Identification of CCp5 and FNPA as Novel Non-canonical Members of the CCp Protein Family in Babesia bovis

Bovine babesiosis, caused by Babesia bovis, is an economically significant tick-borne disease that imposes restrictions to livestock production worldwide. Current methods to control bovine babesiosis have severe limitations and novel approaches, including transmission-blocking vaccines, are needed. Members of the widely conserved CCp family are multidomain adhesion proteins containing LCCL motifs, which are differentially expressed on gametocytes of apicomplexans, including Babesia spp. and Plasmodium spp. While Plasmodium parasites contain 6 distinct CCp genes, only three members (CCp 1-3) were previously identified in B. bovis. In this study, we describe the identification and characterization of two novel non-canonical members of the CCp gene family in B. bovis, named CCp5 and FNPA. The genes were identified in silico by TBLASTN using P. falciparum CCp family domains as queries. Unlike CCp1-3, the B. bovis CCp5 and FNPA proteins lack the LCCL canonical domain but contain other typical multidomain adhesion motifs which are present in classical CCp proteins. In addition, the B. bovis CCp5 and FNPA are in synteny with known CCp genes in related apicomplexans. Sequence analysis of these two proteins demonstrated high sequence conservation among B. bovis different isolates. Transcription, immunoblot, and immunofluorescence analyses demonstrated expression of CCp5 and FNPA in blood and in vitro induced sexual stages of B. bovis. The FNPA, in contrast to CCp5, has a predicted transmembrane domain, suggesting that it might be expressed in the surface of sexual stage parasites. Altogether, finding of this study support FNPA as a possible target of a transmission-blocking vaccine against B. bovis.


INTRODUCTION
Babesia bovis, which is transmitted by Rhipicephalus spp. ticks, is a hemoparasite responsible for bovine babesiosis, a disease that causes enormous economic losses to the cattle industry in tropical and subtropical regions worldwide. Babesia parasites have a complex lifecycle that includes the development of asexual stages in vertebrate hosts and sexual stages inside their definitive tick vectors (1)(2)(3). Sporozoites, the infectious form of B. bovis, are introduced via tick saliva into the bovine host by infected Rhipicephalus spp., where they invade and reproduce asexually as merozoites in red blood cells (RBC). Upon feeding on infected cattle, ticks may ingest B. bovis-infected RBCs and B. bovis gametogenesis is induced while in the tick midgut lumen, leading to zygote formation (4,5). In some species of Babesia, gametocytes have been identified in host RBCs, but it is impossible to distinguish them under the light microscope (6,7). Babesia bovis acute infection in cattle results in high fever, anorexia, inappetence, and severe intravascular hemolytic anemia. In addition, B. bovis expresses proteins that facilitate cytoadhesion of infected RBCs to capillaries, causing neurological symptoms and general organ failure, leading to rapid death of cattle, especially in immunologically naïve adult (more than one-year old) animals (3,8,9).
Commonly used methods to control acute bovine babesiosis combine approaches for tick management, immunization with live, attenuated, blood based Babesia vaccines, and babesicidal drugs. Despite being relatively effective, these interventions have various disadvantages and could be improved. Production of live B. bovis vaccines is expensive and laborious, requiring serial rapid passages of virulent strains in splenectomized bovines to obtain an attenuated parasite strain. In addition, these vaccines require a cold chain for deployment, are often transmissible by competent ticks, and there is a risk for the parasites to revert to virulence (1)(2)(3)8). Thus, alternative subunit vaccines that can contribute toward effective control and eradication of bovine babesiosis are urgently needed. The current arsenal of measures against bovine babesiosis could also be greatly enhanced by the addition of transmission-blocking vaccines (TBV). Such TBV could be based on parasite sexual stage proteins, such as HAP2 (10), members of the CCp family (11) and 6cys (12)(13)(14).
CCp is highly conserved family of six multidomain adhesion proteins, containing LCCL motifs, which are differentially expressed on the surface of apicomplexan gametocytes. Previous experiments performed in the Babesia-related Plasmodium parasites demonstrated that knocking out CCp genes leads to the blocking of sexual stage development of the parasite in the mosquito vector (15)(16)(17). Three CCp genes, denominated CCp1-3, were previously identified in B. bovis and B. bigemina (11). In the current study we describe the identification of two novel genes of the B. bovis CCp family, which we named CCp5 and FNPA. The newly identified B. bovis FNPA and CCp5 proteins lack the LCCL domain characteristic of the other members of the CCp family, but do contain other functionally relevant motifs found in Plasmodium CCps (PfCCps), and their coding genes have conserved synteny with other CCp genes in related apicomplexans. Therefore, we propose CCp5 and FNPA as noncanonical members of the B. bovis CCp family. The findings of this work contribute toward a better understanding of the biology of B. bovis and support the rationale for using members of the CCp family in the development of novel TBV against B. bovis.

