Molecular data reshape our understanding of the life cycles of three digeneans (Monorchiidae and Gymnophallidae) infecting the bivalve, Donax variabilis: it’s just a facultative host!

The coquina, Donax variabilis, is a known intermediate host of monorchiid and gymnophallid digeneans. Limited morphological criteria for the host and the digeneans’ larval stages have caused confusion in records. Herein, identities of coquinas from the United States (US) Atlantic coast were verified molecularly. We demonstrate that the current GenBank sequences for D. variabilis are erroneous, with the US sequence referring to D. fossor. Two cercariae and three metacercariae previously described in the Gulf of Mexico and one new cercaria were identified morphologically and molecularly, with only metacercariae occurring in both hosts. On the Southeast Atlantic coast, D. variabilis’ role is limited to being a facultative second intermediate host, and D. fossor, an older species, acts as both first and second intermediate hosts. Sequencing demonstrated 100% similarities between larval stages for each of the three digeneans. Sporocysts, single tail cercariae, and metacercariae in the incurrent siphon had sequences identical to those of monorchiid Lasiotocus trachinoti, for which we provide the complete life cycle. Adults are not known for the other two digeneans, and sequences from their larval stages were not identical to any in GenBank. Large sporocysts, cercariae (Cercaria choanura), and metacercariae in the coquinas’ foot were identified as Lasiotocus choanura (Hopkins, 1958) n. comb. Small sporocysts, furcocercous cercariae, and metacercariae in the mantle were identified as gymnophallid Parvatrema cf. donacis. We clarify records wherein authors recognized the three digenean species but confused their life stages, and probably the hosts, as D. variabilis is sympatric with cryptic D. texasianus in the Gulf of Mexico.

Ré sumé -Les données moléculaires remodèlent notre compréhension des cycles de vie de trois Digènes (Monorchiidae et Gymnophallidae) parasites du bivalve Donax variabilis : ce n'est qu'un hôte facultatif !. La telline, Donax variabilis, est un hôte intermédiaire connu de digènes Monorchiidae et Gymnophallidae. Le nombre limité de critères morphologiques pour identifier les hôtes et les stades larvaires des parasites sont à la base de confusion dans la littérature. Dans cette étude nous avons identifié par séquençage les tellines collectées sur la côte Atlantique des États-Unis. Nous démontrons que les séquences pour D. variabilis dans GenBank sont incorrectes et que celle du spécimen américain est en fait celle de D. fossor. Deux cercaires et trois métacercaires décrites au préalable dans le Golfe du Mexique ainsi qu'une nouvelle cercaire ont été identifiées morphologiquement et par séquençage. Les métacercaires seules infectent les deux hôtes. Sur la côte sud-est Atlantique, D. variabilis a un rôle limité à seulement celui de second hôte intermédiaire facultatif, et D. fossor, une espèce plus ancienne, est premier et second hôte intermédiaire. Le séquençage a montré 100 % de similarité entre les stades larvaires de chacun des trois digènes. Des sporocystes avec des cercaires à queue simple et les métacercaires dans le siphon inhalant ont des séquences identiques à celles de Lasiotocus trachinoti, pour qui nous donnons le cycle complet. Les adultes des deux autres digènes ne sont pas connus et les séquences des stades larvaires ne sont identiques à aucune dans GenBank. Des sporocystes de grande taille, les cercaires (Cercaria choanura) et les métacercaires enkystées dans le pied des tellines sont identifiées comme étant Lasiotocus choanura (Hopkins, 1958) n. comb. Des petits sporocystes avec des furcocercaires et des métacercaires dans le manteau des tellines sont identifiées comme étant le gymnophallide Parvatrema cf. donacis. Nous clarifions les Introduction Coquinas, Donax spp., are small hard-shell clams that live in large colonies in the surf zone of sandy beaches. Their life span is short, they are fast developing, and they are ecologically significant as they are filter feeders that are preyed upon by numerous species of crustaceans, shore birds, and marine fishes [11,33,36,65]. In particular, D. variabilis Say, 1822 is common on the southeastern US Atlantic coast (south of Chesapeake Bay to mid-east Florida) and throughout the Gulf of Mexico from southwest Florida, USA to Campeche State, Mexico [2,66]. It has been reported as host to a variety of parasites [in 4,32] including several digenean larvae [29,30,38,62,63]. However, because these clams live in sympatry with congeneric cryptic species [1,2], such host reports are questionable. Furthermore, while