WO2001098500A1 - Receptor binding molecules - Google Patents

Abstract

Disclosed is a receptor binding molecule comprising an incomplete portion of a colicin R domain, the incomplete portion comprising fewer than 300 amino acid residues of the colicin R domain yet nevertheless retaining the receptor binding activity of the complete R domain, and comprising at least amino acid residues 343-418 of colicin E9 or the equivalent residues of the R domain of another colicin, the equivalent residues being determined by sequence alignment using the program Megalign, with the alignment being maximised by the CLUSTAL method.

Description

RECEPTOR BINDING MOLECULES

Fiel of the Invention

This invention relates to receptor binding molecules, methods of making the molecules, and various uses thereof.

Background of the Invention

Colicins are plasmid-encoded antibacterial protein toxins produced by strains of Escherichia coli and closely related bacteria. They are classified into groups according to the cell surface receptor to which they bind in sensitive E. coli cells; for example E colicins bind to the BtuB receptor which is an essential component of the high affinity vitamin B₁ transport system (James et al, 1996 Microbiology 142, 1569-1580). Killing of E. coli cells by E colicins requires three stages: receptor-binding, translocation and cytotoxicity. The E colicin proteins have three functional domains, each of which is implicated in one of the stages. The central R domain is responsible for receptor-binding activity, whilst the N-terminal translocation (T) domain mediates translocation, the process by which the cytotoxic (C) domain is transported from the receptor to the site of its cytotoxicity. Cytotoxicity of E colicins is due to one of three known mechanisms: (a) the formation of ion channels in the cytoplasmic membrane [e.g. colicin El] (Elkins et al, 1997 Structure 5, 443-458); or (b) an endonuclease activity which degrades DNA [e.g. colicins E2, E7, E8 and E9] (Kleanthous et al, 1999 (Nat. Struct. Biol. 6, 243-252); Ko et al, 1999 (Structure 7, 91-102); Schaller and Nomura, 1976 (Proc. Natl. Acad. Sci. USA 73, 3989-3993)); or (c) a ribonuclease activity which specifically cleaves 16S ribosomal RNA [e.g. colicins E3, E4 and E6] (Boon, 1971 (Proc. Natl. Acad. Sci. USA 68, 2421-2425); Senior and Holland, 1971 (Proc. Natl. Acad. Sci. USA 68, 959-963)), or specific tRNAs [e.g. colicin E5] (Ogawa et al, 1999 (Science 283, 2097-2100)).

E colicin producing bacteria are protected from killing by co-synthesizing immunity (Im) proteins which bind to the C-terminal cytotoxic domains and neutralize their activity. In the case of colicin El this process occurs in the cytoplasmic membrane, whilst in the case of the enzymatic E colicins (DNases and RNases) it occurs on synthesis in the cytoplasm and the resulting colicin/immunity protein complex is then released from the producing cells.

After binding to cell surface receptors colicins are then translocated to their site of toxicity by one of two translocation systems in E. coli cells (Lazdunski et al, 1998 J. Bacteriol. 180, 4993-5002). Group A colicins like the E colicins, colicin A and cloacin DF13 require the tø/-dependent translocation system (consisting of the proteins TolA, TolB, TolQ and TolR), whilst Group B colicins like colicin B and la use the ton-dependent translocation system. The process by which the cytotoxic C-terminal domains of E colicins are translocated to the cytoplasm of E. coli cells, across both the outer and the cytoplasmic membranes, is probably unique in bacteria (James et al, 1996 cited above) but is at present poorly understood.

The precise location of the R domain within colicins is unknown, but is generally shown in published work as being between residues 120-450.

Ohno-Iwashita & Imahori (1980 Biochemistry 19, 652-659) conducted some investigations on proteolytic fragments of colicins E2 and E3. They found one fragment, of 41kD, which retained receptor binding activity. However, the fragment (E3-BP) was rather large, being calculated in excess of 380 amino acid residues, and the authors were unable to indicate which particular residues within this large fragment were necessary for receptor binding. Further the fragment was obtained by proteolytic cleavage of the complete, pre-folded colicin and it was by no means obvious or predictable that smaller fragments, prepared by de novo expression from recombinant nucleic acid sequences, would adopt a similar conformation and retain receptor binding activity.

There are no reports of single mutations located in the R domain of the enzymatic E colicins which are defective in receptor binding, so there is no information as to which residues of the R domain are important for BtuB binding. Similarly, although there are a number of btuB mutants described which confer E colicin immunity to E. coli cells, these all appear to be stop-codon mutants and provide little help in understanding the nature of the R domain- BtuB protein-protein interaction. Summary of the Invention

The present inventors have surprisingly found that a less than complete portion of a colicin receptor binding (R) domain may retain the receptor binding activity of the full length R domain.

Accordingly in a first aspect the invention provides a receptor binding molecule comprising an incomplete portion of a colicin R domain, the incomplete portion comprising fewer than 300 (preferably fewer than 225, more preferably fewer than 200, most preferably fewer than 175) amino acid residues of the colicin R domain yet nevertheless retaining the receptor binding activity of the complete R domain, and comprising at least amino acid residues 343-418 of colicin E9 or the equivalent residues of the R domain of another colicin (especially another Group A colicin, preferably an E colicin), the equivalent residues being determined by sequence alignment using the program Megalign (from DNASTAR Inc, Madison WI, USA) with the alignment being maximised by the CLUSTAL method.

Typically the incomplete R domain will comprise between 70 and 100 amino acid residues, conveniently between 75 and 95 residues, preferably between 75 and 90 amino acid residues.

