Assessment and Molecular Characterization of Human Intestinal Parasites in Bivalves from Orchard Beach, NY, USA

Bivalves have been shown to be carriers of the human intestinal parasites Cryptosporidium parvum and Toxoplasma gondii. The goal of this study is to determine the prevalence of protozoan parasites in mollusks of New York City using a polymerase chain reaction (PCR)-based assay. Four species of mollusks, Mya arenaria, Geukensia demissa, Crassostrea virginica, and Mytilis edulis, were collected from Orchard Beach, NY in the fall of 2014, totaling 159 specimens. Each individual mollusk was dissected to harvest the digestive gland, the mantle, the gills, the foot and the siphon. The tissues were assayed for the presence of Cryptosporidium parvum, Giardia lamblia, and Toxoplasma gondii DNA by using primers that target parasite-specific genes. C. parvum was found at a prevalence of 50%, 11.3%, and 1%, respectively, in Mya arenaria, G. demissa, and Mytilis edulis. C. parvum DNA was detected in all the tissues of these bivalve species, except the gills. Furthermore, G. lamblia was detected in Mya arenaria, G. demissa, Crassostrea virginica and Mytilis edulis at a prevalence of 37.5%, 4.5%, 60%, and 20.6%, respectively, while T. gondii DNA was not detected.


Introduction
Bivalve mollusks are frequently used as bio-indicators in the appraisal of environmental quality. Among several features, their extensive geographic distribution and their stationary nature make them valuable indictors for chemical and biological pollution. As suspension-feeding organisms, they are exposed and accumulate considerable amount of pollutants, although those pollutants might be present at low concentration [1][2][3]. They have previously been used to monitor metal contamination of water [4,5].
Recent studies have found Toxoplasma gondii and Cryptosporidium parvum, two important human parasites, in several species of marine bivalves [6][7][8][9][10]. Thus, bivalves could be used as bio-sentinels for human parasites. According to the Center for Diseases Control (CDC), Cryptosporidium is the leading cause of water-borne disease among humans in the United States [11]. The mode of transmission for Cryptosporidium is through contaminated water and food or direct contact with an infected host. The oocysts have a thick shell that help the parasite to survive for long periods of time until it infects a definitive host.
Orchard beach is a public beach located in Pelham Bay Park in the Bronx borough of New York City. The goal of the present study is to determine the prevalence of human intestinal parasites in wild populations from Orchard Beach using molecular diagnostics tools.
Orchard beach is a public beach located in Pelham Bay Park in the Bronx borough of New York City. The goal of the present study is to determine the prevalence of human intestinal parasites in wild populations from Orchard Beach using molecular diagnostics tools.

Collection and Sample Processing
Bivalves were collected on Orchard beach, a public beach located in the Bronx in New York City. Collection was restricted to the Twin Island section of the park (40.871186° N, 73.784389° W) ( Figure  1). A total of 159 bivalves were collected at low tide on 9 September, 2014 belonging to four species. The bivalves collected were longneck steamers (Mya arenaria, 8), ribbed mussels (Geukensia demissa, 44) Atlantic oysters (Crassostrea virginica, 10), and blue mussels (Mytilis edulis, 97). They were all burrowed in the sand at the time of collection, except for Crassostrea virginica, which were attached to rocks. The samples were kept alive moist in sea water at 4 °C prior to dissection and DNA isolation. Following dissection, the tissues were stored at −80 °C.

DNA Extraction and Molecular Analysis
After thawing of the tissues, DNA was extracted from each tissue by using the DNAeasy Blood and Tissue extraction kit from Qiagen (Germantown, MD, USA). DNA purity was measured using the Nanodrop. The bivalves were tested for C. parvum, Giardia lamblia, and Toxoplasma gondii. All PCR assays were performed using 300 ng of purified DNA per reaction, PCR master mix (Qiagen, Germantown, MD, USA), parasite specific primers at 20 pmol/µL in a final volume of 25 µL. To detect G. lamblia, nested-PCR was performed using primers that target the β-giardin gene following a previously described protocol [12]. The forward primer for the first reaction was Gia7 (5′-AAGCCCGACGACCTCACCCGCAGTGC-3′) and the reverse primer was Gia759 (5′-GAGGCCGCCCTGGATCTTCGAGACGAC-3′). The nested PCR reaction was amplified using the forward primer (5′-GAACGAACGAGATCGAGGTCCG-3′) and the reverse primer (5′-CTCGACGAGCTT CGTGTT-3′). The genotype was assessed by digesting the amplicons with the restriction endonuclease HaeIII. Briefly, 20 µL of the total volume PCR reactions were digested with 10 unit of HaeIII for 4 h at 37 °C. To detect the presence of C. parvum DNA, PCR was performed using the LAX primer pairs previously described to specific for C. parvum [13,14]. The primer sequences are as follows forward: LAX469F 5′-CCGAGTTTGATCCAAAAAGTTACGAA-3′, and LAX869R.

