Transmission and diversity of Schistosoma haematobium and S. bovis and their freshwater intermediate snail hosts Bulinus globosus and B. nasutus in the Zanzibar Archipelago, United Republic of Tanzania

Background The Zanzibar Archipelago (Pemba and Unguja islands) is targeted for the elimination of human urogenital schistosomiasis caused by infection with Schistosoma haematobium where the intermediate snail host is Bulinus globosus. Following multiple studies, it has remained unclear if B. nasutus (a snail species that occupies geographically distinct regions on the Archipelago) is involved in S. haematobium transmission on Zanzibar. Additionally, S. haematobium was thought to be the only Schistosoma species present on the Zanzibar Archipelago until the sympatric transmission of S. bovis, a parasite of ruminants, was recently identified. Here we re-assess the epidemiology of schistosomiasis on Pemba and Unguja together with the role and genetic diversity of the Bulinus spp. involved in transmission. Methodology/Principal findings Malacological and parasitological surveys were conducted between 2016 and 2019. In total, 11,116 Bulinus spp. snails were collected from 65 of 112 freshwater bodies surveyed. Bulinus species identification were determined using mitochondrial cox1 sequences for a representative subset of collected Bulinus (n = 504) and together with archived museum specimens (n = 6), 433 B. globosus and 77 B. nasutus were identified. Phylogenetic analysis of cox1 haplotypes revealed three distinct populations of B. globosus, two with an overlapping distribution on Pemba and one on Unguja. For B. nasutus, only a single clade with matching haplotypes was observed across the islands and included reference sequences from Kenya. Schistosoma haematobium cercariae (n = 158) were identified from 12 infected B. globosus and one B. nasutus collected between 2016 and 2019 in Pemba, and cercariae originating from 69 Bulinus spp. archived in museum collections. Schistosoma bovis cercariae (n = 21) were identified from seven additional B. globosus collected between 2016 and 2019 in Pemba. By analysing a partial mitochondrial cox1 region and the nuclear ITS (1–5.8S-2) rDNA region of Schistosoma cercariae, we identified 18 S. haematobium and three S. bovis haplotypes representing populations associated with mainland Africa and the Indian Ocean Islands (Zanzibar, Madagascar, Mauritius and Mafia). Conclusions/Significance The individual B. nasutus on Pemba infected with S. haematobium demonstrates that B. nasutus could also play a role in the local transmission of S. haematobium. We provide preliminary evidence that intraspecific variability of S. haematobium on Pemba may increase the transmission potential of S. haematobium locally due to the expanded intermediate host range, and that the presence of S. bovis complicates the environmental surveillance of schistosome infections.