It will be appreciated that the invention does not extend to entire, naturally-occurring colicin molecules, as these will include a complete R domain. Thus, whilst a molecule in accordance with the invention may comprise other components, these will not be such as to provide a complete, naturally-occurring colicin molecule. Thus, for example, the invention may in one embodiment provide an isolated portion of an incomplete R domain, and in another embodiment may comprise an incomplete R domain joined to other amino acid sequences which may comprise sequences of a colicin, but the molecule will not comprise an entire colicin R domain. It is highly preferred that the receptor binding molecule of the invention is prepared by expression from a recombinant nucleic acid rather than by, say, proteolytic cleavage of a colicin molecule and the invention explicitly encompasses a method of making a receptor binding molecule by means of expression thereof from a recombinant nucleic acid. Retention by molecules in accordance with the invention of the "binding activity" of the complete R domain is not to be understood as meaning that the incomplete R domain will necessarily possess all the receptor binding characteristics of the naturally-occurring R domain. For example, a molecule in accordance with the invention may well be altered (either due to changes within the R domain sequence, or as a result of the influence of additional functional moieties which may optionally be present in the molecule, as described below), so as to have a different (preferably increased) binding affinity for the receptor, and/or a slightly altered "fine specificity", although molecules according to the invention will still bind to the receptor recognised by the complete, naturally-occurring, R domain, which is what is intended by "retaining receptor binding activity", generally with a binding affinity at least 75 % of that displayed by the compete R domain.

By way of explanation, the present inventors have identified a specific portion of the colicin E9 R domain which is found to be sufficient to bind to the colicin E9 receptor, BtuB. This specific portion comprises amino acid residues 343-418.

It will be apparent to those skilled in the art that the incomplete portion of the colicin R domain need not comprise the --mino acid sequence identical to residues 343-418 of colicin E9. Firstly, there is a high degree of homology between certain colicins, so the benefits of the invention may also be obtained by using the equivalent portion of the R domain from other colicins. As defined above, the equivalent portion from other colicins may readily be identified by use of a computer program, such as Megalign (referred to above) which aligns amino acid sequences so as to enable identification of equivalent residues. In particular, the equivalent portion of other nuclease-type colicins may be used (such as that of DNase type colicin or E2 or E7, or RNase type colicins E3 or E6, the complete gene sequences of which are all known). It should be noted that the numbering of the residues of equivalent portions may be slightly different in other colicins. For example, the equivalent portion of the R domain from colicin E3 is residues 342-417.

Secondly, the person skilled in the art will appreciate that minor variants from the sequence of residues 343-418 of colicin E9 may be tolerated without disrupting or abolishing the receptor binding activity of the molecule. Thus, for example, certain amino acid residues may be altered or omitted, if desired. Typically no more than ten or so amino acid residues will be altered, and typically any such alterations will comprise conservative substitutions (i.e. substituting one residue for another with similar properties e.g. substituting leucine for isoleucine, or lysine for arginine).

With the benefit of the present disclosure it will be apparent to those skilled in the art how to make variant sequences and test them for retained receptor binding activity. For example, nucleic acid sequences encoding amino acid residue variants of 343-418 of colicin E9 may be prepared by in vitro synthesis, or by site directed or PCR-mediated mutagenesis, and the resulting sequence inserted into a suitable expression construct and expressed in a host cell to produce the encoded polypeptide, which may then be screened for receptor binding activity by one of the methods described herein. Alternatively the variant amino acid sequence may be expressed using one of the known bacteriophage display systems, and phage particles then screened for the ability to bind to the relevant receptor.

Conveniently, a molecule in accordance with the invention will comprise a portion of a colicin R domain which has binding specificity for BtuB. BtuB is a protein present in the outer membrane of E. coli and related bacteria, such as Salmonella spp. BtuB is required for uptake of vitamin B₁₂ by bacteria.

Other preferred receptor binding activities for molecules in accordance with the invention include specific binding to IutA, FepA or FhuA.

Whilst molecules which essentially consist of the incomplete R domain of a colicin may be useful, as described below, it will normally be preferred for the molecule to comprise one or more additional functional moieties. Such additional moieties may be chemically conjugated or joined in any acceptable manner to the incomplete R domain. Advantageously however, the other functional moiety or moieties will comprise amino acid residues, which can be joined to the incomplete R domain by peptide bonds. Typically, the molecule will comprise a chimeric fusion protein, encoded by a corresponding nucleic acid sequence. The additional functional moiety or moieties may be derived from any source and may have any desirable biological activity. The additional functional moieties may be added to the N and/or C terminal of the incomplete R domain.

Methods of making nucleic acid constructs encoding fusion proteins, and causing expression thereof in suitable host cells, are now well-known to those skilled in the art, such that substantially any desired fusion protein in accordance with the invention could be prepared by those of normal skill in the art with the benefit of the present disclosure.

By way of illustration, the molecule may comprise a moiety which facilitates expression in a host cell (especially a micro-organism, such as a bacterium, yeast, fungus or plant cell), and/or purification therefrom. For example, it is known to add a polyhistidine (HIS) tag, or other tag (e.g. glutathione - GST; maltose binding protein - MBP; or cellulose binding domain - CBD) to polypeptides to facilitate their purification. In addition, a cysteine-rich acidic peptide has been used successfully to facilitate expression in bacteria of the antimicrobial peptide buforin II (Lee et al, 1998 Protein Expression and Purification 12, 53-60; Lee et al, 1999 J. Microbiology and Biotechnology 9, 303-310). An identical or similar acidic peptide might usefully be employed to facilitate expression of molecules in accordance with the invention. Equally, it may be desired to add a sequence to direct the expressed molecule to a particular compartment of the host cell (e.g. a signal sequence to cause secretion). Signal sequences of this sort are well known to those skilled in the art.

One particular example of an additional functional moiety which may be included in a molecule in accordance with the invention is a labelling moiety. Such a labelling moiety may be any moiety which facilitates detection of the molecule e.g. a biological activity such as an enzyme activity (e.g. β-galactosidase, alkaline phosphatase, horseradish peroxidase), or a chromophore or fluorophore. A preferred fluorophore is green fluorescent protein (GFP).

In particular, because molecules in accordance with the invention retain the ability to bind to a particular receptor which, by definition, will include at least a portion exposed at or near the surface of a bacterium, it may be possible to target a functional moiety to bacterial cells expressing the relevant receptor. Such a technique may be especially useful if the relevant bacterial cells are present in a complex mixture (e.g. a biological sample), possibly containing other bacteria or microorganisms which do not express the relevant receptor.

