Vaccines for Plague

The present invention relates to novel vaccines for provision of protection against infection with the organism Yersinia pestis and to methods for administering these. Particularly provided are parenterally and orally active vaccines capable of offering protection against bubonic and pneumonic plague. particularly by induction of mucosal immunity in both humans and other animals.

Yersinia pestis is the highly virulent causative organism of plague in a wide range of animals, including man. Infection with such organisms results in a high rate of mortality. Studies have shown that the high virulence is due to a complex array of factors encoded by both the chromosome and three plasmids, including the Lcr genes, a fibrinolysin and a capsule.

Man is an occasional host in the natural cycle of the disease, and bubonic plague, characterised by the swelling of local lymph nodes, may occur following the bite of an infected flea. One of the complications of bubonic plague is secondary pneumonia, and in these cases the disease is readily transmitted between humans by airborne droplets. Plague is endemic in regions of North and South America, Africa, China and Asia (see Butler (1983) Plague and Other Yersinia Infections; Plenum Press, New York). Current outbreaks are believed to be part of the fourth world pandemic of the disease, with a clear need to protect individuals living or travelling in endemic areas, and laboratory workers handling the bacterium.

The current whole cell vaccines available for prevention of plague are highly heterogeneous. resulting in side effects which make them unsuitable for widespread use (eg Meyer et al (1974) J. Infect Dis 129 supp 13-18 and 85-120: Marshall et al (1974) ibid supp 19-25).

One current vaccine for plague is the Cutter vaccine which comprises formaldehyde killed plague bacilli and is administered to the body by intramuscular injection. However, parenteral immunisation, although effective in inducing systemic immunity, does not effectively induce mucosal immunity (McGhee et al, (1992) Vaccine 10, 75-88) and cases of pneumonic plague have been reported in vaccinated individuals (Meyer (1970) Bull WHO 42 p663-668). So far no vaccine capable of producing a protective immune response at mucosal surfaces has been reported.

The live attenuated EV76 vaccine was tested extensively and used in the former Soviet Union from 1939, although its efficacy in evoking an immune response in man is questionable (Meyer et al (1974) J. Infect. Dis. 129 Supp: 13-18). It has been shown that the virulence of EV76 differs in several animal species. and non-human primates are particularly susceptible to a chronic infection with this strain. In the Western World the vaccine is considered to be unsuitable for mass vaccinations due to the extreme severity of the side effects and the possibility of the strain reverting to full virulence.

Two known Y. pestis antigens are the Y. pestis LcrV (V antigen), and the Fl antigen: both of which have now been found to be capable of evoking protective immune responses in animals. The present inventors have previously provided live orally active vaccine microorganisms capable of expressing V antigen and F1 antigen respectively which each provide good protection against challenge with Y. pestis at up to 10.sup.3 cfu. These vaccines are the subject of copending patent applications PCTIGB94/02818 and GB 9404577.0.

The present inventors have now surprisingly found that whereas only the unacceptably hazardous EV live vaccine had been shown to be capable of giving good protection against challenge with 10.sup.9 cfu or more with Y. pestis GB strain, and V and F1 antigens alone only provide full protection against challenge with about 10.sup.5 cfu. by administering a combined vaccine comprising V and F1 antigens they can at least match the protection afforded by EV76 without any of the hazards that have kept the EV vaccine from general use.

Still more advantageously, they have found that the vehicle for administration may be a simple mixture of the two protein components, rather than as a more complex attenuated whole organism. For long term protection and for the purposes of producing mucosal immunity, they have provided preferred forms of vaccine compnsing the two components in the form of live attenuated vaccine such as the F1 and V expressing Aro A or C Salmonella typhi referred to in the aforesaid copending applications, and in more preferred forms a single or double mutant expressing these antigens separately, or a fusion protein comprising both antigens.

Further provided are micro-organisms comprising both of Fl and V types of construct or plasmids of the applicants copending applications referred to above. These contain constructs that are capable of transforming a human or animal gut colonising micro-organism such that it is enabled to express proteins that are equivalent in sequence to F1 and V antigens respectively; these producing a protective immune response against Yersinia pestis in a human or animal body when the micro-organism is administered by oral or parenteral routes, and preferably allow the micro-organism to maintain its ability to colonise the human or animal gut.

A particularly preferred recombinant DNA, plasmid or human or animal gut colonising organism encodes for or expresses all or a protective epitopic part of the mature V protein of Yersinia pestis and all or a protective epitopic part of the mature Fl protein of Yersinia pestis. DNA encoding the whole or part of the F1 antigen and DNA encoding the whole or part of the V antigen could be used directly as a genetic vaccine.

Particularly preferred recombinant DNA encoding for V comprises a DNA sequence as described in SEQ ID No 1 or SEQ ID No 3, more preferably positioned in frame with a promoter such as lacz or nir.beta., and preferably in a vector capable of expression and replication in a Salmonella. Particularly preferred recombinant DNA encoding for F1 comprises a DNA sequence as described in SEQ ID No 10. SEQ ID No 2 and SEQ ID No 4 show the amino acid sequences of two preferred V antigen proteins; SEQ ID No 2 being the sequence of the V-antigen itself, and SEQ ID No 4 being that of V-antigen with four extra vector defined N-terminal amino acids. SEQ ID No 11 is that of an F1 protein as encoded for by SEQ ID No 10.

The preferred DNA constructs used in microorganisms of the invention allow production of micro-organisms that when orally administered induce local stimulation of the gut-associated lymphoid tissue (GALT) and, by trafficking of lymphocytes through the common mucosal immune system provide a secondary stimulation of the bronchial associated lymphoid tissue (BALT). In this manner a secretory IgA response is achieved at the respiratory mucosal surface.

The micro-organisms provided by transformation using this DNA in vector or directly inserted format, are preferably attenuated, more preferably attenuated salmonella.

Attenuated micro-organisms such as S. typhimurium have been well characterised as carriers for various heterologous antigens (Curtiss, (1990); Cardenas and Clements, (1992)). Attenuation may be effected in a number of ways, such as by use of the aro A and/or aro C mutation approach (see Hosieth et al (1981) Nature 291. 238-239: Dougan et al (1986) Parasite Immunol 9, 151-160; Chatfield et al (1989) Vaccine 7, 495498); multiple mutations such as aro A and aro C mutants as described by Hone et al (1991) Vaccine 9, pp 810-816 may also be used. However, any suitably defective organism that is safe for intended use may be employed.

Many other such attenuated deletions and mutations will be known for these and other microorganisms which will render them suitable for transformation with constructs of the present invention for the purposes of expressing vaccine proteins in the gut and/or gut colonisation in animals to be treated for Y.pestis. For human vaccination vectors containing the constructs of the present invention are placed in attenuated S. typhi and that transformed organism used as active agent for a live oral vaccine.

