Detection of haplosporidian protistan parasites supports an increase to their known diversity, geographic range, and bivalve host specificity.

Haplosporidian protist parasites are a major concern for aquatic animal health, as they have been responsible for some of the most significant marine epizootics on record. Despite their impact on food security, aquaculture and ecosystem health, characterizing haplosporidian diversity, distributions and host range remains challenging. In this study, water filtering bivalve species, cockles Cerastoderma edule, mussels Mytilus spp. and Pacific oysters Crassostrea gigas, were screened using molecular genetic assays using deoxyribonucleic acid (DNA) markers for the Haplosporidia small subunit ribosomal deoxyribonucleic acid region. Two Haplosporidia species, both belonging to the Minchinia clade, were detected in C. edule and in the blue mussel Mytilus edulis in a new geographic range for the first time. No haplosporidians were detected in the C. gigas, Mediterranean mussel Mytilus galloprovincialis or Mytilus hybrids. These findings indicate that host selection and partitioning are occurring amongst cohabiting bivalve species. The detection of these Haplosporidia spp. raises questions as to whether they were always present, were introduced unintentionally via aquaculture and or shipping or were naturally introduced via water currents. These findings support an increase in the known diversity of a significant parasite group and highlight that parasite species may be present in marine environments but remain undetected, even in well-studied host species.


Introduction
The phylum Haplosporidia consists of 36 recognized species in four genera, Urosporidium, Minchinia, Haplosporidium and Bonamia. Certain haplosporidian species have been credited with causing some of the most serious epizootic marine disease breakouts on record, in particular in shellfish species (Carnegie et al., 2016). In recent years, over ten newly detected haplosporidian species have been added to the phylum, including species in the Bonamia and Minchinia lineages (Arzul and Carnegie, 2015). In a previous study, environmental samples (water and sediment) from South Africa, Panama and the UK were molecularly screened, and revealed previously undescribed phylogenetic lineages within the Haplosporidia (Hartikainen et al., 2014). Besides their low detection prevalence, a major reason for the lack of detection of novel haplosporidian taxa is thought to be the use of (and increasing reliance on) broadly targeted molecular probes that are unsuitable for the highly divergent genes that characterize parasite groups (Hartikainen et al., 2014). Despite these obstacles, novel haplosporidians continue to be discovered in host/carrier species and habitats including the recently detected Haplosporidium pinnae in the fan mussel Pinna nobilis in the western Mediterranean Sea, thus highlighting the possibility that a significant diversity of haplosporidians have yet to be discovered (Lynch et al., 2013Arzul and Carnegie, 2015;Pagenkopp Lohan et al., 2016;Ramilo et al., 2017;Catanese et al., 2018).