species names were often attributed to the cercarial stages (e.g., Cercaria choanura Hopkins, 1958 [30]; Cercaria fragosa Holliman, 1961 [29]), no association between sporocysts, cercariae, and metacercariae could be made accurately at the time of those descriptions due to limitations in morphological characters common to these stages [9]. Similarly, no connection could be established between these larval stages and adults, which infect shore birds and/or marine fishes. Hence, none of the monorchiid and gymnophallid species reported from coquinas in the Gulf of Mexico have known life cycles, and the suprageneric species are species inquirendae until their identity is resolved [9]. In coastal South Carolina (SC), coquinas D. variabilis and D. fossor are cryptic species, and preliminary opportunistic observations over the past decade revealed that infection by digeneans was common although unreported ( Fig. 1). Using morphology, we identified two monorchiids and one gymnophallid previously described from D. variabilis in the Gulf of Mexico, and we discovered one previously unreported cercaria. The advent of molecular identification via DNA sequencing has made the unraveling of complex parasite life cycles much easier compared to experimental infections (e.g., [10,40,67]). The use of molecular tools allowed us to connect the various life cycle stages encountered, to identify cryptic intermediate hosts, and to elucidate the complete life cycle of one of the monorchiid species by identifying the adult stage from local fish, and thus, in part clarifying the current state of confusion. Further, we found that D. variabilis was not the sole intermediate host in the life cycles of these parasites, with, in contrast, its role limited to being a facultative, second intermediate host.

Collection of hosts and parasites
A total of 1095 coquinas were sampled opportunistically in the months of June through October between 2018 and 2020 at low tide from the swash zone along Folly Beach (32°66 0 34 N, 79°91 0 99 W) and Sullivan's Island (32°76 0 60 N, 79°82 0 09 W) on the SC coast. Some specimens were haphazardly collected between 2010 and 2015 from the same localities, but not included in prevalence calculations. Additional coquinas were collected from North Wildwood, New Jersey (NJ) (39°00 0 00 N, 74°78 0 70 W) on September 6, 2020. Adult digeneans were collected from Florida Pompano Trachinotus carolinus (L.) in October 2014 at the same localities in SC and were identified and sequenced (see below) in an attempt to match resulting sequences with those from larval stages found infecting coquinas.
Coquina specimens were maintained on a bed of sand in aerated seawater in the laboratory and examined within 3 days post-collection. After the clams were shucked, gonads, digestive gland, mantle, foot, and both inhalant and exhalant siphons were isolated, squashed separately on individual slides, and examined for infection under a compound microscope. The inner side of the shells were also examined under the dissecting microscope. Clam tissue (uninfected gonad and adductor muscle) was fixed in 95% ethanol or directly frozen at À20°C (SC), or whole clams were fixed in 100% EtOH (NJ) for molecular study. Parasites from SC coquinas and fish were isolated in filtered seawater, and parasites from NJ were isolated from the EtOH-fixed clams. There was no attempt to observe the natural emergence of cercariae out of the coquinas. Some specimens were examined and photographed fresh while others, including manually-excysted metacercariae, were heatkilled by slowly passing a flame under the slides on which they were placed. Parasites were fixed in either 95% ethanol or 5% neutral buffered formalin (NBF) for molecular and morphological studies, respectively.

Morphological study
Fixed parasitic stages were stained with Semichon's acetocarmine or Mayer's hematoxylin, dehydrated according to standard protocol, and mounted in Canada balsam or Permount. Measurements were taken using an ocular micrometer and are reported in micrometers unless otherwise specified; ranges are provided with averages in parentheses. We used the morphological criteria of shell outline and size of folds and papillae of the incurrent siphon described by Simone and Dougherty [56] to verify molecular identification of the clams (see below; Fig. 1

Molecular study DNA extraction
Genomic DNA of the clam hosts and parasite specimens was extracted using the DNeasy Blood and Tissue kit (Qiagen, Hilden, Germany) based on the manufacturer's protocol, with the exception of decreasing the elution volume to 100 lL for the parasites. Parasite DNA was concentrated to~50 lL using an Eppendorf VacufugePlus (Hamburg, Germany) prior to PCR.

