The blue crab, Callinectes sapidus, is an ecologically and economically significant crustacean with a geographic distribution from maritime Canada to Uruguay. It is a key epibenthic link in the near shore marine environment where it feeds on a diverse range of invertebrates, and is a component of the diet of many coastal fish species. In the Atlantic and Gulf coast of the United States the blue crab supports a commercial fishery with an estimated worth of $165 million/year [1] and a substantial recreational fishery of undetermined magnitude that extends as far north as Massachusetts. There was a time when the blue crab was so abundant the harvests were described as “inexhaustible” [2]. However, the blue crab population in the Chesapeake Bay has been at historic lows for most of the past decade, most likely due to levels of exploitation [3, 4]. Other states such as North Carolina, South Carolina and Georgia have also experienced declines in their blue crab population and harvests [5]. In the Chesapeake Bay, the persistently low blue crab broodstock stimulated the formation of a research consortium to study hatchery-based broodstock enhancement and motivated Maryland and Virginia to implement unprecedented harvest restrictions [6, 7]. Though harvest restrictions have helped alleviate the strain on the blue crab population, it is typical to have year-to-year fluctuations in abundance that cannot be explained by fishing pressure alone. It is now understood that estimates for natural mortality of blue crab are over-simplified and even though many blue crab diseases have been described, their influence on mortality and population structure is poorly understood [8–11].

In mid-Atlantic coastal bays of North America, blue crab populations suffer episodic mortalities associated with infections by a parasitic dinoflagellate in the genus Hematodinium[12, 13]. Currently, there are two formally described species in the genus: H. perezi and H. australis. H. perezi was first described by Chatton and Poisson infecting the portunid crabs C. maenas (green crab) and Liocarcinus depurator (swimming crab) on the French Mediterranean coast [14]. In Australia, H. australis was described in Portunus pelagicus, the sand crab (also confusingly known as “blue swimmer crab”) [15]. Crustaceans residing in deep, stenohaline waters, such as Chionoecetes opilio, C. bairdi, C. tanneri and Nephrops norvegicus, are infected by yet another genotype of Hematodinium sp. distinct from H. perezi and H. australis[16, 17]. For many years, data gaps in descriptions of morphology, host range and sequence data led to uncertainty in species designations, [18] resulting in the species of Hematodinium infecting blue crabs being referred to generically as Hematodinium sp. [12, 19, 20]. Recent work by Small et al. confirms that the Hematodinium sp. infecting blue crabs on the US Atlantic coast is H. perezi[18, 21].

Epizootics of H. perezi in blue crab can recur annually and are often localized in coastal bay “hotspots”, characterized by constrained, high salinity, warm waters, suggesting that there are environmental or biotic reservoirs of the parasite [22]. The parasite is found in the hemolymph and tissue of the blue crab where it can multiply rapidly, eventually causing organ failure and crab death [13]. The life cycle of Hematodinium sp. is complex and includes trophont and sporont cell types within the host and flagellated dinospores that are occasionally observed within and emerging from infected hosts [23–25]. Recently, Li et al. reported on the in vitro culture of the entire H. perezi life cycle including the dinospore stages [26]. The existence of free-living dinospores suggests that the environment (water or sediment) can be a source of new H. perezi infections.

Detection of Hematodinium sp. within host tissues can be achieved by histological methods [11, 13], or by using vital stains for live parasites in fresh hemolymph [11, 14, 27]. These methods, however, are impractical for detecting the parasite in environmental samples. Molecular methods, developed for detection of Hematodinium spp. in commercially important crustaceans [17, 28–31], have the potential to be applied to DNA extracted from environmental samples. Initial research, using either endpoint polymerase chain reaction (PCR) or quantitative PCR (qPCR), targeted the SSU region of the parasite ribosomal RNA gene to assess Hematodinium sp. infections in blue crab hemolymph [28, 29]. Although these primers were sufficient for detection within blue crab, the sequences targeted were not specific enough for analysis of environmental samples, which may contain DNA of free-living dinoflagellate species that possess SSU gene sequences similar to those of Hematodinium sp. [31, 32].

