Modular platform for generating antibodies

Abstract

Provided herein is modular platform for generating antibodies specific for protein and polysaccharide antigens. Also provided herein is a hybrid lipopolysaccharide (LPS) molecule, Gram-negative bacteria comprising the hybrid LPS molecule and methods of producing and using the same, particularly for the prevention or treatment of Gram-negative bacterial infections

Description

MODULAR PLATFORM FOR GENERATING ANTIBODIESField of the invention[1] The present invention relates to a modular platform for generating antibodies specific for protein and polysaccharide antigens.[2] The present invention also relates to a hybrid lipopolysaccharide (LPS) molecule, Gram-negative bacteria comprising the hybrid LPS molecule and methods of producing and using the same, particularly for the prevention or treatment of Gram-negative bacterial infections.[3] Infectious disease poses a serious threat to human health as demonstrated by the Covid-19 pandemic. Whilst viruses, like SARS-CoV2, have the potential to cause swift global outbreaks, increasing levels of antimicrobial resistance (AMR) are now rendering many bacterial infections untreatable by even reserve classes of antibiotics. This has been coined the ‘slow burning pandemic’ and threatens to make everyday bacterial infections untreatable in the future. In 2019 alone it was estimated that 4.95 million deaths were associated with bacterial AMR and 1 .27 million deaths directly attributable to these infections. The global response to this enormous burden has been to invest significant funding at national and international level towards new prevention and treatment strategies, spanning academia and industry. Whilst the search for novel antibiotics forms one pillar of the response, other options under evaluation include preventative strategies, such as vaccines, and novel biological therapies such as monoclonal antibodies (mAbs).[4] Vaccines induce protective responses against pathogens and are used widely to prevent or reduce the severity of infectious diseases in humans. One key component of the response to an effective bacterial vaccine is the induction of specific antibody responses to the pathogen which the vaccine protects against.[5] Antibodies bind to specific targets which are usually polysaccharide or protein in nature. The binding between antibodies, which are produced by B-cells, and target (antigen) is very specific. Each mature B-cell produces one specific antibody. The immune response to a pathogen will select for B-cells which produce antibodies that specifically bind antigens present on the pathogen. Over the course of natural infection and recovery, these B-cell responses will be honed and select for antibodies which bind bacterial targets even more strongly. This occurs by a process called somatic hypermutation and affinity selection. For this process to take place several doses of the same vaccine are often given, the quality of the antibodies improving with each administration.[6] Effective vaccination also requires the induction of a memory response; this means that when the host immune system encounters the pathogen that the vaccine protects against, it remembers the pathogen and does not need to start the antibody selection process from scratch. This memory response is much faster than when the host encounters the pathogen for the first time. Memory B-cells, generated as part of the immune response to the vaccine, play an important role here as they are able to quickly divide and produce specific antibodies that target the pathogen as soon as the infection takes place.[7] Alongside vaccines, treatments relying on monoclonal antibodies (mAbs) have been developed. These represent an emerging field with several mAbs recently reaching the market. In mAb therapy, instead of relying on vaccine-induced antibody generation, the protective antibodies themselves are produced in vitro and administered to patients.[8] mAbs still require a functional immune system to work as once the antibody binds the antigen on a pathogen, other immune cells are required to kill and remove it. One exception here are mAbs that target bacterial toxins, where binding of the antibodies directly inhibits the toxin’s ability to induce damage.[9] The term monoclonal antibody (mAb) refers to an antibody produced from a single B-cell. As a single B-cell produces only one type of antibody, which binds to a specific part of an antigen (called epitope), all the antibodies produced from that B-cell will be identical. Following vaccination or recovery from bacterial infection, the host will produce many different antibodies which bind different epitopes on the same antigen with different affinities. This collection of antibodies is called a polyclonal antibody response.

[0010] For antibodies to be used as drugs, they need to be able to be produced in large quantities and quality controlled. The most straightforward way to do this is by generating monoclonal antibodies which have defined functional binding to antigens (determined during the development stage) and can be produced by growing monoclonal B-cells in large quantities. It should be noted that polyclonal antibodies have been used in attempts to treat human infections too in the form of convalescent sera. Here, patients who have recovered from infections have their blood taken to obtain the antibody-containing component (sera). These sera are then given to other patients as a drug. There are several issues with this technique including the requirement for the initial patients to survive infection, production constraints (i.e. very limited amounts of product and high batch-to-batch variability), concerns regarding the transmission of blood-borne viruses and limited data for actual efficacy. For the above reasons mAbs are now the most desirable format for these immunotherapeutics.

[0011] Several technologies are used to ‘find’ mAbs. These can broadly be divided into target-agnostic and target-specific approaches.

[0012] In target-agnostic approaches the company / researcher looks for all antibodies that bind a specific pathogen. There will be many thousand possible targets on each pathogen. This is usually achieved by looking in the blood from patients who have recovered from a specific infection. The blood of these patients will contain memory B-cells which produce antibodies to specific antigens on the pathogen. If many memory B-cells produce antibodies that bind one antigen it suggests that this is a dominant epitope and thus antibodies generated against it may be protective. The company / researcher will then immortalise these memory B-cells so that they can grow them up in large numbers and produce large quantities of antibodies. Another target-agnostic approach is to use large libraries of B-cells and look for ones which produce antibodies that bind the pathogen of interest. In both these approaches the target of the antibody is identified after it is confirmed to bind the pathogen and hence, this approach is target-agnostic.

[0013] In target-specific approaches the company / researcher goes in the opposite direction. They choose a target antigen and either look for antibodies that bind it (either from patients or B cell libraries) or directly induce a specific antibody response against that antigen. Conventionally, this approach is achieved by first generating an immunogen / vaccine, and immunising either experimental animals (usually mice) or humans (as part of a phase l / ll clinical trial). Memory B-cells are harvested from subjects and those producing antibodies against the specific-antigen are selected for development. These memory B-cells are taken after the complete multicourse vaccination administration so that cells encode antibodies that have undergone affinity maturation, i.e., antibodies that are better at binding their target. The major hurdle / bottleneck in the target-specific approach is that it requires the initial production or purification of the antigen before it is used for immunisation.

[0014] Example antigens. Lipopolysaccharide (LPS) is expressed in almost all Gram-negative bacteria (excluding Mycobacterium tuberculosis all WHO Critical priority pathogens are Gram- negative) and its essential for their survival. Thus, LPS is a major target for vaccines and mAbs. LPS is exported to the bacterial surface using two alternative pathways, the ABC-transporter system used, for example, by Klebsiella pneumoniae (KP) and the ’wzy-dependent’ pathways used, for example, by Escherichia coll (Ec). In order to demonstrate that we can use citrOgen for generating antibodies against the two LPS export types, we substitute the natural Citrobacter rodentium LPS O-antigen locus with the KP and Escherichia coll loci. In addition to LPS, KP capsular polysaccharide (CPS) and protein mAb and vaccine targets have already been validated and therefore we use them for to showcase the citrOgen technology. Of note, LPS O25b is predominant on Escherichia coll ST131 , an ESBL-producing WHO Critical priority organism.

[0015] In view of the above, there remains a need for a platform that generates a robust immune response and efficiently generates specific antibodies.

[0016] Infectious disease poses a serious threat to human health as demonstrated by the Covid-19 pandemic. Whilst viruses, like SARS-CoV2, have the potential to cause swift global outbreaks, mounting antimicrobial resistance (AMR) is now rendering many serious bacterial infections untreatable by even reserve classes of antibiotics. In 2019 alone it was estimated that 4.95 million deaths were associated with bacterial AMR and 1 .27 million deaths directly attributable to these infections. Whilst the search for novel antibiotics forms one pillar of the response, other options under evaluation include preventative strategies, such as vaccines, and novel biological therapies, such as monoclonal antibodies (mAbs).

[0017] Vaccines protect against the severe consequences of infection by exposing hosts to a component (or components) of pathogens in a safe and controlled manner. The administered component(s) of the pathogen are termed immunogens as they stimulate an immune response in the host. Immunogens are most frequently proteins or carbohydrates (e.g., polysaccharides). To be effective these immunogens must be present in the pathogen but absent in the host so that the immune response is directed towards the infection. Vaccines induce a variety of protective responses including the generation of antibodies specific to the immunogens administered. Monoclonal antibodies (mAbs) can also be used in the treatment and prevention of infection. They can be administered prophylactically (like vaccines) or as therapeutics (treating established infection).

[0018] The outer surface of the Gram-negative bacteria has many surface polysaccharides which are often unique to the infecting species. Gram-negative bacteria are responsible for most serious infections in hospitals. These include highly abundant glycoproteins and glycolipids such as capsular polysaccharide (CPS) and lipopolysaccharide (LPS), respectively. These serve as bacterial virulence factors enabling infection but also have proven to be very good vaccine candidates. As they lie on the external surface of bacteria they are also serve as excellent targets for host antibodies, which are unable to penetrate the impermeable outer-membrane of bacteria to reach targets inside the cytoplasm.

[0019] LPS is a glycolipid composed of two major components. A lipid moiety (lipid A) that sits in the lipophilic outer membrane and a complex polysaccharide component that extends away from the cell. The two are covalently linked and the genes that are required to make and assemble both these components are encoded on the bacterial genome. The polysaccharide component has three distinct regions called inner core, outer core and O-antigen. O-antigen is composed of repeating subunits of different sugars and the specific sugars, how they are joined together, and their order is determined by the genes encoding the LPS type. This gives rise to significant complexity in structures across Gram-negative bacteria, however, O-antigen types tend to be restricted within species. This in turn means that O-antigen based vaccines can be targeted to a specific bacterial pathogen without affecting the microbiome. For example, the leading cause of hospital-acquired infections, Klebsiella pneumoniae (KP), has 8 known O-antigen types but only 4 (O1 , 02, 03 and 05) would need to be targeted to cover 80% of invasive infections. Indeed, both mAbs and vaccines targeting KP O1 and 02 O-antigen types have been demonstrated in preclinical models to be effective in preventing and treating infection by several academic groups and drug companies.

[0020] CPS is another carbohydrate antigen that covers the surface of most Gram-negative including common human pathogenic bacteria. In Klebsiella pneumoniae there is significant variation in CPS types with over 80 described in clinical isolates. Whilst LPS is known to be anchored via lipid A the exact mechanism for CPS anchoring is unknown. CPS is very abundant and significantly changes the physical properties of Klebsiella pneumoniae and is regarded one of the prominent virulence factors for infection.

[0021] The lipid A component of LPS is toxic to mammals which limits the administration of whole intact LPS molecules as vaccines. Lipid A has other additional drawbacks; it is ubiquitous across Gram-negative bacteria and not unique to a single species (off target vaccine effects against the microbiome). It is also inaccessible to antibodies as it sits in the outer membrane, where antibodies cannot reach. However, the O-antigen is not only specific, but it can also be safely administered as an immunogen. Therefore, it can be harnessed as a vaccine or to generate reactive B-cells and mAbs. The major issues with O-antigen precluding its widespread use in vaccines are set out below.

[0022] 1. LPS / CPS production. LPS is either harvested from target bacteria or synthetically generated. LPS harvesting is a low efficiency process which has changed little since the original protocols (the hot phenol-water method described by Westphal and Jann in 1965). The laborious nature of the process means that industrial scale production of LPS and derivatives re-mains difficult and economically costly. Despite some advancements (e.g., the phenol / chloroform / light petroleum by Galanos, Luderitz & Westphalin 1969) this remains a major bottleneck in producing LPS O-antigen for vaccine. Cell-free methods such as chemical synthesis are now being employed which have improved throughput but require individually tailored chemical conditions to generate each specific LPS O-antigen type. This means that each new O- antigen candidate requires intensive optimisation to enable enough to be generated through chemical synthesis. CPS isolation from bacteria suffers from the same problems associated with LPS production and requires laborious and bacteria species-specific isolation methods. Cell-free methods such as chemical synthesis are also now being employed here and require specific conditions for every CPS being made.

[0023] 2. Conjugation. O-antigen and CPS is solely composed of carbohydrate (which is not highly immunogenic) and as such makes a poor vaccine candidate. In order to generate a robust and enduring immune response both T and B-cells need to be stimulated but carbohydrate antigens (when administered alone) are unable to stimulate such a response. O-antigen can induce this type of response if it is chemically conjugated to a protein carrier molecule (such as EPA or CRM197). When carbohydrates and proteins are linked together as immunogens they are known as glycoconjugates. These are widely used as human vaccines especially in the case of CPS-based glycoconjugates including widespread vaccines conferring protection against Neisseria meningitidis, Streptococcus pneumoniae and Haemophilus influenzae type B. Whilst many CPS based-glycoconjugate vaccines exist there are no licensed O-antigen based vaccines on the market. Both LPS and CPS glycoconjugates are also used to immunise humans by pharmaceutical companies to derive reactive B-cells and produce mAbs.

[0024] In view of the above, there remains a need for an LPS molecule which can be efficiently produced and generates a robust immune response.Summary

[0025] The present invention (also termed “citrOgen”) is a highly adaptable and modular platform which harnesses the natural infection cycle and clearance of the enteric mouse pathogen Citrobacter rodentium (CR) to generate specific antibodies against user-selected antigens. The platform can generate antibodies against both polysaccharide and protein antigens which are, for example, used widely as the antigens for licensed vaccines.

[0026] Citrobacter rodentium is a mouse-restricted pathogen that mirrors the infection strategy of enteropathogenic Escherichia coll in humans. Oral inoculation of Citrobacter rodentium in mice results in highly reproducible infection cycle. In resistant mice (e.g., C57BL / 6, BALB / c, CD1), Citrobacter rodentium causes self-limiting infection. Resolution of infection is dependent on functional ar\t\-Citrobacter rodentium IgG antibody responses. This premise underpins the technology and contemporary genetic engineering enables modification of the Citrobacter rodentium genome to present heterologous antigens of choice. This means that Citrobacter rodentium itself continuously generates antigen throughout the infection until clearance. Mice start naive to Citrobacter rodentium infection and the course of natural infection is protracted (28 days), so antibodies to the heterologous antigen are affinity selected and matured in the process. Citrobacter rodentium infection is facilitated by injecting ‘effector’ proteins into host cells via a typeIll secretion system injectisome that interfere with cellular processes. Following injection of the effectors, Citrobacter rodentium tightly adheres to the colonic epithelium forming ‘attaching and effacing’ lesions.

