Title of Invention

A N. MENINGITIDIS BACTERIUM, MENINGOCOCCAL VESICLE OBTAINED BY CULTURING SAID BACTERIUM AND PHARMACEUTICAL COMPOSITION COMPRISING SAID VESICLE

Abstract

Knockout of the meningococcal mltA homolog gives bacteria that spontaneously release vesicles that are rich in immunogenic outer membrane proteins and that can elicit cross-protective antibody responses with higher bactericidal titres than OMVs prepared by normal production processes. Thus the invention provides a bacterium having a knockout mutation of its mltA gene. The invention also provides a bacterium, wherein the bacterium: (i) has a cell wall that includes peptidoglycan, and (ii) does not express a protein having the lytic transglycosylase activity of MltA protein. The invention also provides compositions comprising vesicles that, during culture of bacteria of the invention, are released into the culture medium.

Full Text

WO 2006/046143 PCT/IB2005/003494 IMMUNOGENIC BACTERIAL VESICLES WITH OUTER MEMBRANE PROTEINS All documents cited herein are incorporated by reference in their entirety. TECHNICAL FIELD This invention is in the field of vesicle preparation for immunisation purposes. BACKGROUND ART One of the various approaches to immunising against N.meningitidis infection is to use outer membrane vesicles (OMVs). An efficacious OMV vaccine against serogroup B has been produced by the Norwegian National Institute of Public Health [e.g. ref 1] but, although this vaccine is safe and prevents MenB disease, its efficacy is limited to the homologous strain used to make the vaccine. The 'RIVM' vaccine is based on OMVs containing six different PorA subtypes. It has been shown to be immunogenic in children in phase H clinical trials [2]. Reference 3 discloses a vaccine against different pathogenic serotypes of serogroup B meningococcus based on OMVs which retain a protein complex of 65-kDa. Reference 4 discloses a vaccine comprising OMVs from genetically-engineered meningococcal strains, with the OMVs comprising: at least one Class 1 outer-membrane protein (OMP) but not comprising a Class 2/3 OMP. Reference 5 discloses OMVs comprising OMPs which have mutations in their surface loops and OMVs comprising derivatives of meningococcal lipopolysaccharide (LPS). As well as serogroup B N.meningitidis, vesicles have been prepared for other bacteria. Reference 6 discloses a process for preparing OMV-based vaccines for serogroup A meningococcus. References 7 and 8 disclose vesicles from N.gonorrh.oeae. Reference 9 discloses vesicle preparations from N.lactamica. Vesicles have also been prepared from Moraxella catarrhalis [10,11], Shigellaflexneri [12,13], Pseudomonas aeruginosa [12,13], Porphyromonas gingivalis [14], Treponema pallidum [15], Haemophilus influenzae [16 & 21] and Helicobacterpylori [17]. The failure of OMVs to elicit cross-protection against, non-homologous strains is not well understood, particularly as most N.meningitidis isolates share a small number of conserved protective surface antigens that, if present in OMVs, would be expected to provide broad protective coverage. One possible explanation for the failure is the existence of variable immune-dominant surface antigens that prevent the conserved antigens from exerting their protective action, and the presence of immune-dominant hyper-variable proteins such as PorA has been extensively documented and demonstrated. Other possible explanations are that the methods for OMV preparation result in contamination with cytoplasmic and/or inner membrane proteins that dilute the protective outer membrane proteins, or that antigens are lost by the detergent extraction. There have been various proposals to improve OMV efficacy. Reference 18 discloses compositions comprising OMVs supplemented with transferrin binding proteins (e.g. TbpA and TbpB) and/or Cu,zn-superoxide dismutase. Reference 19 discloses compositions comprising OMVs supplemented by various proteins. Reference 20 discloses preparations of membrane vesicles obtained from -1- WO 2006/046143 PCT/IB2005/003494 N.meningitidis with a modified fur gene. Reference 21 teaches that nspA expression should be up-regulated with concomitant porA and cps knockout Further knockout mutants of N.meningitidis for OMV production are disclosed in references 21 to 23. In contrast to these attempts to improve OMVs by changing expression patterns, reference 24 focuses on changing the methods for OMV preparation, and teaches that antigens such as NspA can be retained during vesicle extraction by avoiding the use of detergents such as deoxycholate. It is an object of the invention to provide further and improved vesicle preparations, together with processes for their manufacture. In particular, it is an object of the invention to provide vesicles which retain important bacterial immunogenic components from N.meningitidis. DISCLOSURE OF THE INVENTION The invention is based on the surprising discovery that disruption of the pathways involved in degradation of peptidoglycan (the murein layer) gives bacteria that release vesicles into then- culture medium, and that these vesicles are rich in immunogenic outer membrane proteins and can elicit broad-ranging bactericidal immune responses. The vesicles are different from the OMVs that can be prepared by disrupting whole bacteria (e.g. by sonication and sarkosyl extraction [25]), and can be prepared without even disrupting bacterial cells e.g. simply by separating the vesicles from the • bacteria by a process such as centrifugation. In particular, the inventors have found that knockout of the meningococcal mltA homolog (also referred ro as 'GNA33' or 'NMB0033' [26]) leads to the spontaneous release of vesicles that are rich'-, in immunogenic outer membrane proteins and that can elicit broadly cross-protective antibody responses with higher bactericidal titres than OMVs prepared by normal production processes. This enhanced efficacy is surprising for two reasons: first, the NMB0033 protein has previously been reported to be highly effective in raising bactericidal antibodies (e.g. see table 1 of ref. 196) and to be a strong vaccine candidate (e.g. see table 2 of ref. 27), with a recommendation in reference 28 that it should be upregulated for vesicle production, so its loss would a priori be expected to reduce bactericidal efficacy rather than to increase it; second, the knockout strains do not have the correct topological organisation of the cellular membrane, and the main constituent proteins of normal OMVs (e.g. the PorA, PBS, class 4 and class 5 outer membrane proteins) had previously been reported to be released into culture medium [25], The inventors have now found that the previously- reported release does not involve secretion of discrete proteins, but that instead the outer membrane proteins are released in the form of vesicles. These vesicles are advantageous over OMVs prepared by prior art means because they are released spontaneously into the culture medium and can thus be prepared simply and efficiently without the complicated and time-consuming disruption and purification methods that are normally used for preparing OMVs. -2- WO 2006/046143 PCT/IB2005/003494 Thus the invention provides a bacterium having a knockout mutation of its mltA gene. The bacterium preferably also has a knockout mutation of at least one further gene e.g. iheporA and/orporB and or ipxA genes. ' The invention also provides a bacterium, wherein: (i) the bacterium has a cell wall that includes peptidoglycan; and (ii) the bacterium does not express a protein having the lytic transglycosylase activity of MltA protein. The bacterium is preferably a mutant bacterium i.e. the bacterium is a mutant strain of a wild-type species that expresses MltA protein. The bacterium preferably also does not express at least one further protein e.g. the PorA and/or PorB and/or LpxA proteins. Preferred bacteria of the invention are in the genus Neisseria, such as N.meningitidis, and so the invention provides a meningococcus bacterium having a knockout mutation of its gna33 gene. A preferred meningcoocus is gra33 lpxA PorA. The invention also provides a composition comprising vesicles that, during culture of bacteria of the invention, are released into the culture medium. This composition preferably does not comprise any living and/or whole bacteria. This composition can be used for vaccine preparation. The invention also provides a composition comprising vesicles, wherein the vesicles are present in the filtrate obtainable after filtration through a 0.22μm filter of a culture medium in which a bacterium of the invention has been grown. This composition can be used for vaccine preparation. The invention also provides a meningococcal vesicle, wherein the vesicle does not include at least one of {i.e. does not include 1, 2 or 3 of) MinD, FtsA, and/or phosphoenolpyruvate synthase. The invention also provides a meningococcal vesicle, wherein the vesicle does not include at least one of NMB proteins 0126, 0154, 0J57, 0171, 0219, 0359, 0387, 0426, 0595, 0617, 0618, 0631, 0757, 0763, 0875, 0876, 0943, 0946, 0957, 1131, 1252, 1323, 1341, 1445, 1497, 1574, 1576, 1869, 1934, 1936,2096 and/or 2101. The invention also provides a meningococcal vesicle, wherein the vesicle is substantially free from ribosomes. The invention also provides a meningococcal vesicle, wherein the vesicle is substantially free from any aminoacid-tRNA-synthetases. The invention also provides a meningococcal vesicle, wherein the vesicle is substantially free from any enzyme from the Krebs cycle. These vesicles will also not include MltA (because of the knockout mutation), but will include outer membrane proteins. The vesicles may include trimeric outer membrane proteins (Figure 13). The invention also provides a meningococcal vesicle, which includes the following 47 proteins: NMB0035, NMB0044, NMB0086, NMBOO88, NMB0109, NMB0124, NMB0138, NMB0182, NMB0204, NMB0278, NMB0294, NMB0313, NMB0345, NMB0346, NMB0382, NMB0460, NMB0461, NMB0550 , NMB0554, NMB0623, NMB0634, NMB0663, NMB0703, NMB0787, NMB0873, NMB0928, NMB1030, NMB1053, NMB1057, NMB1126, NMB1285, NMB1301, NMB1332, NMB1429, NMB1483, NMB1533, NMB1567, NMB1612, NMB1710, NMB1870, NMB1898, NMB1949, NMB1961, KMB1972, NMB1988, NMB2039 andNMB2091. -3- WO 2006/046143 PCT/IB2005/003494 The invention also provides a meningococcal vesicle, -which includes one or more (i.e. 1, 2, 3,4, 5, 6, 7, 8, 9, 10,11,12,13,14,15,16,17,18, or 19) of the following 19 proteins: NMB0044, NMB0086, NMB0204, NMB0278, NMB0294, NMB0313, NMB0345, NMB0346, NMB0460, NMB0550, NMB0873, NMB0928, NMB1030, NMB1057, NMBI483, NMB1870, NMB1898, NMB1961. and/or NMB2091. See also Table 4 below. The invention also provides a composition comprising a first set of vesicles of the- invention and a second set of vesicles of the invention, waerein said first and second sets are prepared from different strains of meningococcus. The invention also provides a process for preparing a mixture of vesicles, comprising: (a) preparing vesicles of the invention from a first meningococcal strain; (b) preparing vesicles of the invention from a second meningococcal strain; and (c) combining the vesicles from (a) and (b). Combining vesicles from different strains can improve the coverage of clinical strains. The invention also provides a process for preparing bacterial vesicles, comprising the steps of: (i) culturing a MltA bacterium in a culture medium such that the bacterium releases vesicles into said medium; and (ii) collecting the vesicles from said medium. The MltA bacterium is preferably a ? MltA knockout mutant The vesicles can be collected by size separation (e.g. filtration, using a filter which allows the vesicles to pass through but which does not allow intact bacteria to pass through), which can conveniently be performed after centrifugation to preferentially pellet cells relative to the smaller vesicles (e.g. low speed centrifugation). Peptidoglycan metabolism Peptidoglycan (also known as murein, mucopeptide or glycosaminopeptide) is a heteropolymer found in the cell wall of most bacteria. Peptidoglycan is the component that is primarily responsible for the mechanical strength of the bacterial cell wall and for maintaining cellular shape. In Gram-positive bacteria it is the major component of the cell wall. In Gram-negative bacteria it occurs as a layer between the cytoplasmic and outer membranes, and is covalently linked to the outer membrane via the Braun lipoprotein. Peptidoglycan consists mainly of linear heteropolysaccbaride backbone chains that are cross-linked by 'stem' peptides to form a lattice structure. It is a polymer so large that it can be thought of as a single immense covalently linked molecule. In E. coli the saccharide backbone is formed from alternating N-acetylglucosamine (GlcNAc) and N-acetylmuramic acid (MurNAc) residues. A MurNAc residue may be linked to a stem tetrapeptide. Cross-links between backbone chains are usually formed direcily between D-alanine in one stem peptide and a meso DAP of another. The E.coli structure is typical for Gram-negative bacteria, but there is more variation within Gram- positive bacteria e.g. in S.aureus 30-50% of the muramic acid residues are not acetylated, the stem peptide often has L-lysine in place of meso-DAP and isoglutamine in place of D-glutamate, and cross-links can occur between stem peptides. -4- WO 2006/046143 PCT/IB2005/003494 The initial step in E.coli peptidoglycan biosynthesis is the formation of the UDP derivative of GlcNAc, which occurs in the cytoplasm. Some UDP-GlcNAc is converted to UDP-MurNAc in a reaction of UDP-GlcNAc ' and phosphoenolpyruvate (PEP), catalysed by PEP:UDP-GlcNAc enolpyruvyl transferase. Still within the cytoplasm, amino acids are added sequentially to UDP-MurNAc to form a UDP-MurNAc-pentapeptide known as the 'Park nucleotide' that includes a terminal D-alanyl-D-alanine. The Park nucleotide is then transferred to bactoprenol monophosphate in the cytoplasmic membrane, where UDP-GlcNAC is also added to make a bactoprenol- disaccharide-pentapeptide subunit The disaccharide-pentapeptide subunit is then transferred into the periplasmic region, with bactoprenol-pyrophosphate remaining in the membrane. Within the periplasm the transferred subunit is inserted into a growing peptidoglycan. To allow ceil division, changes in shape, and import/export of large complexes (e.g. during conjugation) then peptidoglycan degradation must occur. In E.coli this degradation is caused by enzymes referred to as murein hydrolases [29], which as a family includes lytic transglycosylases {mltA, mltB, mltC, mltD, slt70, emtA), endopeptidases (pbp4, pbp7, mepA) and amidases (amiC). Muraonidases such as lysozyme cleave the same B-(l-4)-glycosidic linkages between MurNAc and GlcNAc residues; unlike muramidases, however, the transglycosylases cleave the glycosidic bond with concomitant formation of 1,6-anhydromuramoyl residues (AnhMurNAc). The standard peptidoglycan anabolic and catabolic pathways are thus well-characterised, as are the minor variations and modifications that occur between bacteria. The enzymes are well-characterised, and proteins have been readily annotated as being involved in the pathways when newi bacterial genomic sequences have been published. The skilled person can thus easily determine the-enzymes involved in the peptidoglycan metabolic pathways for any given bacterium, can easily identify the enzymes involved, and can easily identify the genes encoding those enzymes. The invention is based on the knockout of the mltA gene, which encodes a membrane-bound lytic transglycosylase. The MltA family is recognised in INTERPRO (entry 'ipr005300') and PFAM (entry 'MltA' or 'PF03562'), and the PFAM record lists MltA proteins in bacteria as diverse as Rhizobiwn loti, Bradyrhizobium j'aponicum, Brucella melitensis, Brucella suis, Rhizobium meliloti, Agrobacteriurn tumefaciens, Zymomonas mobilis, Caulobacter crescentus, Yersinia pestis, Salmonella