Parasite PCR and sequencing
Portions of nuclear and mitochondrial genes were amplified and sequenced for comparison to one another and to those in the National Center for Biotechnology Information (NCBI) GenBank database. Primers and primer sequences used for each locus are in Table 1. Primers 3S [12] or GA1 [5] and ITS2.2 [20] were used to amplify the second internal transcribed spacer (ITS2) region of the ribosomal RNA (rRNA) gene of digeneans. For the 3S + ITS2.2 PCR, a 20-lL total reaction contained 0.5Â GoTaq Flexi PCR Buffer (Promega, Madison, WI, USA), 2 mM MgCl 2 , 0.4 mM dNTPs, each primer at 0.4 lM, 0.06 U lL À1 Promega GoTaq DNA polymerase, and 1 lL template DNA. PCR cycling conditions were done as in Cutmore et al. [22]  All products were electrophoresed on 1% agarose gels (100 V, 30 min), subsequently stained with GelRed (Biotium, Inc., Hayward, CA, USA), and visualized under UV light. PCR products were cleaned using ExoSAP-IT (Affymetrix, Santa Clara, CA, USA), following the manufacturer's protocol and sent to Eurofins MWG Operon LLC (Louisville, KY, USA) for direct bi-directional sequencing using the same primers as above. Complementary sequences were compared to one another and to their chromatograms using Sequencher version 5.4 (Gene Codes, Ann Arbor, MI, USA). Sequence alignments were performed in MEGA X [34] using MUSCLE [25]. Alignments were trimmed so that all sequences were of equal length with no terminal gaps, and p-distance calculations were also done in MEGA X [34] with 1000 bootstrap replicates. For ITS2 alignments, Gblocks was used to remove poorly aligned positions and divergent regions [14]. A portion of the small subunit (18S) rRNA gene was also amplified and sequenced using primers 600F and A27R [37], but sequences from the two monorchiid species were only 0.3% different from one another in a 356-bp alignment and thus not useful in differentiating the two monorchiid species found. While the 18S sequences obtained were deposited into GenBank (see Supplementary Tables 1 and 2), we did not pursue sequencing for this marker.
For the phylogenetic analysis of monorchiid 28S data, we aligned the Lasiotocus sequences from this study with those from other monorchiids in GenBank (namely those used in the phylogeny of Wee et al. [64] with some exceptions due to lack of overlap with sequences generated herein) and sequences from sister family Lissorchiidae (to be used as an outgroup) using MUSCLE [25]. Indels larger than 5 bases and affecting >10% of the sequences were removed, resulting in an 848-bp alignment. Maximum parsimony analysis was conducted in MEGA X [34] with 1000 bootstrap replicates and 100 random additions.
Host PCR, sequencing, RAPD All coquinas infected with sporocysts (except one collected in 2018 that was lost and 5 collected in 2010-2015; see Supplementary Tables 1 and 2) and subsamples of those infected with metacercariae in the foot, siphon, and/or mantle were identified molecularly. We amplified and sequenced a portion of COI using primers LCO1490Mba and HCO2198Mba [43] for a total of 65 hosts. A 20-lL total reaction contained 1Â Promega GoTaq Flexi PCR Buffer, 2.5 mM MgCl 2 , 0.2 mM dNTPs, 0.25 mg mL À1 bovine serum albumin (BSA), each primer at 0.5 lM, 0.05 U lL À1 Promega GoTaq DNA polymerase, and 3 lL template DNA. The PCR protocol was as follows: 5 min at 95°C; 35 cycles of 95°C for 1 min, 40°C for 1 min, 72°C for 1 min; and 72°C for 7 min. Products were electrophoresed, visualized, purified, and sequenced as above. Sequences were also edited as above and compared to sequences available in the NCBI database using BLAST [3].