Babesia bovis in vitro Culture
The B. bovis S79-T3Bo strain was propagated in continuous microaerophilic stationary-phase culture, as previously described (10,14) and used for gene and protein expression analyses. Genomic DNA from the B. bovis strains T2Bo (20), Mo7 (21,22), and L17 (20) were used for gene amplification by PCR and followed by comparison of CCp5 and FNPA gene sequences.

Polymorphism and Genetic Analysis
PCR products derived from the full length CCp5 and FNPA genes from different strains of B. bovis were cloned into pCR-TOPO 2.1 vector (Invitrogen, CA, USA) and sequenced (Eurofins MWG Operon, Louisville, KY). Sequencing of PCR products was performed using the primers indicated in the Supplementary Table 1. The complete gDNA sequence for the newly identified genes were compared among four geographically distinct B. bovis strains including Texas (T2Bo) attenuated, T2Bo virulent (20), Mo7 (21,22), and L17 (Argentina) (20). Strain-specific single nucleotide polymorphisms (SNPs) were then estimated to calculate the ratio of synonymous to non-synonymous changes (12)(13)(14). SNAP (http://hcv.lanl.gov/ content/sequence/SNAP/SNAP.html) was used to investigate ω (dN/dS ratio) as follows: ω > 1 indicates positive selection, as the selection has caused some amino acid substitution; and ω < 1 indicates occurrence of purifying selection and a high degree of sequence conservation (23). The Multiple Alignment using Fast Fourier Transform (MAFFT) (24) was used for DNA sequence alignment. The Molecular Evolutionary Genetics Analysis (MegaX) (19) was used to generate phylogenetic relationship among CCp family members. Nucleotide substitutions were manually calculated.

In vitro Induction of B. bovis Sexual Stages
Babesia bovis infected-RBCs were induced to form sexual stages in vitro using 100 µM xanthurenic acid (XA) (Sigma, St. Louis, MO, USA), as previously described (10,25). Cultures were incubated up to 48 h at 27 C with 5% CO 2 , and collected at 0, 12, 24-and 48-h post-induction for RNA extraction. In addition, blood smears were prepared from induced and noninduced parasite cultures at different time points and stained with Giemsa stain to be further analyzed under light microscope for morphological changes.

RNA Isolation, CDNA Synthesis and Transcription Analysis for CCp Gene Family Members
Total RNA was extracted from induced and non-induced sexual stage B. bovis cultures. Parasites were collected in Trizol (Thermo Fisher Scientific, Waltham, MA, USA) and RNA extractions were performed using a phenol-based protocol. Two hundred ng of total RNAs were utilized for cDNA synthesis using the Superscript III TM cDNA Synthesis Kit (Thermo Fisher Scientific, Waltham, MA, USA) following the manufacturer's protocol. Synthesized cDNA was used for PCR and quantitative PCR (qPCR) to test the transcript levels of CCp5 and FNPA in B. bovis parasite. The qPCRs were performed in a CFX96? Real-Time PCR Detection System using the SsoFast TM EvaGreen R Supermix (Bio-Rad, USA). The cycling conditions consisted of an enzyme activation step of 95 • C for 30 s followed by 40 cycles of 95 • C denaturation for 5 s and annealing/extension of 60 • C for 5 s. Reactions were performed in duplicate in 20 µl using 200 nM of each primer and 2 µl of a 1/20 dilution of cDNA as template. The B. bovis gapdh gene (BBOV_II002540) was used as a reference gene and transcription level of CCp genes was normalized to blood stage (non-induced) parasites using the formula: relative expression (sample) = 2 [Cq(control)−Cq(sample)] , where time zero was used as Cq control. Supplementary Table 1 shows the primers used to evaluate the cDNA levels for CCp5 and FNPA. RT-PCR amplicons were cloned into pCR-TOPO 2.1 vector (Invitrogen, CA, USA) and sequenced (Eurofins MWG Operon, Louisville, KY). CCp5 and FNPA transcript levels were compared in blood and kinete stages using previously reported B. bovis RNAseq datasets (26).

