Identification of Anaplasma marginale adhesins for entry into Dermacentor andersoni tick cells using phage display

ABSTRACT

Anaplasma marginale is an obligate, intracellular, tick-borne bacterial pathogen that causes bovine anaplasmosis, an often severe, production-limiting disease of cattle found worldwide. Methods to control this disease are lacking, in large part due to major knowledge gaps in our understanding of the molecular underpinnings of basic host–pathogen interactions. For example, the surface proteins that serve as adhesins and, thus, likely play a role in pathogen entry into tick cells are largely unknown. To address this knowledge gap, we developed a phage display library and screened 66 A. marginale proteins for their ability to adhere to Dermacentor andersoni tick cells. From this screen, 17 candidate adhesins were identified, including OmpA and multiple members of the Msp1 family, including Msp1b, Mlp3, and Mlp4. We then measured the transcript of ompA and all members of the msp1 gene family through time, and determined that msp1b, mlp2, and mlp4 have increased transcript during tick cell infection, suggesting a possible role in host cell binding or entry. Finally, Msp1a, Msp1b, Mlp3, and OmpA were expressed as recombinant protein. When added to cultured tick cells prior to A. marginale infection, all proteins except the C-terminus of Msp1a reduced A. marginale entry by 2.2- to 4.7-fold. Except OmpA, these adhesins lack orthologs in related pathogens of humans and animals, including Anaplasma phagocytophilum and the Ehrlichia spp., thus limiting their utility in a universal tick transmission-blocking vaccine. However, this work greatly advances efforts toward developing methods to control bovine anaplasmosis and, thus, may help improve global food security.

INTRODUCTION

The family Anaplasmataceae includes a number of obligate, intracellular, tick-borne bacterial pathogens including Ehrlichia spp. and Anaplasma spp. that cause disease in many species of animals and humans (1). Anaplasma marginale is one of the most common, has a worldwide distribution, and causes bovine anaplasmosis (2). Affected animals often develop severe anemia, which leads to reduced productivity and, in some cases, death. Methods to prevent bovine anaplasmosis are limited and, in some parts of the world, include vaccination with Anaplasma centrale, a related but poorly virulent bacterium (3). Because A. centrale is a live product produced in cattle, it poses the risk of transmission of other blood-borne pathogens and, thus, its use is limited. Additionally, bovine anaplasmosis can be prevented by the administration of long-acting tetracyclines and the use of acaricides. The former is strongly discouraged due to efforts to improve antibiotic stewardship. The latter is expensive and repeated application is necessary, particularly in tropical and subtropical regions in which tick burdens are heavy for many months of the year (4). Thus, additional methods to prevent bovine anaplasmosis are needed and would improve animal health and, thus, food security worldwide.

One underexplored method for preventing bovine anaplasmosis is to block pathogen transmission. This approach was used successfully with the first licensed vaccine to prevent Lyme disease (5, 6). However, our lack of understanding of the molecular interface between the tick and the pathogen limits our ability to develop such tools targeting A. marginale. Importantly, the life cycle of A. marginale is well defined in Dermacentor spp. ticks, which are natural vectors of the pathogen (7–9). Following ingestion of the blood meal by the tick, A. marginale must first enter the midgut epithelial cells, establish a vacuole to avoid digestion along with the blood meal, and replicate (10). The pathogen then transits to the salivary glands, where it invades salivary gland cells, replicates, and is released when the tick ingests a second blood meal (11).

However, with a few exceptions, the molecules used by A. marginale for tick cell adhesion, typically the first step in cell entry, are largely unknown (12–16). Adhesins are bacterial outer membrane proteins that selectively target specific proteins or glycoproteins on the host cell surface, define tissue tropism, and may trigger pathogen entry (17–19). The first goal of this work was to identify a broad array of candidate adhesins using cultured Dermacentor andersoni cells, DAE100. Because we have limited ability to genetically manipulate A. marginale, we developed a phage display library to accomplish this goal (20). We next determined if a subset of the adhesin candidates have increased transcript during infection of tick cells, suggestive of a role in cell binding or entry. We then expressed the adhesin candidates to determine if the recombinant protein reduced A. marginale entry into tick cells through competition with A. marginale for host cell binding. Finally, we asked if these identified adhesins are conserved among the Anaplasmataceae.

In this paper, we discuss the findings from this work, the utility of phage display to identify bacterial adhesins, and the potential to identify broadly conserved adhesins for tick cells that could be used as vaccine targets to prevent tick transmission of the Anaplasmataceae.