DNA Extraction and Molecular Analysis
After thawing of the tissues, DNA was extracted from each tissue by using the DNAeasy Blood and Tissue extraction kit from Qiagen (Germantown, MD, USA). DNA purity was measured using the Nanodrop. The bivalves were tested for C. parvum, Giardia lamblia, and Toxoplasma gondii. All PCR assays were performed using 300 ng of purified DNA per reaction, PCR master mix (Qiagen, Germantown, MD, USA), parasite specific primers at 20 pmol/µL in a final volume of 25 µL. To detect G. lamblia, nested-PCR was performed using primers that target the β-giardin gene following a previously described protocol [12]. The forward primer for the first reaction was Gia7 (5 1 -AAGCCCGACGACCTCACCCGCAGTGC-3 1 ) and the reverse primer was Gia759 The nested PCR reaction was amplified using the forward primer (5 1 -GAACGAACGAGATCGAGGTCCG-3 1 ) and the reverse primer (5 1 -CTCGACGAGCTT CGTGTT-3 1 ). The genotype was assessed by digesting the amplicons with the restriction endonuclease HaeIII. Briefly, 20 µL of the total volume PCR reactions were digested with 10 unit of HaeIII for 4 h at 37˝C. To detect the presence of C. parvum DNA, PCR was performed using the LAX primer pairs previously described to specific for C. parvum [13,14]. The primer sequences are as follows forward: LAX469F 5 1 -CCGAGTTTGATCCAAAAAGTTACGAA-3 1 , and LAX869R. 5 1 -TAGCTCCTCATATGCCTTATTGAGTA-3 1 . The following conditions were used: 94˝C for 3 min, 94˝C for 45 s, 52˝C for 45 s, and 72˝C for 1 min, a final extension at 72˝C for 7 min. C. parvum purified DNA was used as a positive control [14]. T. gondii DNA was detected by using the T. gondii GRA6 primer sets (forward 5 1 -GTAGCGTGCTTGTTGGCGAC-3 1 , reverse primer, 5 1 -ACAAGACATAGAGTGCCCC-3 1 ). The PCR reactions were performed as described by Fazaeli et al. [15]. The PCR products were detected by ultraviolet light using agarose gels stained with ethidium bromide. Ten nanograms of purified parasite DNA was used as a positive control and water was used as a negative control for all PCR reactions. The resulting PCR products were sequenced and the genotype was analyzed with Sequencher 5.3.

Prevalence of Human Intestinal Parasites in Bivalves from Orchard Beach
A total of 159 specimens of bivalves belonging to four species were collected at low tide in the fall of 2014 to test for exposure to human intestinal parasites. We found that specimens of Mya arenaria (4/8), Geukensia demissa (5/44) and Mytilis edulis (1/97) tested positive for C. parvum DNA, while all the specimens of the Atlantic oysters were negative for C. parvum DNA (Table 1). In total, 50% of the tested Mya arenaria were positive for C. parvum, while 11% of the tested Geukensia demissa and 1% of Mytilis edulis were infected with C. parvum DNA, respectively. Sequence analysis confirmed the presence of C. parvum DNA in all positive samples. In contrast, 60% of the tested Crassostrea virginica was positive for G. lamblia DNA, while 37.5% of Mya arenaria, and 4.5% of the Geukensia demissa were positive, respectively (Table 1). Our data indicate contamination of the water of Orchard beach with both G. lamblia and C. parvum oocysts. We did not detect T. gondii DNA in any of the specimens suggesting no contamination with T. gondii oocysts at Orchard beach.

Parasite DNA Tissue Distribution
We determined the tissue distribution of C. parvum and G. lamblia by dissecting the foot, mantle, digestive glands, gills, siphon and abductor muscle (Crassostrea virginica only) of each collected bivalve specimen. We found that parasite DNA was found predominantly in the gills and digestive glands (Tables 2 and 3). Interestingly, G. lamblia DNA was not found in the mantle and digestive glands of Geukensia demissa. C. parvum. DNA was only detected in the foot and gills of Mytilis edulis (Tables 2 and 3). Although the sample size for Mya arenaria consisted of eight specimens, we found co-infection of two specimens of Mya arenaria with C. parvum and G. lamblia (Tables 2 and 3). The tissue distribution was distinct in the co-infected specimens. In specimen 4, G. lamblia DNA was detected in the siphon, while C. parvum DNA was detected in the foot and digestive glands (Tables 2 and 3). In contrast, G. lamblia DNA was found in the mantle of specimen 5, while C. parvum DNA was detected in the foot (Tables 2 and 3). Parasite DNA was not detected in every dissected tissues of the bivalve specimens, suggesting specific tissue distribution and little cross-contamination during the dissection (Tables 2 and 3).