Immunofluorescence Assays (IFA)
Non-induced and induced B. bovis cultures were collected at 0, 12, 24, and 48 h and used for IFA. Cells were washed with 1X PBS at 900xg and pellets suspended with 1X PBS containing 3% bovine serum albumin (BSA). A smear was made by taking 5 µl from each cell preparation to a glass microscope slide. The slides were air dried, fixed for 5 min in cold acetone, and blocked with 1X PBS containing 10% BSA for 30 min. Then, slides were incubated with rabbit anti-CCp5 or -FNPA sera (1:50) in 1X PBS-10% BSA for 1 h. The slides were washed 3 times with 1X PBS and incubated in 10% BSA with goat anti-rabbit IgG AlexaFluor R 555 (Thermo Fisher, Waltham, MA, USA), and then again washed 3 times with 1X PBS and mounted with a drop of Prolong TM Gold Anti-fade with 4 ′ ,6-diamidino-2-phenylindole (DAPI) (Thermo Fisher, Waltham, MA, USA) and cover slip. The slides were analyzed using a Leica Sp8-X White Light Laser point scanning confocal microscope (Leica Microsystems, Wetzlar, Germany).

Identification of Two New Non-canonical CCp Genes in B. bovis
Three CCp gene family members were previously identified in B. bovis and denominated CCp1, CCp2 and CCp3 (11). However, database searches on the B. bovis genome using TBLASTN with the P. falciparum CCp family domains as queries resulted in the identification of two additional and previously unnoticed CCp-like genes, which we termed here as CCp5 and FNPA. The CCp5 and FNPA genes were originally annotated as BBOV_III007200 and BBOV_I002720 (PA14 domain containing proteins), respectively ( Table 1) Table 2). The CCp5 and FNPA genes encode for proteins with predicted molecular weights of 117.5 and 98.9 KDa, respectively ( Table 2). Synteny maps for CCp5 and FNPA genes show conservation among several distinct piroplasm species, including B. bigemina (Figure 1). However, similar homology and synteny analysis searches failed in identifying a gene homologous to CCp4 in the B. bovis genome (Supplementary Figure 1).
Analysis of domain architectures, using two different programs, confirmed the lack of LCCL domains in the CCp5 and FNPA proteins in B. bovis, However, a fibronectin type 2 domain (FN2), a collagen-binding domain, followed by a domain (PA14), similar to anthrax protective antigen, with potential roles in carbohydrate recognition were found in identical arrangements among these two proteins (Supplementary Figure 2). These domains are known to mediate calcium dependent interactions (27). A transmembrane domain was found only in B. bovis FNPA (Table 1; Figure 2), suggesting that this protein might be inserted in the cell membrane, and the possible exposure of functionally relevant protein domains on the surface of the parasite. A phylogenetic analysis performed with the five known    CCp members of B. bovis together with similar CCp proteins of other related piroplasm species showed formation of five different clades corresponding to CCp1-3, FNPA, and CCp5 (Figure 3) Figure 3). Moreover, the calculated synonymous and non-synonymous S/N ratios with the parameter, ω (ω = dN/dS), as an indicator of potential selection pressures are shown in Table 3. In all cases, ω parameters <1 was obtained, providing no support for positive selection for the CCp5 and FNPA genes. Altogether, these observations suggest that these two genes are not likely exposed to immune selection in the vertebrate host. The full DNA sequences of the CCp5 and FNPA genes derived from these four distinct strains, and their alignment are shown in Supplementary Figures 4A,B, respectively. In addition, the CCp5 and FNPA genes are also well-conserved among different Babesia spp. and other piroplasm species (Supplementary Figure 5).

