Genetic investigation of glycosylphosphatidylinositol (GPI) anchored Bd37 orthologs in Babesia divergens group and potential use of recombinant protein for ecological survey in deer

Abstract

The Babesia divergens/B. capreoli group includes parasites with confirmed or possible zoonotic potential to cause human babesiosis. Currently, diagnostic antigen of the group has not been established. In this study, we investigated the ortholog of Bd37, a glycosylphosphatidylinositol (GPI)-anchored major merozoite surface protein of B. divergens sensu stricto, in the Asia lineage of the group. From two genomic isolates from sporozoite/sporoblast stages, three Bd37 gene variants, namely Bd37 JP-A, JP-B, and JP-C, were isolated with 62.3–64.1% amino acid sequence identity. Discriminative blood direct PCR revealed that Bd37 JP-A was encoded in all parasites infecting wild sika deer examined (n=22). While Bd37 JP-B and JP-C genes were randomly detected in 12 and 11 specimens, respectively. Sequencing of all JP-A variants revealed that the gene was polymorphic, with a low ratio of non-synonymous to synonymous substitutions (dN/dS) and that a highly polymorphic region was not related to predicted B-cell epitopes. A recombinant JP-A-based ELISA showed an overall positive rate of 13.9% in sika deer in Japan from north (Hokkaido) to south (Kyushu island) across 24 prefectures (n=360). This positive rate was twice as high as that examined by 18S rRNA-based PCR (6.6%). The geographical trends in infection rates were consistent. This study demonstrated that direct examination was informative for revealing genetic background and selecting antigen candidates. Bd37 orthologs may serve diagnostic purposes in combination with indirect fluorescence assay, which requires biological isolates.

Research and management of wildlife diseases are becoming increasingly important. Underlying this trend is the recognition that emerging zoonotic diseases are on the rise and that wildlife can be an important factor in controlling infectious diseases in humans, livestock, and companion animals. In particular, the increasing wild cervid population has presented both direct and indirect public health challenges. Expanding distribution of venison, a meat of wildlife, has the potential to cause food-borne zoonoses [7, 28]. Cervids are major blood sources for ticks and the expansion of their habitat into human living areas threatens efforts to control tick-borne infectious diseases [8, 27], even though cervids are not always reservoirs of tick-borne infectious diseases.

Piroplasmida, a tick-borne erythrocytic protozoa, includes various species with zoonotic or zoonotic potential, some of which infect cervids [16]. In Japan, we previously reported that sika deer (Cervus nippon) carried a Babesia parasite genetically closely related to but different from European (EU) B. divergens (B. divergens s.s.), Babesia sp. MO1 in the United States (US) and B. capleoli [32]. In phylogenetic analysis, these parasites assembled into a monophyletic clade (B. divergens/B. capreoli group) and were divided into distinct lineages that reflected geographical origin of the parasites (EU, US and Asia lineages). With the exception of the Japanese B. divergens (Asia lineage), the three lineages in the B. divergens/B. capreoli group appear to be zoonotic.

B. divergens s.s. (B. divergens EU lineage in this study) is the main causal agent of human babesiosis in European countries. Since the first case was reported in 1957 in a splenectomized Yugoslavian farmer, approximately 40 human cases have been reported and attributed to B. divergens infection, although not all cases were diagnosed molecularly [15]. Most of which were splenectomized, thus, clinical outcome of the patients tends to be severe to sometime death. A nationwide serological survey using whole parasite as antigen in indirect fluorescence assay (IFA) revealed that presence of antibodies against B. divergens was evident in blood donors in Austria (2.1%) [25], in tick-exposed patients in Germany (4.9%) and Belgium (33.2%) [17, 21], and in forestry workers in France (0.1%) [22], suggesting widespread subclinical infections in healthy individuals. The Babesia spp. from the United States (B. divergens US lineage, or B. divergens-like MO1) emerged in 2002 in a patient in Kentucky, who had undergone splenectomy. Since then, five severe cases including a possible transfusion transmission have been reported in the United States [20]. Recent emergence of “MO1-like” in a splenectomized patient in France (Babesia sp. FR1) [2] suggested genetic diversity and widespread distribution of the MO-1. B. capreoli (B. capreoli lineage in this study) has long been recognized as a parasite of cervids and chamois. In addition, the retrospective study in Gansu province, northwestern China, revealed parasites in undiagnosed patients [29]. The reported 18S rRNA sequence (MK256977) from the parasites is identical to those of the B. capreoli lineage from reindeer (Rangifer tarandus) in Germany (KM657248) and roe deer in France (FJ 944827).

B. divergens Asia lineage is maintained between wild sika deer (Cervus nippon) and Ixodes persulcatus in the eastern part of Japan [32,33,34]. I. persulcatus is predominantly distributed in entire Hokkaido and mountain areas in the eastern part of Japan. It is common species causing tick bites and transmitting many tick-borne diseases, including Tick-borne encephalitis [30], Lyme disease and relapsing fever [9]. Thus, the Asia lineage may be transferred to humans to some extent.

The serological examination is widely used for diagnosis, screening and epidemiological surveys. IFA by using whole parasites as antigens is the principal method for detecting antibodies against Babesial infection in human. The IFA antigen slides are prepared by spreading densely infected erythrocytes with Babesia isolates, in which antigen epitopes in the blood stage are expressed. Isolation and propagation of the parasite, either in vitro or in vivo, are essential for the IFA. However, the isolation of B. divergens Asia lineage has not succeeded.

