Haemoplasma Prevalence and Diversity in Three Invasive Rattus Species from Gauteng Province, South Africa

Invasive Rattus species are carriers of haemotropic Mycoplasmas (haemoplasmas) globally, but data from Africa are lacking. Using a PCR-sequencing approach, we assessed haemoplasma prevalence and diversity in kidney and buccal swabs collected from three invasive Rattus species (Rattus rattus, R. norvegicus and R. tanezumi) in Gauteng Province, South Africa. Whilst the overall sequence-confirmed haemoplasma prevalence was 38.4%, infection rates in R. rattus (58.3%) were significantly higher (χ2 = 12.96; df = 2; n = 99 p < 0.05) than for R. tanezumi (14.3%). Differences between host sex (χ2 = 3.59 × 10−31; df = 1; n = 99; p = 1.00) and age (χ2 = 4.28; df = 2; n = 99; p = 0.12) were not significant. Whilst buccal (1.01%) and ectoparasite positivity (2.13%) were low, these results suggest that multiple transmission routes are possible. Three phylogenetically distinct lineages, consistent with global rat-associated strains described to date, were detected, namely, ‘Candidatus Mycoplasma haemomuris subsp. Ratti’, and two Rattus-specific haemoplasmas that are yet to be formally described. These results expand the known distribution of invasive rat-associated haemoplasmas and highlight the potential for pathogen co-invasion of new territories together with invading rodent hosts.


Introduction
Haemotropic mycoplasmas or haemoplasmas represent a group of potentially emergent, unculturable bacteria that have garnered interest as the causative agents of infectious anaemia or haemoplasmosis in a wide range of mammalian species [1,2]. Originally classified into two separate genera, viz. Haemobartonella and Eperythrozoon, within the family Anaplasmataceae, members of this group have been reclassified within the genus Mycoplasma based solely on the 16S rRNA gene phylogeny, a decision which remains contentious [3][4][5][6]. This over-reliance on the highly conserved ribosomal gene is due to the difficulties with in vitro culture, low levels of detection associated with microscopic examination of blood smears, and the lack of polymerase chain reaction (PCR) assays with a broad species recognition range [6,7]. As a result, 16S rRNA amplification and sequencing are currently the gold standards for detecting and characterising haemoplasmas [8]. PCR assays targeting the 23S rRNA and the Rnase P gene regions are used to a lesser degree [7,8], and alternative targets such as gapA, dnak, and gyrB have limited applications as these assays have a narrow species-recognition range. Consequently, 16S rRNA gene Gauteng Province, the economic hub of South Africa, represents a densely populated region with the highest levels of HIV/TB co-infections in the country, placing immunocompromised individuals at risk of more severe infection should they encounter zoonotic pathogens [41,42]. Within this region, many HIV-and TB-positive patients live in informal rural, urban, and peri-urban settlements, landscapes in which three synanthropic and invasive Rattus species, viz. R. rattus, R. norvegicus and R. tanezumi thrive where direct and indirect inter-specific contact is frequent [39]. The distributional range of these three Rattus species varies in South Africa [43]. Rattus Rattus is predicted to occur mostly in coastal areas in the Western Cape, Eastern Cape, and KwaZulu-Natal Provinces and in urban inland areas in the Gauteng and Free State Provinces of South Africa, R. norvegicus is predicted to occur in coastal areas as well as urban inland areas in the Gauteng and Limpopo Provinces of South Africa, and R. tanezumi is predicted to occur predominantly inland in areas such as Gauteng and Mpumalanga Provinces and along the northeast and southeast coast of the country, with the distributions of these three Rattus species shown to overlap, especially in urban areas, such as in informal settlements in Gauteng Province, where niche partitioning of food resources and shelter likely allows them to co-exist with each other and indigenous species such as Mastomys spp. [43].
Furthermore, considering the need to better understand the dynamics of haemoplasmas cycling in free-living species occurring on the African continent, the aim of this study was to investigate the prevalence, diversity, and potential transmission routes of haemoplasmas in three Rattus species (R. rattus, R. norvegicus, and R. tanezumi) sampled from Gauteng Province of South Africa using conventional PCR methods. It was predicted that haemoplasmas detected in the three Rattus species would be closely related to haemoplasmas previously detected in other Rattus species sampled from Japan, Hungary, and Brazil. As R. norvegicus is the most aggressive of the three and has been shown to be infected by up to 1.4 times more ectoparasites compared to R. rattus [43,44]. It was predicted that R. norvegicus would display the highest haemoplasma prevalence of the three Rattus species assessed. Furthermore, it was predicted that haemoplasmas would be detected both in buccal swabs and associated ectoparasites collected from the Rattus samples screened for haemoplasmas, and that they would likely use more than one concurrent transmission route [7]. Finally, it was predicted that male individuals would likely show a higher haemoplasma prevalence than females and that adult Rattus individuals would display a higher haemoplasma prevalence compared to younger individuals.