There are two frequently encountered challenges when using qPCR to look for organisms in the environment. First, the accuracy of a qPCR assay rests on the ability to produce a reliable standard curve, which is usually based on DNA extracted from a pure culture of the organism itself or on plasmid DNA carrying the target sequence of the organism of interest. Until recently [26], H. perezi, cultures were not readily available for use in generating standard curves, and reliable passage of infections from wild isolates posed significant technical hurdles. The second challenge is that, in practice, accuracy of a qPCR method also depends on reproducible extraction of target DNA from environmental samples. An assay based on purified or cloned target DNA may not be representative of how parasite DNA is recovered from environmental samples in which there is a heterogeneous mixture of DNA from other species, in addition to potential qPCR inhibitory compounds. Therefore, to develop methods suitable for environmental assessment and prevalence studies, we sought to design and validate a qPCR assay targeting the species-specific ITS2 region of the rRNA gene, and establish a standard curve using a surrogate for H. perezi cells mixed with environmental sediment. We then used this assay to look for H. perezi in environmental samples on a temporal and spatial scale.

Along the US Atlantic coast, H. perezi infections in blue crabs display an annual peak of prevalence and intensity in late summer or autumn [22]. Conversely, H. perezi infections disappear between peaks, and therefore raise the question of whether there is a source (reservoir) of new H. perezi infections. Possible H. perezi reservoirs include other blue crabs, alternate hosts, and the environment. There have been conflicting reports as to the role of cannibalism in transmission [32, 35]. Hematodinium sp. has been detected in numerous alternate hosts sympatric with (and possible prey for) blue crab [36], but only at low prevalence and intensity [37]. The existence of a water-borne dinospore stage is well documented, and recent work supports a model in which Hematodinium spp. may have a resting stage (cyst, or another overwintering stage) that could be a source of re-infection [23, 32, 34, 38].

PCR is a practical technique for investigating the spatial and temporal presence of H. perezi in environmental reservoirs. A number of PCR assays have been developed to detect Hematodinium spp. DNA, all of which target one of the regions within the ribosomal RNA locus [28, 29, 31, 32, 39]. The SSU region of the rDNA locus shares high sequence identity between different species of Hematodinium (including the various clades described by others [17, 39, 40]). The Hematodinium sp. SSU also shares significant identity to other, free living, dinoflagellates, as we discovered at the initiation of the present study. As reflected in Additional file 1: Table S1, BLAST analysis revealed that the primers and probe used in Nagle et al.[29] are predicted to amplify the SSU sequences from presumptive free-living dinoflagellates on both the Gulf and Atlantic coasts.

More species-specific sequences are found in ITS1 and ITS2 regions, which are transcribed as part of the polycistronic rRNA product, but are not part of the mature ribosomal RNAs (SSU, 5.8S and LSU) [21]. We chose to develop a PCR assay based on the ITS2 region for two reasons: firstly, when rDNA clones derived from several H. perezi infections were compared, the ITS2 region displayed few clone-to-clone SNPs and length polymorphisms, as compared to the ITS1 region; and secondly, the intra-species variation of the sequence of the ITS2 region is constrained by its role in processing of the final rRNA molecules. The ITS2 region of the rRNA transcript folds into a cruciform structure that interacts with splicing enzymes. Coleman [41] and others [42] have reported on the advantages of the ITS2 as a molecular feature for taxonomic comparisons between congeneric species. Our finding that a single H. perezi-infected crab had two versions of an ITS1 polymorphism could be interpreted either to mean that the crab carried a dual infection of two strains that differ in ITS1 sequence, or that a single strain of H. perezi carried both versions, similar to within-strain ITS1 length variation in insects [43].