typhimurium, Buchnera aphidicola, Photorhabdus luminescens, Escherichia coli, Shigella flexneri, Salmonella typhi, Pseudomonas aeruginosa, Pseudomonas putida, Pseudomonas syringae, Coxiella bumetii, Vibrio chohrae, Vibrio vulnificiis, Vibrio parahdemolyticus, Haemophilus ducreyi, Pasteurella mtiltocida, Chromobacterium violaceum, Neisseria meningitidis; Neisseria gonorrhoeae, Bordetella parapertussis, Bordetella bronchiseptica, Bordetella pertussis, Nitrosomonas europaea, Ralstonia solanaceanan, Synechococcus elongatus, Gloeobacter violacens, and Leptospira interrogans. -5- WO 2006/046143 PCT/IB2005/003494 Preferred bacteria for MtA knockout are in the Naisseria genus, vnihN.memngitidis being the most preferred 'bacterium. The MltA gene in serogroup B N.meningitidis has been referred to in the literature as 'GNA33' [25,26,196],' and an example sequence has GenBank accession number 'AF226391.1\ The MltA gene in serogroup A (cNMA0279') has GenBank accession number NP_283118.1. Aligned polymorphic forms of meningococcal MltA can be seen in figures 7 and 18 of reference 30. Two full genome sequences of N.meningitidis are available [31,32]. For any given strain of N.meningitidis, therefore, the skilled person will be able to identify the mltA gene. For meningococcus, the knocked-out mltA gene is preferably the gene which, in the wild-type strain, has the highest sequence identity to SEQ ID NO: 1 herein. MltA is a lipoprotein in meningococcus [26]. Knockout of mltA can result in reduced virulence, abnormal cell separation, abnormal cell morphology, undivided septa, double septa, cell clustering and sharing of outer membranes [25]. At the same time, however, the knockout mutation has surprisingly been found to give bacteria that can spontaneously produce vesicles that are immunogenic and enriched in outer membrane proteins. Bacteria The bacterium from which vesicles are prepared may be Gram-positive, but it is preferably Gram- negative. The bacterium may be from genus Moraxella, Shigella, Pseudomonas, Treponema, Porphyromonas or Helicobacter (see above for preferred species) but is preferably from the Neisseria genus. Preferred Neisseria species are N.meningitidis and N.gonorrhoeae. Within N.meningitidis, any of serogroups A, C, W135 and Y may be used, but it is preferred to prepare vesicles from serogroup B. Where relevant, the meningococcus can be of any serotype (e.g. 1, 2a, 2b, 4, 14, 15, 16, etc.), of any serosubtype (P1.2; P1.4; P1.5; Pl.5,2; Pl.7,16; P1.7,16b; P1.9; Pl.9,15; Pl.12,13; PI.13; P1.14; P1.15; PI.21,16; Pl.22,14; etc.) and of any immunotype (e.g. LI; L3,3,7; L10; etc.), and preferred bacteria include: B:4:P1.4; B:4:P1.15; B:15:P1.7,16. The meningococcus may be from any suitable lineage, including hyperinvasive and hypervirulent lineages e.g. any of the following seven hypervirulent lineages: subgroup I; subgroup III; subgroup IV-i; ET-5 complex; ET-37 complex; A4 cluster; lineage 3. These lineages have been defined by multilocus enzyme electrophoresis (MLEE), but multilocus sequence typing (MLST) has also been used to classify meningococci [ref. 33] e.g. the ET-37 complex is the ST-11 complex by MLST, the ET-5 complex is ST-32 (ET-5), lineage 3 is ST-41/44, etc. Preferred strains within serogroup B are MC58, 2996, H4476 and 394/98. In some embodiments of the invention, however, the meningococcus is not strain MC58 and is not strain BZ232. As well as having a knockout of mltA, the bacterium may have one or more knockout mutations of other gene(s). To reduce pyrogenic activity, for instance, the bacterium should have low endotoxin (LOS/LPS) levels, and this can be achieved by knockout of enzymes involved in LPS biosynthesis. Suitable mutant bacteria are already known e.g. mutant Neisseria [34,35] and mutant Helicobacter -6- WO 2006/046143 PCT/IB2005/003494 [36]. The IpxA mutant of rneningococcus is preferred. Processes for preparing LPS-depleted outer membranes from Gram-negative bacteria are disclosed in reference 37. In N.meningitidis, a preferred further knockout is the PorA class I outer membrane protein. Advantageously, such knockouts will not display the immunodominant hypervariable strain-specific PorA protein, thereby focusing a recipient's immune response on other antigens. In a specific aspect, the invention provides a N.meningitidis bacterium, comprising both a knockout mutation of MltA and a knockout mutation of PorA. The bacterium can also carry further knockout mutations e.g. in LOS/LPS synthetic pathways {e.g. IpxA), immunodominant variable proteins, PorB, OpA, OpC, etc. As well as having knockouts of particular endogenous genes, the bacterium may express one or more genes that are not endogenous. For example, the invention may use a recombinant strain that expresses new genes relative to the corresponding wild-type strain. Although it is preferred to knockout PorA expression, in an alternative approach it is possible to engineer a rneningococcus to express multiple PorA subtypes {e.g. 2, 3, 4, 5 or 6 of PorA subtypes: PI.7,16; PI.5,2; PI. 19,15; P1.5c,10; Pl.12,13; and P1.7h,4 [e.g. refs. 38, 39]). Expression of non-endogenous genes in this way can be achieved by various techniques e.g. chromosomal insertion (as used for introducing multiple PorA genes [40]), knockin mutations, expression from extra-chromosomal vectors {e.g. from plasmids); etc. As well as down-regulating expression of specific proteins, the bacterium may over-express (relative * to the corresponding wild-type strain) immunogens such as NspA, protein 287 [19], protein 741 [41], TbpA [18], TbpB [18], superoxide dismutase [18], etc. The bacterium may also include one or more of the knockout and/or over-expression mutations disclosed in reference 16, 21-24 and/or 42-43. Preferred genes for down-regulation and/or knockout include: (a) Cps, CtrA, CtrB, CtrC, CtrD, FrpB, GalE, HtrB/MsbB, LbpA, LbpB, LpxK, Opa, Ope, PilC, PorA, PorB, SiaA, SiaB, SiaC, SiaD, TbpA, and/or TbpB [16]; (b) CtrA, CtrB, CtrC, CtrD, FrpB, GalE, HtrB/MsbB, LbpA, LbpB, LpxK, Opa, Ope, PhoP, PilC, PmrE, PmrF, PorA. SiaA, SiaB, SiaC, SiaD, TbpA, and/or TbpB [21]; (c) ExbB, ExbD, rrnpM, CtrA, CtrB, CtrD, GalE, LbpA, LpbB, Opa, Ope, PilC, PorA, PorB, SiaA, SiaB, SiaC, SiaD, TbpA, and/or TbpB [42]; and (d) CtrA, CtrB, CtrD, FrpB, OpA, OpC, PilC, PorA, PorB, SiaD, SynA, SynB, and/or SynC [43]. For meningococcai compositions, the selection criteria of reference 44 may be used. Preferred vesicles are prepared from menirtgococci having one of the following subtypes: PI .2; Pl.2,5; P1.4; P1.5; Pl.5,2; P1.5,c; P1.5c,10; Pl.7,16; P1.7,16b; P1.7h,4; P1.9; PI.15; Pl.9,15; Pl.12,13; PI.13; PI.14; Pl.21,16; Pl.22,14. The meningococcus is preferably in serogroup B. Vesicles may also be prepared from the Escherichia genus, such as from the E.coli species. E.coli strains have traditionally been classified as either commensal or pathogenic, and pathogenic strains are then sub-classified as intestinal or extraintestinal strains. Classification may also be based on the -7-V WO 2006/046143 PCT/IB2005/003494 'K' antigens. The best-studied 'K' antigen is 'K' which is considered to be the major determinant of virulence among those strains of E.coli that cause neonatal meningitis. Vesicles of the invention can be prepared from any of these E.coli strains, but are preferably from a pathogenic strain, including an extraintestinal pathogenic ('ExPEC [45]) strain, a uropathogenic (UPEC) strain or a meningitis/sepsis-associated (MNEC) strains. Genome sequences of pathogenic strains are available in the databases tinder accession numbers AE005174, BA000007 and NC-004431. Rather than use a mltA knockout, it may be preferred to knockout one or more of the components of the E.coli Tol-Pal complex [46], such as tolA, tolQ, tolB, pal and/or tolR. Knockout of tolR is preferred. The meningococci do not have a homolog of the Tol-Pal system. Vesicle compositions The invention provides the vesicles that arc spontaneously released into culture medium by bacteria of the invention. These vesicles are distinct from the vesicles that can be prepared artificially from the same bacteria, such as the sarkosyl-extracted OMVs prepared in reference 25 from '7GNA33' meningococci. They are also distinct from microvesicles (MVs [47]) and 'native OMVs' ('NOMVs' [64]), although vesicles of the invention seem to be more similar to MVs and NOMVs than to sarkosyl-extracted OMVs. The vesicles are also distinct from blebs, which are outer-membrane protrusions that remain attached to bacteria prior to release as MVs [48,49]. The vesicles of the invention have a diameter of 50-100nm by electron microscopy, which is smaller than that of artificial meningococcai OMVs (diameter ~270nm [50]). The diameter is roughly the same as that of artificial OMVs that have been heat-denatured (~105nm [50]), but the vesicles of the invention retain antigenicity whereas heat-denatured artificial OMVs lose their antigenicity. Moreover, vesicles of the invention (unlike MVs, OMVs and NOMVs) are substantially free from cytoplasmic contamination. Vesicles of the invention preferably contain no more than 20% by weight of LOS/LPS, measured relative to the total protein (i.e. there should be at least 4x more protein than LOS/LPS, by weight). The maximum LOS/LPS level is preferably even lower than 20% e.g. 15%, 10%, 5% or lower. Unlike the starting culture, the vesicle-containing compositions of the invention will generally be substantially free from whole bacteria, whether living or dead The size of the vesicles of the invention means that they can readily be separated from whole bacteria by filtration through a 0.22um filter e.g. as typically used for filter sterilisation. Thus the invention provides a process for preparing vesicles of the invention, comprising filtering the culture medium, from bacteria of the invention through a filter that retards whole bacteria but that lets the vesicles pass through e.g. a 0.22um filter. Although vesicles will pass through a standard 0.22um filters, these can rapidly become clogged by other material, and so it is preferred to perform sequential steps of filter sterilisation through a series of filters of decreasing pore size, finishing with a standard sterilisation -8- WO 2006/046143 PCT/IB2005/003494 filter (e.g. a Q.22μm filter). Examples of preceding filters would be those with pore size of 0.8μm, 0.45μm, etc The filtrate can be farther treated e.g. by ultracehtrifugation. Vesicles of the invention contain Iipids and proteins. The protein content of meningococcal vesicles has been analysed, and substantially all of the proteins in the vesicles are classified as outer membrane proteins by bioinformatic analysis. Outer membrane proteins seen in the vesicles include: PilE; IgA-specific serine endopeptidase; PorA; FrpB; P1B; etc. Unlike artificial OMVs, which have previously been analysed proteomically [51], the vesicles of the invention were found to lack proteins such as MinD, FtsA and phosphoenolpyruvate synthase. The vesicles also lack MtA. The vesicles of the invention are advantageous when compared to vesicles prepared by disruption of cultured bacteria because no artificial disruption is required. Simple size-based separation can be used to separate bacteria and vesicles, without any need for chemical treatments, etc. As well as being a simpler process, this avoids the risk of denaturation caused by the detergents etc. that are used during prior art OMV preparative processes. As mentioned above, vesicles of the invention may be similar to microvesicles (MVs) and 'native OMVs' ('NOMVs'), which are naturally-occurring membrane vesicles that form spontaneously during bacterial growth and are released into culture medium. MVs can be obtained by culturing Neisseria in broth culture medium, separating whole cells from the broth culture medium (e.g. by filtration or by low-speed centrifugation to pellet only the cells and not the smaller vesicles) and then collecting the MVs that are present in the cell-depleted medium {e.g. by filtration, by differential precipitation or aggregation of MVs, by high-speed centrifugationto pellet the MVs). Strains for use in production of MVs can generally be selected on the basis of the amount of MVs produced in culture. References 52 and 53 describe Neisseria with high MV production. Vesicle combinations The invention allows the production of irnmunogenic vesicles from a bacterium of choice. The bacterium will typically have been generated by mutation of a chosen starting strain. "Where there are multiple starring strains of interest then the invention provides methods for preparing vesicles from each of the strains, and the different vesicles can be combined. This combination strategy is particularly .useful for bacteria where strain-to-strain variation means that a single strain usually does not offer clinically-useful protection e.g. serogroup B meningococcus. Thus the invention provides a composition comprising a mixture of n sets of vesicles of the invention, prepared from n different strains of a bacterium. The value of n can be 1, 2, 3, 4, 5, etc! The different strains can be in the same or different serogroups. Preferred mixtures of serogroups include: A+B; A+C; A+W135; A+Y; B+C; B+W135; B+Y; C+W135; C+Y; W135+Y; A+B+C; A+B+W135; A+B+Y; A+C+W135; A+C+Y; A+W135+Y; B+C+W135; B+C+Y; C+W135+Y; A+B+C+W135; A+B+C+Y; B+C+W135+Y; and A+B+C+W135+Y. -9- WO 2006/046143 PCT/IB2005/003494 The invention also provides a kit comprising vesicles of the invention prepared from n different strains of a "bacterium. The vesicles can be kept and stored separately in the kit until they are required to be used together e.g. as an admixture, or for simultaneous separate or sequential use. The invention also provides a process comprising: preparing n sets of vesicles of the invention, one from each of n different strains of a bacterium; and combining then sets of vesicles. The different sets can be combined into a kit or into an admixture. The invention also provides the use of vesicles from a first strain of a bacterium in the manufacture of a medicament for immunising a patient, wherein the medicament is administered simultaneously separately or sequentially with vesicles from a second strain of the bacterium. The invention also the use of vesicles from a first strain of a bacterium in the manufacture of a medicament for immunising a patient, wherein the patient has been pre-immunised with vesicles from a second strain of the bacterium. The bacterium is preferably N.meningitidis, and is more preferably from serogroup B. The different strains may be selected according to various criteria. Example criteria include: subtype and/or serosubtype [e.g. ref. 47]: immunotype; geographical origin of the strains; local prevalence of clinical strains; hypervirulent lineage e.g. one or more of subgroups I, IH and IV-1, ET-5 complex, ET-37 complex, A4 cluster and lineage 3; multilocus sequence type (MLST) [54]. Preferred criteria for selecting strains are: selection of more than one PorB serotype (class 2 or 3 OMP); selection of more than one PorA serosubtype (class 1 OMP); selection of more than one different immunotype (lipopolysaccharide or lipoollgosaccharide); selection of more than one of the three different NMB1870 variants [55]. NMB1870 is seen in the vesicles of the invention, shows distinct variants, and is a good candidate antigen for vaccination [55-57]. A combination of vesicles covering two or three different NMB1870 variants is particular advantageous. As well as being selected from different meningococcal strains, vesicles can be selected from, different pathogens. Thus the invention provides a composition comprising a mixture of n sets of vesicles of the invention, prepared from n different species of bacteria. Similarly, the invention provides a kit comprising vesicles of the invention prepared from n different species of bacteria, and provides a process comprising the step of preparing n sets of vesicles of the