BACKGROUND OF THE INVENTION The present invention relates to the sequence of genes from the E. canis bacterium, and the development of a vaccine against this organism.

Ehrlichia canis (E. canis) is a small gram-negative, obligately intracytoplasmic rickettsia. This bacteria is the agent which causes canine monocytic ehrlichiosis (CME), a tick-borne disease which predominantly affects dogs. The most common carrier of E. canis is the brown dog tick Rhipicephalus sanguineus. The disease was described originally in Algeria in 1935. It was subsequently recognized in the United States in 1962, but is now known throughout much of the world. Canine monocytic ehrlichiosis caused much concern during the Vietnam War, when 160 military dogs died from the E. canis infection. There is no vaccination currently available against E. canis. It is a life threatening disease that continues to be an important health concern for veterinarians and pet owners alike.

Canine monocytic ehrlichiosis is an infectious blood disease. A reduction in cellular blood elements is the primary characteristic of the disease. E. canis lives and reproduces in the white blood cells (leukocytes). It eventually affects the entire lymphatic system, and devastates multiple organs. By targeting the white blood cells, these cells die off rapidly. These dead blood cells migrate primarily to the spleen, which enlarges as a result. The bone marrow recognizes the loss of the white blood cells and works to form new, healthy cells. It sends out the cells prematurely, and these immature cells do not work properly. Often, these immature cells mimic those in leukemic patients, so the disease is misdiagnosed as leukemia. Canine monocytic ehrlichiosis may also predispose dogs to various cancers.

There are three stages of canine monocytic ehrlichiosis. The first, acute stage mimics a mild viral infection. During the acute stage, most, if not all, of the damage is reversible and the animal is likely to recover. This is the stage where treatment is the most effective, stressing the need for early detection. Without treatment, however, the animal will progress into a subclinical (second) stage and/or to the chronic (final) stage. When the animal has reached the chronic stage, the bacterial organism has settled within the bone marrow. Many dogs in this stage suffer massive internal hemorrhage, or develop lethal complications such as sudden stroke, heart attack, renal failure, splenic rupture or liver failure.

E. canis can be cultured in vitro in a mammalian-derived cell line (DH82).

Continued maintenance of these cells is difficult because the cell culture must be supplemented with primary monocytes (white blood cells found in bone marrow) every two weeks. The cultures are very slow growing, and the culture media is expensive.

Data concerning the genes in the E. canis genome has concentrated primarily on the 16S rRNA gene. Previous work has sequenced this gene, which is a ubiquitous component of the members of the ehrlichia family, as well as the majority of organisms worldwide. The high sequence homology between this gene throughout the living world and its poor immunogenicity makes it an unsuitable candidate for vaccine development. It is necessary to find other genes within this genome if hope for a vaccine against this deadly disease can ever be realized.

Sequencing of the 16S rRNA gene indicates that E. canis is closely related (98. 2% homology) to E. chaffeensis, the novel etiologic agent of human ehrlichiosis.

Western blots of E. canis are similar when probed with antisera to E. canis, E. chaffeensis, and E. ewingi (another cause of human ehrlichiosis) indicating a close antigenic relationship between these three species (Chen et al., 1994).

The indirect fluorescent antibody test (IFA) has been developed for detecting canine monocytic ehrlichiosis. IFA detects the presence of antibodies against the invading organism in a dog's blood. Unfortunately, this test is not always accurate. Sometimes, dogs will test negative in the acute phase because their immune system is delayed in forming antibodies. Another false negative may occur if there is a low titer in the chronic stage. An additional drawback of this test is the cross-reactivity found. The anti E. canis polyclonal antibody positively reacts with E. chaffeensis, undermining the specificity of the test. An alternative test, the Giesma smear, has been used to locate the actual organism in a dog's blood. Unfortunately, despite appropriate staining techniques and intensive film examination, the organisms frequently can not be located. The fallibility of these tests makes it essential to provide better diagnostic tools for this disease.

Due to difficulties in the detection of a tick bite, early diagnosis of infection, the suppression of host defenses and the nature of persistent infection of the disease, an effective vaccine against E. canis is urgently needed for dogs.