Introduction
Schistosomiasis is a snail-borne neglected tropical disease (NTD), that can cause severe morbidity and mortality in both humans and animals [1,2]. Schistosoma haematobium and S. mansoni are the two species responsible for most cases of human schistosomiasis in sub-Saharan Africa, causing urogenital and intestinal schistosomiasis, respectively. Community or school-based treatment of schistosomiasis using the only recommended preventive chemotherapeutic drug currently available, praziquantel (Merck KGaA), is the most common and effective means of alleviating disease burden [3,4]. As prevalence of the infection in humans moves towards elimination in parts of sub-Saharan Africa following large-scale multi-year treatment programmes, it is apparent that low-level transmission is continuing, facilitated by the presence of freshwater snail intermediate hosts enabling reinfection [5,6]. Monitoring schistosome infections in snails as part of control and elimination surveillance, either supplementing human/veterinary parasitology surveys or as a stand-alone measure, will aid in establishing whether Schistosoma spp. transmission is persisting, interrupted or re-established following elimination.
Unguja and Pemba islands, collectively known as Zanzibar (United Republic of Tanzania), are endemic for urogenital schistosomiasis. Zanzibar has a long history of pioneering urogenital schistosomiasis research and control dating back to the 1920s. This ranges from investigating the freshwater snails involved in disease transmission to early trials of schistosomacidal drugs and assessing disease prevalence through low cost diagnostics [7][8][9][10][11][12][13][14][15]. More recently, the islands have been targeted for elimination with concerted efforts within the Zanzibar Elimination of Schistosomiasis Transmission (ZEST) project (2012)(2013)(2014)(2015)(2016)(2017). This trialled the additive impact of integrated interventions (such as mollusciciding against the snail intermediate host or educational measures for behavioural change) in combination with bi-annual mass drug administration (MDA) and MDA alone [16][17][18][19]. Over the course of the ZEST project, prevalence was significantly reduced across the islands, but transmission was not interrupted [18], leaving focal endemicity in hotspot areas that require new methods of surveillance and tailored interventions [19][20][21] Of the four endemic snail species of Bulinus on Zanzibar (B. globosus, B. nasutus, and two B. forskalii group taxa: B. forskalii; and a taxon currently undescribed and presented previously as Bulinus sp.), it has been concluded from earlier studies investigating local intermediate host compatibility that only B. globosus is a compatible intermediate host involved in S. haematobium transmission, with B. nasutus being refractory to S. haematobium infection on Zanzibar [22][23][24][25]. Schistosoma haematobium transmission in the past couple of decades was therefore considered to be restricted across the islands to only freshwater bodies where B. globosus resided (Fig 1), with the distribution of B. globosus on the islands also being constrained by this species' habitat preferences such as water hardness [23]. Therefore, freshwater habitats (e.g., in the southern part of Unguja) previously identified as containing only B. nasutus (a freshwater snail species closely related in the Bulinus africanus species group and morphologically overlapping with B. globosus) were considered free of S. haematobium transmission (Fig  1), despite there being evidence that B. nasutus can harbour pre-patent Schistosoma spp. on Pemba [26]. Schistosoma haematobium populations on Zanzibar are considered more genetically diverse in comparison to mainland African populations, with the presence of both Group 1 (mainland Africa) and the more diverse Group 2 (Indian Ocean Islands) strains [27,28]. However, it was not until recently that a second Schistosoma species, S. bovis which infects ruminants, was identified as being transmitted by B. globosus on Pemba [29]. Surveillance of Schistosoma transmission in Zanzibar is therefore complicated by not only the presence of two morphologically indistinguishable larval Schistosoma species shed from Bulinus snails, but also two morphologically overlapping Bulinus species.
Despite several studies demonstrating the incompatibility of S. haematobium with B. nasutus on Unguja [22][23][24][25], overlapping areas of B. nasutus and urogenital schistosomiasis endemicity have been shown to exist predominantly on Pemba (Fig 1), and reports up until 1962 demonstrate high S. haematobium prevalence and only B. nasutus presence in the south of Unguja [9,10,23]. This suggests that B. nasutus could be acting as an intermediate host for S. haematobium on Pemba and in the past on Unguja, as is the case in nearby coastal regions of Kenya [31] and Tanzania [32][33][34][35][36][37], with the addition of the closely related snail species B. (nasutus) productus also identified as involved in transmission in mainland Tanzania [38].  Tanzania) showing shehias where malacological surveys for Bulinus species were conducted in the current study and predicted distributions of Bulinus spp. and Schistosoma haematobium endemicity based on previous findings. Bulinus spp. distribution inferred from Stothard et al. [23] and Pennance et al. [29,30]. Schistosoma haematobium infection distribution interpreted from Knopp et al. [18]. Digital shape files for Unguja and Pemba administrative regions were obtained from DIVA-GIS (https://www.diva-gis.org). https://doi.org/10.1371/journal.pntd.0010585.g001 In this study we aimed to investigate the current transmission status of urogenital and bovine schistosomiasis on Pemba and Unguja islands by monitoring the species distributions and genetic diversity of B. globosus and B. nasutus and their associated Schistosoma spp. collected between 2005 and 2019. Inferences are made on how the findings may impact schistosomiasis control, surveillance and elimination in Zanzibar.

Ethics statement
Ethical approval for the collection and analyses of the snail and schistosome samples collected during the ZEST project were obtained from the Zanzibar Medical Research Ethics Committee in Zanzibar, United Republic of Tanzania (ZAMREC, reference no. ZAMREC 0003/Sept/011), the "Ethikkommission beiber Basel" (EKBB) in Basel, Switzerland (reference no. 236/11) and the Institutional Review Board of the University of Georgia in Athens, Georgia, United States of America (project no. 2012-10138-0) [18]. The ZEST study is registered with the International Standard Randomized Controlled Trial Number register (ISRCTN48837681). Additional sampling and analyses of snails in Unguja and Pemba were conducted in agreement with the Neglected Diseases Program of the Zanzibar Ministry of Health and the Public Health Laboratory-Ivo de Carneri respectively. All other snail and cercariae samples used were accessioned in the Schistosomiasis Collection at the Natural History Museum [39].