Recently, González et al. reported the positive serological reactivity in a patient against E. coli-based recombinant protein of B. divergens Bd37 [10]. The Bd37 was initially identified as a soluble immunodominant antigen of 37 kDa, which conferred protective immunity against homologous lethal challenges with the Rouen87 human isolate [11]. The biochemical characteristics of the Bd37 include the presence of an N-terminal signal peptide, a C-terminal (glycosylphosphatidylinositol) GPI-anchoring sequence, a disulfide bond bridging the N-terminal and C-terminal parts of the protein via cysteines, and charged residues involved in salt bridges [4]. The Bd37 sequences vary among different isolates, in particular, the N-terminal region (after signal peptide cleavage site to valine 80) is the most polymorphic. Nevertheless, recombinant Bd37 from the Rouen87 conferred complete protection against lethal heterologous challenges with 5 different strains, which represented five major polymorphic groups identified by PCR-RFLP [13]. Sequence polymorphism among isolates was also observed in the Bd37 ortholog, Bcp37/41, of the B. capreoli. The sequence identity between 40.8 kDa (clone 2801 F10) and 36.8 kDa (clone 2770 F6) with 44 amino acid residues deleted is 50.3% [26]. In all cases, the Bd37 and Bcp37/41 are considered to be expressed as single molecules from their respective isolates.

The glycosylphosphatidylinositol (GPI) anchored protein group has great potential as an excellent immunodiagnostic marker, because of its high expression and essential role in parasite survival [3, 12, 18]. Since diagnostic antigens have not been developed in the B. divergens/ B. capreoli group, we hypothesized that the Bd37 orthologs could serve as candidates. We examined the presence of the Bd37 ortholog(s) in the Asia lineage genomes, directly prepared from field samples. This novel approach allowed us to detect up to three orthologs with unique compositions. Serological examination using the selected ortholog revealed that antibodies could be detected in reservoirs throughout Japan.

MATERIALS AND METHODS

DNA, RNA and cDNA preparation

The genomic isolates of B. divergens Asia lineage, IpSG13-13-1 and IpSG10 from I. persulcatus ticks (Table 1), were prepared in our previous study [32] and used for primary isolation of Bd37 gene sequence. Both ticks were collected in the eastern part of Hokkaido, Japan. Briefly, the genomic DNA was isolated from activated salivary glands of ticks, which was fed on gerbil for 3 days. For RNA extraction, the salivary glands kept in −80°C were used. Note that IpSG10 RNA was not available for this experiment since the salivary gland was not remained in freezer. Total RNA was extracted by using Isogen (Nippon Gene, Tokyo, Japan). Contamination of genomic DNA in the total RNA was assessed by nested PCR. Complementary DNA was synthesized from the total RNA using Bd37 specific primer, Bd37-TAA, which contains stop codon and upstream sequence (Supplementary Table 1) (SuperScript III reverse transcriptase, Invitrogen, Carlsbad, CA, USA). Genomic DNA of MRNK strain [33] (Table 1) was used for isolation of Bd37 sequence as a reference for the B. divergens EU lineage.

Table 1. Materials of primery genetic investigation.

Babesia divergens genetic isolate	Origin	Country	Material	Reference
IpSG13-13-1	Ixodes persulcatus (field collected), salivary glands	Japan	gDNA, mRNA	[32]
IpSG10	Ixodes persulcatus (field collected), salivary glands	Japan	gDNA	[32]
MRNK	Bos taurus (maintained in gerbil)	Ireland	gDNA	[33]

PCR cloning and sequencing

The full-length of Bd37 sequence (approx. 1 kb), except for stop codon, was amplified by Ex Taq polymerase (Takara Bio, Kusatsu, Japan) using the primers, Bd37-ATG and Bd37-TAA removed (Supplementary Table 1), according to the manufacture’s protocol. The resulting PCR amplicons were purified (QIAquick PCR Purification Kit, Qiagen, Venlo, the Netherlands) and ligated into pBAD/Thio-TOPO plasmid vector (Thermo Fisher Scientific, Waltham, MA, USA). The plasmids were transformed into TOP10 competent cells (Thermo Fisher Scientific). Transformants were screened by colony PCR using Bd37 internal primer Bd37-SQF1 (Supplementary Table 1) and vector specific primer pBAD-Reverse (Thermo Fisher Scientific). Plasmids were purified by using Spin Miniprep Kit (Qiagen). Sequencing (Eurofins Genomics, Tokyo, Japan) was performed by using vector primer, Trx-Forward and pBAD-Reverse (Thermo Fisher Scientific). As a control, pBAD/Thio-TOPO plasmid vector carrying full length of Venus (YFP variant) fluorescent protein gene was generated accordingly [31].

Analysis of Bd37 gene sequences

DNA sequences were aligned with each other and with closely related sequences available in GenBank by ClustalW implemented in BioEdit7.2.5 [14]. DnaSP version 6 [23] was used to determine the number of segregating sites (S), number of haplotypes (H), haplotype diversity (Hd) and observed nucleotide diversity per site (π). The dN/dS (ratio of non-synonymous substitution / synonymous substitution) value was used as measure of evolutionary pressures on protein-coding regions by using Nei-Gojobori method with Jukes and Cantor correction. A ratio dN/dS >1 results when changes in the protein sequence are favored by natural selection (evidence of positive selection), while a ratio <1 is expected if natural selection suppresses protein changes (evidence of negative selection). A dN/dS ratio equal to 1 represents a situation of neutral evolution. A two tailed Z-test was performed on the difference between dN and dS by MEGA-X 10.0.5 [19].

