Targeted mutagenesis in Ehrlichia canis deleting the phage head-to-tail connector protein gene and its assessment as a vaccine candidate preventing canine ehrlichiosis

Abstract

Ehrlichia canis is primarily a Rhipicephalus sanguineus tick-borne rickettsial pathogen initially identified as causing canine monocytic ehrlichiosis, and infections in people have also been reported in Venezuela, Mexico, and parts of Europe. It is of high importance to have a vaccine suitable in protecting the canine host, which will aid in lessening E. canis infections also in people. Gene inactivation mutations in the phage head-to-tail connector protein genes (phtcp) from E. chaffeensis and A. marginale caused attenuated growth, and prior infection with the mutated bacteria induced protective immunity against wild-type bacterial infections in the natural hosts, independent of infection acquisition from blood-borne infection or tick-transmission. In the current study, we describe the development of targeted mutagenesis for the first time in E. canis genome and with a novel modification to avoid introducing antibiotic resistance cassettes to delete the phtcp ortholog from E. canis. The mutated E. canis was then assessed for its in vivo growth and the induction of host immunity exerted following the mutant infection aiding to protect against wild-type infection challenge in the canine host. We assessed systemic pathogens, hematological parameters, immune responses, and plasma cytokines following the mutant infection relative to uninfected dogs. Similarly, the assessments were carried out following wild-type pathogen infections in dogs with or without prior mutant infection challenges. The study demonstrates that prior infection of dogs with the mutant induces immunity to prevent infection establishment by wild-type E. canis. Similarly, the mutant infection resulted in clear biological differences compared to the wild-type infection. This study establishes that the molecular genetic methods are broadly applicable to pathogens belonging to the family Anaplasmataceae and that the modified live vaccines with phtcp gene orthologs are valuable in reducing the diseases caused by the tick-borne rickettsial pathogens belong to Anaplasmataceae, including E. canis.

1. Introduction:

Emerging and reemerging tick-borne rickettsial diseases remain a major public health concern in the USA and many parts of the world [1–5]. The diseases pose a significant risk to companion animal health, as well as contributing to global economic losses within the food animal production industry [6–9]. Despite some progress made, molecular genetic methods applicable to obligate intracellular pathogens remain a challenge [10–13]. Targeted gene inactivation methods are of great importance for investigations into defining the functional roles of genes [10, 14–16]. The availability of targeted mutagenesis methods greatly aids in advancing research in understanding the pathogenesis of obligate intracellular bacteria and extending research into the development of therapeutics and prevention methods [10, 12, 17–21].

Canine monocytic ehrlichiosis (CME), resulting from Ehrlichia canis infections, is the most prevalent tick-borne disease of dogs throughout the world, including in the USA [22–26]. The primary transmitting vector is Rhipicephalus sanguineus sensu lato (s.l.) complex which comprises of multiple, closely related, and morphologically similar species which are commonly known as the brown dog tick [27–29]. While the tick complex prefers dogs as the primary host and resides indoors, it also feeds on people [27, 28]. Human infections with E. canis are reported from several Western hemisphere and European countries, such as Venezuela, Mexico, Costa Rica, Italy, and Montenegro [30–35]. E. canis tropism is to the immune cells, monocytes, and macrophages, leading to a multisystemic disease ranging from acute to subclinical and chronic phases [24, 36, 37]. CME is characterized by nonspecific clinical signs such as fever, lethargy, inappetence, and several hematological abnormalities, including thrombocytopenia, leukopenia, and anemia [38]. Most immunocompetent dogs will naturally clear the clinical signs without any treatment. However, the disease can potentially progress to a chronic phase, causing aplastic anemia or severe bone marrow aplasia, known as myelosuppression, developing into life-threatening pancytopenia [39–41]. If left untreated, death can occur, resulting from septicemia or other complications related to the disease [39–41]. Considering the impact of E. canis infections on dogs and humans and the limited treatment options, mainly the use of doxycycline [42–46], it is of critical importance to have a vaccine that protects the canine host [37], which will also aid in reducing the risk of human infections with the pathogen. Despite a prior study reporting that a cell culture-derived attenuated strain provided immune protection against virulent challenge [36], to date, there are no follow-up investigations to advance vaccine research in preventing CME. Additionally, this study did not describe how bacterial attenuation is achieved at the genomic level [36].

We previously developed a homologous recombination-based molecular genetic system in Ehrlichia chaffeensis [11, 47], a member of the Anaplasmataceae family of pathogens and responsible for causing infections in people and several other vertebrates, including dogs [2, 22, 48–53]. The mutational methods have been valuable in creating mutations in several bacterial genes, including disrupting and restoring a gene function [11]. We also reported that inactivation of the gene encoding a membrane-bound phage head-to-tail connector (phtcp) of E. chaffeensis causes attenuated growth in vivo in both reservoir (white-tailed deer) and incidental (dog) hosts [17, 47](naturally known to acquire infections from ticks, while inducing a long-lasting immunity conferring protection against wild-type infection challenge by both intravenous (IV) inoculation and tick transmission [17, 18, 20, 21, 47]. Further, we reported that the E. chaffeensis phtcp gene inactivation results in the pathogen’s inability to obtain metal ions to support its growth in infected host macrophages [54]. Since the phtcp gene orthologs are well-conserved in all known Anaplasma and Ehrlichia species [54], we extended similar targeted mutagenesis development by deleting the phtcp ortholog from another related Anaplasmataceae pathogen, A. marginale [12]. Furthermore, we reported that vaccination of cattle with the genetically modified A. marginale mutant lacking the phtcp displays significantly attenuated growth and induces immunity to prevent severe clinical bovine anaplasmosis resulting from IV infection or from infected ticks [12, 19].

In the current study, we developed the first targeted deletion mutation in E. canis while not introducing any antibiotic resistance genes as part of the mutational strategy. With no vaccines available to prevent CME, we assessed whether the phtcp deletion mutant, as a modified live attenuated vaccine (MLAV), can serve as a vaccine candidate to support improving companion animal health (supplementary Fig S1). Vaccination of dogs with the E. canis phtcp gene deletion mutant resulted in the canine host inducing immunity that is sufficient in preventing wild-type pathogen infection establishment.

2. Materials and methods:

2.1. Culturing E. canis.

Wild-type and the genetically modified mutant E. canis Jake isolate were cultivated in Ixodes scapularis cell line culture (ISE6) (ATCC# CRL-3576) at 34°C in the absence of CO₂ as described earlier [55]. Wild-type and mutant E. canis were also cultivated in the canine macrophage cell line, DH82 (ATCC# CRL-10389), at 37°C in an incubator with 5% CO₂.