When DNA is used to transform the attenuated micro-organism by direct insertion into microorganism DNA this may be by direct integration into a gene. Alternatively when incorporated in the form of a plasmid that expresses V or F1 protein or epitopic fragments thereof this may be such that only the V or F1 protein or fragment is expressed or that this is expressed as a fusion peptide with a further protein or peptide fragment, preferably including the other one of the antigenic F1/V components. Such further protein or peptide fragment might be such as to promote export of mature proteins or peptide through the cell membrane or might be a further Y. pestis antigen.

The Icr gene was cloned from Y. pestis strain KIM by Price et al and its nucleotide sequence published in J Bacteriol (1989) 171, pp 5646-5653. In the examples below this information was used to design oligonucleotide primers which could amplify the gene from Y. pestis (strain GS) using the polymerase chain reaction (PCR). PCR primers were designed to be complementary to respective sequences flanking the 5' and 3' ends of the Icrv gene but also having 5' end tails containing a restriction enzyme recognition site to enable amplified IcrV gene to be cloned directionally into a plasmid vector (the 5' PCR primer containing an EcorRI site and the 3' primer containing a SacI site). These restriction enzyme sites are examples only and should not be seen as excluding other restriction enzymes.

In the examples below the constructs of the invention include a lac promoter, but other promoters such as the macrophage promoter (nir.beta.) may be used.

The production of Fl has been described fully in Oyston et al (1995) Infect. Immun. Vol. 63 No 2 p563 - see page 564 under results: Cloning and Expression of caf1.

The dosage of the antigen components in a vaccine may vary dependent upon an individual animals immune characteristics, but for immunisation in the mouse animal model of the examples below it has been found that 10 .mu.g of each of V and F1 per dose were effective in providing full protection when administered in a standard primer and booster schedule.

The antigens may be incorporated into a conventional pharmaceutically acceptable carrier, no particular limitation being imposed here. Conveniently the antigens have been incorporated into an oil in water emulsion. Adjuvants may be included in the vaccine composition, and particularly Freund's Incomplete adjuvant IFA has been found to be effective when treating the mouse model.

The carrier may be one suited to parenteral administration, particularly intraperitoneal administration but optionally oral, in the case of micro-organism based vaccines, or administration in the form of droplets or capsules, such as liposomes or microcapsules as would be effective in delivering the composition to the airways of an individual for the purpose of evoking mucosal immune response. The carrier may also comprise a slow-depot release system e.g. Alhydrogel.

Another method of encapsulation includes the use of polymeric structures in particular linear block co-polymers. Biodegradable polymers for example poly-lactic acid with or without glycolic acid or block co-polymer may be used; these may contain the following repeat unit: (POP-POE).sub.n where POP is polyoxypropylene and POE is polyoxyethylene. Block copolymers which contain (POP-POE).sub.n are of particular use.

The method, constructs, micro-organisms and vaccines of the invention will now be exemplified by way of illustration only by reference to the follcwing Sequence using Figure and Examples. Still further embodiments will be evident to those skilled in the art in the light of these.

FIG. 1 illustrates in bar-chart form the survival rates of a number of groups against a challenge of Y.pestis.

FIG. 2 illustrates in graphical form, IgG priming responses to intramuscular BSA immunisation in Balb/c mice.

SEQUENCE LISTING

SEQ ID No 1: Shows the nucleotide and derived amino acid sequence of a V-encoding DNA with 6 bases of vector pMAL-p2 or pMAL-c2 into which it is cloned at the 5' end using the EcoRI site in sequence GAATTC (derived from the 5' end PCR primer) and at the 3' end at the Sall site in sequence GTCGAC (derived from the 3' end PCR primer). The base at position 1006 has been altered by PCR mutagenesis to a T to create a second in frame stop codon. The start of the amino acid sequence is C-terminal to a factor Xa cleavage site.

SEQ ID No 2: Shows the amino acid sequence of the peptide expressed by the insert DNA of the invention, with an additional four amino acids encoded for by the vector (IE+FS) at the N-terminal end.

SEQ ID No 3: Shows the nucleotide and derived amino acid sequence of a second V-encoding DNA of the invention with 10 bases of a vector pGEX-5X-2 into which it is cloned shown at the 5' end using the EcoRl site in sequence GAATTC (GA derived from the 5' end PCR primer) and the Sall site in sequence GTCGAC (GTCGAC derived from the 3' end PCR primer). The base at position 1006 has been altered by PCR mutagenesis to create a second in frame stop codon; the base at position 16 has been altered to a C from an A to create the EcoRI site. The start of the amino acid sequence is C-terminal to a factor Xa cleavage site.

SEQ ID No 4: Shows the amino acid sequence of the peptide expressed by the DNA of SEQ ID No 3, with four amino acids encoded by the vector (G, I, P and G) at the N-terminal end.

SEQ ID No 5: Shows the nucleotide sequence of a gene 5' end primer oligonucleobde used to generate V-encoding DNA used in SEQ ID No 1.

SEQ ID No 6: Shows the nucleotide sequence of a gene 3' end primer oligonucleotide used to generate V-encoding DNA used in the Examples.

SEQ ID No 7: Shows the nucleotide sequence of a PCR primer oligonucleotide corresponding to the first 21 bases encoding for mature caf1 with an additional 5' region encoding for a SacI site.

SEQ ID No 8: Shows the nucleotide sequence of a PCR primer oligonucleotide corresponding to the sequence of caf1 which encodes a `stem loop` downstream of the termination codon with an added 5' region encoding SacI and AccI sites.

SEQ ID No 9: Shows the nucleotide sequence of a PCR primer oligonucleotide corresponding to an internal end region of the caf1 gene starting 107 bases downstream from the end of the first oligonucleotide.

SEQ ID No 10: Shows the nucleotide sequence of the pFGAL2a construct showing the fusion of the first few bases of the .beta.-galactosidase sequence in the vector with caf1 minus its signal sequence and having a 5' tail including a Sac I restriction site: the sequence is shown up to the caf1 AACC 3' end with some vector bases.

SEQ ID No 11: Shows the amino acid sequence of the protein encoded by OFGAL2a. This sequence may be proceeded by Met, Thr, Met, lie, Thr. Asn.

SEQ ID No 12: is that of primer FIOU2 used to amplify the F1 operon.

SEQ ID No 13: is that of primer M4D used to amplify the F1 operon.

SEQ ID No 14: is that of primer M3U used to amplify the F1 operon.

SEQ ID No 15: is that of primer FIOD2 used to amplify the F1 operon.