Analysis of Bd37 protein sequence

The putative protein sequences were aligned with each other and with closely related sequences available in GenBank by ClustalW implemented in BioEdit7.2.5. Phylogenetic analysis was performed by the neighbor-joining (NJ) method with 1,000 bootstrap replicates in MEGA-X 10.0.5. Putative signal peptide and GPI-anchor were searched by SignalP-4.1 (http://www.cbs.dtu.dk/services/SignalP/), Phobius (http://phobius.sbc.su.se/) and PredGPI (http://gpcr.biocomp.unibo.it/predgpi/pred.htm).

Discriminative PCR

To detect and discriminate the Bd37 gene sequence by the sequence types (JP-A, JP-B, JP-C and EU-A1 and EU-A2) (Table 2), PCR primers specific for each type were designed. Primers and their specificity were listed in Supplementary Table 1. It is noted that the specific primer JP-B/EU-A2 F2 anneals to both JP-B and EU-A2, as appropriate site could not be identified. The specificity of the designed primers was examined by using plasmid clones carrying either of JP-A, JP-B, JP-C, EU-A1 or EU-A2 type sequence. In this test, the plasmids of 1 × 10⁸ copies in a PCR mixture were used. For examining sensitivity of the nested PCR, the 1st PCR was performed using universal primers Bd37-ATG and Bd37-TAA (Supplementary Table 1) and 1 × 10¹, 10² or 10³ copies of plasmids in a PCR mixture (20 μL). One microliter of the product was used for 2nd PCR in which either of the specific primers (forward) and universal primer Bd37-R1 (reverse) (Supplementary Table 1) were used. Ex Taq polymerase (Takara Bio) was used according to the manufacturer’s instructions.

Table 2. Number of plasmid clones carrying Bd37 sequence types.

Isolate	Material	No. sequenced	Bd37 sequence type
			JP-A			JP-B	JP-C	EU-A
			1	2	3			1	2
IpSG13-13-1	Genome	10	5	2	0	3	0	0	0
	cDNA	10	3	0	0	7	0	0	0
IpSG10	Genome	10	0	3	4	0	3	0	0
MRNK	Genome	10	0	0	0	0	0	6	4
Total		40	8	5	4	10	3	6	4

Deer and tick samples used for the discriminative PCR

Hunting, sample collection and screening of B. divergens infection were performed in our previous study [33, 34]. Whole blood was taken from wild deer to heparinized tubes. Erythrocytes were separated by centrifugation and kept at −80°C until use. B. divergens genomic DNAs extracted from host seeking I. persulcatus ticks (whole body) or additional salivary glands, IpSG 17-6-3, IpSG 14-12-2 were prepared in our previous study [32].

Blood direct PCR and sequencing

To screen the presence of Bd37 sequence in B. divergens-positive wild deer, frozen erythrocytes were thawed and directly used as a template (Phusion Blood direct PCR kit, Thermo Fisher Scientific) as described previously [34]. Direct sequencing was performed by using inner primers Bd37-SQF1 and Bd37-SQR1 (Supplementary Table 1). Quality of base-calling was checked by Quality score in Sequencher (GeneCodes, Ann Arbor, MI, USA), then chromatograph was manually checked in case of low-quality score (below 40%). When PCR amplicons were directly sequenced and overlapped peaks in chromatograph were observed, the corresponding residues were collected manually according to IUPAC nucleotide code. The threshold of detecting the minor population was 20%, which was calculated by comparing chromatograph of direct sequencing of PCR amplicon and sequencing of plasmids of PCR clones.

Expression of recombinant Bd37 proteins

Recombinant Bd37 (rBd37) protein was expressed as a thioredoxin fusion protein from the pBAD/Thio-TOPO vector (Thermo Fisher Scientific), which encodes thioredoxin and a histidine-tag at the N-terminus and C-terminus of multiple cloning sites, respectively. These additional sequences increase the size of expressed recombinant protein by 13 kDa and 3 kDa, respectively. Protein expression was induced with 0.2–2.0% of arabinose at 25°C for 6 hr. Recombinant Bd37 and Venus (control) proteins were purified by using ProBond Purification system (Thermo Fisher Scientific) according to the manufacturer’s instructions. Briefly, cells were harvested by centrifugation and lysed by sonication in native buffer including Complete Mini (Roche, Mannheim, Germany) and Lysozyme (Sigma Aldrich, Tokyo, Japan) (50 mg/mL) on ice. Recombinant proteins were purified by nickel-chelating resin and eluted under native conditions. The eluted fractions were concentrated by filter (Amicon Ultracel-30K, Merk Millipore, Burlington, MA, USA) and dialyzed against TNE buffer (20 mM Tris-HCl pH 8.0, 0.5 M NaCl, 0.1 mM EDTA) at 4°C. Protein concentration was measured using Pierce BCA Protein Assay Kit (Thermo Fisher Scientific).

Anti-sera against recombinant Bd37 protein

The purified rBd37 was dialyzed against PBS (−). Rabbits were injected with 200 μg of the rBd37 or rVenus in an equal volume of TiterMax Gold (TiterMax USA, Inc., Norcross, GA, USA) subcutaneously. Three weeks following primary immunization, the same amount of antigen emulsified in TiterMax Gold was injected as a booster. Whole blood was harvested after 7 weeks post initial immunization and serum was separated by centrifugation.