Thus, for example, where the functional moiety comprises a labelling moiety (e.g. GFP), it will be possible specifically to label those bacterial cells present in a sample (possibly containing other bacteria) which express the relevant receptor. Equally, it will be possible to identify whether or not particular bacteria express the relevant receptor, by determining if they became labelled upon contact with molecules in accordance with the invention comprising a receptor binding portion of an R domain joined to a labelling moiety.

In particular it may be preferred to use a molecule in accordance with the invention to direct a substance having antimicrobial properties to a bacterial cell. In such an embodiment the functional moiety may be any molecule (preferably a peptide or polypeptide) which has a toxic or inhibitory effect on bacteria, especially on Gram negative bacteria. Such a moiety may be described as an antimicrobial moiety. Joimng the antimicrobial moiety to the receptor binding portion of the R domain will allow the antimicrobial moiety to become locally concentrated around the bacterial cell, such that the antimicrobial moiety has a higher effective concentration. Particularly preferred are those moieties which have a toxic mode of action involving the outer and/or cytoplasmic membrane of Gram negative bacteria (e.g. pore-forming or transporter-blocking moieties).

A number of anti-microbial peptides are known to those skilled in the art which could be readily expressed as part of a chimeric fusion protein with a receptor binding portion of a colicin R domain. A considerable number of suitable antimicrobial peptides are listed in US Patent No. 6,025,326. Other peptides are disclosed in the following US Patents: 6,015,941; 6,008,195; 5,994,308; 5,945,507; and 5,830,993. A specific example is the antimicrobial peptide buferin. Another example is lysostaphin, a 29kDa polypeptide which disrupts the cell wall of Staphylo coccus aureus. The C terminus of lysostaphin is thought to be the targeting domain so, by removing this and replacing it with the incomplete R domain of a colicin in accordance with the invention, it should be possible to re-target the molecule.

Alternatively, as achieved by the inventors, it is possible to replace the R domain of a Cloacin (such as Cloacin DF13) with an incomplete Colicin R domain in accordance with the invention, thereby retargeting the Cloacin from its natural IutA receptor, the resulting fusion protein having receptor binding specificity for BtuB.

Thus, in some embodiments (as mentioned above), the incomplete R domain can be 'transplanted' into other molecules to alter their receptor binding specificity. However, in other embodiments, such transplanting might result in expansion of the receptor binding specificity, rather than its alteration. For example, the inventors have also made fusion proteins comprising the T and R domains of a Cloacin and the R and C domains of Colicin E9. The resulting chimera retained the ability to bind to both IutA (the receptor for the Cloacin R domain) and to BtuB (the receptor for the ColE9 R domain). Such an approach might be useful in enlarging the range of organisms to which a cytotoxic moiety or other functional moiety could be delivered.

In a second aspect the invention provides a nucleic acid encoding a receptor binding polypeptide comprising an incomplete portion of a colicin R domain, the incomplete portion comprising fewer than 300 (preferably fewer than 225, more preferably fewer than 200, most preferably fewer than 175) amino acid residues of the colicin R domain yet nevertheless retaining the receptor binding specificity of the complete R domain and comprising at least --mino acid residues 343-418 of colicin E9 or the equivalent residues of the R domain of another colicin, the equivalent residues being determined by sequence alignment using the Megalign program (as described above). Conveniently the nucleic acid encodes a polypeptide molecule, especially a fusion protein, in accordance with the first aspect of the invention.

The invention further provides a recombinant nucleic acid construct comprising the nucleic acid sequence defined above. Conveniently the construct will be a plasmid or other construct which comprises a promoter operably linked to the nucleic acid sequence, so as to enable expression thereof in a suitable host. Numerous such promoters are known to those skilled in the art.

The invention further provides a host cell (e.g. mammalian cell, plant cell, or microorganism, such as a bacterium, yeast, virus or fungal cell) comprising a nucleic acid sequence in accordance with the second aspect of the invention, especially a host cell which expresses a molecule in accordance with the first aspect of the invention as a result of the presence of the nucleic acid.

Molecules in accordance with the invention have a wide variety of uses. An incomplete portion of a colicin R domain has been found by the inventors to have antibacterial properties in its own right, even in the absence of any additional antimicrobial moiety. Thus, for example, the presence of high concentrations of the incomplete R domain can successfully compete for binding to the bacterial cell receptor, preventing its natural ligand from binding to the receptor. If binding of the natural ligand is essential for growth of the bacterium, then a sufficiently high concentration of the incomplete R domain can prevent binding of the natural ligand thereby inhibiting the growth of the microorganism. For example, where the incomplete R domain binds to the BtuB receptor, it can successfully inhibit uptake of vitamin B₁₂ by the bacterial cell. Salmonella spp, among others, have an absolute requirement for B₁₂, so blocking the BtuB receptor will kill or inhibit the growth of Salmonella spp. A concentration of about lμM of incomplete R domain may exert a lethal effect. It may be anticipated that fusion of the incomplete R domain with an additional antimicrobial moiety will result in a lethal effect at even lower concentrations.

Molecules in accordance with the invention could be used as a preservative, being incorporated into, or added to the surface of, or otherwise mixed with foodstuffs or other substances to prevent microbial growth causing contamination, degradation or spoilage thereof. Alternatively the molecules may find therapeutic application to prevent or treat microbial (especially bacterial) infections, in human or animal subjects. In such embodiments it may be desirable to use a chimeric molecule which includes an antimicrobial peptide or polypeptide as described above. In particular, molecules in accordance with the invention may be employed to treat or prevent gut infections, especially infections caused by Salmonella spp.

It will be apparent to those skilled in the art that the molecules may be administered to the substance or subject in any convenient manner. For example, the molecule may be administered as part of a complex mixture (e.g. a crude preparation obtained from a host cell culture), or may be applied in substantially pure form, and/or in substantially sterile form. A composition for injection into an animal (especially mammalian) subject will normally be in such a substantially pure, sterile form. Another method of administration is to administer a plurality of harmless host cells, each of which is expressing the relevant molecule, typically as a chimeric fusion protein. A large number of attenuated or harmless micro-organisms are known to those skilled in the art, (generally recognised as safe, or "GRAS" organisms) such as non-pathogenic strains of Lactobacillus or E. coli, or attenuated vaccine strains of

Salmonella spp (especially double "aro" mutants thereof). Preferably the organism will be secreting the molecule, which will desirably therefore include a signal peptide sequence recognised by the organism in question. Conveniently a culture containing the organisms will be added to a foodstuff or other substance to be protected or may be administered

(typically orally) to an animal subject.