RESULTS

Evaluation of phage display library

For phage display, we used T7 bacteriophages, which express proteins up to 1,200 aa on their outer coat. We selected all A. marginale genes for inclusion in the library that encode the following: (i) outer membrane proteins (21–24), (ii) proteins recognized by antibody in protectively immunized animals (25–27), and (iii) proteins of unknown function identified through cross-linking and mass spectrometry on the surface of A. marginale (23) (Table S1). The genes encoding the largest proteins (Am072, Am366, Am387, Am712, Am811, Am1051) were divided into two to four segments, while Msp1a, BamA, VirB11, and Am366_bp2832-6832 were absent from the library. Thus, there were 74 genes or gene segments, encoding 66 proteins included in the library. The average size of the cloned genes and gene segments was 1,422 bp with a range of 207–3,103 bp.

Sequential rounds of biopanning were used to select phages expressing A. marginale proteins that bound tick cells (Fig. S1). Empty wells were used as a negative control for binding plastic. Following biopanning, A. marginale genes in the population of recovered phages were identified using real-time PCR. The percentage of replicate wells from which a particular A. marginale gene was recovered, termed percentage recovery, was calculated.

Factors that can influence the outcome of biopanning include the size of the DNA insert, quantity of each DNA insert in the input pool, and non-specific adhesion of phage-expressed proteins to plastic wells. Thus, we first calculated the correlation between size of the insert and percentage recovery from either tick cells or empty wells following biopanning (Fig. 1A). For tick cells, there was a moderate correlation (r = –0.61, R² = 0.37) with 37% of the variance shared between insert size and percentage recovery. For the empty wells, there was a weak correlation between insert size and percentage recovery from plastic (r = –0.39, R² = 0.15).

Fig 1.

Correlations between features of the phage display library and percentage recovery of each A. marginale gene from tick cells or empty wells following biopanning. (A) There is a moderate and poor negative correlation between the size in base pairs (bp) of an insert and its percentage recovery (% recovery) from tick cells or empty wells, respectively. (B) There is a good and poor negative correlation between the amount of an insert in the input pool prior to biopanning and its percentage recovery from tick cells or empty wells, respectively, after biopanning. (C) There is a strong negative correlation between the amount of an insert in the output pool following biopanning and its percentage recovery from tick cells or empty wells. r, Pearson’s correlation coefficient.

Next, we determined the strength of correlation between the amount of each insert in the input pool and the percentage recovery from tick cells or empty wells (Fig. 1B). There was a good correlation (r = –0.70, R² = 0.50) between the amount of each insert in the input pool and its recovery from tick cells following biopanning, and a weak correlation (r = –0.32, R² = 0.10) between this variable and recovery from empty wells. Thus, both size and quantity of the insert may influence the outcome of biopanning with tick cells, but not with empty wells.

The two measures of output were the Ct value and the percentage recovery of each A. marginale gene following biopanning. There was a strong correlation between these two output values with tick cells (r = –0.87, R² = 0.76) and empty wells (r = –0.75, R² = 0.56), thus providing confidence in the methodology (Fig. 1C). To narrow down the comprehensive library to candidate adhesins that consistently bound tick cells, even if recovered in low quantities, we used percentage recovery as the basis for selection of candidate adhesins.

Identification of candidate adhesins

There were two criteria for an A. marginale gene encoding a phage-displayed protein to be selected as a candidate adhesin. First, the A. marginale gene had to be in the 75th percentile corresponding to ≥78% recovery following biopanning with tick cells (Fig. S2). Second, the percentage recovery of a particular A. marginale gene following biopanning with tick cells had to be twice that of the percentage recovery from empty wells. For analysis, the encoded proteins were divided into categories based on annotation including A. marginale multiprotein families (Msp1 and Msp2), type IV secretion system (T4SS) proteins, proteins with predicted functions or conserved domains (other), outer membrane proteins not in superfamilies (Omps), and proteins with unknown function (unknown).

Sixty-eight percent of screened proteins were eliminated as adhesins due to poor binding to tick cells (Table S2 to S7). Proportionally, the majority of those were outer membrane proteins (83%) and proteins of unknown function (82%) (Fig. 2).

Fig 2.

Summary of results from biopanning with phage-displayed A. marginale proteins. For analysis, the screened proteins were grouped into the Msp1 and Msp2 superfamilies (SF), type IV secretion system (T4SS), proteins with predicted functions or conserved domains (Other), proteins with unknown function and lacking conserved domains (Unknown), or outer membrane proteins not included in superfamilies (Omps). Except for the Msp1a SF, most proteins in each category were eliminated due to poor binding to tick cells (gray bars). Between 8% and 75% of proteins in each category were deemed adhesin candidates (blue bars). A small proportion of proteins (7%–25%) were recovered from empty wells in 100% of replicates (yellow bars).