[0027] citrOgen exploits this well-orchestrated and reliable response by employing genetically modified Citrobacter rodentium in which the genetic code for heterologous antigens has been introduced into the genome. Heterologous antigens are expressed and presented to the host in a physiologically relevant fashion in the context of an active and natural infection. The result is a polyclonal antibody response to Citrobacter rodentium, including antibodies to the heterologous antigen of choice. Surprisingly, it was discovered by the inventors that the magnitude of the antibody response can be boosted by overexpressing the type III secretion system effector protein EspO to levels above those observed during normal Citrobacter rodentium infection.

[0028] In view of the above, citrOgen therefore overcomes many of the issues in triggering the initial responses required to generate specific monoclonal antibodies against surface-expressed antigens such as polysaccharides and proteins. Other advantages of the invention are described herein.

[0029] Accordingly, in first aspect of the invention there is provided a Citrobacter rodentium strain which expresses higher levels of EspO, or a homologue thereof, compared to a control.

[0030] In a second aspect of the invention there is provided a composition comprising a Citrobacter rodentium strain described herein and a pharmaceutically acceptable excipient or diluent.

[0031] In a third aspect of the invention there is provided a method of producing antibodies comprising infecting a non-human animal with a Citrobacter rodentium strain described herein or a composition described herein, comprising incubating the non-human animal and extracting the antibodies from the non- human animal.

[0032] In a fourth aspect of the invention there is provided a method of producing a hybridoma comprising infecting a non-human animal with a Citrobacter rodentium strain described herein or a composition described herein, comprising incubating the non-human animal, extracting B cells from the non-human animal and using the B cells to produce the hybridoma.

[0033] In a fifth aspect of the invention there is provided a method of producing the Citrobacter rodentium strain described herein, comprising providing a Citrobacter rodentium strain and inserting the mrk operon genes of the different Gram-negative bacterium into the Citrobacter rodentium strain.

[0034] In a sixth aspect of the invention there is provided a method of producing the Citrobacter rodentium strain described herein, comprising providing a Citrobacter rodentium strain and substituting the rfb locus of the Citrobacter rodentium strain with a rfb locus from the different Gram-negative bacterium.

[0035] In a seventh aspect of the invention there is provided a method of producing the Citrobacter rodentium strain described herein, comprising providing a Citrobacter rodentium strain and substituting thecolanic acid biosynthesis locus of the Citrobacter rodentium strain with a cps locus from the different Gramnegative bacterium.

[0036] In an eighth aspect of the invention there is provided a Citrobacter rodentium strain described herein or a composition described herein for use in medicine.

[0037] In a ninth aspect of the invention there is provided a Citrobacter rodentium strain described herein or a composition described herein for use in the prevention or treatment of a Gram-negative bacterial infection.

[0038] The use of sub-titles of section headings herein is for ease of understanding the content of the application only and the subject matter under any specific heading or sub-heading is not to be construed as limited only to that described by the heading or sub-heading unless specifically indicated.Brief description of the drawings

[0039] Figure 1. The citrOgen platform induces antibody responses to an LPS O-antigen of choiceA. Wild type (non-genetically altered) Citrobacter rodentium (OR) lipopolysaccharide (LPS) is composed of three regions. Lipid A (anchoring the molecule in the membrane) covalently linked to inner / outer core then a specific signature of sugars called O-antigen. Genes involved in O-antigen production are encoded in the genome at a site called the rfb locus.B. citrOgen strains (e.g., CitroKP02a) express a hybrid LPS molecule. The rfb gene locus is replaced by advanced genetic engineering with those from a pathogen of choice. Citrogen acts to present the chosen pathogen’s O-antigen (here Klebsiella’s O-antigen in red) forming a hybrid LPS type.C. (1) In normal infection with wild type Citrobacter rodentium the natural LPS is presented on the surface the bacteria. (2) Experimental infection takes place with Citrobacter rodentium inoculated into the gut. (3) Citrobacter rodentium is a gut pathogen and reaches very high level by days 8-15 post oral inoculation. (4) The normal mouse immune response matures and results in antibody production from B-cells (around 2-3 weeks). (5) Some B-cells will produce antibodies which are specific to Citrobacter rodentium’s natural O- antigen (Fig 1A). (6) At 28 days post infection the immune response has cleared Citrobacter rodentium infection from the gut and this clearance requires IgG. (7) In CitrOgen infections the Citrobacter rodentium administered expresses a pathogen of choices O-antigen in a hybrid LPS (Fig 1 B). (8) The Citrogen strain is administered orally and (9) reaches high levels in the gut. (10) A robust immune response is triggered inducing reactive B-cell and antibody formation (11) against the hybrid LPS’s O-antigen (e.g., Klebsiella pneumoniae). (12) Clearance occurs at around 28 days post infection, but a reactive B-cell memory pool remains in the mouse to protect against repeated infection.

[0040] Figure 2. citrOgen induces protective anti-Klebsiella pneumoniae antibodiesA. Combining Citrogen infection (to vaccinate / induce specific antibodies) followed by Klebsiella pneumoniae (KP) pathogen challenge this platform can both elicit antibodies and test efficacy in one workflow. (1) Infection takes place as in Fig 1C. In this experiment three different inoculum were used with either wild type Citrobacter rodentium, CitroKPG2a (expressing the O2a O-antigen from KP) or CitroKP02aoverexpressing EspO. (2) Infection peaks, (3) induces an immune response and is cleared (4). (5) A blood sample is taken at clearance to test for antibodies. (6) 7 days later mice were infected with KP into the lungs and after 36 hours (7) lungs were collected and KP infection severity was assessed by quantification of bacterial numbers.B. Blood samples (from Fig 2A) (5) were analysed for antibodies to KP that expressed the O2a antigen. In Citrobacter rodentium infected mice low levels of antibodies are present. In CitroKP02a infected mice higher levels are induced. In CitroKP02a mice over expresses EspO the highest levels are induced and are very significantly increased. Here the responders above the WT Citrobacter rodentium mean are coloured in red for clarity.C. (i) Lungs were tested for the number of KP from Fig2A(5). When mice were challenged following CitrogenKP02a infection boosted with EspO a significant reduction in KP numbers (CFU, colony forming units) was observed, (ii) when the mice which were high responders were analysed, the response was even more significant, indicating that the titre of antibodies was contributing to the effect.

[0041] Figure 3. citrOgen can be used for other wzz LPS types on key pathogensBlood samples were analysed for antibodies to ST131 Escherichia coll that expressed the O25b antigen. In Citrobacter rodentium infected mice low levels of antibodies are present. In CitroECo25b<wzzcR) and CitroECo25b(wzzEcsTi3i)) infected mice the levels were very high with the highest levels observed in CitrOECo25b(wzzECST131) .

[0042] Figure 4. Strategy of constructing Citrobacter rodentium expressing the Klebsiella pneumoniae O1v1 -antigen (CRKPOI )The wild type Citrobacter rodentium (CRwr) O-antigen biosynthesis (rib) locus (A) was replaced with the Klebsiella pneumoniae rfb locus (KP O2a) (B) by seamless, scarless homologous recombination on the Citrobacter rodentium genome. For expression of KP O1v1 we subsequently inserted the KP glycosyltransferases wbbY and wbbZ downstream of the Citrobacter rodentium glmS gene to generate CRKPOI .The gene locus tags for all genes are listed in Table 1 . The Citrobacter rodentium and KP strains used for this work were ICC169 and ICC8001 , respectively.A. A schematic of the CRwr rfb locus bound by galF (5’) and gnd (3’).B. A schematic of the CRKPOI rfb locus in which the KP O2a locus replaces the genes between galF and gnd. KP O2a O-antigen requires modification by the glycosyltransferases wbbY and wbbZ to generate KP O1 -antigen These were inserted into the CRKPOI by transposon mutagenesis at the high frequency Tn7 insertion site downstream of the glmS gene.

[0043] Figure 5. EspO overexpression enhances colonisation of CRKPOI and augmenting antibody responses to the heterologous O-antigenA. The CRwr strain used for this work is ICC169, the KPwr used for this work is ICC8001. A schematic for testing the heterologous antigen presentation platform. 109colony forming units (CFUs) of Citrobacter rodentium were administered to mice by oral gavage. Specific pathogen free, female CD-1 mice aged 6-8 weeks were used for all infections. A stool sample was taken at 8 days post infection (dpi) to evaluate colonisation when stool Citrobacter rodentium shedding peaks. Themice recover from Citrobacter rodentium infection and a blood sample was taken at 30 dpi to assess antibody responses. Mice were then challenged with KPwr delivered by intratracheal inoculation. At 72 hours post infection (hpi) mice were killed and samples taken to assess the functional protection against KP.B. At 8dpi, CRKPOI results in a colonisation defect compared to CRwr This defect can be reversed by over expressing EspO in CRKPOI pespO. Graph shows median from 3 biologically independent repeats. Raw data was log 10 transformed to achieve normal distribution as assessed by D’Agostino and Pearson’s test for normal distribution. Significance was tested by ordinary one-way ANOVA, corrected for multiple comparisons to CRwr with Dunnett’s multiple comparison test. * p=0.0146.C. CRKPOI and CRKPOI pespO gut infection and clearance results in anti-KP antibodies. Antibody levels were measured by Enzyme Linked Immunosorbent Assay (ELISA). There is a significant increase in KP reactive antibody titres compared to control animals infected with CRwr (not expressing the heterologous KPO1 LPS). There is a trend to increased titres with EspO overexpression. Graph shows median from 3 biological repeats, statistical significance was determined by ordinary oneway ANOVA, where multiple comparisons were performed with c. **** p<0.001 .D. Antibodies raised by CRKPOI and CRKPOI pespO infection at clearance are specific to the KPO1 antigen. The sera from mice infected with CRwr, CRKPOI and CRKPOI pespO were incubated with KPwr (expressing KP O1 LPS) or KPartb (an isogenic mutant in which the rfb locus was deleted resulting in absent O-antigen). Binding was assessed by flow cytometry and detection with a fluorescently labelled anti-mouse IgG secondary antibody. Sera from CRwr infected mice do not bind KPwr or KPartb. Sera from CRKPOI and CRKPOI pespO infected mice specifically binds KPwr but not KPartb indicating that heterologous antibodies are directed to the KP O1 O-antigen. Graph shows mean±SEM and significance was determined by unpaired T-test, **** p<0.001 .

[0044] Figure 6. CRKPOI induces antibodies that bind KP LPSImmunofluorescence microscopy confirms the binding of sera from mice infected with CRKPOI and CRKPOI pespO to KPwr in a pattern matching that of a validated anti-KP O1 antibody. Sera from CRKPOI and CRKPOI pespO infected mice were pooled as both resulted in specific anti-KP O1 -antigen responses (Fig 2D). The mouse monoclonal antibody C13 was uses as an anti-KP O1 reference (PMID: 18725260). The CRwr strain used for this work is ICC169, the KPwr used for this work is ICC8001 .A. Sera from mice infected with CRwr does not bind KP.B. The anti-KP O1 reference antibody C13 binds KP with a pattern of staining around the circumference of the bacterial cells.C. Pooled sera from CRKPOI and CRKPOI pespO infections binds KP in the same pattern as (B), confirming that infection with CRKPOI and CRKPOI pespO result in KP O1 -antigen antibodies.

[0045] Figure 7. CRKPOI infection results in functionally protective responses against KP challenge, augmented by EspO overexpressionMice infected CRwr, CRKPOI and CRKPOI pespO were challenged with KP following Citrobacter rodentium clearance (as per Fig 2A). 500 CFU were administered by intratracheal inoculation. All graphs show median of 3 biologically independent repeats; ns=not significant p>0.05 throughout. The CRwr strain used for this work is ICC169, the KPwr used for this work is ICC8001 .A. Mice pre-infected with either CRKPOI or CRKPOI pespO experience significantly less weight loss compared to CRwr. ** p=0.0036, *** p=0.0002, Kruskall- Wallis test corrected for multiple comparisons using Dunn’s test.B. Blood glucose levels were tested at baseline in CRwr, CRKPOI and CRKPOI pespO pre KP challenge. Following KP challenge, CRwr and CRKPOI treated mice presented significant hypoglycaemia. CRKPOI pespO treated mice presented no hypoglycaemia. **** p<0.0001 ; * p=0.0350. ns=not significant. Ordinary one-way ANOVA, Dunnett’s multiple comparison test, with a single pooled variance.C. CRKPOI and CRKPOI pespO-infected mice have reduced KP bacterial burdens in the lung compared to CRwr. Raw data was Iog10 transformed to achieve normal distribution as assessed by D’Agostino and Pearson’s test for normal distribution. ** p=0.0067, **** p <0.0001 Ordinary one-way ANOVA, Dunnett’s multiple comparison test, with a single pooled variance.D. CRKPOI -infected mice have reduced KP bacterial burdens in the blood compared to CRwr. No KP was detected in the blood of CRKPOI pespO animals. *** p=0.0005, **** p<0.0001 . Kruskall-Wallis test, Dunn’s multiple comparison on Iog10-transformed data.E. CRKPOI and CRKPOI pespO treated mice have significantly higher total white cell counts in the blood follow KP challenge compared to CRwr. * p=0.0270, ** p=0.083. Kruskall-Wallis test, Dunn’s multiple comparison.F. CRKPOI and CRKPOI pespO treated mice have significantly higher lymphocyte counts in the blood follow KP challenge compared to CRwr. * p=0.0287, *** p=0.0009. Kruskall-Wallis test, Dunn’s multiple comparison.G. No changes in blood neutrophil counts were observed following KP challenge in CRwr, CRKPOI and CRKPOI pespO treated mice. Ns=not significant. Kruskall-Wallis test, Dunn’s multiple comparison.H. CRKPOI and CRKPOI pespO treated mice have significantly higher platelet counts in the blood follow KP challenge compared to CRwr. CRwr vs. CRKPOI* p=0.0125. CRwr vs. CRKPOI pespO * p=0.0465. Kruskall-Wallis test, Dunn’s multiple comparison.I. CRKPOI and CRKPOI pespO treated mice have lower levels of serum C-reactive protein compared to CRwr following KP challenge. ** p=0.0023, *** p=00.0003. Kruskall-Wallis test, Dunn’s multiple comparison.J. The CRKPOI pespO treated mice have higher levels of serum albumin compared to CRwr following KP challenge. *** p=0.0009. Kruskall-Wallis test, Dunn’s multiple comparison.K. CRKPOI and CRKPOI pespO treated mice have lower levels of serum creatinine compared to CRwr following KP challenge. ** p=0.0021 *** p=0.0002. Kruskall-Wallis test, Dunn’s multiple comparison.L. CRKPOI and CRKPOI pespO treated mice have lower levels of serum surfactant protein D compared to CRwr following KPwr challenge. * p=0.0118 **** p<0.0001. Kruskall-Wallis test, Dunn’s multiple comparison.

