Bulinus snails in the Lake Victoria Basin in Kenya: Systematics and their role as hosts for schistosomes

The planorbid gastropod genus Bulinus consists of 38 species that vary in their ability to vector Schistosoma haematobium (the causative agent of human urogenital schistosomiasis), other Schistosoma species, and non-schistosome trematodes. Relying on sequence-based identifications of bulinids (partial cox1 and 16S) and Schistosoma (cox1 and ITS), we examined Bulinus species in the Lake Victoria Basin in Kenya for naturally acquired infections with Schistosoma species. We collected 6,133 bulinids from 11 sites between 2014–2021, 226 (3.7%) of which harbored Schistosoma infections. We found 4 Bulinus taxa from Lake Victoria (B. truncatus, B. tropicus, B. ugandae, and B. cf. transversalis), and an additional 4 from other habitats (B. globosus, B. productus, B. forskalii, and B. scalaris). S. haematobium infections were found in B. globosus and B. productus (with infections in the former predominating) whereas S. bovis infections were identified in B. globosus, B. productus, B. forskalii, and B. ugandae. No nuclear/mitochondrial discordance potentially indicative of S. haematobium/S. bovis hybridization was detected. We highlight the presence of Bulinus ugandae as a distinct lake-dwelling taxon closely related to B. globosus yet, unlike all other members of the B. africanus species group, is likely not a vector for S. haematobium, though it does exhibit susceptibility to S. bovis. Other lake-dwelling bulinids also lacked S. haematobium infections, supporting the possibility that they all lack compatibility with local S. haematobium, thereby preventing widespread transmission of urogenital schistosomiasis in the lake’s waters. We support B. productus as a distinct species from B. nasutus, B. scalaris as distinct from B. forskalii, and add further evidence for a B. globosus species complex with three lineages represented in Kenya alone. This study serves as an essential prelude for investigating why these patterns in compatibility exist and whether the underlying biological mechanisms may be exploited for the purpose of limiting schistosome transmission.


Introduction
One of the fascinating aspects of the biology of infectious diseases is that, in some cases, the parasites responsible and their hosts (including vectors) do not comprise a single parasite or host species, but complex arrays of related species [1][2][3][4]. Such arrays might reveal a checkerboard of host-parasite interactions, ranging from pairs of host and parasite species being fully compatible and supporting transmission to marginally compatible and fully incompatible pairs. The Schistosoma haematobium species group and species within the genus Bulinus comprise such an array [5]. We have directed our attention to representatives of these two groups of organisms that occupy the Kenyan waters of the Lake Victoria Basin (LVB), in hopes of eventually revealing the factors that dictate the various outcomes of such associations.
Like most digenetic trematodes, schistosomes depend on a gastropod intermediate host to complete their life cycles, within which vertebrate-infective cercariae are asexually produced in prolific numbers. Such gastropods are often termed intermediate hosts because of their obligatory role in schistosome larval development, and although they do not directly deliver parasites to their vertebrate hosts as, for example, a mosquito transmits malaria parasites, they nonetheless play an indispensable vector role. Successful transmission of vector-borne parasites like schistosomes is dependent on a variety of factors, including ecological circumstances which impact the encounter rates between parasite and vector [6][7][8], associations with particular symbionts including those that prey on the free-living forms of certain parasites [9], facilitated susceptibility where prior infection with a specific parasite allows a second parasite to develop in a non-typical host [10,11], and nuanced physiological and immunological interactions which dictate the outcome of an infection [2,[12][13][14][15][16][17][18][19][20].
To begin to fully appreciate the intimate relationships between vector and parasite that influence compatibility and ultimately transmission, a sound understanding of the underlying systematics of both parasite and vector host is critical. This task is complicated when the species involved cannot consistently be accurately differentiated using morphology alone, or when a lack of clear morphological differences belie the presence of large genetic differences, that is, when cryptic species are involved [21]. from the B. reticulatus group, B. reticulatus. Several of the species he reported are hard to differentiate from one another, and some are rarely encountered or studied and in general are poorly known, including browni, scalaris and reticulatum. More recently, based largely on the useful discrimination provided by the cytochrome c oxidase subunit 1 (cox1) gene, Chibwana et al. [50] found 7 species in the LVB, including B. globosus (described as a complex), B. truncatus, B. tropicus, B. nasutus productus and B. forskalii as well as two taxa, Bulinus sp. 1 and 2, provisionally identified as B. trigonus and B. ugandae, respectively. Currently accepted species are associated (with some variation) with ephemeral pools or ponds (e.g. B. forskalii, B. scalaris, B. reticulatus), seasonal ponds or springs (Bulinus productus), more permanent habits such as streams or dams (Bulinus globosus), the lakeshore and associated papyrus swamps (Bulinus ugandae), or the deeper waters of the lake (B. tropicus, B. truncatus and B. trigonus). Of particular note is a growing body of evidence that the pan-African species B. globosus is not a single species, but a complex of multiple species [35,38,46,50].
Based on an examination of the relevant literature, coupled with sequence data of marker genes to aid in the identification of both snails and the schistosomes they host, we provide an overview of the bulinid species we have recovered from various habitats in the LVB. We highlight some difficulties regarding Bulinus systematics and identify some peculiarities regarding the role of bulinids in the transmission of S. haematobium and S. bovis in western Kenya. This study serves as a prelude to investigations aimed at understanding the underlying causes dictating the patterns of compatibility posed by the complex interacting arrays of Schistosoma and Bulinus species in western Kenya.