SUMMARY OF THE INVENTION This invention discloses novel sequence data for E. canis genes. Specifically, a clone has been identified and sequenced. Four proteins termed ProA, ProB, mmpA (for morula membrane protein, which is an ORF), and a cytochrome oxidase homolog, have been identified within this clone. In addition, a partial gene encoding a lipoprotein signal peptidase homolog has been discovered.

An embodiment of this invention includes the creation of a vaccine with this sequence and protein information. The proteins disclosed in this invention are extremely antigenic. Therefore, they have the potential to be extremely useful as a vaccine. The types of vaccine made available by this novel technology include a DNA vaccine, a recombinant vaccine, and a T cell epitope vaccine.

BRIEF DESCRIPTION OF THE DRAWINGS Fig. 1 shows the three clones identified in the library screen.

DESCRIPTION OF THE PREFERRED EMBODIMENT E. canis causes a devastating canine disease. Currently, there is no vaccine available to prevent this disease. This invention provides the tools necessary to develop such a vaccine. More specifically, four genes have been identified from a genomic fragment of E. canis, named ProA, ProB, mmpA, and a cytochrome oxidase homolog. In addition, a partial gene coding for a lipoprotein signal peptidase homolog has been found.

Any of these proteins can be utilized in an embodiment of this invention to develop a vaccine.

Screening an E. canis library To identify genes in the E. canis genome, a genomic DNA expression library was constructed. An E. canis strain isolated from dogs with canine ehrlichiosis was grown in the dog cell line DH82 by a technique being known in the art, and incorporated by reference (Dawson et al., 1991; Rikihisa, 1992). The cells were harvested and the chromosomal DNA extracted as described by a technique known in the art (Chang et al., 1987; Chang et al., 1989a; Chang et al., 1989b; Chang et al., 1993a; Chang et al., 1993b).

To construct the library, 200 pg of DNA was partially digested with Sau3A. DNA fragments from 3 to 8 kb were isolated and ligated to a plasmid, pHG165 (Stewart et al., 1986). The plasmids were transformed into E. coli TB1 (Chang et al., 1987).

The library was screened with polyclonal antibodies against E. canis. Polyclonal antibodies were generated from dogs that had been bitten by a tick harboring E. canis.

The polyclonal antibodies were preabsorbed with the lysate of an E. coli host strain. The library was plated on petri plates at a density of 1,000 colony forming units. Colonies were transferred to nitrocellulose and each filter was probed with 1 ml of the preabsorbed polyclonal antibodies. Positive colonies were identified with a second antibody consisting of an alkaline phosphatase-conjugated goat anti-rabbit IgG (Kirkegaard and Perry Laboratories, Gaithersburg, MD), followed by color development with a substrate solution containing nitroblue tetrazolium (NBT) and 5-bromo-4-chloro-3-indolyl phosphate (BCIP). Positive clones were rescreened three times.

Three clones were isolated from this screening procedure, pCH2,4, and 5 (Figure 1). Within these three positive clones, an overlapping reading frame was found that codes for mmpA. The longest genomic fragment (pCH4) encodes four complete genes and one partial gene. It completely encodes the proteins ProA, ProB, mmpA, and a cytochrome oxidase homolog, as well as containing the partial sequence of a lipoprotein signal peptidase homolog. ProA and ProB are located on a single operon. Restriction endonuclease digestion mapping and DNA sequencing were done by techniques known in the art, and incorporated by reference (Chang et. al., 1987; Chang et. al., 1989a ; Chang et. al., 1989b; Chang et. al., 1993a; Chang et. al., 1993b). Briefly, the DNA sequence was determined by automated DNA sequencing on the ABI PRISM Model 377 DNA system.

The complete nucleotide sequences were determined on both strands by primer walking.

The thermal cycling of the sequencing reactions utilized the Taq DyeDeoxy Terminator Cycle sequencing kit. Databases were searched for homologous proteins through the use of the BLAST network service of the National Center for Biotechnology Information (NCBI) (Althchul et al., 1990; Gish et al., 1993).