Sample Collection
Rodents belonging to the genus Rattus (R. rattus, R. norvegicus, and R. tanezumi) in this study were sampled from in and around office buildings, storage facilities, and formal and informal human dwellings in 11 different localities across Gauteng Province, South Africa ( Figure 1; Table 1). These samples were sourced from two prior studies investigating the genetics and/or pathogen diversity and transmission routes of invasive Rattus species [39,45].
Samples were obtained under permit number CPF6 0032 from Gauteng Department of Nature Conservation, Johannesburg, South Africa. Permit numbers 13788 and 12/11/1/1(2022MvA) to conduct this study were granted under the terms of Section 20 of the Animal Diseases Act (Act no. 35 of 1984) of the South African Department of Agriculture, Forestry and Fisheries. An animal ethics approval number ECO25-10 and NAS059/2021 to conduct this study was granted by the Animal Ethics Committee of the University of Pretoria, Pretoria, South Africa.   Hammanskraal  0  0  12  12  Hatfield  33  0  1  34  Tembisa  0  32  0  32  Villeria  0  0  1  1  Centurion  0  0  1  1  Mountain View  0  0  6  6  Diepsloot  0  3  0  3  Garsfontein  1  0  1  2  Menlyn  2  0  0  2  Boschkop  0  0  5  5  Rietfontein  0  0  1  1   Total  36  35  28  99 DNA was extracted from buccal swabs collected from 99 rats and stored in 1:1 phosphatebuffered saline (PBS)/glycerol solution and from kidney samples of the corresponding animals. Buccal swabs were selected to test if haemoplasmas could be detected in the saliva of free-living rodent species based on a prior report of haemoplasma presence in saliva of experimentally infected rodents [11]. Kidney samples were selected as (a) they were the only tissue samples available from previous studies, (b) represent a blood-rich organ and (c) kidney samples have tested positive for haemoplasma presence, in some cases up to 11 months after inoculating test animals with haemoplasma-infected blood [46,47]. The Roche High Pure PCR Template Preparation Kit (Roche Diagnostics, Johannesburg, South Africa) was used to prepare the DNA extracts made available for this study (Table S1).
In addition, DNA extracts of 47 individual ectoparasites collected from 18 rats were included in this study. Ectoparasites were individually removed by combing the fur of Rattus specimens in a biological safety hood. All ectoparasites collected were preserved in 100% ethanol and subsequently sorted into mites, ticks, lice, and fleas for further identification. Ectoparasites were removed from 100% ethanol and rinsed three times with double distilled water. Whole ectoparasites were individually crushed using sterile sand and a pestle and eluted in 200 µL of PBS. Genomic DNA was extracted from ectoparasite homogenates of ticks using the High Pure PCR Template Preparation Kit (Roche Diagnostics, Johannesburg, South Africa), according to the manufacturer's protocol, with the exception that ectoparasite DNA was eluted in 100 µL instead of 200 µL. All extracts were stored at −20 • C until further processing.

Rattus Age Classes
Based on the level of maxillary molar toothrow cusp eruption and wear [48], samples belonging to the three Rattus species were grouped into five relative age classes (tooth-wear class I-V). Following the methods of others [39,49], these tooth-wear classes were further divided, where class I represented juvenile individuals, age classes II and III represented sub-adults, and age classes IV and V represented adults.

Rattus and Ectoparasite Species Identification
Species identification of all Rattus samples was confirmed through the amplification of the cytochrome b (cyt b) gene region as part of a prior study [39]. The ectoparasites used in the current study were identified by using published morphological keys for fleas [50][51][52][53], ticks [54][55][56][57][58][59] and mites [60][61][62]. Fleas and ticks were identified to species level while mites were identified to genus level.

Additional Gene Regions
A subset of sequence-confirmed haemoplasma positive samples (n = 10), which were representative of all haemoplasma variants recovered during the initial 16S rRNA screening phase, were selected for amplification with additional published primer sets targeting the Rnase P, dnaK, and gapA gene regions ( Table 2).