SDS-PAGE and western blot

One hundred micrograms of the purified rBd37 or rVenus was loaded onto an acrylamide gel (e-PAGEL 10–20% gradient, ATTO, Tokyo, Japan) with a protein marker (Blue star, Nippon gene). After electrophoresis, separated proteins on the gel were stained with Coomassie Blue/ethanol/acetic acid and destained by methanol/acetic acid. For western blot, the separated proteins were transferred onto Immobilon-P PVDF membrane (Merck Millipore). The membrane was blocked with 10% skim milk in Tris Buffered Saline, with Tween 20, pH 8.0, and reacted with mouse anti-thioredxin antibodies (MBL, Nagoya, Japan). Alkaline phosphatase conjugated goat anti-mouse IgG antibody (Sigma Aldrich) was used as a secondary antibody and proteins on the membrane were visualized using BCIP/NBT substrate (Sigma Aldrich). When deer plasma specimens were used as primary antibody, the plasma was diluted 1:50 with ImmunoBlock (DS Pharma Biomedical Co., Osaka, Japan). ProteinA/G-Calf Intestinal Alkaline phosphatase conjugated (Thermo Fisher Scientific) was used as a pseudo-secondary detection reagent. Plasma specimens of wild sika deer were randomly chosen by prefecture from the sample collected in previous studies [32, 33].

ELISA

Recombinant proteins were diluted with Bicarbonate/carbonate coating buffer (100 mM) at 0.2 μg/mL, and 100 μL/well was added to Polysoap Immunoplate (Nunc, Roskilde, Denmark). The recombinant protein was immobilized for 2 days at 4°C. For serological examination of deer, plasma specimens frozen at −80°C was diluted 1:100 with ImmunoBlock and used for ELISA. Horseradish peroxidase conjugated Protein G (Sigma Aldrich) was used as a pseudo-secondary detection reagent. OPD (o-Phenylenediamine dihydrochloride) tablets (Sigma Aldrich) was used as substrate for the detection of peroxidase activity. The absorbance at 450 nm was measured by a microplate reader (Multiskan FC, Thermo Fisher Scientific). OD value of rVenus (control) was subtracted as background. Mean OD values were derived from 3 independent experiments. For the ELISA positive threshold, the end-point cut-off was established as the mean OD450 value of the plasma of PCR-negative deer, plus three standard deviations.

Experimental animals

Specific pathogen-free rabbits (JW strain) were purchased from Kitayama Rabes, Nagano, Japan. Animal experimentation was carried out according to the Laboratory Animal Control Guidelines of the National Institute of Infectious Diseases (institutional permission no. 116106, 111072 and 212008).

RESULTS

Bd37 gene sequences of the isolates, IpSG13-13-, IpSG10 and MRNK

Bd37 gene sequences were obtained by PCR cloning from genomic DNA (Table 1) IpSG13-13-1, IpSG10 and MRNK and cDNA of IpSG13-13-1. Full-length sequences were determined from 10 clones each (Table 2) (total 40 clones). All aligned DNA sequences are shown in Supplementary Fig. 1 and accession numbers of representative sequences are shown in Supplementary Table 2. Multiple alignments of the Bd37 sequence showed heterogeneity and length polymorphism. Nevertheless, the various Bd37 sequences from Japanese strains were distinctive and grouped into three types, namely JP-A, JP-B and JP-C types (Fig. 1 and Table 2), among which sequence identities (%) were 69.89–71.43% (Supplementary Table 3). The Bd37 sequences of MRNK strain (EU-A type) were distinct from those of the Japanese types. The identities between the JP types and EU-A type were 72–78% (Supplementary Table 3). JP-A and EU-A were more polymorphic than the others, and could be grouped into subtypes, as JP-A1, A2 and A3 and EU-A1 and A2 respectively (Table 2). Genomic DNA and cDNA of IpSG13-13-1 contained JP-A and JP-B types, while genomic IpSG10 contained JP-A and JP-C types. European strain, MRNK, contained only EU-A (EU-A1 and EU-A2) type sequences.

Fig. 1.

Alignment of Babesia divergens Bd37 amino acid sequences (A, B) and phylogenetic tree (C). (A, B) Yellow and purple circles indicate disulfide bridges and salt bridges identified in Rouen1987 strain [4, 6]. Cleavage site predicted in signal peptide (solid line) and glycosylphosphatidylinositol (GPI) anchoring site (circle) are shown (see METHODS). (C) Phylogenetic tree of various Bd37 orthologs from the Babesia divergens/B. capreoli group. Bootstrap values are shown at each node. Sequences determined in this study are shown in bold. Red, blue and green circles indicate Bd37 from Asia lineage, EU lineage and B. capreoli, respectively.

Comparison of Bd37 protein sequences

Putative protein sequences were obtained from all DNA sequences determined above (Supplementary Fig. 2). ClustalW alignments of representative protein sequences of JP-type and EU-type (Fig. 1A) and JP-A1, A2 and A3 (Fig. 1B) are shown. Bd37 of the Rouen87 strain (CAD19563) was included as a reference (Fig. 1A). Regardless of the observed DNA sequence polymorphism, all putative protein sequences contained a signal peptide at the N-terminus and a GPI anchor domain at the C-terminus with high probabilities (99.9%). In other regions, conserved and variable regions were mosaic throughout the sequences. High level of polymorphism was especially observed along forty amino acid residues after signal peptide sequence, where substitutions and in/dels were occurred frequently. Cysteines involved in disulfide bridges and residues involved in salt bridges [11] were conserved among JP-A, JP-B, JP-C and EU-A with few exceptions (Fig. 1). Percent identities (Table 3) among the JP-A, JP-B and JP-C type protein sequences were distantly related (60.12–77.04%) to the reported sequences from B. divergens (Rouen1987, Y5, Weybridge and 71/07/B strains) and B. capreoli. The EU-A1 and A2 types from MRNK strain were closely related to the B. divergens Y5 strain (CAD48926) (89.91% and 99.69%, respectively).

Table 3. Protein sequence percent identities.

