A European nudibranch new to the Gulf of Maine: Doris pseudoargus Rapp, 1827

The Gulf of Maine (GOM) has seen an increasing number of introduced species, some of which have significantly impacted benthic community structure. In 2017, a number of specimens of the European dorid nudibranch, Doris pseudoargus, were observed on rocky ledges in waters off Cape Ann, Massachusetts. The presence of numerous specimens and egg masses suggested the species had been established before 2017. Additional field observations and literature searches revealed specimens occurring from Nova Scotia to the northern end of the Cape Cod Canal, which poses the question, when and where did the initial introduction take place and how? Genetic analysis confirmed the species as Doris pseudoargus with genomic similarities to specimens from Northern Europe. Until this introduction, the GOM had only one species of sponge feeding nudibranch, Cadlina laevis, which is a trophic specialist on one genus of sponge. Unlike this endemic, stenotrophic feeding dorid, the introduction of another large sized, sponge predator known to feed on diverse sponge species, that occur from the intertidal to at least 25 m in depth has the potential to significantly impact community structure over a wide variety of hard bottom habitats.


Introduction
Introduced nonendemic species are a continuing and growing issue as global trade expands (Elton 1958;Simberloff 2013;Carlton 1989Carlton , 2003Fofonoff et al. 2009). The Gulf of Maine (GOM) has experienced a number of introductions of both animals (Berman et al. 1991;Pederson et al. 2005;Dijkstra et al. 2007;Harris and Mathieson 2000) and marine plants (Dijkstra et al. 2017b;Mathieson et al. 2008), some of which have become conspicuous members of marine communities (Harris and Tyrrell 2001;Pederson et al. 2005). One group of marine mollusks for which there is little information are members of the Order Nudibranchia. The diversity of nudibranch species in the GOM and the Atlantic Maritime Provinces is significantly lower than that of comparable regions in Europe (for GOM : Bleakney 1996;Moore 1964;Shine 2012; for Europe: Picton and Morrow, 1994;Brown 1976, 1984). Most of the known GOM nudibranch species are also found in Europe, thus it is difficult to determine if the endemic fauna is due to amphiatlantic distributions, or whether some of the species are cryptic introductions from longestablished trade routes between Europe and North America (Carlton 2003;Simberloff 2013).
Until recently, the only documented introduction of a European nudibranch species to the GOM, was the tritonid, Duvaucelia plebeia (Johnston, 1828). It appeared in the early 1980's and then rapidly disappeared within a few years (Allmon and Sebens 1988). A population of D. plebeia has been rediscovered in Eastport, Maine in recent years (Harris et al. manuscript in preparation). A unique feature of the GOM nudibranch fauna has been the paucity of dorid species that prey on sponges. Only one species, Cadlina laevis (Linnaeus, 1767), which preys on Halisarca sp., is known (Bleakney 1996;Shine 2012). More sponge-feeding dorids are known in Europe Thompson and Brown 1984), which is typical of marine communities in many regions of the world (Behrens and Hermosillo 2005;Chavanich et al. 2010;Coleman et al. 2015;Debelius and Kuiter, 2007;Gosliner 1987;Gosliner et al. 2008;Valdes et al. 2006).
According to Thompson and Brown (1984) D. pseudoargus is the largest dorid nudibranch in European waters, has an annual life cycle and occurs in a diversity of color patterns from yellow, orange and red to mottled browns. The color patterns are more variable than most species of nudibranchs and the annual life cycle is typical of larger nudibranchs with stable prey populations, such as sponges (Thompson 1964(Thompson , 1966(Thompson , 1976Todd 1981Todd , 1983.
This study on D. pseudoargus synthesizes additional information obtained since the initial 2017 observation, confirms their identity using molecular techniques, and discusses the ecological implications of its arrival in the GOM.