Sampling of Bulinus spp. on Pemba and Unguja islands
Following oral approval to conduct surveys by local Shehas, community leaders that locally govern each area (Shehia), human-freshwater contact sites were located from previous reports or with the help of local residents. Coordinates were taken at 112 freshwater sites (109 on Pemba and three on Unguja) across 20 shehias (Fig 1) using a Garmin GPSMAP 62sc device (Garmin, Kansas City, USA) and each water body was surveyed for the presence of intermediate host snails. Between 1 and 5 surveys were conducted at each human-freshwater contact site across Pemba and Unguja in October 2016, October 2017, February, July and November 2018 and January 2019 (Table 1).
At each site, snails were identified using shell morphology to their genera, with any non-Bulinus africanus group snails (namely B. forskalii group snails) being returned to the collection site. Snails were collected by hand predominantly from submerged vegetation and tree roots around the water's edge that were in close proximity to access points to the water. Each site was surveyed for 15 minutes by three collectors, starting from access points and then searching as much of the accessible perimeter of the waterbody during this time. Snails morphologically identified as either B. globosus or B. nasutus (species differentiation on morphology alone not possible) were placed in collection pots and transported back to either the Public Health Laboratory-Ivo de Carneri (Chake Chake, Pemba) or the Neglected Diseases Program laboratory (Zanzibar Town, Unguja) where they were counted and housed in plastic trays with bottled water and covered by a glass lid overnight to acclimatise. The following morning (before 08:00), snails were rinsed with bottled water (to mitigate carry over of any cercariae between snails) and examined for cercarial shedding by placing individuals in wells of 12-well ELISA plates filled to approximately two thirds with bottled water and placed under indirect sunlight. Each well was checked using a dissection microscope after two hours and again eight hours after first sunlight to capture schistosomes with different shedding patterns [40]. An experienced microscopist distinguished furcocercous schistosome cercariae from other species using descriptions of Schistosoma spp. under a dissecting microscope [41]; a subset of any shed cercariae were individually captured and pipetted in 3.5 μl onto Whatman FTA cards (Whatman, Part of GE Healthcare, Florham Park, USA) for molecular characterisation. All snails were preserved in 100% ethanol for subsequent molecular characterisation as previously described (see [29]). A targeted malacological survey was conducted in late January 2019 to collect B. nasutus snails at one site in Kangagani on Pemba previously identified as inhabited by B. nasutus (see [26]) (Fig 1). These snails were maintained in laboratory aquaria (dimensions 45x30x30cm, 40.5L, filled to approximately two thirds full with water from the collection site and equipped with an air pump for continuous aeration) at a maximum density of 100 Bulinus individuals per aquarium. Snails were re-checked for shedding of schistosome cercariae three weeks later in an effort to capture any infections that may have been pre-patent during the initial screen. Aquarium water was replaced twice a week using water from the site of collection (Kangagani). Water was only used in aquaria after storage for at least 48 hours in transparent 15L water containers, allowing for any sediment to settle and eliminate the risk of introducing live schistosome eggs, miracidia or cercariae into the aquaria. Snails were fed on dried lettuce when all lettuce in the tank had been eaten. Dead snails were removed from the aquaria daily.

Archived samples included in the analysis
Bulinus samples, from Unguja and Pemba, accessioned within the Schistosomiasis Collection at the Natural History Museum (SCAN) [39] were included in the study providing material from areas not covered in the malacological surveys described above. Samples were only included if associated geographical information was available. Six snails were included, five from Unguja (Jendele, Miwani, Kinyasini, Mtopepo, Chaani) and one from Pemba (Wingwi) as presented in Table 1. Two of these snails from Miwani and Kinyasini on Unguja were recorded as patent with S. haematobium when they were collected, as this was later confirmed by cercarial molecular analysis as described below. The remaining four Bulinus snails from the SCAN collection were negative for patent Schistosoma infections (S1 Table). One of the infected B. globosus identified in the SCAN repository (MCF389B0F0286, S1 Table) could not be associated with its emerging S. haematobium cercariae as it was preserved together with two other B. globosus also infected with S. haematobium from the same site.

