claim 7 wherein the additional polypeptide is glutathione-S-transferase or a polypeptide substantially similar to glutathione-S-transferase.

8. 9. A polypeptide according to claim 7 wherein the additional polypeptide is a polypeptide rich in histidine residues.

9. 10. An isolated and purified DNA molecule which has a nucleotide sequence coding for a polypeptide according to claim 1, 3 or 7.

10. 11. A DNA molecule according to claim 10, which nucleotide sequence is identical to, or substantially similar to, SEQ ID NO: 1, 3 or 5 in the Sequence Listing.

11. 12. A replicable expression vector which carries and is capable of mediating the expression of a DNA molecule according to claim 10.

12. 13. A vector according to claim 12 which is the vector pARC0558 (NCIMB No. 40666), pARC0559 (NCIMB No. 40667), pARC0512 (NCIMB No. 40665), pARC0438 (NCIMB No. 40661), pARC0468 (NCIMB No. 40662), pARC0469 (NCIMB No. 40663), 9 5 0 * 1 2 .’ -88pARC0571 (NCIMB No. 40668), pARC0593 (NCIMB No. 40670), pARC0392 (NCIMB No. 40659), pARC0499 (NCIMB No. 40664), or pARC0400 (NCIMB No. 40660).

13. 14. A cell harbouring a vector according to claim 12.

14. 15. A process for production of a polypeptide which is a derivative of penicillin binding protein, comprising growing a cell according to claim 14 in or on a culture medium for expression of the polypeptide and optionally recovering the polypeptide.

15. 16. A process for the production of a water soluble polypeptide according to claim 1 which comprises culturing Escherichia coli cells harbouring an expression vector wherein a DNA coding sequence for said polypeptide is under the control of an isopropyl thiogalactoside (IPTG) - inducible promoter, said culturing being carried out in the presence of a sub-optimal concentration of IPTG for induction of the said promoter and at a temperature in the range of 20 to 24°C.

16. 17. A method of identifying an antibody capable of binding a bacterial bifunctional penicillin binding protein which includes the step of employing a polypeptide according to claim 1 or 3 in an antibody binding assay and selecting antibodies that bind to the polypeptide.

17. 18. A method of assaying for compounds which bind to a penicillin binding protein, said method comprising (a) contacting a polypeptide according to claim 1, 3 or 7 with a compound to be investigated; and (b) detecting whether said compound binds to the penicillin binding protein. 950412. -8919. A method of assaying for compounds which bind to a penicillin binding protein, said method comprising (a) culturing cells according to claim 14; (b) lysing the said cells and isolating the crude cell extract; (c) exposing the said cell extract to potential inhibitors of a penicillin binding protein; (d) introducing an agent, known to bind a penicillin binding protein, to the said cell extract; (e) removing the unbound fraction of said agent; and (f) assaying the presence of said agent remaining in the cell extract.

18. 20. A method of assaying for compounds which bind to a penicillin binding protein, said method comprising (a) exposing a polypeptide according to claim 1, 3 or 7, immobilised on a solid support, to a potential inhibitor of a penicillin binding protein; (b) exposing an agent, known to bind a penicillin binding protein, to the immobilised polypeptide; (c) removing the unbound fraction of said agent; and (d) assaying the presence of said agent bound to the immobilised polypeptide.

19. 21. A method of assaying for compounds which bind to a penicillin binding protein, said method comprising (a) exposing a polypeptide according to claim 1, 3 or 7 to a potential inhibitor of a penicillin binding protein; (b) exposing the said polypeptide to an agent, known to bind a penicillin binding protein, which agent is immobilised on a solid support; and (c) assaying the presence of polypeptide bound to the immobilised agent.

20. 22. A method of assaying for compounds which bind to the transglycosylase domain of a penicillin binding protein, said method comprising (a) exposing the transglycosylase domain of a polypeptide according to claim 1 or 7, said polypeptide being immobilised on a solid support, to a potential inhibitor of the transglycosylase activity of a penicillin binding protein; (b) exposing -90 95 04 11 an agent, known to bind the transglycosylase domain of a penicillin binding protein, to the immobilised polypeptide; (c) removing the unbound fraction of said agent; and (d) assaying the presence of said agent bound to the immobilised polypeptide. A method of assaying for compounds which bind to the transglycosylase domain of a penicillin binding protein, said method comprising (a) exposing the transglycosylase domain of a polypeptide according to claim 1 or 7 to a potential inhibitor of a penicillin binding protein; (b) exposing the said polypeptide to an agent, known to bind to the transglycosylase domain of a penicillin binding protein, which agent is immobilised on a solid support; and (c) assaying the presence of polypeptide bound to the immobilised agent. A method according to any one of claims 19 to 23 wherein the agent known to bind a penicillin binding protein is a monoclonal antibody. A method according to any one of claims 19 to 23 wherein the agent known to bind a penicillin binding protein is a labelled antibiotic compound. A method of determining the protein structure of a penicillin binding protein, characterized in that a polypeptide according to claim 1 or 3 is utilized in X-ray crystallography. 950*12.

21. 27. A polypeptide according to claim 1, substantially as hereinbefore described and exemplified.

22. 28. A polypeptide according to claim 3, substantially as hereinbefore described and exemplified.

23. 29. An isolated and purified DNA molecule according to claim 10, substantially as hereinbefore described.

24. 30. A replicable expression vector according to claim 12, substantially as hereinbefore described.

25. 31. A process according to claim 15 for production of a polypeptide, substantially as hereinbefore described and exemplified.

26. 32. A process according to claim 16 for the production of a water soluble polypeptide, substantially as hereinbefore described and exemplified.

27. 33. A method according to claim 17 of identifying an antibody, substantially as hereinbefore described and exemplified.

28. 34. A method according to any one of claims 18-21 of assaying for compounds which bind to a penicillin binding protein, substantially as hereinbefore described and exemplified.

29. 35. A method according to claim 22 or 23 of assaying for compounds which bind to the transglycosylase domain of a penicillin binding protein, substantially as hereinbefore described and exemplified.

30. 36. A method according to claim 26 of determining the protein structure of a penicillin binding protein, substantially as hereinbefore described.