Morphological identification of tick species
Of the 35 ticks collected from Kwale, 16 specimens were semi-engorged, 11 were fully engorged, 3 were partially fed and 5 were slightly fed (Table 1). On microscopic examination, 14 specimens were found to have damaged or missing mouth parts while 21 specimens were intact. All the study tick specimens were examined for key morphological characteristics defining Boophilus species. In Kenya, R. decoloratus is endemic and widely distributed and is likely to be confused with R. microplus. Thus, reference R. microplus and R. decoloratus ticks were also analysed and compared to the Kwale ticks. Since most the ticks were semi- or fully engorged and some had missing mouth parts, only those specimens that were intact could be tentatively identified based on morphological features described in the identifications guides and keys used.
Microscopically, the principal features used for identification and discrimination of ticks in the sub-genus Boophilus from other Rhipicephalus species included an inornate scutum, short mouth-parts with palps slightly extending beyond the mouth, small eyes, pale yellow legs, a round-shaped spiracular plate, absence of festoons, and presence of ventral anal plates with spurs in males. Among the features used to discriminate species within the Boophilus sub-genus are the structure of denticles in the hypostome in both males and females, presence of seta on palp article one in females, the length of spurs on coxa I and presence or absence of caudal appendages in males [21]. The basis capituli was the only feature that was clear in intact semi- or fully engorged adults as most of the other features become unclear when the tick body stretches after feeding.
As expected, reference R. microplus female specimens analysed in this study had four rows of denticles on each side of the hypostome and a concavity with no setae on the medial aspect of article one of the palpal segment (Figure 2). Eleven of the twenty-one ticks that had an intact mouth part were found to have morphological features characteristic of female R. microplus while 12 of the 14 ticks with damaged mouth parts had a spiracular plate (Figure 2) consistent with that of Boophilus, they were thus tentatively grouped as Boophilus with no species designation. There were no male R. microplus in the Kwale collections. Reference R. australis males from Queensland had four rows of denticles on each side of the hypostome and indistinct adanal plates with short spurs which did not reach the posterior body margin and were not visible dorsally (Figure 3). On the other hand, reference R. decoloratus females and males had three rows of denticles on each side of the hypostome. Females had a convex protuberance with setae on the first palpal segments while males had long distinct adanal plates with long spurs which extended beyond the posterior body margin and were clearly visible outside the scutum (Figure 3 and 4). No R. decoloratus specimen was found among the 35 ticks from Kwale. Ten tick specimens had features consistent with R. appendiculatus while two specimens had features consistent with the Amblyomma genus. Figure 5 shows mouth part images of representative specimens from 11 ticks from Kwale with features consistent with R. microplus. The tentative species identity of all the 35 tick specimens from Kwale based on observed morphological features is listed in Table 1.
Molecular identification and genetic characterization based on COI
Molecular analysis involving partial amplification and sequencing of cytochrome oxidase subunit I (COI) gene was performed on each of the 23 tick specimens tentatively identified as Boophilus, eight identified as R. appendiculatus, one A. variegatum and the 11 reference specimens. A total of 43 ticks were characterized based on the COI sequence. BlastN analysis of the COI sequences confirmed the 23 specimens from Kwale to be R. microplus as well as the R. appendiculatus, A. variegatum and the species identity of reference specimens.
The 23 R. microplus COI sequences exhibited nucleotide similarity >99.5%. Only two informative polymorphic sites were found, resulting in three haplotype sequences differing in two variable polymorphic sites at position 51 and 483 with a haplotype diversity (Hd) of 0.5692. Haplotype 1 consisted of six sequences which had the bases TG while fourteen sequences formed haplotype 2 with bases CG at the two variable positions. Haplotype 3 consisted of 3 sequences with bases CA at the two polymorphic sites. COI sequences of the 23 Kwale R. microplus and 11 COI sequences of reference ticks appear in GenBank under the Accession numbers MT181192-MT181227 while the three haplotype sequences appear under Accession numbers MT181228-30 ( Table 1).
Pairwise identity comparisons between the three R. microplus haplotype sequences and reference sequences obtained in this study or from GenBank are indicated in Supplementary Table S1. The three Kwale haplotypes had a high similarity of >99.6%. They matched with >99% to each of the 10 R. microplus reference sequences from Africa. These included a Kenyan R. microplus sequence (Rm_KE_KX228549), sequences from Cameroon (Rm_CF4, Rm_CF5, Rm_CM__MK648412, Rm_CM_MG983832, Rm_CM_MG983831), Congo (Rm_DRC_MF45873), South Africa (Rm_SA_KY678117), Benin (Rm_BN_KY678120) and Madagascar (Rm_MA_KY678118). They also had a high similarity (99.5%) with GenBank R. microplus reference sequences from USA (Rm_USA_KP143546), Colombia (Rm_CL_KT906181) and Philippines (Rm_PH_KX228548) (Supplementary Table S1).
