Targeted mutagenesis in Anaplasma marginale to define virulence and vaccine development against bovine anaplasmosis

Tick-borne Anaplasma species are obligate, intracellular, bacterial pathogens that cause important diseases globally in people, agricultural animals, and dogs. Targeted mutagenesis methods are yet to be developed to define genes essential for these pathogens. In addition, vaccines conferring protection against diseases caused by Anaplasma species are not available. Here, we describe a targeted mutagenesis method for deletion of the phage head-to-tail connector protein (phtcp) gene in Anaplasma marginale. The mutant did not cause disease and exhibited attenuated growth in its natural host (cattle). We then assessed its ability to confer protection against wild-type A. marginale infection challenge. Additionally, we compared vaccine protection with the mutant to that of whole cell A. marginale inactivated antigens as a vaccine (WCAV) candidate. Upon infection challenge, non-vaccinated control cattle developed severe disease, with an average 57% drop in packed cell volume (PCV) between days 26–31 post infection, an 11% peak in erythrocytic infection, and apparent anisocytosis. Conversely, following challenge, all animals receiving the live mutant did not develop clinical signs or anemia, or erythrocyte infection. In contrast, the WCAV vaccinees developed similar disease as the non-vaccinees following A. marginale infection, though the peak erythrocyte infection reduced to 6% and the PCV dropped 43%. This is the first study describing targeted mutagenesis and its application in determining in vivo virulence and vaccine development for an Anaplasma species pathogen. This study will pave the way for similar research in related Anaplasma pathogens impacting multiple hosts.


Introduction
Over the last few decades, several new, emerging and reemerging tick-borne rickettsial diseases have been identified as a major public health concern in the USA and globally [1][2][3][4][5]. Similarly, tick-borne diseases are responsible for major economic losses in food animal production throughout the world [6,7]. The much-needed targeted mutagenesis methods are yet to be fully developed and are of great importance for defining genes essential for several obligate intracellular bacterial pathogens, including tick transmitted Anaplasmataceae pathogens [8][9][10][11]. Notably, the lack of targeted mutagenesis methods is regarded as a major obstacle for advancing research in defining the pathogenesis of obligate intracellular bacteria and developing effective interventions, including vaccines [11]. Mutagenesis methods can aid in developing efficacious vaccines against such pathogenic bacteria [11][12][13][14][15]. We recently reached a milestone in establishing targeted genetic manipulation methods by developing successful allelic exchange-based targeted mutagenesis in Ehrlichia chaffeensis, a member of the Anaplasmataceae; the method aided in creating mutations in several genes and supported both disruption and restoration of a gene function [14]. Similar research, however, is yet to be advanced in other related rickettsial pathogens, including the genera Anaplasma and Neorickettsia. We reasoned that the targeted mutagenesis methods we developed for E. chaffeensis could be adapted to other Anaplasmataceae pathogens, including Anaplasma species.
In our recent studies, the functional disruption in the gene encoding the membrane-bound phage head-to-tail connector protein (phtcp) of E. chaffeensis (gene tag # ECH_0660) causes rapid pathogen clearance from a host, while inducing sufficient immune response to confer protection against wild-type infection challenge by both intravenous inoculation and by tick transmission [16,17]. Our recent data also suggest that the functional disruption in the E. chaffeensis phtcp gene is likely detrimental in altering the pathogen's ability to obtain metal ions to support its growth within a phagosome of infected host macrophages [18]. Since the phtcp gene orthologs are well-conserved in all known Anaplasma and Ehrlichia species [18], we reasoned that targeted mutagenesis disrupting the function of phtcp orthologs in other related Anaplasmataceae pathogens could facilitate establishing mutagenesis methods for application in developing live attenuated vaccines.
Anaplasma marginale infects bovine erythrocytes and causes diverse clinical signs, including persistent high fever, anemia, icterus, weight loss, abortion in pregnant cows, lowered milk production in dairy cattle, and reduced meat production in beef cattle. A. marginale is also reported to cause high mortalities in beef and dairy cattle globally [6,19,20]. Despite the estimated major global economic losses due to bovine anaplasmosis, amounting to billions of US dollars annually, efficacious and safe vaccines for worldwide use in controlling the disease are currently not available [6,21,22]. Although a heterologous live A. centrale blood vaccine is used as a means of reducing disease severity caused by A. marginale in parts of sub-Saharan Africa, Israel, Uruguay, and Australia, it is not in use in parts of the world, including Europe and North America as it may result in introducing other blood-borne pathogens [6,23]. Whole cell inactivated vaccines have been reported with limited efficacy [24,25]. While a similar experimental killed vaccine is marketed by the Louisiana State University Agricultural Center, there are no scientific reports describing its production and efficacy. Additionally, development of subunit vaccines against bovine anaplasmosis has had limited success [6,[26][27][28][29][30][31][32][33][34]. A recent study reported the feasibility of developing a live attenuated vaccine using an A. marginale strain with a random insertion mutation, though the work has yet to progress beyond the initial description [35]. Given the insufficiencies of the current control options, coupled with the high economic burden associated with bovine anaplasmosis, investigations focused on developing alternative vaccines for control of this disease remain a high-priority goal.
In the current study, we report the development of a targeted mutagenesis method for an Anaplasma species; we deleted the phtcp gene from the A. marginale genome and demonstrated that cattle infected with this mutant (MLAV) exhibited no clinical disease and the mutant had an in vivo growth defect. Further, we present data showing that prior infection with the mutant offered protection against virulent disease caused by the wild-type A. marginale and kept erythrocyte infection below detectable levels, as assessed by light microscopy. We also tested an A. marginale whole cell antigen-based inactivated vaccine (WCAV) to compare its efficacy with that of the MLAV and report that it failed to prevent clinical disease following wild-type infection challenge. The current study represents the first description of targeted mutagenesis in an Anaplasma species and illustrates its application in defining a gene essential for in vivo bacterial growth, making advancement towards developing a live attenuated vaccine to protect cattle from severe anaplasmosis.