2.2. Generation of Ecaj_0381 deletion plasmid construct.

Homologous arms of 0.84 kb and 0.86 kb in length spanning the 5’ and 3’ to the targeted deletion region of Ecaj_0381 gene (the phtcp gene homolog) of E. canis (GenBank # CP000107) were amplified by PCR using specific primers (Table S1) and using wild-type bacterial genomic DNA as the template. The amplicons were cloned upstream and downstream to the previously described pGGA plasmid vector (New England Biolabs, Ipswich, USA) containing the A. marginale amtr promoter to drive the expression of the mCherry gene coding sequence and aadA gene encoding resistance to streptomycin and spectinomycin (pLox cisA7Himar Ch-SS) [56]. A Golden Gate Assembly kit (New England Biolabs, Lpswich, MA, USA) was used to reassemble the fragments in the following order: 5’ E. canis homology arm, Amtr-mCherry-aadA segment, and 3’ E. canis homology arm. The 5’-3’ homologs are referred to here as the Left homologous arm (LHA) and the Right homologous arm (RHA). The recombinant plasmid was transformed into the DH5α strain of E. coli and used to generate plasmid DNA by following the standard molecular biology protocols [57]. The proper assembly of the recombinant plasmid was confirmed by DNA sequence analysis using T7 and SP6 promoter primers (Integrated DNA Technologies, Coralville, IA, USA) having binding sites upstream and downstream to the insertion segment within the pGGA plasmid backbone. Subsequently, the aadA gene was deleted from the recombinant plasmid using the Q5 Site-Directed Mutagenesis Kit (New England Biolabs, Ipswich, USA) according to the manufacturer’s protocol, resulting in a new recombinant construct designated pGGA-Ecaj_0381-KO-Amtr-mCh. This modified plasmid was then used as the template to amplify the insertion segment containing the LHA, the Amtr-mCherry segment, and the RHA using the primers (RRG2015 and RRG2048) listed in Table S1. The PCR amplicons were then purified using a QIAquick PCR Purification Kit (Qiagen, Germantown, MD) as in [11].

2.3. Generation, clonal purification, verification, and propagation of E. canis phtcp mutant.

The E. canis phtcp deletion mutation was generated as per prior reported methods [11, 12] with a modification to exclude the use of an antibiotic selection marker. The cultures were maintained in the media without added antibiotics, and the media were changed once a week for the first three weeks and twice a week thereafter. Cultures were assessed for the mutant expressing mCherry by fluorescent microscopy. To achieve clonal purity of the mutant from wild-type E. canis, mutant bacteria were enriched through repeated serial dilutions of cultured bacteria, which took several months. When approximately 10% of the culture was expressing mCherry, the media supernatant from infected cultures containing cell-free bacteria was collected and used to infect a monolayer of ISE6 tick cells in six-well plates. The day after infecting the monolayer, all media was removed and replaced with 2 ml of a modified infection medium containing 0.4% low-melting-temperature agarose, which created a thin agarose overlay allowing the development of a plaque of cells expressing mCherry. After the agarose had set overnight, 4 mL of culture media was added to the agarose layer, and cultures were monitored for 4 weeks. When colonies of mCherry expressing fluorescent bacterial colonies were visible (between 3 and 4 weeks), liquid media was removed, and a pipette tip was used to carefully recover mCherry expressing fluorescent cells from beneath the agarose layer. The recovered cells were then diluted in 0.5 ml culture media in sterile 1.5 ml microcentrifuge tubes, pipetted to disrupt cell clumps and agarose, and added to wells of a fresh 6-well plate having an ISE6 cell monolayer maintained in infection media. A total of three sequential passages were performed to isolate the purified mutant. Presence of the mutant expressing mCherry was monitored and confirmed by fluorescence microscopy, and images were captured using BioTek Cytation C10 Confocal Imaging Reader (Agilent, USA). The clonal purity of the mutant was independently established by PCR, Southern blot analysis, and whole-genome sequencing analysis as outlined below.

A PCR assay was performed using genomic DNA recovered from the mutant to define the clonal purity using the primers (RRG2015 and RRG2048) listed in Table S1. The primers were designed to target the genomic regions upstream and downstream of the gene deletion region (listed in Table S1). The PCR assays were performed in 25 μl reactions in 1x Q5 reaction buffer containing 2 mM MgCl₂, 0.5 mM of each dNTP, 0.2 μM of each of the primers, 1 unit of Q5 high-fidelity DNA polymerase (New England Biolabs, Ipswich, MA, USA), and genomic DNA from wild-type or mutant organisms as the templates. The PCR cycling conditions for the first two PCRs were 98°C for 30 s, followed by 35 cycles of 98°C for 10 s, 65°C for 30 s, and 72°C for 2 min 30 s, then 72°C for 3 min, and a final hold at 4°C. The presence of a specific PCR amplicon was confirmed by resolving DNAs on a 1 % agarose gel containing ethidium bromide and visualizing it under a UV transilluminator. To independently confirm the clonal purity of the mutant, Southern blot analysis was performed using the mutant culture-derived genomic DNA as well as the wild-type E. canis genomic DNA; the DNAs were digested with BamH1, Sph1, or a combination of BamH1and Sph1 restriction enzymes by following the standard molecular biology protocols [57]. The digested DNAs were resolved on a 0.9 % agarose gel, transferred to a nylon membrane, and subjected to blot analysis using a DNA segment to be retained in the restriction-digested DNA fragments surrounding the genetically modified region spanning DNA from the LHA or insertion-specific mCherry gene segment DNA as a probe. Whole-genome sequencing analysis was performed to further confirm the clonal purity of the mutant. Genomic DNA extracted from the mutant-infected DH82 was sequenced using a hybrid sequencing approach that combined the Oxford Nanopore MinION and Illumina NextSeq 2000 platforms (Plasmidsaurus Inc., Kentucky, USA). To remove Nanopore adapters and to perform quality filtering, we used Porechop v.024 [58] and Filtlong v.0.2.1 (https://github.com/rrwick/Filtlong) with default settings. The remaining high-quality reads were mapped to the E. canis Jake isolate reference genome (CP000107) and for the presence of Ecaj_0381deletion construct sequences including the Amtr promoter and mCherry using Minimap2 v.2.28 [59]. Mapped reads were extracted via samtools v.1.21 [60] and assembled de novo using Flye v.2.9.6 [61]. To improve the assembly, the Illumina NextSeq2000 reads were groomed and filtered for quality using Trimmomatic v.0.39 [62]. The high-quality Illumina reads were then mapped to the draft assembly using Bowtie2 v.2.5.3 [63], followed by polishing using Polypolish v.0.6.0 [64].