invention, one from each of n different species of bacteria. MltA expression Bacteria of the invention do not possess functional MltA enzymatic activity. Prevention of MltA protein expression can be achieved in two main ways: removal or disruption of the endogenous mltA gene (including its control regions) to give a MltA" strain; or suppression of MltA expression in a MltA+ strain. It is preferred to use a MltA- strain. -10- WO 2006/046143 PCT/IB2005/003494 MltA- strains can be constructed by conventional knockout techniques. Techniques for gene knockout are well known, and meningococcus knockout mutants of have been reported previously [e.g. refs. 25 & 58-60]. The knockout is preferably achieved by deletion of at least a portion of the coding region (preferably isogenic deletion), but any other .suitable technique may be used e.g. deletion or mutation of the promoter, deletion or mutation of the start codon, etc. The bacterium may contain a marker gene in place of the knocked out gene e.g. an antibiotic resistance marker. Where suppression of expression from an endogenous mltA gene is used then techniques such as antisense inhibition and inhibitory RNA can be used, although these techniques are moTe typically used in eukaryotic hosts, in the resulting bacterium, mRNA encoding the knocked-out protein will be substantially absent and/or its translation will be substantially inhibited {e.g. to less than 1% of the itvzl of expression tinat would bs seen in the absence of suppression). As an alternative to knockout or suppression of expression, site-directed rnutagenesis of the endogenous mltA gene can be used. Reference 61 discloses mutants of meningococcal MltA in which residues Glu255, Glu323 and Asp362 were mutated and then tested for MltA catalytic activity. An E255G mutant of showed a 50% reduction in activity, and an E323G mutant showed a 70% reduction in activity. Mutagenesis of specific residues within the MltA coding region can therefore be used as a technique to knockout the lyric transglycolase enzymatic activity without knocking out the coding region. Whichever technique (or combination of techniques) is chosen, the resulting bacterium will be substantially free from MltA enzymatic activity. Pharmaceutical compositions The invention provides a pharmaceutical composition comprising (a) vesicles of the invention and (b) a pharmaceutically acceptable carrier. The invention also provides a process for preparing such a composition, comprising the step of admixing vesicles of the invention with a pharmaceutically acceptable carrier. Typical 'pharmaceutically acceptable carriers' include any carrier that does not itself induce the production of antibodies harmful to the individual receiving the composition. Suitable carriers are typically large, slowly metabolised macromolecules such as proteins, polysaccharides, polylactic acids, polyglycolic acids, polymeric amino acids, amino acid copolymers, and lipid aggregates (such as oil droplets or liposomes). Such carriers are well known to those of ordinary skill in the art. The vaccines may also contain diluents, such as water, saline, glycerol, etc. Additionally, auxiliary substances, such as wetting or emulsifying agents, pH buffering substances, sucrose, and the like, may be present Sterile pyrogen-free, phosphate-buffered physiologic saline (e.g. pH 7.4) is a typical carrier. A thorough discussion of pharmaceutically acceptable excipients is available in reference 62. -11- WO 2006/046143 PCT/IB2005/003494 Compositions of the invention will typically be in aqueous form (i.e. solutions or suspensions) rather than in a dried form (e.g. lyophilised). Aqueous compositions are also suitable for reconstituting other vaccines from a lyophilised form (e.g. a lyophilised Hib conjugate vaccine, a lyophilised meningococcal conjugate vaccine, etc.). Where a composition of the invention is to be used for such extemporaneous reconstitution, the invention provides a kit, which may comprise two vials, or may comprise one ready-filled syringe and one vial, with the aqueous contents of the syringe being used to reactivate the dried contents of the vial prior to injection. Compositions of the invention may be presented in vials, or they may be presented in ready-filled syringes. The syringes may be supplied with or without needles. Compositions may be packaged in unit dose form or in multiple dose form. A syringe will generally include a single dose of the composition, whereas a vial may include a single dose or multiple doses. For multiple dose forms, therefore, vials, are preferred to pre-fiiled syringes. Effective dosage volumes can be routinely established, but a typical human dose of the composition has a volume of about 0.5ml e.g. for intramuscular injection. The RTVM OMV-based vaccine was administered in a 0.5ml volume [63] by intramuscular injection to the thigh or upper arm. Similar doses may be used for other delivery routes e.g. an intranasal OMV-based vaccine for atomisation may have a volume of about lOOul or about 130Μ1 per spray. [64], with four sprays administered to give a total dose of about 0.5ml. The pH of the composition is preferably between 6 and 8, and more preferably between 6.5 and 7.5 (e.g. about 7 or about 7.4). The pH of the RTVM OMV-based vaccine is 7.4 [65], and a pH (preferably use of a buffer e.g. a Tris buffer, a phosphate buffer, or a bistidine buffer. Compositions of the invention will generally include a buffer. If a composition comprises an aluminium hydroxide salt, it is preferred to use a bistidine buffer [66] e.g. at between 1-l0mM, preferably about 5mM. The RIVM OMV-based vaccine maintains pH by using a l0mM Tris/HCl buffer. The composition may be sterile and/or pyrogen-free. Compositions of the invention may be isotonic with respect to humans. Compositions of the invention are immunogenic, and are more preferably vaccine compositions. Vaccines according to the invention may either be prophylactic (i.e. to prevent infection) or therapeutic (i.e. to treat infection), but will typically be prophylactic. Immunogenic compositions used as vaccines comprise an immunologically effective amount of antigen(s), as well as any other components, as needed. By 'immunologically effective amount', it is meant that the administration of that amount to an individual, either in a single dose or as part of a series, is effective for treatment or prevention. This amount varies depending upon the health and physical condition of the individual to be treated, age, the taxonomic group of individual to be treated (e.g. non-human primate, primate, etc.), the capacity of the individual's immune system to synthesise antibodies, the degree of protection desired, the formulation of the vaccine, the treating doctor's assessment of the medical -12- WO 2006/046143 PCT/IB2005/003494 situation, and other relevant factors. It is expected that the amount will fall in a relatively broad range that can be determined through, routine trials. The antigen content of compositions of the invention will generally be expressed in terms of the amount of protein per dose. A dose of about 0.9 mg protein per ml is typical for OMV-based intranasal vaccines [64]. The MeNZB™ OMV-based vaccine contains between 25 and 200 μg of protein per millilitre e.g. between 45 and 90 jig/ml, or 50±10 p.g/ml. Compositions of the invention preferably include less than 100μg/ml of OMV per strain of bacterium. Meningococci affect various areas of the body and so the compositions of the invention may be prepared in various forms. For example, the compositions may be prepared as injectables, either as liquid solutions or suspensions. The composition may be prepared for pulmonary administration e.g. as an. inhaler, using a fine powder or a spray. The composition may be prepared as a suppository or pessary. The composition may be prepared for nasal, aural or ocular administration e.g. as spray, drops, gel or powder [e.g. refs 67 & 68]. Compositions of the invention may include an antimicrobial, particularly when packaged in multiple dose format. Antimicrobials such as thiomersal and 2-phenoxyethanol are commonly found in vaccines, but it is preferred to use either a mercury-free preservative or no preservative at all. Compositions of the invention may comprise detergent e.g. a Twcen (polysorbate), such as Tween 80. Detergents are generally present at low levels e.g. Compositions of the invention may include sodium salts {e.g. sodium chloride) to give tonicity. A concentration of 10+2 mg/ml NaCl is typical. The concentration of sodium chloride is preferably greater than 7.5 mg/ml. Compositions of the invention will generally be administered in conjunction with other immunoregulatory agents. In particular, compositions will usually include one or more adjuvants, and the invention provides a process for preparing a composition of the invention, comprising the step of admixing vesicles of the invention with an adjuvant e.g. in a pharmaceutically acceptable carrier. Suitable adjuvants include, but are not limited to: A. Mineral-containing compositions Mineral containing compositions suitable for use as adjuvants in the invention include mineral salts, such as aluminium salts and calcium salts. The invention includes mineral salts such as hydroxides (e.g. oxyhydroxides), phosphates {e.g. hydroxyphosphates, orthophosphates), sulphates, etc. [e.g. see chapters 8 & 9 of ref. 69], or mixtures of different mineral compounds, with the compounds taking any suitable form {e.g. gel, crystalline, amorphous, etc.), and with adsorption being preferred. The mineral containing compositions may also be formulated as a particle of metal salt [70]. A typical aluminium phosphate adjuvant is amorphous aluminium hydroxyphosphate with PO4/AI molar ratio between 0.84 and 0.92, included at 0.6mg Al3+/ml. Adsorption with a low dose of -13- WO 2006/046143 PCT/IB2005/003494 aluminium phosphate may be used e.g. between 50 and 100p.g Al3+ per conjugate per dose. Where an aluminium phosphate it used and it is desired not to adsorb an antigen to the adjuvant, this is favoured by including free phosphate ions in solution (e.g. by the use of a phosphate buffer). The RTVM vaccine was tested with adsorption to either an aluminium phosphate or an aluminium hydroxide adjuvant, and the aluminium phosphate adjuvant was found to give superior results [65]. Hie MeNZB™, MenBvac™ abd VA-MENINGOC-BC™ products all include an aluminium hydroxide adjuvant A typical dose of aluminium adjuvant is about 3.3 mg/ml (expressed as Al3+ concentration). B. - Oil Emulsions Oil emulsion compositions suitable for use as adjuvants in the invention include squalene-water emulsions, such as MF59 [Chapter 10 of ref. 69; see also ref. 71] (5% Squalene, 0.5% Tween 80, and 0.5% Span 85, formulated into submicron particles using a microfluidizer). Complete Freund's adjuvant (CFA) and incomplete Freund's adjuvant (IFA) may also be used. C. Saponin formulations [chapter 22 of ref. 69 J Saponin formulations may also be used as adjuvants in the invention. Saponins are a heterologous group of sterol glycosides and triterpenoid glycosides that are found in the bark, ieaves, stems, -roots and even flowers of a wide range of plant species. Saponin from the bark of the Quillaia saponaria Molina tree have been widely studied as adjuvants. Saponin can also be commercially obtained from Smilax ornata (sarsaprilla), Gypsophilla paniculata (brides veil), and Saponaria officianalis (soap root). Saponin adjuvant formulations include purified formulations, such as QS21, as well as lipid formulations, such as ISCOMs. QS21 is marketed as Stimulon™. Saponin compositions have been purified using HPLC and RP-HPLC. Specific purified fractions using these techniques have been identified, including QS7, QS17, QS18, QS21, QH-A, QH-B and QH-C. Preferably, the saponin is QS21. A method of production of QS21 is disclosed in ref. 72. Saponin formulations may also comprise a sterol, such as cholesterol [73]. Combinations of saponins and cholesterois can be used to form unique particles called immunostimulating complexs (ISCOMs) [chapter 23 of ref. 69]. ISCOMs typically also include a phospholipid such as phosphatidylethanolamine or phosphatidylcholine. Any known saponin can be used in ISCOMs. Preferably, the ISCOM includes one or more of QuilA, QHA and QHC. ISCOMs are further described in refs. 73-75. Optionally, the ISCOMS may be devoid of extra detergent [76]. A review of the development of saponin based adjuvants can be found in refs. 77 & 78. D. Virosomes and virus-like particles Virosomes and virus-like particles (VLPs) can also be used as adjuvants in the invention. These structures generally contain one or more proteins from a virus optionally combined or formulated with a phospholipid. They are generally non-pathogenic, non-replicating and generally do not contain any of the native viral genome. The viral proteins may be recombinantly produced or isolated from -14- WO 2006/046143 PCT/IB2005/003494 whole viruses. These viral proteins suitable for use in virosomes or VLPs include proteins derived from influenza virus (such as HA or NA), Hepatitis B virus (such as core or capsid proteins), Hepatitis E virus, measles virus, Sindbis virus, Rotavirus, Foot-and-Mouth Disease virus, Retrovirus, Norwalk virus, human Papiiloma virus, HIV, RNA-phages, Q15-phage (such as coat proteins), GA- phage. fr-phage, AP205 phage, and Ty (such as retrotransposon Ty protein pi). VLPs are discussed further in refs. 79-84. Virosomes are discussed further in, for example, ref. 85 E. Bacterial or microbial derivatives Adjuvants suitable for use in the invention include bacterial or microbial derivatives such as non-toxic derivatives of enterobacterial lipopolysaccharide (LPS), Lipid A derivatives, immunpstimulatory oligonucleotides and ADP-ribosylating toxins and detoxified derivatives thereof. Non-toxic derivatives of LPS include monophosphoryl lipid A (MPL) and 3-O-deacylated MPL (3dMpL). 