SEQ ID No 16: is the nucleotide and derived amino acid sequence of a DNA fragment encoding an F1-V fusion protein. There is a SacI cloning site at the 5' end and a Hind III cloning site at the 3' end. Bases 452-472 is a sequence contained in the cloned insert, but derived from PCR primers (not found in Y. pestis DNA).

SEQ ID No 17: is the amino acid sequence of SEQ ID No 16.

SEQ ID No 18: is that of primer 5'FAB2 used to amplify the F1 operon including signal sequence.

SEQ ID No 19: is that of primer 3'FBAM used to amplify the F1 operon including signal sequence.

SEQ ID No 20: is the nucleotide and amino acid sequence for F1 antigen as defined by PCR primers detailed in exemplified SEQ ID No 18 and 19 including signal sequence.

SEQ ID No 21: is the amino acid sequence of SEQ ID No 20.

SEQ ID No 22: is the nucleotide and amino acid sequence of F1/V fusion protein including a 5 amino acid linker region. The T at position 1522 was modified from G to create a second in frame stop codon.

SEQ ID No 23: is the amino acid sequence of SEQ ID No 22. The linker region referred to in SEQ ID No 22 is at amino acid position 171-176 (bases 523-540 in SEQ. ID No 22).

SEQ ID No 24: shows the nucleotide sequence of a gene 5' end primer oligonucleotide used to generate V-encoding DNA used in SEQ ID No 3.

Cloning of SEQ. ID No 3

Materials and Methods: Materials for the preparation of growth media were obtained from Oxoid Ltd. or Difco Laboratories. All enzymes used in the manipulation of DNA were obtained from Boehringer Mannheim UK Ltd. and used according to the manufacturer's instructions. All other chemicals and biochemicals were obtained from Sigma chemical Co unless otherwise indicated. Monospecific rabbit polyclonal anti-V and mouse anti-GST sera were prepared by Dr R Brubaker (Department of Microbiology, Michigan State University) and Dr E D Williamson (Chemical and Biological Defence Establishment), respectively.

Bacterial strains and cultivation: Yersinia pestis GB was cultured aerobically at 28.degree. C. in a liquid medium (pH 6.8) containing 15 g proteose peptone, 2.5 g liver digest. 5 g yeast extract. 5 g NaCl per liter supplemented with 80 ml of 0.25% haemin dissolved in 0.1M NaOH (Blood Base Broth). Escherichia coli JM109 was cultured and stored as described by Sambrook et al. Molecular Cloning. A Laboratory Manual.

Production of recombinant V and F1 proteins: Manipulation of DNA. Chromosomal DNA was isolated from Y. pestis by the method of Marmur.

Production of recombinant V-antigen: The gene encoding V-antigen (1crV) was amplified from Y. pestis DNA using the polymerase chain reaction (PCR) with 125pmoi of primers homologous to sequences from the 5' and 3' ends of the gene (see Price et al (1989). J. Bacteriol 171 p5646-5653).

The sequences of the 5' primer (V/5'E: GATCGAATTCGAGCCTACGAACAA) and the 3' primer (GGATCGTCGACTTACATAATTACCTCGTGTCA) also included 5' regions encoding the restriction sites EcoR1 and Sal1, respectively. In addition, two nucleotides were altered from the published sequence of IcrV (Price et al, 1989), so that the EcoRI site was completed and the amplified gene encoded an extra termination codon (TAA). The PCR pnmers were prepared with a DNA synthesiser (392 Applied Biosystems) Applied Biosystems. A DNA fragment was obtained after 30 cycles of amplification (95.degree. C., 20secs, 45.degree. C., 20 secs, 72.degree. C. 30 secs: Perkin 9600 GeneAmo PCR System). The fragment was purified, digested with EcoR1 and Sal1. ligated with suitably digested plasmid pGEX-5X-2 and transformed into E. coli JM109 by electroporation (see Sambrook et al 1989). A colony containing the recombinant plasmid (pVG 100) was identified by PCR using 30-mer primers (5' nucleotides located at positions 54 and 794: see Price et al 1989) which amplified an internal segment of the IcrV gene.

Expression of rV in E. coli. Cultures of E. coli JM109/pVG100 were grown in LB containing 100 .mu.gml.sup.-1 ampicillin at 37.degree. C. until the absorbance (600 nm) was 0.3. Isopropyl-.beta.-D-thiogalactopyranoside (IPTG) was then added to the culture to a final concentration of 1 mM and growth was continued for a further 5 hours. Whole cell lysates of the bacteria were prepared as described in Sambrook et al and expression of the GSTN fusion protein was examined by staining 10% SDS-polyacrylamide gels (Mini-Protean II, BioRad) with Coomassie Brilliant Blue R250 and by Western Blotting (see Sambrook et al). Western Blots were probed with rabbit anti-V serum diluted 1/4000 or mouse anti-GST serum in dilution at 1/1000 and protein bands were visualised with a colloidal gold labelled secondary antibody (Auroprobe BLplus, Cambio).

Quantification of GSTN expression in vitro. Cultures of E. coli JM109 containing pVG1OO or pGEX-5X-2 were grown as described above. One ml aliquots were removed from the cultures in logarithmic and stationary phases and the number of viable cells determined by inoculating onto L-agar containing ampicillin. The cells were harvested from a second 1 ml aliquot by centrifugation and resuspended in 1 ml of phosphate buffered saline (PBS). The cell suspension was frozen at -20.degree. C. for 1 hour, thawed and then sonicated on ice at 10.degree. C. lower for 3.times.30 secs (model XL2015 sonicator, 3.2 mm Microtip probe; Heat Systems Inc.). The sonicates and a standard solution of rV (5ugml.sup.-1) were serially diluted in PBS in a microtitre plate and allowed to bind overnight at 4.degree. C. The quantity of GST/V fusion protein in eacn sonicate was determined in a standard ELISA using rabbit anti. V serum as the primary antibody. Antibodies were incubated in 1% skimmed milk powder in PBS.

Purification of rV. E. coii JM109/pVG100 was grown in 5.times.100 ml volumes of LB as described above. The cells were harvested by centrifugation ana resuspended in 3 ml. phosphate buffered saline (PBS). After the addition of 80 .mu.l lysozyme (10 mgml.sup.-1 ; the cell suspension was incubated for 10 min at 22.degree. C. Triton X-100 was added to final concentration of 1% and the cells were frozen (-20.degree. C.), thawed and sonicated on ice at 70%, power for 3.times.30 s (model XL2015 sonicator). The lysed cells were centrifuged, and the supernatant was made up to 30 ml with PBS and mixed with 5 ml of Glutathione Sepaharose 4B (Pharmacia Biotech) which had been washed three times with PBS+0.1% Triton X-100. The mixture was stirred for 18 hours at 4.degree. C., centrifuged and washed twice in 100 ml PBS. and then packed into a chromatography column (Poly-Prep: Bio-rad) as a 50% slurry. The GST/V fusion protein was eluted with 10 ml of 50 mM Tris pH 8.0 containing 5 mM reduced glutathione (Pharmacia Biotech). After dialysis against PBS, the fusion protein was cleaved with factor Xa (Boehringer Mannheim UK Ltd) for 18 hours at 22.degree. C., according to the manufacturer's instructions. Cleaved GST and excess uncleaved GSTN were removed from the solution by affinity adsorption, as described above, to leave purified recombinant V (rV).