Sequence Information The E. canis genes were sequenced. The cloned fragment contains 5,299 nucleotides, and codes for four proteins. There is also one partial gene at the carboxy terminus. SEQ. ID. NO. 1 is the entire nucleotide sequence. SEQ. ID. NO. 2 and 3 are the translation of nucleotides 12 through 533 from SEQ. ID. NO. 1 and code for a cytochrome oxidase homolog, which was deemed ec3 and encodes a product of 174 amino acids.

Cytochrome oxidase is important in virulence, and therefore is a strong candidate for use in a vaccine. SEQ. ID. NO. 4 and 5 are the translation of nucleotides 939 through 2,252 from SEQ. ID. NO. 1 and code for ProA. ProA begins 402 nucleotides downstream from the end of ec3 and encodes a product of 438 amino acids. The ProA protein shares extensive characteristics with the putrilysin family, which is a subset of metallopeptidases.

The highest homology is seen in the N-terminal portion, which includes conserved putative metal binding sequence, His-X-X-Glu-His as well as conserved Glu for catalysis.

ProA was also shown to share 27% of its characteristics over 284 amino acids with MPP-ß (mitochondrial processing peptidase) in Solanum tuberosum (Genbank accession number X80237) and 27% of its characteristics over 222 amino acids with E. Coli pitrilysin (Genbank accession number M17095). Many bacterial zinc proteases are also found to share homology with ProA protein, such as PqqF in M. extrorquens AM1 where 36% of 376 amino acids are shared (Genbank accession number L43135), Y4wA in Rhizobium sp. where 34% of 416 amino acids are shared (Genbank accession number AE000103), 27% of 419 amino acids of the putative gene product PA0372 in Pseudomonas aeruginosa (Genbank accession number AE004475), and 26% of 389 amino acids of the putative gene product of HP 1012 in Helicobacter pylori, strain 26695 are shared (Genbank accession number AE000609). SEQ. ID. NO. 6 and 7 are the translation of nucleotides 2,258 through 3,610 from SEQ. ID. NO. 1 and code for ProB. ProB begins 6 nucleotides after the end of proA and encodes a protein of 451 amino acids long. Through Blast analysis it was shown that ProB does not contain the conserved Zinc-binding domain. ProB was shown to share homology with both subunits of the MPP subfamily and some bacterial putative zinc proteases such as PqqF, Y4wA, and the gene products of PA0372. ProB also shared homology with bacterial proteases that do not contain Zinc-ligand motif, but show similarities with the pitrilysin family, such as PqqG, Y4wB, gene product of PA0371, and the gene product of HP0657 in M. extrorquens AM1, Rhizobium sp., P. aeruginosa, and H. pylori 26695. Preliminary evidence indicates that ProA and ProB are proteases. SEQ.

ID. NO. 8 and 9 are the translation of nucleotides 4,132 through 4,794 from SEQ. ID. NO.

1 and code for ORF, designated mmpA. The polypeptide that is generated consists of 221 amino acids and does not have a significant identity to any proteins found in existing databases. SEQ. ID. NO. 10 and 11 are the translation of the complementary sequence of nucleotides 4,883 through 5,299 from SEQ. ID. NO. 1 and code for the partial sequence of a lipoprotein signal peptidase homolog. Lipoprotein signal peptidases are membrane proteins, and by nature may be less desirable for vaccine development. However, this protein is still worth pursuing in the creation of a vaccine.