Construction of the homologous recombination cassette for use in the A. marginale phtcp gene deletion
Targeted mutagenesis methods that we reported previously for E. chaffeensis [14] were successfully adapted in this study to create a targeted deletion mutation in the A. marginale genome. To generate a gene deletion mutation, 1.1 kb each of A. marginale St. Maries strain (GenBank # CP000030) genomic DNA segments upstream and downstream to the phtcp gene (gene tag # AM581) were engineered to serve as the homology arms in the mutagenesis construct. The fragments were positioned upstream and downstream to mCherry and gentamicin resistance gene coding sequences to be transcribed from the E. chaffeensis tuf promoter [14]. The recombinant plasmid construct (AM581-KO-tuf-mCherry-Gent) was used for the homologous recombination experiments (Fig 1).
inserted mCherry sequence confirmed the presence of the deletion mutation ( Fig 2C). The AM581 gene deletion was further verified by a 3 rd PCR assay amplifying the targeted insertion region, which produced the expected larger amplicon from the mutant compared to that for wild-type ( Fig 2C). Clonal purity of the mutant bacterium was further confirmed by Southern blot analysis following digestion of the mutant genomic DNA with HindIII or EcoRV restriction enzymes and hybridized with mCherry gene segment probe. Only the predicted restriction digested fragments were observed in the DNA recovered from the mutant, but not in the genomic DNA recovered from wild-type A. marginale (Fig 2D).

Development of a virulent bovine anaplasmosis infection model
Infection progression in an 18-month-old steer receiving blood stabilate infection was assessed for 92 days by observing for clinical signs, complete blood count (CBC) analysis, and in light microscopy examination for the presence of infection in erythrocytes. A. marginale erythrocytic inclusions were observed from day 8 post infection and peaked by day 29 to 25% of RBCs containing inclusions. The animal also had occasional spikes of temperature. The PCV and hemoglobin levels also steadily declined falling to below normal range by day 28 to 22.8% (normal range: 24-37%) and 7.1 g/dL (normal range: 8-15 g/dL), respectively (not shown). Blood was collected on alternate days when the bacteremia levels were about 11%, for use in preparing stabilates for infection experiments described below.

Impact of A. marginale phtcp gene disruption on in vivo infection in erythrocytes and clinical disease progression
To test the hypothesis that gene deletion in the A. marginale phtcp gene is detrimental for the pathogen's in vivo growth, approximately 3 x 10 8 mutant organisms each were intravenously

PLOS PATHOGENS
Anaplasma targeted mutagenesis aids vaccine development (IV) inoculated into each steer. The steers were monitored daily for 28 days for clinical symptoms, changes in complete blood counts and for the presence of infected erythrocytes. The mutant-inoculated steers did not develop any clinical signs or anemia, and did not exhibit infected erythrocytes detectible by light microscopy. The animals also did not appear to have erythrocyte anisocytosis (not shown).

Assessment of attenuated mutant and whole cell inactivated antigens as vaccines
We then assessed if the prior inoculation of the A. marginale gene deletion mutant (as a modified live attenuated vaccine; MLAV) in cattle (n = 3) confers protection against disease progression resulting from virulent infection challenge. Similarly, we tested inactivated A. marginale whole cell antigens as another vaccine candidate (WCAV) (n = 3). The WCAV was injected twice subcutaneously, after combining with AddaVax oil-in-water emulsion as the adjuvant, on day 0 and day 21. The infection challenges with the virulent A. marginale St. Maries strain were performed on day 35 using the same source of blood stabilate for MLAV, WCAV, and non-vaccinated steers. The non-vaccinated infection control group steers

PLOS PATHOGENS
Anaplasma targeted mutagenesis aids vaccine development (n = 3) didn't receive either vaccine, however, they were inoculated with the adjuvant in a similar manner as the WCAV steers. All three animals in the infection control group and two from the WCAV group developed occasional spikes of fever and exhibited clinical signs such as lethargy, inappetence, pale mucous membranes, anemia, and jaundice, while the MLAV animals had no recognized clinical signs before or after infection challenge (Table 1).