3dMPL is a mixture of 3 de-O-acyiated monophosphoryi lipid A with 4, 5 or 6 acylated chains. A preferred "small particle" form of 3 De-O-acylated monophosphoryi lipid A is disclosed in ref. 86. Such "small particles" of 3dMPL are small enough to be sterile filtered through a 0.22μm membrane [86]. Other non-toxic LPS derivatives include monophosphoryi lipid A mimics, such as aminoalkyl glucosaminide phosphate derivatives e.g. RC-529 [87,88]. Lipid A derivatives include derivatives of lipid A from Escherichia coli such as OM-174. OM-174 is described for example in refs. 89 & 90. Immunostimulatory oligonucleotides suitable for use as adjuvants in the invention include nucleotide sequences containing a CpG motif (a dinucleotide sequence containing an unrnethylated cytosine linked by a phosphate bond to a guanosine). Double-stranded RNAs and oligonucleotides containing palindromic or poly(dG) sequences have also been shown to be immunostimulatory. The CpG's can include nucleotide modifications/analogs such as phosphorothioate modifications and can be double-stranded or single-stranded. References 91, 92 and 93 disclose possible analog substitutions e.g. replacement of guanosine with 2'-deoxy-7-deazaguanosine. The adjuvant effect of CpG oligonucleotides is further discussed in refs. 94-99. The CpG sequence may be directed to TLR9, such as the motif GTCGTT or TTCGTT [100]. The CpG sequence may be specific for inducing a Thl immune response, such as a CpG-A ODN, or it may be more specific for inducing a B cell response, such a CpG-B ODN. CpG-A and CpG-B ODNs are discussed in refs. 101-103. Preferably, the CpG is a CpG-A ODN. Preferably, the CpG oligonucleotide is constructed so that the 5' end is accessible for receptor recognition. Optionally;, two CpG oligonucleotide sequences may be attached at their 31 ends to form "immunomers". See, for example, refs. 100 & 104-106. Bacterial ADP-ribosylating toxins and detoxified derivatives thereof may be used as adjuvants in the invention. Preferably,, the protein is derived from E.coli (E.coli heat labile enterotoxin "LT"), cholera ("CT"), or pertussis ("PT"). The use of detoxified ADP-ribosylating toxins as mucosal adjuvants is -15- WO 2006/046143 PCT/IB2005/003494 described in ref. 107 and as parenteral adjuvants in ref. 108. The toxin or toxoid is preferably in the form, of a holotoxin, comprising both A and B subunits. Preferably, the A subunit contains a detoxifying mutation; preferably the B subunit is not mutated. Preferably, the adjuvant is a detoxified LT mutant such as LT-K63, LT-R72, and LT-G192. The use of ADP-ribosylating toxins and detoxified derivaties thereof, particularly LT-K63 and LT-R72, as adjuvants can be found in refs. 109-116. Numerical reference for amino acid substitutions is preferably based on the alignments of the A and B subunits of ADP-ribosylating toxins set forth in ref. 117, specifically incorporated herein by reference in its entirety. F. . Human immunomodulators Hitman immunomodulators suitable for use as adjuvants in the invention include cytokines, such as interleukins (e.g. 3L-1, EL-2, 3L-4, IL-5, IL-6, IL-7, IL-12 [118], etc.) [119], interferons (e.g. interferon-?), macrophage colony stimulating factor, and tumor necrosis factor. G. Bioadhesives and Mucoadhesives Bioadhesives and mucoadhesives may also be used as adjuvants in the invention. Suitable bioadhesives include esterified hyaluronic acid microspheres [120] or mucoadhesives such as cross-linked derivatives of poly(acrylic acid), polyvinyl alcohol, polyvinyl pyrollidone, polysaccharides and carboxyraethylcellulose. Chitosan and derivatives thereof may also be used as adjuvants in the invention [121]. H. Microparticles Microparticles may also be used as adjuvants in the invention. Microparticles {i.e. a particle of ~100nm to ~150um in diameter, more preferably ~200nm to ~30um in diameter, and most preferably ~500nm to ~10um in diameter) formed from materials that are biodegradable and non-toxic (e.g. a poly(a-hydroxy acid), a polyhydroxybutyric ' acid, a polyorthoester, a polyanhydride, a polycaprolactone, etc.), with poly(lactide-co-glycolide) are preferred, optionally treated to have a negatively-charged surface {e.g. with SDS) or a positively-charged surface {e.g. with a cationic detergent, such as CTAB). I. Liposomes ('Chapters 13 & 14 of ref. 69) Examples of liposome formulations suitable for use as adjuvants are described in refe. 122-124. J. Polvoxyethylene ether andpolyoxyethylene ester formulations Adjuvants suitable for use in the invention include polyoxyethylene ethers and polyoxyethylene esters [125]. Such formulations further include polyoxyethylene sorbitan ester surfactants in combination with an octoxynol [126] as well as polyoxyethylene alkyl ethers or ester surfactants in combination with at least one additional non-ionic, surfactant such as an octoxynol [127]. Preferred polyoxyethylene ethers are selected from the following group: polyoxyethylene-9-lauryi ether (laureth 9), polyoxyethylene-9-steoryl ether, polyoxytheylene-8-steoryl ether, polyoxyethylene-4- lauryl ether, polyoxyethylene-35-lauryl ether, and polyoxyethylene-23-lauryl ether. -16- WO 2006/046143 PCT/IB2005/003494 K. Polvphosphazene (PCPP) PCpp formulations are described, for example, in refs. 128 and 129. L. Muramvl peptides Examples of muramyl peptides suitable for use as adjuvants in the invention include N-acetyl- muramyl-L-threonyl-D-isoglutamine (thr-MDP), N-ace^l-normuramyl-L-alanyl-D-isoglutamine (nor- MDP), and N-acetylmuramyl-L-alanyl-D-isoglutaminyl-L-alanine-2-(l'-2'-dipalmitoyl-.sn-grycero-3- hydroxyphosphoryloxy)-ethylamine MTP-PE). M. Imidazoquinolone Compounds. Examples of imidazoqumolone compounds suitable for use adjuvants in the invention include hniquamod and its homologues (e,g. "Resiquimod 3M"), described further in refs. 130 and 131. The invention may also comprise combinations of aspects of one or more of the adjuvants identified above. For example, the following adjuvant compositions may be used in the invention: (1) a saponin and an oil-in-water emulsion [132]: (2) a saponin (e.g. QS21) + a non-toxic LPS derivative (e.g. 3dMPL) [133]; (3) a saponin (e.g. QS21) + a non-toxic LPS derivative (e.g. 3dMPL) + a cholesterol; (4) a saponin (e.g. QS21) + 3dMPL + IL-12 (optionally + a sterol) [134]; (5) combinations of 3dMPL with, for example, QS21 and/or oil-in-watcr emulsions [135]; (6) SAP, containing 10% squalane, 0.4% Tween 80™, 5% pluronic-block polymer L121, and thr-MDP, either microfluidized into a submicron emulsion or vortexed to generate a larger particle size emulsion. (7) Ribi™ adjuvant system (RAS), (Ribi Immunochem) containing 2% squalene, 0.2% Tween 80, and one or more bacterial cell wall components from the group consisting of monophosphorylipid A (MPL), trehalose dimycolate (TDM), and cell wall skeleton (CWS), preferably MPL + CWS (Detox™); and '(8) one or more mineral salts (such as an aluminum salt) + a non-toxic derivative of LPS (such as 3dMPL). Other substances that act as immunostimulating agents are disclosed in chapter 7 of ref. 69. ' The use of aluminium salt adjuvants is particularly preferred, and antigens are generally adsorbed to such salts. It is possible in compositions of the invention to adsorb some antigens to an aluminium hydroxide but to have other antigens in association with an aluminium phosphate, hi general, however, it is preferred to use- only a single salt e.g. a hydroxide or a phosphate, but not both. Not all vesicles need to be adsorbed i.e. some or all can be free in solution. Methods of treatment The invention also provides a method for raising an immune response in a mammal, comprising administering a pharmaceutical composition of the invention to the mammal. The immune response is preferably protective and preferably involves antibodies. The method may raise a booster responsein a patient that has already been primed against N.meningitidis. Subcutaneous and intranasal prime/boost regimes for OMVs are disclosed in ref. 136. The mammal is preferably a human. Where the vaccine is for prophylactic use, the human is preferably a child (e.g. a toddler or infant) or a teenager; where the vaccine is for therapeutic use, the -17- . WO 2006/046143 PCT/IB2005/003494 human is preferably an adult. A vaccine intended for children may also be administered to adults e.g. to assess safety, dosage, immunogenicity, etc. The invention also provides vesicles of the invention for use as a medicament The medicament is preferably able to raise an immune response in a mammal (i.e. it is an immxmogenic composition) and is more preferably a vaccine. The invention also provides the use of vesicles of the invention in the manufacture of a medicament for raising an immune response in a mammal. The invention also the use of vesicles of the invention in the manufacture of a medicament for immunising a patient, wherein the patient has been pre-immunised with at least one of the following: diphtheria toxoid; tetanus toxoid; acellular or cellular pertussis antigens; a conjugated Hib capsular saccharide; hepatitis B virus surface antigen: a conjugated rnenicgococcal capsular saccharide; and/or a conjugated pneumococcal capsular saccharide. These uses and methods are preferably for the prevention and/or treatment of a disease caused by N.meningitidis e.g. bacterial (or, more specifically, meningococcal) meningitis, of septicemia. One way of checking efficacy of therapeutic treatment involves monitoring Neisserial infection after administration of the composition of the invention. One way of checking efficacy of prophylactic treatment involves monitoring irnmune responses against the vesicles' antigens after administration of the composition. Immunogeniciry of compositions of the invention can be determined by administering them to test subjects (eg. children 12-16 months age, or animal models [137]) and then determining standard parameters including serum bactericidal antibodies (SBA) and ELISA titres (GMT). These immune responses will generally be determined around 4 weeks after administration of the composition, and compared to values determined before administration of the composition. A SBA increase of at least 4-fold or 8-fold is preferred. Where more than one dose of the composition is administered, more than one post-administration determination may be made. Preferred compositions of the invention can confer an antibody titre in a patient that is superior to the criterion for seroprotection for an acceptable percentage of human subjects. Antigens wifh an associated antibody titre above which a host is considered tcbe seroconverted against the antigen are well known, and such titres are published by organisations such as WHO. Preferably more than 80% of a statistically significant sample of subjects is seroconverted, more preferably more than 90%, still more preferably more than 93% and most preferably 96-100%. Compositions of the invention will generally be administered directly to a patient Direct delivery may be accomplished by parenteral injection (e.g. subcutaneously, intraperitoneally, intravenously, intramuscularly, or to the interstitial space of a tissue), or by rectal, oral, vaginal, topical, transdermal, intranasal, ocular, aural, pulmonary or other mucosal administration. Intramuscular administration to the thigh or the upper arm is preferred. Injection may be via a needle (e.g. a -18- WO 2006/046143 PCT/IB2005/003494 hypodermic needle), but needle-free injection may alternatively "be used. A typical intramuscular dose is 0.5 ml. Dosage treatment can be a single dose schedule or a multiple dose schedule. Multiple doses may be used in a primary immunisation schedule and/or in a booster immunisation schedule. A primary dose schedule may be followed by a booster dose schedule. Suitable timing between priming doses (e.g. between 4-16 weeks), and between priming and boosting, can be routinely determined. The OMV-based RTVM vaccine was tested using a 3- or 4-dose primary schedule, with vaccination at 0, 2&8 or 0, 1,2&8 months. MeNZB™ is administered as three doses at six week intervals. These schedules can be used according to the invention. The vesicle preparations given at each dose stage can be the same or different In methods of the invention, where a first dose is given at time zero then a second and a ihiru dose may be given over the next two months, and a fourth dose may be given between 11 and 13 months after time zero. The first, second and third doses may comprise vesicles with the same serosubtype as each other, and the fourth dose may comprises vesicles with a different serosubtype from the first three doses. The fourth dose may contain vesicles only with a different serosubtype from the first three doses, or it may contain two types of vesicle, one with a different serosubtype from the first three doses and one with the same subtype. The first, second and third doses are preferably given, at intervals of between 6 and 8 weeks. The fourth dose is preferably given about 1 year after time zero. The patient preferably receives the same quantity of vaccine at each of the four doses. As described above, the invention may involve administration of vesicles from more than one: subtype and/or serosubtype of N.meningitidis [e.g. ref. 47], either separately or in admixture. The invention may be used to elicit systemic and/or mucosal immunity. In general, compositions of the invention are able to induce serum bactericidal antibody responses after being administered to a subject. These responses are conveniently measured in mice and are a standard indicator of vaccine efficacy [e.g. see end-note 14 of reference 196]. Serum bactericidal activity (SBA) measures bacterial killing mediated by complement, and can be assayed using human or baby rabbit complement. WHO standards require a vaccine to induce at least a 4-fold rise in SBA in more than 90% of recipients. MeNZB™ elicits a 4-fold rise in SBA 4-6 weeks after administration of the third dose. Rather than offering narrow protection, compositions of the invention can induce bactericidal antibody responses against more than one hypervirulent lineage of serogroup B. In particular, they can preferably induce bactericidal responses against two or three of the following three hypervirulent lineages: (i) cluster A4; (ii) ET5 complex; and (iii) lineage 3. They may additionally induce bactericidal antibody responses against one or more of hypervirulent lineages subgroup I, subgroup m, subgroup IV-1 or ET-37 complex, and against other lineages e.g. hyperinvasive lineages. This does not necessarily mean that the composition can induce bactericidal antibodies against each and -19- WO 2006/046143 PCT/IB2005/003494 every strain of serogroup B meningococcus within these hypervirulent lineages e.g. rather, for any given group of four of more strains of serogroup B meningococcus within a particular hypervirulent lineage, the antibodies induced by the composition are "bactericidal against at least 50% (e.g. 60%, 70%, 80%, 90% or more) of the group. Preferred groups of strains will include strains isolated in at least four of the following countries: GB, AU, CA, NO, IT, US, NZ, NL, BR, and CU. The serum preferably has a bactericidal titre of at least 1024 (e.g. 210, 211, 212,213, 2M, 215, 216, 217, 218 or higher, preferably at least 214) e.g. the serum is able to kill at least 50% of test bacteria of a particular strain when diluted 1/1024, as described in reference 196. Preferred compositions can induce bactericidal responses against the following strains of serogroup B meningococcus: (i) from cluster A4, strain 961-5945 (B:2b:P1.21,16) and/or strain G2136 (B:-); (ii) from ET-5 complex, strain MC5S (B:15:P1.7,16b) and/or strain 44/76 (B:15:P1.7.16): (iii) from lineage 3, strain 394/98 (B:4:PL4) and/or strain BZ19S (B:NT:-). More preferred compositions can induce bactericidal responses against strains 961-5945, 44/76 and 394/98. Strains 961-5945 and G2136 are both Neisseria MLST reference strains [ids 638 & 1002 in ref. 138]. Strain MC58 is widely available (e.g. ATCC BAA-335) and was the strain sequenced in reference 32. Strain 44/76 has been widely used and characterised (e.g. ref. 139) and is one of the Neisseria MLST reference strains [id 237 in ref. 138; row 32 of Table 2 in ref. 33]. Strain 394/98 was originally isolated in New Zealand in 1998, and there have been several published studies using this strain (e.g. refs. 140 & 141). Strain BZ198 is another MLST reference strain [id 409 in ref. 138; row 41 of Table 2 in ref. 33]. Further antigenic components As well as containing antigenic vesicles of the invention, compositions of the invention may include . further non-vesicular antigens. For example, the composition may comprise one or more of the following further antigens: - a saccharide antigen from N.meningitidis serogroup A, C, W135 and/or Y, such as the oligosaccharide disclosed in ref. 142 from serogroup C or the oligosaccharides of ref. 143. The VA-MENINGOC-BC™ product contains serogroup C polysaccharide. - a saccharide antigen from Streptococcus pneumoniae [e.g. refs. 144-146; chapters 22 & 23 of ref. 153]. - an antigen from hepatitis A virus, such as inactivated virus [e.g. 147, 148; chapter 15 of ref. 153]. - an antigen from hepatitis B virus, such as the surface and/or core antigens [e.g. 148,149; chapter 16 of ref. 153]. - an antigen from hepatitis C virus [e.g. 150]. -20- WO 2006/046143 PCT/IB2005/003494 - an antigen from Bordeiella pertussis, such as pertussis holotoxin (PT) and filamentous haemagglutinin (filA) fromlpertussis, optionally also in combination with pertactin and/or agglutinogens 2 and 3 [e.g. refs. 151 & 152; chapter 21 of ref. 153]. - a diphtheria antigen, such as a diphtheria toxoid [e.g. chapter 13 of ref. 153]. - a tetanus antigen, such as a tetanus toxoid [e.g. chapter 27 of ref. 153]. - a saccharide antigen from Haemophilus influenzae B [e.g. chapter 14 of ref. 153] - an antigen from N.gonorrhoeae [e.g. ref. 154]. - an antigen from Chlamydia pneumordae [e.g. 155-161]. - an antigen from Chlamydia trachomatis [e.g. 162]. - an antigen from Porphyromonas gingivalis [e.g. 163]. - polio antigen(s) [e.g. 164,165; chapter 24 ofref. 