Genotyping of G. lamblia from Bivalves Collected at Orchard Beach
Genetic variation is common among G. lamblia strains [16,17]. Seven genotypes are commonly accepted as distinctive evolutionary lineages by restriction fragment length polymorphism (RFLP) [18,19]. Several studies have revealed multiple differences among these lineages, including drug sensitivity, metabolism, and host preference [18,19]. Humans are only infected by assemblages A and B, while assemblages C and D infect dogs; hoofed animals are infected by assemblage E [18].
Moreover, assemblage F infects cats, while assemblage G only infects rodents [18]. Restriction fragment length polymorphism of G. lamblia positive samples, revealed assemblage A in all of the Mya arenaria specimens (Figure 2A and Table 4) as evidenced by the expected fragments of 201, 150, 110, and 50 bp. Sequence analysis of the PCR fragment confirmed that they belonged to sub-assemblage AII previously shown to be associated with humans and cats [20]. Likewise, assemblage A was the only genotype observed in the Crassostrea virginica positive samples ( Figure 2B and Table 4). Sequence analysis revealed G. lamblia of the sub-assemblage AI was revealed in all the oyster samples. Sub-assemblage AI is found in cats and humans [20]. We were able to assess the genotype of two of the G. demissa G. lamblia-positive specimens (2/3) and found that one specimen belonged to assemblage C as evidenced by DNA fragments of 194, 150, 102, 50, 15 bp, while the other belonged to assemblage D (200, 194, and 117 bp) ( Figure 2C and Table 4). The sequencing data support that finding. We genotyped 16/20 Mytilis edulis positive for G. lamblia DNA. Eight belonged to sub-assemblage AII, while the remainder belonged to assemblage C ( Figure 2D and Table 4). Overall, 68% of the bivalves collected at Orchard beach in the fall of 2014 had the G. lamblia sub-assemblage AII genotype, while 36% belonged to assemblage C, and 4% belonged to assemblage D ( Table 4). The DNA sequences from all the samples were exact matches to reference genotype.  Although, the four species of bivalves examined in this study were collected within feet of each other, a difference in the prevalence of C. parvum and G. lamblia assemblages was observed. G. lamblia assemblage A was detected in Mya arenaria, while assemblages C and D were found in Geukensia demissa. Interestingly, these two species of bivalves were within inches of each other at the time of collection. Future work with a larger sample size from each species will assess the differential degree of susceptibility of each bivalve species to C. parvum and G. lamblia genotypes. The Crassostrea virginica specimens were collected at different location on Orchard beach. In contrast to previous survey of C. parvum in Crassostrea virginica collected from the gulf of Maine, we found that the Crassostrea virginica specimens were negative for C. parvum DNA, but were positive for G. lamblia of the assemblage A genotype at a prevalence of 60% [10]. Although a link between eating raw oysters and infection with intestinal parasites such as C. parvum and G. lamblia has not been established, the screening of oysters for these parasites have been recommended [8].
Assemblages A and B are connected with human infections, whereas assemblages C and D are dog-specific [20][21][22]. This suggests that humans and dogs might have played a role in the contamination of Orchard beach. Intriguingly, Orchard beach was closed during the month of August 2015 because of fecal contamination from humans [23]. This fits well with our data. Additionally, our data indicate that Mya arenaria is an excellent bio-indicator species for detection of C. parvum and G. lamblia in aquatic environment. Mya arenaria might be more suitable bivalve species for the screening of human intestinal parasites than the commonly used Mytilis edulis. In the future, we will assess the effect of seasonal variation as well as size and age of the bivalve on parasite prevalence.  Although, the four species of bivalves examined in this study were collected within feet of each other, a difference in the prevalence of C. parvum and G. lamblia assemblages was observed. G. lamblia assemblage A was detected in Mya arenaria, while assemblages C and D were found in Geukensia demissa. Interestingly, these two species of bivalves were within inches of each other at the time of collection. Future work with a larger sample size from each species will assess the differential degree of susceptibility of each bivalve species to C. parvum and G. lamblia genotypes. The Crassostrea virginica specimens were collected at different location on Orchard beach. In contrast to previous survey of C. parvum in Crassostrea virginica collected from the gulf of Maine, we found that  2-12) and G. demissa (lanes 13-15) by digestion of secondary PCR with HaeIII. Lane 1, 100 bp marker; lanes 2, G. lanblia assemblage A, lanes 3-4 undetermined genotype, lanes 5-9 G. lanblia assemblage C, lanes 10 undetermined genotype, lanes 11-12 G. lanblia assemblage A, lanes 13, G. lanblia assemblage D, lanes 14, G. lanblia assemblage C, lanes 15, undetermined genotype.

Conclusions
We have demonstrated the presence of C. parvum and G. lamblia DNA in tissues of four bivalve species collected at Orchard beach, NY in the fall of 2014. Our study is the first report of these protozoan human intestinal parasites in bivalves from both Orchard Beach and New York City.
These results are relevant to public health because some mussels are often consumed raw and Orchard beach is well frequented during the warm summer months.