Low genetic diversity and population structuring of Amblyomma hebraeum and Rickettsia africae from coastal and inland regions in the Eastern Cape Province of South Africa

Amblyomma hebraeum is the main vector of Rickettsia africae, the causative agent of African tick bite fever in southern Africa. Because pathogen dispersal is known to be influenced by tick adaptations to climate or host species, this study aimed to analyse the genetic diversity of A. hebraeum and R. africae infection of ticks collected from cattle in the Eastern Cape province of South Africa. DNA was extracted, amplified, and sequenced for the COI and ITS2 markers from A. hebraeum samples and the 17 kDa and ompA genes for rickettsial detection. Between six and ten haplotypes were identified from 40 COI and 31 ITS2 sequences; however, no population structuring was observed among sites (ΦST = 0.22, p < 0.05). All A. hebraeum isolates clustered with southern Africa GenBank isolates. Rickettsia africae was detected in 46.92% (95% CI = 41%–53%, n = 260) of ticks. All R. africae isolates clustered with strain PELE and Chucks, which were reported previously from South Africa. These results confirm that A. hebraeum populations are undergoing a recent population expansion driven by cattle movement, facilitating local and long dispersal events across the Eastern Cape province.


INTRODUCTION
Amblyomma hebraeum is a three-host tick distributed across southern Africa's savanna and thicket biomes (Estrada-Peña et al., 2008). In South Africa, A. hebraeum is found throughout the northern and eastern provinces, mainly distributed along the south-eastern coastal regions in the savanna and grassland biomes. The coastal belt environment in South Africa is most favourable to A. hebraeum due to the abundance of grasslands, more stable temperature ranges, and higher annual rainfall compared to inland regions (Marufu et al., 2011;Nyangiwe et al., 2011). However, its distribution has been reported to spread into the northeastern Northern Cape regions (Spickett et al., 2013;Yawa et al., 2018). Amblyomma hebraeum parasitizes mainly cattle, other domestic animals, and wildlife, creating an interface for human disease emergence (Horak et al., 2018;Iweriebor et al., 2020).
Rickettsia africae is an obligate, intracellular bacteria that is the etiological agent responsible for African tick bite fever (ATBF) (Raoult & Roux, 1997;Williams et al., 2007). Livestock husbandry practices play a significant role in the transmission dynamics of R. africae to humans, with a prevalence of up to 92.6% in cattle (Mutai et al., 2013;Yssouf et al., 2014). The high infection of R. africae in A. hebraeum is maintained by transovarial and transstadial transmission resulting in the larval, nymph, and adult life stages being able to infect humans (Kelly & Mason, 1991;Mutai et al., 2013;Yssouf et al., 2014). Amblyomma hebraeum is the principal vector of ATBF in southern Africa (Jensenius et al., 2003).
The rapid growth in tourism to South Africa offered by game reserves, heritage sites, and beaches has led to an increase in international ATBF case reports (Mxunyelwa & Lloyd, 2019;Silva-Ramos & Faccini-Martínez, 2021). However, autochthonous case reports of ATBF are scarce due to the lack of awareness of the disease or the presentation of mild febrile symptoms that are often misdiagnosed (Tomassone et al., 2016). Molecular analysis of R. africae isolates have revealed high interspecies homogeneity, making identification and analysis difficult Kimita et al., 2016). Therefore, investigating intraspecific variation among R. africae isolates may help resolve discordance among the different gene fragments used to classify R. africae. Consequently, this may help diagnose the acute phase of the disease, which is often missed by serological diagnosis (Fournier et al., 2002).
Previous studies examined a single mtDNA marker and did not find any clear A. hebraeum population structuring between different coastal and a few inland regions in South Africa (Cangi et al., 2013;Matthee, 2020). Exploring the genetic diversity between inland and coastal A. hebraeum populations may help identify sites for evolutionary divergence. Thus, an in-depth and extensive study into the genetic diversity of A. hebraeum populations with more variable markers may clarify the factors driving the dispersal of the A. hebraeum-R. africae system. This study aimed to determine the population genetic structure of coastal and inland A. hebraeum ticks and R. africae isolated from the ticks using the COI and ITS-2 markers for A. hebraeum populations and the ompA and 17 kDa genes for R. africae.