153] such as IPV. - rabies antigen(s) [e.g. 166] such as lyophilised inactivated virus [e.g. 167, RabAvert™]. - measles, mumps and/or rubella antigens [e.g. chapters 19,20 and 26 of ref. 153]. - influenza antigen(s) [e.g. chapters 17 & 18 of ref. 153], such as the haemagglutinin and/or neuraminidase surface proteins. - an antigen from Moraxella catarrhalis [e.g. 168]. - a protein antigen from Streptococcus agalactiae (group B streptococcus) [e.g. 169, 170]. - an antigen from Streptococcus pyogenes (group A streptococcus) [e.g. 170,171,172]. Where a saccharide or carbohydrate antigen is used, it is preferably conjugated to a carrier in order to enhance immunogenicity. Conjugation of H.influenzae B, meningococcal and pneumococcal saccharide antigens is well known. Toxic protein antigens may be detoxified where necessary (e.g. detoxification of pertussis toxin by chemical and/or genetic means [152]). Where a diphtheria antigen is included in the composition it is preferred also to include tetanus antigen and pertussis antigens. Similarly, where a tetanus antigen is included it is preferred also to include diphtheria and pertussis antigens. Similarly, where a pertussis antigen is included it is preferred also to include diphtheria and tetanus antigens. DTP combinations are thus preferred. Saccharide antigens are preferably in the form of conjugates. Preferred carrier proteins for conjugates are bacterial toxins or toxoids, such as diphtheria toxoid or tetanus toxoid The CRM197 mutant of diphtheria toxin. [173-175] is a particularly preferred carrier for, as is a diphtheria toxoid. Other suitable carrier proteins include the N.meningitidis outer membrane protein [176], synthetic peptides [177,178], heat shock proteins [179,180], pertussis proteins [181,182], cytokines [183], lymphokines [183], hormones [183], growth factors [183], artificial proteins comprising multiple human CD4+ T cell epitopes from various pathogen-derived antigens [184] such as N19, protein D from H. influenzae -21- WO 2006/046143 PCT/IB2005/003494 [185,186], pneumococcal surface protein PspA [187], pneumolysin [188], iron-uptake proteins [189], toxin A or B from C.difficile [190], etc. Preferred compositions include meningococcal Vesicles as described above, plus a conjugated capsular saccharide from one or more (i.e. 1,2, 3 or 4) of meningococcal serogroups A, C, W135 and Y. Where the Vesicles are from serogroup B then this approach allows the following serogroups to be covered: B+A; B+C; B+W135; B+Y; B+C+W135; B+C+Y; B+W135+Y; B+A+C+W135; B+A+C+Y; B+A+W135+Y; B+C+W135+Y; and B+A+C+W135+Y. Two preferred combinations use serogroup B Vesicles plus conjugate antigens from either serogroups A+W135+Y or serogroups A+C+W135+Y. In general, it is possible to cover all five of serogroups A, B, C, W135 and Y by choosing Vesicles for x serogroup(s) and conjugated saccharides for the remaining 5-x serogroups. Specific meningococcal protein antigens (preferably from serogroup B) may also be added to supplement the vesicle compositions. In particular, a protein antigen such as disclosed in refs. 41 & 191 io 199 may be added. A small number of defined antigens may be added (a mixture of 10 or fewer (e.g. 9, 8, 7, 6, 5, 4, 3, 2) purified antigens). Preferred additional immunogenic polypeptides for use with the invention are those disclosed in reference 199: (1) a TSiadA' protein; (2) a '741' protein; (3) a '93.6' protein; (4) a '9537 protein; and (5) a '287' protein. Other possible supplementing meningococcal antigens include transferrin binding proteins (e.g. TbpA and TbpB) and/or Cu,Zn- superoxide dismutase [18]. Other possible supplementing meningococcal antigens include ORP40 (also known as 'Rsf or 'NhhA' [200,201]).. LctP [202] and ExbB [2021. Other possible . supplementing meningococcal antigens include proteins comprising one of the following amino acid sequences: SEQ ID NO:650 from ref. 191; SEQ ID NO:878 from ref. 191; SEQ ID NO:884 from ref. 191: SEQ ID NO:4 from ref. 192; SEQ ID NO:598 from ref. 193; SEQ ID NO:818 from ref. 193; SEQ ID NO:864 from ref. 193; SEQ ID NO:866 from ref. 193; SEQ ED NO: 1196 from ref. 193; SEQ ID NO:1272 from ref. 193; SEQ ID NO:1274 from ref. 193; SEQ ID NO:1640 from ref. 193; SEQ ID NO:1788 from ref. 193; SEQ ID NO:2288 from ref. 193; SEQ ID NO:2466 from ref 193; SEQ ID NO:2554 from ref. 193: SEQ ID NO:2576 from ref. 193; SEQ ID NO:2606 from ref. 193; SEQ ID NO:2608 from ref. 193; SEQ ID NO:2616 from ref. 193; SEQ ID NO:2668 from ref. 193; SEQ ID NO:2780 from ref. 193; SEQ ID NO:2932 from ref. 193; SEQ ID NO:2958 from ref. 193; SEQ ID NO.-2970 from ref. 193- SEQ ID NO:2988 from ref. 193, or a polypeptide comprising an amino acid sequence which: (a) has 50% or more identity (e.g. 60%, 70%, 80%, 90%, 95%, 99% or more) to said sequences; and/or (b) comprises a fragment of at least n consecutive amino acids from said sequences, wherein n is 7 or more (eg. 8,10, 12,14,16, 18, 20, 25, 30, 35, 40, 50, 60, 70, 80, 90, 100, 150, 200, 250 or more). Preferred fragments for (b) comprise an epitope from the relevant sequence. More than one (e.g. 2, 3, 4, 5, 6) of these polypeptides may be included. The meningococcal antigens transferrin binding protein and/or Hsf protein may also be added [203]. -22- WO 2006/046143 PCT/IB2005/003494 Supplementation of the OMVs by defined meningococcal antigens in this way is particularly useful where the OMVs are from a serosubtype P1.7b,4- meningococcus or a serosubfype Pl.7,16 meningococcus. Supplementation of a mixture of OMVs from both these serosubtypes is preferred. It is also possible to add vesicles that are not vesicles of the invention e.g. OMVs, MVs, NOMVs, etc. that are prepared by methods other than those of the invention (e.g. prepared by methods involving disruption of bacterial membranes, as disclosed in the prior art). Antigens in the composition will typically be present at a concentration of at least lug/mi each. In general, the concentration of any given antigen will be sufficient to elicit an immune response against that antigen. As art alternative to using protein antigens in the composition of the invention, nucleic acid encoding the antigen may be used. Protein components of the compositions of the invention may thus be replaced by nucieic acid (preferably DN A e.g. in the form of a plasmid) that encodes the protein. New meningococcal proteins The genome sequence of serogroup B meningococcus is reported in reference 32. The initial annotation of the genome has not been accepted for all of the >2000 genes e.g. the start codon on NMBI870 has subsequently been re-assigned [41,55]. The inventors have found that the start codons for NMB0928, NMB0109 and NMB1057 should also be re-assigned: ? The original sequence of NMB0928 is shown in Figure 6 (SEQ ID NO: 3). The inventors believe that the true start codon for NMB0928 is the ATG encoding the methionine at residue 24 of Figure 6. With the new start codon (SEQ ID NO: 6), NMB0928 presents a typical signature of a '• surface-exposed protein, characterised by a signal peptide with a lipo-box motif (underlined). ? The .original sequence of NMBOl 09 is shown in Figure 7 (SEQ ID NO: 4). The inventors believe that the true start codon for NMB0109 is the ATG encoding the Met at residue 39 of Figure 7. (SEQ ID NO: 7) ? The original sequence of NMB1057 is shown in Figure 8 (SEQ ID NO: 5). The inventors believe that the true start codon for NMB1057 is the GTG encoding the Val at residue 14 of Figure 8. (SEQ ID NO: 8) Thus the invention provides a polypeptide comprising: (a) the amino acid sequence of SEQ ID NO:6; (b) an amino acid sequence having at least 50% (e.g. 60%, 70%, 80%, 85%, 90%, 95%, 96%, 97%, 98%, 99%, 99.5% or more) sequence identity to SEQ ID NO:6, and/or comprising an amino acid sequence consisting of a fragment of at least 7 {e.g. 8, 9, 10,11,12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 35, 40, 45, 50, 60, 70, 80, 90, 100, 120, 140, 160, 180, 200, 225, 250) contiguous amino acids from SEQ ID NO:6. Preferred polypeptides have aN-terminus cysteine residue, preferably corresponding to Cys-19 of SEQ ID NO:6, and the N-terminus cysteine is - 23- WO 2006/046143 PCT/IB2005/003494 preferably lipidated. Preferred polypeptides do not include the arnino acid sequence MTHIKPVIAALALIGLAA (SEQ ID NO: 9) within 30 amino acids of their N-terminus. The invention also provides a polypeptide comprising: (a) the amino acid sequence of SEQ ID NO:7; (b) an amino acid sequence having at least 50% (e.g. 60%, 70%, 80%, 85%, 90%, 95%, 96%, 97%, 98%, 99%, 99.5% or more) sequence identity to SEQ ID NO:7, and/or comprising an amino acid sequence consisting of a fragment of at least 7 (e.g. 8, 9, 10,11, 12, 13, 14, 15, 16,17,18,19,20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 35, 40, 45, 50, 60, 70, 80, 90, 100, 120, 140,160, 180, 200, 225, 250) contiguous amino acids from SEQ ED NO:7. Preferred ,polypeptides do not include the amino acid sequence MLKCGTFFITRHIPRGCRRFFQPNQARQTEIYQIRGTV (SEQ ID NO: 10) within 20 amino acids of their N-teiminus. The invention also provides a polypeptide comprising: (a) the amino acid sequence of SEQ ED NO: 8; (b) .an amino acid sequence having at least 50% (e.g. 60%, 70%, 80%, 85%, 90%, 95?/o, 96%, 97%, 98%, 99%, 99.5% or more) sequence identity to SEQ ID NO:8, and/or comprising an amino acid sequence consisting of a fragment of at least 7 (e.g. 8, 9,10,11,12,13,14,15, 16,17,18,19, 20,21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 35, 40, 45,^50, 60, 70, 80, 90, 100, 120, 140, 160, 180, 200, 225, 250) contiguous amino acids from SEQ ID NO: 8. Preferred polypeptides have a N-terminus cysteine residue, preferably corresponding to Cys-Gln of SEQ ID NO:8, and the N-terminus cysteine is preferably lipidated. Other preferred polypeptides do not include the anuno acid sequence ' MPCMNHQSNS (SEQ ID NO: 11) within 20 arnino acids of their N-terminus. Polypeptides can be prepared by various means e.g. by chemical synthesis (at least in part), by digesting longer polypeptides using proteases, by translation from RNA, by purification from cell culture (e.g. from recombinant expression or from N.meningitidis culture), etc. Heterologous expression in an E.coli host is a preferred expression route. Polypeptides of the invention may be attached or immobilised to a solid support. Polypeptides of the invention may comprise a detectable label e.g. a radioactive label, a fluorescent label, or a biotin label. This is particularly useful in immunoassay techniques. Polypeptides can take various forms (e.g. native, fusions, glycosylated, non-glycosylated, lipidated, disulfide bridges, etc.). Polypeptides are preferably meningococcal polypeptides. Polypeptides are preferably prepared in substantially pure or substantially isolated form (i.e. substantially free from other Neisserial or host cell polypeptides) or substantially isolated form. In general, the polypeptides are provided in a non-naturally occurring environment e.g. they are separated from their naturally-occurring environment In certain embodiments, the subject polypeptide is present in a composition that is enriched for the polypeptide as compared to a control. As such, purified polypeptide is provided, whereby purified is meant that the polypeptide is present in a composition that is substantially free of other expressed polypeptides, where by substantially -24- WO 2006/046143 PCT/IB2005/003494 free is meant that less than 50%, usually less than 30% and more usually less than 10% of the composition is made up of other expressed polypeptides. The term "polypeptide" refers to amino acid polymers of any length. The polymer may be linear or branched, it may comprise modified amino acids, and it may be interrupted by non-amino acids. The terms also encompass an amino acid polymer that has been modified naturally or by intervention; for example, disulfide bond formation, glycosylation, lipidation, acetylation, phosphorylation, or any other manipulation or modification, such as conjugation with a labeling component. Also included within the definition are, for example, polypeptides containing one or more analogs of an amino acid (including, for example, unnatural amino acids, etc.), as well as other modifications known in the art. Polypeptides can occur as single chains or associated chains. General The term "comprising" encompasses "including" as well as "consisting" e.g. a composition "comprising" X may consist exclusively of X or may include something additional e.g. X + Y. The term "about" in relation to a numerical value x means, for example, x±10%. The word "substantially" does not exclude "completely" e.g. a composition which is "substantially free" from Y may be completely free from Y. Where necessary, the word "substantially" may be omitted from the definition of the invention. References to a percentage sequence identity between two amino acid sequences means that, when aligned, that percentage of amino acids are the same in comparing the two sequences. This alignment and the percent homology or'sequence identity can be determined using software programs known in ' the art, for example those described in section 7.7.18 of reference 204. A preferred alignment is determined by the Smith-Waterman homology search algorithm using an affine gap search with a gap open penalty of 12 and a gap extension penalty of 2, BLOSUM matrix of 62. The Smith- Watennan homology search algorithm is well known and is disclosed in reference 205. References to 'NMB' proteins with a four digit number refers to the standard nomenclature of reference 32, assigned on the basis of a genome sequence of a prototypic strain of serogroup B meningococcus. The public sequence databases include these NMB sequences. For any given meningococcus, the skilled person can readily and unambiguously find the gene corresponding to a NMBnnnn sequence by using the existing sequence from the database and/or the genetic environment of the NMBnnnn ORF in the prototype strain e.g. to design primers, probes, etc. The terms 'GNA33', 'NMB0033' and cmltA' can be used interchangeably when referring to meningococcus. -25- WO 2006/046143 PCT/IB2005/003494 BRIEF DESCRIPTION OF DRAWINGS Figure 1 shows the amino acid sequence (SEQ ID NO: 1) and micleotide sequence (SEQ ID NO: 2) of the membrane-bound lytic murein transglycosylase A (mltA) from the genome sequence of strain MC58 of serogroup B Neisseria meningitidis, taken from GenBank accession AAF40504.1 [32]. Figure 2 shows 2D-PAGE of vesicles of the invention. Figure 3 shows the gel filtration outputs with standard proteins (bottom) and with the centrifugation pellet (top) of the culture supernatant of the ?mltA strain. The y-axis shows absorbance at 280nm. Figure 4 shows electron microscopy of vesicles of the invention. Figure 5 shows western blot analysis of vesicles of the invention. Six different antibodies (A-F) were used to stain the blots: A= mouse serum raised against OMVs prepared from the NZ strain by deoxycholate extraction; B= mouse serum raised against ? GNA33 knockout mutants; C— mouse anti-PorAp1.4 monoclonal; D~ mouse anti-NMB2132 serum; E= mouse anti-NMB1030 serum; F= mouse anti-NMB1870 serum. Figures 6 to 8 show amino acid sequences of NMB0928, NMB0109 andNMB1057. Figures 9 to 11 show amino acid sequences of NMB0928, NMB0109 and NMB1057 with shifted start codons. Figure 12 compares the proteins released into culture supematants by wild-type or ? GNA33 bacteria. Lane 1: Molecular weight markers; lane 2: culture medium control; lane 3: 20 ug proteins collected by high speed centrifugation of ? GNA33 culture medium at OD620nn =0.5,corresponding to 6.5 ml of culture medium; lane 4: proteins collected by high speed centrifugation from 6.5 ml of