Complete blood count (CBC) analysis and clinical disease.
In infection control animals, the PCV dropped to as low as 15% (a 57% drop) between days 26-31 post inoculation from 35% PCV observed in all groups of steers prior to infection challenges ( Fig 3A). Disease severity in the infection control group warranted all three animals requiring close monitoring by the attending veterinarian, with one animal euthanized prior to study completion on day 31. The WCAV vaccinees also developed similar clinical disease as in non-vaccinated infection control group animals, while the MLAV vaccinees stayed healthy with no detectible clinical signs. The PCV in the WCAV animals also dropped below normal range to 20% (a 43% drop) ( Fig 3B) with no significant difference noted between these two groups. In contrast, the MLAV animals had no notable variation in the PCV prior to infection challenge, while a transient mild drop in PCV (25%) was observed following wild-type infection challenge. The PCV values for MLAV were significantly higher than the non-vaccinated infected group animals for several days of assessment, spefically between days 20 and 35 post infection challenge (identified in Fig 3 as days [55][56][57][58][59][60][61][62][63][64][65][66][67][68][69][70]. As with PCV, RBC counts also decreased below normal ranges for the infection controls and WCAV vaccinees, while the MLAV animals had significantly higher RBC counts and the values were within the normal range, as anticipated for healthy steers ( Fig  3D-3F).

Peripheral blood smear analysis by light microscopy.
Blood sampled over several weeks from animals in all three groups was assessed for the presence of A. marginale inclusions in erythrocytes by light microscopy (Figs 4 and 5); this assay is regarded as a gold standard for monitoring clinical bovine anaplasmosis [36][37][38]. Erythrocyte anisocytosis was evident in the blood smears of both the infection controls and WCAV vaccinees about four weeks after infection challenge, but not in the MLAV vaccinees (Fig 4). The mean percent of infected erythrocytes four weeks following challenge was >10% for the infection control group steers and declined thereafter (Fig 5A). Similarly, infected erythrocytes were observed in the WCAV vaccinees, although the peak infection was lower at about 6% ( Fig 5B). In contrast, the MLAV vaccinees had no detectiable infected erythrocytes throughout the 80 days of assessment ( Fig 5C).

PLOS PATHOGENS
Anaplasma targeted mutagenesis aids vaccine development

Systemic bacterial copy numbers assessed by quantitative PCR.
We then assessed A. marginale DNA copies in blood by a more sensitive real-time quantitative PCR (qPCR) assay targeted to the 16S rDNA gene for the infection controls and WCAV vaccinees (Fig 6).  At significance level α = 0.05, One-way ANOVA with repeated measures showed significant differences for PCV (P = 0.0128 � ) and RBC (P = 0.0375 � ) levels for MLAV compared to infection controls or WCAV group animals for the days 60-70. However, no significant difference was observed for WCAV and infection control groups throughout the study period. Tukey's multiple comparisons test also showed similar significant differences in PCV (P = 0.0056 �� ) and RBC P = 0.0294 � ) for MLAV group animals compared to infection controls or WCAV animals between days 60 and 70. This method also did not yield significant difference between infection controls and WCAV animals. https://doi.org/10.1371/journal.ppat.1010540.g003

PLOS PATHOGENS
Anaplasma targeted mutagenesis aids vaccine development  The 16S rDNA TaqMan probe-based qPCR assays were performed for samples collected from all three groups animals over the study period (A-C). The MLAV group animal samples were also assessed by mCherry qPCR assay (D). Significance for days 55-65; at significance level α = 0.05, one-way ANOVA with repeated measures showed that 16S rDNA copy numbers of wild-type A. marginale bacteria were significantly different between the infection control, WCAV and MLAV groups (P = 0.0375 � ), for several days post-challenge. Additionally, Tukey's multiple comparisons test showed a significant difference between infection control and MLAV groups (P = 0.0294 � ). No significant difference was observed between infection control and WCAV, or WCAV and MLAV. https://doi.org/10.1371/journal.ppat.1010540.g006

