Confirmation of the shell-boring oyster parasite Polydora websteri (Polychaeta: Spionidae) in Washington State, USA

Invasions by shell-boring polychaetes such as Polydora websteri Hartman have resulted in the collapse of oyster aquaculture industries in Australia, New Zealand, and Hawaii. These worms burrow into bivalve shells, creating unsightly mud blisters that are unappealing to consumers and, when nicked during shucking, release mud and detritus that can foul oyster meats. Recent findings of mud blisters on the shells of Pacific oysters (Crassostrea gigas Thunberg) in Washington State suggest a new spionid polychaete outbreak. To determine the identity of the polychaete causing these blisters, we obtained Pacific oysters from two locations in Puget Sound and examined them for blisters and burrows caused by polychaete worms. Specimens were also obtained from eastern oysters (Crassostrea virginica Gmelin) collected in New York for morphological and molecular comparison. We compared polychaete morphology to original descriptions, extracted DNA and sequenced mitochondrial (cytochrome c oxidase I [mtCOI]) and nuclear (small subunit 18S rRNA [18S rRNA]) genes to determine a species-level molecular identification for these worms. Our data show that Polydora websteri are present in the mud blisters from oysters grown in Puget Sound, constituting the first confirmed record of this species in Washington State. The presence of this notorious invader could threaten the sustainability of oyster aquaculture in Washington, which currently produces more farmed bivalves than any other US state.

www.nature.com/scientificreports www.nature.com/scientificreports/ the sides of the caruncle (midway and between the palps), and as granular patches on dorsal surface near base of branchiae from middle segments posteriorly.
Chaetiger 1 with neurochaetae, without notochaetae, with digitiform notopodial lobes (Figs. 3A-D and 4A-D). Cilia of lateral organs present between notopodial lobe and neuropodial lobes of chaetiger 1 and present between notopodial and neurochaetae of chaetiger 2 (additional lateral organs may be present on more posterior chaetigers but have been lost during fixation). Winged capillary notochaetae of chaetigers 2-4, 6 and subsequent   www.nature.com/scientificreports www.nature.com/scientificreports/ chaetigers arranged in three successive rows, reduced to thin notochaetae in posterior chaetigers; no specialized posterior notochaetae. Winged capillary neurochaetae of chaetigers 2-4, 6 and subsequent chaetigers arranged in two vertical rows; 5-8 bidentate hooded hooks begin on chaetiger 7, not accompanied by capillaries, increasing to 8-10 in series at chaetiger 9; hooks with approximately right angle between main fang and shaft, with constriction www.nature.com/scientificreports www.nature.com/scientificreports/ on shaft; glandular pouches near base of ventral-most hooded hook in chaetigers 7-8, observed by the external portion of secretory cells which appear as small papillae.
Chaetiger 5 almost twice as large as chaetigers 4 and 6, with slightly curved row of 5-7 exposed major spines and additional embedded spines, major spines alternating with pennoned companion chaetae, sometimes exhibiting frayed tips; anterior dorsal fascicle of 4-6 geniculate notochaetae present and tips directed posteriorly, ventral fascicle of 4-6 winged capillary neurochaetae below row of major spines (Figs. 3F-H and 4F-H). Major spines  falcate, with shallow lateral flange, most visible in younger, posterior spines (Figs. 3G,H and 4G,H); older, anterior spines may appear to have lateral tooth but this is the remains of the worn flange (Fig. 4H).
Remarks. The specimens of Polydora websteri from WA and NY match the taxonomically important features of those in the original description (Hartman in 27 ), redescription 54 , and more recent reports 9,10,52,63 . Although the caruncle was described as extending to end of chaetiger 2 in the lectotypes of Polydora websteri 54 , others have found it reaching mid-chaetiger 3 9 , end of chaetiger 3 63 or to chaetiger 4 10 . In the present specimens the caruncle extended to mid-chaetiger 2 in some and through end of chaetiger 3 in others. As noted by 54 Fig. 1a in 10,52 ). After fixation in formalin and preservation in ethanol, the differences in palp pigmentation patterns are retained (e.g., USNM 1606136 from NY with line of pigmentation; USNM 1606128 from WA with bands of pigmentation). The methyl green staining pattern is similar to that observed by Read 9 , although he noted granular staining in anterior branchiae. Major spines are falcate, with a shallow lateral flange (Figs. 3G,H and 4G,H); although older anterior spines may appear to have a lateral tooth, this is the remains of the worn flange (Fig. 4H). Lateral organs (=lateral ciliated organs; see 64 ) were present on chaetigers 1 and 2, but presence/absence on posterior chaetigers should be confirmed based on specimens fixed in glutaraldehyde. One of the specimens from WA (USNM 1606127; Fig. 3B) had hooded hooks beginning on chaetiger 6, but this seems to be an abnormal specimen; all other reports and specimens examined herein show that the hooded hooks begin on chaetiger 7.
Prevalence. Of the 183 oysters collected from south Puget Sound, 40% (74 individuals) were infested with at least one blister or burrow. Among oysters from Oakland Bay, in South Puget Sound (Fig. 2), 53% were infested; among oysters from Totten Inlet, 34% were infested.