wild-type MC58 culture medium at OD620mn==0.5. Figure 13 shows SDS-PAGE of a wild-type MC58 total extract (lanes 2 and 4) and of vesicles released by ? GNA33 knockout mutant (lanes 3 and 5). Lanes 2 and 3 are proteins not denatured at 95°C prior to SDS-PAGE; lanes 4 and 5 were denatured at 95°C. Figures 14 and 15 show ID and 2D SDS-PAGE of vesicles prepared from strain 394/98. In Figure 15, the horizontal axis runs from pl 3 to 10 and the vertical axis runs from 10 to 200 kDa. Figures 16 & 17 show ID SDS-PAGE of vesicles prepared from tolR ExPEC knockout strains. Figures 18 to 20 show ID and 2D SDS-PAGE of vesicles from ? mltA knockout meningococci. MODES FOR CARRYING OUT THE INVENTION Preparation of meningococcal ? mltA knockout sti'ain A meningococcal strain was prepared in which the mltA gene is replaced by allelic exchange with an antibiotic cassette. N.meningitidis strain MC58 was transformed with plasmid pBSUDGNA33ERM- This plasmid contains upstream and downstream flanking regions for allelic exchange, a truncated mltA gene, and the ermC gene (encoding erythromycin resistance). The upstream flanking region (including the start -26- WO 2006/046143 PCT/IB2005/003494 codon) from position -867 to +75 and the downstream flanking region (including the stop codon) from position +1268 to +1744 were amplified from MC58 by using the primers U33FOR, U33REV, D33FOR and D33REV [25]. Fragments were cloned into pBluescript™ and transformed'into E.coli DH5 by using standard techniques. Once all subcloning was complete, naturally competent Neisseria strain MC58 was transformed by selecting a few colonies grown overnight on GC agar plates and mixing them with 20ul 10 mM Tris-HCl (pH 6.5) containing lp.g plasmid DNA. The mixture was spotted onto a chocolate agar plate, incubated for 6 h at 37°C with 5% CO2, and then diluted in phosphate buffered-saline (PBS) and spread on GC agar plates containing 7ug/ml erythromycin. Allelic exchange with the chromosomal mltA gene was verified by PCR, and lack of MltA expression was confirmed by Western blot analysis. As reported in reference 25, the ?mltA knockout strain does not have the correct topoiogical organisation of the cellular membrane, has abnormal cell separation, abnormal cell morphology, undivided "septa, "double sepfa,~c'ell clustering, sharing of outer membranes arid-reduced virulence: Reference 25 also reports that the knockout strain releases various membrane proteins into the culture supernatant, including the PqrA, PIB, class 4 and class 5 outer membrane proteins. A ?mltA knockout was also made from New Zealand strain 394/98 (Kn3; B:4:P1.4), which is the strain from which the MeNZB™ product is produced. Analysis of released proteins The ?mltA strain was grown in GC culture medium in a humidified atmosphere containing 5% CO2 until OD600nm 0.5. Bacteria were collected by 10 minutes of centrifugation at 3500 x g. The supernatant {i.e. culture medium) was filtered through a 0.22 pm pore size filter (Millipore), and the cell-free filtrate was subjected to high-speed centrifiigation (200,000 x g, 90 nan). This centrifugation resulted in formation of a pellet, with about 8-12 mg protein per litre of culture medium. No such pellet was seen if wild-type MC58 bacteria were treated in the same way, and so the pellet formation is a result of the ?mltA knockout. The pellet was washed twice with PBS (centrifiigation 200,000 x g, 30 min) for further analysis. In a first analysis, material from the pellet was re-suspended in PBS and applied to a Superdex 200 PC3.2/30 gel filtration column, run on a SMART system (Amersham Biosciences) that had been equilibrated in PBS. The flow rate was 40 pl/min, and eluate was monitored at 280 nm. The column was calibrated with 20 ug Bleu dextran (2,000 kDa), 10 u.g ferritine (440 kDa), 140 ug bovine serum albumin. (65 kDa)- and 200 fig ribonuclease A (15 kDa). As shown in Figure 3,.most offhe-proteiDS eluted in a major peak corresponding to a molecular weight substantially higher than 2,000 kDa. This result suggests that the various proteins are associated. In a second analysis, the material present in the high molecular weight peak was subjected to negative staining electron microscopy. This analysis revealed the presence of well-organised membrane vesicles with a diameter of about 50-100 nm (Figure 4). -27- ' WO 2006/046143 PCT/IB2005/003494 These experiments suggest that deletion of the mltA gene perturbs the normal assembly of the bacterial membrane, and that this results in the spontaneous release into the culture supernatant of membrane structures which assemble in spherical, homogeneous vesicles. Figure 12 shows SDS-PAGE analysis of culture media after growth of wild-type or ? GNA33 bacteria, and shows the different protein release characteristics. Analysis of vesicles The ?mltaA-derived vesicles were compared to meningococcal vesicles prepared by the 'normal' detergent extraction method. Meningococcal strains MC58, NZ394/98 and NZ98/254, and their respective isogenic ?mltaA mutants, were grown in 20 ml or 200 ml GC culture medium in humidified atmosphere containing 5% CO2 until OD620nm 0,5. Bacteria were collected by 10-rninute centrifugation at 3500g. Vesicles (DOMVs) were prepared from the wild-type bacteria by detergent extraction as described in reference 206. Vesicles of the invention ('mOMVs') were prepared from knockout strains by filtration through a 0J22um pore size filter, followed by high-speed centrifiigation (200,000g, 90 min) of the filtrates, washing of the vesicle-containing pellets (centrifugation 200,000g, 30 min) twice with phosphate buffer saline, (PBS), and then re-suspension with PBS. Both the mOMVs and the DOMVs were analysed by denaturing mono-dimensional electrophoresis. Briefly, 20ug of vesicle proteins were resolved by SDS-PAGE and visualised by Coomassie Blue staining of 12.5% gels. Denaturing (2% SDS) and semi-denaturing (0.2% SDS, no ditbiothreitol, no heating) conditions were used mono-dimensional electrophoresis. The amount of protein (20ug) was determined by DC protein assay (Bio-Rad), using bovine serum albumin as a standard protein. The vesicles were denatured for 3 minutes at 95°C in SDS-PAGE sample buffer containing 2% SDS. 20|ig of protein were then loaded onto 12,5% acrylamide gels, which were stained with Coomassie Blue R-250. 2-dimensional electrophoresis was also performed on 200ug of protein brought to a final volume of 125U1 with re-swelling buffer containing 7M urea, 2M thiourea, 2% (w/v) (3-((3-cholanudopropyl)dimethylammonio)-l-propane-sulfonate), 65 mM ditbiothreitol, 2% (w/v) amidosulfobetain-14, 2 mM tributylphosphine, 20mM Tris, and 2% (v/v) carrier ampholyte. Proteins were adsorbed overnight onto Immobiline DryStrips (7 cm; pH-gradient 3-10 non linear). Proteins were then 2D-separated. The first dimension was run using a IPGphor Isoelectric Focusing Unit, applying sequentially 150 V for 35 min., 500 V for 35 min., 1,000 V for 30 min, 2,600 V for 10 min., 3,500 V for 15 mm., 4,200 V for 15 min", and finally 5,000 V to reach 12kVh. For the second dimension, the strips were equilibrated and proteins were separated on linear 9-16.5% polyacrylamide gels (1.5-mm thick, 4x7 cm ). Gels were again stained with Coomassie Brilliant Blue G-250.266 protein spots could be seen after Colloidal Coomassie Blue staining (Figure 2). -28-' WO 2006/046143 PCT/IB2005/003494 The ID and 2D gels were then subjected to in-gel protein digestion and sample preparation for mass spectrometry analysis. Protein spots were excised from the gels, washed with 100 mM ammonium bicarbonate/acetonitrile 50/50 (V/V), and dried using a SpeedVac centrifuge. Dried spots were digested 2 hours at 3TC in. 12 pi of 0.012 ug/ul sequencing grade trypsin (Promega) in 50 mM ammonium bicarbonate, 5 mM. After digestion. 5 ul of 0.1 % trifiuoacetic acid was added, and the peptides were desalted and concentrated with ZTP-TIPs (CIS, Millipore). Sample were eluted with 2 ul of 5 g/1 2,5-dihydroxybenzoic acid in 50% acetonitrile/0.1% trifluoroacetic acid onto the mass spectrometer Anchorchip 384 (400 urn, Bruker, Bremen, Germany) and allowed to air dry at room temperature. MALDI-TOF spectra were acquired on a Bruker Biflex HI MALDI-TOF equipped with a 337 nmN2 laser and a SCOUT 384 muMprobe ion source set in a positive-ion reflector mode. The acceleration and reflector voltages were set at 19 kV and 20 kV, respectively. Typically, each spectrum was determined by averaging 100 laser shots. Spectra were externally calibrated using a combination of four standard peptides, angiotensin II (1,046.54 Da), substance P (1,347.74 Da), Bombensm (1,619.82 Da) and ACTH18-39 Clip human (2,465.20 Da), spotted onto adjacent position to the samples. Protein identification was carried out by both automatic and manual comparison of experimentally-generated monoisotopic values of peptides in the mass range of 700-3000 Da with computer-generated fingerprint's using the Mascot software. Results from the MC58 ?mltA 'mutant are shown in Figure 18. From the 20 excised bands on just the ID gel, 25 unique proteins were identified; 24 (96%) of which were predicted to be outer-membrane proteins by the PSORT algorithm (Table 1 below). 170 protein spots on the 2D gel, corresponding to 51 unique proteins, were unambiguously identified by MALDI-TOF (Table 1). 44/51 identified proteins have been assigned to the outer membrane compartment by the genome annotation [32]. The 7 remaining proteins were analysed for possible errors in the original annotation. Four proteins (the hypothetical proteins NMB1870, NMB0928 and NMB0109, and the glutamylrranspeptidase NMB1057) could be classified as outer membrane proteins using different start codons from those in ref. 32 e.g. for NMB1870, using the start codon assigned in reference 55. The combined ID and 2D eiectrophoresis experiments identified a total of 65 proteins in the MC58 ?mltA mutant-derived vesicles. Of these, 6 proteins were identified in both ID and 2D gels, whereas 14 and 45 were specific for the ID and 2D gels, respectively (Table 1). Moreover, 61 out of the 65 identified proteins were predicted as membrane-associated proteins by current algorithms, indicating that the ?mltA vesicles (mOMVs) are mostly, and possible exclusively, constituted by membrane proteins.. . The ?mltA knockout of strain NZ394/98 was similarly subjected to ID and 2D SDS-PAGE (Figures 14 & 15). Table 2 shows 66 proteins that were identified in one or both of the gels, together with the predicted location of the proteins. Again, most of the proteins were predicted as membrane- associated. The 47 proteins common to Tables 1 and 2 are shown in Table 3. -29- WO 2006/046143 PCT/IB2005/003494 Results from the NZ98/254 ?mltA mutant are shown in Figure 19. 66 proteins were identified from these two gels, 57 of which were assigned to the outer membrane compartment. Again, therefore, the mOMVs are highly enriched in outer membrane proteins. 46 of the 57 proteins had also been identified in the MC'58-derived mOMVs. For comparison, Figure 20 shows the results from NZ98/254 DOMVs. Proteomic analysis revealed 138 proteins, only 44 of which were assigned to the outer membrane compartment. The remaining 94 proteins belonged to the cytoplasmic and inner membrane compartments. Of these 44 membrane proteins, 32 were also found in the 57 outer membrane proteins found in the mOMVs from the isogenic strain. While mOMVs are largely constituted by outer membrane proteins, therefore, about 70% of DOMV proteins are either cytoplasmic or inner membrane proteins. DOMVs differ from mOMVs not only . for the_ proportion of cytoplasmic proteins but also for the different profile of their outer membrane proteins. Of the 44 outer membrane proteins seen in DOMVs, only 32 were also seen in mOMVs. 19 proteins seen in mOMVs from both MC58 and NZ98/254, but not in the DOMVs from NZ98/254, are listed in Table 4 below. A total cell extract of bacteria was prepared as follows: Bacterial cells were washed with PBS, and the bacterial pellet was resuspended in 8 ml of 50 mM Tris-Hd pH 7.3 containing protease inhibitor cocktail (Roche Diagnostic).-2 mM EDTA and 2000 units of benzonase (Merck) were added, ceils were disrupted at 4°C with Basic Z 0.75V Model Cell Disrupter equipped with an "one shot head" (Constant System Ltd) by 2 cycles, and the unbroken cells were removed by centrirugationl0 min at 8 000 x g at 4°C. This extract was analysed by SDS-PAGE, for comparison with a protein extract of the vesicles produced by ? GNA33 bacteria. As shown in Figure 13, the porins PorA and PorB (identities verified by MALDI-TOF sequencing) are seen in the wild-type bacterial outer membrane (lanes 2 & 4) and also in the ? GNA33 knockout mutant's vesicles (lanes 3 & 5). Moreover, these proteins are retained as stable trixners in the vesicles that do not dissociate into monomers in SDS-PAGE sample buffer with a low concentration of SDS (0.2%) under seminative conditions (no heating before electrophoresis; lanes 2 & 3), but that do denature at 95°C (lanes 4 & 5). LPS levels in detergent-extracted OMVs are typically 5-8% by weight, relative to protein [207]. When tested with the Limulus assay, the endotoxin content of the vesicles was about twice as high as found in-detergent-extracted OMVs. Finally, the yield of vesicles in a growing culture was assessed. It was found that up to 20 mg of OMV-associated proteins could be recovered per gram of cells (wet weight) in culture supernatants of early exponentially growing cultures (OD620nm=0.5). -30- WO 2006/046143 PCT/IB2005/003494 Vesicle immunogenicity As the ?7n&4-derived vesicles are highly enriched in outer membrane proteins, their ability to elicit bactericidal antibodies capable of killing a broad panel of MenB clinical isolates was investigated. The strain chosen for the testing was 394/98. This strain was chosen because it is the strain from which the MeNZB™ OMV-based vaccine is prepared, thereby aiding a direct comparison of ?mltA- vesicles of the invention with OMVs prepared from the wild-type strain by typical prior art methods. 