Materials and methods
Exploratory dives were made in the region of Cape Ann, Massachusetts by A. Shure. Photographs of individuals and their habitat were taken along with observational notes, including depth and associated fauna. On July 26, 2017, Shure collected four live specimens and took them to the University of New Hampshire (UNH), Durham, New Hampshire where they were maintained in a recirculating sea water system. The live specimens were photographed and then preserved in 95% EtOH. Two preserved specimens were deposited at the California Academy of Sciences, San Francisco, California, and one was used for genetic analysis. The resulting genomic data were compared with comparable sequence data sets from European samples in GenBank. Additional observations and field collections have since been made by Shure and others. The information obtained has been synthesized and is presented below.

Genetic Taxon sampling
A 658 bp region of the mitochondrial gene cytochrome oxidase I (COI) was sequenced for one Massachusetts D. pseudoargus specimen (CASIZ 223159A) and submitted to GenBank with the following accession number: MZ389059. Genetic comparisons were made with available GenBank material including the following seven sequences of European D. pseudoargus collected from the North Sea: one specimen from Esjerg, Denmark (KR084907); two specimens from Heligoland, Germany (KR084616 & KR084586); one specimen from Inverness, Scotland (KR084378); one specimen from Kingsbarns, Scotland (AY345030); one specimen from Kattegat, Sweden (MG935320); and one specimen from Skagerrak, Sweden (MG935407). Outgroup comparisons included Doris ocelligera (Bergh, 1881) and Doris adrianae Urgorri and Señarís, 2021 from the Mediterranean and Doris montereyensis J.G. Cooper, 1863 from the Eastern Pacific.

DNA extraction, amplification, and sequencing
DNA extraction was performed on a small tissue sample from CASIZ 223159A using the Qiagen Dneasy Blood and Tissue Kit (Qiagen, Valencia, CA, USA) spin column extraction method. Polymerase chain reaction (PCR) was used to amplify a 658 bp fragment of COI using Folmer et al. (1994) universal primers (HCO2198: 5′-TAA ACT TCA GGG AGA CCA AAA AAT CA-3'; LCO1490: 5′-GGT CAA CAA ATC ATA AAG ATA TTG G-3'). The reaction mixture contained the following: 2.5 μL of 10 × PCR buffer, 0.5 μL dNTPs (10 mM stock), 0.5 μL of each primer (10 μM stock), 0.25 μL DreamTaq™ Hot Start DNA Polymerase (5 U/μL, Thermo Fisher), 5 μL betaine, 2 μL bovine serum albumin (BSA), 4 μL of template DNA, and then filled to a final volume of 25 μL with Millipore-H 2 O. The following PCR protocol was run on a BioRad MyCycler Thermocycler (Bio-Rad Laboratories) at the California Academy of Sciences Center for Comparative Genomics (CCG): an initial denaturing for 3 min at 94 °C, followed by 40 cycles of denaturing for 30 s at 94 °C, annealing for 30 s at 46 °C and extension for 45 s at 72 °C with a final extension period of 10 min at 72 °C. Amplified DNA was stained with ethidium bromide and examined using gel electrophoresis on a 1% TBE agarose gel. Successfully amplified products were cleaned using an ExoSAP-IT protocol (USB Scientific) before being sequenced at ELIM Biopharmaceuticals (Hayward, CA, USA).

Sequence editing and alignment
Successfully sequenced fragments were assembled, trimmed to remove primers, and edited using Geneious v11.1.5 (Kearse et al. 2012) and Mesquite v3.61 (Maddison and Maddison 2018) before alignment with MAFFT (Katoh et al. 2009). Bayesian Inference (BI) and Maximum Likelihood (ML) analyses were used to estimate the evolutionary relationships within D. pseudoargus. Best-fit evolution model partition definitions for BI and ML analyses were determined for each codon position in COI using ModeltTest-NG (Darriba et al. 2020;Flouri et al. 2014) on XSEDE via the online CIPRES Science Gateway (Miller et al. 2010). The following best-fit evolution models were applied for the Bayesian Inference analysis: SYM + I for codon position 1, F81 + I for codon position 2, and HKY + G for codon position 3.