Bulinus spp. molecular characterisation
Since a significant degree of morphological overlap exists between B. globosus and B. nasutus, a molecular marker was used to unequivocally identify a subset of the Bulinus snails collected (n = 504), and those taken from archived specimens (n = 6). The shell was removed by crushing and the use of sterile forceps from the preserved sample and gDNA from whole snail tissue was extracted using either the Qiagen BioSprint 96 DNA Blood Kit following manufacturer's instructions (Qiagen, Manchester, UK) or the Qiagen DNeasy Blood & Tissue Kit modified protocol (Qiagen, Manchester, UK) using double volumes of the lysis buffers [29]. Since not all snails could be identified using molecular characterisation due to cost and time constraints, a minimum of three non-patent snails per site per malacological survey (except for those collected on Pemba during October 2016 and from Chambani and Uwandani in 2017/2018) were randomly selected for identification. Molecular characterisation was performed for all snails with patent Schistosoma infections and snails retrieved from SCAN. DNA was extracted from a total of 510 Bulinus spp. snails, from Unguja (11 snails collected from 8 sites) and Pemba (499 snails collected from 63 sites). A 623 bp partial region of mitochondrial cox1 DNA was amplified and Sanger sequenced following previously described methods (see [29]). Sanger sequence data were edited, manually trimmed to 463-621 bp and aligned in Sequencher v5.4.6 (GeneCodes Corp., Michigan, USA) before being collapsed into cox1 haplotype groups. Species identification were confirmed by alignment and phylogenetic analysis (see below) of Bulinus cox1 haplotypes, as presented in S2 Table, with reference data for B. globosus and B. nasutus [42].

Schistosoma spp. cercariae molecular characterisation
From each infected snail either two or six Schistosoma cercariae were processed individually for molecular identification. Six cercariae were individually processed from each infected snail collected between 2016 and 2019 (n = 20 snails) and two cercariae for snails collected as part of the ZEST study (n = 69 snails) [16][17][18] and made available via SCAN (S3 Table). Following elution of parasite DNA from Whatman FTA cards [43], Schistosoma species identification was confirmed by mito-nuclear genetic profiling targeting the partial mitochondrial cox1 region (956 bp) and the complete nuclear ITS (1-5.8S-2) rDNA region (967 bp) from each individual cercariae as described in [27,28]. Both mitochondrial and nuclear DNA were analysed for species identification and to identify any hybridisation [44]. The cox1 data were also used for genetic diversity and phylogenetic analyses. The sequence data were manually edited and trimmed to 750 bp for cox1, and 880bp for ITS, using Sequencher v5.4.6 (GeneCodes Corp., Michingan, USA). The cox1 species identity was confirmed by comparison to nucleotide sequences using NCBI-BLAST [45] and the ITS species ID was confirmed by comparison to reference data as described [28,29]. Species identification of each cercariae was confirmed by concordance between the cox1 and ITS genetic profiles. Cercariae of identical cox1 sequences were collapsed into cox1 haplotype groups for further phylogenetic analysis (S4 Table). ITS data were not used for phylogenetic analysis since no intra-species diversity was observed.

Statistical analysis
Within the molecular analyses S. haematobium cercariae were assigned to either the Group 1 or Group 2 cox1 haplotype group [27], and Chi-squared tests were performed in R v.4.0.0 [58] to investigate any differences in the abundance of the two groups in relation to their snail host species and geographical distribution.

Patent schistosome infections of Bulinus
Over the six malacological surveys conducted on Pemba between 2016 and 2019, and the one survey conducted in Unguja, a total of 11,116 Bulinus spp. were collected from 65 of the 112 sites. From the subset of the snails that were identified by molecular analysis from each of these sites (n = 504), and those identified from the SCAN repository (n = 6), the majority were B. globosus (n = 433 snails), the remainder were B. nasutus (n = 77 snails). The other 10,606 Bulinus collected remain only morphologically identified as being within the Bulinus africanus species group (Bulinus genus), since morphological differentiation between B. globosus and B. nasutus is not possible. Of the 11,116 Bulinus, 0.2% (n = 20) shed Schistosoma spp. cercariae. These were collected from eight sites: four sites in Kinyasini (n = 16 snails), two in Kizimbani (n = 2 snails), one in Chambani (n = 1 snail) and one in Kangagani (n = 1 snail) as presented in Table 1. The infected snail from Kangagani was collected during the targeted malacological B. nasutus survey, in which 198 individual Bulinus spp. specimens were collected and at the time of collection were not shedding. Although having no observed patent schistosome infections during the first round of shedding, a single B. nasutus was found shedding Schistosoma cercariae 21 days later. No follow up shedding was attempted on the other snails collected in Pemba as they were preserved within 48 hours of collection.
For the malacological surveys conducted at three sites in two shehias (Mtende and Kizimbani Dimbani) on Unguja, only six B. nasutus (Table 1) were collected in total, of which none shed Schistosoma cercariae within 24 hours of their collection. No further shedding attempts were made on these snails which were preserved after the first round of shedding.