Ethics statement
Informed written consent was obtained from all individual participants included in the study. The Kenya Medical Research Institute Scientific and Ethics Review Unit (KEMRI/SERU/ CBRD/173/3540) and the University of New Mexico Institution Review Board (IRB 821021-12, IRB 821021-9) approved all aspects of this project involving human subjects. Ethical approval for the collection and analyses of snail and schistosome samples were obtained from the National Commission for Science, Technology and Innovation (permits number NACOSTI/P/21/9648 and NACOSTI/P/22/17142), and National Environmental Management Authority (permit number NEMA/AGR/149/2021).

Sampling
We collected Bulinus snails from 11 different localities (S1 Table and S1 Fig); some localities include endemic transmission sites where we have collected from Jan 2014 -Mar 2021. Two methods were used to collect snails: scooping from the shore and dredging from a boat [66]. From the shore, two experienced lab members scooped snails for 30 minutes per sampling site using long-handled scoops (steel sieve with a mesh size of 2 × 2 mm, supported on an iron frame). Offshore from a boat, snails were collected for 30 min by passing a dredge (0.75 m long and 0.4 m wide with an attached sieve, 2 × 2 mm mesh size) along the bottom. Dredge hauls were made, beginning at 1 m depth and extending perpendicular to the shore to a maximum of 10 m depth, typically covering a distance of about 150 m. Live snails were transported to the Kenya Medical Research Institute (KEMRI), Center for Global Health Research, Kisian, Kisumu.
Snails were provisionally identified using keys [32,67]. Snails were rinsed and placed one snail per well in 12-well cell culture plates in 3 ml of aged tap water. The plates were placed in ambient outdoor lighting for 2 hr to induce cercarial shedding. Cercariae were identified morphologically [68]. Each shedding snail was preserved in one sample tube, and the cercariae they released in a corresponding tube, all in 95% ethanol. Non-shedding snails were maintained in the lab to allow cercariae-shedding infections to develop and re-shed 1-5 weeks later. Snails were maintained in 20 L tanks with oyster shells, aeration, and fed boiled lettuce and shrimp pellets.
S. haematobium miracidia were sourced from the urines of local schoolchildren enrolled in this study (see ethics statement below) or from discarded clinical samples and were used for phylogenetic analyses and comparisons with schistosomes shed from infected snails.

Additional sampling records
Additional specimens were obtained by a loan from collections held at the Division of Parasites, Museum of Southwestern Biology, University of New Mexico.