Seventy-five percent of proteins from the Msp1 superfamily met the criteria as candidate adhesins as compared to 8%–31% for the other functional categories (Fig. 2). Although only four proteins from the Msp1 superfamily were screened, Msp1b and Mlp3 were each recovered from 100% of the wells with tick cells with no plastic binding. Mlp4 had 78% recovery from tick cells and no plastic binding. Mlp2 bound plastic in all replicates, thus, its capacity to adhere to tick cells could not be measured using phage display (Fig. 3A; Table S2).

Fig 3.

Results from biopanning for individual phage-displayed A. marginale proteins. Percentage recovery (% recovery) is the proportion of replicates from which the A. marginale gene was recovered from the phage display library following biopanning using tick cells (solid blue bars) or empty wells (striped bars). To meet the criteria as a candidate adhesin (bold type), the percentage recovery from tick cells must be ≥78% (dotted line) and more than twice the percentage recovery from empty wells. Each graph represents a group of proteins including the (A) Msp1 superfamily, (B) type IV secretion system, (C) proteins with predicted functions or conserved domains, (D) Msp2 superfamily, (E) proteins with unknown function and lacking conserved domains, and (F) outer membrane proteins not included in superfamilies. Data represent three independent experiments.

Proteins from the type IV secretion system (T4SS) were also overrepresented as candidate adhesins. Phage-displayed VirB3, VirB7, and both copies of VirB9 met the criteria as candidate adhesins (Fig. 2 and 3B; Table S3). VirB4, VirB6, VirB10, and VirD4 had low recovery from tick cells and, thus, were eliminated as candidate adhesins. Both copies of VirB2 had 100% recovery from tick cells but also had a high level of binding to plastic. Similarly, VirB8 could not be assessed as it was a high plastic binder.

Among the proteins with predicted functions or conserved domains, 64% were eliminated as candidate adhesins (Fig. 2) including PepA, Eftu, PurD, and GroEL. Am573, which has a type 2 periplasmic binding domain (Pbp type II), Aaap and Alp2 proteins involved in actin filament formation, and SucC, which functions in the tricarboxylic acid (TCA) cycle, met the criteria for inclusion as candidate adhesins (Fig. 3C; Table S4) (28, 29).

In the Msp2 superfamily, the identified candidate adhesins include Msp3, Msp2, Omp14, and Omp1 (Fig. 3D; Table S5). The remaining 75% of these proteins were eliminated as candidate adhesins (Fig. 2). Apart from Am197, all screened proteins of unknown function were eliminated or bound plastic (Fig. 2 and 3E; Table S6). Of the Omps that are not part of a multigene family, only OmpA met the criteria as a candidate adhesin for tick cells (Fig. 3F; Table S7).

In summary, 17 of the 66 screened proteins met the criteria for selection as candidate adhesins (Table 1). These candidates were overrepresented in the Msp1 superfamily with three (Msp1b, Mlp3, Mlp4) of the four screened proteins identified as candidate adhesins. The fourth protein (Mlp2) could not be screened by phage display as it bound plastic. Interestingly, none of these proteins are known to bind tick cells. Msp1a, a member of the Msp1 superfamily, was not included in the phage display library but is an adhesin for Dermacentor variabilis midguts and Ixodes scapularis cells (14, 30). Additionally, OmpA is an adhesion for I. scapularis cells and retinal endothelial cells (RF/6A) but has not been tested in a biologically relevant cell type (12). Thus, we next focused on all members of the Msp1 superfamily and OmpA.

TABLE 1.

Adhesin candidates as identified by phage display

			% recovery
Locus tag	Annotation	Insert size (bp)	Tick cells	Empty wells
Am180	Msp1b	2,220	100%	0%
Am535	Mlp3	876	100%	0%
Am536	Mlp4	936	78%	0%
Am1315	VirB9-2	840	78%	0%
Am815	VirB3	297	78%	17%
Am097	VirB9-1	813	78%	33%
Am306	VirB7	244	78%	33%
Am573	Pbp type II	207	100%	33%
Am231	SucC	1,107	89%	17%
Am880	Alp2	843	89%	17%
Am878	Aaap	1,260	89%	0%
Am1063	Msp3	2,832	100%	33%
Am1144	Msp2	1,224	100%	33%
Am075	Omp14	1,179	89%	33%
Am1139	Omp1	993	78%	17%
Am197	Hypothetical	510	100%	17%
Am854	OmpA	711	100%	17%

Differential gene expression of candidate adhesins during tick cell entry

OmpA and asp14, adhesins of Anaplasma phagocytophilum, have increased transcript during infection of mammalian cells (31, 32). To determine if this is the case for A. marginale msp1a, msp1b, mlp2-4, and ompA, we measured the relative expression of each of these genes during cell entry and days 1–5 following infection. GroEL, a chaperonin that aids in folding large proteins, was used as a non-adhesin control. Unlike in other bacteria, there is no evidence that GroEL moonlights as a surface-exposed protein in A. marginale (21, 23, 27).