PLOS PATHOGENS
Anaplasma targeted mutagenesis aids vaccine development For the MLAV vaccinees, qPCR was also performed targeting the mutant insertion segment; the mCherry gene ( Fig 6D). The systemic bacterial loads, which peaked by 30 days post infection challenge, were higher for the infection control group than for the WCAV vaccinees, though the differences were not significant. The average estimated bacteria at peak were about 7 x 10 5 per microliter of blood in non-vaccinated infection control group animals. In the WCAV vaccinees, the bacterial loads were also similarly observed but the peak bacterial numbers were two-thirds less (2.5 x 10 5 per microliter of blood). The MLAV group animals had undetectable levels of mutant bacteria prior to virulent infection challenge, as determined by 16S rDNA qPCR and mCherry gene qPCR assays. Post infection, a significantly lower bacterial load was detected by qPCR assays in the MLAV vaccinees compared to WCAV and non-vaccinate infection control animals (Fig 6), with peak bacterial numbers detected as <10 5 per microliter of blood in the MLAV animals. The differences between the MLAV group and the WCAV and infection control groups were most notable for days 55-65.

Xenodiagnosis.
To further assess the A. marginale infection status in animals post virulent infection challenges, a more sensitive xenodiagnosis assay was performed by allowing D. variabilis nymphal ticks to acquisition feed on all three groups of animals. Following blood feeding of nymphs and molting to adults, 20 randomly selected ticks (10 males and 10 females) per animal (60 total per group) were tested to determine the presence of A. marginale infection by performing the 16S rDNA gene qPCR. Fifty-nine out of 60 ticks from both the non-vaccinated infection group and the WCAV group animals were positive for A. marginale. Similar numbers of ticks from two MLAV group animals tested positive for two animals (animals #s, HH5 and 4506), while all 20 ticks from the third animal were negative for A. marginale infection (animal # DP324) ( Table 2). To confirm the absence of A. marginale infection in the third

PLOS PATHOGENS
Anaplasma targeted mutagenesis aids vaccine development for the wild-type and mutant, respectively. We used 20 randomly selected ticks recovered from steers; #HH5 and #4506 and 10 ticks from steer #DP324 (Fig 7). Ticks recovered from animals HH5 and 4506 were positive primarily for the mutant-specific larger amplicon, while a few ticks were positive for the smaller wild-type A. marginale-specific amplicon. Some ticks were negative for both amplicons and similarly few ticks were positive only for the wild-type amplicon. Consistant with the qPCR assays, all 10 ticks from steer DP324 were negative for the conventional PCR assay.

Gross and histopathological analysis.
Animals in all three experimental groups had no lesions identified in visceral organs by gross or light microscopic examination. Histological examination of bone marrow from the WCAV vaccinees had fewer to no identifiable megakaryocytes, contained cholesterol clefting/atrophy of adipose, edema, and one animal lacked progenitor cells (Fig 8). Contrary to this, bone marrow from both non-vaccinated infection controls and MLAV vaccinees were histopathologically unremarkable (Fig 8).

Total IgG expression assessed by enzyme-linked immunosorbent assays.
A. marginale whole cell antigen-specific IgG response was assessed by ELISA on plasma samples collected at multiple time points post vaccination and post challenge for the WCAV vaccinees and MLAV vaccinees, and post infection challenge for samples collected from non-vaccinated infection controls (Fig 9). A. marginale-specific IgG response was observed for both the WCAV and MLAV vaccinees, although the response in WCAV animals was greater and increased following booster vaccination; it was about three-fold higher compared to MLAV vaccinees. The IgG levels were significantly lower for the MLAV animal samples compared to WCAV. The non-vaccinated infection controls had the lowest IgG response.

Discussion
Targeted mutagenesis in pathogenic bacteria having the ability to inactivate a gene and also to restore a gene function, including for pathogens of the order Rickettsiales, is a heavily sought

PLOS PATHOGENS
Anaplasma targeted mutagenesis aids vaccine development after goal [11]. The order Rickettsiales has two families: Rickettsiaceae and Anaplasmataceae. To date, only two publications reported targeted mutations for Rickettsia species belonging to the Rickettsiaceae family; one targeted homologous recombination (allelic exchange) mutation in Rickettsia prowazekii reported in 2009 by Driskell et al. [39] and another targeted LtrA group II intron retrohoming system-based mutation in Rickettsia rickettsii in 2015 by Noriea et al. [9]. For the Anaplasmataceae family, we have previously reported the development of targeted mutagenesis by allelic exchsnge in creating mutations at few genomic sites in Ehrlichia chaffeensis [12,14].