RNA analysis was performed using total RNA isolated from the mutant and the wild-type E. canis, as described elsewhere [65], to assess mRNA expression from the gene Ecaj_0381 and from a gene upstream and downstream to the mutant site; Ecaj_0383 and Ecaj_0380, respectively. After verifying the removal of any residual genomic DNA from the RNAs by DNase I treatment using Q1 RNase-Free DNase (Promega, Madison, WI, USA) per the manufacturer’s recommendation, first-stand cDNAs were synthesized using SuperScript III First-Strand Synthesis System for RT-PCR kit (Invitrogen, Carlsbad, CA, USA ) and used as templates to perform RT-PCR analysis. Primers for the three gene targets were listed in Table S1. Polymerization was performed with 2 μl of cDNA template using GoTaq Green DNA polymerase per the manufacturer’s instructions (Promega, Madison, WI, USA). The PCR products were resolved in a 1.5% agarose gel stained with ethidium bromide, and a UV transilluminator was used for visualization to identify the target PCR products.

2.4. Infection experiments in dogs with the mutant E. canis.

Infection experiments in dogs of both sexes were performed. The study design was reviewed and approved by the University of Missouri’s Animal Care and Use Committee (UM ACUC) and executed as per the approved protocol and as per compliance with the Public Health Service (PHS) policy on the humane care and use of laboratory animals and the U.S. Department of Agriculture’s Animal Welfare Act and Animal Welfare Regulations (https://olaw.nih.gov/policies-laws/phs-policy.htm). We obtained purpose-bred laboratory-reared beagle dogs of both sexes about 8 months of age from a USDA approved commercial vendor (Ridglan Farms, Inc. Mt. Horeb, WI). All dogs were allowed to acclimate for two weeks. During this period, 4 ml each of blood samples in EDTA tubes were collected from all dogs and used to confirm that none of the animals had current or past E. canis infections. The analysis was performed by subjecting total DNAs recovered from blood by pathogen-specific PCR. Additionally, plasma samples recovered from the bloods were evaluated by ELISA analysis for the pathogen-specific antibodies by ELISA. All dogs tested negative by PCR and ELISA. The infections were performed using either the in vitro cultured mutant organisms or the wild-type E. canis cultured in DH82 cells. Before mutant or wild-type infection, animals were administered Benadryl orally at 2 mg/kg approximately 30 min prior to the IV infection challenge. For mutant E. canis infection experiments, dogs were administered about 1.25 × 10⁸ culture-derived mutant organisms re-suspended in 1 ml of 1 x PBS (n=5) intravenously (Group 1). Similarly, a group of dogs (n=4) that did not receive the mutant were injected with 1 x PBS only and served as controls (Group 2). Both groups of dogs were then challenged with wild-type E. canis on day 29 following the mutant infection by IV using 2 ml each of cultured organisms (about 1.25 x 10⁸ bacteria per dog). All dogs were monitored for signs of infection, changes in body temperature, weight, and Complete Blood Count (CBC) analysis. Blood samples in EDTA tubes were collected over time and used to define the presence of infection by in vitro culture recovery and PCR analysis as described in the previous section. Further, total RNAs were isolated from 250 μl of each blood sample at each time point throughout the study period. Quantitative reverse transcriptase polymerase chain reaction (qRT-PCR) assays targeting the bacterial 16S rRNA (using primers listed in Table S1) were performed as described in our previous work to further define the infection status [66]. After day 72 post-challenge, all animals received doxycycline at a dose of 10 mg/kg body weight once a day orally for 28 days as per Merck veterinary manual. Blood samples collected on days 11 and 32 post-doxycycline treatment were assessed by PCR using primers RRG2015 and RRG2048, listed in Table S1.

2.5. qRT-PCR.

Two microliters of total RNA recovered from 250 μl of blood samples collected in EDTA tubes from dogs at various time points were assessed by qRT-PCR to detect and quantify E. canis RNA. This was performed according to the manufacturer’s instructions using the SuperScript III Platinum One-Step RT-qPCR Kit (Thermo Fisher Scientific, USA). The qRT-PCR conditions were followed as previously described [66].

2.6. Assessment of IgG response by ELISA.

Whole-cell antigen was prepared and used to perform an enzyme-linked immunosorbent assay (ELISA) as described earlier [17]. Briefly, ninety-six-well Immulon^™ 2HB microtiter plates (Thermo Fisher Scientific, Waltham, MA) were coated with 2 μg/ml E. canis WCA in the ELISA coating buffer and incubated overnight. Plates were then blocked with 1% BSA for one hour at 37°C, followed by 2-3 hours of incubation at 37°C with 1:200 diluted plasma collected at different time points post-mutant infection, or following infection challenges with wild-type E. canis. The plates were washed three times with 1X PBS containing 0.05% Tween 20 (1X PBST), then incubated with 100 μl of horseradish peroxidase (HRP)-conjugated anti-canine total IgG at a 1:50,000 dilution (PA1-29738, Thermo Fisher Scientific, USA). The plates were washed three times, then incubated with TBM substrate for about 5-8 min at room temperature. The reactions were stopped with the stop buffer having 1 M phosphoric acid, and the absorbance at 450 nm was measured using a Tecan Infinite M Nano (Männedorf, Switzerland).

2.7. Plasma cytokines analysis.

The commercially available Cytokine/Chemokine/Growth Factor 11-Plex Canine Procarta Plex^™ Panel 1 kit (Thermo Fisher Scientific, USA) was used to measure secreted cytokines and chemokines using Luminex xMAP technology. The assays were performed according to the manufacturer’s instructions using plasma samples collected in EDTA tubes on the day before mutant infection and on days 3, 7, and 14 post-immunization/infection. Samples were also taken after E. canis wild-type infection challenges for the same time points from all groups of dogs.

2.8. Preparation of peripheral blood mononuclear cells (PBMC) to perform ELISPOT and ELISA assays for measuring the IFN-γ expression.

Blood samples were collected from dogs at different time points in 8 ml sodium citrate CPT tubes (BD Biosciences, San Jose, CA, USA) on the days preceding any infection or vaccination, and then on days 7 and 14 post-mutant infection challenges, and after wild-type infection challenges. The samples were centrifuged at 1500 x g for 30 min at room temperature, and the separated cells were thoroughly mixed before being shipped overnight to Iowa State University in cold packs. The harvested PBMCs were washed twice with cRPMI media, followed by cell counting and viability assessment. The cells were then adjusted to 4 × 10⁶ PBMCs/mL, and ELISPOT and ELISA assays were performed to measure IFN-γ expression by stimulating PBMCs with E. canis whole-cell antigens, as described in [20].