Molecular characterization
Snail sequences. Prior to extraction, snails to be processed for sequencing were photographed to provide a record of shell size and shape. Snail genomic DNA was extracted from a small portion of the head foot using the E.Z.N.A. Mollusc DNA Kit (Omega Bio-Tek, Norcross, GA) according to manufacturer's instructions. Partial sequences of the cytochrome c oxidase subunit I (cox1) and 16S rRNA genes were obtained for molecular identification and differentiation among Bulinus species.
Cox1 partial sequences (706 bp) were amplified using universal primers [69] and occasionally using reverse primer COR722b [70]. 16S partial sequences (481 bp) were amplified using forward primer 16Sar and reverse primer 16Sbr [71]. Thermocycling conditions for both cox1 and 16S were as follows: preheat at 94˚C for 5 min followed by 45 cycles of denaturation at 94˚C for 15 sec, annealing at 45˚C for 30 sec and extension at 72˚C for 1 min; final extension step at 72˚C for 10 min. All snail and parasite PCR reactions had a volume of 25 μL with 1 μL of 40 ng of DNA, 0.8 mM/l dNTPs, 2.5 mM/l MgCl 2 , 0.25 units of Ex Taq DNA (Clontech, Mountain View, CA), and 0.4 μM/L of each primer.
Schistosome sequences. Partial cox1 mtDNA and partial internal transcribed spacer 1 (ITS1) + 5.8S + partial internal transcribed spacer 2 (ITS2) rRNA sequences were used to identify and differentiate among Schistosoma species. A single cercaria was removed from the ethanol preserved cercariae obtained from a single snail and used for DNA extraction. Genomic DNA was extracted from parasite specimens using QIAamp DNA Micro kit (Qiagen, Valencia, CA) according to manufacturer's instructions with a 40 μL final elution volume.
Cox1 partial mtDNA (423 bp) sequences were generated using a modified forward primer designed from the S. bovis/S. haematobium universal primer [72] (ModShAsmit1: 5' TTTTTTGGKCATCCTGAGGTGTAT3'), and the reverse primer Cox1_schist_3' [73]. Thermocycling conditions were as follows: preheat at 94˚C for 5 min followed by 30 cycles of denaturation at 94˚C for 30 sec, annealing at 40˚C for 30 sec and extension at 72˚C for 2 min followed by a final extension period of 72˚C for 5 min. ITS1 + 5.8S + ITS2 partial rRNA (981 bp) sequences were amplified using forward primer ITS5 and reverse primer ITS4 [74]. Thermocycling conditions were as follows: preheat at 94˚C for 5 min followed by 30 cycles of denaturation at 94˚C for 30 sec, annealing at 54˚C for 45 sec and extension at 72˚C for 1 min; followed by a final extension period of 72˚C for 5 min.
For both snails and schistosomes. PCR fragments were separated by 1% agarose gel electrophoresis and visualized with 0.5% GelRed Nucleic acid gel stain (Biotium, Hayward, CA). PCR products were purified using ExoSap-IT (Affymetrix, Santa Clara, CA). Both strands were sequenced using an Applied Biosystems 3130 automated sequencer and BigDye terminator cycle sequencing kit Version 3.1 (Applied Biosystems, Foster City, CA). DNA sequences were verified by aligning reads from the 5 0 and 3 0 directions using Sequencher 5.1 and manually corrected for ambiguous base calls (Gene Codes, Ann Arbor, MI).
Additional Bulinus (MT707420.  [78][79][80][81]. Indoplanorbis exustus and Schistosoma mattheei were selected as outgroups for the Bulinus and Schistosoma analyses, respectively. Genbank sequences MH037061 and MH037083, and GU451744 and GU451726 were concatenated to produce outgroup sequences for the bulinid cox1 + 16S concatenated alignment [76,77]. GenBank accession numbers for bulinid sequences provided in this study can be found in Table 1. Multiple sequence alignments were performed using the program MUSCLE [82] in MEGA X [83]. The best fit maximum likelihood (ML) nucleotide substitution model was chosen for all genes in MEGA X using BIC criterion. Phylogenetic relationships were inferred using ML in MEGA X using 1000 bootstrap replicates. Uncorrected pairwise distance values (p-distances) were calculated in MEGA X [83]. Data were summarized within and between groups (Tables 2 and S2).
Specimens sequenced as part of this study were deposited as vouchers in the Division of Parasites, Museum of Southwestern Biology at the University of New Mexico. Snail and parasites specimens were designated a MSB:Host: or a MSB:Para: number, respectively (Tables 1 and 3).