Recent discoveries have highlighted that the geographic range of Phylum Haplosporidia is much greater than originally appreciated. Bonamia ostreae, which has caused significant mortalities in the European flat oyster Ostrea edulis, was thought to exclusively occur in the Northern hemisphere in both western and eastern North America and Europe but is now known to extend to the southern hemisphere in New Zealand where it has parasitized the native oyster Ostrea chilensis (Lane et al., 2016). Bonamia exitiosa was originally described in O. chilensis in New Zealand (Dinamani et al., 1987;Hine, 1996) but is now known to have an extensive geographic range in the southern and northern hemisphere and can infect several oyster species (Hill-Spanik et al., 2015). Haplosporidium nelsoni, the causative agent of MSX disease in the eastern oyster Crassostrea virginica in North America, was detected for the first time in Irish and Spanish Pacific oysters Crassostrea gigas and in O. edulis in Ireland (Lynch et al., 2013). In addition, Minchinia mercenariae, reported to cause infections in the hard clam Mercenaria mercenaria from the Atlantic coast of the United States (Ford et al., 2009), was detected in the common cockle Cerastoderma edule in the Netherlands (Engelsma et al., 2014) and the UK (Longshaw and Malham, 2013) where it was implicated in host population crashes, and an M. mercenariae-like parasite was recently confirmed in C. edule in Galicia, Spain (Ramilo et al., 2017).
Parasites in the phyla Haplosporidia have been reported infecting a number of bivalve hosts from across Europe. Species known to infect C. edule are Haplosporidium edule, Minchinia tapetis and M. mercenariae (Longshaw and Malham, 2013;Ramilo et al., 2017) and in mussels Minchinia sp. in Mytilus galloprovincialis along the Mediterranean coast of France (Comps and Tige, 1997), Haplosporidium sp. in M. edulis in Maine (Figueras and Jardon, 1991), USA, a haplosporidian-like parasite in M. edulis in Atlantic Canada (Stephenson and McGladdery, 2002) and Minchinia mytili in Mytilus edulis (Ward et al., 2019). Lynch et al. (2014) assessed the health status of Mytilus spp. around the coasts of Ireland and Wales, and detected a previously undescribed haplosporidian (Haplosporidia sp. SAL-2014) belonging to the Minchinia clade in a single M. edulis from Wales . The sequence of this haplosporidian was most similar to Minchina chitonis detected in the chiton Lepidochitona cinereus and an undescribed haplosporidian species parasitizing the Florida marsh clam Cyrenoida floridana (Reece et al., 2004). H. nelsoni has been associated with C. gigas populations in California (Friedman, 1996;Burreson et al., 2000), Korea (Kern, 1976), France (Renault et al., 2000), Japan (Friedman, 1996;Kamaishi and Yoshinaga, 2002) and Ireland (Lynch et al., 2013). In addition to H. nelsoni, Haplosporidium costale, a species associated with seaside organism (SSO) disease in C. virginica, was recently detected in C. gigas in China for the first time (Wang et al., 2010).
The objectives of this study were: determine (1) if Haplosporidia spp. were present in cockles, mussels and oysters at particular sites in Ireland, (2) did coinfection occur and (3) what abiotic and biotic factors associated with site influence and anthropogenic activities, i.e. aquaculture and shipping may influence their presence.