	Isolate1	Accession number	Lineage2	Sequence type	1	2	3	4	5	6	7	8	9	10	11	12	13
1	Babesia divergens IpSG13-13-1#5	LC601771	Asia lineage	JP-A1	100	98.19	92.17	63.22	64.42	70.72	69.11	70.72	74.92	75.23	76.76	59.45	61.45
2	B. divergens IpSG13-13-1#10	LC601780	Asia lineage	JP-A2	98.19	100	93.67	63.22	64.72	70.4	69.11	70.4	74.92	75.23	76.76	59.15	61.45
3	B. divergens IpSG10#2	LC601804	Asia lineage	JP-A3	92.17	93.67	100	64.13	64.42	69.47	67.89	69.47	74.62	74.92	77.68	59.15	60.24
4	B. divergens IpSG13-13-1#26	LC601787	Asia lineage	JP-B	63.22	63.22	64.13	100	60.54	65.44	66.67	65.44	58.98	59.28	59.21	52.73	52.08
5	B. divergens IpSG10 #11	LC601808	Asia lineage	JP-C	64.42	64.72	64.42	60.54	100	70.86	67.77	70.55	65.56	65.86	64.44	56.13	55.42
6	B. divergens MRNK #5	LC601796	EU lineage	EU-A1	70.72	70.4	69.47	65.44	70.86	100	89.6	99.69	68.4	68.71	66.67	61.37	58.1
7	B. divergens MRNK #7	LC601790	EU lineage	EU-A2	69.11	69.11	67.89	66.67	67.77	89.6	100	89.6	63.86	64.16	62.73	60.55	57.36
8	B. divergens Y5	CAD48926	EU lineage		70.72	70.4	69.47	65.44	70.55	99.69	89.6	100	68.4	68.71	66.67	61.06	57.8
9	B. divergens Weybridge	CAD48924	EU lineage		74.92	74.92	74.62	58.98	65.56	68.4	63.86	68.4	100	99.71	87.92	54.88	54.55
10	B. divergens Rouen1987	CAD19563	EU lineage		75.23	75.23	74.92	59.28	65.86	68.71	64.16	68.71	99.71	100	88.22	55.18	54.84
11	B. divergens 71/07/B	CAD48927	EU lineage		76.76	76.76	77.68	59.21	64.44	66.67	62.73	66.67	87.92	88.22	100	57.49	56.19
12	B.capreoli 2770	ACX30791	B. capreoli lineage		59.45	59.15	59.15	52.73	56.13	61.37	60.55	61.06	54.88	55.18	57.49	100	60.7
13	B.capreoli 2801	ACX30790	B. capreoli lineage		61.45	61.45	60.24	52.08	55.42	58.1	57.36	57.8	54.55	54.84	56.19	60.7	100

1. Genetic isolates are included. 2. Lineage in the B. divergens/B. capreoli group.

The phylogenetic tree based on the ClustalW alignment among the Bd37 protein sequences determined in this study and from GenBank was shown in Fig. 1C. All Bd37 sequence types determined in this study (JP-A, JP-B, JP-C and EU-A) branched independently regardless of the lineages based on the 18S rRNA [32]. Bd37 from B. capreoli 2770 and 2801 strains (referred as Bcp37/41) were clearly separated from the others.

The protein sequences of Bd37 were highly polymorphic (Fig. 1A, 1B and Supplementary Fig. 2) and independently evolved (Fig. 1C). However, it seems not to influence the propensity of the predicted liner B cell epitope probability (Fig. 2A). The pattern was conserved throughout the sequences of all Bd37, except for those of B. capreoli 2801. By using the default cut-off value of 0.5 as suggested in the program, the propensity for B cell epitopes was abundant and encompasses almost all of the entire protein, except for the N- and C-terminal parts (Fig. 2A). Regions where probability was relatively high (score >0.6) were indicated as epitope A, B and C (Fig. 2A). Sequence alignments of the epitopes, except for that of B. capreoli 2801, are shown in Fig. 2B. Protein sequence in epitope A were more polymorphic than those in epitopes B and C.

Fig. 2.

B-cell epitopes of various Bd37 orthologs predicted by BepiPred-2.0 server and window graph of DNA polymorphism in JP-A, JP-B and JP-C. (A) The threshold (default value=0.5) is shown as red line. Predicted B-cell epitopes above 0.6 are designated as Epitope A, B and C. Sequence types with accession numbers are shown in Table 3. Asterisk shows highly polymorphic area (aa112-145) identified in Fig. 2C (JP-A). (B) Amino acid sequences of the epitopes A, B and C. (C) Nucleotide diversity of JP-A, JP-B and JP-C examined by sliding window plot with a window length of 10 bp and a step size of 25 bp.

Genetic analysis

Since protein sequences among JP types were highly polymorphic, we further examined segregating polymorphisms among gene sequences (Table 4 and Fig. 2C). Note that only 3 of JP-C type DNA sequences were available. By comparing JP-A and JP-B, haplotype diversity was comparable (0.897 ± 0.056 vs. 0.933 ± 0.077), while number of segregating sites (S) (74 vs. 9) and nucleotide diversity (Pi) (0.0279 vs. 0.0019) were relatively higher in JP-A. It indicates that multiple nucleotide substitutions within a single JP-A gene sequence frequently occurred and possessed higher levels of genetic diversities. The ratio of non-synonymous (dN) to synonymous (dS) substitutions (dN/dS) was calculated to assess natural selection acting on Bd37. The ratio of JP-A was dN/dS <1, representing potential evidence that a purifying selection pressure might be taking place to shape the populations. This negative (purifying) selection was statistically significant examined by Z-test (dN <dS, P=0.00097). dN/dS value of JP-B was slightly over 1 (1.044), but neither of positive (dN >dS) nor neutral (dN=dS) selections was statistically significant. Figure 2C shows sliding window plot with a window length of 10 bp and a step size of 25 bp providing detailed analysis of the diversity. Pi of JP-A was ranged from 0.0429 to 0.19, and the highest peak diversity was observed within the nucleotide position at 300–500 (actual position at 335–535 of CDS). The highly divergent area, 335–535 of CDS, correspond to position at 112–145 of proteins sequence. In the linear B-cell epitope prediction, the score of the 112–145 (aa) region was slightly above the threshold (Fig. 2A) indicating polymorphic region of Bd37 JP-A may have weak antigenicity. Humoral immunity against Bd37 JP-A may be predominantly induced against epitope A, epitope B and/or epitope C, and DNA sequences encoding these regions are genetically stable.