Immunisation with rV. Six week old female Balb/c mice, raised under specific pathogen-free conditions (Charles. River Laboratories, Margate, Kent. UK), were used in this study. A group of 16 mice received a 0.2 ml primary immunising dose intraperitoneally (i.p.) of 10.13 .mu.g of rV antigen, presented in a 1:1 water-in-oil emulsion with Incomplete Freund's Adjuvant (IFA). On days 14 and 34, each animal received booster immunising doses, prepared as above. On day 64, 6 animals were sacrificed and their tissues were removed for immunological analyses. as described below. The remaining animals were challenged with Y. pestis. An untreated control group of 16 age-matched mice were divided similarly into groups for tissue sampling and challenge. In a subsequent experiment to determine the degree of protection against higher challenge doses of Y pestis groups of 5 or 6 rV-immunised and control mice were used.

Measurement of serum antibody titre. Blood was sampled by cardiac puncture from mice anaesthetised i.p. with a 0.1 ml cocktail containing 6 mg of Domitor (Norden Laboratories) and 27 .mu.g cf Ketalar (Parke-Davies). The samples were pooled and the serum was separated. The serum antibody titre was measured by a modified ELISA (Willamson and Tiball, (1993) Vaccine 11: 1253-1258). Briefly, rV (5 .mu.gml.sup.-1 in PBS) was coated onto a microtitre plate and the test sera were serially diluted in duplicate on the plate. Bound antibody was detected using peroxidase labelled conjugates of anti-mouse polyvalent 1 g. The titre of specific antibody was estimated as the maximum dilution of serum giving an absorbance reading greater than 0.1 units. after subtraction of the absorbance due to non-specific binding detected in the control sera.

Isolation of purified T cells from the spleen. A crude suspension of mixed spleen cells was prepared by gently grinding the spleen on a fine wire mesh. The cells were flushed from the splenic capsule and connective tissue with 2 ml of tissue culture medium (DMEM with 20 mM L-glutamine, 10.sup.5 U1.sup.-1 of penicillin and 100 mgl.sup.-1 of strepomycin). A population of mixed lymphocytes was separated from the spleen cell suspension by density gradient centrifugation of Ficoll-Hypaque (Lymphocyte Separation Medium, ICN Flow). A mixed acridine orange (0.0003% w/v) and ethidium bromide (0.001% w/v) stain was used to determine the percentage of viable cells in the preparation.

The mixed lymphocytes were incubated with sheep anti-mouse IgG-coated Dynabeads (M450), Dynal UK) at a ratio of 1:3 for 30 minutes at 4.degree. C. The Dynabead linked B cells were removed by magnetic separation and the remaining T cells were resuspended in DMEM, supplemented with antibiotic and 10% v/v foetal calf serum (FCS) at the desired cell density for seeding to microtitre plates.

In vitro proliferation of crude spleen cells and purified T cells against rV. Doubling dilutions of rV or Concanavalin A (positive control) in DMEM (range 0-50 .mu.gml.sup.-1) were made in the wells of a microtitre plate, such that 0.1 ml remained in each well. Negative controls consisted of 0.1 ml of DMEM alone. An equal volume of the crude spleen cell or purified T cell suspension was seeded into each well at a minimum density of 5.times.10.sup.4 cells and incubated for 72 hours at 37.degree. C. (5% CO.sub.2). One .mu.Ci of .sup.3 H thymidine ([methyl[.sup.3 H]thymidine S.A. 74 GBqmmol.sup.-1 ; Amersham) in 30 .mu.l of DMEM supplemented with 10% FCS was aliquoted into each well and incubation was continued for 24 h. The well contents were harvested onto glass fibre filters using a cell harvester (Titertek) and discs representing each well were punched from the filter mats into 1.5 mI of scintillation fluid (Cyoscint. ICN Biomedicals Inc.) so measure the incorporation of .sup.3 H thymidine into cells. The cell stimulation index was calculated from a replicates as mean cpm (stimulated)mean cpm (negative control).

Production of recombinant F1 antigen. Cloning of caf1: DNA was isolated from Y. pestis by the method of Marmur et al (1961) J. Mol. Biol. 3: pp 208-218. A DNA fragment encoding the open reading frame of caf1 minus its signal sequence was amplified from this using the polymerase chain reaction (PCR). Oligonucleotides were prepared with a Beckman 200A DNA synthesiser for use in the PCR.

pFGAL2a construct. Oligonucleotide GATCGAGCTCGGCAGA7T'AACTGCAAGCACC (SEQ ID No 7) was synthesised corresponding to the first 21 bases of the caf1 gene immediately following the nucleotides encoding the signal sequence with an additional 5'region encoding a SacI site and the complimentary oligonucleotide CAGGTCGAGCTCGTCGACGGTTAGGCTCAAAGTAG (SEQ ID No 8) corresponding to the sequence which encodes a putative `stem loop` structure downstream of the caf1 termination codon with an added 5'region encoding SacI and AccI sites. A DNA fragment was obtained after 35 cycles of amplification (95.degree. C., 15 secs; 50.degree. C., 15 secs; 72.degree. C., 30 secs using a Perkin Elmer 9600 GeneAmp PCR system). The fragment was purified, digested with SacI and AccI, ligated into a similarly digested pUC18 plasmid and transformed into E. coli JM109 by electroporation. Electroporation was carried out using a Biorad Gene Pulser with 0.2 cm cuvettes at 1.25kV. 25pF, 800 Ohms with a time constant of 20.

A pFGAL2a colony containing the cloned caf1 gene was identified by PCR using an oligonucleotide TGGTACGCTTACTCTTGGCGGCTAT (SEQ ID No 9) corresponding to an internal region of the gene 128 to 153 nucieotides from the site identified as the signal sequence cleavage site (see Galyov et al (1990)) and the SEQ ID No 2. An F1 expressing E. coli culture containing the pFGAL2a was grown at 37.degree. C. with shaking in Luria Broth with 1 mM isopropyl-.beta.-D-thiogalactopyranoside (IPTG) for 18 hours. Whole cell lysates and periplasmic and cytoplasmic fractions of the bacteria were prepared as described by Sambrook et al (1989).