Materials and methods
Study sites, bivalve spp. sampled and site description A range of Irish coastal sites was sampled with different environmental (abiotic and biotic) factors and anthropogenic influences (aquaculture and shipping) from nature reserve sites to key economic areas influenced by frequent and heavy anthropogenic effects (Table 1, Fig. 1). Multiple samples of cockles (n = 1,604), mussels (n = 516) and oysters (n = 420) were collected from five of the fourteen Irish sites over several months and years resulting in a higher overall number of individuals being screened at those sites. A minimum sample size of thirty individuals was collected on a single occasion from five other sites.
Cockle, mussel and pacific oyster samples Wild C. edule was sampled at a nonculture site (March 2010-June 2011) and at two C. gigas culture sites (April to August 2015) ( Table 1). Cultured C. gigas were sampled from both culture sites in 2015. Wild Mytilus spp. were sampled from both culture sites in 2015, from a third culture site in September 2017 and from 11 nonculture sites in May to November 2017.

Histology
A cross section of tissue (mantle, gill, connective, digestive and gonad) was taken from each cockle, mussel and oyster, and was fixed in Davidson's solution for 24-48 h and subsequently stored in 70% ethanol. The fixed tissue was then dehydrated fixed in paraffin cut at 5 µm and stained using haematoxylin and eosin (H&E) and mounted in dibutyl phthalate xylene. Slides were examined using a Nikon light microscope at 4×, 10×, 20×, 40× and 100× magnification.
Several polymerase chain reactions (PCRs) using generic and specific primers and thermocycling conditions for Haplosporidia spp. were utilized in the molecular genetic screening. Additionally, a PCR to detect the nuclear DNA markers Me15/Me16 (Inoue et al., 1995) was carried out on mussels that amplified a PCR product to determine if they were M. edulis, M. galloprovincialis or hybrids of both parent species.
Two generic PCRs using similar mastermix and thermocycling conditions with primer pair HAP-F1 ′ /HAP-R3 ′ (Renault et al., 2000;Lynch et al., 2014) and ssu980/HAP-R1 ′ (Molloy et al., 2012;Lynch et al., 2014) to amplify the Small subunit ribosomal deoxyribonucleic acid (SSU rDNA) region of Haplosporidia spp. were carried out on C. edule, Mytilus spp. and C. gigas genomic DNA. A third PCR using specific H. nelsoni MSX-A ′ / MSX-B ′ primers to amplify the small subunit ribosomal ribonucleic acid (SSU rRNA gene) and similar mastermix and thermocycling conditions (Renault et al., 2000) was carried out on DNA from C. gigas. Negative controls containing double distilled water (ddH 2 O) were used in each PCR to control for contamination and infected Haplosporidia (B. ostreae, H. nelsoni, Haplosporidia sp. SAL-2014) genomic DNA was used as a positive control. Initially, in the screening of cockles from 2010/ 2011 no M. mercenariae-like positive material was available as it had not been detected before, however amplification in that single PCR occurred with the cockle deemed positive in the histology and that cockle's DNA was subsequently used as the positive control in the screening of the 2015 samples.
Electrophoresis of amplified products was carried out in a 2% agarose gel and was run with an electrical current of 110 V for 45 min. The expected product size for the HAP-F1 ′ /HAP-R3 ′ PCR was 350 bp (Renault et al., 2000), for the ssu980/HAP-R1 ′ was 430 bp (Molloy et al., 2012) and for the MSX-A ′ /MSX-B ′ PCR was 573 bp (Renault et al., 2000).
Pooled PCR products using replicates (×3) from individual cockles [Flaxfort (n = 1), Dungarvan (n = 1) and Carlingford (n = 1)] using the HAP-F1 ′ /HAP-R3 ′ primers and mussels (Clonea, n = 1) using the ssu980/HAP-R1 ′ primers were used to increase the amplicon concentration for Sanger sequencing, as recommended by EurofinsMWG. Both forward and reverse DNA sequences that were optimally generated by EurofinsMWG Sanger sequencing laboratory were matched against the National Center for Biotechnology Information (NCBI) nucleotide database with Basic Local Alignment Search Tool (BLASTn), which finds regions of local similarity between sequences to identify and confirm the DNA being detected in the PCRs. Percentage query coverage in BLAST refers to how much of the query sequence is aligned with results from the database sequence or, in other terms, the size of the sequence fragments that are comparable, while % identity measures the extent to which the nucleotide sequences relate to one another.
Phylogenetic analysis 18S SSU rDNA sequence data for 28 operational taxonomic units from GENBANK (Table 2) were downloaded, to which, the two 18S sequences were added. These data were aligned using Clustal Omega (Sievers et al., 2011) at the European Bioinformatics Institute portal (https://www.ebi.ac.uk/Tools/msa/ clustalo/). The final alignment was 2221 bp in length. The best-fit evolutionary model for the aligned sequence data was assessed in the jModelTest (v2.1.10; Darriba et al., 2012), using the small sample corrected Akaike information criterion (Hurvich and Tsai, 1989). This returned the generalized time reversible (GTR) model, with a four-category gamma rate distribution and invariant sites (GTR + G + I) as the best-fit model. The phylogeny was reconstructed in Mr Bayes (v3.2.5; Ronquist et al., 2012). A total of 379 base pairs in three distinct regions of the alignment were not able to be unambiguously aligned, and were excluded from the analysis. Two runs of four chains each were run for 5 000 000 generations, saving every thousandth tree. Nodal posterior probabilities were assessed using the 50% consensus tree topology (Huelsenbeck et al., 2002), discarding the first 25% of trees as burnin. Aligned sequences and commands used in the phylogenetic analysis are provided in ***.nex (Supplementary Information).
Resulting forward and reverse sequences were aligned and manually edited to resolve any ambiguities in base calling. The resulting alignments were matched against the NCBI nucleotide database with BLASTn.