The organism expressing a molecule in accordance with the invention might be administered as a vaccine, in order to protect against, or treat, a particular disease. More especially, a live culture of organisms may be administered as a "probiotic" to a subject, such as a human, or more normally a domesticated animal. Probiotics are cultures of safe microorganisms deliberately introduced into subjects (especially domesticated animals such as hens, cows and pigs) in order to compete with, and/or inhibit colonisation by, organisms which might otherwise infect and cause disease in the domesticated animal, or cause disease in humans when products of the domesticated animal (e.g. eggs, milk, meat) are consumed. Methods of growing up suitable cultures and inoculating them into appropriate animal subjects are well known to those skilled in the art.

In a further aspect, the invention provides for a method of specifically separating certain bacteria from a complex mixture. Thus, for example, a molecule in accordance with the invention will generally have a high binding affinity for a surface receptor present on bacteria of interest. By contacting the molecule with the sample, bacterial cells of interest can be caused specifically to adhere to the molecule. If the molecule is then removed in some way from the sample, the adhered bacterial cells will also be removed. Thus, for example, if the molecule in accordance with the invention comprises a member of a specific binding pair of some sort (e.g. biotin), it can be removed by contacting the sample with the corresponding member of the binding pair (i.e. streptavidin). More conveniently, the molecule in accordance with the invention will be immobilised on a solid surface (e.g. a latex or polysaccharide bead, e.g. an affinity chromatography column; a filter or membrane or similar), such that contacting the sample with the solid surface (carrying the molecule) will cause the relevant bacterial cells, expressing the appropriate receptor, to adhere

(indirectly) to the solid surface. Washing steps will remove non-specifically bound material, leaving the bacterial cells in substantially pure form on the solid surface.

The various aspects of the invention will now be further described by way of illustrative example and with reference to the drawings in which:

Figures 1, 2a and 2b are graphs of cell density (OD at 600nm) against time (minutes);

Figure 3 is a schematic representation of various molecules, some of which are in accordance with the invention;

Figure 4 is a graph of fluorescence (arbitrary units) against concentration of chimeric fusion Protein (μM);

Figures 5 a and 5b are graphs of fluorescence (arbitrary units) against concentration (μM) of Protein or vitamin B₁₂ respectively;

Figures 6a and 6b are graphs of cell density (OD₆₀onm) against time (minutes); and

Figure 7 is an amino acid sequence alignment, showing the amino acid sequence of residues 343-418 of colicin E9 (SEQ ID No.l) and the equivalent residues of a number of other colicins. The equivalent portions of colicins E3 and E6 are identical to that of E9, whilst there are 1 and 7 amino acid residue differences respectively in the equivalent portion of colicins E2 (SEQ ID No.2) and E7 (SEQ ID No.3); and

Figure 8 shows the amino acid sequence (SEQ ID No. 4) of a DF19/Col E9 fusion protein comprising an incomplete R domain in accordance with the invention.

EXAMPLES

Example 1

In an attempt to define the minimum E. colicin R domain the inventors first attempted to over-express a polypeptide ("E9TR") comprising residues 1-448 of colicin E9 containing both the T and R domains, but not the DNase domain. Plasmids, bacterial strains and media

E. coli JM83 (ara [wlac-proAB] rpsL Φ80lacZ ::M15) (Lawrence & James, 1984 Gene 29, 145-155) was used as a the host strain for cloning and mutagenesis. E. coli BL21(DΕ3) (Novagen) was used as the host strain for the expression vector pET21a (Novagen), which has a strong, IPTG-inducible T7 polymerase promoter and a C-terminal polyhistidine tag (His-tag) to facilitate the purification of over-expressed proteins. E. coli 113/3 is a metE mutant of the W strain of E. coli (ATCC 9637) (Davis and Mingioli, 1950 J. Bacteriol. 60, 17-28). E. coli RK5016 is a btuB 451 derivative of MC4100 (Heller et al, 1985 J. Bacteriol. 161, 896-903; Heller & Kadner 1985 J. Bacteriol. 161, 904-908). All cultures were routinely grown in Luria-Bertani (LB) broth, or on plates of LB agar, supplemented where required with ampicillin (100 μg ml^"1). All the aforementioned strains are publicly available from the inventors and/or the referenced sources.

Plasmid pNPl l, which encodes the colicin Ε9 structural gene (ceal) and the Im9 immunity gene (ceil), and plasmid pCS4, which encodes the ceal gene, with the introduced restriction sites Ndel at bp 1 and EcoRN at bp 490, together with the ceil gene with a C-terminal His- tag, under the control of an inducible T7 promoter, have both been previously described (Garinot-Schneider et al, 1997 Microbiology 143, 2931-2938). Plasmid pAGl is derived from pML261 and contains a 2.4 kb EcoRI-H dIII fragment that encodes the complete btuB gene in the vector pUC8 (Koster et al, 1991 J. Bacteriol. 173, 5639-5647). The pGFPωv plasmid was purchased from Clontech (UK) and the pΕT21a plasmid from Novagen.

DNA Manipulation

Restriction enzymes and T4 DNA ligase were purchased from Roche Pharmaceuticals, or New England Biolabs (UK) Ltd. Digestion of DNA with restriction endonucleases, electrophoresis of restriction fragments, ligation of DNA fragments and transformation into E. coli were carried out as previously described (Sambrook et al, 1989 Molecular Cloning: a Laboratory Manual, Cold Spring Harbor, NY: Cold Spring Harbor Laboratory).

PCR PCR reactions were carried out in 50 μl volumes using 1 unit of Taq polymerase with 30 cycles of 94°C for 1 min,-55°C for 1 min and 72°C for 1 min, using an Amplitron^® II thermal cycler.