Structure and Expression of ProA, ProB, mmpA, cytochrome oxidase and the lipoprotein signal Structurally, MmpA is extremely hydrophobic with five transmembrane segments and three potential antigenic determinants, which are protein kinase C phosphorylation sites located in position 7 (Ser), 37 (Ser), and 213 (Ser) and one possible casein kinase 1I phosphorylation site in position 177 (Thr), where the above determinants were characterized by PROSITE and the PCgene programs. MmpA was localized primarily in the morula membrane. Since MmpA contains three potential phosphorylation sites, this indicates the possibility of a similar communication system as seen by phosphorylation of one of the chlamydial inclusion membrane proteins, IncA. E. canis grown in vitro (DH82 cells) expressed MmpA, furthermore, sera obtained from dogs that were naturally infected and experimentally infected with E. canis recognized MmpA, which confirms that in vivo and in vitro expression of MmpA as well as the antigenicity of MmpA. The above two results indicate that MmpA is capable of stimulating an immune response, which is necessary for a vaccine to be effective. However, E. canis is an intracellular organism, cell mediated immunity is more important in protecting the dog against this type of infection then humoral immunity and it may be possible to direct these antigens toward a predominant Thl response using an appropriate adjuvant. The mmpA gene was found in E. canis and E. chaffeenis but was not present in the HGE agent. However, the MmpA protein was not expressed by E. clzaffeenis on a western blot. E. canis with MmpA caused cells to lyse, indicating the presence of MmpA protein, where E. chaffeenis with MmpA did not lyse. This result lends to the conclusion that the MmpA protein may be useful for serodiagnosis in differentiating E. canis and E. chaffeenis. Furthermore, MmpA, ProA, and ProB proteins can be used as antigens in ELISA or Western blot analysis to perform a diagnosis of an E. canis infection in animals.

Structurally, ProA and ProB are very similar except for the fact that ProA contains a catalytic zinc-binding motif and ProB does not contain any catalytic residues. ProA and ProB were localized to the soluble cytoplasmic and periplasmic protein portion, where a tiny amount of ProA was detectable in the inner membrane fraction of the bacterial fractions that were collected to do subcellular fractionation to determine a subcellular location. E. canis and E. chaffeenis infected DH82 cells both lysed that contained anti- rProA antibodies, showing that both E. canis and E. chaffeerais express ProA in culture.

Furthermore, both naturally and experimentally infected dogs with E. canis infected DH82 cells recognize rProA and rProB lending to the conclusion that ProA and ProB are expressed in vivo and in vitro. However, ProB was not detectable in a western blot using anti-rProB antibodies with E. chaffeenis. E. canis did detect anti-rProB antibodies. This result shows that ProB may serve as a tool for serological differentiation of E. canis and E. chaffeenis. Antisera from naturally and experimentally infected dogs with E. canis contained antibodies recognizing rProA and rProB. Serum from an uninfected dog did not recognize either of the two proteins. Immunofluorescence staining of E. canis in DH82 cells with rabbit anti-rProA and anti-rProB sera was performed, both ProB and ProA antiserum strongly label the intracellular ehrlichial organisms, showing that ProA and ProB can serve as target antigens and that anti-rProA and anti-rProB sera can be used for indirect immunofluorescent assays (IFA) diagnosis. Recombinant ProB can be used as an antigen in ELISA or Western blot analysis to perform a diagnosis of an E. canis infection in animals.

Overexpression of ProA, ProB, ORF, cytochrome oxidase and the lipoprotein signal peptidase homolog The E. canis antigens are overexpressed in a T7 promoter plasmid. The pRSET vector allows high level expression in E. coli in the presence of T7 RNA polymerase, which has a strong affinity for the T7 promoter. After subcloning the antigen genes into the pRSET vector, the subclones are transformed into an F'E. coli JM109 strain. For maximum protein expression, the transformants are cultured to O. D. 600=0.3, exposed to IPTG (1 mM) for one hour and then transfected with M13/T7 bacteriophages at a multiplicity of infection (MOI) of 5-10 plaque forming units (pfu) per cell. Time course studies indicate that maximum induction is reached two hours after induction.

The pellet is harvested by centrifugation and the cells are resuspended in 6M Guanidinium (pH 7. 8). Cells are ruptured by French press and the total lysate is spun at 6000 rpm to separate cell debris by a technique known in the art, and hereby incorporated by reference (Chang et al., 1993c). Immobilized metal ion affinity chromatography (IMIAC) is used to purify each of the proteins under denaturing conditions as described by the manufacturer (Invitrogen, San Diego, CA). The protein samples are separated by SDS-polyacrylamide gel electrophoresis (SDS-PAGE) and visualized after staining with coomassie blue.