Bulinus spp. genetic diversity and distribution
From the subset of Bulinus spp. successfully sequenced, 433 out of 510 from 3 sites across Unguja and 49 sites across Pemba were B. globosus (S1 Table). The remaining 77 snails were identified as B. nasutus collected from 15 sites in Pemba and 5 in Unguja (S1 Table). Bulinus nasutus was molecularly identified only from freshwater bodies near the east coast of Pemba and the southern districts up to the central west areas of Unguja (Fig 2). Where >1 snail was molecularly identified per site, the snails were identified as either B. globosus or B. nasutus coming from the same site, except for from one waterbody on Pemba (Puj11, Fig 2B), where both species co-occurred in the same seasonal pond (S1 Table). Based on the subset of molecularly identified Bulinus, no other mixed B. globosus and B. nasutus populations were observed, however this would have to be fully confirmed with further genetic analysis of a large set of snails from each water body.
All 11 B. globosus and 9 out of 10 B. nasutus cox1 haplotypes were unique to either Unguja or Pemba, with one exception being B. nasutus haplotype 3, as shown in S2 Table, that was detected on both islands. A single clade of B. nasutus specimens from both Unguja and Pemba was observed, whereas the B. globosus isolates from each island fell into two distinct clades  body cohabited by B. globosus and B. nasutus (Puj11). Digital shape files for Unguja and Pemba administrative regions were obtained from DIVA-GIS (https://www.diva-gis.org).
https://doi.org/10.1371/journal.pntd.0010585.g002 (Fig 3). All B. globosus from Pemba fell into one clade containing two sister groups, with those previously identified from Pemba [42]. The three haplotypes from Unguja fell into a another clade containing two sister groups of B. globosus from Eastern Kenya and those previously identified from Unguja (Fig 3) [42].

Schistosoma spp. cercariae identification
From the subset of cercariae identified from 89 of the infected Bulinus collected from Zanzibar between 2016 and 2019 (n = 20) and identified during the ZEST study (n = 69), 82 were shedding S. haematobium with 18 different haplotypes and seven were shedding S. bovis with two haplotypes (S3 and S4 Tables). Both Group 1 and 2 haplotypes of S. haematobium representing mainland African and Indian Ocean islands respectively were identified [27,28] (Figs 4 and S1). Including Schistosoma coinfections, of which there were seven determined by multiple cox1 haplotypes presented in S5 Table, the proportion of snails shedding S. haematobium Group 1 cercariae (n = 41) was similar to those shedding Group 2 cercariae (n = 45). However, the majority of Group 1 S. haematobium infections occurred in Unguja (n = 35), with significantly fewer (n = 6) from Pemba (χ 2 = 10.0, df = 1, P< 0.01). In contrast, Group 2 S. haematobium cercariae were distributed evenly across the islands (Unguja n = 23 and Pemba n = 22). Most (n = 13 of 18) S. haematobium cox1 haplotypes were unique to either Pemba or Unguja, but five were present across both islands.
Comparison of mitochondrial cox1 and nuclear ITS profiles from S. haematobium cercariae showed no evidence of hybridisation between S. haematobium and S. bovis. No intraspecies variation in the ITS profiles were observed, with 100% match to reference data [27].