2.9. Statistical analysis.

Using GraphPad Software (GraphPad Software, La Jolla, CA, USA), the one- and two-way ANOVA tests with multiple comparisons, along with t-tests, were performed to identify statistical differences among groups of dogs following mutant infections and following wild-type infection challenges. As we found no evidence of sex-specific differences for all parameters assessed, the combined data were presented and described in the study.

3. Results:

3.1. Preparation of the homologous recombination construct for use in deleting the phtcp gene from E. canis.

To generate the targeted deletion mutation of the Ecaj_0381 gene encoding for phtcp from the E. canis strain genome (Jake isolate) (GenBank # CP000107), the homology arms spanning the genomic regions upstream and downstream to the phtcp gene, respectively, were cloned into a plasmid construct (pGGA). A. marginale tr gene promoter (Amtr) and the mCherry reporter gene coding sequence were inserted between the left and right homology arm segments (LHA and RHA) to allow mCherry expression in the mutant. The resulting recombinant plasmid construct, referred to as pGGA-Ecaj_0381-KO-Amtr-mCh, lacks any antibiotic selection markers while the mCherry coding sequence facilitates driving its expression by the Amtr promoter to track the mutant and its isolation (Fig 1A). The plasmid was then used as a template to generate homologous recombination segments by PCR to generate the construct segment, which contained both the 5’ and 3’ homology arms and the Amtr-mCherry segment. Genomic coordinates of the homology arms, position of restriction enzyme sites, and the Amtr-mCherry segments are detailed in (Fig 1B).

Figure 1.

The development of Ecaj_0381 (phtcp) gene deletion mutation from the E. canis genome. A) Plasmid map of pGGA-Eca_0381-KO-Amtr-mCh A. The plasmid was used to amplify the left (LHA) and right (RHA) homology arms flanked the Amtr promoter to drive expression of the mCherry reporter gene segment. B) Schematic representation of the E. canis genomic region targeted mutagenesis generation from wild-type (W) to mutant (M) by allelic exchange. The genomic coordinates of the LHA and RHA, restriction enzyme sites used for mutation verification (BamHI and SphI), and the size of the inserted Amtr-mCherry fragment are shown. Black arrows indicate the primers binding sites; P1 (RRG2051) and P2 (RRG2048) used to confirm Ecaj_0381 gene deletion. C) The gene deletion mutant growth in ISE6 tick cell culture expressing mCherry (confirmed by confocal microscopy using 40x magnification lenses at 400 x magnification. D) PCR analysis confirming the mutation. Primers annealing upstream and downstream of the mutation insertion region were amplified and resolved on an agarose gel which yielded the expected larger and smaller products from the mutant (M) and wild-type (W) DNAs, respectively (L, 1 kb plus DNA ladder). E) Southern blot analysis of W and M genomic DNAs digested with BamHI or double digested with BamHI and SphI. The digested genomic DNAs resolved on an agarose gel and transferred to a nylon membrane were hybridized with probes specific to the LHA or mCherry gene coding sequences, which verified the presence of the predicted segments in the clonally purified mutant DNA and for wild-type DNA. F) Nanopore sequencing reads used to assemble the genome of Ecaj_0381 mutant compared with the E. canis str. Jake reference genome (CP000107). The genome alignments revealed the expected Ecaj_0381 gene deletion in the mutant and inserted fragment in its place. Sequences flanking the transposon insertion site shared 99% identity. G) The impact of E. canis Ecaj_0381 (phtcp) gene deletion was assessed for the gene expression from it and two genes located upstream and downstream of the deleted gene. Transcriptional analysis of RNA recovered from the mutant and wild-type E. canis was assessed by RT-PCR targeting Ecaj_0380, Ecaj_0381, and Ecaj_0382. The assay included wild-type genomic DNA as the template for the PCR to serve as the positive control (+), no template PCR as the negative control (−), and cDNA as the template to represent RNA samples (R) from the organisms. (L, 1 kb plus DNA markers). The mutant cDNA lacked the expected amplicons only for the Ecaj_0381, but not in the cDNA from wild-type E. canis, while RNA expression from the Ecaj_0380 and Ecaj_0383 genes was similar for both cDNAs.

The targeted mutagenesis protocol for E. canis phtcp gene deletion was followed as per previously reported methodology for E. chaffeensis and A. marginale [11, 12]. The method involved utilizing 20 μg of linear DNA fragments for electroporation into purified E. canis organisms derived from ISE6 cells (Ixodes scapularis embryonic cells) cultures. The presence of E. canis mutant organisms expressing mCherry fluorescent protein in ISE6 cells was monitored; the mutant cultures expressing mCherry appeared after three weeks post-electroporation experiment (Fig 1C). Clonal purification of the mutant was achieved by enrichment of mutant bacteria by serial dilution through several passages of mutant and wild-type infected tick cells, followed by a 0.4% low melt agarose layered on top of the tick cell monolayer to facilitate the recovery of purified mutant colonies expressing mCherry; three sequential passages were performed to isolate the purified mutant. (More details of the method were included in the methods section). The presence of a clonally pure mutant was confirmed by three independent molecular methods: PCR analysis to identify the presence of a predicted larger amplicon in the mutant compared to the wild-type (Fig 1D), Southern blot analysis following restriction enzyme digestions with BamHI (B) or SphI (S) or double digestions with both enzymes (BS) utilizing either a DNA probe specific to the LHA or the mCherry coding sequence (mCh) (Fig 1E), and by whole genome sequencing analysis (Fig 1F). Isolated genomic DNA of the mutant E. canis included only the predicted 2.7 kb mutant-specific amplicon contrary to 2 kb observed for the wild-type genomic DNA, indicating the absence of wild-type genomic DNA (Fig 1D). Similarly, Southern blot analysis with the LHA genomic probe identified predicted restriction fragments of 5.7 kb for BamHI or 1.6 kb with SphI digested genomic DNA from the mutant, while the predicted 4.7 kb with BamHI and 3.2 kb with BamHI and SphI double digestion DNA fragments were evident when wild-type genomic DNA was used (Fig 1E, left panel). The anticipated 5.7 kb size for the BamHI-digested DNA was observed only in the mutant when the mCherry DNA probe was used demonstrating that there were no other locations in the genome where the insertion mutations resulted, and it was absent in wild-type DNA, similarly digested (Fig 1E, right panel). Further, whole genome sequence analysis of the mutant revealed the presence of the inserted sequence of Amtr promoter followed by mCherry coding sequence and the absence of the Ecaj_0381 gene in the mutant (Fig 1F). Further, the whole genome sequence analysis confirmed the presence of mCherry sequences only at the targeted mutational site. We then performed RNA analysis by reverse transcriptase PCR (RT-PCR) which revealed the anticipated absence of transcript for gene Ecaj_0381 in the mutant but not in the wild-type, while the RNA expression from genes upstream and downstream (Ecaj_0383 and Ecaj_0380) remained similar for the mutant and wild-type (Fig 1G).