[0046] Figure 8. Construction of heterologous Escherichia coli O25b-antigen expressing Citrobacter rodentiumThe CRwr O-antigen biosynthesis (rfb) locus is replaced with the genes encoding the Escherichia coli (EC O25b) rfb locus by seamless, scarless homologous recombination on the Citrobacter rodentium genome. Citrobacter rodentium and Escherichia coli utilise the wzy-dependent lipopolysaccharide (LPS) export pathway and O-antigen chain length is determined by the wzz gene which lies adjacent but outside the rfblocus. An additional chain length determinant (fepE) is found in some bacterial species such as Salmonella enterica (SE) subspecies enterica serovar Paratyphi. The gene locus tags for all genes are listed in Table 1 . The CRwr strain used for this work is ICC169, the KPwr used for this work is ICC8001 , the Escherichia coli ST131 strain used for this work is EC958, the SE strain used for this work is ATCC9150.A. A schematic of the CRwr rfb locus bound by galF (5’) and gnd (3’).B. A schematic of the CRECO25DWZZCR rfb locus in which the Escherichia coli O25b locus replaces the genes between galF and gnd. The wzz gene is the CRwr version in this strain.C. A schematic of the CREco25bwzzsTi3i rfb locus in which the Escherichia coli O25b locus replaces the genes between galF and gnd. The wzz gene is replaced in this strain by the wzz gene from an Escherichia coli isolate (strain EC958).D. A schematic of the CREco25bwzzsn3i + fePE rfb locus in which the Escherichia coli O25b locus replaces the genes between galF and gnd. The wzz gene is replaced in this strain by the wzz gene from an Escherichia coli isolate (strain EC958) as per (C). In addition, the additional chain length determinant fep ) that is found in some bacterial species such as SE subspecies enterica serovar Paratyphi has been inserted at the glmS site.

[0047] Figure 9. CREco25bwzzcR, CREco25bwzzsri3i and CREco25bwzzsri3i + fepE have different O-antigen chain lengths and result in anti-Escherichia coli ST131 antibodiesThe CRwr strain used for this work is ICC169, the KPwr used for this work is ICC8001 , the Escherichia coli ST131 strain used for this work is EC958, the SE strain used for this work is ATCC9150.A. Lipopolysaccharide (LPS) preparations were separated by polyacrylamide gel electrophoresis and stained with silver stain. CREco25bwzzcR results in an LPS staining pattern with higher modal molecular mass than EC ST131 O25b. CREco25bwzzsTi3i LPS molecular mass matches that of Escherichia coli ST131 O25b. CREco25bwzzsTi3i + fepE results in very high molecular mass LPS.B. Mice were infected with CRWT, CREco25bwzzcR, CREco25bwzzsTi3i and CREco25bwzzsTi3i + fepE as per Figure 2A. At 8dpi, stool colonisation was the same in all mice. CREco25bwzzcR, CREco25bwzzsTi3i and CREco25bwzzsTi3i + fepE strains all contained the plasmid pespO. Ns=not significant. Raw data was Iog10 transformed to achieve normal distribution as assessed by D’Agostino and Pearson’s test for normal distribution. Graph represents median from 2 biological repeats; statistical significance was determined by ordinary one-way ANOVA, multiple comparison to CRwr with Dunnett’s multiple comparison test, with a single pooled variance.C. CREco25bwzzcR, CREco25bwzzsTi3i and CREco25bwzzsTi3i + fepE infection results in anti- Escherichia coli ST131 antibodies. Antibody levels were measured by Enzyme Linked Immunosorbent Assay (ELISA). There is a significant increase in anti- Escherichia coli ST131 reactive antibody titres compared to control animals infected with CRwr. Raw data was log 10 transformed to achieve normal distribution as assessed by D’Agostino and Pearson’s test for normal distribution. Graph represents median from 2 biological repeats; statistical significance was determined by ordinary one-way ANOVA, multiple comparisons with Tukey’s multiple comparison test. **** p<0.0001 .

[0048] Figure 10. Construction of heterologous Klebsiella pneumoniae K2 capsular antigen expressing Citrobacter rodentium (CRKPK2)The wild type Citrobacter rodentium (CRwr) colanic acid biosynthesis locus is replaced with the genes encoding the Klebsiella pneumoniae (KP) K2 capsule locus by seamless, scarless homologous recombination on the Citrobacter rodentium genome. The gene locus tags for all genes are listed in Table 1.The CRwr strain used for this work is ICC169, the KPwr used for this work is ICC8001.A. A schematic of the Citrobacter rodentium colanic acid biosynthesis locus bound by ROD21991 (5’) and galF (3’).B. A schematic of the CRKPK2 strain in which the colanic acid biosynthesis genes have been replaced with the KP K2 locus genes from wzi to manB.

[0049] Figure 11. CRKPK2 pespO infection results in Klebsiella pneumoniae K2 specific antibody responsesThe CRwr strain used for this work is ICC169, the KPwr used for this work is ICC8001 . All graphs represent the median of 3 biologically independent repeats.A. Mice were infected as per the schema in Figure 2A. At 8dpi CRKPK2 pespO colonised to the same level as CRwr. ns=not significant, p>0.05. Raw data was Iog10 transformed and groups were compared by Unpaired T-test.B. CRKPK2 pespO infection results in anti-KP antibodies. Antibody levels were measured by Enzyme Linked Immunosorbent Assay (ELISA). There was no binding, or difference in binding, to a KP mutant that does not express K2 capsule (AwcaJ) in sera from mice infected with CRwr or CRKPK2 pespO. Sera from CRKPK2 pespO resulted in antibodies to KP expressing K2 antigen (KPwr). Sera from CRwr did not result in binding to KP expressing K2 antigen (KPwr). Binding between CRwr and CRKPK2 pespO was measured by Mann-Whitney T-test. Ns=not significant, **** p<0.0001 .

[0050] Figure 12. CRKPK2 induces antibodies that bind KP LPSImmunofluorescence microscopy confirms the binding of sera from mice infected with CRKPK2 pespO to KP in a pattern matching that of a reference K2 antiserum (from the Serum Statens Institute). The CRwr strain used for this work is ICC169, the KPwr used for this work is ICC8001 .A. Sera from mice infection with CRwr do not bind KP.B. The reference K2 antiserum binds KP with a pattern of punctate circumferential staining.C. Pooled sera from CRKPK2 pespO infections binds KP in the same pattern as (B) confirming that infection with CRKPK2 pespO result in KP K2-antigen antibodies.

[0051] Figure 13. Construction of heterologous Klebsiella pneumoniae type 3 fimbrial expressing Citrobacter rodentium (CRpmapmrkAp).The structural genes required for Klebsiella pneumoniae type 3 fimbriae biosynthesis and export were cloned from ICC8001 , fused with the promoter for the Citrobacter rodentium type 3 secretion system effector gene map. The construct was inserted into the CRpmapmrkAF genome by transposon mutagenesis at the high frequency Tn7 insertion site downstream of the glmS gene. The gene locus tags for all genes are listed in Table 1. The CRwr strain used for this work is ICC169, the KPwr used for this work is ICC8001.

[0052] Figure 14. CRpmapmrkAF pespO infection results in Klebsiella pneumoniae type 3 fimbrial specific antibody responsesThe CRwr strain used for this work is ICC169, the KPwr used for this work is ICC8001 . All graphs represent the median from 3 biologically independent repeats.A. Mice were infected as perthe schema in Figure 2A. At 8dpi CRpmapmrkAF colonisation was statistically equivalent to that of CRwr. ns=not significant, p>0.05. Raw data was log 10 transformed and groups were compared by Unpaired T-test.B. CRpmapmrkAF pespO infection results in anti-KP antibodies. Antibody levels were measured by Enzyme Linked Immunosorbent Assay (ELISA). There was no binding, or difference in binding, to a KP mutant that does not express type 3 fimbriae (AmrkA-F) in sera from mice infected with CRwr or CRpmapmrkAF pespO. Sera from CRpmapmrkAF pespO resulted in antibodies to KP expressing type 3 fimbriae (KPwr). Sera from CRwr did bind to KP type 3 fimbriae (KPwr). Statistical significance between the levels of binding between CRwr and CRpmapmrkAF was assessed by Mann-Whitney T-test. Ns=not significant, **** p<0.0001 .

[0053] Figure 15. CRpmapmrkAF pespO induces antibodies that bind Klebsiella pneumoniae type 3 fimbriaeImmunofluorescence microscopy confirms the binding of sera from mice infected with CRpmapmrkAF pespO to KP in a pattern matching that of a reference MrkA polyclonal antibody (https: / / ww.cusabio.com / Polyclonal-Antibo y / mr A-Antibody-12552167.ht l). The CRwr strain used for this work is ICC169, the KPwr used for this work is ICC8001 .A. Sera from mice infection with CRwr do not bind KP.B. The reference MrkA polyclonal antibody binds KP with a pattern of filamentous circumferential staining.C. Pooled sera from CRpmapmrkAF pespO infections binds KP in the same pattern as (B) confirming that infection with CRpmapmrkAF pespO result in type 3 fimbrial antibodies.Detailed Description

[0054] In order for the present invention to be more readily understood, certain terms are first defined below. Additional definitions for the following terms and other terms are set forth throughout the specification.

[0055] As used herein, the term “approximately” or “about,” as applied to one or more values of interest, refers to a value that is similarto a stated reference value. In some embodiments, the term “approximately” or “about” refers to a range of values that fall within 25%, 24%, 23%, 22%, 21 %, 20%, 19%, 18%, 17%, 16%, 15%, 14%, 13%, 12%, 11 %, 10%, 9%, 8%, 7%, 6%, 5%, 4%, 3%, 2%, 1 %), or less in either direction (greater than or less than) of the stated reference value unless otherwise stated or otherwise evident from the context (except where such number would exceed 100% of a possible value).

[0056] The term “Gram-negative”, as used herein, refers to bacterium that does not retain the crystal violet stain used in the Gram staining method of bacterial differentiation.

[0057] The terms “cell-surface protein” and “protein complexes” usually refer to proteins that are exported and incorporated into the cell surface and displayed to the extracellular environment.Citrobacter rodentium

[0058] The present invention relates to the Gram-negative bacterium Citrobacter rodentium. In some embodiments the Citrobacter rodentium strain is ICC168 (GenBank ref.: FN543502.1). In some embodiments, a gene / protein referred to herein has at least 50%, 60%, 70%, 80%, 90%, 91 %, 92%, 93%, 94%, 95%, 96%, 97%, 98% or 99% identity to a sequence in Table 1 . In some embodiments, a gene / protein referred to herein comprises a sequence in Table 1. In some embodiments, a gene / protein referred to herein consists of a sequence in Table 1 .

[0059] Table 1 : Citrobacter rodentium (GenBank ref.: FN543502.1) genes / proteins described herein

[0060] In some embodiments, the protein Map is defined by UniProt reference D2TKE9 and has the amino acid sequence of SEQ ID NO: 75. In some embodiments, Map has at least 50%, 60%, 70%, 80%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98% or 99% identity to the sequence of SEQ ID NO: 75. In some embodiments, Map comprises a sequence of SEQ ID NO: 75. In some embodiments, Map consists of a sequence of SEQ ID NO: 75.

[0061] In some embodiments, the Citrobacter rodentium strain is DBS100 (GenBank ref.: CP038008.1). Citrobacter rodentium strain DBS100 genes and proteins described herein are identical to those of Citrobacter rodentium strain ICC168 (see Table 1).Klebsiella pneumoniae

[0062] The present invention relates to the Gram-negative bacterium Klebsiella pneumoniae. In some embodiments, the Klebsiella pneumoniae strain is KPPR1 (GenBank ref.: CP009208.1) which is derived from Klebsiella pneumoniae strain ATCC43816. In some embodiments, a gene / protein referred to herein has at least 50%, 60%, 70%, 80%, 90%, 91 %, 92%, 93%, 94%, 95%, 96%, 97%, 98% or 99% identity to a sequence in Table 2. In some embodiments, a gene / protein referred to herein comprises a sequence in Table 2. In some embodiments, a gene / protein referred to herein consists of a sequence in Table 2.

[0063] Table 2: Klebsiella pneumoniae (GenBank ref.: CP009208.1) genes / proteins described herein

[0064] In some embodiments, the wbbY gene is from Klebsiella pneumoniae strain CWK2 (GenBank ref.: MG458672.1) and has the nucleotide sequence of SEQ ID NO: 134 and amino acid sequence of SEQ ID NO: 135. In some embodiments, wbbY has at least 50%, 60%, 70%, 80%, 90%, 91 %, 92%, 93%, 94%, 95%, 96%, 97%, 98% or 99% identity to the sequence of SEQ ID NO: 134. In some embodiments, wbbY comprises a sequence of SEQ ID NO: 134. In some embodiments, wbbY consists of a sequence of SEQ ID NO: 134. In some embodiments, WbbY has at least 50%, 60%, 70%, 80%, 90%, 91 %, 92%, 93%, 94%, 95%, 96%, 97%, 98% or 99% identity to the sequence of SEQ ID NO: 135. In some embodiments, WbbY comprises a sequence of SEQ ID NO: 135. In some embodiments, WbbY consists of a sequence of SEQ ID NO: 135.