Polymerase Chain Reaction (PCR) Amplification and Nucleotide Sequencing
PCRs were performed in a final reaction volume of 40 µL, containing a final concentration of 1 × Dream Taq Buffer, 0.2 µM dNTPs (Fermentas), 1.5 U of Dream Taq (Thermo Fisher Scientific, Waltham, MA, USA), 0.4 µM of each primer, and 3 µL of template DNA. Touchdown PCRs with an initial denaturation at 96 • C for 12 s, primer annealing for 30 s at annealing temperatures optimised for each assay (Table 2), elongation at 70 • C with variable time (depending on the size of amplicon targeted) and final elongation at 70 • C for 1 min, were performed on the same ABI 2720 thermal cycler (Applied Biosystems, Foster City, CA, USA)). The PCR products were separated by 1.5% agarose gel electrophoresis and size was estimated against a DNA molecular weight marker (Fermentas, Waltham, MA, USA). All products of the correct size were purified directly from the tube using the Roche High Pure PCR product purification kit (Roche Diagnostics GmbH, Mannheim, Germany) and supplier-prescribed protocols.
Purified DNA was cycle sequenced using BigDye Terminator Cycle Sequencing Ready Reaction Kit (Applied Biosystems, Foster City, CA, USA). Sequence chromatograms viewed in the Chromas programme in MEGA version 6 [64] were edited and aligned to generate contiguous sequences and used in nucleotide BLAST (BlastN) searches against the GenBank database (www.ncbi.nlm.nih.gov/blast), accessed on 10 October 2021) to identify the closest sequence matches [65].

Phylogenetic Analyses
Both the 16S rRNA and Rnase P nucleotide sequences generated through screening Rattus kidney samples in this study were complemented with reference sequences and aligned with ClustalW in Mega 6 [64]. For the 16S rRNA dataset, Mycoplasma fastidiosum, a closely related sister taxon to the haemotropic Mycoplasma species lineage, was included as an outgroup. For the Rnase P gene region, Mycoplasma fastidiosum and Mycoplasma leachii were included as outgroups. The final aligned 16S rRNA and Rnase P datasets were used to identify the best-fit model of sequence evolution under the Bayesian Information Criterion (BIC) in Mega 6 [64]. For the 16S rRNA gene region, the best-fit model of sequence evolution was the Tamura-Nei model, while the best-fit model of sequence evolution was the Tamura 3-parameter model for the Rnase P gene region. These models of sequence evolution were used for Minimum Evolution (ME) and Maximum Likelihood (ML) analyses performed in Mega 6 [66] and phyML [67], respectively. Nodal support was evaluated through 10,000 non-parametric bootstrap replications. The best-fit models guided the selection of priors for Bayesian Inference (BI) performed in MrBayes [68,69]. For both the 16S rRNA and the Rnase P BI analyses, trees were sampled every 100th iteration, and MCMC Trace Analysis Tool version 1.6.0 [70] was used to confirm a 25% burn-in.

Statistical Analyses
Chi-square (χ 2 ) test was used to determine differences in haemoplasma prevalence between host species, sex, and age. Multi-comparison Fisher's exact test with a Bonferroni correction was used post hoc to find which host species had significantly different haemoplamsa prevalence. Fisher's exact test was used to determine differences in haemoplasma prevalence between host sex per Rattus species. Statistical analyses were performed using algorithms in the statistical program R with the use of R Studio [71].

Haemoplasma Prevalence in Ectoparasites
Of the 47 ectoparasites screened, only one of the nine positive extracts produced a 120 nucleotide (nt) stretch of unambiguous sequence with the MyChlo-F/Mycop-R primer set (sequence-confirmed haemoplasma prevalence of 2.1%). All other sequences are comprised of mixtures. Nucleotide BlastN searches confirmed that the 120 nt fragment had a 94.1% sequence identity to a haemoplasma sequence (Genbank accession number: MK295631), which has previously been detected in the small big-eared brown bat (Histiotus montanus) ( Figure S1). Of the 47 ectoparasites screened with the Myco16S-322s/HemMycop16S-1420as primer set, 13 produced amplicons of the correct size and were cycle sequenced, with BlastN searches revealing that the closest matches for seven samples were to non-target bacterial genera (Table S2).

Haemoplasma Prevalence in Buccal Swabs
Of the 99 Rattus buccal swabs screened, one sample of the 14 positive amplicons detected produced a clean haemoplasma sequence during the initial 16S rRNA screening (sequence-confirmed haemoplasma prevalence of 1.0%). The haemoplasma genome presence in this buccal swab sample from an R. tanezumi individual was confirmed through two separate PCR amplifications and nucleotide sequencing events with the MyClos-1F/Mycop-1R primer set. Similarly, the corresponding kidney sample of this individual was confirmed to be haemoplasma-positive through nucleotide sequencing and to be identical to the 16S rRNA gene fragment detected in the buccal swab. Despite multiple attempts, it was not possible to generate a larger 16S rRNA fragment or generate sequence data for alternative gene regions for this haemoplasma-positive buccal swab.