Statistics
A χ 2 test was used to determine whether prevalence differences of parasites observed were significant (P values < 0.05) between sample sites in 2015. R packages used were dplyr, tidyr, ggplot2, car and NCStats.

Cockles
Haplosporidia-like single cell and plasmodia-like life stages identical to those described in Ramilo et al. (2017) and a spore-like stage ( Fig. 2A-F) (Table 1).

Mussels
In the M. edulis, haplosporidia-like single cell and plasmodia life stages similar to those described in the cockles were also observed in mussels deemed to be positive in the PCR (Clonea 1.7% (1/60) and Ringaskiddy 3.3% (2/60)). No spores were observed (Table 1).
A single haplosporidian spp. M. mercenariae-like parasite in C. edule at Flaxfort Strand (2010) and at the C. gigas culture sites Dungarvan and Carlingford Lough (2015) was identified in the PCR products amplified in cockles from the three sites [Accession # KY522823.1 (Ramilo et al., 2017)] with 97% query coverage and 100% maximum identity and are referred to as 'Isolates 221, 226 and 228' in this study.
Mussels: No PCR products were produced using both primer pairs (Renault et al., 2000;Molloy et al., 2012) in the wild Mytilus spp. screened at the three culture sites at Dungarvan, Carlingford and Ballymacoda.
Pacific oysters: No PCR products were produced using either primer pairs (Renault et al., 2000;Molloy et al., 2012) in the C. gigas screening or in the MSX-A ′ /MSX-B ′ screening for H. nelsoni (Renault et al., 2000).
Direct sequencing: The PCR product of the single mussel at Clonea was successfully sequenced however, the PCR products for both mussels at Ringaskiddy were also sent for sequencing but only the reverse sequences (which were a match) were amplified and not the corresponding forward sequence. As both sequences could not be aligned the result was not considered robust.
Due to the cost of sequencing all of the cockle PCR products amplified (n = 60), a subsample of cockle PCR products was sent for sequencing (representative of each sample site and years). Additionally, a similar morphology of each parasite was observed in the cockle and mussel histology for each respective parasite. More recent sequencing of PCR products (n = 40) from additional cockles at each location (study currently being carried out by the authors) has confirmed the findings of this study.
Phylogenetic analyses: The maximum likelihood phylogeny of haplosporidian taxa conclusively places Isolate 376, which is identical to the undescribed isolate SAL-2014 [Haplosporidia sp. Accession # KC852876.2 ], and Isolates 221, 226 and 228, which are identical to M. mercenariae-like parasite [Accession # KY522823.1 (Ramilo et al., 2017)], in a clade with species of the genus Minchinia (Fig. 4). Also included in this clade is an undescribed haplosporidian parasite of the Florida marsh clam C. floridana. The internal topology of this clade is  The phylogeny also recovers unambiguous support for the monophyly of Urosporidium (Fig. 4). Several of the internal nodes of the phylogeny are poorly supported in the bootstrap (Fig. 4), although, these involve branching events among the Bonamia Group, the Minchinia Group and the paraphyletic genus Haplosporidium.

Discussion
Detection of Haplosporidia spp.
Two haplosporidian species were detected for the first time in C. edule and in M. edulis in a new geographic range. When detected, both Haplosporidia spp. were observed in individual samples of mussels and cockles that consisted of at least 30 individuals. The prevalence of the M. mercenariae-like parasite was significantly greater in C. edule at both aquaculture sites (late spring, summer and early autumn samples) compared to C. edule at the nonculture site (summer sample), which may indicate that the extended presence of this parasite has some association with anthropogenic inputs and activities in these areas. Oyster seed/ spat consignments are routinely imported to both culture sites from France and the UK for on-growing in late spring. A low overall prevalence of the M. mercenariae-like parasite was detected in this study, which is similar to that observed in the Ramilo et al. (2017) study and for M. mercenariae infecting North American hard clams (Ford et al., 2009). Haplosporidia sp. SAL-2014 , a novel species first detected in a Welsh M. edulis in 2012, was also detected in this study for the first time at a similar prevalence in wild Irish M. edulis at two nonculture sites. One of those sites is a busy ferry/shipping terminal in Cork Harbour and a ferry between Wales and Cork Harbour was in operation from the 1980s up until recently.