PCR mutagenesis was used to introduce an Ndel restriction site at the start codon of colicin E9, and an Xhol restriction site after residue 448 just before the start of the DNase domain. The 1.3 kbp Ndel-Xhol PCR product was cloned into pET21a, resulting in the plasmid pCS7 in which the N-terminal 448 residues of the colicin E9 gene was fused with vector sequences encoding a Glu residue and a C-terminal His-tag. Induction of E. coli BL21 (DE3) pCS7 with IPTG allowed the over-expression and subsequent purification of a 47.4 kDa E9TR polypeptide by metalchelate chromatography. N-terminal truncations of the 448 residue his- tagged T + R domain polypeptide encoded by pCS7 were made by introducing an Nde I site, which included an in-frame ATG start codon, at various locations by means of PCR mutagenesis. The Nde l-Xhol fragments of the resulting plasmids were then sub-cloned back into pET21a.

The protocol for His-Tag protein purification was adapted from that provided in the manufacturer's (Novagen) user manual. Two litres of LB broth (100 μg ml^"1 ampicillin) was inoculated with 20 ml of an overnight culture of E. coli BL21 (DE3) containing the appropriate plasmid and incubated at 30°C with shaking. At an OD₆₀o of 0.6 over-expression was induced by adding IPTG to a final concentration of 1 mM. Growth was continued for a further 16 hours. The cultures were centrifuged for 15 min at 10 000 rpm. The cells were washed in 50 ml of 50 mM Tris-HCl, pH 8.0, and the cell pellet was stored at -80°C for 24 h. The cell pellet was resuspended in 100 ml ice cold binding buffer, and sonicated to break the cells. The post-centrifugation supernatant was filtered through a 0.45 micron membrane. A 10 ml Ni²⁺ column was charged with 50 ml charge buffer (50 mM NiSO₄), and then equilibrated with 20 ml binding buffer. The prepared cell extract was then loaded onto the column. The column was washed with 20 ml binding buffer, 20 ml wash/bind (1:1) buffer mixture (32.5 mM imidazole, 500 mM NaCl, 20 mM Tris-HCl, pH 7.5) and the His-tagged protein complex was eluted with lx elute buffer (1 M ididazole, 500 mM NaCl, 20 mM Tris-HCl, pH 7.5). The elution was monitored by Bradford assay (see below) of aliquots at 595nm and by ninning samples on SDS-PAGE. Protein containing fractions were dialysed exhaustively against 20 mM Tris-HCl, 100 mM NaCl, pH 7-9 depending on the isoelectric point of the protein. After dialysis the proteins were subjected to gel filtration on a Superdex S-75 26/60 column (Pharmacia LKB) in 20 mM Tris-HCl, lOOmM NaCl (pH 7.0 or 9.0 depending on the protein being purified).

Protein determinations

Protein fractions were concentrated to approximately 8-10 mg ml^"1 using an Amicon^® ultrafiltration cell (Amicon Ltd) and then lyophilised as 1 ml aliquots. Samples were either stored at -20°C until required, or resuspended in 1 ml of dH₂O prior to use. Protein concentrations were measured against a series of BSA standards using the Bradford protein assay at OD_595nπι, and expressed as average values from, at least, three independent replicates.

To determine if the truncated protein retained receptor binding, it was assayed for receptor binding activity using an in vivo competition assay, similar to that previously described (Ohno-Iwashita and Imahori, 1980 Biochemistry 19, 652-659). In brief, an overnight culture of sensitive E. coli JM83 (pUC18) cells was diluted 100 times into 100 ml of LB broth containing ampicillin, and grown at 37°C to an OD₆₀o of 0.1. Ten ml aliquots were then transferred to clean, sterile flasks and the cells were incubated with 1.2 μg of native ColΕ9/Im9 and either 1, 10 or 100-fold molar equivalents of mutant colicin at 37°C for at least 5 h. Cultures incubated with 1.95nM of native or mutant colicin only, or no colicin at all, were included as controls. The OD₆₀o of each culture was taken at 30 min intervals to monitor cell growth. Each competition assay was performed at least twice with each mutant ColE9/Im9 complex being tested. Typical results of the assay are shown in Figure 1.

The negative control (filled triangle) shows growth of the E. coli cells in the absence of added substance. The positive control (empty triangle) shows killing of the cells by added E9/Im9 complex, preventing growth of the culture. This killing could be competed out by increasing relative amounts of the 47.4kDa T/R domain polypeptide ("E9TR") (E9/Im9:E9TR of 1:1, 1:10 and 1:100 denoted by empty squares, filled squares, and empty circles respectively).

The E9TR polypeptide protected sensitive E. coli cells against killing by the native colicin E9/Im9 complex at slightly lower relative concentrations than a full size mutant ColE9/Im9 complex which is biologically inactive due to a H575A mutation in the DNase domain (Garinot-Schneider et al, 1996 J. Mol. Biol. 260, 731-742) (data not shown). The polypeptide containing the T+R domains of colicin E9 could therefore bind to the BtuB receptor. In an attempt to localize precisely the BtuB-binding domain, the inventors constructed deletion sub-clones in order to determine the minimum size polypeptide that could bind to BtuB.

Example 2

Defining the N-terminus of the R domain of colicin E9

Plasmid pCS4 has been previously described and consists of the ceal gene, with the introduced restriction sites Ndel at bp 1 and EcoRV at bp 490, together with the ceil gene with a C-terminal His-tag, under the control of an inducible T7 promoter (Garinot- Schneider et al, 1997, cited previously). On induction of E. coli BL21 (DE3) cells containing pCS4 with IPTG, the complex of colicin E9/Im9 can be readily purified using metal-chelate chromatography, due to the high affinity of Im9 for the colicin E9 DNase (Wallis et al, 1995 Biochemistry 34, 13743-13750), and the presence of the C-terminal His-tag in the Im9 protein. The inventors used pCS4 in an attempt to isolate mutants in the T or R domain which resulted in the loss of biological activity of colicin E9, using a hydroxylamine or PCR mutagenesis strategy described previously (Garinot-Schneider et al, 1996, cited previously). When attempting to isolate random mutants, located in either the Ndel-EcόRY or the EcoRY-SacI fragment of the ceal gene of pCS4, the inventors surprisingly obtained some mutants that encoded a polypeptide which was smaller than the control colicin E9 on SDS-PAGE. These were subsequently shown to be in-frame deletion mutants and proved of value in locating the N-terminal boundary of the R domain. The truncated proteins were purified by metal-chelate chromatography and then assayed for receptor-binding activity to detej-mine if they could bind to BtuB, as described above. The results are shown in Figures 2a and 2b (legend as for Figure 1). Figure 2a shows the results obtained with a polypeptide expressed as a result of an in-frame deletion which removes amino acid residues 164-343 ("Δ 164-343"). Figure 2b shows results for a mutant lacking residues 330-513 ("Δ 330-513"). The Figures clearly show that whilst in- frame deletion of amino-acids 164-343 of colicin E9 had no effect in the competition binding assay, deletion of amino-acids 330-513 abolished the receptor-binding activity of the mutant colicin protein. This strongly suggests that the N-terminal boundary of the R domain is localised downstream of residue 343.