Diagnosis Techniques: KELA ELISA, Western Blotting (Immunoblots), and PCR The recombinant ProA, ProB, and MmpA proteins are useful diagnostic agents. One diagnostic technique where the ProA, ProB, and MmpA proteins are used is kinetic enzyme linked immunosorbent assay (KELA ELISA), as described by a technique known in the art (Appel et al., 1993 ; Chang et al., 1995). KELA measures the levels of serum antibodies to E. canis that is present. In this diagnostic technique, diluted serum (1: 100 dilution) is added to duplicate wells in microtiter plates that contain antigens of MmpA, ProA, and ProB. The antigens are prepared by French-pressing them. The bound antibodies are then detected by using second antibodies of a goat anti-canine antibody of heavy and light chain specificity conjugated to horseradish peroxidase (HRP). Color development is seen and measured using the chromogen tetramethylbenzidine with H202 as a substrate, which is measured kinetically and expressed as the slope of the reaction rate between the enzyme and substrate solution.

Each unit of slope is designated as a KELA unit. The cutoff point between positive and negative samples is then confirmed by Western blotting against French-pressed E. canis.

The procedure for Western blot analysis, as described by a technique known in the art (Appel et al., 1993; Chang et al., 1995), is performed. Recombinant ProA, ProB, and MmpA are used as antigens and are subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Western blot analysis is performed in a miniblotter.

Test sera from experimental animals are used as a first antibody, followed by goat anti- dog IgG conjugated to HRP as a second antibody. Bands are developed by using substrates, such as 24 yg of 4-chloro-1-naphthol in 8 ml of methyl alcohol, 40 ml of Tris-buffer solution, and 24 y1 of 30% H202.

Another diagnostic technique recombinant ProA, ProB, and MmpA proteins can be used for is PCR diagnosis. The DNA from biopsy samples of skin or from post mortem tissues or blood are extracted as described by a technique known in the art (Chang et al., 1998a; Chang et al., 1998b). To prevent contamination of the mixtures and samples, DNA extraction, amplification, and detection of PCR products are all performed in different rooms. Amplification of E. canis MmpA, ProA, or ProB-specific target sequences is carried out in a 50-tel reaction mixture. As a positive control, E. canis genomic DNA is used. As a negative control distilled water is used. The reaction mixture is then put through 40 cycles of amplification using an automated DNA thermal cycler. Each cycle involves heating the reaction mixture to 94°C from 1 minute, to cause the DNA to denature; cooling of the reaction mixture to 69°C for 1 minute, to allow the primers to anneal; and then heating the reaction mixture to 72°C for 2 minutes, to allow primer extension to occur. Gel electrophoresis on a 1. 5% agarose gel is done in order to get visualization of the PCR amplification products.

Vaccine Development Prior to the present invention, no vaccine against E. canis had been developed.

E. canis is endemic in dogs and closely related canidae in many parts of the world. Dogs in North America are also increasingly at risk and the application of the present invention can potentially save the lives of thousands of dogs each year. An E. canis vaccine that can elicit cell-mediated immunity against this tick-borne disease of dogs is desperately needed.

DNA Vaccine A DNA vaccine is constructed by subcloning the gene of interest into a eukaryotic plasmid vector. Candidate vectors include, but are not limited to, pcDNA3, pCI, VR1012, and VR1020. This construct is used as a vaccine.

Each of the newly identified genes, ProA, ProB, mmpA, the cytochrome oxidase homolog, or the partial lipoprotein signal peptidase homolog can be used to create a DNA vaccine (reviewed in Robinson, 1997). In addition, any immunologically active portion of these proteins is a potential candidate for the vaccine. A plasmid containing one of these genes in an expression vector is constructed. The gene must be inserted in the correct orientation in order for the genes to be expressed under the control of eukaryotic promoters. Possible promoters include, but are not limited to, the cytomegalovirus (CMV) immediate early promoter, the human tissue plasminogen activator (t-PA) gene (characterized in Degen et al., 1986), and the promoter/enhancer region of the human elongation factor alpha (EF-1 a) (characterized in Uetsuki et al., 1989). Orientation is identified by restriction endonuclease digestion and DNA sequencing.