Overview of Bulinus collections
A total of 6,133, Bulinus snails were collected from 11 locations (S1 Fig) in in the LVB between January 2014 and March 2021 and initially provisionally identified, in some cases just to species group (S1 Table). Recovered snails included B. globosus (n = 2994), B. ugandae (n = 889), B. productus (n = 1302), B. tropicus/truncatus group species (n = 245), and B. forskalii group species (n = 685). Bulinid species presence and trematode composition and prevalence varied by site (S1 Table). The highest schistosome prevalence was recovered from B. globosus at Asao stream (6.5% prevalence) and few to no schistosome infections were recovered from the various lake shore habitats. Further sequence-based specifications of species identities for both bulinids and schistosomes are found below.

Molecular identification of bulinids
Partial portions of the cox1 gene were sequenced from 62 bulinids. Because some clades were initially overrepresented, 58 sequences were used in the final phylogenetic analysis (Fig 1). 16S sequences were produced for 70 bulinid specimens. Some specimens did not produce amplicons for both genes and therefore concatenated (cox1 +16S) sequences were produced for 57 bulinid specimens (Fig 2). Specimens were chosen for sequencing to include representative species from the widest variety of habitats possible. Specimen information can be found in Table 1.

cox1 phylogenetic analysis
A total of 62 cox1 sequences were generated as a part of this study. Because some clades were overrepresented, 58 cox1 sequences from this study and 25 sequences from GenBank were used to hypothesize phylogenetic relationships among bulinid specimens we collected. The cox1 sequence analysis discriminated each of the four Bulinus species groups as well as each of the 17 species included in the phylogenetic analysis (Fig 1). Intraspecific p-distances were less than 1% with the exceptions of the East African B. globosus complex (2.27%) and B. scalaris (3.26%). Interspecific p-distances ranged from 5.78% for within species group (ex. B. globosus and B. ugandae) to 19.48% between species groups (B. forskalii and B. productus). Exceptionally, members of the B. truncatus/tropicus group exhibited low interspecific p-distances as compared to other closely related bulinids ( Table 2).

Combined (cox1 + 16S) dataset analysis
Concatenated cox1 and 16S sequences from GenBank representing outgroup sequences, and 58 sequences from this study (9 species) were used to infer phylogenetic relationships among bulinids. The concatenated sequence analysis discriminated among the three Bulinus species groups included in this analysis and additionally allowed interspecific discrimination (Fig 2) with greater resolution than the single cox1 dataset.
Intraspecific species p-distances were less than 1%, with the exception of B. scalaris. Interspecific p-distances were greater than 5%, with the exception of members of the B. truncatus/ tropicus group which exhibited low intraspecific p-distances (S2 Table).

Trematode infections in field-collected snails
Natural infection prevalence varied by collection site and by host species (S1 Table). The highest prevalence of patent mammalian schistosome infections was found in B. globosus (6.37%) followed by B. productus (1.97%) and B. forskalii/scalaris (0.71%) with few infections found in B. ugandae (0.22%) and no schistosome infections observed among B. truncatus/tropicus specimens (S1 Table and Fig 3).
Higher overall trematode diversity was observed among B. globosus, B. ugandae, and B. forskalii (minimum 5 trematode taxa per species) than was observed for B. truncatus/tropicus specimens (2 taxa) (Fig 3). This study did not seek to identify non-schistosome trematode cercariae to the species level and has therefore likely underestimated the diversity of trematode

Phylogenetic analysis of schistosomes
Partial cox1 sequences provided by this study were primarily used to identify cercariae samples to the species level. Specimens were identified as either S. haematobium or S. bovis. Cox1 and ITS sequences were used to examine cercariae samples for nuclear/mitochondrial discordance, which was not observed. Concatenated (cox1 + ITS) alignments were used to infer relationships among specimens (Fig 4), and information relating to specimens provided by this study can be found in Table 3 (Fig 4).