During cell entry, msp1b (3.1-fold, P < 0.0001), mlp2 (2.7-fold, P < 0.0001), and mlp4 (3.3-fold, P < 0.0001) had statistically significantly increased transcript relative to 1 d postinfection (Fig. 4). There was a 1.7-fold increase in msp1a, which was not statistically significant. Through time, transcript levels of mlp3 were variable, and there was little change in ompA. There were small, non-statistically significant decreases in groEL transcript during cell entry (0.61-fold), at 4 d (0.56-fold) and 5 d (0.60-fold) postinfection. The increased msp1b, mlp2, and mlp4 transcript levels during entry provide additional evidence that these genes may have a role in binding or entry into tick cells. Although expected, an increase in transcript levels late in the infection did not occur. Because limited tick cells are available, this is likely due to overgrowth of A. marginale and death of the tick cells prior to completion of an infection cycle.

Fig 4.

Transcriptional analysis of msp1a, msp1b, mlp2-mlp4, and ompA through time. Transcript levels of msp1a, msp1b (blue bars), mlp2, mlp3, mlp4 (green bars), ompA, and groEL (orange bars) were measured during cell entry (2 h) and 1 to 5 d (1d–5d) after infection of tick cells. To calculate the relative normalized expression (2^-ΔΔCt), A. marginale rpoh served as the reference gene, and values were normalized to 1 d postinfection. Statistically significant differences were determined using ANOVA followed by Dunnett’s multiple comparisons test and a minimum of four independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Recombinant Msp1a, Msp1b, Mlp2, and OmpA compete with A. marginale for entry into tick cells

We next determined if these proteins could competitively inhibit A. marginale binding, as measured by reduced entry of A. marginale into D. andersoni tick cells. Recombinant full-length Msp1b, Mlp2, and OmpA proteins were expressed and purified (Fig. S3). Previous experiments indicate the N-terminal repeat region of Msp1a serves as the adhesin, thus, to confirm these findings, Msp1a was expressed as two segments: amino acids 1–300, which encompassed the N-terminal repeats in the St. Maries strain, and amino acids 301–623 (15). We were unable to purify Mlp3 and Mlp4 and, thus, they were excluded from the assays.

There was a statistically significant reduction in infection levels following treatment with OmpA (3.7-fold reduction, P = <0.0001), Msp1a_1-300 (4.7-fold reduction, P < 0.0001), and Msp1b (2.3-fold reduction, P = 0.0018) (Fig. 5). The buffer for Mlp2 negatively affected A. marginale entry. However, after normalization with A. marginale levels in buffer-only treated wells, treatment with Mlp2 resulted in a 2.2-fold decrease (P = 0.0018) in A. marginale entry. In contrast, treatment with Msp1a_301-623 (1.2-fold reduction, P = 0.1084) did not significantly reduce A. marginale entry. Therefore, OmpA, the N-terminus of Msp1a, Msp1b, and Mlp2 are likely adhesins that aid A. marginale binding or entry into tick cells.

Fig 5.

A. marginale levels in tick cells following treatment with recombinant candidate adhesins. Controls include Anaplasma medium only (no treatment), a buffer in which proteins were maintained, an empty vector expressing His-Patch thioredoxin, V5 epitope, and 6x-His tag in appropriate buffer (empty vector). Buffers include native binding buffer (NBB), NBB with 250 mM imidazole (Im), NBB with 500 mM Im, or PBS alone. Rpoh was used to enumerate A. marginale levels relative to tick cells (β-actin). Values were normalized to the no-treatment controls. Graphs are representative data from one of two independent experiments. ANOVA with Dunnett’s multiple comparisons test was used to determine statistical significance. For multiple comparisons, Msp1a, Msp1b, and OmpA were compared to the no-treatment controls, and Mlp2 was compared to the buffer-only treatment. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Conservation of the adhesins and candidate adhesins among the Anaplasmataceae

To assess the potential for conserved adhesins to serve as vaccine targets, we determined the percent amino acid identity between the A. marginale candidate adhesins and their respective orthologs in A. phagocytophilum, Ehrlichia canis, Ehrlichia chaffeensis, and Ehrlichia ruminantium (Fig. S4). A. phagocytophilum and the Ehrlichia spp. lack orthologs of the Msp1 protein superfamily, Msp2, Msp3, Alp2, and Aaap. The exceptions are some members of the T4SS (>65% identity) and OmpA (35%–43% identity). Additionally, Omp1 has some degree of conservation among the Anaplasmataceae (26%–44% identity) and Omp14 is conserved only with A. phagocytophilum.