[0065] In some embodiments, the wbbZ gene is from Klebsiella pneumoniae strain NCTC11682 (GenBank ref.: UGMM01000004.1 , locus wbbZ_UGMM01000004) and has the nucleotide sequence of SEQ ID NO: 136 and amino acid sequence of SEQ ID NO: 137. In some embodiments, wbbZ has at least 50%, 60%, 70%, 80%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98% or 99% identity to the sequence of SEQ ID NO: 136. In some embodiments, wbbZ comprises a sequence of SEQ ID NO: 136. In some embodiments, wbbZ consists of a sequence of SEQ ID NO: 136.Escherichia coli

[0066] The present invention relates to the Gram-negative bacterium Escherichia coli. In some embodiments, the Escherichia coli strain is EC958 (GenBank ref.: HG941718). In some embodiments, a gene / protein referred to herein has at least 50%, 60%, 70%, 80%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98% or 99% identity to a sequence in Table 3. In some embodiments, a gene / protein referred to herein comprises a sequence in Table 3. In some embodiments, a gene / protein referred to herein consists of a sequence in Table 3.

[0067] Table 3: Escherichia coli genes / proteins described hereinSalmonella enterica

[0068] The present invention relates to the Gram-negative bacterium Salmonella enterica subsp. enterica serovar Paratyphi A. In some embodiments, the Salmonella enterica strain is ATCC 9150 (GenBank ref.: CP000026). In some embodiments, a gene / protein referred to herein has at least 50%, 60%, 70%, 80%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98% or 99% identity to a sequence in Table 4. In some embodiments, a gene / protein referred to herein comprises a sequence in Table 4. In some embodiments, a gene / protein referred to herein consists of a sequence in Table 4.

[0069] Table 4: Salmonella enterica genes / proteins described herein

[0070] In some embodiments, the (first) Gram-negative bacterium causes infections of the gastrointestinal tract such as Citrobacter rodentium or enteropathogenic Escherichia coli. In some embodiments the (first) Gram-negative bacterium causes infections of the gastrointestinal tract in rodents such as Citrobacter rodentium. In some embodiments the (first) Gram-negative bacterium causes infections of the gastrointestinal tract in humans such as enteropathogenic Escherichia coli. In some embodiments, the (first) Gram-negative bacterium is enterohaemorrhagic Escherichia coli. In some embodiments, the (first) Gram-negative bacterium is Shigella spp. In some embodiments, the (second / different) Gram-negative bacterium can be from a genus selected from Serratia, Haemophilus, Moraxella, Salmonella, Shigella, Neisseria, Klebsiella, Vibrio, Pseudomonas, Legionella and Helicobacter. In some embodiments the (second / different) Gram-negative bacterium is Klebsiella pneumoniae, Neisseria meningitidis, Haemophilus influenzae (type B), Pseudomonas aeruginosa, Salmonella, Escherichia coli (e.g., strain ST131) or Helicobacter pylori. In some embodiments the (second / different) Gram-negative bacterium is not Acinetobacter baumannii. In some embodiments, the first Gram-negative bacterium has a different O- antigen to the second / different Gram-negative bacterium. In some embodiments, the first Gram-negative bacterium has a different CPS to the second / different Gram-negative bacterium. In some embodiments, the surface polysaccharide, O-antigen or CPS of the first Gram-negative bacterium is structurally different to the surface polysaccharide, O-antigen or CPS, respectively, of the second / different Gram-negative bacterium. In some embodiments, the surface polysaccharide, O-antigen or CPS of the first Gram-negative bacterium is immunologically different to the surface polysaccharide, O-antigen or CPS, respectively, of the second / different Gram-negative bacterium. In some embodiments, the first Gram-negative bacterium is a different genus to the second / different Gram-negative bacterium. In some embodiments, the first Gramnegative bacterium is a different species to the second / different Gram-negative bacterium. In some embodiments the second / different Gram-negative bacterium does not cause infections of the gastrointestinal tract.

[0071] Some embodiments relate to the prevention or treatment of a Gram-negative bacterial infection. Such infection can be caused by any of the Gram-negative bacterium described herein.

[0072] The term "comparable", as used herein, refers to a system, set of conditions, effects, or results that is / are sufficiently similar to a test system, set of conditions, effects, or results, to permit scientifically legitimate comparison. Those of ordinary skill in the art will appreciate and understand which systems, sets of conditions, effects, or results are sufficiently similar to be "comparable" to any particular test system, set of conditions, effects, or results as described herein.

[0073] The term "correlates", as used herein, has its ordinary meaning of "showing a correlation with". Those of ordinary skill in the art will appreciate that two features, items or values show a correlation with one another ifthey show a tendency to appear and / or to vary, together. In some embodiments, a correlation is statistically significant when its p-value is less than 0.05; in some embodiments, a correlation is statistically significant when its p-value is less than 0.01 . In some embodiments, correlation is assessed by regression analysis. In some embodiments, a correlation is a correlation coefficient.

[0074] As used herein, the terms "improve," "increase", “higher”, "reduce," or grammatical equivalents, indicate values that are relative to a reference (e.g., baseline) measurement, such as a measurement taken under comparable conditions (e.g., in the same individual prior to initiation of treatment described herein, or a measurement in a control individual (or multiple control individuals) in the absence of treatment) described herein. In some embodiments, a control Gram-negative bacterium (e.g., Citrobacter rodentium) is a wild type Gram-negative bacterium. In some embodiments, a control Gram-negative bacterium (e.g., Citrobacter rodentium) is an unmodified Gram-negative bacterium. In some embodiments, the control is a wild type Citrobacter rodentium strain. In some embodiments, the control is Citrobacter rodentium strain ICC168 or DBS100. Where higher levels of EspO expression is achieved by cloning the gene espO into a suitable vector (e.g., a commercially available vector such as pACYC184) to generate a plasmid and introducing the plasmid into the Citrobacter rodentium strain, the control may be the same Citrobacter rodentium strain without the plasmid or with a control plasmid (e.g., the same plasmid without the gene espO). Alternatively, the control may be the unmodified Citrobacter rodentium strain. Suitable controls and assays for determining levels of expression of proteins, including EspO, is known in the art. In some embodiments, the Citrobacter rodentium strain expresses higher levels of EspO compared to a control.

[0075] As used herein, the term "subject", "individual", or "patient" refers to any organism upon which embodiments of the invention may be used or administered, e.g., for experimental, diagnostic, prophylactic, and / or therapeutic purposes. Typical subjects include animals (e.g., mammals such as mice, rats, rabbits, non-human primates, and humans; insects; worms; etc.).

[0076] As used herein, “heterologous antigen” refers to antigens not naturally expressed in the context of an organism described herein.

[0077] As used herein, "carbohydrate" means a naturally occurring carbohydrate, a modified carbohydrate, or a carbohydrate derivative. A carbohydrate is a biomolecule including carbon (C), hydrogen (H) and oxygen (O) atoms. Carbohydrates can include monosaccharide, disaccharides, trisaccharides, tetrasaccharides, oligosaccharides or polysaccharides.

[0078] The term "surface polysaccharide", as used herein, refers to a polysaccharide which is normally present or expressed on the cell surface. Surfaces polysaccharides include, but are not limited to, O- antigen (part of LPS) and CPS. In some embodiments, the O-antigen is from Klebsiella pneumoniae (e.g., O1v1 , O1v2, O1v3, O2a, O2afg, O2ac, 03, O3a, O3b, 04, 05, 012, 013, OL102, OL103 or OL104, for nomenclature see, for example: Lam et al. Kaptive 2.0: updated capsule and lipopolysaccharide locustyping for the Klebsiella pneumoniae species complex, 2022:(8)3). Preferably, the O-antigen from Klebsiella pneumonia is O1v1 or O2a. In other embodiments, the O-antigen is from Pseudomonas aeruginosa and the O-antigen is 05. In other embodiments, the O-antigen is from Escherichia coll and the O-antigen is O25b. In some embodiments, the CPS is from Klebsiella pneumoniae (e.g., K2).

[0079] As used herein, "expression" means the process by which a gene ultimately results in a protein. Expression includes, but is not limited to, transcription, post-transcriptional modification (e.g., splicing, polyadenylation, addition of 5 '-cap), and translation.

[0080] As used herein, "transcription" or "transcribed" refers to the first of several steps of DNA based gene expression in which a target sequence of DNA is copied into RNA (especially mRNA) by the enzyme RNA polymerase. During transcription, a DNA sequence is read by an RNA polymerase, which produces a complementary, antiparallel RNA sequence called a primary transcript.

[0081] As used herein, "pharmaceutically acceptable excipient or diluent" means any substance suitable for use in administering to an animal. In some embodiments, a pharmaceutically acceptable excipient or diluent is sterile saline. In some embodiments, such sterile saline is pharmaceutical grade saline.

[0082] A (pharmaceutical) composition according to the present invention may be presented in a form that is ready for immediate use. Alternatively, the composition may be presented in a form that requires some preparation prior to administration.

[0083] (Pharmaceutical) compositions of the invention may be adapted for administration by any appropriate route, for example by the oral (including buccal or sublingual), topical (including buccal, sublingual or transdermal), or parenteral (including subcutaneous, intramuscular, intravenous, intraperitoneal or intradermal) route. Typically, the composition will be adapted for intravenous administration.

[0084] The pharmaceutically acceptable carrier, diluent or excipient that is present in the pharmaceutical compositions of the invention may be any suitable pharmaceutically acceptable carrier, diluent or excipient that is known in the art.

[0085] (Pharmaceutical) compositions adapted for intravenous administration may include aqueous and non-aqueous sterile injection solution which may contain anti-oxidants, buffers, bacteriostats and solutes which render the formulation substantially isotonic with the blood of the intended recipient; and aqueous and non-aqueous sterile suspensions which may include suspending agents and thickening agents.

[0086] Excipients which may be used for injectable solutions include water, alcohols, polyols, glycerine and vegetable oils, for example. The compositions may be presented in unit-dose or multidose containers, for example sealed ampoules and vials, and may be stored in a freeze-dried (lyophilized) condition requiring only the addition of the sterile liquid carried, for example water for injections, immediately prior to use.Extemporaneous injection solutions and suspensions may be prepared from sterile powders, granules and tablets.

[0087] The (pharmaceutical) compositions may contain preserving agents, solubilising agents, stabilising agents, wetting agents, emulsifiers, sweeteners, colourants, odourants, salts, buffers, coating agents or antioxidants.

[0088] The (pharmaceutical) compositions may be formulated as a vaccine. A Gram-negative bacterium formulated in a vaccine may be live attenuated by any suitable means.

[0089] Dosages of the Gram-negative bacterium or (pharmaceutical) composition of the invention can vary between wide limits, depending upon the disease or disorder to be treated, the age and condition of the individual to be treated, whether the treatment is prophylactic or therapeutic, the type, onset, progression, severity, frequency, duration, or probability of the disease to be treated, the clinical endpoint desired, previous or simultaneous treatments, etc. Dosages can be based upon current existing protocols, empirically determined, using animal disease models or optionally in human clinical trials. A physician will ultimately determine appropriate dosages to be used. The dosage may be increased or decreased depending on any adverse side effects, complications or other risk factors of the treatment or therapy and the status of the subject.

[0090] The Gram-negative bacterium or (pharmaceutical) composition is typically administered to the subject in need thereof in a “therapeutically effective amount”. By “therapeutically effective amount” is meant an amount sufficient to show a therapeutic benefit to the subject, i.e., to prevent, reduce or relieve the symptoms of a (Gram-negative) bacterial infection.

[0091] The subject is typically a human subject, but the invention may also find use in veterinary medicine and the subject may therefore be an animal, typically a mammal, for example a companion animal such as a dog, cat, rabbit, rat or mouse or an agricultural animal such as a cow, sheep, pig, horse, deer, chicken or goat, or a primate such as a chimpanzee, gorilla or monkey.

[0092] As used herein, “treatment” is also intended to cover preventative treatment, i.e., prophylaxis.

[0093] The term "comprising" is used herein to mean including the method steps or elements identified, but that such steps or elements do not comprise an exclusive list and as such there may be present additional steps or elements.

[0094] Further, to the extent that the term "includes" is used in either the detailed description or the claims, such term is intended to be inclusive in a manner similar to the term "comprising" as "comprising" is interpreted when employed as a transitional word in a claim.

[0095] “Identity” as known in the art is the relationship between two or more polypeptide sequences or two or more polynucleotide sequences, as determined by comparing the sequences. In the art, identity alsomeans the degree of sequence relatedness between polypeptide or polynucleotide sequences, as the case may be, as determined by the match between strings of such sequences. While there exist a number of methods to measure identity between two polypeptide or two polynucleotide sequences, methods commonly employed to determine identity are codified in computer programs. Preferred computer programs to determine identity between two sequences include, but are not limited to, GCG program package (Devereux, et al., Nucleic Acids Research, 12, 387 (1984), BLASTP, BLASTN, and FASTA (Atschul et al., J. Molec. Biol. 215, 403 (1990)).

[0096] One can use a program such as the CLUSTAL program to compare amino acid sequences. This program compares amino acid sequences and finds the optimal alignment by inserting spaces in either sequence as appropriate. It is possible to calculate amino acid identity or similarity (identity plus conservation of amino acid type) for an optimal alignment. A program like BLASTx will align the longest stretch of similar sequences and assign a value to the fit. It is thus possible to obtain a comparison where several regions of similarity are found, each having a different score. Both types of identity analysis are contemplated in the present invention.

[0097] The percent identity of two amino acid sequences or of two nucleic acid sequences is determined by aligning the sequences for optimal comparison purposes (e.g., gaps can be introduced in the first sequence for best alignment with the sequence) and comparing the amino acid residues or nucleotides at corresponding positions. The “best alignment” is an alignment of two sequences which results in the highest percent identity. The percent identity is determined by the number of identical amino acid residues or nucleotides in the sequences being compared (i.e., % identity = number of identical positions / total number of positions x 100).