Haemoplasma Prevalence in Rattus Kidneys
Of the 99 Rattus kidney samples screened, 38 of the positive amplicons detected produced unambiguous haemoplasma 16S rRNA sequences, corresponding to an overall sequence-confirmed haemoplasma prevalence of 38.4% in all Rattus kidney samples screened. In total, three distinct haemoplasma genotypes were detected in the three Rattus species screened. Genotype 1, which was detected in 21 R. rattus kidney samples (58.3%), 13 R. norvegicus kidney samples (37.1%), and four R. tanezumi kidney samples (14.3%), was the most prevalent genotype, occurring at an overall prevalence of 34.3%, followed by genotype 3 (overall prevalence of 2.0%), which was detected in two R. norvegicus samples, and genotype 2 (overall prevalence of 1.0%), which was identified from a single R. rattus kidney sample ( Figure 2). 58

16S rRNA Nucleotide Searches and Phylogenetic Analyses
Through amplification and sequencing using primer sets Myco16S-322s/HemMycop16S-1420as and 27F/Mycop-1R, contiguous 16S rRNA sequences, 1063 nt in length, were generated for all three haemoplasma genotypes detected in kidney samples, based on the initial 16S rRNA screening with the MyChlo-1F/Mycop-1R primer set. These were complemented with reference sequences obtained from the Genbank database and resulted in a final aligned 16S rRNA dataset, 1063 nt in length and comprising 67 taxa. Phylogenetic analyses revealed that all three Rattus-associated haemoplasmas detected in the current study were phylogenetically distinct from each other and fell into three separate, well-supported clusters ( Figure 3). Nucleotide BlastN searches revealed that Genotype 1, which was detected in all three Rattus species, showed 100% sequence identity to a haemoplasma sequence previously detected in R. rattus sampled from Brazil (KT215635). The 16S rRNA phylogeny further confirmed that Genotype 1 clustered with Mycoplasma haemomuris strains detected in R. rattus from Brazil and Japan (96-100% nodal support; Figure 3).   -Genotype 1, which formed a well-supported clade with M. haemomuris, is indicated in green, Genotype 2 is indicated in orange and Genotype 3 is indicated in blue. The three haemoplasma genotypes detected in this study fell within the Haemofelis cluster (circled at the node) and formed three, separate well-supported clades with haemoplasmas strains previously detected in Rattus species (highlighted in grey). The tree was inferred using the aligned 1063 nucleotide data set of the 16S rRNA gene region. Bootstrap support values ≥70% from the minimum evolution analysis (ME; 10,000 bootstrap replicates), maximum likelihood (ML; 10,000 bootstrap replicates); posterior probability support values ≥90% from the Bayesian inference (BI; 1,000,000 generations sampled every 10 generations with a 25% burn-in) are indicated as ME/ML/BI next to each node. Genotype 1 also formed a well-supported sister clade to the subspecies "Candidatus Mycoplasma subsp. musculi" (sequence identity of 99.4% to U82963; AB758436 and AB75837), previously detected in Apodemus argenteus in Japan. Phylogenetic analyses and nucleotide BlastN searches showed that Genotype 2 formed a well-supported cluster with novel haemoplasma sequences previously detected in Rattus species from Brazil (100% sequence identity to MN423263, KT215643, KT215640 and KT215642) and was closely related to strains from Hungary (99.8% sequence identity to a haemoplasma detected in R. norvegicus, KJ739312) and Japan (99.5% sequence identity to a haemoplasma detected in R. norvegicus, AB752303) ( Figure 3). Genotype 3 formed a well-supported cluster with haemoplasma sequences previously detected in R. norvegicus samples from Hungary (97.8% sequence identity to KJ739311) and Brazil (97.8% sequence identity to KM203857). Together, these Rattus-associated haemoplasma strains formed a lineage sister to a novel haemoplasma sequence previously detected in Micromys minutus from Hungary (97.7% sequence identity to KC863983). All 16S rRNA gene sequences generated in this study were submitted to Genbank under accession numbers ON733032-ON733034.