Anaplasma targeted mutagenesis aids vaccine development
The current study is the first in describing a targeted mutation in an Anaplasma species. The disruption mutation in the E. chaffeensis phage head to tail connector protein (phtcp) gene (gene tag # ECH_0660) has minimal impact for its in vitro growth, while inducing attenuated growth in two different vertebrate hosts [12,13]. As the phtcp gene orthologs are broadly conserved in Anaplasma and Ehrlichia species [18], we reasoned that it serves as an ideal target for defining targeted mutagenesis in diverse Anaplasmataceae organisms. Indeed, the current study demonstrated such feasibility in A. marginale. In this study, we demonstrated that the targeted mutagenesis methods we developed for E. chaffeensis [14] are applicable for mutagenesis experiments in both Anaplasma and Ehrlichia species. Considering the current success, it is likely that the allelic exchange mutagenesis can similarly be applied to other related Anaplasmataceae organisms, including the genera Neorickettsia and Wolbachia [2,5,40].
Previous studies involving Anaplasma species reported the use of transposon mutagenesis and it has remained the only option available for creating mutations [41][42][43]. The allelic exchange-based targeted mutagenesis will aid in defining genes essential for bacterial pathogenesis in a host, defining host-pathogen interactions, and developing prevention methods for diseases caused by several emerging tick-borne rickettsial diseases. Earlier, we reported that the phtcp gene disruption of E. chaffeensis causes an in vivo growth defect, possibly resulting from limiting the bacterial zinc and iron acquisition [18]. The data presented in the current study extends our prior data reporting that the functional phtcp protein is also critical for A. marginale in vivo growth. Previously, Munderloh's group and Kocan's group reported that cattle infected with I. scapularis tick cell culture-derived A. marginale develop the clinical disease similar to that observed when animals received blood stabilate infections [44,45]. The bacterial immunogenic surface proteins' expression is also similar in A. marginale cultured in tick cell to that observed in infected erythrocytes [46]. Tick cell culture-derived A. marginale infection caused a significant drop in the PCV and a steady increase in bacteremia similar to cattle infected with blood stabilates. Contrary to these data, cattle infected with the phtcp gene deletion mutant resulted in the attenuated growth of A. marginale with no significant drop in PCV and rise in bacteremia [44,45]. While the possibility of attenuation of wild-type A. marginale during tick cell culture growth cannot be ruled out, the phtcp gene disruption mutation causing growth defect in A. marginale is consistent with our prior published data for E. chaffeensis having a similar gene mutation [12,16,17].
Bovine anaplasmosis continues to cause high economic losses throughout the world resulting from the reduced milk and meat production. Furthermore, the excessive use of tetracycline

PLOS PATHOGENS
Anaplasma targeted mutagenesis aids vaccine development derivatives added as a food additive for reducing A. marginale infections also contributes to the economic burden and increases the antibiotic resistance risk to animals and humans [47,48]. Thus, a vaccine to prevent bovine anaplasmosis will be most valuable in both containing the disease and in reducing the antibiotic prophylactic used as a food additive. Our current study demonstrated that animals receiving one dose of the phtcp gene deletion mutant as a live vaccine offers the best protection in clearing the clinical disease, improving hematological parameters and also in reducing the systemic bacterial loads. Contrary to this, the WCAV vaccinees developed clinical disease similar to the non-vaccinated animals, although some improvements were noted in reducing both the bacterial infection in erythrocytes and anemia. Inadequate protection with WCAV against infection challenge is similar to a previous study by de la Fuente et al. [25]. This study prepared a whole cell inactivated antigen vaccine using A. marginale cultured in I. scapularis tick cell cultures. Over five decades ago, Brock et al. reported that vaccine-induced protection with inactivated bacterial antigens is not sufficient in protecting animals from bovine anaplassmosis [24].
A. marginale was undetectable in MLAV vaccinees in erythrocytes when assessed by light microscopy and lacked anemia. A more sensitive qPCR assay demonstrated the presence of both the mutant and wild-type A. marginale in the blood of MLAV vaccinees although the bacterial numbers were significantly lower compared to WCAV vaccinees and non-vaccinated animals. Further, xenodiagnosis substantiated the presence of low-level circulation of the mutant and wild-type A. marginale. The infection-persistence, however, was observed in only two of the three MLAV vaccinees. The data suggest that despite the absence of clinical disease and recovery from anemia, the MLAV did not offer complete sterile immunity at least in two of the three animals assessed.
The bone marrow was normal in MLAV vaccinees, thus the vaccine also helped to keep the bone marrow healthy as in comparison to non-vaccinated animals. It is unclear why WCAV vaccinees had the loss of megakaryocytes in bone marrow, and other changes, such as adipocyte atrophy, cholesterol clefts, and edema. One possible explanation is that the vaccineinduced immunity in WCAV vaccinees may have adversely impacted animals when receiving the virulent pathogen challenge. Modified live vaccines are likely to activate all arms of the immune system and provide immunity to combat clinical disease. We reported previously that E. chaffeensis phtcp gene mutant as the live vaccine provided complete protection for dogs against virulent pathogen infection challenge by IV inoculation and by tick transmission [16,17]. The current study assessed only IV infection challenge with a homologous virulent strain of A. marginale. Live A. centrale blood stabilate vaccine is generally regarded as having the ability to confer protection against A. marginale infections by both mechanical and ticktransmission challenge [6]. Thus, it is highly plausible that the A. marginale phtcp gene deletion mutant as a live vaccine offers sufficient protection against the disease resulting from diverse A. marginale strains transmitted from ticks and by mechanical transmission. This hypothesis, however, remains to be tested. Induction of T cell responses during intracellular bacterial infections is known to play a greater role in generating protection against infection than B cell responses [49,50]. Consistent with the previous observations [51], higher antibody response observed in the WCAV vaccinates did not aid in preventing the clinical disease, neither in reducing infection in erythrocytes nor in restoring the loss of erythrocytes. Protective response against bovine anaplasmosis, therefore, is more than just the induction of the B cell response; the present study is the critical first step in furthering studies to define the immune mechanisms of protection. The study is also important in determining if MLAV offers protection against diverse A. marginale strains transmitted mechanically or from an infected tick.