SDS-PAGE and Western blotting: SDS-polyacrylamide gel electrophoresis (PAGE) and Western blotting were performed as described by Hunter et al (1993) Infec. Immun. 61. 3958-3965. Blots were probed with polyclonal antisera raised in sheep (B283) against killed Y pestis (EV76 strain grown at 37.degree. C.) and bound antibody was detected with a horseradish peroxidase-labelled donkey anti-sheep IgG (Sigma).

Animals: Barrier bred female 6-week old Balb/c mice free of mouse pathogens were obtained from Charles-Reiver Laboratories, Margate. Kent. UK and were used throughout these Examples.

Immunisation: Mice were divided into groups of 5 or 6 and immunised as follows.

Comparative vaccine EV76: A total of 50 mice received a single subcutaneous (s.c.) priming dose of live EV76 vaccine on day 0 of the schedule delivered in a total volume of 100 .mu.l.

Comparative vaccine Cutter USP vaccine: A further group of 50 mice were primed intramuscularly (i.m.) with 100 .mu.l of Cutter vaccine; this representing about 0.2 of the human dose and comprised approximately 2.times.10.sup.8 formaldehyde killed plague bacilli. This dose was administered to each animal again on day 16 of the schedule to effect booster immunisation.

F1 and V vaccines: Groups of 62 mice received a primary immunising dose intra-peritoneally (i.p.) of either recombinant V-antigen or recombinant F1-antigen, presented in a 1:1 water-in-oil emulsion with incomplete Freunds adjuvant (IFA; Sigma). Animals were primed with a 18 .mu.g dose of the respective antigen in a total volume of 0.1 ml emulsion. Animals were boosted with the respective antigen as appropriate on days 14 (V and Fl groups) and 28 (all F1 group and a subgroup of 12 of the V group).

A further group of 12 mice were primed and boosted on days 0. 14 and 28 with a combination of 10 .mu.g F1 and 10 .mu.g V jointly incorporated into the aqueous phase of a 1:1 water-in-oil emulsion with IFA (final volume 0.1 ml per mouse).

On day 50 of the immunisation schedule, 6 animals were selected at random from each of the treatment groups for subsequent analysis of spleen cell responses. The remaining animals in each group were challenged with Y. pestis. An untreated control group of age-matched mice was similarly split.

Multiple LD challenge to determine limits of protection: Mice from each of the immunised groups and untreated controls were divided into groups of 5 or 6 for challenge by the s.c. route with Y. pestis GB strain in the dose range 20 to 2.times.10.sup.9 viable organisms. Challenged mice were closely observed over a 14-day period for the development of symptoms and where appropriate time to death was carefully recorded.

Animals which succumbed to the challenge were autopsied and blood smears, livers and spleens removed for bacteriological analysis.

The results of these challenges on control and test animals is given in Table 1 below; two sets of results being given corresponding two experimental runs with respective controls. From these results it can clearly be seen that while the V antigen is more effective than Cutter, it is inferior to EV76. However, when combined with the less effective F1 the combination is as effective as EV76 without side effects.

Expression of recombinant V-antigen from S. typhimurium and typhi.

pMAL-p2: a vector designed to express the cloned gene as a fusion product with a maltose binding protein (MBP). The C-terminus of the MBP is fused to the N-terminus of the V-antigen. The fusion protein so produced on expression is exported to the periplasm. Vector including the V-antigen DNA sequence was designated pVMP100.

pMALC2: a vector similar to pMAL-p2 except that MBP-V antigen fusion protein is expressed cytoplasmically. The recombinant plasmid was designated pVMC100.

pGEX-5X-2: a vector designed to express the cloned gene as a fusion protein with glutathione-S-transferase (GST). The C-terminus of GST is fused to the N-terminus of V antigen and the fusion protein is expressed cytoplasmically. The recombinant plasmid was designated pVG100.

All the vectors contain the P.sub.tac promoter and the lacI.sup.Q gene: the latter encoding the 1ac repressor which turns off transcription from P.sub.tac in Escherichia coli until IPTG is added. The plasmids contain the origin of replication from pBR322 and as a result replicate to a low copy number in the bacterial cell.

Each of the recombinant plasmids was electroporated into Salmonella typhimurium strain SL3261, an attenuated strain that has been used extensively as a live vaccine vector for the expression of foreign antigens. It contains a specific deletion mutation in the aroA gene which makes the mutant dependent upon certain aromatics for growth (see Hosieth et al). For producing microorganism suitable for human vaccination use electroporation is into attenuated Salmonella typhi.

The recombinant plasmids all expressed V antigen as shown by Western blotting of S. typhimurium cultures and probing with a monospecific anti-V antigen polyclonal antiserum supplied by R Brubaker, Dept Microbiology, Michigan State University, East Lansing, Mich. 48824-1101, USA. Recombinant S. typhimurium were innoculated intravenously into mice at 5.times.10.sup.7 cfu/dose and shown to colonise the liver and spleen at high levels: between 8.times.10.sup.6 and 5.times.10.sup.8 cfu per organ were recovered. The majority of the bacterial cells recovered were also ampicillin resistant indicating retention of recombinant plasmids.

Expression of F1 in S. typhimurium: The pFGAL2a plasmid was isolated using general techniques described in Sambrook et al (1989) Molecular Cloning; a Laboratory Manual, 2nd Edition. Cold Spnng Harbour Laboratory, New York. Purified plasmid was electroporated into S. typhimurium LB5010 (restriction, modification) and methylated pFGAL2a was subsequently isolated from the LB5010 for electroporation into S. typhimurium SL3261 (aro A). Peeplasmic and cyptoplasmic fractions were prepared for SDS-PAGE and Western blotting as described above.

Stability of constructs: Five female Balbic mice were inoculated intravenously with either 5.times.10.sup.5 or 5.times.10.sup.7 cfu S. typhimunum containing pFGAL2a in 200 .mu.L phosphate buffered saline. Control mice were inoculated similarly with S. typhimurium containing pUC18 with no insert. After 7 days the mice were killed by cervical dislocation and their livers and spleens removed. The organs were homogenised in 10 ml phosphate buffered saline using a stomacher on maximum setting for 2 minutes and the homogenate was serially diluted in phosphate buffered saline and placed onto L agar or L agar containing 55 .mu.g ml.sup.-1 ampicillin.