Alternative Gene Regions
Despite multiple attempts, including varying the annealing temperatures (Table 1) and extension times, as well as trying multiple primer combinations ( Table 1), none of the samples confirmed positive by 16S rRNA screening, amplified with primer sets targeting the dnaK and gapA gene regions. Similarly, despite multiple attempts, only one positive R. norvegicus kidney sample was amplified with the Rnase P gene primers, with all other samples screened either failing to amplify, producing mixed amplicons, or showing nonspecific amplification (Table S3). The Rnase P sequence was generated for a sample identified as genotype 3 by the 16S rRNA gene analysis (Figures 3 and 4). The final aligned Rnase P dataset consisted of 28 taxa and had a total length of 225 nt. Phylogenetic analyses showed that this sequence fell within the haemofelis cluster with a high level of support (98%, 95%, and 100 nodal support for ME, ML, and BI, respectively) but could not resolve close phylogenetic relationships between the detected sequence and reference haemoplasma sequences obtained from the Genbank database ( Figure 4). Nucleotide BlastN searches indicated that the sequence detected had the highest sequence match (84.5%) to 'Candidatus Mycoplasma haemohominis' (Genbank accession number: GU562825). The RnaseP gene sequence generated in this study was submitted to Genbank under accession number ON684278.   . Minimum evolution tree depicting genetic relationships of a Rattus-associated hemoplasma genotype, indicated in blue, detected in a R. norvegicus sampled in the Gauteng Province of South Africa, with relevant reference sequences. Genbank accession numbers of the three haemoplasmas detected in the current study are indicated. This haemoplasma genotype fell within the Haemofelis cluster (circled at the node), but close phylogenetic relationships could not be resolved. The tree was inferred using the aligned 225 nucleotide data set of the Rnase P gene region. Bootstrap support values ≥70% from the minimum evolution analysis (ME; 10,000 bootstrap replicates), maximum likelihood (ML; 10,000 bootstrap replicates); posterior probability support values ≥90% from the Bayesian inference (BI; 1,000,000 generations sampled every 10 generations with a 25% burn-in) are indicated as ME/ML/BI next to each node.