[0098] The determination of percent identity between two sequences can be accomplished using a mathematical algorithm known to those of skill in the art. An example of a mathematical algorithm for comparing two sequences is the algorithm of Karlin and Altschul (1990) Proc. Natl. Acad. Sci. USA 87:2264- 2268, modified as in Karlin and Altschul (1993) Proc. Natl. Acad. Sci. USA 90:5873-5877. The NBLAST and XBLAST programs of Altschul, et al. (1990) J. Mol. Biol. 215:403-410 have incorporated such an algorithm. BLAST nucleotide searches can be performed with the NBLAST program, score = 100, wordlength = 12 to obtain nucleotide sequences homologous to nucleic acid molecules. BLAST protein searches can be performed with the XBLAST program, score = 50, wordlength = 3 to obtain amino acid sequences homologous to protein molecules for use in the invention. To obtain gapped alignments for comparison purposes, Gapped BLAST can be utilised as described in Altschul et al. (1997) Nucleic Acids Res. 25:3389-3402. Alternatively, PSI-Blast can be used to perform an iterated search which detects distant relationships between molecules (Id.). When utilising BLAST, Gapped BLAST, and PSI-Blast programs, the default parameters of the respective programs (e.g., XBLAST and NBLAST) can be used. See http: / / www.ncbi.nlm.nih.gov. Another example of a mathematical algorithm utilised for the comparison of sequences is the algorithm of Myers and Miller, CABIOS (1989). The ALIGN program (version 2.0) which is part of the CGC sequence alignment software package has incorporated such an algorithm. Other algorithms for sequence analysis known in the art include ADVANCE and ADAM as described in Torellis and Robotti (1994) Comput. Appl. Biosci., 10 :3-5; and FASTA described in Pearson and Lipman (1988)Proc. Natl. Acad. Sci. 85:2444-8. Within FASTA, ktup is a control option that sets the sensitivity and speed of the search.

[0099] As used herein, “EspO” refers to a virulence factor which is injected into intestinal epithelial cells by the type III secretion system during infection and facilitates adhesion to host cells in vivo. In some embodiments, the EspO is from Citrobacter rodentium. In some embodiments, espO has the sequence: atgccattgtcaataagaaatatattttcacgtgcttctacacatcgaccagaaatatccgggcctgtaattgataaaccaatacctaaaaattgcactc tgatttcaagcacatgtaatttggatggtataatggtcataaacagaaggaccagcttttatgatataaagccgcctggtgccggagaacgacagcc atccctaaaaatttctgcctcagaagctcagtggatgtgcaaaattatagagaccgagattaataactcaaataaatcctga (SEQ ID NO: 164). In some embodiments, espO has at least 50%, 60%, 70%, 80%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98% or 99% identity, at the nucleotide level, to the nucleotide sequence of SEQ ID NO: 164. In some embodiments, espO comprises the sequence of SEQ ID NO: 164. In some embodiments, espO consists of consists essentially ofthe sequence of SEQ ID NO: 164. In some embodiments, EspO has the sequence defined by Uniprot reference D2TI12. In some embodiments, EspO has the sequence defined by GenBank accession CBG90789.1. In some embodiments, EspO has the sequence defined by genetic locus tag ROD40881 from strain ICC168 (GenBank genome reference FN543502.1). In some embodiments, the EspO has the sequence:MPLSIRNIFSRASTHRPEISGPVIDKPIPKNCTLISSTCNLDGIMVINRRTSFYDIKPPGAGERQPSLKISASE AQWMCKIIETEINNSNKS (SEQ ID NO: 165). In some embodiments, EspO has at least 50%, 60%, 70%, 80%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98% or 99% identity, at the amino acid level, to the amino acid sequence of SEQ ID NO: 165. In some embodiments, EspO comprises the sequence of SEQ ID NO: 165. In some embodiments, EspO consists of consists essentially of the sequence of SEQ ID NO: 165. In some embodiments, EspO is a homologue of the EspO from Citrobacter rodentium (for example, from pathogenic Escherichia coli species). In some embodiments, the strain of the invention (e.g., of Citrobacter rodentium) expresses higher levels of EspO compared to a control described herein. Such higher levels of EspO expression can be achieved by any means known in the art (e.g., general molecular biology techniques known in the art). For example, such higher levels of expression can be achieved by cloning the gene espO (e.g., with a sequence described herein, for example, SEQ ID NO: 164) into a suitable vector (e.g., a commercially available vector such as pACYC184), generating a plasmid and then introducing the plasmid into the strain by any means known in the art. The plasmid can also include a selective marker (e.g., a resistance gene such as a tetracycline resistance gene). Thus, a strain of the invention can comprise a plasmid comprising the gene espO, or a homologue thereof. In some embodiments, the homologue of EspO is a protein with an equivalent function in a different Gram-negative bacterium. Alternatively, homologue of EspO can be a protein with at least 50%, at least 60%, at least 70%, at least 80%, at least 90%, at least 90%, at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98% or at least 99% identity, at the amino acid level, to the amino acid sequence of SEQ ID NO: 165, or another sequence of EspO described herein. In some embodiments, the homologue of EspO is OspE (from Shigella spp.). In some embodiments, the homologue of EspO is SopO (Salmonella spp.). Evidence of homologues of Citrobacter rodentium EspO working in accordance with the invention is found, for example, in Chatterjee S, et al. The type III secretion system effector EspO of enterohaemorrhagic Escherichia coli inhibits apoptosis through an interaction with HAX-1. Cell Microbiol. 2021 Sep;23(9):e13366.

[0100] In some embodiments, the Gram-negative bacterium has a substituted (rfb) locus with a (rfb) locus from a second / different Gram-negative bacterium. The rfb loci can be one of 2 types: 1) ATP-binding Cassette (ABC) transporters (for example, used by Klebsiella to produce O1v1 , O1v2, O2a, O2afg, 03, O3a, O3b and 05) or 2) Wzy-dependent pathways (as used by the majority of Gram- negatives bacteria, for example, Escherichia coli). Substitution can be performed in any appropriate manlCC168ner, e.g., by using insertion cassettes, or enzymatically through the use of a recombinase such as, for example, Cre- loxP system, Cre recombinase or a lambda phage recombinase system. The Cre recombinase can utilize known lox sites compatible with Cre recombinase (e.g., Iox66 and Iox71 sites). Alternatively, substitution can be performed through the action of a nuclease such as, for example, Type II CRISPR Cas9. The substitution can also be performed with a homologous recombination vector or insertion cassette containing regions of sequence homology to a region in the genome where an insertion or deletion is desired. The homologous recombination vector can be incorporated by a single cross-over event or a double cross-over event. The substitution can also be performed through use of an integrase, such as, for example, PhiC31 or bxb1 , or through the use of a suicide vector. In some examples the vector is assembled in vitro and subsequently transformed and amplified in Escherichia coli. In some examples the vector is assembled in S. cerevisiae. In some examples, the amplified vector is introduced into the Gram-negative bacteria by conjugation, electroporation, chemical competence, biolistics, transduction, or via natural competence. Transformation of cells with any construct described herein can be performed by any suitable method, for example bacterial conjugation, electroporation, or chemical transformation.EmbodimentsThe Citrobacter rodentium strain

[0101] The Citrobacter rodentium strain may comprise one or more heterologous molecules present on the cell surface.

[0102] The one or more heterologous molecules may comprise lipopolysaccharide (LPS).

[0103] The one or more heterologous molecules may comprise capsular polysaccharide (CPS).

[0104] The one or more heterologous molecules may comprise a cell-surface protein or protein complex.

[0105] The cell-surface protein or protein complex may comprise type 3 fimbriae.

[0106] The control may be a wild type Citrobacter rodentium strain.

[0107] The wild type Citrobacter rodentium strain may be ICC168 or DBS100.

[0108] The one or more heterologous molecules may be from a different Gram-negative bacterium.

[0109] The different Gram-negative bacterium may be a pathogenic Gram-negative bacterium.

[0110] The different Gram-negative bacterium may be from a genus selected from the group consisting of Serratia, Haemophilus, Moraxella, Salmonella, Shigella, Neisseria, Klebsiella, Vibrio, Pseudomonas, Escherichia, Helicobacter and Legionella.

[0111] The different Gram-negative bacterium may be Serratia marcescens, Haemophilus influenzae type B, Moraxella catarrhalis, Salmonella enterica, Shigella dysenteriae, Shigella flexneri, Shigella boydii, Shigella sonnei, Klebsiella pneumoniae, Klebsiella aerogenes, Klebsiella oxytoca, Klebsiella variicola, Klebsiella quasipneumoniae, Klebsiella michiganensis, Vibrio cholerae, Vibrio parahaemolyticus,, Pseudomonas aeruginosa, Neisseria meningitidis, Escherichia coll, Helicobacter pylori or Legionella pneumophila.

[0112] The different Gram-negative bacterium may be Escherichia coll.

[0113] The different Gram-negative bacterium may be Escherichia coll sequence type ST131 .

[0114] The different Gram-negative bacterium may be Klebsiella pneumoniae.

[0115] The O-antigen of the LPS may be selected from the group consisting of: Klebsiella pneumoniaeO1v1 , O1v2, O1v3, O2a, O2afg, O2ac, 03, O3a, O3b, 04, 05, 012, 013, OL102, OL103 and OL104.

[0116] The O-antigen may be O1 v1 .

[0117] The O-antigen of the LPS may be O25b.

[0118] The CPS may be K2.

[0119] The rfb locus may be substituted with a rfb locus from the different Gram-negative bacterium.

[0120] The colanic acid biosynthesis locus of the Citrobacter rodentium strain may be substituted with a cps locus from the different Gram-negative bacterium.

[0121] The Citrobacter rodentium strain may further comprise the cps locus from the different Gramnegative bacterium.

[0122] The Citrobacter rodentium strain may further comprise the genes wbbY and wbbZ.

[0123] Rhe wzz gene of the Citrobacter rodentium strain may be substituted with the wzz gene from the different Gram- negative bacterium.

[0124] The wzz gene of Citrobacter rodentium may be wzzB, or a gene with at least 50% sequence identity to wzzB.

[0125] The Citrobacter rodentium strain may further comprise the gene fepE.

[0126] The Citrobacter rodentium strain may further comprise the mrk operon genes mrkA, mrkB, mrkC, mrkD and mrkF of the different Gram-negative bacterium.

[0127] The mrk operon genes may be inserted into the Tn7 site of Citrobacter rodentium.

[0128] The mrk operon genes may be from Klebsiella pneumoniae.

[0129] The one or more genes may be under control of the / er-regulon.

[0130] The one or more of the genes may be selected from the group consisting of wbbY, wbbZ, wzz, fepE, mrkA, mrkB, mrkC, mrkD and mrkF and may be under control of the / er-regulon.

[0131] The genes mrkA, mrkB, mrkC, mrkD and mrkF may be under control of the / er-regulon.Methods

[0132] The non-human animal may be a rodent.

[0133] The rodent may be a mouse.

[0134] Infection may comprise introducing the Citrobacter rodentium strain into the gut of the non-human animal.

[0135] The method may further comprise screening for the antibodies or the B cells.

[0136] The antibodies or B cells may be extracted from secondary lymphoid tissue from the non-human animal.

[0137] The antibodies may be extracted from the blood of the non-human animal.

[0138] The non-human animal may be incubated for a sufficient amount of time to clear the Citrobacter rodentium strain from the non-human animal prior to extracting the antibodies or B cells.

[0139] The non-human animal may be incubated for 28 days prior to extracting the antibodies or B cells.

[0140] The non-human animal may be maintained for a sufficient amount of time to clear the Citrobacter rodentium strain from the non-human animal prior to extracting the antibodies or B cells.

[0141] The non-human animal may be maintained for 28 days prior to extracting the antibodies or B cells.

[0142] The antibodies or antibodies produced by the B cells may be specific for LPS.

[0143] The antibodies or antibodies produced by the B cells may be specific for CPS.

[0144] The antibodies or antibodies produced by the B cells may be specific for type 3 fimbriae.

[0145] The method may further comprise infecting the non-human animal with the different Gram-negative bacterium and screening the non-human animal for a immunogenic response against the different Gramnegative bacterium.

[0146] The different Gram-negative bacterium may be Klebsiella pneumoniae and infecting the non- human animal with Klebsiella pneumoniae may be infecting the lung of the non-human animal.

[0147] Infection with the Citrobacter rodentium strain may results in higher titres of antibodies specific for the one or more heterologous molecules present on the cell surface in the non-human animal compared to a control.

[0148] The control may be Citrobacter rodentium strain ICC168 or DBS100.

[0149] The provided Citrobacter rodentium strain may be strain ICC168 or DBS100.

[0150] The different Gram-negative bacterium may be Klebsiella pneumoniae.LPS

[0151] The surface polysaccharide may be an O-antigen.

[0152] The surface polysaccharide may be a capsular polysaccharide (CPS).

[0153] The first Gram-negative bacterium may be from the genus Citrobacter.

[0154] The first Gram- negative bacterium may be Citrobacter rodentium.

[0155] The second / different Gram-negative bacterium may be a pathogenic Gram-negative bacterium.

[0156] The second / different Gram-negative bacterium may be from a genus selected from the group consisting of Serratia, Haemophilus, Moraxella, Salmonella, Shigella, Neisseria, Klebsiella, Vibrio, Pseudomonas, Escherichia, Helicobacter and Legionella.

[0157] The second / different Gram-negative bacterium may be Klebsiella pneumoniae, Pseudomonas aeruginosa, Neisseria meningitidis, Haemophilus influenzae type B, Escherichia coli or Helicobacter pylori.

[0158] The second / different Gram-negative bacterium may be Escherichia coli strain ST131 .

[0159] The second / different Gram-negative bacterium may be Klebsiella pneumoniae and the O-antigen may be O1v1 , O1v2, O2a, O2afg, 03, O3a, O3b or 05.

[0160] The second / different Gram-negative bacterium may be Klebsiella pneumoniae and the O-antigen may be O1v1 or O2a.

[0161] The second / different Gram-negative bacterium may be Pseudomonas aeruginosa and the O- antigen may be 05.

[0162] The second / different Gram-negative bacterium may be Escherichia coli and the O-antigen may be O25b.

[0163] The second / different Gram-negative bacterium may be Klebsiella pneumoniae and the CPS may be K2.Gram-negative bacterium

[0164] The rfb locus may be substituted with a rfb locus from the second / different Gram-negative bacterium.

[0165] The rfb locus may be substituted with an cps locus from the second / different Gram-negative bacterium. Alternatively, rfb locus may not be substituted with a cps locus from the second / different Gramnegative bacterium.