Based on information gathered from the in-frame deletion sub-clones, a series of plasmid constructs were made by introducing an Ndel site, which included an ATG start codon, at various locations through the T and R domain of colicin E9 in pCS7. The resulting clones encoded N-terminal deletions of the T+R domain, as illustrated schematically in Figure 3. In Figure 3, the Sαcl and Ncol restriction sites are marked, R, T and D signify the receptor-binding, translocation and DΝase domains respectively. Phenotype (+ or -) refers to receptor-binding activity (as judged by the E9/Im9 competition assay). The indicated values for the molecular weight of the encoded polypeptides are calculated for the colicin-derived polypeptide and do not include the hexahistidine tag.

The smallest polypeptide encoded by one of these constructs which retained receptor binding activity was 12.6 kDa which comprised of the residues from Met343 to Glu448 (11.6 kDa, Fig. 3) plus a (« lkDa) hexahistidine tag. Ν-terminal deletions of this 12.6 kDa polypeptide, which resulted in the production of stable polypeptides missing a further 20, 41, 50 or 59 residues, were all inactive in the competition assay. The results of the competition binding assay with the purified His-tagged polypeptides therefore suggest that the Ν-terminal boundary of the R domain is located between residues 343 and 363.

Example 3

Defining the C-terminus of the R domain of colicin E9

Since the Ν-terminus of the R domain polypeptide had been localised to between residue

343 and 363, the inventors now sought to identify the C-teπninus of the BtuB-binding domain. Using site-directed mutagenesis by PCR, Xhoϊ sites were introduced at various points from 15 to 150 bp from the 3' end of the 1,340 bp Ndel - Xhόl fragment, of pCS7.

Cloning of the resulting truncated Ndel - Xhόl fragments into pET21a allowed the purification of stable, truncated T+R domain polypeptides which were missing the C- terminal 5, 10, 20, 30, 40 or 50 amino-acids, as His-tagged proteins. In vivo competition assays (as described above) with these polypeptides showed that removal of up to 30 amino-acids from the C-terminus had no effect on receptor-binding activity, removal of up to 40 residues reduced receptor binding activity by about 50%, whereas removal of 50 residues completely abolished receptor-binding activity (data not shown). These results suggest that a polypeptide which consists of residues Met 343 to Glu 418 of colicin E9 (8.4 kDa) is the smallest polypeptide that retains receptor binding activity. A plasmid expressing this polypeptide ("E9R") was constructed in pET21a and was shown to protect

E. coli cells in the competition assay (data not shown).

Example 4

A fluorescent BtuB binding assay

In order to complement in vivo receptor binding assay a more quantitative fluorescence method was developed using a chimeric ColE9::GFP protein, consisting of 497 residues of colicin E9 (the T and R domains and the N-terminus of the DNase domain) fused to GFPwv (green fluorescent protein). The construction and expression of this fusion protein was achieved by a series of genetic manipulations of the plasmids pNPll and pGFPttv. Firstly, the complete ceal gene was removed from pNPll using the restriction enzymes Pstl and EcøRI and cloned into complementary sites of pUC18 in which the H dIII site had been filled-in, resulting in plasmid pNPll-Η. The GFPwv gene was then removed from pGFP-w using the restriction enzymes EcøRI and H diπ and cloned into complementary sites of pNPll-Η to produce pNP79. The HinάTIl site in pNP79 is located in the DNase domain at nucleotide 1488 of the ceal gene, whilst the EcøRI site is located in the multiple cloning site of the vector. This produced an in-frame fusion between ColΕ9 and GFP at residue 497 of the ColE9 protein.

The resulting chimera, ColE9::GFP, could be expressed from pNP79 using mitomycin C induction, but only to relatively low levels. High level expression of ColE0: :GFP was later achieved by cloning the recombinant gene into the Ndel and Xhol sites of pET21a which carries the IPTG-inducible T7/αc promoter. Before this was possible, Ndel and Xhol sites internal to GFPuv were removed using PCR mutagenesis, and the stop codon of GFPuv was replaced with an Xhόl site to allow a one step purification of ColE9::GFP by metal chelate chromatography of the hexahistidine tagged chimeric protein.

The fluorescent receptor-binding assay was performed as follows: E. coli RK5016 (btuB) and E. coli RK5016 (pAGl) were grown overnight in 5 ml of lxMMA minimal medium (lx salts solution, 100 μg ml^"1 L-Met and L-Arg, 1 mM MgSO₄,

0.2% glucose) at 37°C with aeration. The cells were diluted 65 times into 50 ml of lx

MM A and grown to an OD^^ of 1.5 for approximately 5.5 h. The cells were collected by centrifugation at 10K rpm for 5 mins and resuspended in 10 ml of 20 mM Tris-HCl, pH 9.0. One ml aliquots were added to eppendorf tubes containing an aliquot of test protein(s), and mixed immediately by inverting the tube three times. The cells were collected by centrifugation for 40 s, washed in 1 ml of 20 mM Tris-HCl, pH 9.0 and then resuspended in 1.5 ml of 20 mM Tris-HCl, pH 9.0. Fluorescence measurements were made of the entire sample in 3 ml quartz cuvettes, using a Shimatzu RF-5000 spectrofluorophotometer at excitation (Εx) and emission (Εm) wavelengths of 395nm and

505nm, respectively, and Εx/Εm band widths of 1.5 and 3nm, respectively.