DISCUSSION

In the absence of a mutant library or ability to do targeted gene knockouts, phage display enabled a screen of 66 A. marginale surface-exposed proteins and vaccine candidates for their ability to bind tick cells leading to the identification of 17 candidate adhesins. Furthermore, we determined that OmpA and multiple members of the Msp1 superfamily, including Msp1a, Msp1b, and Mlp2, have a role in A. marginale binding or entry into tick cells. Additional experiments are required to determine if the two remaining members of the Msp1 superfamily, Mlp3 and Mlp4, similarly play a role in host cell binding or entry. Based on phage display, both have the capacity to adhere to tick cells, and Mlp4, although not Mlp3, has increased transcript levels during tick cell entry.

Msp1a is a well-studied surface protein of A. marginale. The N-terminus is composed of tandem repeats frequently used to genotype A. marginale strains (33). Previous studies demonstrate that this tandem repeat region is an adhesin for tick cells (15, 30). In contrast, Msp1b was demonstrated to bind erythrocytes but not ticks cells (14, 34). However, Msp1a is not sufficient for tick cell entry as the non-tick transmissible Mississippi strain of A. marginale does not enter and replicate in Dermacentor spp. ticks but retains Msp1a repeat structure identical to tick transmissible strains (35), thus supporting the findings here that multiple members of the Msp1 protein family contribute to adhesion to tick cells. Little is known about the structural relationship of these proteins, although Msp1a and Msp1b are non-covalently and covalently linked in the outer membrane (36, 37).

Based on the concordance between the identification of candidate adhesins using phage display and the ability of the candidate adhesins, when expressed as recombinant proteins, to block A. marginale entry, the phage display screen was an effective means to identify adhesins.

The designation of a protein as a candidate adhesin following biopanning provides evidence that the displayed protein interacts with tick cell membranes, although additional experiments are required to determine if that interaction plays a direct role in host cell entry, including triggering bacterial uptake. For example, phage display identified VirB3, VirB7, and both copies of VirB9 as candidate adhesins. The structure of the A. marginale T4SS has not been resolved, although predictions based on PSORTb v.3.0.3 and the Kyoto Encyclopedia of Genes and Genomes (KEGG) indicate VirB3 forms the T4SS pilus, and VirB7 and VirB9 are surface exposed (Table S1). Thus, these proteins likely interact with tick cell membranes. In contrast, VirB10 is predicted to associate with the inner bacterial membrane. VirB6 is predicted to span the periplasm, and VirB4 and VirD4 are components of the ATPase. All four of these proteins were eliminated as candidate adhesin.

An additional caveat is that the correlation analysis identified several biases in phage display. Specifically, phage-displayed proteins encoded by smaller genes or in greater abundance in the input pool were more readily recovered from the tick cells following biopanning (Fig. 1). Thus, it is likely that some of the candidate adhesins were falsely identified. For example, four proteins with known or predicted functions, including Am573 with a Pbp type II domain, Aaap and Alp2, thought to form F-actin appendages on A. marginale containing vacuoles in bovine erythrocytes, and SucC, involved in the TCA cycle, were encoded by smaller-than-average inserts (207–1,260 bp) in the phage display library and were abundant in the input pool with a Cts < 20 (28, 29).

Additional candidate adhesins identified through phage display include Am197, an immunogenic member of the protective outer membrane preparation, Msp2 and Msp3, which play a role in immune evasion, and Omp1 and Omp14 with unknown functions. Due to non-specific binding to plastic, Msp5, RlpA, OpAg1, Omp9, VirB8, and Am529 could not be screened. Additional experiments are needed to determine if any of these proteins are adhesins or play a role in tick cell binding or entry.

The adhesins characterized in this work primarily including the Msp1 superfamily, lack orthologs in other members of the Anaplasmataceae, thus excluding these proteins as candidates for a broadly applicable transmission-blocking vaccine. However, OmpA and other proteins identified as candidate adhesins, including most members of the T4SS, Omp1, and Am197, have some conservation among the Anaplasmataceae and, thus, may serve as key vaccine targets. Additionally, this work advances efforts toward developing a vaccine to control bovine anaplasmosis, which ultimately would improve global food security.

MATERIALS AND METHODS

A. marginale strains, culture conditions, and cryopreservation

The St. Maries strain of A. marginale was used in all experiments (38). Embryonic D. andersoni cells (DAE100) were grown in L15B complete medium at 34°C (39). A. marginale was grown in DAE100 cells using Anaplasma medium composed of L15B complete medium buffered with HEPES (MilliporeSigma, Burlington, VT, USA) and NaHCO₃ (MilliporeSigma, Burlington, VT, USA) (40, 41). A. marginale was isolated from DAE100 cells, enumerated using real-time PCR, aliquoted, and cryopreserved in sucrose-phosphate-glutamate buffer (SPG) at −80°C and as previously described (42). All experiments were carried out in CELLSTAR 24-well plates (Greiner Bio-One, Monroe, NC, USA) with 1 × 10⁵ DAE100 cells per well.