Expression of these gene products is confirmed by indirect immunofluorescent staining of transiently transfected COS cells, CHO cells, or other suitable cells. The same plasmid without these genes is used as a control. Plasmid DNA is transformed into Escherichia coli DH5a. DNA is purified by cesium chloride gradients and the concentration is determined by a standard protocol being known in the art, and incorporated by reference (Nyika et al., 1998).

Once the DNA is purified, the vector containing the insert DNA can be suspended in phosphate buffer saline solution and directly injected into dogs. Inoculation can be done via the muscle with a needle or intraveneously. Alternatively, a gene gun can be used to transport DNA-coated gold beads into cells by a technique known in the art, and hereby incorporated by reference (Fynan et al., 1993). The rationale behind this type of vaccine is that the inoculated host expresses the plasmid DNA in its cells, and produces a protein that raises an immune response. Each of the newly identified genes can be used to create a vaccine by this technique.

CpG molecules can be used as an adjuvant in the vaccine. This technique is known in the art, and is hereby incorporated by reference (Klinman et al., 1997).

Adjuvants are materials that help antigens or increase the immune response to an antigen. The motifs consist of an unmethylated CpG dinucleotide flanked by two 5' purines and two 3'pyrimidines. Oligo nucleotides containing CpG motifs have been shown to activate the immune system, thereby boosting an antigen-specific immune response. This effect can be utilized in this invention by mixing the CpG oligonucleotides with the DNA vaccine, or physically linking the CpG motifs to the plasmid DNA.

Immunofluorescence staining of E. canis in DH82 cells with rabbit anti-rProA and anti-rProB sera was performed, both ProB and ProA antiserum strongly label the intracellular ehrlichial organisms, showing that ProA and ProB can serve as target antigens and that anti-rProA and anti-rProB sera can be used for indirect immunofluorescent assays (IFA) diagnosis, making the DNA vaccine a viable option to combat this disease.

Recombinant Vaccine In order to develop a recombinant vaccine, each of the genes is individually subcloned into overexpression vectors, and then purified for vaccine development. ProA, ProB, mmpA, the cytochrome oxidase homolog, or the partial lipoprotein signal peptidase homolog is expressed in a plasmid with a strong promoter such as the tac, T5, or T7 promoter. Alternatively, immunologically active fragments of these proteins are used in the development of a vaccine. Each of these genes is subcloned into a plasmid and transformed into an E. coli strain as described above.

The recombinant protein is overexpressed using a vector with a strong promoter.

Vectors for use in this technique include pREST (Invitrogen Inc. , CA), pKK233-3 (Pharmacia, CA), and the pET system (Promega, WI), although any vector with a strong promoter can be used. After overexpression, the proteins are purified and mixed with adjuvant. Potential adjuvants include, but are not limited to, aluminum hydroxide, QuilA, or Montamide. The purified protein is used as immunogen to vaccinate dogs by a technique being known in the art, and incorporated by reference (Chang et al., 1993c; Chang et al., 1995). Briefly, the individual protein is expressed and purified from E. coli. Then, the dogs are injected intramuscularly or subcutaneously with the purified recombinant vaccine and adjuvant. This injection elicits an immune response.

T Cell Epitope Vaccine Direct cell cytotoxicity mediated by CD8+ T lymphocytes (CTL) is the major mechanism of defense against intracellular pathogens. These effector lymphocytes eliminate infected cells by recognizing short peptides associated with MHC class I molecules on the cell surface. Exogenous antigens enter the endosomal pathway and are presented to CD4+ T cells in association with class II molecules whereas endogenously synthesized antigens are presented to CD8+ T cells in association with MHC class I molecules. E. canis is an intracellular pathogen that resides in monocytes and macrophages. The present invention develops novel ways of generating an E. canis- specific CTL response that would eliminate the organism from monocytes or macrophages of infected animals.

A strategy for increasing the protective response of a protein vaccine is to immunize with selective epitopes of the protein. The rationale behind this is that an epitope vaccine contains the most relevant immunogenic peptide components without the irrelevant portions. Therefore, a search is performed for the most highly antigenic portions of the newly identified proteins.