Phylogenetic construction and species delimitation analyses
Bayesian Inference was performed in MrBayes v3.2.7a (Ronquist and Huelsenbeck, 2003) and the dataset was run for 2.5 × 10 7 generations. Markov chains were sampled every 1000 generations and the standard 25% burn-in calculated before a 50% majority rule consensus tree of calculated posterior probabilities (pp) was created from the remaining tree estimates. Maximum likelihood was performed using randomized accelerated maximum likelihood (RA×ML) v8.2.12 (Stamatakis 2014) and non-parametric bootstrap values (bs) were estimated from 2.5 × 10 4 fast bootstrap runs set with the evolution model GTR + GAMMA. Tree branches were considered strongly supported if pp ≥ 0.95 and bs ≥ 70, while pp ≤ 0.94 and bs ≤ 70 were considered to have low support (Alfaro et al. 2003).

Results and observations
Observations of Doris pseudoargus from the GOM are summarized in Table 1. The first confirmed sightings were made in June 2017, when Shure found more than 10 adults on rocky ledges below 13 m off Cape Ann, Massachusetts. Egg masses were present, and individuals were observed mating. Figure 1 illustrates the wide diversity of color patterns exhibited by specimens occurring in the GOM. These color variations also include egg masses. Feeding wounds indicated the animals were grazing on the sponge Isodictya palmata (Ellis and Solander, 1786), the dominant occurring sponge at this site.
Following the original sighting, additional observations at sites in Massachusetts and Maine have been recorded by Shure and others (Table 1; Fig. 2). The initial observations were all at subtidal locations on rocky ledges. However, additional observations in Maine and Massachusetts have included intertidal sites. Subsequent dives by Shure recorded specimens at a series of additional GOM sites including Halfway Rock, Salem Sound, MA and the wreck of the ship/vessel Chester Poling, Gloucester, MA (Table 1). Nudibranchs at most of the deeper sites were observed feeding on I. palmata. Additional specimens were observed at several locations in 2018 through 2022. In each case, the habitat consisted of rocky ledges below the thermocline that supported significant sponge populations, particularly the palmate sponge, I. palmata. Specimens were observed once on artificial substrates on the wreck of the Chester Poling. In 2020, Shure observed specimens of D. psuedoargus at Graves Light Station, Boston Harbor at depths as shallow as 4 m (Table 1). Those animals were feeding on Halichondria panicea. Other specimens had been observed at this site in recent years (Captain J. Sullivan, personal communication to A. Shure 08/14/2020).
In 2019, specimens, including egg masses, were collected off Broad Cove Appledore Island, Isles of Shoals (Table 1). Adult D. pseudoargus and egg masses were observed at Mingo Rock, Isles of Shoals in 2019 and 2020. A single specimen was observed at Cape Neddick, York Beach, Maine in 2021. The southernmost reported collections have been from the northern end of the Cape Cod Canal at Scusset Beach, MA. The initial find was only a single animal, however for the second collection, three animals were procured. Nudibranchs from both collections were intertidal and were found coincident with populations of the sea anemone, Metridium senile. Very few sponges were observed in the vicinity. Recent postings on iNaturalist of specimens observed at intertidal locations included Calf Island near Plymouth, MA (Putnam 2022 https:// www. inatu ralist. org/ obser vatio ns/ 11257 7858) and Appledore Island (Bluedorn 2022 https:// www. inatu ralist. org/ obser vatio ns/ 12429 9989). The geographic distribution pattern for the species was recently expanded by Shure discovering a posted web citing of an animal likely to be D. pseudoargus collected off Yarmouth County, Nova Scotia, Canada in 2017 (phycophile, 2017 https:// www. inatu ralist. org/ obser vatio ns/ 70940 660). The specimen recorded from Nova Scotia was also intertidal.