Discussion
Our long-term goal is to understand the underlying biological processes that influence the complex interrelationships between bulinid snails and trematodes, especially schistosomes, in the LVB. Towards that end, we identified 8 distinct Bulinus taxa, 2 of which were naturally  Table).  (Figs 1 and 2). The presence of these taxa in areas we did not sample surely cannot be ruled out. We note that many of the previous identifications of B. africanus in the LVB were based largely on morphological criteria [  Phylogenetic relationships of schistosomes from this study and from GenBank based on 1166 bp of concatenated cox1 + partial ITS1 + 5.8S + partial ITS2 sequences inferred from ML analysis. Bootstrap values over 95% are indicated by an asterisk. Specimens are listed by MSB:Para: number followed by the host species. Bolded sequences were generated during this study and additional information for specimens can be found in Table 3. https://doi.org/10.1371/journal.pntd.0010752.g004

PLOS NEGLECTED TROPICAL DISEASES
The phylogenetic inferences we generated for west Kenyan specimens of the related species B. globosus grouped with East African specimens designated as B. globosus on GenBank. Similar to observations of Kane et al.
[38], we found (Fig 1) that type locality specimens for B. africanus [88] from Port Durban, South Africa, and for B. globosus [89] from Angola, belong to lineages separate from the East African B. globosus lineage. This suggests that the East African B. globosus is not conspecific with those from the type localities. In our phylogenetic analysis, representatives from the LVB did not group with the type-locality specimens for either B. [46] assembled the mitogenome of a B. ugandae sample from Lake Victoria. Based on examination of the shell photographs, similarities in habitat types, and phylogenetic analysis of sequences, we agree that these sequences represent B. ugandae specimens.
The phylogenetic relationships inferred in our study indicate that B. ugandae is sister to the East African B. globosus lineage as we have described it above. Figs 1 and 2 differ slightly in their topology, which may be resolved in the future with increased taxon sampling. However, both phylogenetic analyses support B. ugandae and East African B. globosus as separate lineages. In agreement with earlier studies [32,36,84,85], we did not find B. globosus in lacustrine habitats, while B. ugandae was found commonly from the shore of Lake Victoria or in marshes and swamps along the lakes edge.
It is of more than passing interest to correctly discriminate B. ugandae from B. globosus [96], and the application of molecular criteria is recommended. Bulinus ugandae is the only member of the B. africanus group not implicated in the transmission of S. haematobium [59,60]. The relationship between B. ugandae and S. bovis is more nuanced with field studies suggesting that Kenyan B. ugandae is refractory to S. bovis [84] whereas other studies suggested that B. ugandae from Sudan or Uganda are vectors of S. bovis [94,97]. Bulinus ugandae from Western Kenya was found to be compatible with S. bovis in experimental infections [98], and we found B. ugandae to be naturally infected with S. bovis at two of our lakeshore study sites (S1 Table). The low prevalence of S. bovis in B. ugandae we observed may explain why some studies did not report natural infections. Alternatively, perhaps S. bovis relies on facilitation by other trematodes to successfully infect B. ugandae as has been reported in other bulinid species [10]. Pennance [35] also noted a natural infection of B. ugandae with an oft-overlooked member of the S. haematobium group, S. kisumuensis, previously known only from West Kenya based on anatomical characteristics and sequence data for adult worms recovered from rodents [24].
B. ugandae hosts a variety of other trematode species in the LVB (S1 Table and Fig 3). Amphistomes were not recovered during this study nor from a Tanzanian survey [85]. However, in Sudan, B. ugandae was found shedding amphistome cercariae [99], raising the possibility that significant intraspecific differences within B. ugandae may occur with resultant differences in compatibility with trematodes, further contributing to the complex patchwork of Bulinus-trematode compatibility so often noted.