CCp5 and FNPA Genes Are Expressed in Blood and Sexually Induced Stages of B. bovis
The levels of CCp5 and FNPA transcripts were investigated by RT-PCR analysis performed on RNA extracted from cultured non-induced and in vitro sexual-stage induced B. bovis parasites. The RAP-1 gene was also similarly amplified by nested PCR as control for cDNA quality of all stages (30), and the sexual stage marker 6cysA gene (14) was amplified for sexual stage development control (Figure 4A). The qPCR transcription analysis demonstrated differential expression of CCp5 and FNPA genes in blood and sexual stages ( Figure 4B). The transcription levels of CCp5 were similar between sexually induced and blood stage parasites at 12 and 24 h post-sexual stage induction, but it significantly increased in sexually induced parasites at 48 h post induction. However, the levels of expression of the FNPA gene significantly increased at 12, 24 and 48 h post-sexual stage induction, compared to the non-induced blood stage parasites. The peak for FNPA gene expression was observed at 48 h after sexual stage induction by temperature. In addition, the B. bovis RNAseq datasets were used to compare the transcript levels of CCp5 and FNPA at the blood and kinete stages (Figure 5). Overall, these comparisons showed that the blood stage levels of CCp5 and FNPA transcript are significantly higher than the kinete stage levels (Figure 5).
Expression of CCp5 and FNPA proteins were also examined in in vitro non-induced and sexual stage induced (48 htemperature induction) parasites by immunoblot analysis using rabbit polyclonal anti-CCp5 and FNPA peptides   (Figure 6). Results showed that anti-CCp5 and anti-FNPA peptide antibodies weakly reacted with blood stage proteins of expected molecular weight of 117 and 97 KDa, respectively (Figures 6A,B). In contrast, reactivity of the monoclonal antibody BABB35 against MSA-1 was strongly evident in both, blood and sexual stage lysates ( Figure 6C). Therefore, consistent with the transcription analysis, CCp5 and FNPA protein expression were also confirmed in blood and sexual stage parasites. Immunofluorescence assays using rabbit polyclonal anti-CCp5 and FNPA peptides also revealed expression of CCp5 and FNPA proteins in in vitro non-induced (blood stage-BS) and induced (12-24-48 h-XA induction) parasites. Fluorescence signals were more intense for sexually in vitro induced forms of the parasite compared to blood stage or non-induced forms (Figure 7). Pre-immune rabbit sera did not show any IFA reactivity against blood stage or in vitro induced sexual stages (Supplementary Figure 6).