3.2. Impact of the phtcp gene deletion on the progression of E. canis infection in a physiologically relevant host:

The canine infection model study flow chart is included (supplementary Fig S1). We assessed the impact of the phtcp gene deletion on the pathogen’s in vivo growth in the canine host, as it acquires the pathogen naturally. Five beagles (Group 1) were injected IV with approximately 1.25 x 10⁸ mutant organisms per dog and monitored for the systemic infection-associated changes twice a week for four weeks. The assessment included evaluating clinical symptoms, changes in CBC, presence of the mutant in circulation by PCR analysis, plasma cytokine secretion, and the pathogen-specific IgG response. Four beagles (Group 2) were included as uninfected controls. Over the course of the study, none of the mutant-infected or uninfected control beagles developed clinical signs such as fever or weight changes. Mutant genomic DNA was not detected in any of the controls (Group 2) but was detected in the blood samples of all five mutant-inoculated beagles (Group 1) by mutation region-specific PCR analysis (Table 1). Blood-derived RNA was similarly assessed by qRT-PCR targeting the E. canis 16S rRNA, and RNA copies were estimated as per our previously described method [66]. We detected the presence of bacterial RNA with a sharp increase over time, reaching its highest level on day 14 post-infection, followed by a marked decline thereafter (Fig 2 and supplementary Table S1). Since E. canis infections in dogs are known to alter hematological parameters, we performed CBC analysis, which revealed significant changes in the blood profiles of mutant-infected dogs compared to uninfected dogs (Fig 3). The mutant infection-specific changes included a substantial decline in the platelets (Fig 3A), red blood cells (RBC) (Fig 3B), and white blood cells (WBC) (Fig 3C). The mutant-infected animals developed an IgG response against the E. canis antigens, as evidenced from day 7 post-infection, while no IgG response was detected in uninfected dogs (Fig 4).

Table 1.

E. canis mutant infection assessment in the canine host (Mutant-specific PCR).

	Dog ID	Days post mutant infection
		0	3	7	10	14	21	29
Mutant infected dogs (Group 1)	AHL (M)	−	−	+	+	+	+	−
	BAK (F)	−	−	+	+	+	+	−
	BGK (F)	−	−	+	+	+	+	+
	CUL (M)	−	−	+	+	+	+	+
	DWL (M)	−	−	+	+	+	+	−
Dogs that did not receive the mutant infection (Group 2)	AVK (F)	−	−	−	−	−	−	−
	BCK (F)	−	−	−	−	−	−	−
	ECL (M)	−	−	−	−	−	−	−
	GVL (M)	n/a	n/a	n/a	n/a	n/a	−	−

^*+ and − refer to PCR positives and negatives, respectively, and n/a refers to samples not available. M and F refer to male and female dogs, respectively.

Figure 2.

E. canis 16s rRNA copy numbers in blood-derived RNA assessed by TaqMan qRT-PCR following infection with the phtcp mutant. Total RNA was isolated from blood samples collected at different time points following mutant infection in the canine host (n = 5). The qRT-PCR analysis was performed as per our previous study [66] to quantify E. canis small subunit ribosomal RNA copies per 25 μl of blood.

Figure 3.

The complete blood count (CBC) assessment following the mutant E. canis infection in the canine host. The CBC analysis performed over time following mutant E. canis infection revealed a significant drop in the platelet counts (P < 0.0001 ****) (panel A), RBC counts (P = 0.0007 ***) (panel B), and WBC counts (P < 0.0001 ****) (panel C), compared to uninfected dogs. The dotted lines indicate normal canine reference values, as per the Merck Veterinary manual. Statistical significance was measured using an unpaired t-test with Welch’s correction. Statistical significance was calculated from day 10 to 29, when values started to decline following mutant infection, and compared with the uninfected control group (right panels).

Figure 4.

Total IgG response to E. canis antigens in dogs was measured by ELISA using plasma samples collected weekly following the mutant infection. Specific antibody responses were observed from day 7 post-infection in the mutant-infected group of dogs but not in the uninfected control dogs.

3.3. Evaluation of the E. canis phtcp gene deletion mutant as a vaccine:

We then evaluated if the prior injection with the E. canis phtcp gene deletion mutant serves as an MLAV in offering protection against wild-type E. canis infection challenge. Approximately 1.25 x 10⁸ cultured E. canis wild-type bacteria were IV administered into all five Group 1 beagles on day 29 post-mutant infection. Similarly, the control group beagles (Group 2) received the wild-type E. canis to serve as the infection controls.

3.3.1. Prior mutant infection offered protection against Wild-type E. canis:

The mutant-infected group following wild-type infection challenge remained positive for the mutant-specific PCR throughout the 72-day assessment period when assessed using blood-derived DNAs, although the number of dogs testing positive declined with time. The Group 2 controls tested negative for the mutant-specific DNA (Table 2). All dogs in the MLAV group remained negative for wild-type E. canis throughout the 72-day monitoring period following the wild-type infection challenge. This was consistent regardless of the detection methods used: culture recovery, which was performed on days 14 and 21, or by a conventional PCR assay (Table 3). In contrast, all dogs in the infection control group (Group 2) challenged with wild-type E. canis tested positive by both PCR beginning on day 10 post-challenge and by culture recovery assessment (assessed only on days 14 and 21) (Table 3). The detected positives were significantly different for the presence of mutant and wild type DNA for both groups of dogs (Fig 5). To ensure an active infection is established in the infection control dogs and further to monitor the potential dynamics of wild-type E. canis infection, we performed qRT-PCR targeting 16S rRNA on the RNA samples recovered several days post wild-type E. canis infection challenge were assessed in all dogs, and RNA copies were estimated as in [66] (Fig 6 and Table S2). The data revealed the presence of significantly lower 16S rRNA copies in the Group 1 group throughout the assessment period (data shown in Table S2), whereas the infection control group dogs (Group 2) receiving wild-type E. canis infection had a rapid spike in the 16S rRNA copies, which peaked around day 14 post-infection challenge and declined to low levels there after which remained persisted till the study end point of assessment (Table S2). Animal body temperatures and weights remained similar for both groups throughout the study.