B. forskalii species have received less attention than other bulinids in East Africa, likely because they are not associated with S. haematobium transmission in that area, unlike in West Africa [32, 100,101]. Three B. forskalii group species: B. forskalii, B. scalaris, and B. browni have been reported from the LVB, and all have been observed to occur in sympatry [32]. In addition to finding B. forskalii commonly among our Kenyan samples, we found a juvenile of a second genetically distinct taxon that differed substantially from B. forskalii. It differed to a lesser extent from B. scalaris obtained from Ukerewe Island, Tanzania, the latter snail conforming conchologically to B. scalaris based on having rounded shoulders on the shell whorls [32]. The unknown juvenile tended to group with B. scalaris phylogenetically, yet intraspecific p-distances of these two sequences were higher than what has been reported within most Bulinus species (Tables 2 and S2). One possibility is that this snail is of the poorly known species B. browni, reported as being morphologically indistinguishable from B. forskalii but with unique enzyme banding patterns [102]. Its status remains uncertain as it has not been identified in any previous sequence-based analyses.
Neither B. forskalii nor B. scalaris are experimentally compatible with S. haematobium nor have been found to host natural infections in Western Kenya [59][60][61]. It is believed that B. browni similarly is not involved in transmission of S. haematobium [103], but both B. forskalii and B. browni have been implicated in the transmission of S. bovis [84,103,104]. These observations were supported by our findings which genetically identified S. bovis from natural infections in B. forskalii, yet we found no S. haematobium infections from any B. forskalii group snails. B. forskalii is known to vector a wide variety of other trematodes including amphistomes [55,105], echinostomes [106], and others [85]. Interestingly, the long periods of estivation that this species undergoes, which are associated with the ephemeral nature of their habitats, do not preclude it from frequently being parasitized by larval trematodes.
We found members of the B. truncatus/tropicus group only in Lake Victoria, an environment for which our accumulated taxonomic understanding for this species group is complicated. Based on morphological, enzymatic, and ploidy criteria, Brown  The cox1 p-distances between B. truncatus and B. tropicus was the lowest among any two bulinid species we examined (Tables 2 and S2). Our presumptive B. transversalis and the presumptive B. trigonus of Chibwana et al. [50] differ to a greater extent from either B. tropicus or B. truncatus, and from each other, suggesting they are distinct species. The low p-distances between B. truncatus and B. tropicus has also been noted by others [39,43,107] and is somewhat paradoxical when considering their differences in ploidy, morphology and role as vectors of schistosomes.
Among the 245 individuals of the B. truncatus/tropicus group we examined, only 2 were positive for natural trematode infections (S1 Table and Fig 3). Neither Kenyan B. truncatus nor B. tropicus are known to vector local S. haematobium isolates [59, 60,108]. However, Kenyan B. truncatus has been found compatible with allopatric S. haematobium isolates [60,109].
Experimental infections with what was likely a laboratory population of B. transversalis also proved refractory to East African S. haematobium infection [59]. B. truncatus has been found compatible with local isolates of S. bovis [84,110]. B. tropicus was found compatible with S. bovis only if it is previously infected with Calicophoron microbothrium [10,111]. No natural schistosome infections were documented for any member of the B. truncatus/tropicus group as part of our study.
As recently noted by Chibwana et al.
[50], a range of Bulinus species are present in Lake Victoria and surrounding waters and they also noted that bulinid presence in the lake potentially implies the presence of S. haematobium and health risks from urogenital schistosomiasis for people living along the shore, or on the lake's islands. A considerable body of work has been undertaken over the years to examine the role of lake-associated bulinids in schistosome transmission (see the several papers cited above). Evidence from surveys and experimental infections, in agreement with data provided by this study, indicate that common lake species like B. ugandae, B. tropicus and B. truncatus are not found to be infected with local S. haematobium isolates, nor are members of the B. forskalii species group. Common africanus group species members like B. productus and B. globosus found in habitats other than lake shore are found to naturally host S. haematobium. The lake-dwelling B. ugandae, along with B. forskalii, B. globosus and B. productus have been found to naturally host S. bovis infections. At this time, unlike the situation for S. mansoni, the shorelines of Lake Victoria do not seem to pose a strong risk of S. haematobium infection.