F1 operon construct: Attempts to PCR replicate the entire F1 operon as one piece were unsuccessful, so a strategy was developed whereby it was amplified using PCR to produce two discrete fragments using primer pairs (A) of SEQ ID No 12 and 13 and (b) of SEQ No 14 and 15 respectively to produce fragments of 3.36 kb and 1.89 kb from Y. pestis MP6 template DNA. Marmur extract of DNA was used without CsC1.sub.2 purification. The PCR cycle conditions used were 96.degree. C. for 30 seconds, 57.degree. C. for 30 seconds and 72.degree. C. for 1 minute; total of 30 cycles.

These two fragments were digested using Nhe1 and joined together.

This fused fragment, encoding the full length operon (5.25kb), was digested with EcoR1 and Sal1 and then cloned into a number of vectors. When this fragment was cloned into pBR322 and expressed in E. coli, S. typhimudum LBE5010 or SL3261 instability of the recombinant plasmid was noted. To circumvent this problem the operon was cloned into plasmid pLG339, a low copy number plasmid km.sup.R. The entire F1 operon was also inserted into AroC gene on the chromosome of gene on the chromosome of S. typhimurium using vector pBRD1084.

As discussed previously in this application. the V and F1 antigen may be microencapsulated. The V and F1 antigens may be separately microencapsulated. The combined microencapsulated sub-units may be used for immunisation. The present inventors beiieve that the protection afforded by the combined microencapsulated sub-units is superior to that provided by existing plague vaccines and that there is an additive effec: in combining the sub-units. The protective efficacy of the combined microencapsulated sub-un-its may further be enhanced by co-administering an adjuvant for example cholera toxin B sub-unit (CTS). Microencapsulation of the sub-unit vaccine prolongs release of the vaccine in vivo. permits oral. intra-nasal or inhalational delivery and gives scope for targetting.

The microencapsulation of the combined V and F1 sub-unit vaccine is described below. Also demonstrated is that this formulation is able to induce both mucosal and systemic immunity against plague.

The microencapsulation of sub-units was effected in PLA 2000 using a solvent evaporation technique.

Immunisation with the microencapsulated sub-units was carried out as follows.

Groups of 21 mice received a primary immunising dose of 25 .mu.g of either V antigen or F1 antigen, presented in microspheres resuspended in PBS for intra-peritoneal (i.p.) injection. Further groups of 21 mice received a combination of 25 .mu.g of each of the F1 and V antigens, presented in microspheres. A dose of 25 .mu.g of F1 was delivered in a total mass of 5.42 mg of spheres, whilst 25 .mu.g of V was contained in 2.08 mg of spheres. The required mass of microspheres was re-suspended in 100 .mu.l PBS per animal for injection. Animals were boosted with the respective antigen(s), as appropriate on days 14 and 28.

Two further groups of 21 mice were primed and then boosted i.p. on days 14 and 28 with a combination of 10 .mu.g F1 and 10 .mu.g V. jointly incorporated in the aqueous phase of 25% (v/v) suspension of alhydrogel (Alhydrogel 1.3%, Superfos. Denmark) in PBS

Selected groups of animals received in addition a dose of 10 .mu.g CTB (Sigma, Poole) incorcorated into the delivery vehicle at each dosing point.

Control groups, each of 21 mice, received either alhydrogel only (100 .mu.l of 25% solution) or CTB only (10 .mu.g in 100 .mu.l PBS) or remained untreated.

In order to compare the protective efficacy of immunisation with combined sub-units, free or microencapsulated, against that provided by the Greer vaccine, (purchased from Greer laboratories) animals were challenged s.c. with virulent Y.pestis.

There was a 60% survival rate in Greer vaccinees against a challenge of 2.times.10.sup.5 cfu Y.pestis (FIG. 1). By comparison, 80% of the combined microencapsulated F1+V (group 1) survived this challenge and there was 100% survival in group 2 (.mu.V+pF1+10kg CTB). Thus the combined microencapsulated formulation was protective against virulent Y. pestis with no evidence of side effects.

In summary treatment groups were:

Micro-encapsulation may also be carried out with block co-polymers, in the following experiments, model protein antigen BSA was used. The preparation and characterisation of microspheres is as follows. Protein-loaded microspheres were prepared by an oil/water solvent evaporation method procedure previously described see R. L. Hunter and B. Bennet. The J. Immunol., 133(6), 3167-3175 (1984), with some modifications. Polymer (poly-D-lactic acid): Resomer 206. Boehringer Ingelheim, Germany; 125 or 250 mg) solution in acetone (22.5 ml), containing model protein (antigen) BSA (at 15-25% theoretical loading level) and 0.11%w/v Pluronic L101 (or 0.09% w/v L121) available from Zeneca, probe sonicated for 10 seconds and then added to the aqueous phase (22.5 ml), mixing at 100 rpm for 5 minutes and rotary evaporated until the organic solvent had been removed. The resulting coiloid was washed and freeze-dried. Microspheres with an average diameter of .about.1 .mu.m (as determined by Malvern Mastersizer) and protein loadings .about.0.5-1.0% produced in this fabrication condition. External morphology of the resulting microspheres were analysed by scanning electron microscopy (SEM). Surface characteristics were defined in terms of zeta potential and hydrophobicity.

Hydrophobicity measurements: hydrophobicity of microspheres was quantified using hydrophobic interaction chromatography (HIC) as previously reported see H. O. Alpar and A. J. Almeida, Eur. J. Pharm. Biopharm. 40, (4), 198-202 (1994). Microspheres were eluted from a series of agaroses which were modified with hydrophobic residues. The retention of microspheres in octyl agarose was used as an index of hydrophobicity.

Immunisation: A study was designed to establish the effects of differences in the type of microsphere and surface properties of the immune response. Female Balb/c mice (five per group) were injected i.m. with a single dose of BSA either encapsulated in Pluronic formulated microspheres or free in 100 .mu.l alone or suspended in the presence of surfactant. The control group of mice received the same amount of antigen encapsulated into microspheres containing PVA as emulsion stabiliser. Tail tip blood samples were removed periodically for 2 months. The serum from each sample was analysed for anti-BSA antibody using an enzyme-linked immunosorbent assay (ELISA). Results are represented graphically in FIG. 2.

The presence of Pluronic L101 and L121 endows the surface with a much higher degree of hydrophobicity compared to PVA formulations (70% retained on octyl agarose column as opposed to 30% latex control 95% retained). A hydrophobic surface would facilitate macrophage interactions and subsequent uptake and would therefore be much more likely to mediate increased immune response. FIG. 2 shows the effect of different BSA formulations in eliciting an immunoresponse. Batches formed by Pluronic L101 had a more enhanced effect on the plasma anti-BSA antibody titre than those formed by PVA. Batches formed by Pluronic L121 were slightly inferior to those of PVA microspheres in inducing good primary antibody response after delivery of only a small dose of antigen (1 .mu.g). The higher serum IgG level obtained with Pluronic L101 preparations as compared to other preparations is noted and may partly be due to the higher surface hydrophobicity.