There were two COI sequences from D. variabilis in GenBank: one obtained from a specimen collected in New York (NY), USA (MH012241; [43]) and another from Mauritius (MN178795; [direct submission]); however, these sequences were only 81% similar to one another. Given this dissimilarity, we assumed that the appointed organism name of the sequence from Mauritius was an error, as this locality is outside this species' geographic range [2]. Unexpectedly, however, the NY D. variabilis sequence (MH012241) was 99-100% similar to sequences of specimens we identified morphologically as D. fossor, but only 83-84% similar to sequences of specimens that we identified as D. variabilis, which themselves were 78-81% similar to other Donax spp. COI sequences in GenBank (KY951446, MF668317, MT334589). The discrepancy between the sequence from the specimens we identified morphologically as D. variabilis and the NY D. variabilis COI sequence in GenBank led us to suspect yet another error in the identity of the specimen associated with this GenBank sequence. This contention was further supported by the fact that the specimen used to generate the D. variabilis COI sequence in GenBank was collected much further north [43] than the northernmost range limit of this species [2,56]. To confirm this suspected error, specimens of D. fossor were collected from NJ, USA (see above), where it occurs in the absence of other Donax species [2]. Ten individuals were dissected, their larval digeneans collected, and genomic DNA of clam tissues (n = 6) and of the parasites was extracted as described above. PCR and sequencing also were performed as above. Additionally, we further verified the host species identity using a random amplified polymorphic DNA (RAPD) marker following the protocol of Adamkewicz and Harasewych [1]. Marker E18 (Table 1) is diagnostic for D. variabilis, producing a 500-bp product [1]. A 25-lL total reaction contained 1Â Fermentas Buffer + KCl (Vilnius, Lithuania), 1 mg mL À1 BSA, 1.9 mM MgCl 2 , 0.25 mM dNTPs, primer at 0.5 lM, 0.05 U lL À1 Promega GoTaq DNA polymerase, and 1 lL (15 ng) template DNA. Cycling was as follows: 94°C for 5 min followed by 45 cycles of 94°C for 1 min, 36°C for 1 min, 72°C for 2 min, and a final extension at 72°C for 5 min. Products were electrophoresed on 2% agarose gel stained with GelRed alongside a 100 bp ladder (IBI Scientific, Dubuque, IA, USA) at 75 V, 60 min and visualized under UV light. Four individuals of each Donax species were assayed.   [30], and while larger in all dimensions, resembled the ones we found. However, he erroneously suggested that they were encysted cercariae Cercaria choanura (a cercaria with a collar-like tail that we also report below) (Figs. 3, 4A, 4B). Wardle [62], who correctly suggested that the monorchiid L. trachinoti infected individuals of D. variabilis, also mistakenly associated C. choanura with this species (Fig. 3). Adults found herein fit the description of L. trachinoti from its type host, T. carolinus (see [47]).  Adult: unknown Remarks: Two bi-ocellate, distome cercariae with protusible collar-like tails have been described from D. variabilis, both from the Gulf of Mexico: Cercaria choanura [30] (same as Cercaria A of Loesch [38] according to Hopkins [30]) and C. pocillator [29]. The adult stage of neither species is known. Herein, sporocysts and cercariae were found only in individuals of D. fossor (Fig. 3). Cercariae are smaller in all dimensions compared to both previously described cercariae, but most closely resemble C. choanura in having a conspicuous genital primordium, which C. pocillator lacks, and in having very short ceca. We could not observe penetration glands. Based on our observations, which are further supported by molecular data (see below), C. choanura is a valid taxon belonging to Lasiotocus Looss, 1907. We thus propose the new combination Lasiotocus choanura (Hopkins, 1958) n. comb. Although Holliman [29] stated that C. pocillator "may encyst in the first intermediate host", he did not observe metacercariae in the clams that he examined. Hopkins [30] and Loesch [38] incorrectly attributed metacercariae they found in high prevalence in the inhalant siphon of D. variabilis to C. choanura (Fig. 3). Therefore, based on our results, these authors have confused the two species of monorchiids that we have identified. Loesch [38] did report the presence of "A" type metacercariae (associated with cotylocercous Cercaria A) in the foot of the coquinas that were examined, and Wardle [62] inferred that metacercariae in the foot of the coquina were encysted cercariae (Cercaria XXVII), which, based on the presence of short ceca, closely resembled C. choanura. Our results confirm these suggestions. However, Wardle [62] further suggested that this species could be the monorchiid L. trachinoti from the Florida Pompano, which our results showed was not the case (see above; Fig. 3). Significantly, both Loesch [38] and Hopkins [30] reported metacercariae in specimens of D. tumida Philippi, 1849 (now D. texasianus Philippi, 1847; [17,45]). Remarks: Loesch [38] called the metacercariae and furcocercous cercariae he encountered in specimens of D. variabilis from the Gulf of Mexico (Texas) Gymnophallus metacercariae and Gymnophallus cercariae, respectively. Hopkins [30] revisited these identifications and described the metacercariae as being Parvatrema donacis (Fig. 3); he also inferred that the Gymnophallus cercariae of Loesch [38] were cercariae of Pa. donacis. Holliman [29] described another gymnophallid cercaria, Cercaria fragosa, from individuals of D. variabilis in the Gulf of Mexico (Florida) and differentiated it from the cercaria described by Hopkins [30], mainly by the presence of bristles and spines on its body and tail and by the large number of embryos (over 50) per sporocyst compared to about 6 in putative sporocysts of Pa. donacis. Herein, we found sporocysts and cercariae only in D. fossor.