Bulinus observed shedding Schistosoma haematobium and S. bovis
Of the 20 infected snails collected in Pemba listed in S6 Table, 12 were identified as B. globosus infected with S. haematobium and seven as B. globosus infected with S. bovis. The remaining infected Bulinus collected during the targeted survey in Kangagani was identified as B. nasutus (Haplotype 3 in S2 Table), matching that previously reported on Pemba (GenBank Accession: AM921812, see [42]). The two cercariae identified from this snail were S. haematobium of a single cox1 haplotype (Sh3). This S. haematobium haplotype has been previously identified as 'Group 1' (GenBank Accession: GU257343, see [28]) representing those from mainland Africa and Zanzibar (Figs 4 and S1 Fig). The infected B. globosus from Unguja from the SCAN repository was shedding 'Group 2' cercariae as presented in S6 Table.

Discussion
Here we update Schistosoma and Bulinus species distributions in Zanzibar as a schistosomiasis surveillance resource, while also investigating specific associations between Bulinus and Schistosoma spp. on both Pemba and Unguja islands. Of the 11,116 Bulinus spp. collected from Pemba and Unguja between 2016 and 2019, 0.2% were infected with S. haematobium group species; B. globosus shed S. haematobium (n = 12) and S. bovis (n = 7) and B. nasutus shed S. haematobium (n = 1). The latter host-parasite relationship is the first to confirm preliminary findings by Ame [26] whilst also refuting that B. nasutus is refractory to S. haematobium infection across Zanzibar [23]. The distribution of B. globosus and B. nasutus across Pemba and Unguja inferred from molecular identifications of a subset of those collected confirmed previous findings of separate distributions [23,29,30], with the exception that the two Bulinus species were present in the same waterbody at one site and in close proximity across two other neighbouring sites along the east coast of Pemba (potentially overlapping distribution) where only B. nasutus was known to be abundant from previous reports. Identification of a second compatible intermediate snail host for S. haematobium (B. nasutus) changes our understanding of the snail and Schistosoma spp. biology on Zanzibar. In addition, the confirmed presence of B. globosus infected with S. bovis almost two years following the first recording of this pathogen on the island [29] and in combination with the presence of infected cattle [54], is cause for concern since increased transmission could lead to significant animal health and economic impacts, as well as a potential risk for hybridisation with S. haematobium (see [59,60]).

Distribution and diversity of Bulinus globosus and B. nasutus
Co-occurrence of B. globosus and B. nasutus was observed at just one site on Pemba (Puj11), generally supporting previous observations that species distribution in freshwater bodies is dictated by species specific ecological factors (such as water conductivity) determined by the geological zones of Zanzibar (see [61,62]). However, it is noteworthy that during the current study it was only feasible to identify a proportion (n = 510 of 11,116) of the B. globosus and B. nasutus accurately through partial cox1 sequencing. Therefore, it is possible that other sites containing both B. globosus and B. nasutus may be identified on the Zanzibar Archipelago in future species identification. The development of a cheap, easily interpreted, rapid diagnostic assay to distinguish between B. globosus and B. nasutus, such as is available for the differentiation of S. haematobium and S. bovis [63], would provide a much needed solution for identifying large numbers of field collected specimens.
The 11 B. globosus cox1 haplotypes identified from the snails collected across Zanzibar fell into two distinct clades representing Unguja and Pemba taxa [24]. The B. globosus from Unguja were more closely related to those previously identified from East Kenya [42], whilst those from Pemba form an independent group distinct from the other East African isolates, suggesting independent origins. This agrees with our current understanding of Zanzibar's geological formation, whereby Pemba island separated from mainland Africa earlier, during at least the early Pliocene compared to Unguja island during the Pleistocene [61,64]. In contrast, there was no phylogenetic distinction between the B. nasutus from Unguja, Pemba and mainland East Africa, with samples from each region all forming a single clade of multiple haplotypes. As recorded previously, B. globosus is more genetically diverse than B. nasutus (see [65]). Identical haplotypes of B. nasutus were also present on Unguja and Pemba. At this stage of investigation, we postulate that since B. nasutus infected with S. haematobium observed here in Pemba matched haplotypes of B. nasutus in the south of Unguja (see below), it might serve as an intermediate host of urogenital schistosomiasis across the Archipelago, as is the case in Kenya [31,66] and Tanzania [32][33][34][35][36][37].