Test samples, comprising total cell protein extracts of cultures producing the relevant test polypeptide, were prepared as follows:

Standing overnight cultures were diluted 1 in 100 in 5 ml LB broth containing ampicillin and shaken at 37°C for 2 h (OD^ -0.2-0.4). At this point 2.5ml of culture was transferred to a second flask containing mitomycin C (0.5 μg ml^"1). About two hours after induction with mitomycin C, when the ODgoo was more than 1.0, 2 ml of the culture from each flask was centrifuged at 13,000 g for 30 sees. The cell pellets were then resuspended in 100 μl of 1 x SDS-loading buffer (25 mM Tris/HCl pH6.8, 1% SDS, 5% glycerol, 0.1% bromophenol blue) and were then boiled for 2 minutes before centrifugation. The supernatants (total cell proteins) we restored at -20°C until required.

Binding of the chimeric protein to E. coli cells over-expressing BtuB was found to be concentration dependent, as shown in Figure 4.

Figure 4 shows the results for three different assays on RK 5016 (pAGl) cells (over- expressing the BtuB receptor), as indicated by filled or empty triangles, and filled squares, and the results for negative control RK5016 cells, which carry a mutation in the chromosomal btuB gene, and therefore display a very low level of binding of the fluorescent chimeric polypeptide.

A time course experiment showed that maximal binding of the fluorescent protein was achieved within 1 min of its addition to E. coli RK5016 (pAGl) cells (data not shown). This incubation time was used in the subsequent competition experiments to observe the effect of unlabelled polypeptides on the binding of the fluorescent protein. Competition between the fluorescent ColΕ9::GFP polypeptide and various truncated forms of the ColE9 R domain, was measured as the loss of fluorescence that accompanied an increase in the concentration of competing polypeptide. The results for the minimal R domain ("E9R"), residues 343-418, are shown in Figure 5a, confirming the results of the in vivo competition assay.

It was also shown that vitamin B12 could compete with ColE9::GFP for binding to BtuB (see Figure 5b).

Example 5

As vitamin B₁₂ (Di Masi et al, 1973 J. Bacteriol. 115. 505-513) gains entry into E. coli cells after first binding to BtuB receptor molecules, protects E. coli cells against killing by E colicins (Cavard 1994 FEMS Microbiol. Lett. 116, 37-42) and inhibits binding of the chimeric, fluorescent protein (Fig. 5b), the inventors investigated whether the minimum R domain (E9R) could inhibit the growth of vitamin B₁₂-dependent E. coli cells. E. coli 113/3 is a metE mutant that is unable to make methionine unless supplied with vitamin B₁₂, and will therefore only grow in minimal media when supplied with either methionine or vitamin B₁₂ (Davis & Mingioli 1950 J. Bacteriol 60, 17-28).

E. coli 113/3 and E. coli 113/3 (pAGl) were grown overnight in 5 ml 1 x MMA minimal media supplemented with 1 nM vitamin B₁₂ at 37°C with aeration. The cells were diluted 1: 100 in 50 ml 1 x MMA supplemented with 1 nM B12 and various concentrations of Ε 9R domain. The cells were grown for approximately 6 h and measurements of growth were taken at 30 min intervals by estimating cell density at OD _mπm. The results for E. coli 113/3 or 113/3(pAGl) cells are shown in Figures 6a and 6b respectively. In each case, negative controls (no Ε9R) are shown by filled triangles, and various concentrations of added B_!2 (InM, 10, 100, 250 or 500nM) are indicated by empty squares, filled squares, empty circles, filled circles, or empty triangles respectively.

E. coli 113/3 requires approximately 1 nM of vitamin B₁₂ for optimum growth (data not shown), but its growth is inhibited if 1 μM of E9R is provided in the same medium (Fig. 6a). By varying the concentration of E9R, with the concentration of vitamin B₁₂ kept at InM, it was found that growth of E. coli 113/3 could be partially inhibited with lOnM E9R and completely inhibited with E9R concentrations above 25nM. To test whether the levels of E9R necessary for growth inhibition was affected if the cells were producing more BtuB molecules, E. coli 113/3 was transformed with pAGl and the experiment repeated. Optimum growth of E. coli 113/3 (pAGl) cells again occurred at 1 nM B₁₂ (data not shown). However, E9R could only inhibit growth when present at concentrations in excess of 25nM, and concentrations greater than lOOnM were required for complete inhibition of growth.

Example 6

Preparation of synthetic variants of the E9 minimum R domain.

Random mutagenesis of residues 389-419 of the R domain was achieved by incorporating a spiked oligonucleotide in a two step PCR. The spiked oligonucleotides consisted of 45 nucleotides and were synthesised in a procedure in which all four of the phosphoramidite solutions were "doped" with a 3% (w/v) mixture of the other three nucleotides, thus resulting in random mutation(s) in the oligonucleotides. It was used in a PCR with a reverse primer to amplify a product that contained one or more random nucleotide substitutions. This product was then used as a mega primer with a forward primer to amplify a second product containing flanking restriction sites. The mutated DNA was then cloned into pCS4 using the flanking restriction sites, and transformed into appropriate host cells. The colicin activity of the various mutant forms of E9R was then screened by a stab test assay, as described below.

Stab test assay

Test cultures were stabbed into LB agar plates containing ampicillin (100 μg ml^"1) and were incubated at 37°C overnight. The plates were exposed to chloroform vapour for 15 minutes to break the cells and were then overlaid with 4 ml molten 0.1% (w/v) non-nutrient agar containing 50 μl of an overnight culture of the colicin-sensitive indicator strain E. coli JM83

(pUC18). The plates were incubated at 30°C for at least six hours and then inspected for a clear zone of growth inhibition around the test culture. Each culture was tested independently at least three times. The radius of the zone of inhibition around each stab was measured for a semi-quantitative estimate of colicin activity.