Molecular identification.
All of the specimens identified as Polydora websteri by morphological analysis were confirmed as belonging to that species by molecular analyses. Of the 13 specimens collected from Oakland Bay sequenced at 18S rRNA, 12 were identified as Polydora websteri (Table 1). Our 18S rRNA neighbor-joining phylogeny indicated that these 12 sequences clustered in the same clade as the Polydora websteri sequences from Genbank. As Rice et al. 10 reported for sequences of Polydora websteri from several Atlantic coast, Gulf coast, and Hawaiin specimens, all Polydora websteri 18S rRNA sequences in our study were identical. The Oakland 18S rRNA sequences also match the four sequences from Long Island ( Fig. 5) with the exception of sequence LI4B which had several unresolved bases. There was more structure evident in the phylogeny based on sequences from the mtCO1 gene. Even so, ten worms from Oakland Bay and the four from Long Island that were sequenced with mtCOI also clustered with Polydora websteri in the mtCOI neighbor-joining phylogeny (Fig. 6, Table 1) and were clearly divergent from all other published mtCO1 sequences for Polydora sp on Genbank. Some worms collected from Oakland Bay and all Totten Inlet worms were not included in our molecular analyses as we do not have clear morphological identifications and there are not matching, published molecular data available on Genbank for these specimens. In summary, however, both 18S rRNA and mtCOI sequence analysis indicates that Polydora websteri is present in Oakland Bay, Puget Sound, Washington.
Haplotype diversity, nucleotide diversity and the average number of nucleotide differences were all substantially lower for the 18S rRNA gene sequences relative to the mtCO1 gene sequences (Table 2) as would be expected for the more conserved nuclear locus. For mtCOI, the mean intraspecific for Polydora websteri was 0.002 (n = 21, Table 3) and the interspecific distances between Polydora websteri and the other species ranged between 0.185 and 0.240 (Table 3). For nuclear 18S, the mean intraspecific distance for Polydora websteri was 0.00 (n = 21, Table 4) and the interspecific distances between Polydora websteri and the other species was 0.02 (Table 4).

Discussion
Our findings constitute the first report of Polydora websteri in Washington State, United States. The presence of this shell-boring polydorin poses a danger to the region's valuable oyster aquaculture industry. All worms from Oakland Bay that were identified as Polydora websteri based on diagnostic morphological features also clustered with GenBank sequences of Polydora websteri both in the 18S rRNA and mtCOI phylogenetic trees (Table 1, Figs. 5 and 6). Based on detailed morphological analysis, specimens of Polydora websteri from Oakland Bay (Fig. 3) matched previous descriptions and the newly collected material from Long Island, NY near the type locality (Fig. 4); the same specimens that we morphologically identified were also sequenced, and morphological and molecular diagnoses agreed. We therefore confirm the presence of Polydora websteri, a shell-boring mud worm, in the shells of Washington State Pacific oysters. Polydora websteri has never before been reported from Washington.
The fact that Polydora websteri has never before been documented in Washington State oysters suggests a recent introduction, but it is also possible that the species has been present in the region for some time and has undergone a recent increase in prevalence perhaps associated to the aquaculture industry or environmental changes. Extensive exchange of shell and live oysters among regions in Washington continues to the present day, and to such an extent that Polydora websteri populations are genetically homogenous across broad swathes of their contemporary range 10 . Washington State has a long history of exchange with other oyster-growing regions 65 and polydorin pelagic larvae may also have been introduced through ballast water 66,67 . Although it is likely that Polydora websteri is native to Asia and exotic to North America 10 , we suggest that Polydora websteri be considered cryptogenic in Washington State 68 until further research can resolve its origins. Considering the species is distributed north and south of Washington (e.g. 28,29,[57][58][59][60], it is likely that the species has been present in this region but has never before been reported because it occurred only at low prevalence until recently. The prevalence of Polydora websteri is sensitive to environmental change. For example, increasing siltation can increase the susceptibility of Crassostrea virginica to Polydora websteri 69 . In contrast, reducing pH actually decreases susceptibility to infestation 70 . Because Polydora websteri can recruit to both live and dead oyster shells 30 , the expansion of the oyster aquaculture industry, oyster restoration, and increased density of oysters in beds across the state might have promoted an increase in transmission and prevalence if the polychaete was already present. Whatever their origin or how affected they are by changing conditions, the blister-forming polychaetes we document here are a new challenge for Washington State oyster growers and the government agencies charged with management of shellfish stocks. Because Polydora websteri is a generalist pest 9,32,33 , it may impact other shellfish species of ecological, economic, and cultural importance to Washington State. An important example is the Olympia oyster (Ostrea lurida), an overexploited native species that is the focus of intensive restoration efforts 71 . Mussels [11][12][13] , scallops [14][15][16] , and abalone ( 17 see review in 4 ) are also at risk. Given the important ecosystem services provided by filter-feeding shellfish species 48 , a polydorin outbreak could affect more than just the bottom line of the shellfish industry; ecosystem functioning is also at risk.
We were not able to definitively identify the majority of worms collected from Totten Inlet (Fig. 2) using our combined morphological and molecular approach. However, our work indicates that the Puget Sound region hosts several cryptogenic spionid polychaete species, all of which may pose a danger to the regions oyster aquaculture industry. In our research, we positively identified the notorious shell-boring polydorin, Polydora websteri, in commercially farmed Pacific oysters, providing the first formal documentation of this globally distributed pest in Washington State. The pathology caused by shell-boring mud worms results in unsightly blisters that reduce the market value of infested oysters, especially those served on the half-shell. Washington's Pacific oyster industry is dominated by the half-shell market 61 , and given the high prevalence of infestation found in this study, these pests have the potential to threaten the valuable Pacific oyster aquaculture operations in Washington State.