10 μg of each type of vesicle was adsorbed to an aluminium hydroxide adjuvant (3mg/ml) and injected into mice 5-week old CD1 female mice (5-10 mice per group). The vesicles were given intraperitoneally on days 0 and 2 L Blood samples for-analysis were taken on day 34, and were tested for SBA against \ 5 different serogroup B strains corresponding to 11 different sub-types, including ±e four major hypervirulent lineages, using pooled baby rabbit serum as the complement source. Serum bactericidal titers were defined as the serum dilution resulting in 50% decrease in colony forming uniis (CFU) per ml after 60 minutes incubation of bacteria with reaction mixture, compared to control CFU per ml at time 0. Typically, bacteria incubated with the negative control antibody in the presence of complement showed a 150 to 200% increase in CFU/ml during the 60 min incubation. Tilers were as follows, expressed as the reciprocal of the serum dilution yielding =-50% bacterial killing: /The results show that serum from ?mltA-derived vesicles were at least as bactericidally effective, and usually better than, OMVs prepared by chemical extraction, except for the homologous strain. The vesicles of the invention thus give much better cross-strain reactivity than typical OMVs. Moreover, talcing a 1:1024 dilution as the threshold for bactericidal efficacy, the vesicles of the invention were effective against 87% of the strains, whereas the artificial OMVs were only 40% effective. -31- WO 2006/046143 PCT/IB2005/003494 Thus mOMVs are better than DOMVs for eliciting complement-dependent antibody killing when tested over a panel of 15 different serogroup B strains. The anti-mOMV mouse sera showed high bactericidal activities against the homologous strain and against 14 additional strains, including i 0 different PorA subtypes. In contrast, mouse sera raised against DOMVs show high bactericidal titers only against six MenB strains, belonging to two PorA subtypes. These results indicate that the protection of anti-mOMV sera was not only due to the elicitation of bactericidal antibodies against PorA, which is one of the most abundant outer membrane proteins and the most potent inducer of bactericidal antibodies, but also to other bactericidal antigens which in mOMVs are present in higher amounts than in DOMVs. Western blot To confirm that the ?mltA-derived vesicles do contain conserved, protective antigens, ihey were run on an SDS-PAGE, iransferred onto a PDF ulier and immunobiotted using specific anti-sera against six proteins antigens previously shown to be protective and highly conserved, including l287J, '953', '741' (GNA1870) and 'NadA'. The vesicles were separated onto 10% acrylamide .SDS-PAGE gels employing a Mini-Protean H eiectrophoresis apparatus (Bio-Rad). After protein separation, gels were equilibrated with 48 mM Tris-PICi, 39 mM glyciae, pK 9.0,20% (v/V) nierhanoi and transferred to a nitrocellulose membrane (Bio-Rad) using a Trans-Blot™ semi-dry electrophoretic transfer cell. The nitrocellulose membranes were blocked with 10% (w/v) dammed milk in PBS containing 0.2% (w/v) sodium azide. As shown in figure 5, all six proteins were abundant in the vesicles. In contrast, the same six proteins' were poorly represented in the DOMVs. In conclusion, the ?mltA-derived vesicles are predominantly constituted by outer membrane proteins, whereas DOMVs are heavily contaminated by cyloplasmic proteins. When used to immunize mice, sera raised against ?mltA-derived vesicles .showed a higher and wider strain coverage than DOMVs. Extraintestinal pathogenic E. coll A knockout strain of ExPEC CFT073 was prepared by isogenic deletion of the toIR gene, replacing it with a kanamycin resistance marker. The knockout strain was grown to OD600nm 0.4, and the culture was then ccntrifuged. The supernatant was filtered through a 0.22μm filter and the filtrate was prepciptated using TCA. The pellet was then resuspended in Tris buffer. The same growth and purification procedure was used for the parent's (rain, without the knockout, and SDS-PAGE analysis of the two final preparations is shown in Figure 16. The right-hand band is from the knockout srrain.and shows enrichment of several protein bands. Further toIR knockout ExPEC strains were prepared from strains DH5a, 536 and 1HE3034. Vesicles were prepared as before, and SDS-PAGE analysis of TCA precipitates is shown in Figure 17. - 32 - WO 2006/046143 PCT/IB2005/003494 The knockout mutant produces high amounts of vesicles, and these vesicles were subjected to proteomic analyses, including ID and 2D SDS-PAGE and tryptic digestion of surface-exposed proteins in the vesicles followed by sequence analysis of released peptides. It will be understood that the invention has been described by way of example only and modifications may be made whilst remaining within the scope and spirit of the invention. -33- WO 2006/046143 PCT/IB2005/003494 WO 2006/046143 PCT/IB2005/003494 -35- -36- WO 2006/046143 PCT/IB2005/003494 WO 2006/046143 PCT/IB2005/003494 REFERENCES (the contents of which are hereby incorporated by reference) [1] Bjune etal (1991) Lancet 338(8775):1093-1096. [2] de Kleijn et al. (200-1) Vaccine 20:352-358. [3] US patents 5,597,572 & 5,747,653; see also European patent 0301992. [4] European patent 0449958 (granted from WO90/06696). [5] US patent 5,705,161; see also WO94/08021. [6] WO01/9178S. [7] Parmar et al. (1997) Vaccine 15:1641-1651. [8] WO99/59625. [9] WO 00/50074. [10] US patents 5,552,146,5,981,213 & 5,993,826; see also WO93/03761. [11] Zhou et al (1998) FEMS MicrobiolLett 163:223-228. [12] Kaduntgarnuwa & Beveridge (1999)Microbiology 745:2051-2060. [13] WO97/05899. [14] Kesavalu et al. (l992) Infect Immun. 60:1455- 1464. [15] Blanco et al. (1999) J.Immunol 153:2741-2746. [16] WO01/09350. [17] Keenan et al. (1998) FEMS Microbiol Lett 161:21-27. [18] WO00/258H. [19] WO01/52885. [20] WO98/56901. [21]WO02/09746. [22] WO02/062378. [23] WO2004/014417. [24] WO2004/019977. [25] Adu-Bobie et al. (2004) Infect Immun 72:1914-1919. [261 Jennings et al. (2002) Ear JBiochem 269:3722-3731. [27] Pollard & Moxon (2002) Arch Dis Child 87:13-17. [28]WO2004/014417. [29] Shockman & Holtje (1994) Microbial peptidoglycan (murein) hydrolases. Pages 131-166 in Bacterial Cell Wall (eds. Ghuysen & Hakenbeck). [30]WO00/6674I. [31] Parkhill et al. (2000) Nature 404:502-506. [32] Tettelin et al (2000) Science 287:1809-1815. [33] Maiden et al (1998) PNAS USA 95:3140-3145. [34] WO99/10497. [35] Steeghs et al. (2001) The EMBO Journal 20:6931-6945. [36] WO02/07763. [37] European patent 0624376. [381 Claassen et al. (1996) Vaccine 14:10014008. [39]Peeters et al (1996) Vaccine 14:1009-1015. [40] van der Ley et al. (1995) Vaccine 13:401 -7. [41] WO2004/048404. [42] WO 02/062378. -37- WO 2006/046143 PCT/IB2005/003494 [43] WO 2004/014417. [44] UK patent application 0419627.5. [45] Russo & Johnson (2000) J Infect Dis 181:1753-4. [46] Bemadac etal. (1998) JBacterial 3 80(18):4872-3. [47] WO02/09643. [48] Beveridge (1999) JBacteriol 181:4725-33. [49] Moe et al. (2002) Infect Immun 70:6021-31. [50] Arigita et al. (2003) Vaccine 21:950-960. [51] WO2004/046177 [52] US patent 6,180,111. [53] WO01/34642. [54] Maiden etal (1998) PNAS USA 95:3140-3145. [55] Masignani et al (2003) J Exp Med 197:789-799. [56] WO03/063766. [57j Fletcher et al. (2004) Infect Immun 72:2088-2100. [58] Zhu et al (2000) J Bacteriol 182:439-47. [59] Moe et al (2001) Infect Immun 69:3762-71. [60] Seib et al (2003) FEBS Lett 546:411-5. [61] WO01/64922 [62] Gcnnaro (2000) Remington: The Science and Practice of Pharmacy, 20th edition, ISBN: 0683306472. [63] RTVM report 124001 004. [64J Katial et al (2002) Infect Immun 70:702-7. [65] RIVM report 000012 003. [66] WO03/009869. [67] Almeida & Alpar (1996) .1 Drug Targeting 3:455-467. [68] Agarwal & Mishra (1999) Indian JExp Bio! 37:6-16. [69] Vaccine Design... (1995) eds. Powell & Newman. ISBN: 030644867X. Plenum. [70] WO00/23105. [71] WO90/14837. [72] US patem 5,057,540. [73] WO96/33739. [74] EP-A-0109942. [75]WO96/11711. [76] WO00/0762L [77] Barr et al (1998) Advanced Drug Delivery Reviews 32:247-271. [78] Sjolanderet et al. (1998) Advanced Drug Delivery Reviews 32:321-338. [79] Niiknra et al. (2002) Virology 293:273-280. [80] Lenz etal. (2001) J Immunol 166:5346-5355. [81] Pinto etal (2003) J Infect Dis 188:327-338. [82]Gerber et al. (2001) Virol 75:4752-4760. [83] WO03/024480 [84] WO03/024481 [85] Gluck et al (2002) Vaccine 20:B10-B16. -38- WO 2006/046143 PCT/IB2005/003494 [86]EP-A-0689454. [87] Johnson et al. (1999) Bioorg Med Chem Lett 9:2273-2278. [88] Evans et al. (2003) Expert Rev Vaccines 2:219-229. [89] Meraldi et al. (2003) Vaccine 21:2485-2491. [90] Pajak et al. (2003) Vaccine 21:836-842. [91] Kandimalla etal. (2003) Nucleic Acids Research 31:2393-2400. [92] WO02/26757. [93]WO99/62923. [94] Krieg (2003) Nature Medicine 9:831-835. [95] McCluskie et al. (2002) FEMS Immunology and Medical Microbiology 32:179-185. [96] WO98/40100. [97] US patent 6,207,646. [98] US patent 6,239,116. [99] US patent 6,429,199. [1001 Kandimalla et al. (2003) Biochemical Society Transactions 31 (part 3):654-658 [101] Blackwell et al. (2003) J Immunol 170:4061-4068. [102] Krieg (2002) Trends Immunol 23:64-65. [103] WO01/95935. [104] Kandimalla etal (2003) BBRC 306:948-953. [105] Bhagat et al. (2003) BBRC 300:853-861. [106]WO03/035836. [107] WO95/17211. [108]WO98/42575. [109]Beignon et al. (2002) Infect Immun 70:3012-3019. [110] Pizza et al. (2001) Vaccine 19:2534-2541. [111 ] Pizza et al (2000) Int J Med Microbiol 290:455-461. [112]Scharton-Kersten et al. (2000) Infect Immun 68:5306-5313. [113] Ryan et al (1999) Infect Immun 67:6270-6280. [114] Partidos et al. (1999) Immunol Lett 67:209-216. [115] Peppoloni et al. (2003) Expert Rev Vaccines 2:285-293. [116] Pine et al. (2002) J Control Release 85:263-270. [117] Domenighini et al. (1995) MolMicrobiol 15:1165-1167. [118]WO99/40936. [119]WO99/44636. [120] Singh et al] (2001) J Cont Release 70:267-276. [121]WO99/27960. [122] US patent 6,090,406 [123] US patent 5,916,588 [124]EP-A-0626169. [125] WO99/52549. [126]WO01/21207. [127]WO01/21152. [128] Andrianov et al (1998) Biomaterials 19:109-115. [129] Payne et al (1998) Adv Drug Delivery Review 31:185-196. -39- WO 2006/046143 PCT/IB2005/003494 [130] Stanley (2002) Clin Exp Dermatol 21:511-577. [131] Jones (2003) Curr Opin Investig Drugs 4:214-218. [132] WO99/11241. [133] WO94/00153. [134]WO98/57659. [135] European patent applications 0835318, 0735898 and 0761231. [136] Bakke et al (2001) Infect. Imimn. 69:5010-5015. [137]WO01/30390. [138] http://neisseri.ci. org/nm/typing/mlst/ [139] Pettersson et al (1994) MicrobPathog 17(6):395-408. [140] Welsch et al. (2002) Thirteenth International Pathogenic Neisseria Conference, Norwegian Institute of Public Health, Oslo, Norway; Sept. 1-6, 2002. Genome-derived antigen'(GNA) 2132 elicits protective serwn antibodies to groups B and C Neisseria meningitidis strains. [141] Santos et al. (2002) Thirteenth International Pathogenic Neisseria Conference, Norwegian Institute of Public Health, Oslo, Norway; Sept. 1-6, 2002. Serum bactericidal responses in rhesus macaques Immunized with" novel vaccines containing recombinant proteins derived from the genome ofN. meningitidis. [142] Costantino et al (1992) Vaccine 10:691-698. [143]WO03/007985. [144] Watson (2000) Pediatr Infect Dis J 19:331-332. [145] Rubin (2000) Pediatr Clin North Am 47:269-285, v. [146] Jedrzejas (2001) MicrobiolMolBiolRev 65:187-207. [147] Bell (2000) Pediatr Infect Dis J 19:1187-11X8. [148] Iwarson (1995) APMS 103:321-326. [149] Gerlich et al. (1990) Vaccine 8 Suppl:S63-68 & 79-80. [150] Hsu et al. (1999) Clin Liver Dis 3:901-915. [151] GusmSson et al (1996) N. Engl J. Med. 334:3.49-355. [152] Rappuoli et al. (1991) TIBTECH 9:212-218. [153] Vaccines (2004) eds. Plotkin & Orenstein. ISBN 0-7216-9688-0. [154] WO02/079243. [155]WO02/02606. [156] Kalman et al (1999) Nature Genetics 21:385-389. [157] Read et al (2000) Nucleic Acids Res 28:1397-406. [158] Shirai etal (2000) J. Infect. Dis. 181(Suppl 3):S524-S527. [159]WO99/27105. [160] WO00/27994. [161] WO00/37494. [162] WO99/28475. [163] Ross etal (2001) Vaccine 19:4135-4142. [164] Sutter et al (2000) Pediatr Clin North Am 47:287-308. [165] Zinitnexman & Spann(l 999)AmFam Plrysician 59:113-112,125-126. [166] Dreesen (1997) Vaccine IS Suppl:S2-6. [167] MMWR Morb Mortal Wkly Rep 1998 Jan 16;47(1):I2, 19. [168] McMichael{2000) Vaccine 19 Suppl l:S101-107. [169] Schuchat.(1999) Lancet 353(9146):51 -6. -40- WO 2006/046143 PCT/IB2005/003494 [170]WO02/34771. [171] Dale (1999) Infect Dis Clin North Am 13:227-43, viii. [172] Ferretti etal (2001) PNAS USA 98: 4658-4663.- [173] Anonymous (Jan 2002) Research Disclosure, 453077. [174] Anderson (1983)InfectImmiin 39(l):233-238. [175] Anderson etal (1985) /Clin Invest 76(l):52-59. [176] EP-A-0372501. [177]EP-A-0378881. [178]EP-A-0427347. [179]WO93/17712 [180]WO94/03208. [181]WO98/58668. [182] EP-A -0471177 [183] WO91/01146 [184] Fulugi et al. (2001) Eur J Immnol 31:3816-3824. [185] EP-A-05946J0. [186]WO00/56360. [187]WO02/091998. [188] Kuo et at. (1995) Infect Immun 63:2706-13. [189]WO01/72337 [190]WO00/61763. [191]WO99/2-1578. [192]WO99/36544. [193] WO99/57280. [194]WO00/22430. [195] WO96/29412. [196] Pizza et al. (2000) Science 287:1816-1820. [197] WO01/64920. [198]WO03/020756. [199] WO2004/032958. [200] Peak et al. (2000) FEMS Immunol Med Microbiol 28(4):329-34, [201]WO99/31132. [202] Sun et al. (2005) Vaccine 23(32):4136-41. [203] WO2004/014419. [204] Current Protocols in Molecular Biology (F.M. Ausubel et al., eds., 1987) Supplement 30. [205] Smith & Waterman (1981) Adv. Appl Math. 2:482-489. PQ6]Fredriksen et al (1991) NIPH Ann. 14:67-79. [207] Guthrie et al (2004) Infect Immum 72:2528-37. -41- WO 2006/046143 PCT/IB2005/003494 CLAIMS 1. A bacteriunvwherein: (i) the bacterium has a cell wall that includes peptidogrycan; and (ii) the bacterium does not express a protein having the lytic transglycosylase activity of MtA protein. 2. A bacterium liaving a knockout mutation of its mltA gene. 3. The bacterium of claim 2, also having a knockout mutation of at least one further gene. 4. The bacterium of any preceding claim, which is in the Neisseria or Escherichia genus. 5. The bacterium of claim 4, which is Kmeningiiidis. 6. The bacterium of claim 5, wherein the N.meningitidis is from serogroup A, B, C, W135 or Y. 7. The bacterium, of claim 5 cr claim 6, wliicu is agna33~ ipxA" PorA~ menineococcus. o. The bacterium of claim 4, which is E.coli 9. The bacterium of claim 8, which is a pathogenic Ecoli. 10. The bacterium of claim-9, wherein the pathogenic E.coli is an. extraintestinai .pathogenic bacterium, a uropathogenic bacterium, or a meningitis/sepsis-associated bacterium. 11. A pathogenic Escherichia coli bacterium, which does not express a protein of the Tol-Pal complex. 12. The E.coli of claim 11, which is a toIR strain. 13. A composition comprising vesicles that, during culture of the bacterium of any preceding claim,' are released into the culture medium. 14. The composition of claim 13, wherein the composition does not comprise any living and/or whole bacteria. 15. A composition comprising vesicles, wherein the vesicles are present in the filtrate obtainable after filtration through a 0.22 μm filter of a culture medium in which-a bacterium of any one of claims 1 to 12 has been grown. 16. A meningococcal vesicle obtainable by cuituring.the bacterium of any one of claims 5 to 7. 17. The meningococcal vesicle of claim 16, which does not include at least one of MinD, FtsA, and/or phosphpenolpyruvate synthase proteins. 18. The meningococcal vesicle of claim 16, which is substantially free from ribosomes. 19. The meningococcal vesicle of claim 16, which is substantially free from any aminoacid tRNA-synthetases. -42- WO 2006/046143 PCT/IB2005/003494 20. The meningococcal vesicle of claim 16, which is substantially free from any enzyme from the Rrebs cycle. 21. The meningococcal vesicle of claim 16, which includes the following 47 proteins: NMB0035, NMB0044, NMB0086, NMB0088, NMB0109, NMB0124, NMB0138, NMB0182, NMB0204, NMB0278, NMB0294, NMB0313, NMB0345, NMB0346, NMB0382, NMB0460, NMB0461, NMB0550 , NMB0554, NMB0623, NMB0634, NMB0663, NMB0703, NMB0787, NMB0873, NMB0928, NMB1030, NMB1053, NMB1057, NMB1126, NMB1285, NMB1301, NMB1332, NMB1429, NMB1483, NMB1533, NMB1567, NMBI612, NMB1710, NMB1870, NMB1898, NMB1949, NMB1961, NMB1972, KMB1988, NMB2039 and NMB2091. 22. A pharmaceutical composition comprising meningococcal vesicles of any one of claims 16 to 21. 23. A composition comprising (i) a first set of meninRococcal vesicles of any one claims 26 to 21 and (ii) a second set of memingocoocal vesicles of any one claims 16 to 21, wherein ssid first and second sets are prepared from different strains of meningococcus. 24. The composition of any one of claims 13,14,15,22 or 23, including an adjuvant, 25. A process for preparing bacterial vesicles, comprising the steps of; (i) culturing the bacterium of any one of claims I to 12 in a culture medium such that the bacterium releases vesicles into said medium; and (ii) collecting the vesicles from said medium. -43- Knockout of the meningococcal mltA homolog gives bacteria that spontaneously release vesicles that are rich in immunogenic outer membrane proteins and that can elicit cross-protective antibody responses with higher bactericidal titres than OMVs prepared by normal production processes. Thus the invention provides a bacterium having a knockout mutation of its mltA gene. The invention also provides a bacterium, wherein the bacterium: (i) has a cell wall that includes peptidoglycan, and (ii) does not express a protein having the lytic transglycosylase activity of MltA protein. The invention also provides compositions comprising vesicles that, during culture of bacteria of the invention, are released into the culture medium.