Table 4. Genetic diversity of Bd37 JP-A, JP-B and JP-C types.

	No	Length	Position	h	Hd	SD	S	Pi	dN/dS	p (Z-test)
JPA	17	930	34–963	10	0.897	0.056	74	0.02786	0.334	0.00097*	dN <dS
JPB	10	951	34–984	8	0.933	0.077	9	0.00189	1.044	0.31910	dN >dS
JPC	3	930	34–963	3	1	0.272	4	0.00287	0.295	0.21027	dN <dS

Number of Haplotypes, h; Haplotype (gene) diversity, Hd; Number of polymorphic (segregating) sites, S; Nucleotide diversity, Pi; Total number of nonsynonymous changes, dN; Total number of synonymous changes, dS. *Statistically significant.

Development of discriminative PCR

For detailed molecular analysis of B. divergens in deer, a discriminative PCR was developed (Fig. 3A) by designing primers specific for variable regions in JP and EU types (Supplementary Table 1). The specificity of this PCR system was examined by using plasmids encoding each type. The results are shown in Fig. 3B, demonstrating formation of a single positive signal only when the specific primer matched, except for JP-B/EU-A2. When the JP-B/EU-A2 primer was used, amplicons were generated from the plasmids encoding JP-B and EU-A2. Thus, we confirmed the sequence type, JP-B or EU-A2 by direct sequencing. Sensitivity of the type-specific PCR was examined by using 10-fold dilutions of plasmid (1 × 10¹ to 1 × 10³ copies in the PCR mixture) as template in the nested PCR. In the first-round PCR (Fig. 3C top panel) by using universal (top panel) primers, products were observed when 1 × 10² and 1 × 10³ copies were used as template. In the nested PCR, in which the first-round PCR products of 1 × 10¹ to 1 × 10³ copies respectively were used as the template, amplicons were visible at all concentrations (Fig. 3C, middle and bottom panels) regardless of using universal or specific primers.

Fig. 3.

Discriminative PCR targeting the Bd37 genes of Babesia divergens Asia and EU lineages. (A) Schematic diagram of the PCR primer design. Black arrows show primers for primary and secondary PCR. Black and white boxes indicate conserved and variable regions. Primer sequences are shown in Supplementary Table 1. (B) Specificity of primers. PCR amplification using universal primers Bd37-ATG, or specific primers Bd37-JPA F1, Bd37-JPB/EUA2 F2, Bd37-JPC F1, or Bd37-EUA1 F1 and Bd37-TAA. Plasmids carrying each Bd37 gene of genetic strains shown in Table 3 were used as the template. M, marker. (C) Sensitivity of specific nested PCR. Conventional PCR using universal primers Bd37-ATG and Bd37-TAA or Bd37-R1 (upper and middle panel, respectively) or Bd37-JPA F1, Bd37-JPB/EUA2 F2, Bd37-JPC F1, or Bd37-EUA1 F1 and Bd37-TAA (lower panel) with the plasmid carrying the Bd37 gene described above. Plasmids were diluted 10-fold from 1 × 10¹ to 1 × 10³ copies and used (lanes 10¹ to 10³). M, marker. (D) Phylogenetic trees of Bd37 JP-A, JP-B and JP-C DNA sequence from sika deer and Ixodes persulcatus tick. Bootstrap values are shown in at each node. Geographical origin of deer is shown in Supplementary Table 4.

Examination of the field samples

The discriminative PCR developed above was applied to examine B. divergens in wild sika deer (Cervus nippon, erythrocytes) (22 samples) and additional field-collected ticks (14 samples) (Table 5). The samples of deer and ticks were previously tested for presence of B. divergens by 18S rRNA-based PCR [32,33,34]. By using the primers targeting all Bd37 sequence types (universal), PCR amplicons were indeed developed from all samples. Further discriminative PCR revealed JP-A specific amplicons were exclusively identified in all 36 samples, while JP-B/EU-A2 and JP-C specific amplicons were generated only from 17 samples, respectively. EU-A1 specific amplicon was not detected. Table 5 shows the number of positives in each combination pairs.

Table 5. Summary of field survey examined by discriminative PCR.