Genetic analyses
Our Bayesian and Maximum Likelihood analyses (Fig. 3a) reveal that the specimen sequenced from Massachusetts (CASIZ 223159A) is nested within a well-supported (pp = 1, bs = 97) cluster of European specimens clearly identifiable as Doris pseudoargus. Within this cluster of D. pseudoargus, there is no clear geographical separation between the European populations and the relationships between all specimens are tentative due to low Bayesian internal node support and the resulting polytomies. The specimen from Massachusetts (MZ389059) groups together with specimens from the North Sea of Scotland (KR084378 & AY345030), one individual from Heligoland, Germany (KR084586) and one from Kattegat, Sweden (MG935320) in a weakly supported polytomy; however, this relationship does correspond with the relationships suggested in the TCS haplotype network (Fig. 3b).
The TCS haplotype network of eight specimens of D. pseudoargus suggest seven unique haplotypes with no geographical separation and one shared haplotype between a specimen from Inverness, Scotland (KR084378) and one of two specimens from Heligoland, Germany (KR084586). It also suggests that the Massachusetts specimen, the four European specimens from the shared polytomy, and the specimen from Skagerrak, Sweden (MG935407) are more closely related than the specimen from Esjerg, Denmark (KR084907) and the second specimen from Heligoland, Germany (KR084616). The ABGD analysis of the mitochondrial COI gene also supports the relationships suggested in the phylogenetic analyses and the TCS haplotype network (Fig. 3b). The genetic variation between the Massachusetts specimen and the five more closely related European specimens (KR084378, AY345030, KR084586, MG935320, MG935407) suggested in the TCS haplotype network was 0.15-0.31%. In contrast, the genetic variation between the Massachusetts specimen and the specimen from Denmark (KR084907) was 0.76% and the genetic variation between the Massachusetts specimen and one specimen from Heligoland, Germany (KR084586) was also 0.76%. Since the specimens within European waters exhibit a greater range of genetic variation between themselves (0.0-0.92%) than to the specimen from the western Atlantic (0.15-0.76%), it strongly supports the hypothesis that the Massachusetts specimen indeed represents an introduction from Europe.