Discussion
Haemoplasmas have been detected in a diverse range of mammalian species, including species of artiodactyls, carnivores, primates, marsupials, chiroptera, and rodents [7]. Haemoplasmas vary from being asymptomatic in some hosts, to causing potentially fatal haemoplasmosis or haemolytic anaemia in other hosts, such as domestic cats, dogs, and humans [7]. However, data on hemoplasma prevalence in free-living species worldwide and an understanding of the potential routes of hemoplasma transmission are limited [2,72]. This study represents the first report of haemoplasmas cycling in free-living Rattus species sampled from South Africa. Rodents are well-known carriers of various pathogens, including haemoplasmas, which have been reported in diverse rodent species sampled from Hungary, Brazil, Japan, Israel, Chile, and South Africa [8,11,21,23,25,40]. The haemoplasma prevalence of 38.4% found in the current study falls within the range of haemoplasma prevalence reported previously in free-living Rattus species sampled from Japan (overall prevalence of 11.1% detected in R. norvegicus [21]), Hungary (overall prevalence of 92.9%, detected in R. norvegicus [23]), and Brazil (overall prevalence of 63.5%, detected in R. norvegicus [8], the overall prevalence of 46.2% detected in R. rattus [22] and the overall prevalence of 30.7% detected in R. rattus [16]). However, it should be noted that these previous studies screened for haemoplasma presence using DNA extracted either from blood [8,21] or spleen samples [16,22,23], while the current study used DNA extracted from kidney samples. Whilst this may have influenced the overall haemoplasma prevalence detected [7,46,47], it nevertheless serves to confirm haemoplasma presence and diversity in Rattus species sampled in the Gauteng Province of South Africa.
The observation that the three haemoplasma genotypes detected in the current study fell within the haemofelis cluster with high levels of support is in accordance with other Rattus-associated strains detected previously [8,16,[21][22][23]25]. Currently, only two rodentassociated haemoplasmas have been described, namely, Mycoplasma coccoides and M. haemomuris; however, in the last decade, novel haemoplasmas have been detected in a wide range of rodent species worldwide [7,25,73]. Moreover, while rodent-associated haemoplasmas are generally considered to be host-specific, a growing body of evidence suggests that rodent-associated haemoplasmas are capable of host switching and may hold zoonotic potential, making it imperative to understand the dynamics of haemoplasma cycling in various rodent species [15,25,26]. All three genotypes detected in the current study formed well-supported clades with haemoplasmas previously detected in Rattus species (Figure 3), supporting our prediction that the haemoplasmas detected by the current study would be phylogenetically similar to haemoplasmas previously detected in Rattus species. These results also support, in part, the prediction that Rattus-associated haemoplasmas are likely restricted to causing infection in closely related Rattus species, with unknown barriers expected to prevent haemoplasma transmission between synanthropic Rattus species and wildlife occupying the same area [16,22].
Certainly, the inter-and intra-specific routes of haemoplasma transmission remain poorly understood [7,11]. Traditionally, emphasis has been placed on shared ectoparasite vectors as the main route of haemoplasma transmission [22,23]. The current study could only detect one sequence-confirmed haemoplasma-positive in a Haemaphysalis elliptica tick, representing the first detection of haemoplasmas in an ectoparasite sampled from South Africa. This tick is a generalist species, associated with a wide range of rodent species during its immature stages, while adults are found on both wild and domestic carnivores [59,74]. However, the fact that the current study could only detect one haemoplasma-positive case in H. elliptica collected from a haemoplasma-negative R. tanezumi individual, is in accordance with other studies reporting discrepancies in haemoplasma infection status between ectoparasites and their host species [24,26,75]. Previous studies either report discrepancies in haemoplasma infection status between ectoparasites and their host species, fail to detect haemoplasmas in ectoparasites collected from positive hosts or find no difference in haemoplasma prevalence between ectoparasite-free and ectoparasite-infested hosts [24,26,75]. For these reasons, some studies have suggested that ectoparasites may play a limited role, if any, in haemoplasma transmission [11,16].
In the current study, one buccal swab of a confirmed haemoplasma-positive R. tanezumi individual tested positive for a Mycoplasma haemomuris-like genotype (Figure 2), lending support to assertions that direct transmission during grooming or aggressive host behaviour may represent an important route of haemoplasma transmission [7,11]. This finding in a wild-caught rat complements a prior report of haemoplasma detection in buccal swabs of rodents (Gerbillus andersoni) experimentally infected with a Mycoplasma haemomuris-like haemoplasma [11]. Haemoplasma infection is characterised by an acute and a chronic phase [7], with Cohen et al. (2018) detecting a low haemoplasma load in buccal swabs only during peak haemoplasma infection of hosts. This may explain the low haemoplasma prevalence found in Rattus buccal swabs compared to the prevalence found in Rattus kidney samples in the current study ( Figure 2). Whilst no bite wounds were recorded for the rodents assessed, aggressive behaviour by Rattus norvegicus towards other sympatric congenerics is well-established [44] and thus a possible route of transmission. These results support the prediction in the current study that there may be more than one concurrent route of haemoplasma transmission. However, future studies are needed to determine whether oral haemoplasma presence in rats is transient and only associated with the bacteremic phase.
It has been suggested that males may display higher levels of haemoplasma prevalence due to more aggressive behaviour, which may facilitate the transfer of saliva from infected hosts into the open wounds of uninfected hosts [7]. However, just as in previous studies investigating haemoplasma prevalence in Rattus species [11,21], the current study found no significant difference in haemoplasma prevalence between male and female Rattus individuals for all three Rattus species assessed, lending support to the idea that sex biases in haemoplasma infection likely vary across mammalian taxa [7]. Therefore, the prediction that there would be a significant difference in haemoplasma prevalence between male and female Rattus samples screened was rejected. Similarly, it has been suggested that older individuals of certain mammalian taxa, such as felids and canids, may display a higher haemoplasma prevalence compared to younger individuals [7]; however, just as with other studies on American minks (Neovison vison) [76] and domestic dogs [77], the current study found no difference in haemoplasma prevalence between different age classes assessed, thereby rejecting the prediction that older Rattus individuals would display a higher haemoplasma prevalence compared to younger individuals.
Phylogenetically similar haemoplasma genotypes have been detected in Rattus species sampled from geographically disparate locations (Figure 3), suggesting that Rattus-associated haemoplasmas were likely introduced into novel environments, such as Brazil and South Africa, along with their host species during historic species introductions [16,22]. Members belonging to the Rattus genus likely reached the African continent primarily through the shipping trade but also overland [78]. Molecular analyses indicate that at least three separate R. rattus and two separate R. norvegicus introductions occurred in South Africa, and that R. tanezumi was likely introduced into this country during a single, more recent event [79]. While these three Rattus species have been found to occur in sympatry in urban areas, specifically informal settlements in Gauteng Province of South Africa, they vary in behaviour and likely exhibit niche partitioning in the urban environments where they co-exist [43].