[0166] The Gram-negative bacterium may be Citrobacter rodentium and the Citrobacter rodentium rfb locus may be substituted with a rfb locus from the second / different Gram-negative bacterium.

[0167] The Gram-negative bacterium may be Citrobacter rodentium and the Citrobacter rodentium rfb locus may be substituted with an cps locus from the second / different Gram-negative bacterium.

[0168] The Gram-negative bacterium may express higher levels of EspO, or a homologue thereof, compared to a control.

[0169] The wzz gene may be substituted with the wzz gene from the second / different Gram-negative bacterium.

[0170] The LPS may be expressed on the cell surface.Prevention or treatment of Gram-negative bacterial infections

[0171] The Gram-negative bacterial infection may be an infection caused by the second / different Gramnegative bacterium. The infection may be colitis.

[0172] The (Citrobacter rode nti urn) strain of the invention may be used in the prevention or treatment of a Gram-negative bacterial infection.

[0173] Accordingly, also provided is a method of treating or preventing Gram-negative bacterial infection comprising administering a (Citrobacter rodentium) strain or composition comprising the same to a subject in need thereof.

[0174] Accordingly, also provided is use of a (Citrobacter rodentium) strain or composition comprising the same for the manufacture of a medicament for the treatment of a Gram-negative bacterial infection.Method of producing antibodies

[0175] The term "antibody" includes intact antibodies, fragments of antibodies, e.g., Fab, F(ab') 2 fragments, and intact antibodies and fragments that have been mutated either in their constant and / or variable region (e.g., mutations to produce chimeric, partially humanized, or fully humanized antibodies, as well as to produce antibodies with a desired trait, e.g., enhanced IL-13 binding and / or reduced FcR binding). The antibody may be polyclonal or monoclonal. Preferably, the antibody is an IgG monoclonal antibody.

[0176] The non-human animal may be a rodent. The rodent may be a mouse.

[0177] Immunising may comprise introducing the Gram-negative bacterium into the gut of the non-human animal.

[0178] The method can further comprise screening for antibodies specific for the surface polysaccharide.

[0179] The screening for antibodies specific for the surface polysaccharide may be achieved by whole bacterial cell ELISA (enzyme-linked immunosorbent assay).

[0180] The antibodies may be extracted from secondary lymphoid tissue from the non-human animal.

[0181] The non-human animal may be incubated for a sufficient amount of time to clear the Gram-negative bacterium from the non-human animal prior to extracting the antibodies.

[0182] The non-human animal may be incubated for 28 days prior to extracting the antibodies.

[0183] The non-human animal may be maintained for a sufficient amount of time to clear the Gramnegative bacterium from the non-human animal prior to extracting the antibodies.

[0184] The non-human animal may be maintained for 28 days prior to extracting the antibodies.

[0185] The surface polysaccharide may be an O-antigen.

[0186] The surface polysaccharide may a CPS.molecule embodiments1. A lipopolysaccharide (LPS) comprising a surface polysaccharide, outer core, inner core and lipid A; wherein the outer core, inner core and lipid A are from a first Gram-negative bacterium and the surface polysaccharide is from a second Gram-negative bacterium, wherein the first Gram-negative bacterium is different from the second Gram-negative bacterium.2. The LPS according to paragraph 1 , wherein the surface polysaccharide is an O-antigen.3. The LPS according to paragraph 1 , wherein the surface polysaccharide is a capsular polysaccharide (CPS).4. The LPS according to any one of paragraphs 1 to 3, wherein the first Gram-negative bacterium is from the genus Citrobacter.5. The LPS according to paragraph 4, wherein the first Gram-negative bacterium is Citrobacter rodentium.6. The LPS according to any one of paragraphs 1 to 5, wherein the second Gram-negative bacterium is a pathogenic Gram-negative bacterium.7. The LPS according to any one of paragraphs 1 to 6, wherein the second Gram-negative bacterium is from a genus selected from the group consisting of Serratia, Haemophilus, Moraxella, Salmonella, Shigella, Neisseria, Klebsiella, Vibrio, Pseudomonas, Escherichia, Helicobacter and Legionella.8. The LPS according to paragraph 7, wherein the second Gram-negative bacterium is Klebsiella pneumoniae, Pseudomonas aeruginosa, Neisseria meningitidis, Haemophilus influenzae type B, Escherichia coll or Helicobacter pylori.9. The LPS according to paragraph 8, wherein the second Gram-negative bacterium is Escherichia coli strain ST131 .10. The LPS according to any one of paragraphs 1 to 8, wherein the second Gram-negative bacterium is Klebsiella pneumoniae and the O-antigen is O1 v1 , O1 v2, O2a, O2afg, 03, O3a, O3b or 05.11. The LPS according to paragraph 10, wherein the second Gram-negative bacterium is Klebsiella pneumoniae and the O-antigen is O1v1 or 02a.12. The LPS according to any one of paragraphs 1 to 8, wherein the second Gram-negative bacterium is Pseudomonas aeruginosa and the O-antigen is 05.13. The LPS according to any one of paragraphs 1 to 8, wherein the second Gram-negative bacterium is Escherichia coll and the O-antigen is O25b.14. The LPS according to any one of paragraphs 1 to 8, wherein the second Gram-negative bacterium is Klebsiella pneumoniae and the CPS is K2.15. A Gram-negative bacterium expressing the LPS according to any one of paragraphs 1 to 14.16. The Gram-negative bacterium according to paragraph 15, wherein the rfb locus has been substituted with a rfb locus from the second Gram-negative bacterium.17. The Gram-negative bacterium according to paragraph 15, wherein the rfb locus has been substituted with an cps locus from the second Gram-negative bacterium.18. The Gram-negative bacterium according to paragraph 15 or 16, wherein the Gram-negative bacterium is Citrobacter rodentium and the Citrobacter rodentium rfb locus has been substituted with a rfb locus from the second Gram-negative bacterium.19. The Gram-negative bacterium according to paragraph 15 or 17, wherein the Gram-negative bacterium is Citrobacter rodentium and the Citrobacter rodentium rfb locus has been substituted with an cps locus from the second Gram-negative bacterium.20. The Gram-negative bacterium according to any one of paragraphs 15 to 19, wherein the Gramnegative bacterium expresses higher levels of EspO, or a homologue thereof, compared to a control.21. The Gram-negative bacterium according to any one of paragraphs 15, 16, 18 or 20, wherein the wzz gene has been substituted with the wzz gene from the second Gram-negative bacterium.22. A composition comprising the Gram-negative bacterium according to any one of paragraphs 15 to 21 and a pharmaceutically acceptable excipient or diluent.23. The Gram-negative bacterium according to any one of paragraphs 15 to 21 or the composition according to paragraph 22 for use in medicine.24. The Gram-negative bacterium according to any one of paragraphs 15 to 21 or the composition according to paragraph 22 for use in the prevention or treatment of a Gram-negative bacterial infection.25. The Gram-negative bacterium or composition for use according to paragraph 24, wherein the Gram-negative bacterial infection is an infection caused by the second Gram-negative bacterium.26. A method of producing antibodies specific for a surface polysaccharide from a Gram-negative bacterium comprising immunising a non-human animal with the Gram-negative bacterium according to any one of paragraphs 15 to 21 , incubating the non-human animal and extracting the antibodies from the non- human animal.27. The method according to paragraph 26, wherein the non-human animal is a rodent.28. The method according to paragraph 27, wherein the rodent is a mouse.29. The method according to any one of paragraphs 26 to 28, wherein immunising comprises introducing the Gram-negative bacterium into the gut of the non-human animal.30. The method according to any one of paragraphs 26 to 29, further comprising screening for antibodies specific for the surface polysaccharide.31. The method according to any one of paragraphs 26 to 30, wherein the antibodies are extracted from secondary lymphoid tissue from the non-human animal.32. The method according to any one of paragraphs 26 to 31 , wherein the non-human animal is maintained for a sufficient amount of time to clear the Gram-negative bacterium from the non-human animal prior to extracting the antibodies.33. The method according to any one of paragraphs 26 to 31 , wherein the non-human animal is maintained for 28 days prior to extracting the antibodies.34. The method according to any one of paragraphs 26 to 33, wherein the surface polysaccharide is an O-antigen.35. The method according to any one of paragraphs 26 to 33, wherein the surface polysaccharide is a CPS.36. Anti-O-antigen antibodies obtained from the method according to paragraphs 26 to 35.37. Anti-CPS antibodies obtained from the method according to paragraphs 26 to 35.38. A method of generating the Gram-negative bacterium according to any one of paragraphs 15 to 21 , comprising substituting the rfb locus from the first Gram-negative bacterium with the rfb locus from the second Gram-negative bacterium.39. A method of generating the Gram-negative bacterium according to any one of paragraphs 15 to 21 , comprising substituting the rfb locus from the first Gram-negative bacterium with the cps locus from the second Gram-negative bacterium.

[0187] The present invention is further illustrated in the following Examples. It should be understood that these Examples, while indicating embodiments of the invention, are given by way of illustration only. From the above discussion and these Examples, one skilled in the art can ascertain the essential characteristics of this invention, and without departing from the spirit and scope thereof, can make various changes and modifications of the invention to adapt it to various usages and conditions. Thus, various modifications of the invention in addition to those shown and described herein will be apparent to those skilled in the art from the foregoing description. Such modifications are also intended to fall within the scope of the appended claims.EXAMPLES - PART AExample 1 - Generation of CitroKPO2a - a strain targeting O2a LPS of Klebsiella pneumoniae

[0188] The genetics of the rfb locus are complex and there are two major systems involved. These are the wzy-dependent pathway and the ABC-transporter pathway; Citrobacter rodentium naturally uses the wzy-dependent pathway. We started by making the most complex substitution possible and replaced the natural Citrobacter rodentium rfb locus with that encoding the Klebsiella pneumoniae O2a O-antigen. This represents an ABC-transporter pathway and is therefore dissimilar to Citrobacter rodentium’s natural wzy- dependent pathway. We demonstrated in vitro that the genetically engineered Citrobacter rodentium strain (CitroKPo2a) now expresses the KP O2a O-antigen. When we subsequently infected mice (gut infection) with this strain we were able to show that mice developed a polyclonal response in their blood that was cross reactive with O2a O-antigen expressed naturally on a KP strain expressing this LPS (KPo2a) (Figure 2A&B). In summary, the genetically engineered CitroKPo2a induced an antibody response in mice following gut infection. Some of this antibody response was directed towards Klebsiella pneumoniae O2a O-antigen indicating the formation of specifically reactive B-cells. These reactive B-cells are the material required to generate monoclonal antibodies. This response was directed to the Klebsiella pneumoniae O2a O-antigen as a proof of principle; our platform could be modified to target other LPS O-antigens (see Example 4).Example 2 - Augmentation of responses with EspO

[0189] We then built in another step into the platform by overexpressing a virulence factor called EspO (a small 10 kDa bacterial protein that is injected into intestinal epithelial cells during infection) in CitroKPo2a. When we overexpressed this EspO we significantly improved the titre of serum antibody that was raised against the KP O2a antigen (Figure 2B).Example 4 - Vaccination with CitroKPO2a protects against subsequent Klebsiella pneumoniae infection due to induced functional antibody responses.

[0190] We then proceeded to demonstrate that the antibodies in this model were protective by first infecting mice with CitroKPo2a in the gut to vaccinate them. Subsequently, we infected mice with KP expressing KP O2a O-antigen, KPo2a, into the lungs (which is the one of the natural sites of infection in mice and humans). Mice developed significantly less severe infection upon KPo2a challenge with less KPo2a in the lungs after 36 hours (Figure 2Ci), importantly, this required EspO augmentation. The best protection was associated with higher levels of antibodies to KP O2a O-antigen prior to infection with KPo2a (Figure 2Cii). This indicates that the antibodies induced by our CitroKPo2a vaccination were potent, effective and protective. This indicates the CitroKPo2a infected mice, especially with EspO boosting, would be a very good source of memory B-cells that produce antibodies that are specific and protective against KP infection.Example 4 - citrOgen can be used to target other priority pathogens.

[0191] In order to demonstrate our ability to target other organisms we also generated a citrOgen strain that expressed the O25b O-antigen from Escherichia coli (CitroECo25b). This Escherichia coli O-antigen is encountered in the highly successful antibiotic resistant ST131 Escherichia coli clade, responsible for widespread extraintestinal infections. ST131 Escherichia coli uses a wzy-dependent LPS pathway and we used this to demonstrate the adaptability of our platform. In the wzy-dependent LPS pathway there is an additional complication where the O-antigen length is determined by an additional gene called wzz. We made two versions of CitroECo25b, one with the wzz from Citrobacter rodentium (CitroECo25b<wzzcR)) and one with the wzz from an Escherichia coli ST131 (CitroECo25b<wzzEcsTi3i)). This adds an additional complication to our platform which can be modified with our genetic techniques.

[0192] When we infected mice with the two strains (CitroECo25b<wzzcR) or CitroECo25b<wzzEcsTi3i) into the gut and allowed mice to recover we observed very high levels of antibody suggesting an even better response than that observed with CitroKPo2a. The CitroECo25b<wzzEcsTi3i) appeared to induce the highest responses. Importantly, this demonstrates the adaptability of our platform for future development.Example 4 - citrOgen works with both O-antigen and CPS

[0193] In the case of CPS, the rfb locus is replaced but instead of substituting O-antigen LPS genes we replace it with CPS encoding genes. Klebsiella pneumoniae K2 CPS is functionally expressed on the surface of CPS based citrOgen strains which has been demonstrated by immunofluorescence microscopy.Examples Part A conclusions

[0194] The present invention solves many issues in LPS O-antigen and CPS vaccine / antibody production. These include: (1) no requirement to harvest LPS and CPS from bacteria with the inherent issues of high economic cost and low efficiency. Genetically-engineered citrOgen bacteria express a chosen LPS or CPStype continuously on their surface (2) no requirement to chemically synthesise O-antigen or CPS with optimisation of chemistry for each selected antigen. Genetically-engineered citrOgen bacteria express a chosen LPS or CPS type continuously on their surface (3) no complex conjugation protocols to carrier proteins to generate glycoconjugate vaccines for robust immune responses. The citrOgen infection itself induces a robust immune response (4) no requirement for long time courses in animal experiments which require multiple repeated dosing (and dose optimisation) of glycoconjugate vaccines as a single experimental infection and recovery is sufficient for robust antibody responses (approximately 28 days) (5) no requirement for adjuvant administration to boost immune responses. citrOgen bacterial infection itself acts as the adjuvant.