To identify T-cell epitopes from the newly discovered proteins, an initial electronic search for homologous sequences to known T-cell epitopes is performed. In addition, extensive T-cell epitope mapping is carried out. Each of the proteins, ProA, ProB, mmpA, the cytochrome oxidase homolog, and the partial lipoprotein signal peptidase homolog, is tested for immunogenic peptide fragments. Mapping of T cell epitopes by a technique known in the art is hereby incorporated by reference (Launois et al., 1994; Lee and Horwitz, 1999). Briefly, short, overlapping peptide sequences (9-20 amino acids) are synthesized over the entire length of the protein in question. These short peptide fragments are tested using healthy dogs, which have been immunized with the protein of interest. Peripheral blood mononuclear cells from the dogs are tested for T cell stimulatory and IFN-y inducing properties. Those fragments which elicit the strongest response are the best candidates for a T-cell epitope vaccine.

Once fragments are identified which will make the best epitopes, a recombinant adenylate cyclase of Bordetella bronchiseptica is constructed carrying an E. canis CD8+ T cell epitope. The adenylate cyclase toxin (CyaA) of Bordetella bronchiseptica causes disease in dogs and elicits an immune response. In addition, CyaA is well suited for intracytoplasmic targeting. Its catalytic domain (AC), corresponding to the N-terminal 400 amino acid residues of the 1,706-residue-long protein, can be delivered to many eukaryotic cells, including cells of the immune system. Also, toxin internalization is independent of receptor-mediated endocytosis, suggesting that the catalytic domain can be delivered directly to the cytosol of target cells through the cytoplasmic membrane.

The Pseudomonas aeruginosa exotoxin A (PE) is another toxin which could be used in this procedure to deliver peptides or proteins into cells, by a technique known in the art, and hereby incorporated by reference (Donnelly et al., 1993).

Foreign peptides (16 residues) have been inserted into various sites of the AC domain of CyaA without altering its stability or catalytic and calmodulin-binding properties. Thus, protein engineering allows the design and delivery of antigens that specifically stimulate CTLs. The induction of specific CD8+ T cells can play an important role in canine ehrlichiosis control due to the intracellular persistence of E. canis in monocytes.

The adenylate cyclase (AC) toxin (cya) gene of B. bronchiseptica has been cloned. A synthetic double-stranded oligonucleotide encoding a 9 to 20 amino acid class I T cell epitope of either ProA, ProB, mmpA, the cytochrome oxidase homolog, or the partial lipoprotein signal peptidase homolog, is designed according to B. bronchiseptica codon usage. The complementary oligonucleotides are inserted in the hypervariable region of the cloned AC-coding sequence of the cya. This technique is known in the art in other systems, and is incorporated by reference (Sebo et al., 1995; Guermonprez et al., 1999).

Recombinant plasmids carrying the chimeric cya gene are sequenced to determine the copy number and orientation of the inserted epitope. A plasmid with a complete copy of the insert that specifies the T-cell epitope (CD8+) in the correct orientation is chosen from the sequenced plasmids. The ability of the new chimeric protein to enter eukaryotic cells is necessary to ensure intracellular targeting of the epitopes (Fayolle et al., 1996).

A vaccine can be created in one of two ways. Recombinant chimeric protein can be purified and used to inoculate dogs. Alternatively, an attenuated B. bronchiseptica strain that carries a T-cell epitope or E. canis gene by in-frame insertion into adenylate cyclase is created by allelic-exchange. Allelic-exchange is a technique known in the art, and is hereby incorporated by reference (Cotter and Miller, 1994).

Finally, protection against E. canis infection in dogs vaccinated with the adenylase cyclase-ProA, ProB, mmpA, cytochrome oxidase homolog, or lipoprotein signal peptidase homolog chimeric protein is determined. Wild type and recombinant ACs and CyAs are diluted to working concentrations in PBS and the chimeric protein is injected into dogs either intramuscularly or subcutaneously. Alternatively, the T-cell epitope is inserted into the adenylate cyclase gene of an attenuated B. bronchiseptica strain in frame, and the dogs are given the live bacteria.