Species	Material	No. of PCR-positive				Total
		A only	A and B	A and C	A, B and C
Ixodes persulcatus	Whole body	2	3	5	2	12
	Salivary glands	1	1	0	0	2
Cervus nippon	Erythrocytes	5	6	5	6	22
Total		8	10	10	8	36

The amplicons amplified from deer sample by the discriminative PCR above were further sequenced to examine single nucleotide polymorphisms (SNPs) within individual genomic samples (Supplementary Fig. 3). Among 12, 7 and 6 of JP-A, JP-B and JP-C, respectively, SNPs were observed in 8 and 1 of JP-A (n=8) and JP-C (n=1) (Supplementary Fig. 3A, 3C). SNP within individual genomic samples was not observed in the JP-B (Supplementary Fig. 3B). Length polymorphism was observed only in a JP-C (deer 08#18) where the length is 9nt shorter than others (Supplementary Fig. 3C). Furthermore, JP-C from deer 07#12 contained high polymorphic region in latter half of the sequence (87/335) (Supplementary Fig. 3C). Further examination of 07#12 sequences by cloning revealed that (Supplementary Fig. 3D), 2 sequence types were mixed (JPC-1 and JPC-2) (Supplementary Fig. 3D). Identities between the two deduced amino acid sequences was 80–84%.

By using all haplotype sequences from deer and ticks in this study, phylogenetic tree (DNA, 738 nt) was constructed (Fig. 3D). Geographical origin of the deer and ticks were 9 prefectures and one (Hokkaido), respectively. While regional structure was not observed in the tree demonstrating the genetically closely related parasites were transmitted between the tick vector I. persulcatus and reservoir Cervus nippon throughout Japan.

Expression of recombinant Bd37 proteins

To examine whether Bd37 was antigenically recognized in deer, recombinant Bd37 proteins (rBd37) fused with thioredxin (Trx) at N-terminus were expressed in E. coli. The rBd37 of all types and control (rVenus) were expressed as soluble proteins and purified under native conditions. The SDS-PAGE and western blot analysis using anti-Trx antibody (Fig. 4A) showed that all recombinant proteins were expressed as expected size. The observed proteins were 11.7 kDa larger than the expected size because of fused thioredoxin.

Fig. 4.

Expression, cross-reactivity and use in ELISA of recombinant Bd37 from Babesia divergens Asia and EU lineages. (A) SDS-PAGE and western blot analysis of purified rBd37 proteins conjugated with thioredoxin (Trx) and His-tag. Note that the molecular weight of Trx and His-tag is 16 kDa. Anti-Trx antibody was used as primary antibody. (B) Cross reactivity of PCR-positive and PCR-negative deer (upper left), and positive rates shown by bar chart and geographical map (upper right and lower panels, respectively). Wild sika deer were serologically examined by rJP-A-based ELISA (n=360). Positive rates in each prefecture (n=24) are shown. PCR-positive rate is shown for comparison. (C) Different serological profiles of deer (upper panel) and cross-reactivities of rBd37 (lower panel). Numbers above the line indicate designated sample number of wild sika deer.

Recombinant Bd37 JP-A-based ELISA

Genetic analysis revealed JP-A1 and JP-A2 were exclusively detected in all B. divergens examined in Japan (Table 2, Fig. 3D). Thus, JP-A2 was used as representatives of rBd37 JP-A as antigen in the ELISA. Rabbit serum raised against rBd37 JP-A2 discriminated the rJP-A from rJP-B, rJP-C, EU-A1 and EU-A2 (Supplementary Fig. 4).

The presence of cross-reactive antibody in deer against the rBd37 JP-A was examined by using plasma from PCR-positive (50 specimens) and PCR-negative deer (12 specimens). Latter of which were chosen from B. divergens-PCR free prefectures, examined in our previous study [34]. In the ELISA, mean OD values of PCR-positive and negative deer were 0.12 ± 0.16 and 0.002 ± 0.01, respectively. Out of 50 PCR-positive deer examined, OD values of 32 samples were above the end-point cut-off value, which was established by the mean OD value of the PCR-negative plasma plus 3 standard deviations (mean + 3SD) (Fig. 4B).

Serological screening of B. divergens infection in deer

Presence of PCR-positive, ELISA-negative samples suggests that the test may produce some false negative results. While nation-wide survey was conducted to see antibody detection throughout Japan. Fifteen deer plasma samples were randomly chosen from each of the 24 prefectures (total 360). Overall positive rate of the serological test was 13.9%, which was approximately twice as high as the positive rate derived from PCR tests (6.8%) [33, 34]. Higher positive rates (>20%) (Fig. 4B) were observed in Hokkaido and prefectures having mountain area including Iwate, Miyagi, Gunma, Yamanashi, Nagano and Gifu. The trend of the positive rate among prefecture was comparable to that examined by PCR (Fig. 4B), and related to the geographical distribution of vector I. persulcatus tick [32].

Detection of antibodies against rBd37 JP-B and JP-C in deer

In addition to antibodies against rBd37 JP-A, those of rBd37 JP-B and JP-C in deer were examined by western blot analysis. Three deer plasma showing relatively high OD value (#64, #275 and #324) in the rJP-A-based ELISA were tested (Fig. 4C). Major bands were recognized at approximately 50 kDa of rBd37 JP-A in all specimens tested. Additional bands against rBd37 JP-B and JP-C were also recognized in the plasma of #275 and #324, respectively. Recombinant Bd37 of JP-A2 as well as rBd37 JP-A1 and A3 also reacted to all plasma specimens of wild sika deer, suggesting JP-A1, A2, and A3 could be considered antigenically similar in this system.