Morphological study of the parasites
The presence of a conspicuous genital pore located at a distance from the anterior edge of the ventral sucker in metacercariae allowed the identification of a species of Parvatrema Cable, 1953 [19]. Although smaller in all dimensions, metacercariae in D. fossor and D. variabilis and cercariae in D. fossor from SC closely resemble those of Pa. donacis [30]. In particular, the number of embryos per sporocyst in our specimens was between 5 and 20 and neither spines nor bristles were observed on the body and tail of cercariae. Therefore, based on metacercaria and cercaria morphology, we identify these three stages as Pa. cf. donacis until DNA sequences of Pa. donacis from the type locality in the Gulf of Mexico are available.

Molecular study
Lasiotocus trachinoti Overstreet and Brown, 1970 Sequences obtained from adult specimens (n = 8) from Florida Pompano were identical to those from the metacercariae found in the inhalant siphon of coquinas of both species (n = 11) and the simple-tailed cercariae with sporocysts (n = 13) based on alignments of ITS2 (312 bp), 28S (893 bp), and COI sequences (433 bp) (Tables 2 and 3; Supplementary Tables 1  and 2). Intraspecific sequence variation among COI sequences ranged from 0 to 0.1%, but appeared random and not host-or life-stage dependent; the other markers did not demonstrate any variation. ITS2 and 28S sequences were 100% similar to those of L. trachinoti from an adult in the Gulf of Mexico  Lasiotocus choanura (Hopkins, 1958) n. comb Sequences from metacercariae found in the foot of coquinas of both species (n = 19) were identical to those from C. choanura with sporocysts (n = 7) based on alignments of ITS2 (304 bp), 28S (870 bp), and COI sequences (433 bp) (Tables 2 and 3; Supplementary Tables 1 and 2). Intraspecific sequence variation among COI sequences ranged from 0 to 0.5%, but again appeared random and not host-or life-stage dependent; the other markers did not show any intraspecific variation. The ITS2 sequence from this species differed from that of L. trachinoti ITS2 sequence by 1.3% based on a 304-bp alignment, 1.7% based on a 870-bp alignment of 28S sequences, and 8.1% based on a 433-bp alignment of COI sequences. ITS2, 28S, and COI sequences differed from L. mulli sequences by 13.4% (352 bp), 6.6% (870 bp), and 13.6% (308 bp), respectively. The maximum parsimony analysis of monorchiid 28S data produced a consensus tree with a topology similar to that generated by Wee et al. [64] (Fig. 6). The placement of this species within Lasiotocus was based on this phylogeny where its sequence formed a well-supported monophyletic clade with those of L. mulli and L. trachinoti. Sequences were deposited into GenBank and accession numbers can be found in Supplementary Tables 1 and 2.

Host identification
As mentioned above, partial COI gene sequences of specimens morphologically identified as D. fossor (both from SC and NJ) were !99% similar to a D. variabilis COI sequence in GenBank (MH012241) collected from NY [43], whereas D. variabilis specimens had sequences that were only 83-84% similar to this NY D. variabilis sequence. Specimens morphologically identified as D. variabilis produced the diagnostic 500-bp product using RAPD marker E18, while specimens identified as D. fossor did not (Fig. 7), confirming the identity error of D. variabilis in GenBank. Coquinas infected with sporocysts of both monorchiids (L. trachinoti n = 13; L. choanura n. comb. n = 20) and of the gymnophallid (n = 1) were all identified molecularly as D. fossor (Table 2); note: one coquina infected by gymnophallid sporocysts and cercariae was lost and thus identified as Donax sp. (Supplementary  Table 2). One individual D. fossor (MW628253, Table 2) had a double infection with sporocysts and cercariae of both monorchiid species. Identities of another 32 coquinas with metacercariae of either species were also verified molecularly, showing that both D. variabilis and D. fossor were infected by the three types of metacercariae (Tables 2 and 3; Fig. 3). Partial COI gene sequences from hosts were deposited into GenBank as accession numbers MW628241-MW628287 (D. fossor; Table 2) and MW628288-MW628305 (D. variabilis; Table 3

Discussion
Coquinas D. fossor and D. variabilis from the US southeastern Atlantic coast were found to be infected by larval stages of two monorchiid and one gymnophallid digenean species that were previously described from D. variabilis in the Gulf of Mexico [30,38,62]. One of these monorchiids was also reported to occur in D. tumida [30,38], now D. texasianus (see [17,45]) in the Gulf of Mexico. Significantly, however, D. variabilis was identified as both first and second intermediate hosts of these digeneans in the Gulf of Mexico, but neither sporocyst nor cercaria was found in this species on the SC coast, a finding brought to light only after we molecularly identified all the individual clams serving as first intermediate hosts of these three digeneans in our collections. Thus, given the long term