It was also mentioned earlier in the present application that DNA encoding the whole or part of the F1 antigen and DNA encoding the whole or part of the V antigen could be used directly, as a genetic vaccine. By way of example this may be carried out as follows.

1/ DNA encoding Y.pestis F1 and V is obtained by Polymerase Chain Reaction (PCR, amplification of specific regions of the Y.pestis genome, or by isolation of these genes from previously constructed plasmid clones e.g. for V exemplified sequence l.D. no 3.

2/ The F1 and V genes are cloned into mammalian expression vector plasmids such that the genes are situated downstream of a eukaryotic promoter. Suitable plasmids include pCMV.beta. (purchased from Clontech), in which the cytomegalovirus Immediate Early promoter is used. F1 and V may be cloned individually or in combination, and may be cloned as fusions with such genes as glutathione S-transferase, or eukaryotic signal sequences, which may stabilise the expressed protein and may facilitate export from mammalian cells.

3/ The recombinant plasmids are propagated in Escherichia coli and stocks are purified for transfection into a mammalian animal model and for immunisation of experimental animals by the intramuscular or intra-dermal or inhalational routes.

Example 1 To construct a DNA vaccine expressing V antigen, the plasmid vector pCMV was digested with restriction enzyme Not 1 to remove the lac Z gene coding sequence. The digested plasmid was treated with Klenow enzyme to create blunt-ended vector DNA. An ssp 1 restriction fragment containing the coding sequence for a fusion protein of V antigen and glutathione-S transferase was isolated from recombinant plasmid pVG100 and ligated to the vector DNA. The V sequence used in this case is that described in exemplified Seq. ID No. 3. The recombinant plasmid was transformed into E coli strain Nova Blue. Purifiea plasmid was inoculated into Balbic mice by intra-muscular injection. Immunoglobulin responses to V antigen were detected in the serum of inoculated animals.

Example 2: To construct a DNA vaccine expressing F1 antigen, PCR primers were designed to amplify the complete caf1 open reading frame. This encodes F1 and its signal peptide which directs export of the protein from the bacterial cell. The PCR primers had "tails" at their 5' ends which contained restriction enzyme recognition sites to allow directional insertion into a plasmid vector. The sequences of the PCR primers, 5'FAB2 and 3'FBAM, are given in exemplified Seq.ID. No. 18 and Seq.ID No. 19, respectively.

5'FAB2 and 3'FBAM were used to amplify a PCR fragment, the sequence of which is given in exemplified Seq.ID.no. 20. The PCR fragment was digested with restriction enzymes Nhe 1 and Bam HI and cloned into the plasmid pBKCMV which had been digested with the same enzymes. The resulting plasmid, pF1AB was transformed into E. coli Nova Blue and purifies plasmid was used to inoculate Balb/c mice by intramuscular injection. Immunoglobulin responses to F1 were found in inoculated animals.

Example 3: To construct a DNA vaccine expressing both F1 and V, the coding sequence for V was inserted into the DNA vaccine expressing F1 . detailed in example 2. A linker region coding for 6 amino acids was positioned between the F1 and V coding sequences to allow each of the proteins to attain their conformational shape independantly. The linker-V coding sequence was obtained by digesting the recombinant plasmid placF /6 with Bam HI and Hind III. The linker-V DNA was ligated with the plasmid pFI.sup.AB which had also been digested with Bam HI and Hind III. The resulting plasmid. pFVAB. was transformed into cells of E. coli Nova Blue and stocks of plasmid were purified for further use. The nucleotide and derived amino acid sequence of the F1 /V fusion are given in exemplified Seq. ID No. 22.

An example of a fusion protein comprising both F1 and V antigens is described below.

Enzymes and reagents.

Materials for the preparation of growth media were obtained from Oxoid Ltd or Difco Laboratories. All enzymes used in the manipulation of DNA were obtained from Boehrinoer Mannheim UK Ltd and used according to the manufacturer's instructions. All other chemicals and biochemicals were obtained from Sigma Chemical Co unless otherwise indicated. Monospecific rabbit polyclonal anti-V serum was supplied by Dr R Brubaker (Department of Microbiology, Michigan State University) and mouse anti-F1 IgA monoclonal antibody (Mab) F13G8-1 was obtained from the American Type Culture Collection.

Bacterial strains and cultivation.

Yersinia pestis GS was cultured aerobically at 28.degree. C. in a liquid medium (pH 6.8) containing 15 g proteose peptone, 2.5 g liver digest, 5 g yeast extract, 5 g NaCl per liter, supplemented with 80 ml of 0.25% haemin dissolved in 0.1M NaOH (Blood Base broth). Eschedchia coli JM109 was cultured and stored as described by Sambrook et al (Sambrook J et al. 1989. Molecular cloning, a laboratory manual. Cold Spring Harbor Laboratory Press, New York).

Manipulation of DNA.

Chromosomal DNA was isolated from Y. pestis by the method of Marmur (Marmur. J. 1961. J Mol Biol 3: 208-218). The genes encoding F1 antigen (caf1) and V antigen (IcrV) were amplified from Y. pestis DNA using PCR with 125 pmol of primers homologous to sequences from the 5' and 3' ends of the gene (Galyov, E E et al. 1990. FEBS Lett 277:230-232: Price S B et al. 1989. J Bacteriol 171: 5646-5653), although for caf1 only the region encoding the mature F1 antigen was amplified. The sequences of the primer (F/5'B: GATCGAGCTCGGCAGATTTAACTGCAAG CACC), the F1 3' primer (Flink/3'A: GCATGGATCCTTGGTTAGATACGGT), the V 5' primer (Vlink/5'A: ATGGATCCATCGAAGGTCGTATTAGAGCCTACGAACAA), and the V 3' primer (VG/3'A. GCATAAGCTTCTAGTGTCATTTACCAGACGT) also included 5' tails encoding the restriction sites SacI BamHI, BamHI and HindIII, respectively. In addition the nucleotide A was altered from the published sequence of IcrV (Price S B et al. 1989. J Bacteriol 171: 5646-5653) to include an extra termination codon (TAA) in the amplified DNA. The tail of primer Vlink/5'A also included nucleotides encoding the factor Xa cleavage sequence lle-Glu-Gly-Arg. The PCR primers were prepared with a DNA synthesiser (model 392: Applied Biosystems). DNA fragments were obtained after 30 cycles of amplification (95.degree. C., 20 s; 45.degree. C., 20 s: 72.degree. C., 30 s: model 9600 GeneAmp PCR System; Perkin Elmer) and the fragments were purified. The caf1 PCR product was digested with SacI and BamHI, ligated with suitably digested plasmid pUC18 and transformed into E. coli JM109 by electroporation. Subsequently, the IcrV-linker PCR product was digested with BamHI and HindIII, and ligated into the intermediate plasmid to form the recombinant plasmid placFV6. A colony containing placFV6 was identified by PCR using 30-mer primers (5' nucleotides located at positions 54 and 794 (Price S B et al. 1989. J Bacteriol 171: 5646-5653) which amplified an internal segment of the IcrV gene. To confirm the nucleotide sequence of the cloned insert, sequencing reactions containing placFV6 and primers designed from the caf1 and IcrV genes were performed using an automated Taq polymerase cycle sequencing protocol with fluorescently labelled dideoxy nucleotides (CATALYST Molecular Biology Labstation: Applied Biosystems). The reaction products were analysed using an automated DNA sequencer (model 373A: Applied Biosystems).