Rattus norvegicus is an aggressive, mostly ground-dwelling species, which has been found to be infected with up to 1.4 times more ectoparasites compared to R. rattus, a less aggressive species, which is an adept climber, closely associated with urban environments, preferring upper floors and rooftops, which may limit its exposure to ectoparasites, such as ticks, which wait for their hosts on the ground [43,44]. Rattus tanezumi, which forms part of the R. rattus species complex and is known to hybridise with R. rattus and may outcompete the other two Rattus species with scent marking [43,44,80]. As is the case for R. rattus, this species is arboreal but also lives in and around human dwellings, agricultural lands, and natural sites [43,81].
We found that more than one haemoplasma genotype was present in both R. norvegicus and R. rattus, viz. genotypes 1 and 3, and genotypes 1 and 2, respectively, while only one haemoplasma genotype (genotype 1) was detected in R. tanezumi (Figures 2 and 3); however, the latter species also had the lowest infection rate. It should be noted that the R. norvegicus samples were collected from separate sampling sites compared to R. rattus and R. tanezumi (Figure 1), which may, in part, restrict the transmission of genotypes 2 and 3 between the different host species. Nevertheless, it is of interest that the current study found that R. rattus, which is both less aggressive and displays lower ectoparasite loads compared to R. norvegicus [44], had the highest haemoplasma prevalence, thereby rejecting the prediction that R. norvegicus would display the highest haemoplasma prevalence, lending credence to doubts regarding the importance of vector-borne transmission.
Importantly, the results of the current study also expand on species-specific differences reported for other pathogenic bacterial genera cycling within the three Rattus species in Gauteng Province metropoles. Julius et al. (2021) found a high overall Streptobacillus prevalence, with specific Streptobacillus species being associated with R. rattus/R. tanezumi hosts or R. norvegicus hosts, but not both, which is in contrast to the results of the current study that showed that genotype 1 was present in all three Rattus species sampled ( Figure 3). Moseley et al. (2020) found a high prevalence of Leptospira borgpetersenii (overall prevalence of 44%) cycling in, primarily R. norvegicus and one R. rattus individual within this geographical region [82]. Together, these studies confirm the presence of a broad range of potentially zoonotic bacterial pathogens cycling in Rattus species within Gauteng Province [80]. Moreover, the significantly lower levels of haemoplasma prevalence found in R. tanezumi compared to R. rattus is in accordance with what has been found when assessing broad-range pathogenic bacteria [83] as well as Streptobacillus [39]. In Rattus species sampled from the same geographical area, Julius et al. (2021) found an almost two-fold higher Streptobacillus prevalence in R. rattus compared to R. tanezumi, whilst in the current study, R. rattus had a haemoplasma prevalence of almost four times higher than R. tanezumi ( Figure 2). As R. tanezumi belongs to the same species complex as R. rattus [39] and given that in the current study, R. rattus and R. tanezumi were sampled in close proximity to each other (Figure 1), the results of the current study further support the suggestion by Julius et al. (2021) that there may be unknown genetic or behavioural factors, such as varying levels of aggressive behaviour or ectoparasite load, which may contribute to differences in pathogenic bacterial prevalence between these two morphologically indistinguishable host species.
Both phylogenetic analyses and nucleotide BlastN searches confirmed that the genotype detected in all three Rattus species (genotype 1) is closely related to M. haemomuris, which has been detected in R. rattus from Japan, Hungary, and Brazil [8,21,23]. Moreover, the relatively high prevalence of 38.4% in all samples screened indicates that, just as in other countries, this genotype is likely widespread in free-living Rattus species occurring within the Gauteng Province metropoles of South Africa, as this genotype was detected in Rattus species sampled from five of the 11 sampling localities (Figure 3). Mycoplasma haemomuris has been divided into two subgroups, viz. "Candidatus Mycoplasma haemomuris subsp. ratti" and "Candidatus Mycoplasma haemomuris subsp. musculi" [21,73]. This division is based on genetic differences in the 16S rRNA, 16S-23S rRNA intergenic spacer (ITS), Rnase P, and dnak gene regions, with "Candidatus Mycoplasma haemomuris subsp. ratti" associated with Rattus species sampled from Japan, Hungary, and Brazil, while "Candidatus Mycoplasma haemomuris subsp. musculi" is associated with wild mice sampled across Japan ( Figure 3) [21,73]. Mycoplasma haemomuris is not considered to cause clear clinical symptoms in infected host species [8].
However, in addition to M. haemomuris, two additional haemoplasma genotypes were also detected in the Rattus kidney samples screened (Figure 3). Genotype 2, detected in one R. rattus individual, is identical to Rattus-associated haemoplasmas detected in Brazil, Hungary, and Japan ( Figure 3), with the authors of prior studies suggesting that this genotype likely represents an undescribed Rattus-associated haemoplasma species [16,[21][22][23]. Furthermore, in contrast to Genotypes 1 and 2, which showed 100% identities to haemoplasmas detected previously, phylogenetic analyses and nucleotide BlastN searches revealed that Genotype 3, detected in two R. norvegicus individuals, represents a novel haemoplasma genotype ( Figure 3). This novel genotype was most closely related to a haemoplasma genotype detected in R. norvegicus individuals sampled from Brazil [8] and Hungary [23]. Phylogenetic analyses of the 16S rRNA gene region indicate that the Rattus-associated genotypes are closely related to haemoplasmas detected in wild rodents, domestic cats, and grey foxes (Figure 3), suggesting that, whilst rare, cross-species transmission of Rattusassociated haemoplasmas occurred at some point. This again emphasises the need to better understand the inter-and intra-specific routes of haemoplasma transmission in wild and synanthropic rodent species [7,8].
Based on its phylogenetic distance from Mycoplasma haemomuris, Conrado et al. (2015) suggest that the clade within which genotype 3 falls (Figure 3) may represent a novel haemoplasma species, but that the generation of additional gene regions is needed to confirm species status. While the PCR-assays targeting the dnaK and gapA gene regions failed to produce any amplicons in the current study, we were able to successfully generate a single Rnase P sequence for a genotype 3 strain detected in the kidney sample of an R. norvegicus individual (Figure 4). Both nucleotide BlastN searches and phylogenetic analyses for this gene region showed that, currently, no close sequence matches are available for this genotype for the Rnase P gene region ( Figure 4). While the 16S rRNA gene region is widely used in the identification of unculturable microorganisms and currently forms the bulk of the genetic data available for haemoplasmas on online databases [7,8], these results emphasise the need for data on phylogenetically more informative gene regions such as Rnase P, dnak, and gapA for haemoplasmas. This is currently constrained by the lack of sensitive haemoplasma-specific PCR assays for alternative gene regions. This was underscored by the fact that we were only able to generate Rnase P data for one of the three genotypes detected. Other authors have reported similar difficulties in generating sequence data for alternative gene regions for rodent-associated haemoplasma strains [22].