Methods
Oyster collections. To assess whether shell-boring polychaetes were present in Washington Pacific oysters (Crassostrea gigas) and to confirm the species identity of these worms, we purchased 183 commercially grown oysters from retail shellfish farms in Washington State, USA. Of these, 69 individuals came from Oakland Bay (47°13′45.93″, −123°3′19.43″, Fig. 2, Table 1), and 114 individuals were from Totten Inlet (47°9′43.09″, −122°59′19.62″, Fig. 2 Worm collections. All oysters were shucked, and the soft tissues removed. We observed right and left valves under a stereomicroscope for indications of mud worm infestation, such as burrows and blisters. All oysters (with or without infestation) were photographed and measured (height and length of the shell) using a digital caliper (results in Supplementary Table 2). We removed any worms present in blisters or burrows with a probe or forceps, or by fracturing shells with a hammer to expose worms in their burrows. Once removed from the shell, we photographed the worms and fixed them whole in 95% ethanol for molecular analysis or, in some cases, sectioned worms such that molecular analysis of a worm (typically middle and posterior chaetigers) could be linked with morphological analysis of the same worm (typically anterior ends).

Morphological examination.
For morphological examination, worms were fixed in 4% formalin/seawater overnight, washed in warm tap water, and transferred to 70% ethyl alcohol (EtOH) for storage. For examination with a scanning electron microscope (SEM), the specimens were dehydrated in an ascending ethanol series through 100% EtOH. Drying was accomplished with a Samdri 795 Critical Point Dryer. Once dried, the specimens were mounted on aluminum stubs, coated with gold using an EMS-550 Sputter coater, and viewed with a FEI Quanta 250 SEM. Voucher specimens (Table 1)  Infestation prevalence. We considered any oyster that had at least one blister or burrow to be infested.
Prevalence was calculated as the proportion of infested oysters in each sample. We also calculated the number of blisters/burrows per oyster.  Figure 5. Neighbor-joining phylogeny based on Kimura 2-parameter distances using trimmed 18S1 rRNA sequences (1000 replicates). The optimal tree with the sum of branch length = 0.087 is shown. Clades which were recovered in greater than 80% of replicate trees in the bootstrap test are shown along the branches leading to the clade nodes. The tree is drawn to scale, with branch lengths in the same units as those of the evolutionary distances used to infer the phylogenetic tree. The rate variation among sites was modeled with a gamma distribution (shape parameter = 1). Pseudopolydora dayii (KY677907) was used as an outgroup. New sequences reported in this study labeled with OAK and LI were collected in Oakland Bay and Long Island respectively. Figure 6. Neighbor-joining phylogeny based on Kimura 2-parameter distances using trimmed mtCO1 sequences (1000 replicates). The optimal tree with the sum of branch length = 1.20 is shown. Clades which were recovered in greater than 80% of replicate trees in the bootstrap test are shown along the branches leading to the clade nodes. The tree is drawn to scale, with branch lengths in the same units as those of the evolutionary distances used to infer the phylogenetic tree. The rate variation among sites was modeled with a gamma distribution (shape parameter = 1). Pseudopolydora dayii (KY677907) was used as an outgroup. New sequences reported in this study labeled with OAK and LI were collected in Oakland Bay and Long Island respectively. (2020) 10:3961 | https://doi.org/10.1038/s41598-020-60805-w www.nature.com/scientificreports www.nature.com/scientificreports/ sequencing and phylogenetic analysis of variation at the nuclear 18S rRNA and mitochondrial cytochrome c oxidase I [mtCOI] genes 73 . We followed the protocol of 73 in using a molecular approach to identify worms recovered from blisters and burrows.