Study sites, collecting, sampling, and taxonomy
Two-hundred-and-sixty adult A. hebraeum specimens were collected randomly from multiple cattle of mixed breeds grazing on communal pastures in rural areas of the Eastern Cape province of South Africa.
Populations collected included 20 ticks each from seven coastal and six inland sites (Figure 1). Ticks were collected from predilection sites and preserved in 70% ethanol until morphological and molecular analyses. Morphological identification of ticks was conducted at the Döhne Agricultural Research Institute in Stutterheim, South Africa, using Walker et al. (2003). Individual ticks were surface sterilized with 70% ethanol and rehydrated in distilled water before DNA extraction.

DNA extraction, amplification, and sequencing
Individual whole A. hebraeum ticks were bisected, and one half was placed into a 1.5 ml microcentrifuge tube, which was utilized for DNA extraction. The other half was frozen at À80 C until further use. Tick F I G U R E 1 Location of study sites in the Eastern Cape Province in South Africa. extracts were homogenized using glass microbeads in a Disruptor Gene bead-beater (Scientific Industries, US) to facilitate the release of pathogens and protein digestion for nucleic acid extractions. Genomic DNA was extracted from individual ticks using the standard phenolchloroform protocol (Sambrook & Russell, 2006). The standard protocol was modified by incubating ticks overnight at 56 C to complete deproteinization. DNA was stored at 4 C until further use.
Amplified products were separated by electrophoresis on a 1.5% agarose gel. Purification of amplicons and sequencing reactions were performed at Inqaba Biotech (Pretoria, South Africa). COI, 17 kDa, ITS2, and ompA sequences were translated to amino acids in MEGAX to check for the presence of stop codons (Kumar et al., 2018). Individual COI and ITS2 mitochondrial haplotypes were identified using DNAsp v6 (Rozas et al., 2017). A BLASTN search was conducted on retrieved sequences to draw comparisons with similar published sequences (Altschul et al., 1997). Sequences were aligned using Multiple Sequence Comparison by Log -Expectation (MUSCLE), and consensus sequences were constructed using MEGAX (Kumar et al., 2018).
The 95% confidence intervals (CIs) were computed using the MetaXL add-in for Microsoft Excel (www.epigear.com). The effect of tick sex and sampling location (inland or coastal) on prevalence was tested using Fisher's exact test, and a p-value <0.05 was considered statistically significant. Statistical analyses were conducted using IBM SPSS V26.

Amblyomma hebraeum haplotype variation
Haplotype accumulation curves were constructed in R version 4.1.2 to assess whether our sample sizes provided total coverage of the haplotypes present in the populations sampled using the HaploAccum function in the package Spider with 1000 permutations (Brown et al., 2012). The Chao 1 richness estimator of total haplotype diversity and an abundance-based coverage estimator (ACE) was calculated using the Spider package's chaohaplo function.

Phylogenetic analysis
The jModelTest (Posada, 2008) was used to determine the most suitable nucleotide substitution model using the Akaike information criteria (AIC) for Maximum Likelihood (ML) and Bayesian Inference (BI) analyses. The ML trees for COI and ITS2 haplotypes were constructed using the Tamura 3-parameter model in MEGA X. A BI topology was inferred under the Generalized time-reversible model + Gamma in MrBayes 3.2.7 (Huelsenbeck & Ronquist, 2001) for the CO1 and ITS2 haplotype datasets.
The ML trees for the 17 kDa and ompA sequences were constructed using the Jukes-Cantor and Tamura 3-parameter model,

Population structure and diversity
Haplotype networks were constructed using the haploNet function in the R package pegas to visualize population structure (Paradis, 2010).
Haplotype diversity, nucleotide diversity, and haplotype network branch diversity (HBd) were calculated in R version 4.1.
2. An analysis of molecular variance (AMOVA) was conducted in Arlequin v 3.5 (Excoffier & Lischer, 2010) to determine genetic structure within and among population groups. We assessed population structure by generating pairwise Φ ST values among sites.