The Kwale R. microplus haplotypes had an identity of 94.5% to R. australis sequences (Rm_QLDF1, Rm_QLDF4, Rm_QLDF6) obtained in this study and two other Australian GenBank reference sequences (Rm_AU_KC503255 and Rm_QLD_AF132827). Sequences from reference R. microplus specimens from Laos and GenBank reference sequences from China matched with a lower identity of 92%. The Kwale haplotypes matched to GenBank R. annulatus reference sequences from India (Ran_IN_KX228542), Cameroon (Ran_CM_MK648411) and Burkina Faso (Ran_BF_KY678123) with between 92-93% identity. They matched to R. decoloratus sequences from reference specimens from Kenya (Rd_KBF6 and Rd_KBF7) and Cameroon (Rd_CdF1 and Rd_CdF6) with about 88%, the same with other R. decoloratus reference sequences from GenBank (Rd_SA_KY678130, Rd_SA_AF132826) from South Africa, (Rd_CM_MK648413) from Cameroon and (Rd_Bfaso_KY678127) from Burkina Faso (Supplementary Table S1).
COI phylogeny and genetic relationships
Phylogenetic analysis was undertaken to determine the genetic relationships between the Kwale R. microplus haplotype sequences and reference R. microplus, R. australis and R. decoloratus isolates analysed in this study and other annotated reference sequences available in GenBank. A tree (Figure 6) was constructed using the three R. microplus COI haplotype sequences, the 11 reference sequences analysed in this study and 28 GenBank reference sequences. Two R. appendiculatus sequences, one from GenBank (Rap_AF132833) and another obtained in this study (Rap_KF18) were used as outgroups. Only significant bootstrap values above 70% are shown.
In the COI tree (Figure 6), four major clusters (Clade A-D) of R. microplus complex were observed. The Kenyan sequences (Rm_KW_H1-H3) clustered in a major clade (Clade A) with sequences from Cameroon (Rm_CF4 and Rm_CF5) analysed in this study and three GenBank reference sequences (Rm_CM__MK648412, Rm_CM_MG983832, Rm_CM_MG983831), one from Congo (Rm_DRC_MF45873), South Africa (Rm_SA_KY678117), Benin (Rm_BN_KY678120) and Madagascar (Rm_MA_KY678118), USA (Rm_USA_KP143546), Colombia (Rm_CL_KT906181) and Philippines (Rm_PH_KX228548). The cluster was strongly separated (100%) from a GenBank R. microplus sequence from Kenya (Rm_KE_KX228549). Clade A was strongly separated (92%) from a R. australis sister clade (clade B). Clade A and B formed a larger R. microplus complex cluster which was moderately separated (65%) from a R. annulatus clade. Two sequences of R. microplus specimens from Laos (clade C) analysed in this study clustered with GenBank reference sequences from China (clade D). The Lao sequences were moderately separated from the Chinese sequences (68%). The Laos/China cluster was strongly separated (96%) from the R. microplus/R. annulatus cluster (Figure 6). The Kenyan R. decoloratus sequences analysed in this study (Rd_KBF6, Rd_KBF7) clustered closely in one clade with those from Cameroon (Rd_CdF1 and Rd_CdF6) and other reference sequences from South Africa (Rd_SA_KY678130, Rd_SA_AF132826), Cameroon (Rd_CM_MK648413) and Burkina Faso (Rd_Bfaso_KY678127).
Mitochondrial genomes characterization and phylogeny
MITOS, a web server for automatic annotation of metazoan mitochondrial genomes was used to annotate proteins, tRNAs and non-coding RNAs in the four complete mtDNA genomes analysed in this study. Multiple sequence (MS) analysis was performed on the four mtDNA sequences and 13 reference genomes from GenBank. These included six R. microplus, one R. annulatus, one unverified R. decoloratus, one R. geigyi, a partial R. appendiculatus, one R. sanguineus, one R. turanicus and one H. longicornis reference genome sequences. Figure 7 shows the 13 proteins, 22 tRNAs and two rRNAs annotated by MITOS for the four mtDNA sequences analysed in this study.
From the MITOS prediction, the arrangement and length of the annotated mitochondrial features in the two R. microplus (Rm_KF13 and Rm_KSF2) genomes was very similar (Figure 7). The program predicted the presence of two trnE genes (65bp) in the two genomes. The first occurred in the plus strand in position 4614-4678 while the second is in position 4744-4808 for both Rm_KF13 and Rm_KSF2. Both predictions had a similar e-value of 3.313e x 10-05. In the MS alignment, this region was part of control region I which occurs within the tandem repeat region annotated between trnE and nad1 in other reference R. microplus genomes.
The program also predicted pseudo copies of tRNA genes for trnA (gca), trnR (cga), trnN (aac), trnS1(aga) and trnE (gaa) (Figure 7) in position 4854-5184 of the R. decoloratus (Rd_KBF6) mt genome. The MS alignment showed the first prediction in position 4368-4610 to be conserved and common to the reference mtDNAs. The pseudo prediction from position 4854-5184 was a 331bp long AT rich sequence lacking in the unverified R. decoloratus reference sequence and in other reference genomes. It occurs as part of the tandem repeat region annotated between trnE and nad1 in other reference R. microplus genomes.