DISCUSSION

To seek immunodiagnostic genes and antigens of Asia lineage in the Babesia divergens/B. capreoli group, we investigated GPI anchored protein Bd37, a major surface protein of B. divergens sensu strict. Direct examination of parasite genome in the vector (Table 1) and reservoir (Table 5) identified unique and novel features of the Bd37 gene family: 1. Unlike other lineages, the Asia lineage carries up to three distinct Bd37 ortholog genes, namely Bd37 JP-A, JP-B and JP-C (Table 2). 2. Genetic diversities of them were distinct (Table 4) but predicted B cell epitopes were similar (Fig. 2). 3. The JP-A was exclusively encoded in all genomes isolated, while JP-B and JP-C were encoded 50% of the genomes examined, respectively (Tables 2 and 5). 4. Antibodies against recombinant JP-A (rJP-A), JP-B and JP-C were evident in sika deer indicating expression of all Bd37 orthologs in blood stage (Fig. 4A). 5. Although the sequences were highly polymorphic, the purified rJP-A2 was widely cross-reactive to the JP-A variants (Fig. 4C). Finally, rJP-A2-based ELISA (Fig. 4B) demonstrated that 13.9% of wild sika deer throughout Japan (24 prefectures, n=360) were positive. The rate was 2 times higher than that examined by PCR.

B. divergens parasites have been isolated from mammals including human and bovine, and maintained erythrocytes in vitro or laboratory animals. Such efforts established many strains, and studies for Bd37 were performed on them [1, 4, 6, 10, 11, 13, 25, 26]. Unfortunately, the Asia lineage has not biologically isolated. Thus, in primary investigation, we utilized salivary glands of I. persulcatus ticks as genomic resource. In the salivary glands, multinucleated sporoblast matures and multiple sporozoites develop. From two ticks, as much as three type genes (JP-A, JP-B and JP-C) were isolated (Table 2). Further investigation of deer (Table 5) did not identify additional sequence type. Based on the calculation from semi-quantitative PCR (Fig. 3C), original salivary glands (whole) contained 5 × 10⁵ <copies /tick (data not shown). This high-density population of parasite was rich in diversity and advantageous for deep genomic analysis. Furthermore, isolation of mRNA of JP-A and JP-B from the salivary glands (Table 4) indicates expression of JP-A and JP-B in the sporozoite stage and need for erythrocytes infection in next stage (due to lack of samples, isolation of JP-C mRNA could not be attempted). Although it is challenging to obtain salivary glands infected with matured, multiplied Babesia parasite, this study showed the salivary glands as informative tool to investigate not only interaction between vector ticks and mammals but also population genetics.

Genetic polymorphisms are a characteristic of genes encoding merozoite surface proteins. The polymorphism of B. divergens Bd37 gene was first reported by Hadj-Kaddour et al. [13], where fourteen isolates of B. divergens were clearly distinguished by five major polymorphic groups examined by PCR- Restriction fragment length polymorphism (RFLP) using RsaI and BglII restriction enzymes. Sun et al. [26] performed a similar analysis and observed extreme polymorphism within B. capreoli (Bcp37/41), which was greater than that among B. divergens isolates. In our study, JP type sequences indeed included the RsaI and BglII recognition sites which were unique to JP-A, JP-B, JP-C, as well as EU-A (Supplementary Fig. 1). Though, RsaI recognition site is absent in some JP-A sequences. Characteristic features of JP-A observed in this study were as follows (Table 5). 1. Higher values of nucleotide diversity (π). 2. High number of segregating sites. 3. Significantly low dN/dS value (0.334). 4. Three B-cell epitopes predicted. Sliding window plot (Fig. 2C) showed nucleotide diversity was increased from 300 to 500 which corresponds to the amino acid residues at 100 to 166 (Figs. 1A, 1B and 2A). The region was not overlapped with the B-cell epitopes (Fig. 4B). Indeed, cross reactivity among JP-A1, JP-A2 and JP-A3 was observed in the western blot (Fig. 4C) and ELISA (data not shown). The selective pressure to keep the stable retention in the genome, low dN/dS and possible structural stability may be due to erythrocyte binding role of Bd37 orthologs [6]. We suppose these characteristics may be advantageous to use JP-A in serological assay as an antigen.

The ELISA using recombinant JP-A2 identified 32 positive plasma samples (64%) in a total of 50 PCR-positive deer samples (Fig. 4B). The reason for the lower detection rate is not clear but the proper folding of the recombinant protein and/or presentation of antigenic epitopes may be interfered with. The recombinant Bd37 JP-A (and other orthologs) in this study was expressed as native protein containing N-terminal and C-terminal hydrophobic regions (signal peptide and GPI respectively). By tagging thioredoxin (Trx) to the N-terminus, the fusion protein was successfully purified as soluble. This construction was similar to the recombinant Bd37 in previous studies where hydrophobic regions were unstructured and required for immune system against B. divergens infection [5, 6]. We decide to retain the hydrophobic regions, though expression of recombinant proteins usually reduce hydrophobic sequence to maximize the yield of soluble protein. González et al. [10] diagnosed a patient by using GST tagged Bd37, which lack signal peptide (1–27). Elimination of N-terminus hydrophobic residues may enhance the sensitivity of the JP-A. Other strategies including sandwich ELISA and using strept-tag [24] can be considered. For future endeavor, we attempted to use multiple recombinant proteins to overcome the low sensitivity (Supplementary Fig. 4). Small number of samples showed higher OD value against the cocktail. Improving assay sensitivity through modification of the test format should be assessed in future.

In conclusion, B. divergens Bd37 orthologs exhibited extensive sequence diversity in Asia lineage. Among 3 types identified, one was stably and others were sporadically retained in the genome. Nucleotide diversity of JP-A was higher than those of JP-B and JP-C. Recombinant JP-A-based ELISA could detect positives throughout Japan. The system needs to be improved in terms of sensitivity, but has potential for serological diagnosis/screening. B. divergens group includes multiple lineage zoonotic or zoonotic potential. Continued surveillance and research of the B. divergens group and its effects on cervid ecosystems is vital for controlling the long-term consequences of this emerging disease.

CONFLICT OF INTEREST

The authors declare there are no conflicts of interest.