Demographic history
Tajima's D statistic was tested against selective neutrality. Population equilibrium and Fu's F statistic were calculated to detect population expansions by estimating departures from neutrality in Arlequin v 3.5.
Mismatch distributions were constructed to visualize potential demographic expansion for the entire population and test the null hypothesis of population growth (Rogers & Harpending, 1992). The raggedness index (r) was calculated to test if the frequency of pairwise nucleotide differences followed a smooth unimodal curve expected from a growing population.

Genetic diversity of A. hebraeum
Forty COI and 31 ITS2 sequences of A. hebraeum were obtained across 13 sites in the Eastern Cape province of South Africa. The COI and ITS2 sequence fragments were 581 and 596 bp, respectively. A total number of four COI and three ITS2 variable sites constituting one COI and three ITS2 parsimony informative sites were determined.
The homology between GenBank and study sequences varied from 98% to 100% (Table S1).
Conversely, the ITS2 Hd (0.675) decreased slightly with the inclusion of GenBank sequences (Table 1). Nucleotide diversity was higher among ITS2 sequences (0.0001) when compared to COI sequences (0.0008). The nucleotide diversity increased when COI GenBank sequences were included (0.0017) and remained the same when ITS2 GenBank sequences were included.

Haplotype rarefaction
The population haplotype rarefaction curves ( Figure S1) did not reach an asymptote for both genes, indicating that only part of the actual haplotype diversity had been sampled.

Phylogenetic relationships
The COI BI/ML trees (Figures 3, 4) were based on unique haplotypes, with added outgroup sister taxa (Hyalomma marginatum for COI and  (Table S3).