PCR amplification of gene targets for phage display library

A. marginale genomic DNA was extracted from bovine blood using DNeasy Blood and Tissue kit (Qiagen, Redwood City, CA, USA). The gene targets were PCR amplified using gene-specific oligonucleotides (Eurofins Genomics, Louisville, KY, USA) designed with SnapGene 5.0 (Dotmatics, Boston, MA, USA). Am072, Am366, Am387, Am712, Am811, and Am1051 are greater than 3.6 Kb, the upper limit for insertion length for the T7 vector. Thus, these genes were amplified and cloned as two to four segments (Table S8). To ensure directional cloning, an EcoR1 and a HindIII restriction site were added to the 5’ and 3’ ends of all amplicons, respectively. Green GoTaq MasterMix (Promega, Madison, WI, USA) and the following parameters were used for amplification: denaturation at 95°C for 2 min, then 35 cycles of 95°C for 30 s, 58°C for 30 s, 72°C for 3.5 min with a final extension of 72°C for 5 min.

Phage library construction

The T7Select Phage Display System and T7Select10-3 cloning kit (Novagen Biotech, Great Neck, NY, USA) were used for phage display library construction and biopanning following the manufacturer’s instructions. The T7Select10-3b vector was digested with EcoR1 and HindIII (New England Biolabs, Ipswich, MA, USA) and was dephosphorylated using Quick CIP (New England Biolabs, Ipswich, MA, USA). Following heat inactivation of the enzymes, the digested plasmid was cleaned and concentrated using Zymo gDNA clean and concentrator kit (Zymo Research, Irvine, CA, USA). T4 ligase, ligase buffer (New England Biolabs, Ipswich, MA, USA), 0.01 pmol of vector arms, and 0.10 pmol of insert DNA were used for ligation at 16°C overnight. The recombinant phages DNA was packaged in vitro using the manufacturer’s instructions (Novagen Biotech, Great Neck, NY, USA) and was amplified by inoculation into log phase BLT5403 or BLT5615 Escherichia coli grown in lysogeny broth (LB) supplemented with 50 μg/mL carbenicillin at 37°C, shaking at 225 rpm until the E. coli lysed. Following lysis, the cultures were clarified by centrifugation at 8,000 × g for 10 min.

To verify the presence of inserts in the phage library, DNA was extracted from clarified phage lysates using a phage DNA isolation kit (Norgen Biotek Corp, Thorold, ON, Canada) following the manufacturer’s instructions. PCR using Green GoTaq MasterMix (Promega, Madison, WI, USA), the cycling conditions listed above, and a combination of T7 Up and Down primers and gene-specific primers were used (Table S8).

Plaque assays

To quantify phages, lysates were serially diluted in LB medium, and 100 µL of each dilution was mixed with 250 µL of fresh overnight culture of BLT5403 and 3 mL of 45°C top agarose. The mixture was poured onto LB agar plates supplemented with 50 μg/mL carbenicillin and incubated at room temperature overnight. Plaques were counted, and plaque forming units (PFU) per milliliter were calculated.

Biopanning

Four rounds of biopanning were conducted in each experiment using triplicate wells with DAE100 cells or empty wells (Fig. S1). The phages in the input libraries were quantitated using plaque assays. The input pool was diluted in L15B complete medium for a multiplicity of infection (MOI) of 100:1. Following addition of the phages to the DAE100 cells, the plates were spun at 200 × g for 5 min and were incubated at 34°C for 2 h. The wells were then washed with 500 mL sterile Tris-buffered saline with 0.1% Tween 20 (TBST) (MilliporeSigma, Burlington, MA, USA) 10 times for 1 min each to remove unbound phages.

To amplify bound phages for the next round of biopanning or for detection via real-time PCR, the washed wells were resuspended in 1 mL of LB and were added to 4 mL of log phase BLT5403 or BLT5615. Cultures were grown by shaking at 37°C for 4 h or until the E. coli lysed. The lysate was clarified and prepared for inoculation into DAE100 cells as described above. Alternatively, DNA for real-time PCR was extracted using a phage DNA isolation kit (Norgen Biotek Corp). Three independent experiments were performed using tick cells. Two independent experiments were carried out using empty wells to identify phage-expressed proteins that non-specifically bound the plastic wells.