Table 2.

E. canis infection assessment by PCR targeting the presence of the mutant segment spanning the mutated region (Mutant-specific PCR).

	Dog ID	Days post-wildtype infection challenge*										Post-Doxycycline**
		3	7	10	14	18	21	28	35	49	72	11	32
Mutant-infected followed by wild-type infection (Group 1)	AHL (M)	+	−	−	+	+	−	+	−	−	−	−	−
	BAK (F)	+	−	+	+	+	+	+	+	−	−	−	−
	BGK (F)	−	−	−	+	+	+	+	+	−	−	−	−
	CUL (M)	+	−	−	−	+	−	−	−	−	+	−	−
	DWL (M)	+	−	+	+	+	−	+	−	+	−	−	−
Dogs receiving only wild-type infection (Group 2)	AVK (F)	−	−	−	−	−	−	−	−	−	−	−	−
	BCK (F)	−	−	−	−	−	−	−	−	−	−	−	−
	ECL (M)	−	−	−	−	−	−	−	−	−	−	−	−
	GVL (M)	−	−	−	−	−	−	−	−	−	−	−	−

^*+ and − refer to PCR positives and negatives, respectively. M and F refer to male and female dogs, respectively.

^**All dogs in these two groups were treated with doxycycline for a month after 72 days of study assessment and then assessed for the presence of infection by PCR at the two time points listed.

Table 3.

E. canis infection assessment by PCR targeting to map the presence of wild-type DNA (Wt-specific PCR).

	Dog ID	Days post-wildtype infection challenge*										Post-Doxycycline**
		3	7	10	14	18	21	18	35	49	72	11	32
Mutant infected (Group 1)	AHL (M)	−	−	−	−	−	−	−	−	−	−	−	−
	BAK (F)	−	−	−	−	−	−	−	−	−	−	−	−
	BGK (F)	−	−	−	−	−	−	−	−	−	−	−	−
	CUL (M)	−	−	−	−	−	−	−	−	−	−	−	−
	DWL (M)	−	−	−	−	−	−	−	−	−	−	−	−
Controls (Group 2)	AVK (F)	−	−	+	+/C	+	+/C	+	+	−	+	−	−
	BCK (F)	−	−	+	+/C	+	+/C	+	+	−	+	−	−
	ECL (M)	−	−	+	+/C	+	+/C	+	+	−	+	−	−
	GVL (M)	−	−	+	+/C	+	+/C	+	+	−	+	−	−

^*+ and − refer to PCR positives and negatives, respectively, and C refers to culture positives. M and F refer to male and female dogs, respectively.

^**All dogs in these two groups were treated with doxycycline for a month after 72 days of study assessment and then assessed for the presence of infection by PCR at the two time points listed.

Figure 5.

Vaccination with MLAV prevented the establishment of wild-type infection, as confirmed by PCR and culture recovery. Mutant- and wild-type-specific PCR analysis were performed on genomic DNAs extracted from blood samples collected at different time points following mutant infection and subsequent wild-type infection (Mutant-Wt), as well as after only wild-type infection (Control-Wt). The percentage of positives was used to determine differences in systemic bacterial presence between the groups. Mutant-specific PCR showing detection of the mutant DNA after wild-type infection only in the Mutant-Wt group (P < 0.0001 ****). Wild-type (Wt)-specific PCR detected wild-type-specific positives only in the blood DNAs of the control group (Control-Wt) (P < 0.0001 ****). Statistical differences between groups were determined using the one-way ANOVA with multiple comparisons.

Figure 6.

E. canis 16s rRNA copy numbers in blood-derived RNA assessed by TaqMan probe-based RT-PCR (qRT-PCR) following wild-type E. canis infection. Total RNA isolated from blood samples collected at various time points following wild-type infection from both groups of dogs was assessed by qRT-qPCR to quantify E. canis small subunit ribosomal RNA copies per 25 μl of blood (panel A). The RNA copies were significantly less in the prior mutant-infected group of dogs following the wild-type infection challenge compared with those observed following wild-type infection alone (P = 0.0065 **) or compared with those observed after the mutant infection only (described in Fig 2) (P = 0.0090 **) (panel B). Statistical significance was measured using an unpaired t-test with Welch’s correction.

3.3.2. The CBC analysis showed an increase in the number of RBC only in the MLAV group dogs after wild-type infection challenge:

We performed the CBC analysis following the subsequent wild-type infection challenge and compared the values to determine any differences between the two groups (Fig 7). Hematological abnormalities caused by E. canis infection, such as thrombocytopenia, may take months to recover, even with early clinical improvements [67, 68]. Platelet and WBC levels remained low in the mutant-infected group also following wild-type E. canis infection. However, the RBC levels improved significantly from day 10 onwards compared to the control group receiving only wild-type E. canis infection.

Figure 7.

The CBC assessment following the wild-type E. canis infection challenge in naïve dogs or in dogs previously infected with the mutant. Both groups of dogs had a platelet drop (panel A). RBC counts significantly improved in dogs that had received prior mutant infection (P = 0.0335 *) (panel B). WBC counts were lower in naïve dogs receiving wild-type E. canis infection on days 21 and 28 post-infection compared with those previously infected with the mutant (panel C). Dotted lines refer to normal canine reference values. Statistical analysis was performed using an unpaired t-test with Welch’s correction for values from day 10 to 28 when the declines were evident after wild-type infection.

3.3.3. Challenge with wild-type E. canis elicited an IgG response in all dogs:

Immune response was measured by ELISA to detect antibodies against E. canis specific IgG over the course of the study. The presence of E. canis-specific IgG antibodies was evident for both groups of dogs following wild-type E. canis infection (Fig 8). The IgG response was similar for the mutant group following wild-type infection compared to the control group receiving only wild-type E. canis infection (Fig 8).

Figure 8.

Total IgG response to E. canis antigens in dogs measured by ELISA. Plasma samples collected weekly following wild-type E. canis infection in the MLAV group and the control group were assessed by ELISA. IgG response following wild-type E. canis infection was similar in both groups.