and recurrent issue with Donax species misidentification [1,2,45], and because D. variabilis lives in sympatry with the cryptic species D. texasianus in the Gulf of Mexico [1,2,60], it is then highly probable that the first intermediate host from these earlier studies was in fact misidentified. This is supported by a detail in Loesch [38], who reported infection by Cercaria A (i.e., presumed Cercaria choanura of Hopkins [30] and Lasiotocus choanura herein) in D. variabilis in the text, but illustrated it as being from D. tumida (= D. texasianus; [17,45]) as he found one infected individual of this species. Loesch [38] further mentioned that all individuals of D. tumida he examined from his collection site in Louisiana, USA and most of those from Mustang Island, Texas, USA were infected by metacercariae type A, thus indicating that infection was indeed observed in both Donax species. One other area of confusion from the literature is that, whereas Loesch [38], Hopkins [30], and Wardle [62] all suggested certain connections between the life stages that they observed, such inferences could only be speculative given that no experimental infection was carried out and that sporocysts and cercariae cannot be identified to species level on the sole basis of morphology. As a result, the associations of these digeneans' larval stages proposed by these authors were also mistaken. Sequencing data allowed us to right these previous records, which we clarify herein. In a nutshell, the monorchiid Cercaria choanura is identified as Lasiotocus choanura n. comb., its metacercariae encyst in the foot of the clam, and its adults are not known. Cercariae of L. trachinoti, on the other hand, have a simple long retractile tail, had not been observed prior to our study, its metacercariae encyst in the inhalant siphon, and its adults infect the digestive tract of the Florida Pompano, T. carolinus. Cercariae of the gymnophallid Parvatrema cf. donacis are furcocercous, its metacercariae do not encyst (as is typical of gymnophallids) and infect the mantle of the coquinas, and its adult remains to be identified. Lastly, for the US Atlantic coast, D. fossor serves as sole first intermediate host and also as second intermediate host, while D. variabilis is only a second intermediate host.
Monorchiidae is a very speciose family [39,66]; however, of the 20 species with a bivalve as intermediate host, only nine have known life cycles: one from the Mediterranean coast [41], one from the European Atlantic coast ( [10]; in [40]), three from the North American Atlantic coast [42,57,58], two from the South American Atlantic coast [8,28], and two from the North American Pacific coast [23,68]. Most of these known life cycles are two species-host cycles, as is the case in D. fossor herein for both Lasiotocus species found, which include the use of the same bivalve species as both first and second intermediate host. For rare species of this family, metacercariae may be found within sporocysts [8,10] or may be extruded into the water [57,58]. More typically, however, metacercariae encyst in the tissues of the bivalve, similar to those we observed in D. variabilis and D. fossor. Cercariae known as Cercaria choanura were identified herein as Lasiotocus choanura n. comb. and encyst in the foot of the coquina and not in the inhalant siphon as Hopkins [30] reported. Adults have yet to be found and/or sequenced, but according to the 28S rRNA gene phylogeny, this species appears to be sister to L. trachinoti. Wardle [62] inferred correctly that these cercariae encyst in the foot, but incorrectly suggested that they were of L. trachinoti, which as adults, commonly infect the digestive tract of the Florida Pompano ( [47,48]; present study). While cercariae of L. trachinoti develop in coquinas, albeit D. fossor on the US Atlantic coast, they had not been observed prior to our study, and we thus demonstrate this cycle in its entirety. There is no significant morphological or molecular difference (based on 1TS2 and 28S sequences) between adult specimens of L. trachinoti from the Atlantic coast and the Gulf of Mexico, which further supports that our data can be applied to the infections of coquinas in the Gulf of Mexico. The cercariae of L. trachinoti have a simple long and retractile tail (similar to those of group 1 in Cremonte et al. [18]), whereas that of L. minutus (Manter, 1931) is microcercous [57,58]. Hence, the cercaria type, which is typically thought to be a reliable indicator of phylogenetic relationships [21], would appear to vary within Lasiotocus, a discrepancy further supporting the contention that some species assignments to the genus Lasiotocus are problematic and need revisiting [39,48,55]. Wee et al. [64] began restructuring this genus, and although cercariae morphology was not taken into account in their study, it may be significant to note that L. elongatus (Manter, 1931), which was reassigned to the genus Proctotrema Odhner, 1911, has microcercous cercariae. Wee et al. [64], however, could only