Ethics statement
All experiments with cattle were performed in accordance with the Public Health Service (PHS) Policy on the Humane Care and Use of Laboratory Animals (https://olaw.nih.gov/ policies-laws/phs-policy.htm), the U.S. Department of Agriculture's (USDA) Animal Welfare Act & Regulations, and with the prior approval of the Kansas State University's Institutional Animal Care and Use Committee (IACUC). At the conclusion of the study, all animals were euthanized according to the institutional IACUC recommendations, which are consistent with the recommendations of the Panel on Euthanasia of the American Veterinary Medical Association.

In vitro cultivation of A. marginale
Both the wild-type and the phtcp mutant A. marginale St. Maries strain were propagated in Ixodes scapularis cell line culture (ISE6) at 34˚C in the absence of CO 2 as described earlier [41], except the media for culturing the mutant included gentamicin at a final concentration of 60 μg/ml.

Generation of AM581 deletion construct
Homology arms of 1.1 kb each from both 5' and 3' to the phtcp gene (gene tag # AM581) of A. marginale St. Maries strain (GenBank #: CP000030) were amplified with the PCR primer sets listed in Table 3 and using the bacterial genomic DNA as the template. The A. marginale PCR products and a previously generated plasmid construct containing the contiguous E. chaffeensis Table 3. List of oligonucleotides used in this study.

Name
Sequence (

PLOS PATHOGENS
Anaplasma targeted mutagenesis aids vaccine development tuf promoter, ORFs of mCherry, and the gentamicin resistance gene (tuf-mCherry-Gent) [14] were cloned into the pGGA plasmid vector (New England Biolabs, Ipswich, USA). The Golden Gate Assembly kit was used to assemble the fragments into the pGGA plasmid in the following order: 5' A. marginale homology arm (1.1 kb), tuf-mCherry-Gent segment (1.6 kb), and 3' A. marginale homology arm (1.1 kb). The final assembled recombinant plasmid is referred to as 'AM581-KO-tuf-mCherry-Gent'. Standard molecular cloning protocols were followed to recover the recombinant plasmid that was transformed into the DH5α strain of E. coli. The integrity of the plasmid DNA, purified from the transformed E. coli, was verified by Sanger's DNA sequencing analysis using the commercially available T7 and SP6 promoter primers (Integrated DNA Technologies, Coralville, IA, USA) annealing to the pGGA plasmid backbone. The recombinant plasmid was used as the template in a PCR to amplify the fragment containing the 5' A. marginale homology arm, the tuf-mCherry-Gent segment, and the 3' A. marginale homology arm (primers listed in Table 3). The PCR products were then purified as reported previously [14].

Generation, clonal purification, verification, and propagation of A. marginale phtcp mutant
Approximately 20 μg of the above purified amplicon from the AM581-KO-tuf-mCherry-Gent plasmid were electroporated into ISE6 tick cell culture-derived A. marginale St. Maries organisms (~3 x 10 8 ), by following our previously described method [14]. The electroporated bacteria were transferred to a cell suspension containing approximately 1 x 10 6 uninfected ISE6 tick cells and propagated at 34˚C in a T25 culture flask containing tick cell media for 24 h and then supplemented with 60 μg/ml final concentration of gentamicin. The cultures were maintained in the media with media changes once a week for the first three weeks and twice a week thereafter. The presence of mutant A. marginale expressing mCherry in cultures was monitored by fluorescence microscopy, while being maintained for several weeks in the presence of gentamicin to clear all wild-type bacteria.
To confirm the clonal purity of the mutant, three different PCR assays were performed using genomic DNA recovered from the mutant cultures; 1) forward primer specific to the inserted gentamicin gene segment and reverse primer targeted to the upstream genomic region 2) forward primer targeted to the downstream genomic region and reverse primer specific to the inserted mCherry gene segment, and 3) and forward and reverse primers targeted to the genomic regions upstream and downstream to the gene deletion-insertion mutation region (all primers were listed in Table 3). The PCR assays were performed in 25 μl reactions in 1x Q5 reaction buffer containing 2 mM MgCl 2 , 0.5 mM of each dNTP, 0.2 μM of each forward and reverse primers, 1 unit of Q5 Taq polymerase (New England Biolabs, Ipswich, MA, USA), and genomic DNA from wildtype 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 mins and a final hold at 10˚C. For the third PCR assay, the annealing temperature was changed to 70˚C. The PCR products were resolved on a 1.5% agarose gel containing ethidium bromide and visualized using a UV transilluminator. Clonal purity of the mutant was further assessed by Southern blot analysis using the mutant culture-derived genomic DNA digested with HindIII or EcoRV restriction enzymes and genomic DNA from wild-type A. marginale was similarly digested and used to serve as the control. The insertion-specific mCherry gene segment-specific DNA probe was used for detecting approximately 4.3 kb and 3.7 kb DNA fragments, respectively, only in genomic DNA recovered from the mutant cultures [12,13].