3.3.4. The MLAV and subsequent wild-type infection stimulated the specific cytokine responses:

The panels of 11 cytokines were analyzed to determine the plasma cytokine profile following the mutant infection and after the wild-type infection challenge in the prior mutant-infected group of dogs, as well as for the wild-type infection controls (Fig 9). Animals from both groups showed detectible changes in four of the 11 cytokines assessed. Interleukin-8 (IL-8), also known as C-X-C motif chemokine ligand 8 (CXCL8), is a chemokine that recruits and activates neutrophils at inflammation sites during bacterial infections [32, 69–71]. Changes in IL-8 secretion levels of dogs post mutant infection did not significantly differ but decreased over the course of 21 days from their baseline (day 0) levels (Fig 9A). Whereas, independent of prior mutant infection or not, wild-type infection resulted in a significant rise in IL-8 starting on days 8 and 14 post infection and its levels in the prior mutant-infected dogs resulted in a significant drop on day 21 after wild-type infection. IL-12p40 is produced by immune cells, such as macrophages and dendritic cells, during intracellular bacterial infections, leading to the differentiation of naive T cells into type 1 helper T cells and stimulating the production of cytokines, including interferon gamma (IFN-γ), which help eliminate bacteria [71–75]. In contrast to IL-8, plasma IL-12p40 concentration steadily increased over time in dogs receiving the mutant as well as for the control dogs receiving the wild-type infection for the 21 days of assessment (Fig 9B). Prior mutant infection, however, caused in the IL-12p40 decline when this group of dogs received wild-type E. canis infection (Fig 9B). The stem cell factor (SCF) plays a significant role in hematopoiesis and is involved in stimulating the production of new peripheral blood cells (WBCs, RBCs, and platelets) during infection and after acute blood loss [76, 77]. Prior infection with the mutant caused no change in the SCF production and prevented its rise following wild-type E. canis infection in the mutant infection group dogs (Fig 9C). Contrary to this, its levels were significantly elevated in wild-type E. canis infection control dogs on days 14 and 21. IFN-γ is a major part of the TH1 response which is classically reported as the protective immune mechanism/response responsible for managing and eliminating infections [71, 73, 74, 78]. IFN-γ was significantly higher only after the mutant infection on day 7 and maintained higher concentration trends on days 14 and 21 (Fig 9D). Independent of the prior infection with the mutant or not, IFN-γ levels remained unchanged during the 21-day assessment period following wild-type E. canis infection challenge (Fig 9D). IFN-γ production was further assessed in PBMCs recovered from dogs and in vitro stimulated with E. canis antigens (Fig 10). We observed a trend towards higher numbers of PBMC producing IFN-γ on day 7 when measured by ELISPOT (Fig 10A) or higher IFN-γ recall responses by ELISA (Fig 10B) assays, independent of post-mutant infection or following wild-type E. canis infection challenge for this group and similarly for the wild-type infection control group dogs.

Figure 9.

Plasma cytokine levels were measured using a Luminex bead-based assay. Cytokines were assessed following infection with the mutant (Mutant), or wild-type infection in naïve dogs (Wt), or wild-type infection following the prior mutant infection (Mutant-Wt). A) Notable changes in IL-8 levels were observed among the groups at different time points after the wild-type infection challenge; IL-8 expression was significantly higher on days 3 and 14 post-wild-type E. canis infection, independent of prior mutant infection or not (P=0.0012 ** and P=0.0144 *, respectively). On day 21 post wild-type infection, mutant infection or wild-type infection following MLAV had significantly less IL-8 compared to the dogs in the wild-type infection alone group (P = 0.0239 * and P = 0.0067 **). B) IL-12p40 levels were significantly lower following the mutant infection group dogs compared to dogs receiving wild-type E. canis infection alone or following the mutant infection on day 21 (P = 0.0003 *** and P = 0.0027 **, respectively). C) SCF levels were significantly higher only in the dogs receiving wild-type E. canis infection on days 14 and 21 post-infection (P = 0.0151 * and P = 0.0317 *, respectively. D) IFN-γ levels were significantly elevated on day 7 following infection with only the mutant strain (P= 0.0322 *), but not after wild-type E. canis infection, independent of prior infection with the mutant or not. Statistical significance was measured using a two-way ANOVA.

Figure 10.

IFN-γ production in peripheral blood mononuclear cells (PBMC) collected over time. PBMCs were stimulated with E. canis antigens following infection with the mutant (Mutant), or wild-type infection in naïve dogs (Control-Wt), or wild-type infection following the prior mutant infection (Mutant-Wt). Groups 1 and 2 animals receiving wild-type infection alone or after mutant infection or following mutant infection alone induced IFN-γ production as measured by ELISPOT assay (panel A) or by ELISA (panel B). SFU indicating Spot Forming Unit.

3.3.5. Treatment with doxycycline resulted in clearance of E. canis in circulation from all infected dogs.

Ehrlichiosis is commonly treated with doxycycline [22, 44, 78], although some studies suggest that the treatment is ineffective in completely clearing the infection [46, 79]. After 72 days of infection monitoring, dogs in both Groups 1 and 2 were treated with doxycycline orally for one month. Subsequently, blood sampled from all dogs at two points post-treatment were tested by PCR to determine if the treatment was effective in clearing both mutant and wild-type E. canis (Tables 2 and 3). Neither the mutant nor the wild-type E. canis was detected in all dogs following doxycycline treatment by PCR, although we cannot rule out the possibility that the pathogen remains in dogs at levels below detection by PCR assays despite the treatment.

4. Discussion:

Genetic modifications of obligate intracellular pathogens, such as those belong to the families Anaplasmataceae and Rickettsiaceae, serve as essential tools in understanding the molecular mechanisms important for pathogenesis and to identify virulence factors, as well as in developing methods of prevention. The ability to inactivate a gene of interest in these pathogens has remained challenging, likely due to the host-dependent nature and the bacterial extensive genome reductions [80, 81]. Despite the limitations, three targeted mutations in Rickettsiaceae family pathogens were reported; one each by an allelic exchange homologous recombination, a group II intron-based targeted mutagenesis, and a CRISPR-mediated mutagenesis are reported in three clinically important pathogenic Rickettsia species: R. prowazekii [82], R. rickettsii [15], and R. parkeri [83]. Similarly, we reported targeted mutations in two Anaplasmataceae family pathogens: A. marginale and E. chaffeensis [11, 12, 47]. Using allelic exchange methods, we generated several targeted mutations in E. chaffeensis. The loss of function mutations in the phtcp genes of E. chaffeensis and A. marginale resulted in growth attenuation, suggesting that the gene is necessary for nutrient scavenging and survival. Indeed, we reported earlier that functional phtcp is required for metal ion homeostasis in E. chaffeensis [54]. As the phtcp is conserved within the Anaplasmataceae family bacteria [12, 54], we inactivated its gene ortholog from E. canis and assessed the mutant as a modified live attenuated vaccine (MLAV). Our previous targeted mutagenesis involved the insertion of an antibiotic resistance gene for a mutant generation. The present study extended the mutant generation with eliminating the need to include an antibiotic resistance gene. The mutated E. canis contained only the marker gene expressing mCherry protein, for which serial dilution coupled with gel agar-based plaque purification of the mutant provided an antibiotic-free mutant of E. canis. Eliminating the introduction of an antibiotic resistance cassette is valuable for a vaccine’s application, as it reduces the risk of antibiotic resistance development by the bacterial pathogens. Three independent experiments were performed to confirm the clonal purity and the presence of the mutation at the desired genomic location following which we tested and documented its value as a mutant modified live vaccine (supplementary Fig S2), along with the study flow chart (supplementary Fig S1)