gather molecular data for the type species, L. mulli, and while these authors suggest that Lasiotocus needs further refinement, for lack of current knowledge in some morphological structures and molecular data, L. trachinoti and L. minutus currently remain in the same genus until further data are available. Examination of two nuclear markers showed that sequences from the two monorchiids we encountered were less than 2% dissimilar from one another, and our phylogenetic analysis of 28S rRNA gene data showed that they form a well-supported monophyletic clade with L. mulli, supporting their inclusion in the genus. However, the COI sequences werẽ 8% different between sequences from the two species and over 14% different from the L. mulli sequence. While there is, to our knowledge, no a priori species delineation for monorchiids using COI sequence data (or any other locus), the difference in COI sequences, which show little to no intraspecific variation, along with the strikingly different cercariae, lead us to question whether these monorchiid species belong in the same genus. Finding and/or molecularly identifying the adults of Lasiotocus choanura n. comb. as well as the cercariae of L. mulli would be of particular interest to clarify this conundrum.
The high prevalence of L. trachinoti metacercariae in the inhalant siphon of the coquinas compared to the rather rare occurrence of sporocysts in D. fossor supports the contention of Bagnato [8] that a long tail in cercariae suggests a free-living lifestyle. This monorchiid may have a two-host life cycle utilizing the same species of bivalve (in this case, D. fossor) as both first and second intermediate hosts but not the same individuals given the rarity of sporocyst-infected coquinas compared to those harboring metacercariae. Maillard [41] experimentally demonstrated such a cycle for the monorchiid, Paratimonia gobii Prévot & Bartoli, 1967 by infecting individuals of the bivalve Abra ovata (now A. segmentum [Récluz, 1843]see [66]). In this case, cercariae leave the first intermediate host (an individual of A. ovata), only to be sucked in by another individual of A. ovata (the second intermediate host) within which they encyst and accumulate in the inhalant siphon. The infected siphons then become autotomic and are eaten by the fish definitive host (a goby). While we do not know if siphons infected by metacercariae of L. trachinoti become autotomic, observations of coquinas with incurrent siphons missing folds indicate that Florida Pompano (known predators of coquinas [6]) may also become infected by partially eating siphons without killing the clam. Because the coquinas would stay alive and accrue metacercariae, Pompano feeding in this manner would enhance these parasites' fitness, especially combined with the occurrence of an additional second intermediate host as metacercariae also commonly infect the siphon of individuals of D. variabilis. Taken together, these factors would then explain the high prevalence of L. trachinoti on the SC coast.
Gymnophallid adult digeneans infect marine charadriiform (shorebirds) or anseriform (ducks) [19,24,54] and typically have a three-host life cycle with bivalves as first intermediate hosts and bivalves or, more rarely, gastropods [13] or polychaetes [35,51] as second intermediate hosts [54]. There are also known examples of unusual cycles that involve parthenogenetic metacercariae [26,27] or no free-living stage [31]. Less rare are cases of shortened life cycles involving the same first and second intermediate hosts (e.g., [40,67]) as we report herein for Pa. cf. donacis in D. fossor. Loesch [38] reported that all 11 of the clams (presumed to be D. variabilis) he examined for gymnophallid metacercariae were infected, and Hopkins [30] described the gymnophallid Pa. donacis from metacercariae from individuals presumed to be D. variabilis on Mustang Island, Texas, wherein 85% of his coquinas were infected compared to the 0.4% we report herein. Hopkins [30] did not find furcocercous cercariae in any of the 100 coquinas he examined, Loesch [38] found them in 0.7% (8/1017), and we found them in only 0.18% (2/1095). Truncated life cycles have a marked seasonality when definitive hosts are migratory birds [52], which is often the case for gymnophallids [16,67]. While all these authors focused most of their parasite observations during summer as we did, the difference in latitude (and thus the onset of seasons) between SC and the Gulf of Mexico could explain in part such discrepancies in prevalence of infection. Differences in the abundance and migratory patterns of the definitive host(s) at different localities could also be highly relevant. Although Loesch [38] did not report finding adult gymnophallids in the plovers, sanderlings, and eastern willets that he observed feed on coquinas, it appears that he only examined them for gut contents. Thus, any of these birds, as well as for instance, Ruddy turnstones that are also known migratory predators of coquinas [53], could be definitive hosts.