MITOS could not locate the gene for trnS1 in the R. apendiculatus (Rap_KF10) genome. In the reference mtDNA annotations, the trnS1 gene is reasonably conserved located upstream of the trnN gene in position 4557-4612 for Rm_KF13 and Rm_KSF2 and 4555-4610 for Rd_KBF6. In the R. appendiculatus Rap_KF10 sequence, the region (4543-4599) is predicted by MITOS to be a pseudo non-standard tRNA feature (trnX) (Figure 7). In the MS alignment, two nucleotides are missing from the 5′ start of the gene in Rap_KF10 and the partial R. appendiculatus Zimbabwe reference genome when compared to the predicted trnS1 gene of the R. microplus reference genomes. MITOS also predicted a short nad4 gene and two nad5 genes in the minus strand in Rap_KF10 (Figure 7). The first nad5 (1500bp long) at position 8251-9750 with a quality value of 2.05 x 108 and a second tiny pseudo fragment (45 bp) at position 11081-11125 with a value of 375.9. With such a lower value, the 2nd prediction is highly unlikely. In the MS alignment, this fragment is part of nad4 which appears to be conserved across the other reference genomes.
Pairwise identity matches and phylogenetic analysis were performed on the four mtDNA sequenced in this study and the 13 reference sequences from GenBank. The mtDNA genome sequences appear in GenBank under the Accession numbers MT430985-88. The percent identities observed are shown in Supplementary Table S1. The two R. microplus sequences Rm_KF13 and Rm_KSF2 sequenced in this study had very high nucleotide similarity (99.97%). They matched with an identity of greater than >98% to the R. microplus reference genome sequence from India, Cambodia, Brazil and USA. They matched to the R. australis reference genome (Rm_AU_KC503255) with a similarity of 96.07% and to the China sequence (Rm_CHI_KC503259) with an identity of 94.34%. Their similarity to the R. decoloratus sequence (Rd_KBF6) was 87.15%, 94.26% to the reference R. annulatus (Ran_ROM_KC503256), 87.32% to the R. geigyi (Rgei_BF_KC503263) and 82.5% to the R. appendiculatus sequence (Rap_KF10). The H. longicornis sequence (HL_MK450606) was used as an outgroup. It matched to the other sequences analysed with an identity ranging between 72-75%.
In the mtDNA phylogenetic tree (Figure 8), four major clusters were observed with all the nodes strongly supported by 100% bootstrap value. The two R. microplus genome sequences (Rm_KF13 and Rm_KSF2) clustered closely with R. microplus reference sequences from Brazil (Rm_BZ_KC503261), USA (Rm_USA_KP143546), Cambodia (Rm_CA_KC503260), India (Rm_MK234703 INDIA) and Australia (Rm_AU_KC503255). This R. microplus cluster was separated from a sister clade consisting of a reference sequence from China (Rm_CHI_KC503259) and a R. annulatus (Ran_ROM_KC503256) sequence. The R. decoloratus sequence (Rd_KBF6) clustered with the unverified sequence (Rd_SA_KY457525) from South Africa and were separated from the R. geigyi sequence. This R. decoloratus sequence matched with an identity of 99.13%, 90.54% and 82.81% to the unverified, R. geigyi and R. appendiculatus (Rap_KF10) sequences respectively (Supplementary Table S1). The R. appendiculatus (Rap_KF10) sequenced in this study clustered with the reference partial sequence (Rap_ZM_KC503257) from Zimbabwe (Figure 8). The percent identity matches between these two sequences was 99.18%. The R. appendiculatus Rap_KF10 sequence matched to the reference R. geigyi sequence with an identity of 83.49% and showed identities of 82.81% and 83.69% to the R. decoloratus (Rd_KBF6) and the reference R. annulatus sequences respectively (Supplementary Table S1). In the tree, R. sanguineus and R. turanicus sequences clustered in a final clade with similarities of 89.2%.
Molecular detection of B. bovis in R. microplus ticks
Genomic DNA samples from 21 specimens confirmed by molecular analysis to be R. microplus were subjected to sensitive molecular quantitative qPCR assays to detect presence of bovine blood and consequently B. bovis parasites. Presence of cattle blood DNA was tested in undiluted (neat) gDNA by amplifying glyceraldehyde 3-phosphate dehydrogenase (GAPDH), a housekeeping gene found in all mammalian cells.
The cycle quantification (Cq) values ranged from a high of 39.45 in KSF3 to a low of 23.40 in KF16. At a threshold of 100 relative fluorescent units (RFU), five samples had Cq values below 30 while 14 had values that ranged from 30-39. Bovine DNA was not detected in 2 samples (KF-K6 and KSF5). Analysis of a 1:10 diluted DNA of these two samples resulted in the detection of bovine DNA (Cq value of 37.26) in KSF5. No DNA was detected in the diluted KF-K6 sample.
Having confirmed the presence of bovine DNA in the extracted tick DNA, B. bovis specific primers were used to amplify specific regions of cytochrome b (Cytb) and 18S rDNA genes to detect presence of B. bovis parasites. At a threshold of 100 RFU, all material was considered negative for B. bovis DNA using either assay (Cq values = >40.00).