E. canis infection leading to CME is the most prevalent tick-borne disease of dogs globally [22, 24, 26]. The disease also gained significant US national attention in late 1960s as severe outbreaks among the US military dogs resulting from E. canis infections are reported [84]. Dogs recovering from the acute and subclinical disease remain positive for the pathogen lifelong and such dogs serve as reservoirs of infection facilitating E. canis life cycle maintained in nature via tick-transmission from the indoor R. sanguineus sensu lato (s.l.) complex [27, 28, 85]. The persistently infected dogs contribute to the risk of transmission to people from infected ticks. Human infections with E. canis are reported from Venezuela, Costa Rica, Mexico, Italy, and Montenegro [30–35]. Considering the global impact of this zoonotic pathogen, it is of critical importance to have a vaccine that protects the canine host which can also reduce zoonotic infections.

A prior study reported the description of cell culture-derived attenuated strain as having offered immune protection against virulent challenge [36]. It is unclear the molecular basis for the attenuation of E. canis resulting from the continuous cell culture passaging. Moreover, it is not uncommon for a cell culture attenuated rickettsial pathogen reverting to virulence [86]. With having no vaccines to prevent CME for nearly a century after its discovery and the lack of follow up investigations in defining the cell culture-attenuated vaccine described previously [36], the phtcp mutant serving as an MLAV offers a new path forward for advancing vaccine research. MLAVs are expected to offer stronger immune protection for intracellular bacterial pathogens as they have the capability to stimulate all arms of immunity: macrophage activation and the MHC II-mediated CD4 T cell activation leading to promoting the B cell activation and long-lasting memory responses [18, 87–90].

The E. canis mutant as an MLAV prevented the establishment of wild-type E. canis infection challenge. This is similar to prior evidence of immune protection reported with the phtcp mutations in A. marginale and E. chaffeensis infections [12, 17–21]. Whereas the wild-type infection in naïve dogs resulted in persistent systemic infection which was confirmed by PCR and culture. Nonetheless, no overt clinical signs such as fever were observed in naive dogs receiving wild-type E. canis infection. This likely reflects an establishment of a subclinical infection course under the controlled experimental conditions with intravenous infection challenge. Future investigations may require an infection challenge by tick transmission. The current study documenting persistent infection with the E. canis mutant in dogs is similar to prior documentation of persistent infection in cattle with the phtcp mutant of A. marginale which also protected wild-type pathogen infection challenges [12, 19]. Opposing to these observations, a similar phtcp mutant E. chaffeensis infection in the canine host results in both clearing the mutant as well as offering immune protection in preventing the wild-type infection progression [17, 18, 20, 21]. These data imply that, despite the close genetic similarities among the three rickettsial species, the pathogens have evolved unique abilities for establishing and maintaining infections in a host.

Although all hematological parameters were similar in dogs following infection with the mutant compared to wild-type E. canis infection, wild-type E. canis infection led to a marked increase in systemic IL-8 levels on days 3, 14, and 21 in the control group, suggesting a strong inflammatory response to wild-type infection, while after mutant infection, IL-8 levels significantly decreased. These results suggest that the immune response differs between infection with the mutant and wild-type E. canis. Conversely, subsequent wild-type infection of dogs previously infected with the mutant demonstrated a shorter period of increased IL-8 that was detected on days 3 and 14 post wild-type challenge. Nevertheless, systemic IL-12p40 secretion following infection was very similar between the mutant and wild-type E. canis, with a significant increase observed on day 21 post-infection. However, the levels remained significantly lower following wild-type infection in the prior mutant-infected group. This finding suggested that the primary systemic response to E. canis infection results in high systemic levels of IL-12p40; however, this robust systemic response is not seen with subsequent infection challenge, likely due to the prevention of the wild-type infection establishment. Interestingly, even though IL-12p40 is increased after both mutant and wild-type pathogen challenges, the downstream increase in IFN-γ is only seen after mutant infection, with detectable systemic IFN-γ in the blood of dogs on day 7 post-infection. This observation illustrates a clear biological difference between mutant and wild-type E. canis on the level of host-pathogen-immune interactions, suggesting that the phtcp deletion mutant did not block systemic IFN-γ production while wild-type infection inhibited its production. The SCF levels were significantly increased in the control group dogs on days 14 and 21 post wild-type infection compared to the dogs challenged with the mutant. While the mutant infection caused a decrease in WBC, RBC and platelet values similar to wild-type pathogen challenge, wild-type infection only animals had significantly higher systemic SCF during infection phase. These data suggest that the blood cell concentrations were only maintained in the wild-type E. canis infected dogs due to a higher rate of cell replacement required to restore blood cell populations as indicated by increased SCF systemic levels. Dogs not previously exposed to the mutant as an MLAV had developed significant pancytopenia compared to MLAV dogs. These findings suggest that the mutant as a vaccine modulated the inflammatory cytokine milieu resulting in a more restrained systemic response which prevented the establishment of wild-type E. canis infection. T cell responses during intracellular bacterial infections play a predominant role in conferring protective immunity against infection than B cell responses [91, 92]. Consistent with this concept, we observed activation of T cell responses and evidence of T-cell memory, as indicated by increased IFN- γ production.

In summary, the data presented in the current study in creating the first mutation in E. canis and its application in assessing as an MLAV are novel. This study suggests that the targeted mutagenesis methods, which we initially developed for E. chaffeensis, are applicable to diverse Anaplasmataceae pathogens [11, 12, 47]. The molecular genetics development facilitated investigating a modified live vaccine in the physiologically relevant canine host, and it resulted in preventing the establishment of wild-type E. canis infection.