Recombinant antigens are promising candidates for human and animal vaccination against various pathogens. However, a serious drawback is the poor immunogenicity of recombinant antigens as compared to native antigens. A major challenge in the development of a new recombinant vaccine is, therefore, to have a new adjuvant system that increases the immunogenicity of antigens. Cytokines are powerful immunoregulatory molecules. Cytokines which could be used as adjuvants in this invention include, but are not limited to, IL-12 (interleukin-12), GM-CSF (granulocyte- macrophage colony stimulating factor), IL-lß (interleukin-lß) andey-IFN (gamma interferon).

These cytokines can have negative side effects including pyrogenic and/or proinflammatory symptoms in the vaccinated host. Therefore, to avoid the side effects of a whole cytokine protein, an alternate approach is to use synthetic peptide fragments with the desired immunostimulatory properties. The nonapeptide sequence VQGEESNDK of IL-lp protein is endowed with powerful immuno-enhancing properties, and is discussed here to illustrate the use of a cytokine to increase immunogenicity.

This nonapeptide is inserted into the ProA, ProB, mmpA, the cytochrome oxidase homolog, or the partial lipoprotein signal peptidase homolog protein and its immunogenicity is compared to that of the native protein. Reportedly, the insertion of this sequence into a poorly immunogenic recombinant antigen increases the chance of a strong protective immune response after vaccination. This peptide could enhance the in vivo immune response against both T-dependent and T-independent antigens. The canine IL-1 (3 sequence may mimic many immunomodulatory activities of the entire molecule of IL-1 (3 while apparently lacking many of its undesirable proinflammatory properties. This strategy is employed to increase the immunogenicity of ProA, ProB, mmpA, cytochrome oxidase, the partial lipoprotein signal peptidase homolog and other E. canis antigens.

Plasmid pYFC199 is derived from a pBR322 plasmid by the insertion of a fragment that includes the ProA, ProB, mmpA, the cytochrome oxidase homolog, or the partial lipoprotein signal peptidase protein from E. canis. This plasmid contains a unique Hindlll site where in-frame insertions encoding exogenous sequences can be inserted. Two complementary oligonucleotides, SEQ. ID. NO. 12 and SEQ. ID. NO.

13, that encode the canine IL-lp 163-171 peptide are annealed, cut with HindIII, and inserted into the pYFC199 Hindlll site. The recombinant plasmid carrying the chimeric IL-lß gene is sequenced to determine the orientation of the inserted epitope.

The efficacy of the recombinant proteins as vaccines is tested in dogs. The purified protein is injected intraperitoneally into dogs. Specific pathogen free (SPF) dogs are divided into five groups: one group is given recombinant adenylate cyclase of Bordetella bronchiseptica carrying E. canis CD8+ T cell epitopes derived from ProA, ProB, mmpA, cytochrome oxidase homolog, or the partial lipoprotein signal peptidase homolog, one group is given recombinant adenylate cyclase of Bordetella bronchiseptica as a control, one group is given the ProA, ProB, mmpA, cytochrome oxidase homolog, or the partial lipoprotein signal peptidase homolog protein plus a canine IL-1 (3 163-171 insert, one group is given a T cell epitope derived from ProA, ProB, mmpA, cytochrome oxidase homolog, or the partial lipoprotein signal peptidase homolog alone, and the last group is given PBS as a negative control.

All animals are vaccinated (30-40 jj. g each) two times. The dogs are challenged ten days after the last vaccination with 107 E. canis. At day five postchallenge, approximately 1 ml blood from each dog is collected in an EDTA tube. Whether the vaccinated groups eliminate the organisms as compared to that of the control group is tested by culture and PCR. Two primers derived from the genes cloned can be used to amplify the gene product from the tissues or blood samples from these dogs. The internal primer can also be designed for use as an oligonucleotide probe to hybridize the PCR gene product.

This invention provides a badly needed vaccine against the E. canis bacterium.

The vaccine can be used to protect dogs throughout the world from canine monocytic ehrlichiosis.

Accordingly, it is to be understood that the embodiments of the invention herein described are merely illustrative of the application of the principles of the invention.

Reference herein to details of the illustrated embodiments are not intended to limit the scope of the claims, which themselves recite those features regarded as essential to the invention.