Real-time PCR

For detection of A. marginale inserts following biopanning, phage DNA was extracted as described above, and Ct values for each insert were measured using real-time PCR and gene-specific oligonucleotides designed using SnapGene 5.0 (Dotmatics, Boston, MA, USA) (Table S8). Efficiency and specificity of oligonucleotides were tested using A. marginale and D. andersoni genomic DNA as positive and negative control templates, respectively. Amplification was carried out using Biorad iQ SYRB Green Supermix, a CFX96 Touch Real-Time PCR Detection System Thermocycler (Bio-Rad, Hercules, CA, USA), and the following cycling conditions: denaturation at 95°C for 5 min, then 40 cycles of 95°C for 10 s, 65°C for 30 s, followed by melt curve analysis from 65°C to 95°C in 0.5 increments for 5 s/step. Amplicons with a Ct of 40 were considered negative.

A. marginale infection of DAE100 cells

For the transcriptional analysis of msp1a, msp1b, mlp2, mlp3, mlp4, OmpA, and GroEL through time, the experiment was performed four independent times with three replicate wells in each experiment. Cryopreserved, host cell-free, quantified A. marginale, as described above, was used to infect cells using published methods with the following minor changes (42). Following a rapid thaw, A. marginale was added to each well at an MOI of 1:100 and was incubated for 2 h at 34°C in L15B complete medium to allow for entry into host cells. Cells for this time point (2 h) were then harvested.

Medium and free bacteria were removed from the remaining wells and were replaced with 50 μg/mL gentamicin in Anaplasma medium for 1 h to kill extracellular bacteria. Gentamicin was removed and Anaplasma medium was added to the cells. Plates were incubated in a sealed BD Campy Container with a GasPak (Scientific Equipment Company, Aston, PA, USA). Cells were harvested every 24 h for 5 d.

For the assays using recombinant adhesin candidates, each experiment was performed two independent times with six replicate wells per experiment. The results of one replicate are represented in the graphs. Two micromolar recombinant fusion proteins or protein expressed from the empty vector was added to the DAE100 cells and was incubated for 1 h. Controls included DAE100 cells treated only with the buffer used for each recombinant protein, and DAE100 cells maintained with Anaplasma medium alone (untreated). Host cell-free A. marginale cryopreserved in SPG, as described above, was then added at an MOI of 50:1 in the continued presence of proteins or to the control wells. Proteins and unattached bacteria were removed after an hour, and the cells were treated with gentamicin as described above and were harvested 48 h later.

Protein expression

A. marginale adhesin candidates were amplified for recombinant protein expression using oligonucleotides listed in Table S9 and Platinum Pfx DNA polymerase (Invitrogen, Carlsbad, CA, USA). Amplicons were cloned into pBAD202/D-TOPO vector and were transformed into One Shot TOP10 chemically competent E. coli cells (Invitrogen, Carlsbad, CA, USA) following the manufacturer’s protocol, thus introducing a His-patch thioredoxin leader at the N-terminal (13 kDa) and a V5 epitope plus a 6x His-tag at the C-terminus (3 kDa) of the proteins. Sequence and orientation of inserts were verified by Sanger sequencing (Eurofins Genomics, Louisville, KY, USA).

Cultures of the transformants were grown in 500-mL LB medium containing 50 µg/mL kanamycin at 37°C with shaking at 220 rpm. Expression conditions are listed in Table S10.

For the first replicate, bacterial pellets were lysed with 20-mL lysis buffer (50 mM NaH₂PO₄ [pH 7.4], 500 mM NaCl, 5% glycerol, 1% Triton X-100, 2 mM β-mercaptoethanol) plus lysozyme (1 mg/mL) and cOmplete Mini EDTA-free Protease Inhibitor Cocktail (Roche, Indianapolis, IN, USA) and were sonicated. Lysates were incubated with 4 mL HisPur Ni-NTA Resin (Pierce Biotechnology, Rockford, IL, USA) for 2 h or overnight at 4°C with gentle end-over-end mixing. The resin was then harvested by centrifugation (500 × g, 2 min, 4°C) and was washed five to six times with six bed volumes of wash buffer (50 mM NaH₂PO₄ [pH 7.4], 500 mM NaCl, 20 mM imidazole). Proteins were eluted with elution buffer (50 mM NaH₂PO₄ [pH 7.4], 500 mM NaCl, 250 mM Imidazole) in 2-mL fractions four times. Select pooled fractions were concentrated using Amicon Ultra-4 Centrifugal Filter Units (10,000 MWCO; MilliporeSigma, Burlington, MA, USA). Endotoxins were removed using Pierce High-Capacity Endotoxin Removal Resin (Pierce Biotechnology, Rockford, IL, USA) following the manufacturer’s instructions. Proteins were maintained in elution buffer.