Conclusions
This study is the first to report haemoplasma prevalence and diversity in three invasive, wild-caught Rattus species from South Africa. While blood and spleen samples and not kidney samples have been traditionally used when testing for haemoplasma prevalence [7], the haemoplasma prevalence of 38.4% found in kidney samples from free-living synanthropic Rattus species within the Gauteng Province provides an initial estimate of haemoplasma prevalence in this region and indicates that haemoplasmas are likely widespread in Rattus species within the Gauteng Province of South Africa. The sequence-confirmed presence of haemoplasma in a buccal swab provides support for the prediction that haemoplasmas may be transmitted through direct host-to-host contact during social or aggressive interactions [11] and, to our knowledge, is the first time haemoplasmas have been detected in the saliva of a free-living rodent species. While the haemoplasma genotypes detected here were likely introduced into South Africa during the historic introduction of Rattus species, phylogenetic analyses suggest that under certain circumstances, historical haemoplasma host-switching events may have occurred [8]. As the three Rattus species overlap in distribution, especially in urban areas, and likely have a much wider distribution within South Africa than previously reported [43], Rattus-associated haemoplasmas may be widespread in this country and require further investigation. Furthermore, the findings of the current study highlight the need for future studies to further investigate the potential routes of intra-and inter-specific haemoplasma transmission in both free-living synanthropic and wildlife species. Future studies should focus on screening different tissue types (including blood, spleen, lungs, and kidney) to identify the optimal blood-rich sample type for haemoplasma detection. In addition, the evaluation of foetal samples, ectoparasites as well as oral and rectal swabs and environmental samples would permit the assessment of the importance of multiple routes of transmission (vertical, vector, environmental) and their relative importance. The difficulties in generating sequence data for gene regions other than the 16S rRNA gene region highlight the need for the development of sensitive haemoplasma-specific PCR assays targeting alternative non-ribosomal gene regions for the accurate phylogenetic placement of novel haemoplasma genotypes.
Supplementary Materials: The following supporting information can be downloaded at: https://www. mdpi.com/article/10.3390/microorganisms10081632/s1, Figure S1. (a) Trimmed 120 nt fragment of the chromatogram generated when screening ectoparasite sampled using the primer set MyClost-F/Mycop-R. The colour lines denote the four nucleotide bases: adenine (A) is indicated in green, cytosine (C) is indicated in blue, guanine (G) is indicated in black and thymine (T) is indicated in red. (b) Alignment result to MK295631 produced when performing a nucleotide blast (BlastN) search for the 120 nt 16S rRNA gene fragment generated when screening a Haemaphysalis elliptica tick for haemoplasma prevalence using the MyClost-F/Mycop-R primer set. Table S1: Sample list of Rattus species and their associated ectoparasites screened in this study; Table S2: Nucleotide BlastN results obtained when performing nucleotide blast searches against the Genbank database for sequences generated when screening ectoparasites with the Myco-16S-322s/HemMycop16S-1420as primer set. Table S3: PCR results obtained when screening DNA extracted from buccal swabs, kidney samples, and ectoparasite samples using haemoplasma-specific PCR assays targeting the Rnase P gene region [12]. PCR screening and nucleotide sequencing revealed that samples either failed to amplify or produced non-target sequences.  Data Availability Statement: Publicly available datasets were analyzed in this study. These data can be found here: [https://www.ncbi.nlm.nih.gov/genbank/] (accessed on 9 August 2021).