[0195] In addition to the above the present invention provides several other advantages above and beyond current techniques: (1) Rapidly adaptable and modifiable - citrOgen can target new LPS O-antigen or CPS types based on genetics alone. This can be used for rapidly emerging infections as genomic information now readily becomes available early in outbreaks (for example the severe Escherichia coli 0104 outbreak in Germany in 2011) (2) Antibody responses can be harnessed to generate diagnostic antibodies (for assays) or research antibodies (3) LPS O-antigen or CPS is presented to the immune system in a physiological fashion (i.e. in the cell membrane) where natural infection by Citrobacter rodentium acts as a strong adjuvant (4) citrOgen requires a single inoculation and stimulation by O-antigen or CPS occurs continuously throughout the infection allowing B-cell and antibody responses to mature. The infected citrOgen bacteria replicate in vivo providing a continuing source of heterologous LPS exposure (5) Combining citrOgen with our expertise in infection modelling we can demonstrate that antibody responses are both elicited AND functional (i.e., protective) in a single experiment (6) It could provide a new live attenuated vaccine platform for human vaccination. It is especially powerful as it is based on a living bacterium. This would mean that the vaccine itself would not require a cold chain and could be delivered orally. This would circumvent many of the issues associated with vaccination (7) We already have completed data supporting its efficacy for KP LPS O-types O1 and 02. These represent some of the most common causes of bacteraemia in critically unwell patients and priority targets for novel treatments.EXAMPLES - PART B

[0196] Citrobacter rodentium is a pathogen of mice and causes infection in the gut. The infection is both robust and prolonged (peaks at day 8-12 post infection, clears from day 21) with most mouse species successfully recovering. It is known that recovery and clearance is mediated, in part, by IgG which is produced by B-cells. It is also known that, following infection, antibodies are produced against the Citrobacter rodentium’s natural LPS O-antigen. The citrOgen platform harnesses this knowledge and uses Citrobacter rodentium as a living vehicle to present the substituted target O-antigen during natural infection of mice. In further modifications of the technology, instead of substituting the LPS O-antigen, we can genetically engineer Citrobacter rodentium to express KP K2 CPS (polysaccharide) or type 3 fimbriae (protein antigens).Example 5 - citrOgen LPS

[0197] LPS is glycolipid composed of lipid A, inner core, outer core and O-antigen. Each region requires a set of genes for production that are encoded in the bacterial genome; the O-antigen is encoded in a region called the rfb locus. The genes in each region encode the proteins required to generate subunits, assemble them and transport them to the outer membrane. Once here LPS carries out its function as a virulence factor and is also surface exposed.

[0198] The genetics of the rfb locus are complex and there are two major systems involved. These are the wzy-dependent pathway and the ABC-transporter pathway; Citrobacter rodentium naturally uses the wzy- dependent pathway, KP the ABC-transporter pathway. We replaced the genes in the Citrobacter rodentium rfb with those from the KP rfb encoding KP 02 LPS. Adding even more complexity, KP uses two genes (wbbY and wbbZ) to convert the product of the KP rfb locus O-antigen from 02 to O1 . We also placed these genes into the genome of Citrobacter rodentium. This resulted in a strain called CRKPOI .

[0199] All citrOgen experiments follow a similar protocol detailed in Figure 5A. The citrOgen strains are administered to mice at the start of the experiment. Each group of mice only receives one strain. Here in the KPO1 experiments we infected with Citrobacter rodentium wild-type (CRwr) as a control which does not express any heterologous antigen, CRKPOI (expressing KP O1 LPS) and CRKPOI pespO (expressing KP O1 LPS and overexpressing the Citrobacter rodentium effector EspO). / . EspO boosting (gut colonisation and antibody responses)

[0200] Citrobacter rodentium naturally reaches peak infection at days 8-12 post inoculation. We tested if heterologous KP O1 expression affected colonisation, which it did with significantly reduced stool CRKPOI count at day 8 post infection (Figure 5B). However, when we overexpressed EspO in CRKPOI pespO the colonisation was returned to normal with no difference between the levels achieved in the stool compared to CRwr. This was found empirically and is not obvious. However, it is key to the invention as EspO overexpression is required for gut colonisation at day 8. This data is complete and robust.

[0201] The above experiments demonstrate and prove that EspO boosting is required for citrOgen. / / . KP 01 specific antibody responses

[0202] We took a blood sample afterthe Citrobacter rodentium gut infections had cleared (Figure 5A). We conducted an ELISA against heat-killed KP. Serum taken from mice infected with CRKPOI and CRKPOI pespO contained antibodies to KP that were absent in mice infected with CRwr (Figure 5C). The highest median antibodies were observed in mice infected with CRKPOI pespO (Figure 5C).

[0203] The above experiments prove that we can substitute the endogenous Citrobacter rodentium LPS antigens with those exported in a ABC-dependent fashion. Moreover, mice infected with CRKPOI pespO raise specific antibodies against KP O1 .Hi. KP 01 specific antibody responses are protective

[0204] Mice were then challenged with KP into the lungs to assess if the antibody responses identified (Figure 5C) would result in functional protection from KP disease.

[0205] After 72 hours mice were weighed to assess weight loss; blood was collected under anaesthesia from the heart then lungs removed post mortem for analysis.

[0206] We used the sera after infection to prove that the antibody response to KP was mediated by binding to KPO1 LPS (Figure 5D). To do this we checked antibody binding (by flow cytometry) to either KPwr (expressing KP O1 LPS) or KPartb (a KP mutant we made that expresses no LPS). Mice infected with CRwr had no detectable binding to either KPwr or KPartb. Mice infected with CRKPOI and CRKPOI pespO had antibodies that bound only KPwr This demonstrated that the IgG response against KP in Figure 5C was mediated by KP O1 LPS.

[0207] We confirmed this specific binding by using immunofluorescent staining of KPwr bacteria. Sera from CRwr had no detectable binding to KPwr (Figure 6A). We then used a reference antibody (known to bind to KP O1 LPS) and observed a circumferential ring of binding to around bacteria consistent with the location of LPS (Figure 6B). We then used sera from mice infected with CRKPOI and observed staining identical to that of the reference KP O1 antibody (Figure 6C).

[0208] We then assessed the protective effect using important metrics of infection:

[0209] Weight loss measures (Figure 7A) showed that mice pre-infected with CRwr lost significantly more weight that those that received CRKPOI and CRKPOI pespO.

[0210] Mice have a normal blood glucose between 4-8 mmol / L (Figure 7B). Severe infection in mice is known to result in hypoglycaemia. CRwr pre-infected mice challenged with KP had significantly lower levels of blood glucose after KP infection than those observed before infection. CRKPOI pre-infected mice were relatively protected from hypoglycaemia but still had blood glucose levels significantly reduced compared to baseline. However, CRKPOI pespO pre-infected animals did not suffer any significant changes in blood glucose.

[0211] CRKPOI and CRKPOI pespO pre-infected animals had significantly reduced levels of KP at 72 hours post-infection in the lungs and the blood compared to CRwr pre-infected animals (Figure 7C&D).

[0212] KPwr infection in mice results in a reduction in total blood white cell count (Figure 7E) and lymphocytes (Figure 7F). Neutrophils typically increase with most infections (Figure 7G) and platelets become lower with the severity of infection (Figure 7H). CRKPOI and CRKPOI pespO pre-infected animals had significantly improved total white cells counted compared to CRwr pre-infected animals and were protected from lymphopaenia. No changes were observed in the neutrophil counts between any groups. Both CRKPOI and CRKPOI pespO pre-infected animals were protected from severe thrombocytopaenia as compared to CRwr pre-infected animals.

[0213] We examined the acute phase response which is characterised by an increase in serum levels of the inflammatory marker C-reactive protein and decrease in the negative acute phase protein albumin. C- reactive protein levels were high in mice pre-infected with CRwr and significantly lower in those pre-infected with CRKPOI and CRKPOI pespO following KPwr challenge (Figure 71). In the case of albumin only CRKPOI pespO significantly protected against the hypoalbuminaemia induced by KPwr infection (Figure 7J).

[0214] Organ failure is hall mark of sepsis. KP is a common cause of sepsis in humans. We looked at creatinine (a marker of renal failure). Both CRKPOI and CRKPOI pespO pre-infected animals had significantly lower levels of serum creatinine compared to CRwr pre-infected animals (Figure 7K). Infection in the lungs causes direct tissue damage and lung injury. To assess lung injury we measured the serum levels of a protein which is usually restricted to the alveolar space called surfactant protein D. Lung damage by KP disrupts the alveolar capillary membrane, allowing this protein to leak into the blood. In CRwr pre-infected animals we observed high levels of this protein in the serum which were significantly reduced by CRKPOI and CRKPOI pespO pre-infection (Figure 7L).

[0215] The above experiments demonstrate and prove that the antibody responses demonstrating in 1 b are functionally protective. iv. Escherichia coli 025b specific responses and wzz (chain length modification).

[0216] We wanted to ensure that this disclosure covered all types of LPS transporter systems and went on to generate a series of strains encoding Escherichia coli O25b, which was taken from the genome of an ST131 Escherichia coli strain. This LPS type is exported in a wzy-dependent fashion. In this system an additional gene called wzz is used to determine the length of the O-antigen.

[0217] We started by replacing the rfb locus with the genes encoding ^^25b. This strain (CRECO25PWZZCR) retains the natural Citrobacter rodentium wzz gene. We made two further modifications by replacing the wzz gene with the version from the ST131 Escherichia coli strain (CREco25bwzzsTi3i) and then also inserting a gene called fepE which results in very long LPS (CRECO25PWZZSTI3I +fePE).

[0218] We extracted LPS from these strains and looked at its size by polyacrylamide gel electrophoresis and silver staining (Figure 9A). Substituting the O25b locus into CRECO25PWZZCR (and keeping the native Citrobacter rodentium wzz gene) increased the length of the LPS compared to Escherichia coli ST131. When we swapped the wzz gene to that from the Escherichia coli ST131 gene CRECO25PWZZSTI3I the LPS was shortened and phenotypically copied that of the Escherichia coli ST131 O25b donor. When we added fepE to this strain CRECO25PWZZSTI3I +fePE the LPS became very long. Some of these changes in LPS length have been demonstrated previously to favourably change the immune response to vaccination against O25b-based vaccines.

[0219] We then infected mice and took blood after clearance of Citrobacter rodentium as per the schema in Figure 5A. As we had demonstrated previously that EspO expression was required to maintain colonisation in the context of heterologous KPO1 expression we overexpressed EspO in all these strains.

[0220] At 8 days post Citrobacter rodentium infection CREco25bwzzcR, CREco25bwzzsTi3i and CREco25bwzzsTi3i +fePEwith EspO overexpression all colonised well with no colonisation defect compared to CRwr (Figure 9B).

[0221] We took sera from the blood sample collected after Citrobacter rodentium clearance and conducted an ELISA against heat-killed ST131 Escherichia coli strain (which expresses O25b LPS, Figure 9C). Serum taken from mice infected with CREco25bwzzcR, CREco25bwzzsTi3i and CREco25bwzzsTi3i +fepE contained antibodies to ST131 Escherichia coli. Whilst no difference in antibody titre was observed between the strains there was significantly higher antibody binding in sera from sera in these groups to that from mice pre-infected with CRwr.

[0222] The above experiments demonstrate and prove that we can substitute and raise specific antibody responses to heterologous LPS antigens exported in a wzy-dependent fashion.

[0223] The above experiments demonstrate that we can change the chain length with functional alterations in the specific antibody response.Example 6 - citrOgen CPS

[0224] Capsular polysaccharide is a extracellular polysaccharide produced as a virulence factor by many pathogenic bacteria that survive in blood. The capsule locus of genes encodes the genes for synthesis and export of this to the cell surface. It is highly antigenic and plays a role in shielding the bacteria from the host immune response. KP K2 capsular antibodies and K2 vaccines have been demonstrated to result in protection from KP infection in mouse models. Our KPwr expresses K2 capsule.

[0225] We took the genes from KPwrthat express K2 CPS (the cps locus) and inserted them into the CRwr genome. We replaced the genes in CRwr that usually synthesise colanic acid. This generated a citrOgen CPS strain called CRKPK2.

[0226] We then infected mice and took blood after clearance of Citrobacter rodentium as per the schema in Figure 5A. As we had demonstrated previously that EspO expression was required to maintain colonisation in the context of heterologous KPO1 expression we overexpressed EspO in CRKPK2. / . KP K2 specific antibody responses

[0227] At 8 days post Citrobacter rodentium infection CRKPK2 overexpressing EspO colonised well with no colonisation defect compared to CRwr (Figure 11 A).

[0228] We took sera from the blood sample collected after Citrobacter rodentium clearance and conducted an ELISA against heat killed KPwr (expressing K2 CPS) and KPwcaj (a mutant we generated that expresses no K2 CPS). CRWT (absent heterologous K2 expression) and CRKPK2 demonstrated no discernible binding to KPwcaj (Figure 11 B). Sera from mice pre-infected with CRKPK2 overexpressing EspO demonstrated binding to KPwr, which was absent in the sera from mice pre-infected with CRwr Taken together this demonstrates that antibodies are raised to heterologous K2 in citrOgen CPS and that the binding is specific to K2.

[0229] We then confirmed this specific binding by using immunofluorescent staining of KPwr bacteria. Sera from CRwr had no detectable binding to KPwr (Figure 12A). We then used a reference antibody (known to bind to KP K2 CPS) and observed a circumferential ‘spiky’ ring of binding to around bacteria consistent with the location and conformation of CPS (Figure 12B). We then used sera from mice infected with CRKPK2and observed staining identical to that of the reference KP K2 antibody (Figure 12C).

[0230] The above experiments demonstrate and prove that we can insert and raise specific antibody responses to heterologous CPS antigens.