For the second replicate, the protein expression remained the same as in Table S9, with the following minor adjustments. Triton X-100 was removed from the lysis buffer for Msp1a_1-300, Mlp2, and OmpA. Lysis buffer for Msp1b and the empty vector lacked Triton X-100, β-mercaptoethanol, and glycerol.

Lysates were sonicated, clarified, passed through a 0.2-µM filter, and loaded onto an AKTA Pure 150 (Cytiva, Marlborough, MA, USA). The samples were purified using HisTrap High-Performance columns (Cytiva, Marlborough, MA, USA) with binding buffer (50 mM NaH₂PO₄ [pH 7.4], 500 mM NaCl) and were eluted with a step gradient of imidazole created using binding buffer and elution buffer (50 mM NaH₂PO₄ [pH 7.4], 500 mM NaCl, 500 mM Imidazole). Following concentration of the eluted fractions as described above, proteins were dialyzed with 10 MWCO Slide-ALyzer (Thermo Fisher Scientific, Waltham, MA, USA) per manufacturer’s instructions with appropriate buffer as follows: Msp1b in PBS, OmpA and Msp1a_1-300 in binding buffer, and Mlp2 in elution buffer. Expression of the empty control vector (His-patch thioredoxin, V5 epitope, and 6x His-tag) was repeated for each condition. Endotoxins were removed as previously described (Fig. S3).

At each step of purification, samples were analyzed via SDS-PAGE and Western blotting to ensure protein expression. Protein concentrations were determined using Bradford Assay (Thermo Fisher Scientific, Waltham, MA, USA), and that of the final protein preparations was adjusted to 20 µM (10x stock) with the appropriate buffer before pre-treating DAE100 cells.

Quantitative reverse transcription real-time PCR (RT-qPCR)

RNA was extracted using RNeasy Mini Kit (Qiagen, Valencia, CA, USA) (transcript analysis) or Quick-RNA Microprep kit (Zymo Research, Irvine, CA, USA) (A. marginale levels) following the manufacturer’s instructions. Extracted RNA was subjected to DNAse treatment twice using TURBO DNase (Invitrogen, Carlsbad, CA, USA) and was cleaned using an RNA Clean and Concentrator-5 kit (Zymo Research, Irvine, CA, USA). RNA concentrations were measured using a NanoDrop ND-1000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA) and were converted to cDNA using a Verso cDNA Synthesis kit (Thermo Fisher Scientific, Waltham, MA, USA) following the manufacturer’s protocol.

RT-qPCR was performed using PerfeCTa SYBR Green FastMix (Quantabio, Beverly, MA, USA), forward and reverse oligonucleotides (Table S8), and performed in a CFX96 Touch Real-Time PCR Detection System (Bio-Rad, Hercules, CA, USA) using the following thermocycling conditions: denaturation at 95°C for 30 s, then 40 cycles of 95°C for 5 s, 60°C for 15 s, followed by a melt curve analysis from 65°C to 95°C in 0.5°C increments for 5 s/step.

Data analysis

GraphPad Prism v.10.0.3 (Dotmatics, Boston, MA, USA) was used for all statistical analyses. To establish criteria for the screened proteins to be considered a candidate adhesin, we calculated a frequency distribution of the percentage recovery from tick cells, excluding the proteins recovered from empty wells in 100% of the replicates, deemed high plastic binders (Fig. S2).

For transcript analysis, Excel was used to calculate 2^−∆∆CT using A. marginale rpoH as the reference gene. Values were normalized to 1 d postinfection. A one-way analysis of variance (ANOVA) followed by Dunnett’s multiple comparisons test was used to determine if there were statistically significant differences in relative normalized expression of msp1a, msp1b, mlp2, mlp3, mlp4, and GroEL through time.

To determine if pretreatment with recombinant proteins reduced A. marginale entry, 2^−∆∆CT was calculated using Excel. RpoH was used to enumerate A. marginale, and DAE100 β-actin served as the refence gene. Values were normalized to untreated wells (42). A one-way ANOVA followed by a Dunnett’s multiple comparisons test was used to determine if differences in A. marginale levels were statistically significantly different in wells pretreated with recombinant adhesin candidates as compared to controls.

DATA AVAILABILITY

All biopanning data are available in Tables S1 to S6.

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/iai.00540-23.

Supplemental figures. iai.00540-23-s0001.pdf.

Figures S1 to S4.

Table S1. iai.00540-23-s0002.docx.

Anaplasma marginale proteins selected to include in phage display library.

Tables S2 to S7. iai.00540-23-s0003.xlsx.

All detailed biopanning results.

Table S8. iai.00540-23-s0004.xlsx.

Oligonucleotides for construction of phage display library and real-time PCR.

Tables S9 and S10. iai.00540-23-s0005.xlsx.

Oligonucleotides for cloning and induction conditions for protein expression.

ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.