In this study a surrogate for H. perezi was used to assess the efficiency of target DNA recovery from environmental samples. DNA preparations from environmental samples are oftentimes inhibitory to PCR amplification. In place of hypothetical environmental forms, or hard-to-maintain hemolymph stages of H. perezi, we used sediment spiked with bacteria carrying the cloned PCR target to validate DNA extractions and PCR assays. Using this approach, we estimated that the PCR method has a detection limit of 13 gene copies per qPCR reaction, which translates to 1300 copies per gram of sediment. However, the minimum number of parasite cells that can be detected by the method is still uncertain, because there is disagreement in the literature as to how many rRNA gene copies there are per H. perezi cell. Nagle et al. [29], reported up to 10,000 copies/cell, while Li et al. [32], reported approximately 300 copies/cell. Interestingly, in a recent study of rRNA gene copy number in a dinoflagellate, Perini et al.[44] found that, depending on strain and stage in culture, Ostreopsis cf. ovata may possess from 81 to 8488 copies/cell of the LSU gene. This example provides one possible explanation for why the two estimates of H. perezi rRNA gene copy number may be so divergent. It would be worthwhile to assess the number of rRNA genes per H. perezi cell from a number of different isolates of the parasite and different developmental stages within the host. The recent development of in vitro culture for H. perezi by Li et al. [26] should provide valuable material for an assessment of rRNA gene copy number through different life stages.

Although the question of how many copies of the rRNA locus are present per cell of H. perezi was not addressed in the current study, the field data we obtained is pertinent to the issue. Many of the environmental samples in the current study contained from several hundred to several thousand ITS2 targets per gram sediment or per liter water/plankton. These values are more consistent with the gene copy numbers of Li et al. [32] unless we postulate that DNA recovery was poor, or that naked H. perezi DNA (representing less than one cell) was present in environmental samples. Using dilutions of ethanol-preserved H. perezi trophonts, we have also generated qPCR data that are more consistent with the lower number of gene copies/cell (Hanif and Schott, unpublished). With the lower gene copy/cell in mind, it should be recognized that samples that were negative by PCR (including replicate samples at sites that also had at least one positive) may indeed carry H. perezi DNA below the limit of detection. At 300 copies per cell, however, our detection limit of 1300 copies per gram sediment would translate to 4 cells. Collections of 0.25 gram would therefore be at the theoretical limit of one cell per sample. Sampling within sediment that had such a low density of H. perezi cells would therefore be expected to produce many samples that fail to capture a cell.

There are several themes that emerge from the application of the H. perezi qPCR assay to the coastal bay study. 1) H. perezi signal was widely detected throughout both coastal bays, and a potential sediment hotspot was detected in Indian River Inlet. 2) H. perezi DNA was detected in water/plankton samples from geographically dispersed areas in both coastal bays and was temporally consistent in prevalence at some locations such as Sinnickson in Chincoteague Bay, where it was detected during all seasons in which samples were collected. 3) H. perezi DNA was not found in non-portunid invertebrates, even those collected in the vicinity of a mass mortality of blue crabs infected with H. perezi.

Repeated detection of H. perezi DNA in sediment is consistent with the hypothesis that there is a sediment reservoir for the parasite. Hamilton et al.[45] suggested that, like other dinoflagellates, Hematodinium sp. may have a resting stage outside of its host as a part of its life cycle. Though a cyst or resting stage has not been observed by microscopy, molecular methods have been used to detect parasite DNA in sediment samples [21, 29]. Frischer et al. [23] presented evidence that the dinospore stage could establish infections in naïve blue crabs upon in situ exposure to water and sediment. Li et al.[46] showed that H. perezi amoeboid trophonts begin to lose viability after 24 h in seawater, but that dinospores survive up to seven days in aquaria [32]. H. perezi dinospores may be a precursor to a resting stage, which could accumulate in undisturbed sediments. Survival time of dinospores and hypothetical resting stages may vary according to factors such as predation, destruction by bacteria or environmental conditions [47, 48].