DISCUSSION
The CCp protein family, which has an important role in sporozoite development and infectivity, consists of six members with a modular structure composed of multiple protein, lipid and carbohydrate binding domains, including LCCL domains in Plasmodium parasites. However, while the Pf CCp1-Pf CCp5 proteins contain at least one LCCL domain, Pf FNPA, which lacks such domain, has also been included as a member of the Pf CCp protein family due to its structural similarity to Pf CCp5 (15)(16)(17). The FNPA acronym is derived from the domains FN2 and PA14 present in this proteins (15). Although six LCCL protein family members have been identified in P. falciparum, only genes encoding homologs of CCp1-3 proteins have been reported in B. bovis and B. bigemina so far (11). In this study, we describe the presence of two additional single gene-copy CCp-related genes (CCp5 and FNPA) in the genome of B. bovis and B. bigemina, as previously described in Plasmodium spp. (16,31). While B. bovis FNPA is structurally similar to Pf FNPA, the B. bovis, B. bigemina and T. equi CCp5 proteins contain no canonical LCCL domains, unlike Pf CCp5. However, the B. bovis CCp5 clusters with other Babesia sp. and Plasmodium parasites in the phylogenetic analysis presented in this study. This finding, together with conserved synteny, supports the denomination of this protein as B. bovis CCp5, despite the lack of the canonical LCCL domain.
Interestingly, no genes encoding for a protein equivalent to Pf CCp4 were found in the genome of Babesia parasites, either by homology searches or by synteny analysis, suggesting that this protein might play a significant role during the life cycle of Plasmodium but not for Babesia parasites. This may correlate with the observation that the CCp gene family encodes for proteins with parasite functions operating during sexual and other stages expressed in their arthropod vectors. Thus, the lack of CCp4 genes might be related to the many differences that occur among the Babesia and Plasmodium life cycles, including the fact that Babesia parasites are transovarially transmitted, and the different nature of their arthropod hosts vectors in comparison with Plasmodium spp. (ticks vs. mosquitos). Alternatively, it is also possible that other CCp genes present in Babesia spp. play similar or overlapping functional roles as the Plasmodium CCp4.
CCp orthologs are also present in the rodent malaria species Plasmodium berghei, where they were termed LAP proteins. Several studies have been carried out to determine the function of this protein family in P. berghei (15,32). Genetic crossover studies have shown that PbCCp/LAP proteins are female-specific markers (33). In addition, recent studies on GFP fusions of PbCCp3, PbCCp1 and PbCCp5 have shown protein expression in macrogametocytes and accumulation of these proteins in crystalloid organelles that form in the ookinete and persist until the early oocyst stage (34,35). Crystalloids are transient specialized structures of malaria parasites which are present in mosquito-specific ookinetes and young oocyst stages of the parasite. They appear to play an important role in protein trafficking and sporozoite transmission and could be exploited as new targets for control of malaria transmission (36). Knockout of LAP1 (CCp3) or LAP3 (CCp5) was shown to completely abolish crystalloid formation in P. berghei (37). While sporozoites are formed in Babesia parasites, crystalloids and oocysts have not been so far reported, and whether CCp proteins may play similar roles in Babesia spp. remains unknown, and more research is needed.
The CCp5 and FNPA nucleotide and amino acid sequences were found to be highly conserved among otherwise distinct B. bovis isolates, and even among other related apicomplexans. High levels of sequence conservation among geographical strains is a very important criteria for selecting effective vaccine candidate antigens. Sequence comparisons of the CCp5 and FNPA genes were also used to investigate whether immune selection pressure in the bovine host affects this gene family. The calculated ω (dN/dS) ratios for CCp5 and FNPA are <1, suggesting that these gene family members are unlikely exposed to host immune system selective forces.
In this study, we demonstrated transcription of the CCp5 and FNPA genes in both blood and in vitro sexual stages of B. bovis. We also showed that the expression of CCp5 and FNPA genes is up regulated in B. bovis sexual stages ( Figure 4B). In addition, according to the RNAseq data, the CCp5 and FNPA transcript levels in the kinete stage were found to be downregulated compared with blood stages. These results demonstrate that the expression of CCp5 and FNPA genes needs to be tightly regulated and exert different transcriptional patterns at developmental stages of B. bovis, which may reflect distinct  functional requirements. Furthermore, relatively low abundance of CCp5 and FNPA proteins were observed in both blood and sexual stages by immunoblot analysis. Consistently, Pf CCp5 and Pf FNPA have been reported to be less abundant in gametocytes than the other four Pf CCp proteins (31). In addition, Pf CCp5 also showed transcript expression in both asexual and sexual stage Plasmodium parasites (16). In this study, we demonstrated expression of CCp5 and FNPA proteins in blood and sexual stages by immunoblot and fixed IFA. Although a more intense fluorescent signal was observed in in vitro induced sexual stages than in blood stage parasites, we are limited to interpret this observation as a strong evidence of increased protein expression because sexual stage parasites tend to aggregate (38), which could result in more intense IFA signals and mislead the interpretation of the results.
Ideally, TBV candidate antigens should be widely conserved and expressed on the surface of the parasites, where they can be possible targets of antibodies. Bioinformatics analysis suggests that the CCp5 and FNPA proteins lack detectable signal peptides, but FNPA, unlike CCp5, has a transmembrane domain, suggesting that it might also be exposed on the surface of sexual stage parasites. Taking this data together, FNPA fulfills, at least in part, the criteria to be considered as a transmission blocking vaccine candidate.
Overall, the data in this study demonstrate the presence of CCp5 and FNPA genes in B. bovis as non-canonical members of the CCp family, which might play important roles for the development of the parasite stages in ticks and vertebrates. Future studies need to be focused on knock out/knock in approaches of these genes in B. bovis for functional analyses.

CONCLUSION
In this study, bioinformatics, pattern of expression, and localization analysis of CCp5 and FNPA genes belonging to the B. bovis CCp family were performed. Findings from this study demonstrated expression of CCp5 and FNPA genes and proteins in both blood and sexual stages of B. bovis, but their expression is increased in in vitro induced sexual stage parasites Furthermore, expression analysis, serology data, and the presence of a TM domain suggest that FNPA may be surface exposed during tick stages of the parasites, and thus, a possible candidate target for TBV.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article, further inquiries can be directed to the corresponding author/s.