Demographic history
Evidence for population expansion was analysed using Tajima's D and  This population expansion scenario has been observed previously in A. hebraeum (Cangi et al., 2013). Local and long dispersal events may be attributed to the high number of domestic cattle in the Eastern Cape (Horak et al., 2017). Communal grazing practices that allow cattle to share pastures with neighbouring herds are likely to support the local dispersal of ticks. However, domestic cattle are also frequently translocated for anthropogenic reasons, facilitating long dispersal events. Amblyomma hebraeum was more abundant at cattle and wildlife interfaces than at pastures grazed by cattle in the Eastern Cape (Smith & Parker, 2010). Previous studies did not find A. hebraeum on rodents and passerine birds, suggesting low dispersal capabilities of A. hebraeum on these animals (Hasle et al., 2009;Horak et al., 2017). Introducing A. hebraeum onto novel hosts, especially wildlife, can have damaging effects such as increased host mortality and decreased reproductive output, causing severe population reductions (Halajian et al., 2016;Portillo et al., 2007).
This shallow, small-scale genetic differentiation may reflect some biological relevance due to the considerable distance between Dontsa and Caquba ($102 km), which may present a barrier to gene flow.
Alternatively, this subtle differentiation might be a mere effect of sampling bias. The Φ ST values suggest genetic homogeneity over the geographic range sampled except for four sites where significant population differentiation was observed. The climate in the Eastern Cape is stable for the maintenance and proliferation of A. hebraeum on various vegetation throughout the year (Yawa et al., 2018). In addition, cattle dipping could play a role in reducing the diversity of A. hebraeum as seen in Rhipicephalus microplus, where resistant populations would be selected over generations (Abbas et al., 2014).
The phylogenetic trees provided poor resolution and were comparatively homogenous, except for one ITS2 clade (Figure 4), where a recent split was observed between haplotypes I1 and I2. The ITS2 split was not observed in the COI phylogeny, which may be due to the differences in nuclear and mitochondrial DNA inheritance, which can affect estimates of gene flow (Presa et al., 2002). PCR and sequencing results showed that only R. africae was detected and was present in 46.92% (122/260) of A. hebraeum. Previous studies in southern Africa support these findings, with a 30%-80% prevalence from A. hebraeum collected from large ruminants (Halajian et al., 2016;Magaia et al., 2020;Mtshali et al., 2016;Pillay & Mukaratirwa, 2020). The higher prevalence of R. africae observed in both males and females could result from feeding on bacteremic cattle hosts (Parola and Raoult, 2001). The higher prevalence of R. africae observed in females than in males could be attributed to a large number of partially engorged females, 29.2% (38/130) collected.
The A. hebraeum female scutum only covers a small portion of the dorsal surface, allowing the females to ingest more blood than males and, therefore, more R. africae (Walker et al. 2003).
The prevalence of R. africae in cattle blood from the Eastern Cape was (22.22%; 20/90), and the prevalence was much lower (10.9%-15.7%) in ticks collected from smaller ruminants (Iweriebor et al., 2020;Jongejan et al., 2020). Our results support previous studies on the adaptability of R. africae to A. hebraeum as a significant reservoir for the pathogen (Fournier et al., 2009). The observed differences in prevalence estimates between cattle-collected ticks and those collected from small ruminants are likely due to rickettsemia in cattle, which may be a source of infection for ticks (Adjou Moumouni et al., 2016).
The 100% similarity among study R. africae isolates and GenBank isolates at two partial gene sequences suggests a single southern Africa genotype. Low genetic diversity was observed previously in R. africae, even with the most discriminatory genotyping method for Rickettsia spp. (Fournier et al., 2009;Mediannikov et al., 2010). The phylogeny supports this result as a clear separation between clades containing R. africae isolates from southern Africa and Sub-Saharan Africa was observed. This could explain the homogeneity observed in our study, that A. hebraeum is highly adapted to R. africae isolates from southern Africa as R. africae isolates from sub-Saharan Africa are more genetically diverse (Kimita et al., 2016;Mediannikov et al., 2012). In addition, the more extensive geographic range of A. variegatum has been shown to restrict A. hebraeum to southern Africa due to interspecific competition resulting in barriers to gene flow (Bournez et al., 2015).
The phylogenetic tree showed that all study R. africae isolates exhibit a monophyletic relationship with GenBank R. africae isolates.
The lack of association between haplotypes and the phylogeny of R. africae positive and negative ticks agreed with a previous study (Kisten et al., 2021), which showed no significant variation in microbial activity diversity between R. africae positive and negative A. hebraeum. These results were attributed to the proximity between geographic sites and similar environmental conditions. However, our study sites were geographically separated; as such, the observed results were likely due to the lack of genetic structuring observed in  (Katswara & Mukaratirwa, 2021).

AUTHOR CONTRIBUTIONS
Samson Mukaratirwa conceived and designed the study and critically revised the manuscript. Alicia Pillay performed the experiments, analysed the data and drafted the manuscript. Nkululeko Nyangiwe participated in collection of samples and in revision of draft manuscript.
All authors read and approved the final manuscript.

ACKNOWLEDGMENTS
The authors thank the farmers at sampling sites in the Eastern Cape for allowing the collection of samples from their cattle. The National Research Foundation (NRF) of South Africa and the NIH grant 1R01AI136035 as part of the joint NIH-NSF-USDA Ecology and Evolution of Infectious Diseases program for financial support. The funding body had no role in the study's design and interpretation of data and writing the manuscript.

CONFLICT OF INTEREST
The authors declare there are no competing interests.

DATA AVAILABILITY STATEMENT
The dataset(s) supporting the conclusions of this article are available in the GenBank repository (http://www.ncbi.nlm.nih.gov/genbank/).
The ITS2 sequences were deposited under the accession numbers OK635793-OK635818 and COI accession numbers OM212676-OM212713. The ompA sequences were deposited under the accession OM249800-OM249865 and 17 kDa accession numbers OM249832-OM249865.

ETHICS STATEMENT
Ethical approval for the study was obtained from the Animal Research Ethics Committee of the University of KwaZulu-Natal (AREC/056/017).

SUPPORTING INFORMATION
Additional supporting information can be found online in the Supporting Information section at the end of this article.