Stab test screening resulted in the isolation of several plasmids encoding mutant colicin E9/Im9 complexes exhibiting either reduced, or zero, colicin activity in the stab test assay. DNA sequencing of one of the mutant plasmids (pNP269) revealed the presence of three mutations in the R domain (A394D, A398P, Q402H). When the mutant protein complex was used in the fluorescent binding assay, the loss of colicin activity was mirrored by a change in the ability of the mutant protein to inhibit BtuB binding by the ColE9::GFPuv fluorescent protein (data omitted for brevity). Since a mutant colicin E9/Im9 complex carrying the M370I and Q402H mutations exhibited full biological activity and BtuB binding (data not shown), this implies that the phenotype encoded by pNP269 is likely to be due to the A394D and/or the A398P mutation. A mutant colicin E9/Im9 complex containing the A394D mutation alone was intermediate in its ability to compete for BtuB binding between that of the control E9 Im9 complex and the triple mutant, and exhibited a significantly reduced colicin activity.

The method described above may, for example, be employed to prepare other mutant forms of the colicin R domain which may similarly be screened for retention (or otherwise) of the ability to bind BtuB.

Example 7

Fusion proteins comprising an incomplete colicin R domain.

This work is described in detail in a paper published by the inventors and their colleagues (Penfold et al., Molecular Microbiology 2000 38, 639-649).

The inventors constructed two plasmids, pNP277 and pNP293. Plasmid NP277 encodes a fusion protein in which the DNase domain of colicin E9 was fused to the T and R domains of cloacin DF13. Plasmid NP293 encodes a fusion protein in which an incomplete R domain (in accordance with the invention) and the DNase domain of colicin E9 was fused to the T domain of cloacin DF13. As discussed in detail in Penfold et al. (cited above), the killing activities of the respective chimeric colicins against cells expressing the BtuB or IutA receptors demonstrated that the 95 amino acid residues of colicin E9 present in the fusion protein encoded by pNP 293 (which include the incomplete R domain) confer BtuB receptor specificity (and also confirm that DNase-mediated killing can be delivered via the IutA receptor).

The inventors made a further construct (pNP297), directing the expression of a fusion protein comprising the T and R domains of cloacin DF13 fused to the R and DNase domains of colicin E9. The plasmid was prepared by engineering Sαcl sites downstream of the R domain of cloacin DF13 at nucleotide 1300, and upstream of the R domain of colicin E9 at nucleotide 988. It was then possible to cut and replace the T domain of ColE9 with the T and R domains of DF13 using the restriction fragments generated by a Ndel and Sαcl double digestion.

The resulting construct directed the expression of a chimeric fusion protein, the amino acid sequence of which is shown in Figure 8 (SEQ ID No. 4). Residues (Met) 1 to (Ala) 336 are from the Cloacin DF13 T domain. Residues (Glu) 337 to (Ala) 443 are from the cloacin DF13 R domain. Residues (Glu) 444 to (Ala) 552 are from the colicin E9 R domain, whilst residues (Met) 553 to (Lys) 686 are from the colicin E9 C domain.

The fusion protein was demonstrated to be capable of killing cells which expressed either one of the Iut A or BtuB receptors.

Claims

Claims 23

1. A receptor binding molecule comprising an incomplete portion of a colicin R domain, the incomplete portion comprising fewer than 300 amino acid residues of the colicin R domain yet nevertheless retaining the receptor binding activity of the complete R domain, and comprising at least amino acid residues 343-418 of colicin E9 or the equivalent residues of the R domain of another colicin, the equivalent residues being determined by sequence alignment using the program Megalign, with the alignment being maximised by the CLUSTAL method.

2. A molecule according to claim 1, comprising the incomplete portion of an R domain of a Group A colicin.

3. A molecule according to claim 1 or 2, comprising the incomplete portion of an R domain of an E colicm.

4. A fusion protein comprising a molecule in accordance with any one of claims 1, 2 or 3, fused to an additional functional moiety.

5. A molecule according to any one of the preceding claims, comprising an additional functional moiety which exhibits antimicrobial activity.

6. A molecule according to claim 5, wherein the additional functional moiety comprises an antimicrobial peptide.

7. A molecule according to any one of the preceding claims comprising amino acid residues 343-418 of colicin E9.

8. A recombinant nucleic acid sequence encoding a polypeptide in accordance with any one of the preceding claims.

9. A recombinant nucleic acid construct comprising a nucleic acid sequence according to claim 8, operably linked to a promoter sequence, so as to enable expression of the polypeptide encoded by the sequence in a suitable host.

10. A host cell comprising a recombinant nucleic acid sequence according to claim 8 or a recombinant nucleic acid construct according to claim 9.

11. A method of inhibiting the growth of a microorganism in or on a substance, the method comprising the step of causing to be present in and/or on the substance a molecule in accordance with any one of claims 1-7, the molecule having binding specificity for a receptor present on the microorganism whose growth is to be inhibited.

12. A method of inhibiting the growth of a microorganism in a human or animal subject, the method comprising the step of causing to be present in the subject a molecule in accordance with any one of claims 1-7, the molecule having binding specificity for a receptor present on the microorganism whose growth is to be inhibited.

13. A method according to claim 12, wherein the molecule is caused to be present by the administration to the subject of a recombinant nucleic acid sequence according to claim 8, or a recombinant nucleic acid construct according to claim 9, optionally within a host cell.

14. A probiotic or vaccine composition comprising live host cells, said host cells comprising a recombinant nucleic acid sequence according to claim 8 or a recombinant nucleic acid construct according to claim 9.

15. A pharmaceutical composition for administration to a human or animal subject, the composition comprising a molecule in accordance with any one of claims 1-7, together with a physiologically acceptable diluent, excipient or carrier.

16. Use of a molecule according to any one of claims 1-7 in the preparation of a medicament to treat a microbial infection in a human or animal subject.

17. Use of a host cell according to claim 10 in the preparation of a probiotic or vaccine composition to treat or prevent a microbial infection in a human or animal subject.

8. A method of making a molecule in accordance with any one of claims 1-7, the method comprising the steps of: preparing a nucleic acid sequence directing the expression of a polypeptide comprising a molecule according to any one of claims 1-7; introducing the nucleic acid sequence into a suitable host cell to cause expression of the sequence; and

(optionally) purifying the expressed polypeptide.