Host partitioning
Interestingly the M. mercenariae-like parasite was not detected in the cohabiting C. gigas nor was it detected in the cohabiting Mytilus spp. Haplosporidia sp. SAL-2014 was exclusively detected in M. edulis even though M. galloprovincialis and Mytilus hybrids were present at the Irish sites where it was detected. This difference in emerging haplosporidian species detection, diversity and abundance in these three bivalve species strongly indicates that these parasite species are host specific and host partitioning is occurring. Additionally, the findings of this study would indicate that the haplosporidian species may be associated more with one environmental niche than another i.e. the sediment rather than with the water column for the M. mercenariae-like parasite, as the cockles were collected on the surface of the sediment and would normally be buried within the sediment similar to clams, while the oysters and mussels were at least 30 cm above the sediment on trestles or rocky outcrops respectively. Hartikainen et al.  . Bayesian phylogenetic tree of haplosporidian taxa based on partial 18S SSU rRNA sequences. Branches marked with an asterisk (*) have 100% posterior probability support for that node, otherwise, the nodal support is indicated by the number given. Inset to the bottom right repeats, for clarity, the Minchinia subclade with very small branch lengths, and gives the nodal posterior probabilities for this topology. A well-supported clade includes all of the 18S SSU rDNA sequences assigned to the genus Minchinia, the previously identified parasite of C. floridana (Reece et al., 2004), as well as the novel sequences for this study.
As has been found in other analyses (Reece et al., 2004;Lynch et al., 2013;Ramilo et al., 2017), we find strong evidence for the paraphyly of the genus Haplosporidium.
the Hartikainen et al. (2014) study would indicate that Minchinia spp. are niche selective or their detections are closely associated with their life stages i.e. in the water column in a planktonic intermediate host or near the sediment associate with primary bivalve host species.

Phylogenetics
The phylogenetic analysis in this study recovers unambiguous support for the grouping of these two haplosporidian isolates into a single clade with members of the genus Minchinia. As such, the four isolates would support the detection of a new geographic range for both of these species within the genus Minchinia. This would represent a substantial increase to the known diversity of this genus, as only six species associated with host species are currently described. The low bootstrap support for the internal nodes in phylogeny involves the genus Haplosporidium, and the interrelationships between its species and the two main haplosporidian clades, the Bonamia Group and the Minchinia Group.
Haplosporidium is paraphyletic (Reece et al., 2004), likely representing a plesiomorphic grade from which the remaining two clades derived. The short internal branch lengths and the low resolving power in the bootstrap potentially indicate a rapid diversification among haplosporidian taxa.

Site effect and shore height influence
A higher prevalence of the M. mercenariae-like parasite was observed in C. edule at Carlingford compared to Dungarvan. Carlingford is a more sheltered site with a lot of shipping activity, while Dungarvan is an oceanic bay that experiences tidal flushing, greater water exchange and some shipping activity Bookelaar et al., 2018). Higher pathogen retention, prevalence and diversity occur at sheltered marine environments, as pathogens are less likely to be swept away on the tides (Lenihan, 1999;Lynch et al., 2016;Bookelaar et al., 2018). A higher prevalence of the M. mercenariae-like parasite was observed in C. edule at the high shore at Carlingford compared to cockles lower down the shore at the oyster trestles. C. edule higher up the shore may experience more stressful abiotic conditions such as air exposure, fluctuating temperatures, precipitation etc., which may make them more susceptible to infections (Wegeberg and Jensen, 2003) or it may be possible that C. edule at the high shore are more likely to be in contact with other host species .