Complexity in parasite life cycles evolves according to costbenefit trade-offs for the parasite [7,15,44]. In digeneans, which encounter repetitive constraints in the transmission of their various larval stages, life cycles often become shorter [50]. The occurrence of three species of digeneans with two host-life cycles using the same individuals of D. fossor as both first and second intermediate hosts may support this contention. On the other hand, the short life span and relatively limited biotope of coquinas add other constraints to parasite transmission that may be circumvented by the addition of another intermediate host, albeit facultative as in this case. Such lateral incorporation of a facultative host can provide extra fitness to the parasites, and is not unusual in digeneans [49]. Our data best support this second scenario because phylogenetic analyses indicate that D. fossor and D. texasianus (the cryptic species that occurs in the Gulf of Mexico) are both older than D. variabilis and display ancestral characters (smaller size, non-migratory behavior, and subtidal habitat) [2]. Since from an evolutionary standpoint, two-host cycles are thought to precede three-host cycles [21], it thus may be inferred that the two monorchiids reported herein have two-host cycles that expand to integrate a third host, either individuals of D. fossor, as our prevalence data indicate, or of D. variabilis when present. This may also occur for the gymnophallid, but we have less evidence given the low prevalence of this parasite. The nonessential role of D. variabilis in the life cycles of these digeneans is evident, as these cycles occur only via D. fossor in NJ where it is the sole Donax species present. The increase in complexity of these parasites' life cycles via the addition of D. variabilis, when present, as a second intermediate host would be expected to alter the transmission dynamics of the parasites and affect their fitness, thereby modifying the selective pressure impacting the evolution of their life history traits.
In conclusion, the combined use of morphological and molecular tools has allowed us to correct the identity and the association of digenean life cycle stages previously described from the clam "D. variabilis" in the Gulf of Mexico. These findings also allowed us to narrow the role of D. variabilis to that of a facultative second intermediate host on the southeastern US Atlantic coast, whereas the more ancient D. fossor is both a first and second intermediate host in the life cycles of these three digenean species encountered. In light of these findings, and because the misidentification of Donax species on the US Atlantic and Gulf of Mexico coastlines has been a long-term recurrent issue, we suspect that sporocyst-infected coquinas in earlier reports were probably misidentified as D. variabilis and that D. texasianus, also older than D. variabilis, acts as first and second intermediate host for these parasites in the Gulf of Mexico. How the addition of D. variabilis in the life cycles of these digeneans impacts the transmission dynamics and whether it results in an increased fitness of these parasites will be the object of future studies.

Supplementary materials
Supplementary material is available at https://www.parasitejournal.org/10.1051/parasite/2021027/olm Supplementary Table 1. GenBank accession numbers of parasite cytochrome c oxidase I (COI) mitochondrial DNA, second internal transcribed spacer region of the ribosomal RNA gene (ITS2), and partial large (28S) and small (18S) subunit ribosomal RNA gene sequences from sporocysts and metacercariae collected from Donax spp. and adults from carangid Trachinotus carolinus in South Carolina and New Jersey (in bold), USA. Italicized accession numbers are sequences that were successfully sequenced in only one direction.
Supplementary Table 2. GenBank accession numbers for parasite cytochrome c oxidase I (COI) mitochondrial DNA, second internal transcribed spacer region of the ribosomal RNA gene (ITS2), and partial large (28S) and small (18S) subunit ribosomal RNA gene sequences from sporocysts and metacercariae from Donax specimens whose identity could not be verified molecularly because they were collected between 2010 and 2015 prior to discovering that D. fossor can be found sporadically in the intertidal zone in South Carolina with D. variabilis. Coquinas infected by sporocysts and cercariae of the Lasiotocus species was inferred to be D. fossor based upon extensive molecular studies since that time (Tables 2  and 3 in text). Data for Parvatrema cf. donacis sporocystinfected Donax were too limited to allow us to make the same inference. Specimens listed here were not accounted for in prevalence data. Italicized accession numbers are sequences that were successfully sequenced in only one direction.