[0231] The above experiments demonstrate the specificity and applicability of the citrOgen K2 response.Example 7 - citrOgen protein

[0232] Protein antigens are the basis of many vaccines and the targets of many mAbs. Type 3 fimbriae are a multicomponent protein complex that assembles on the KP cell membrane. They are known to play a role in binding to abiotic surfaces and biofilm formation. mAbs and vaccines using one component of the type 3 fimbriae (MrkA) have been shown to be protective in models of KP infection. In addition, antibodies to MrkA have been found in both target-agnostic and specific antibody discovery campaigns.

[0233] We took the genes from KPwr that synthesise and export the structural components of KP type 3 fimbriae (mrkABCDF [mrkA-F]) and inserted them into the CRwr genome. We placed these at the 3’ glmS site, which is not a novel insertion site in bacteria. However, we added one complication and used the promoter for the Citrobacter rodentium type III secretion system effector Map to drive expression. This is novel. The strain that we generated was called CRKPmrkA-F. / . KP type 3 specific antibody responses

[0234] We then infected mice and took blood after clearance of Citrobacter rodentium as per the schema in Figure 5A. As we had demonstrated previously that EspO expression was required to maintain colonisation in the context of heterologous KPO1 expression we overexpressed EspO in all these strains.

[0235] At 8 days post Citrobacter rodentium infection CRKPmrkA-F overexpressing EspO colonised well with no colonisation defect compared to CRwr (Figure 14A).

[0236] We took sera from the blood sample collected after Citrobacter rodentium clearance and conducted an ELISA against heat killed KPwr (expressing normal levels of type 3 fimbriae) and two mutants. The first mutant had the genes that encode the type 3 fimbriae deleted (KPamrkA-F) and the second had a stop codon in a gene called mrkJ (KPmrkj*)- As MrkJ is a negative regulator of type 3 fimbriae expression, the KPmrkj* mutant overexpresses type 3 fimbriae. CRwr and CRKPmrkA-F sera did not bind to KPamrkA-F (Figure 14B). CRKPmrkA-F sera bound to KPwr, with the highest binding observed in binding to KPmrkj*. CRwr sera did not bind to either KPwr or KPmrkj*. This indicates specific binding to type 3 fimbriae.

[0237] We then confirmed this specific binding by using immunofluorescent staining of KPwr bacteria. Sera from CRwr had no detectable binding to KPwr (Figure 15A). We then used a reference antibody (known to bind to KP MrkA protein) and observed a circumferential spikes of binding around bacteria consistent with the location and conformation of type 3 fimbriae (Figure 15B). We then used sera from mice infected with CRKPK2 and observed staining identical to that of the reference KP MrkA antibody (Figure 15C).

[0238] The above experiments demonstrate and prove that we can insert and raise specific antibody responses to heterologous protein antigens.Examples Part A conclusions

[0239] The present invention therefore provides a system that harnesses natural infection of the mouse host with the specific mouse pathogen Citrobacter rodentium to generate antibodies. The advantages of the invention are summarised in the table below.

[0240] The present invention therefore overcomes bottlenecks for generating new mAbs. In the case of the LPS O-antigen mAb generation, glycoconjugate immunogens was previously considered by some companies to be the greatest bottleneck.

[0241] citrOgen is also very economical: the reagents to genetically modify strains are modest in price compared to large scale LPS or CPS harvesting or chemical synthesis.

[0242] The impact of a rapidly modifiable immunogen-expression platform like citrOgen is that it allows many LPS O-antigen types, capsular polysaccharides or protein antigens of interest to be generated in a short period of time. In addition, once generated, there is an endless supply of immunogen, as the platform is based on live bacteria they can simply be grown up in liquid culture and does not require expensive and complex equipment.Equivalents and scope

[0243] Preferred features for the second and subsequent aspects of the invention are as for the first aspect of the invention mutatis mutandis. It will be appreciated that all embodiments described herein are considered to be broadly applicable and combinable with any and all other consistent embodiments, as appropriate. Such combinations are considered to fall within the scope of the present invention.

[0244] Those skilled in the art will appreciate that the present invention is defined by the appended embodiments and not by the Examples or other description of certain embodiments included herein.

[0245] Similarly, the singular forms “a”, “an”, and “the” include plural referents unless the context clearly dictates otherwise.

[0246] Unless defined otherwise above, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. Any methods and materials similar or equivalent to those described herein can also be used in the practice or testing of the present invention. Generally, nomenclatures used in connection with, and techniques of, cell and tissue culture, molecular biology, immunology, genetics and protein and nucleic acid chemistry described herein are those well-known and commonly used in the art, or according to manufacturer's specifications.

[0247] All publications, patents and patent applications mentioned in this specification are herein incorporated in their entirety by reference into the specification, to the same extent as if each individual publication, patent or patent application was specifically and individually indicated to be incorporated herein by reference. In addition, citation or identification of any reference in this application shall not be construed as an admission that such reference is available as prior art to the present invention. To the extent that section headings are used, they should not be construed as necessarily limiting.

Claims

Claims1. A Citrobacter rodentium strain which expresses higher levels of EspO, or a homologue thereof, compared to a control.

2. The Citrobacter rodentium strain according to claim 1 , comprising one or more heterologous molecules present on the cell surface.

3. The Citrobacter rodentium strain according to claim 2, wherein the one or more heterologous molecules comprise lipopolysaccharide (LPS).

4. The Citrobacter rodentium strain according to claim 2, wherein the one or more heterologous molecules comprise capsular polysaccharide (CPS).

5. The Citrobacter rodentium strain according to claim 2, wherein the one or more heterologous molecules comprises a cell-surface protein or protein complex.

6. The Citrobacter rodentium strain according to claim 5, wherein the cell-surface protein or protein complex comprises type 3 fimbriae.

7. The Citrobacter rodentium strain according to any preceding claim, wherein the control is a wild type Citrobacter rodentium strain.

8. The Citrobacter rodentium strain according to claim 7, wherein the wild type Citrobacter rodentium strain is ICC168 or DBS100.

9. The Citrobacter rodentium strain according to any one of claims 2 to 8, wherein the one or more heterologous molecules are from a different Gram-negative bacterium.

10. The Citrobacter rodentium strain according to claim 9, wherein the different Gram-negative bacterium is a pathogenic Gram-negative bacterium.

11. The Citrobacter rodentium strain according to claim 9 or 10, wherein the different Gram-negative bacterium is from a genus selected from the group consisting of Serratia, Haemophilus, Moraxella, Salmonella, Shigella, Neisseria, Klebsiella, Vibrio, Pseudomonas, Escherichia, Helicobacter and Legionella.

12. The Citrobacter rodentium strain according to claim 11 , wherein the different Gram-negative bacterium is Serratia marcescens, Haemophilus influenzae type B, Moraxella catarrhalis, Salmonella enterica, Shigella dysenteriae, Shigella flexneri, Shigella boydii, Shigella sonnei, Klebsiella pneumoniae, Klebsiella aerogenes, Klebsiella oxytoca, Klebsiella variicola, Klebsiella quasipneumoniae, Klebsiella michiganensis, Vibrio cholerae, Vibrio parahaemolyticus,, Pseudomonas aeruginosa, Neisseria meningitidis, Escherichia coll, Helicobacter pylori or Legionella pneumophila.

13. The Citrobacter rodentium strain according to claim 12, wherein the different Gram-negative bacterium is Escherichia coll.

14. The Citrobacter rodentium strain according to claim 13, wherein the different Gram-negative bacterium is Escherichia coll sequence type ST131 .

15. The Citrobacter rodentium strain according to claim 12, wherein the different Gram-negative bacterium is Klebsiella pneumoniae.

16. The Citrobacter rodentium strain according to any one of claims 3 to 15, wherein the O-antigen of the LPS is selected from the group consisting of: Klebsiella pneumoniae O1 v1 , O1 v2, O1 v3, O2a, O2afg, O2ac, 03, O3a, O3b, 04, 05, 012, 013, OL102, OL103 and OL104.

17. The Citrobacter rodentium strain according to claim 16, wherein the O-antigen is O1v1 .

18. The Citrobacter rodentium strain according to claims 13 or 14, wherein the one or more heterologous molecules comprise LPS, wherein the O-antigen of the LPS is O25b.

19. The Citrobacter rodentium strain according to claim 15, wherein the one or more heterologous molecules comprise CPS, wherein the CPS is K2.

20. The Citrobacter rodentium strain according to claim 9 to 19, wherein the rfb locus has been substituted with a rfb locus from the different Gram-negative bacterium.

21. The Citrobacter rodentium strain according to claim 19, wherein the colanic acid biosynthesis locus of the Citrobacter rodentium strain has been substituted with a cps locus from the different Gramnegative bacterium.

22. The Citrobacter rodentium strain according to claim 9 to 20, further comprising the cps locus from the different Gram-negative bacterium.

23. The Citrobacter rodentium strain according to any preceding claim further comprising the genes wbbY and wbbZ.

24. The Citrobacter rodentium strain according to any one of claims 9 to 23, wherein the wzz gene of the Citrobacter rodentium strain has been substituted with the wzz gene from the different Gramnegative bacterium.

25. The Citrobacter rodentium strain according to claim 24, wherein the wzz gene of Citrobacter rodentium is wzzB, or a gene with at least 50% sequence identity to wzzB.

26. The Citrobacter rodentium strain according to any preceding claim, further comprising the gene fepE.

27. The Citrobacter rodentium strain according to any one of claims 9 to 26, further comprising the mrk operon genes mrkA, mrkB, mrkC, mrkD and mrkF of the different Gram-negative bacterium.

28. The Citrobacter rodentium strain according to claim 27, wherein the mrk operon genes are inserted into the Tn7 site of Citrobacter rodentium.

29. The Citrobacter rodentium strain according to claims 27 or 28, wherein the mrk operon genes are from Klebsiella pneumoniae.

30. The Citrobacter rodentium strain according to any preceding claim, wherein one or more genes are under control of the / er-regulon.

31. The Citrobacter rodentium strain according to claim 30, wherein one or more of the genes selected from the group consisting of wbb Y, wbbZ, wzz, fepE, mrkA, mrkB, mrkC, mrkD and mrkF are under control of the / er-regulon.

32. The Citrobacter rodentium strain according to claim 31 , wherein the genes mrkA, mrkB, mrkC, mrkD and mrkF are under control of the / er-regulon.

33. A composition comprising the Citrobacter rodentium strain according to any preceding claim and a pharmaceutically acceptable excipient or diluent.

34. The Citrobacter rodentium strain according to any one of claims 1 to 32 or the composition according to claim 33 for use in medicine.

35. The Citrobacter rodentium strain according to any one of claims 1 to 32 or the composition according to claim 33 for use in the prevention or treatment of a Gram-negative bacterial infection.

36. A method of producing antibodies comprising infecting a non-human animal with the Citrobacter rodentium strain according to any one of claims 1 to 32 or the composition according to claim 33,comprising incubating the non-human animal and extracting the antibodies from the non-human animal.

37. A method of producing a hybridoma comprising infecting a non-human animal with the Citrobacter rodentium strain according to any one of claims 1 to 32 or the composition according to claim 33, comprising incubating the non-human animal, extracting B cells from the non-human animal and using the B cells to produce the hybridoma.

38. The method according to claims 37 or 37, wherein the non-human animal is a rodent.

39. The method according to claim 38, wherein the rodent is a mouse.

40. The method according to any one of claims 37 to 39, wherein infection comprises introducing theCitrobacter rodentium strain into the gut of the non-human animal.

41. The method according to any one of claims 36 to 40, further comprising screening for the antibodies or the B cells.

42. The method according to any one of claims 36 to 41 , wherein the antibodies or B cells are extracted from secondary lymphoid tissue from the non-human animal.

43. The method according to any one of claims 36 to 42, wherein the antibodies are extracted from the blood of the non-human animal.

44. The method according to any one of claims 36 to 44, wherein the non-human animal is maintained for a sufficient amount of time to clear the Citrobacter rodentium strain from the non-human animal prior to extracting the antibodies or B cells.

45. The method according to any one of claims 36 to 45, wherein the non-human animal is maintained for 28 days prior to extracting the antibodies or B cells.

46. The method according to any one of claims 36 to 46, wherein the antibodies or antibodies produced by the B cells are specific for LPS.

47. The method according to any one of claims 36 to 47, wherein the antibodies or antibodies produced by the B cells are specific for CPS.

48. The method according to any one of claims 36 to 47, wherein the antibodies or antibodies produced by the B cells are specific for type 3 fimbriae.

49. The method according to any one of claims 36 to 48, wherein the method further comprises infecting the non-human animal with the different Gram-negative bacterium and screening the non- human animal for a immunogenic response against the different Gram-negative bacterium.

50. The method according to claim 49, wherein the different Gram-negative bacterium is Klebsiella pneumoniae and infecting the non-human animal with Klebsiella pneumoniae is infecting the lung of the non-human animal.

51. The method according to any one of claims 36 to 49, wherein infection with the Citrobacter rodentium strain results in higher titres of antibodies specific for the one or more heterologous molecules present on the cell surface in the non-human animal compared to a control.

52. The method according to claim 51 , wherein the control is Citrobacter rodentium strain ICC168 or DBS100.

53. A method of producing the Citrobacter rodentium strain according to any one of claims 27 to 35, comprising providing a Citrobacter rodentium strain and inserting the mrk operon genes of the different Gram-negative bacterium into the Citrobacter rodentium strain.

54. A method of producing the Citrobacter rodentium strain according to any one of claims 20 to 35, comprising providing a Citrobacter rodentium strain and substituting the rfb locus of the Citrobacter rodentium strain with a rfb locus from the different Gram-negative bacterium.

55. A method of producing the Citrobacter rodentium strain according to any one of claims 21 to 35, comprising providing a Citrobacter rodentium strain and substituting the colanic acid biosynthesis locus of the Citrobacter rodentium strain with a cps locus from the different Gram-negative bacterium.

56. The method according to any one of claims 53 to 55, wherein the provided Citrobacter rodentium strain is strain ICC168 or DBS O.

57. The method according to any one of claims 53 to 56, wherein the different Gram-negative bacterium is Klebsiella pneumoniae.