A. marginale WCAV preparation
Wild-type A. marginale St. Maries strain was cultured in tick cell line to~80% infection in 35 ml culture volume. The cultures were harvested and centrifuged 2,860 x g for 15 min at 4˚C. The

PLOS PATHOGENS
Anaplasma targeted mutagenesis aids vaccine development pellet was resuspended in 3 ml of 1 x PBS, passed through a 27 g bent needle for 10 times. The contents were centriguged at 500 x g for 3 min and the supernatant was collected and passed through a two micron filter. The filtrate was then centrifuged at 18,000 x g for 5 min, supernatant was removed and the pellet was resuspended in 1.25 ml of 1 x PBS. The recovered organisms were heat inactivated at 60˚C for 30 min. The protein concentration was estimated by the BCA protein estimation method (ThermoFisher Scientific, Carlsbad, CA, USA). Approximately 200 μg of WCA per 1 ml 1X PBS was mixed with an equal volume of oil-in-water nano-emulsion suspension adjuvant, AddaVax (Invivogen, San Diego, CA, USA), for use as subcutaneously administered vaccine (WCAV). AddaVax is widely used in vaccine studies promoting the vaccine-induced immunity and without causing adjuvant-associated adverse reactions [52,53]. It is known for inducing Th1 response which is essential for the defense against intracellular parasites.

Cattle infection and vaccine studies
Ten Holstein steers, approximately 18 months old, were obtained from an area in North Dakota reported to be free of bovine anaplasmosis (animal numbers were listed in Tables 1  and 2). To confirm no prior exposure, serum and whole blood from each animal were screened by an MSP5-based cELISA (Anaplasma Antibody Test Kit, cELISA v2;VMRD, Pullman, WA, USA) [54] and A. marginale 16S rDNA qPCR [55], respectively. The steers were housed at a vector-free animal facility at Kansas State University with food and water provided ad libitum. Steers could interact and socialize within their respective group animals. Animals were individually housed when tick studies were performed. Adequate space was also given to allow regular exercise/activity.

Infection experiments in steers
Infection experiments were performed using either the in vitro cultured mutant organisms or the wild-type A. marginale St. Maries wild-type strain blood stabilates. For mutant A. marginale infection experiments, steers received~3 x 10 8 ISE6 tick cell culture-derived mutant organisms resuspended in 2 ml of 1X PBS. The infection challenges with wild-type St. Maries strain were performed IV using 2 ml each of blood stabilate (~6 x 10 9 bacteria) (originating from the same batch) as per the previously described protocol [35]. The MLAV vaccinees were challenged with the virulent strain on day 28, while the WCAV group animals were challenged on day 35. Non-vaccinated infection control group steers received AddaVax adjuvant diluted in 1 x PBS (1:1) during the WCAV vaccination days. Prior to infection, blood stabilates were mixed with 5 ml freshly collected homologous blood plasma. Animals in MLAV, WCAV, and non-vaccinated groups received the same batch of inoculum.