Documents:

01844-kolnp-2007-abstract.pdf

01844-kolnp-2007-claims.pdf

01844-kolnp-2007-correspondence others 1.1.pdf

01844-kolnp-2007-correspondence others.pdf

01844-kolnp-2007-description complete.pdf

01844-kolnp-2007-drawings.pdf

01844-kolnp-2007-form 1.pdf

01844-kolnp-2007-form 3 1.1.pdf

01844-kolnp-2007-form 3.pdf

01844-kolnp-2007-form 5.pdf

01844-kolnp-2007-gpa.pdf

01844-kolnp-2007-international publication.pdf

01844-kolnp-2007-international search report.pdf

01844-kolnp-2007-pct priority document notification.pdf

01844-kolnp-2007-sequence listing.pdf

1844-KOLNP-2007-(07-11-2013)-CORRESPONDENCE_.pdf

1844-KOLNP-2007-(08-10-2014)-CORRESPONDENCE.pdf

1844-KOLNP-2007-(08-10-2014)-FORM-1.pdf

1844-KOLNP-2007-(08-10-2014)-FORM-13.pdf

1844-KOLNP-2007-(08-10-2014)-FORM-2.pdf

1844-KOLNP-2007-(08-10-2014)-FORM-3.pdf

1844-KOLNP-2007-(08-10-2014)-FORM-5.pdf

1844-KOLNP-2007-(08-10-2014)-OTHERS.pdf

1844-KOLNP-2007-(08-10-2014)-PA.pdf

1844-KOLNP-2007-(13-02-2012)-CORRESPONDENCE.pdf

1844-KOLNP-2007-(13-06-2014)-ANNEXURE TO FORM 3.pdf

1844-KOLNP-2007-(13-06-2014)-CORRESPONDENCE.pdf

1844-KOLNP-2007-(13-06-2014)-OTHERS.pdf

1844-KOLNP-2007-(20-12-2011)-ABSTRACT.pdf

1844-KOLNP-2007-(20-12-2011)-AMANDED CLAIMS.pdf

1844-KOLNP-2007-(20-12-2011)-DESCRIPTION (COMPLETE).pdf

1844-KOLNP-2007-(20-12-2011)-DRAWINGS.pdf

1844-KOLNP-2007-(20-12-2011)-EXAMINATION REPORT REPLY RECEIVED.pdf

1844-KOLNP-2007-(20-12-2011)-FORM-1.pdf

1844-KOLNP-2007-(20-12-2011)-FORM-2.pdf

1844-KOLNP-2007-(20-12-2011)-FORM-3.pdf

1844-KOLNP-2007-(20-12-2011)-OTHER PATENT DOCUMENT-1.pdf

1844-KOLNP-2007-(20-12-2011)-OTHER PATENT DOCUMENT.pdf

1844-KOLNP-2007-(20-12-2011)-OTHERS.pdf

1844-KOLNP-2007-(23-04-2014)-CORRESPONDENCE.pdf

1844-kolnp-2007-form 18.pdf