As has been noted [60,61], East African B. truncatus are susceptible to what was historically described as B. truncatus-adapted isolates of S. haematobium common to Western Africa and Egypt, and introduction of isolates from these regions into the lake region might pose a new lake-borne S. haematobium problem. Likewise, introductions of exotic species into the lake, altered thermal or water quality regimes or changing populations of snail predators might change the current picture of Bulinus species representation in the lake, as they have in other African lakes [112].
Of further interest to us is to understand the puzzling underlying factors that dictate compatibility with S. haematobium of one Bulinus species, like B. globosus, whereas its close relative, B. ugandae, is seemingly refractory? This characteristic has a great deal to do with keeping S. haematobium transmission from occurring in the lake, thereby averting what could be a massive public health problem. Can this natural resistance to S. haematobium infection, if explained, in some way be used to lessen the vector potential of other bulinid species as a novel means of schistosomiasis control?
Similarly, we are interested in the characteristics of the west Kenyan S. haematobium isolates which favor or disfavor compatibility with certain bulinid species. S. haematobium isolates from across Africa have recently been shown to be genetically homogenous as compared to S. bovis [30], a characteristic that belies the evident heterogeneity in compatibility shown by S. haematobium across Africa with respect to Bulinus species use. One possible explanation is that all S. haematobium isolates tested, with the exception of the Madagascar isolate, have been found to contain varying levels of S. bovis introgression in their genomes [28][29][30]. It will be of interest to determine if the content of such introgressed regions influence the compatibility of S. haematobium to different Bulinus species.
Other avenues of interest for disentangling the Bulinus-schistosome compatibility include the role of symbionts, such as annelids (Chaetogaster), which may prey upon the miracidia or cercariae of trematodes, thereby reducing transmission [9]. Chaetogasters are particularly conspicuous on field-derived specimens of Bulinus [113] and deserve further scrutiny with respect to their impact on influencing infection success of schistosome miracidia.
We are similarly interested in applying the notion of coevolutionary hot and cold spots [114,115] to Lake Victoria shorelines, owing to their intense use by many host species potentially carrying many trematode species [11]. Shoreline locations have been considered coevolutionary hot spots and may dictate certain type of immune or other avoidance strategies by snails to avoid high infection rates. In contrast, deep water locations are considered coevolutionary cold spots because fewer host species (and attendant trematodes) frequent them, which might select for different response strategies among snails living there. We are similarly interested to learn if species like B. forskalii that so often are found in ephemeral habitats and known to be preferential self-crossers [116] have fundamentally different strategies for dealing with pathogens like trematodes than snails that occupy far more stable conditions, like the shoreline habitats of Lake Victoria.

Conclusions
Based on cox1 sequence data, we found 8 distinct taxa of Bulinus in our west Kenyan sampling locations: B. globosus, B. productus, B. ugandae, B. forskalii, presumptive B. scalaris; B. tropicus, B. truncatus and presumptive B. transversalis. We found natural infections of S. haematobium in B. globosus and B. productus, and the ruminant schistosome S. bovis in these two species as well as in B. ugandae and B. forskalii, confirming the vector role for these species outlined in previous studies. We highlight the importance of providing molecularly-based identification, particularly in regards to discriminating S. haematobium vector species like B. globosus from related non-vector species like B. ugandae. Several outstanding issues with respect to Bulinus systematics were noted: the lack of bona fide B. africanus in our samples and the presence of a "B. globosus complex" requiring further resolution; the status of B. productus as a distinct species from B. nasutus; and the need for further collection and resolution among species in both the B. forskalii and B. tropics/truncatus groups, the latter especially as it pertains to the LVB. The complex patterns of Bulinus-Schistosoma compatibilities noted argue for more in-depth study to understand factors dictating the underlying patterns that, at least thus far, have fortuitously kept the immediate shoreline and waters of Lake Victoria largely free of S. haematobium transmission.