Potential drivers of emerging parasites
It is not uncommon for parasites to be widespread in marine environments, especially along near shore coastlines (Raftos et al., 2014). Coastal marine environments are very vulnerable to climate change (Holt et al., 2010) and a changing marine environment can have a direct impact on the distribution, life cycle and physiological status of hosts, pathogens and vectors (Gallana et al., 2013). While a change in host, pathogen or vector does not necessarily translate into a change of the disease, it is the impact of climate change on the interactions between the disease components that impact disease risk (Gallana et al., 2013). In addition, natural coastal disturbance arising from storm surges and high energy systems, which are predicted to increase under future climate change conditions (IPCC, 2018), may also play their part in pathogen emergence. Storms are important episodic events that can resuspend and transport sediments and are known to cause large-scale advection (i.e. transfer of heat or matter), sediment resuspension and transport (Cacchione et al., 1987;Warner et al., 2008). An association between disease outbreaks in marine organisms and storm activity is recognized (Burge et al., 2014). In one study, a strong correlation with hurricane activity [and a strong storm (nor'easter)] and the amoebic pathogen Paramoeba invadens, causative agent of urchin disease in the green urchin Strongylocentrotus droebachiensis in the northwest Atlantic, was modelled and confirmed (Feehan et al., 2015). Other factors such as the movement of non-native species, both intentionally (i.e. for aquaculture) and unintentionally (i.e. as stowaways in ship ballast water or on hulls), brings the threat of 'pathogen spillover' into introduced areas and highly-susceptible host populations (Carnegie et al., 2016;Ek-Huchim et al., 2017). It is also recognized that coastal development may unbalance marine parasite-host systems (Coen and Bishop, 2015), as past emergent and resurgent diseases in wildlife appear to be associated with anthropogenic activities (Harvell et al., 1999;Daszak et al., 2000). Cultured bivalve breeding programmes are designed to mitigate the impacts of pathogens and may be unintentionally resulting in or expediating 'host jumping' from now less susceptible bivalve hosts to new cohabiting and highly susceptible naïve host species (Bookelaar et al., 2018). Such programmes may inadvertently be making selectively bred bivalve hosts more susceptible to novel pathogens. Additionally, due to the very poor biogeographical records available for protistan parasites it is possible that these parasites evolved in these hosts at those locations and were not introduced.

Conclusions
The detection of both these Haplosporidia spp. in the Minchinia clade will contribute to an improved understanding of Haplosporidia diversity, prevalence, host, geographic distribution and a certain degree ecology. This study further supports that a M. mercenariae-like haplosporidan infects C. edule in Europe (Ramilo et al., 2017) and that it appears to be exclusive to C. edule while Haplosporidia sp. SAL-2014  appears to be exclusive to M. edule. As many other species, including protected bird species, rely on C. edule and M. edulis as a food source and the pivotal role both bivalve species play in marine coastal ecosystems, it would be beneficial to better understand the impact that these Haplosporidia spp. may or may not have on cockle and mussel populations currently and under future climate change conditions. Understanding some of the current drivers of parasite introduction, emergence and spread may facilitate a better prediction of future impacts and host or geographical range expansion of Haplosporidia under changing environmental scenarios. In particular with a parasite group such as the Haplosporidia, which have had such a devastating historical impact both commercially and ecologically.
Financial support. This work was supported by the Susfish and Bluefish Projects, both part-funded by the European Regional Development Fund (ERDF) through the Ireland Wales Co-operation Programme. A cross-border programme investing in the overall economic, environmental and social wellbeing of Ireland and Wales.
Conflict of interest. The authors declare that there is no conflict of interest and certify that they have no affiliations with or involvement in any organization or entity with any financial interest or non-financial in the subject matter or materials discussed in this manuscript.
Ethical standards. Not applicable.