The DNA sequence and derived amino acid sequence of the cloned fusion protein is shown in Example 1. The fusion protein consists of F1 and V antigens separated by a six-amino acid linker Gly-Ser-lle-Glu-Gly-Arg. It is cloned downstream of the iac promoter and in-frame with the vector-encoded LacZ' fragment. Thus, the complete fusion protein encodes nine additional amino acids at the N-terminus (Met-Thr-Met-lle-Thr-Asn-er-Ser-Ser), and it accumulates in the cytoplasm.

Expression of the F1 /V fusion protein in E. coli.

Cultures of E. coli JM109/placFV6 were grown in LB containing 100 .mu.gml.sup.-1 ampicillin at 37.degree. C. until the absorbance (600 nm) was 0.3. Isopropyl-D-thiogalactopyranoside (IPTG) was then added to the culture to a final concentration of 1 mM and growth was continued for a further 5 h. Whole cell lysates of the bacteria were prepared as described by Sambrook et al (Sambrook J et al. 1989. Molecular cloning, a laboratory manual. Cold Spring Harbor Laboratory Press, New York) and expression of the F1/V fusion protein was examined by SDS-PAGE on 10-15% gradient gels (Phastsystem, Pharmacia Biotech) and Western blotting (Sambrook J et al. 1989. Molecular cloning, a laboratory manual. Cold Spring Harbor Laboratory Press, N.Y.). Western blots were probed with rabbit anti-V serum at a dilution of 1/4000 or Mab F13G8-1 at a dilution of 1/250, and protein bands were visualised with a colloidal gold labelled secondary antibody (Auroprobe BLplus. Cambio) or an anti-mouse IgA secondary antibody conjugated to horse radish peroxidase (Sigma).

A fusion protein with an approximate molecular weight of 54.2 kDa was detected in lysates of JM109/placFV6 by Western blotting with anti-V and anti-F1 sera. This product was not detected in control lysates of JM109/pUC18.

Electroporation into Salmonella typhimurium SL3261.

Plasmid DNA was extracted and purified from JM109/placFV6 using a Qiaprep kit (Qiagen) and electroporated into S. typhimurium strain LB5010 (r- m.sup.+). Subsequently, modified placFP6 was isolated and electroporated in S. typhimurium strain SL3261 (aroA his). For inoculation into mice. bacteria were grown in LB containing 100 pgml.sup.-1 ampicillin for 18 h without shaking. After washing, the cells were resuspended in 10% glycerol :n phosphate buffered saline (PBS) and stored at -70.degree. C. The cell suspensions were defrosted and diluted in PBS as required prior to injection.

Immunisation with SL3261/placFVG.

Six week old female Balb/c mice. raised under specific pathogen-free conditions (Charles-River Laboratories, Margate, Kent, UK), were used in this study. A group of 19 mice received 0.1 ml immunising doses of approximately 5.times.10.sup.6 cfu of SL3261/pFV6 on days 0 and 14 by the intravenous (iv) route. To retain placFV6 in vivo, mice were also injected subcutaneously (sc) with 50 .mu.l ampicillin trihydrate suspension (150 mgml.sup.-1 ; Penbritin injectable suspension POM: SmithKline Beecham Animal Health) for 5 days after each immunisation. In addition, groups of 15 mice were immunised iv on day 0 with a single 0.1 ml dose of approximately 5.times.10.sup.6 cfu of SL3261 or intraperitoneally (ip) on days 0 and 14 with 0.1 ml of a mixture of 10 .mu.g V and 10 .mu.g F1 adsorbed to Alhydrogel. An untreated group of 10 age-matched mice were used as controls.

On day 7, five mice from the groups receiving SL3261/placFVe or SL3261 were sacrificed and their spleens were removed. The organs were homogenised in 5 ml of PES with a stomacher (Seward Medical Ltd) for 30 sec. The homogenates were serially diluted in PBS and inoculated on to L-agar or L-agar containing 100 .mu.gml.sup.-1 ampicillin to determine the number of bacteria per spleen.

Measurement of serum antibody titre.

On day 42, blood was sampled from the tail vein of mice immunised with SL32611placFV6 and pooled. The serum anti-V and anti-F1 titres were measured by a modified ELISA (Williamson, E D and R W Titball. 1993. Vaccine 11:1253-1258). Briefly, V (5 .mu.gml.sup.-1 in PBS) or Fl (2 ugml.sup.-1) were coated on to a microtitre plate and the test sera were serially diluted in duplicate on the plate. Bound antibody was detected using peroxidase labelled conjugates of anti-mouse polyvalent Ig. The titre of specific antibody was estimated as the maximum dilution of serum giving an absorbance reading greater than 0.1 units, after subtraction of the absorbance due to non-specific binding detected in the control sera. The serum antibody titre was also determined for all groups of mice prior to challenge with Y. pestis.

On day 42, the anti-V and anti-F1 titres of mice receiving SL3261/placFV6 were 1:5120 and 1:2560, respectively.

Challenge with Y. pestis.

On day 57, groups of 5 or 7 mice from the immunised and control groups were challenged subcutaneously with 0.1 ml aliquots of Y. pestis strain GB containing 7.36.times.10.sup.2 or 7.36.times.10.sup.4 cfu. Strain GB was isolated from a fatal human case of plague and has a median lethal dose (MLD) of <1 cfu in Balbic mice by the s.c. route (Russell, P et al. 1995. Vaccine 13: 1551-1556). The mice were observed for 14 days and, where appropriate, the time to death was recorded. A post-mortem was carried out on all animals where possible. To test for the presence of Y. pestis, samples of blood, liver and spleen were smeared on to Congo Red agar and incubated at 28.degree. C. for 48 h.