Discussion
Doris pseudoargus is a common species in European waters and occurs from the intertidal to deeper habitats down to 300 m (Thompson and Brown 1984). The species feeds on several different sponge species in Europe, including Halichondria panicea, that is also common in the Gulf of Maine (GOM). The initial GOM discovery and the majority of subsequent observations have been from deeper subtidal locations (> 15 m) on rocky ledges (Table 1). Additional specimens have now been found at shallower depths including Graves Light Station (4 m), Cape Neddick (5 m), and intertidally in the Cape Cod Canal as well as Calf Island and Appledore Island (Table 1). The sighting in Nova Scotia (phycophile 2017) was also reported as intertidal. Thus, D. pseudoargus appears to be replicating its European natural history within the GOM (Thompson and Brown 1984). It is interesting that D. pseudoargus was not observed at Paddock Rock in 2016 while numerous breeding specimens were observed less than a year later in June 2017. Thompson (1966) described the life cycle of D. (as Archidoris) pseudoargus as recruiting in the summer and reproducing in the following spring. This is typical of many larger nudibranchs in Europe (Thompson 1964(Thompson , 1966Thompson and Brown 1984;Todd 1981Todd , 1983. Larger nudibranchs in the Gulf of Maine tend to follow a similar seasonal pattern (Harris 1973). It is likely that the initial recruitment of the observed 2017 population may actually have occurred earlier in 2016, but individuals were small, cryptic and not seen. The possibility exists that the specimen observed in Nova Scotia in 2017, along with the population at Paddock Rock, suggests the species was present some time prior to 2017, became established, and began to spread throughout the GOM without being detected. The increasing popularity of underwater photography and community science partners increase the likelihood that new populations will be detected and recorded by persons who are informed enough to realize a species may not be typical of that habitat. The presence of specialized photo guides (Martinez 2010;Shine 2012) available to the public also increases the possibilities of new or introduced species being sighted and identified. While the rediscovery of D. plebeia was made by a scientist (Harris) who studies nudibranchs, the discovery of D. pseudoargus was by an informed recreational diver (Shure).
The geographic range of D. pseudoargus that extends from the Mediterranean Sea to the North Atlantic implies that the species is able to tolerate a wide temperature range (eurythermal). Individuals feed entirely on sponges, including Halichondria panicea that occur over a wide depth range from the intertidal to more than 60 m. That D. pseudoargus had initially only been found at depths below 13 m at most sites is ecologically interesting and is probably why it was not detected earlier. However, the recent collections of this species at shallower depths (4-5 m) off Massachusetts Bay and Southern Maine, coupled with the intertidal findings from the Cape Cod Canal, Calf Island and Isles of Shoals and a probable specimen on the outer coast of Nova Scotia (phycophile 2017), where summer water temperatures can be warmer than those found in most of the GOM, begs the question of where D. pseudoargus was originally introduced and when.
The GOM is warming faster than many regions of the world (Pershing et al. 2015) and 2021was the warmest year on record (Gulf of Maine Research Institute (GMRI) 2022). Considering the wide geographic distribution of D. pseudoargus in Europe, it is unlikely that climate change will have a negative  Table 1 impact on the success of the species in the GOM, and its wide thermal tolerances may even allow it to expand south of Cape Cod. This geographic distribution will be predicated however upon the thermal tolerances of its prey species.
The size of the initial population observed off Cape Ann suggested it had been established in the region for some time since there was a range of sizes present, including spawning adults. In Europe, D. pseudoargus is known to have an annual life cycle that includes settling in the later summer, maturation through the winter, and then spawning in early summer (Thompson 1964(Thompson , 1966(Thompson , 1976Thompson and Brown 1984). However, climate change may now be impacting seasonal life history patterns (Pershing et al. 2015;Dijkstra et al. 2011Dijkstra et al. , 2017aStaudinger et al. 2019) as documented for the small dorid nudibranch, Onchidoris muricata (Müller, 1776) (Lambert 2013). In the case of O. muricata, the introduction of the bryozoan Membranipora membranacea (Linnaeus, 1767) as an abundant alternative food source in the summer months may have influenced the shift in its reproductive cycle (Lambert 2013;Lambert et al. 2016). Observations of D. pseudoargus in the GOM have documented large adults mating and producing egg masses from November through June. That suggests a more extended reproductive cycle. It is possible that the initial distribution in deeper, more stable (colder) water masses may have resulted in a breakdown in the typical seasonal reproductive pattern. It begs the question about what other species may be present in the Gulf of Maine, but at depths deeper than normally explored by divers. Bleakney (1996) lists nudibranch species reported for Nova Scotia that have not been documented in the GOM, so it is quite possible that new discoveries are likely as more divers become aware of the possibility of finding additional species.