Plankton DNA samples from Sinnickson were positive for H. perezi in six of the nine months sampled (Additional file 2: Table S2 and Figure 5). Finding evidence of the parasite in the water/plankton, which is highly mobile, implies the existence of a persistent environmental source. Sediment could be that source, consistent with the finding of H. perezi DNA in sediment from Sinnickson in three of the nine dates sampled, and in all seasons. The other site with multiple positive plankton samples (two months out of nine months sampled) was Trappe Creek, which, like Sinnickson, has a sizeable freshwater input. Trappe Creek is close to, and upstream of, Newport Bay, which had positive sediment three of the nine months sampled. A possible mode of H. perezi DNA accumulation in sediments and water/plankton of some areas and not others is through tidal cycles that cause water (and plankton) to oscillate in/out of the small rivers feeding coastal bays. This is in contrast to the channel areas that have faster currents and have less slack water than may be found in the broad mouths of rivers. This is reflected in the sediment types in each area: only sand is present in the channels while silt-containing sediments are adjacent to broad river outlets.

The presence of H. perezi DNA in 2009 water and sediment samples obtained from Indian River Inlet was similar to observations (above) in Chincoteague Bay in 2010. Most of the Indian River Inlet H. perezi signal (including the hotspot) was present in silty sediment with high organic matter content, located in a backwater that experienced relatively little tidal flushing. On 07.29.08, H. perezi was not detected in water from Indian River Inlet, although sampling coincided with what appeared to be a H. perezi outbreak in blue crabs. However, the volume of water collected for analysis at that date (100 ml) was probably insufficient for detection of low amounts of parasite DNA. In 2009, when the volume of water sampled was increased to 1 L, qPCR analysis yielded 82 to 613 gene copies per liter. If the 2008 samples had contained parasite DNA at this abundance, a 100 ml water sample would have had been below the limit of detection for the qPCR assay.

On 07.29.08, the presence of many of dead and dying crabs that contained H. perezi gave us an opportunity to examine the water as well as other aquatic species during a parasite outbreak. H. perezi was detected in the muscle tissue of both lady crab (O. ocellatus) and green crab (C. maenas), which is in agreement with previous reports of Hematodinium sp. in these two species [39, 40]. Both of these crabs were collected (and appeared healthy) during the outbreak, and could represent incidental carriers or reservoir hosts. None of the grass shrimp, amphipods, or snails (grazing on shells of dead C. sapidus) were positive for H. perezi DNA, which argues that these are not major H. perezi reservoirs. This is consistent with the study by Pagenkopp et al.[37], who found no H. perezi in over 200 Palaemonetes sp. shrimp, and only 3% prevalence in caprellid amphipods. It was concluded that these were not likely sources of blue crab infections of H. perezi. Pagenkopp et al.[37] did however suggest that spider crabs (Libinia dubai), which we did not collect in our study, could serve as a biotic reservoir based on a prevalence of 17%.

Lastly, another possibility is that blue crabs with pre-patent H. perezi infections could act as reservoirs of the parasite. Blue crabs collected from winter trawls in Virginia, USA have been reported to be infected with H. perezi[22, 37]. Given that blue crabs are cannibalistic and feed heavily after hibernation, it has been suggested that infections could be transmitted via this mode. However, there have been conflicting results with cannibalism studies in laboratory settings. When Walker et al. [35] fed blue crabs fed infected conspecific tissue, they successfully transmitted the parasite to 7 of 11 blue crabs. A more recent study by Li et al.[46] showed a lack of transmission to naïve blue crabs and amphipods via cannibalism. More research is needed to elucidate if transmission may occur by entry of dinospores through carapace, gills, or breaks in integument, either while scavenging on infected prey or from environmental sources such as the water column or sediment.

Application of a species-specific H. perezi qPCR assay to DNA extracted from coastal bay sediment, plankton, and water samples suggests that H. perezi is continuously present in Delmarva coastal bays, and may present a source of transmission. These assays show that H. perezi DNA in sediments and water in these coastal bays is both temporally and spatially variable. The accumulation of parasite DNA in sediments or “hotspots” may be a consequence of parasite cells settling out of the water column, such as in retentive areas with silty sediments. There are gaps in our understanding of the natural history of the parasite as well as the mode of environmental or biological transmission. Filling some of those information gaps can be accomplished by using the qPCR methodology in a more focused study of the presence of parasite DNA in the environment, paired with detailed studies of water and sediment movements in the vicinity of one or more hotspots. A better understanding of the parasite’s life cycle and modes of transmission will help managers provide more effective fishery forecasts based on environmental conditions.