Animal monitoring, CBC, and assessment of systemic A. marginale
All cattle used in the current study were monitored daily for health and behavioral changes and twice weekly for body temperature. Veterinary care for the animals was overseen by a Kansas State University veterinarian. Throughout the study, 20 ml of blood was collected in EDTA tubes each week from all animals for plasma analysis. About 2 ml of blood was similarly collected twice per week for CBC analysis, performed on a VetScan HM5 Hematology Analyzer v2.3 (Zoetis, Union City, CA, USA). A small fraction of blood also in EDTA tubes was collected every other day to prepare blood smearls for light microscopic analysis of blood smears to monitor for erythrocyte A. marginale inclusions. Blood sampled from all animals were also assessed once per week for the presence of A. marginale by 16S rDNA PCR analysis. Additionally, two ml blood samples were collected daily throughout the acute phase of infection. All blood samples were processed either immediately or stored at 4˚C for a maximum of

PLOS PATHOGENS
Anaplasma targeted mutagenesis aids vaccine development 24 h prior to performing the described analyses. DNeasy Blood and Tissue DNA isolation kit (Qiagen, Germantown, MD, USA) was used to extract total genomic DNA from a 100 μl aliquot of the whole blood samples. Extracted genomic DNA from each sample was eluted in 150 μl of elution buffer. To assess A. marginale infection status, TaqMan probe-based qPCR assays were performed targeting the 16S rDNA as in [56]. Animals receiving the mutant A. marginale strain were also tested for the mutant-specific qPCR assay targeting the mCherry gene. The assay was standardized using the primers and TaqMan probes listed in Table 3. The qPCR assays were performed in 25 μl reactions with final concentrations of 1X reaction buffer containing 0.4 mM of each dNTP, 2.4 mM MgSO 4 , 0.1 μM concentration of both forward and reverse primers and the TaqMan probes, 1 unit of Platinum Taq polymerase (ThermoFisher Scientific, Carlsbad, CA, USA), and by including 2 μl each of genomic DNA as a template. Genomic DNA extracted from the wild-type A. marginale St. Maries strain was included as the positive control for the 16S rDNA assays, while DNA from the mutant A. marginale was used as the positive control for assays targeting the mCherry gene. Negative controls included all reactants and PCR-grade water in place of DNA template. The qPCR cycling conditions for the assays were 95˚C for 3 mins, followed by 45 cycles of 94˚C for 15 s, 50˚C for 30 s and 60˚C for 1 min (signal acquisition stage). Serial dilutions of the 16S rDNA gene-and mCherry genecontaining plasmids were used in the assays to define the copy numbers of molecules in the respective test samples. The Ct values obtained by fluorescence signal detection of the serially diluted plasmid controls ranging from 10 9 to 10 1 copies were used for generating standard curves and all assays were performed in triplicate.

Xenodiagnosis of A. marginale by Dermacentor variabilis
Approximately 250 D. variabilis nymphal stage ticks were placed on all animals on day 19 post A. marginale wild-type infection challenge. Ticks were allowed to feed to repletion (~10 days) [57]. Fed ticks were carefully collected from the tick attachment cells and transferred to a humidified incubator with 14 h day light and 10 h darkness for molting to the adult stage, which took approximately 30 days. Genomic DNAs from molted ticks (equal numbers of males and females) fed on each animal were initially isolated and subjected to qPCR targeting to A. marginale 16S rDNA. The mCherry gene qPCR assays were also performed on DNAs recovered from ticks fed on the MLAV animals. Genomic DNA extractions were performed using a Qiagen DNeasy Blood and Tissue Kit (Qiagen, Germantown, MD, USA). The purified genomic DNA from each tick was recovered in 150 μl each of the elution buffer and stored at -20˚C until use. A conventional PCR assay was also performed on tick DNAs using the forward and reverse primers targeted to the genomic regions upstream and downstream to the homology arm segments used in the mutagenesis experiment (primers listed in Table 3)

Assessment for the presence of A. marginale-specific IgG production in steers by ELISA
Ninety-six well ELISA plates were coated with 10 μg/ml of host-cell free A. marginale total proteins prepared from the ISE6 tick cell-cultured organisms by incubating overnight at 4˚C. The wells were blocked with the blocking buffer (1X PBS containing 1% BSA) and incubated at 37˚C for 1 h. Plasma samples from the animals were diluted 1:200 in the blocking buffer, added to wells and incubated at 37˚C for 1 h. The plates were then washed three times with wash buffer (1X PBS containing 0.05% Tween 20). Finally, the HRP conjugated anti-bovine IgG (Invitrogen, Frederick, MD, USA) at 1:2,000 dilution was added to the wells and incubated at 37˚C for 1 h. The ELISA plates were washed three times with the wash buffer and then TMB substrate (EMD Millipore Corporation, Temecula, CA, USA) was added. After observing color development in the wells, the reactions were stopped by adding 0.1 M phosphoric acid solution (stop solution) and the absorbance at 450 nm was measured using an ELISA reader (Biotek Instruments, Winooski, VT, USA). All assays were performed in triplicate and the mean absorbance values and standard deviations were calculated.

Statistical analysis
One-way ANOVA with repeated measures and Tukey's multiple comparisons tests were performed using GraphPad Software (La Jolla, CA, USA) at significance level, α = 0.05, to assess the differences in wild-type bacteremia (as assessed by 16S qPCR and light microscopy examination of peripheral thin blood smears), PCV, RBC and IgG levels between the three groups at each time point following challenge.