A somewhat unique feature of D. psuedoargus compared to most GOM dorid nudibranch species is Each circle indicates a unique haplotype, size indicates number of specimens sharing that haplotype, lines between circles indicates a single substitution, small black circles indicate hypothetical haplotypes, and colors correspond to geographical locations in a the fact it occurs in a wide range of color patterns, including mottled browns ( Fig. 1; Thompson and Brown 1984). The specimens photographed by Shure and specimens brought for preservation ranged from somewhat uniform colors of yellow and orange (collected specimens) to more expanded color ranges including mottled brown (Fig. 1). Only Acanthodoris pilosa (Müller, 1789) occurs in a range of solid color patterns in the Gulf of Maine (Bleakney 1996;Harris, pers obs). The fact that D. pseudoargus shows a similar diversity of color patterns in the GOM as that seen in Europe suggests that color pattern is genetically programmed and not environmentally induced or dependent on diet.
The genetic analysis suggests that at least one source population was in Northern Europe (Fig. 3), but how the species arrived in North American waters is speculative presently. Doris pseudoargus occurs in Iceland (Lemche 1938;Rudman 2005), although to date, it has not been reported from Greenland or Newfoundland, so natural, westward colonization seems unlikely. The report of a probable specimen in Nova Scotia coincident with the first observed sighting off Cape Ann (Table 1) suggests that a possible route of introduction may have been through the Canadian Maritimes. The alternative would be an anthropogenic introduction into Massachusetts Bay with subsequent expansion of its range north and south. This hypothesis is supported by shared genetic similarities between Northern European specimens and the specimen analyzed from Paddock Rock (Fig. 3). Future collection and sequencing of D. pseudoargus from North America and known locations in Europe may provide further understanding for the introduction of this species and help to narrow the initial source of the species in the GOM.
Ecologically, the introduction of a large sponge predator that has not been present in the GOM and Canadian Maritimes, raises the issue of its potential impact on the endemic sponge fauna that historically has had very limited predation pressure. In shallower depths, only the sea star, Henricia sanguinolenta (O.F. Müller, 1776), has been a dominant sponge predator (Dijkstra et al. 2013;Sheild 1990;Sheild and Witman 1993;Van Volkom et al. 2021). To date, D. pseudoargus has only been observed to feed on Isodictya palmata and Halichondria panicea. Both sponge species occur in European waters (Dyrynda and Dyrynda 1990), but this dorid has not been cited as a predator of I. palmata in Europe (Miller 1961;Swennen 1961;Thompson and Brown 1984;Van Bragt 2004). More indepth ecological studies of this nudibranch in its subtidal habitats both in Europe and the GOM are needed to expand knowledge of the dietary preferences of this species. The documented sponge species that D. pseudoargus preys upon in the GOM have become less common in shallow waters (Harris pers. obs.). This reduction in abundance has been linked to space competition and availability by a series of introduced colonial ascidians (Dijkstra et al. 2007;Dijkstra and Harris 2009). At Paddock Rock, sponge populations do appear to have declined after the presence of large numbers of D. pseudoargus (Shure pers. obs.), and that is likely to occur at Mingo Rock with numerous feeding individuals being present there two years in a row (Table 1). Future directed studies are needed to understand the impact of nudibranch predation on community structure on vertical walls that typically are dominated by suspension feeding, colonial invertebrates of which sponges are an important component.
Acknowledgements This project was not directly supported by outside funding sources. The authors declare that no funds, grants, or other support were received during the preparation of this manuscript. J. Factor and W. Grossman provided specimens collected off Appledore Island and Scusset, MA respectively. E. Kintzing collected one specimen and provided observational information from the Isles of Shoals and Cape Neddick. Live specimens were maintained for observation in the aquatic room in Spaulding Hall, University of New Hampshire and genetic analyses were conducted at the California Academy of Sciences. Cape Ann offshore transportation and dive logistics were provided by D. Stillman, M. Stillman, S. Smith, and D. Shumway. The manuscript has benefitted from the comments and suggestions of external reviewers and their input is much appreciated.
Author contributions All authors contributed to the study conception and design. Material preparation, data collection and analysis were performed by AS, LH, AK, TMG and SAD. TMG and SAD conducted the genetic analyses. The first draft of the manuscript was written by LGH and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.
Funding The authors declare that no funds, grants, or other support were received during the preparation of this manuscript.
Data availability There are no separate data sets that were used in this manuscript other than those cited in Table 1, Figs. 2 and 3 and referenced in the text. Additional questions concerning field observations not provided in iNaturalist citations by A. Shure can be obtained by direct contact (alex. shure@gmail.com). The genetic sequence for the Massachusetts specimen was submitted to GenBank (MZ389059) and is also available through the California Academy of Sciences (CASIZ223159A). Additional information concerning the genetic analysis can be directed to T. Gosliner (tgosliner@ calacademy.org).

Competing interests
The authors have no relevant financial or non-financial interests to disclose. Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Springer Nature or its licensor holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.