A 'new' intermediate host of Schistosoma haematobium on Pemba: Bulinus nasutus
The finding of B. nasutus in one locality naturally infected with S. haematobium on Pemba contradicts previous results suggesting that this snail species was not involved with the transmission of urogenital schistosomiasis on Zanzibar [22][23][24]. Indeed, pre-patent Schistosoma spp. infections previously observed in B. nasutus from Pemba might have been S. haematobium (see [26]) but infections did not appear to result in cercarial production [25].
Schistosoma haematobium cercariae shed from the intermediate snail hosts, analysed here, fall into the two known S. haematobium haplotype groups (1 and 2) previously identified from miracidia collected from infected humans [27,28]. The cox1 haplotypes of cercariae shed from B. nasutus were identified as Group 1, a group predominantly associated with African mainland S. haematobium populations. Possibly only Group 1 S. haematobium is compatible with Pemba B. nasutus, since this same snail species acts as an intermediate host in East Africa [31][32][33][34][35][36][37], however more samples would be needed to test this hypothesis. Schistosoma strain specific interactions/compatibilities with intermediate host snails based on differential immune responses, have been well studied and established in strains of S. mansoni and B. glabrata (as summarised in [67]). However, relatively little is understood regarding co-evolution of host/parasite compatibility between S. haematobium group strains and Bulinus species, save some studies investigating geographical isolates (see [68][69][70][71][72]). Such strain dependency or local adaptation reflects the patchy compatibility of Bulinus snail hosts of Schistosoma generally [73], discussed in a previous study in Tanzania, where experimental infections of B. nasutus using a local strain of S. haematobium that usually infects B. globosus were unsuccessful [74]. As demonstrated again here, B. globosus remains the primary host of S. haematobium on Zanzibar, but endemic B. nasutus may also play a minor role in transmission involving specific S. haematobium strains, complicating future monitoring. This hypothesis also provides some explanation to how 'intermittent' and/or 'unstable' transmission was historically maintained in areas of Zanzibar (such as south Unguja) where only B. nasutus is present currently [7,9,10,75,76].

Study limitations and future work
Several limitations are apparent in the current study. The time and costs required to generate cox1 sequence data limited the number of snail intermediate hosts that could be identified by this means. Development and testing of a rapid diagnostic assay, or alternative, to provide rapid high throughput identification of snails would greatly support future studies. Additionally, few Bulinus specimens were available for analysis from Unguja, so further malacological surveys with molecular sub-sampling here would be beneficial to confirm snail species distributions across the island. Also, since the majority of infected Bulinus spp. collected during the ZEST studies were not accessioned with their associated Schistosoma spp., complete inferences on snail-Schistosoma relationships were not possible. Finally, although cercariae identification was used here to identify Schistosoma species, it would have been of interest to identify a greater number of cercariae per snail infection, as this may have significantly increased the number of coinfections observed from these Bulinus spp. and allow for the potentially immune modulated interactions of Group 1 and Group 2 S. haematobium coinfections, observed in five snails here, to be explored further. In a future study, it would also be of interest to use PCR based methods to identify pre-patent S. haematobium and S. bovis infections in snails to further assess the total number of snails that have been exposed to schistosomes but are not currently contributing to transmission [77].

Implications for future monitoring of schistosomiasis on Zanzibar
The findings discussed here provide implications for future control and elimination efforts of urogenital schistosomiasis on Zanzibar [18,19]. First, it is suggested here that the Ministry of Health, Social Welfare, Elderly, Gender and Children Zanzibar should conduct periodic surveys in areas associated with B. nasutus distribution on Unguja and Pemba. These surveys should include both malacological collections and combined human urine collections with questionnaires (including questions on freshwater usage locally and elsewhere in Zanzibar), followed by targeted treatment of infected individuals and focal snail control, to reduce any ongoing transmission. Second, a monitoring system to check the identity of schistosome cercariae shed from Bulinus spp. snails from Zanzibar to differentiate bovine and human schistosomiasis would enable accurate mapping of both schistosome species, appropriate targeting of urogenital schistosomiasis control interventions, and also monitor any potential hybridization events between S. haematobium and S. bovis that may be occurring in cattle or humans [60].
Nearly a century has passed since the first reports of widespread human urogenital schistosomiasis on Zanzibar, during which time there have been great achievements towards urogenital schistosomiasis elimination. As the battle to eliminate schistosomiasis from the Zanzibar Archipelago continues, our findings emphasize the need to carefully plan future surveillance strategies of transmission on the islands, taking into consideration the presence of bovine Schistosoma species and the capacity for an expanded intermediate host, and therefore geographical, range of S. haematobium.