For a subset (n = 27) of the total number of worms vouchered (n = 107) and for four additional worms collected from Long Island, New York, we extracted DNA using DNeasy 96 Blood & Tissue Kit (Qiagen, Valencia, CA) following the manufacturers' instructions. We used two genes for molecular identification: the nuclear 18S rRNA [18S rRNA] and the mitochondrial cytochrome c oxidase I [mtCOI]. For the 18S rRNA gene, three regions were amplified: 18S-1F1/18S-1R632, 18S-2F576/18S-2R1209, and 18S-3F1129/18S-R1172 74 . For mtCOI, we amplified one region: Dorid_COI.3 F/Dorid_COI.1R 73 . Primer sequences are presented in Table 5. The expected length of the fragments was between 680 and 780 bp. We used polymerase chain reaction (PCR) to amplify DNA using a C1000 Touch (Bio-Rad, Hercules, CA) thermocycler. PCR reactions consisted of 2.5 µM of each primer, 2.0 µl of template DNA, 5 µl of 2X PCR buffer (Phusion ® Hot Start Flex, Thermo Scientific, Foster City, CA), and 0.5 µl MgSO 4 in a 10-µl reaction. 18S rRNA was PCR-amplified with an initial activation step of three minutes at 98 °C, followed by 35 cycles of denaturation (30 seconds at 98 °C), annealing (30 seconds at 54 °C), and extension (30 seconds at 72 °C) with a final extension step (10 minutes at 72 °C). Only the first of the three regions for 18S rRNA (18S-1F1/18S-1R632) was used for analysis because the other two did not amplify consistently. mtCOI was PCR-amplified with an initial activation step of 98 °C, followed by 30 cycles of: denaturation (30 seconds at 98 °C), annealing (30 seconds at 45 °C), and extension (60 seconds at 72 °C) with a final step of five minutes at 72 °C. The size of the PCR amplicons was checked in a 1.5% agarose gel. PCR products were sequenced in both directions using the amplification primers at Molecular Cloning Laboratories (San Francisco, CA).

Molecular analysis.
We combined forward and reverse complementary sequences of 18S rRNA and mtCOI genes using Geneious (version 11.0.5) to create consensus sequences. The consensus sequences were submitted to NCBI and registered in GenBank (accession nos. in Table 1 and Supplementary Table 1). The 18S rRNA sequences we generated with primers 18S-1F1 and 18S-1R632 were approximately 660 bp in length, but were trimmed the final alignment to a common length of 614 bp to remove poorly aligned terminal ends. Similarly,   Table 4. Intraspecific and interspecific Kimura-2-parameter (KP2) distances for the nuclear 18S sequences of Polydora websteri from our dataset and GenBank.
Scientific RepoRtS | (2020) 10:3961 | https://doi.org/10.1038/s41598-020-60805-w www.nature.com/scientificreports www.nature.com/scientificreports/ mtCOI sequences were initially 680 bp in length and were trimmed to 554 bp for analysis. Initially, we aligned our partial consensus sequences of 18S rRNA and mtCOI genes with sequences from the Polydora websteri and other species in the genus Polydora, obtained from GenBank (Table 6). For this alignment we only employed Genbank sequences that have been published along with a clear morphological description of the species. We reconstructed phylogenetic trees using the neighbor-joining method based on Kimura 2-parameter model with 1000 bootstrap replications. We used a gamma distribution (shape parameter = 1) to mode the rate variation among sites. The 18S rRNA analysis involved 25 nucleotide sequences. Codon positions included were 1st+2nd+3rd+Noncoding. All positions containing gaps and missing data were eliminated. There were a total of 566 positions in the final dataset and the optimal total branch length was 0.097. The mtCO1 analysis involved 58 sequences and a total of 540 positions with an optimal total branch length of 1.2. We used the Molecular Evolutionary Genetics Analysis software (MEGA version 7.0.26), with Pseudopolydora dayii Simon as an outgroup. We used MEGA 7.0.26 to determine the haplotype diversity, nucleotide diversity and the average number of nucleotide differences ( Table 2). Pairwise distance for intraspecific and interspecific polydorid species for mt COI (six species, Table 3) and nuclear 18S (three species, Table 4), were also calculated using MEGA 7.0.26 with Kimura's two-parameter method with a gamma rate variation distribution. The Kimura two-parameter metric was chosen to facilitate comparison with previous studies 52 . The sequences used to calculate these distances were retrieved from GenBank and from our own dataset (Tables 3 and 4).