Adaptive radiation in extremophilic Dorvilleidae (Annelida): diversification of a single colonizer or multiple independent lineages?

Metazoan inhabitants of extreme environments typically evolved from forms found in less extreme habitats. Understanding the prevalence with which animals move into and ultimately thrive in extreme environments is critical to elucidating how complex life adapts to extreme conditions. Methane seep sediments along the Oregon and California margins have low oxygen and very high hydrogen sulfide levels, rendering them inhospitable to many life forms. Nonetheless, several closely related lineages of dorvilleid annelids, including members of Ophryotrocha, Parougia, and Exallopus, thrive at these sites in association with bacterial mats and vesicomyid clam beds. These organisms are ideal for examining adaptive radiations in extreme environments. Did dorvilleid annelids invade these extreme environments once and then diversify? Alternatively, did multiple independent lineages adapt to seep conditions? To address these questions, we examined the evolutionary history of methane-seep dorvilleids using 16S and Cyt b genes in an ecological context. Our results indicate that dorvilleids invaded these extreme habitats at least four times, implying preadaptation to life at seeps. Additionally, we recovered considerably more dorvilleid diversity than is currently recognized. A total of 3 major clades (designated “Ophryotrocha,” “Mixed Genera” and “Parougia”) and 12 terminal lineages or species were encountered. Two of these lineages represented a known species, Parougia oregonensis, whereas the remaining 10 lineages were newly discovered species. Certain lineages exhibited affinity to geography, habitat, sediment depth, and/or diet, suggesting that dorvilleids at methane seeps radiated via specialization and resource partitioning.


Introduction
The adaptability of life is truly remarkable, as evidenced by the ability of organisms to exist in most environments on Earth. Certain habitats, however, challenge the persistence of life with adverse environmental conditions, such as extreme temperature, pressure, desiccation, pH, radiation, salinity, oxygen concentration, and/or toxins (reviewed in Rothschild and Mancinelli 2001). Biological diversity in these extreme habitats is often limited (e.g., Gough et al. 2000;Tsurumi 2003;Tobler et al. 2006), and yet certain organisms have evolved physiological tolerance, protective structures, repair capabilities, and other mechanisms that enable survival and success under extreme conditions (reviewed in Grieshaber and Völkel 1998;McMullin et al. 2000;Rothschild and Mancinelli 2001). In complex multi-cellular organisms, such mechanisms can be sophisticated, and presumably energetically expensive, implying that adaptation to extreme environments should be rare.
Methane seeps are one example of an extreme environment. Biological assemblages in these ecosystems interact with methane-and sulfide-rich fluid percolating upward through sediments. As water migrates through these sediments, a series of methane-oxidizing and sulfate-reducing microbial reactions transpire, resulting in extremely high sulfide pore-water concentrations (Sahling et al. 2002;Valentine 2002;Levin et al. 2003). Additionally, little dissolved oxygen penetrates into methane-seep sediments due to strong upward fluid flow as well as reaction with sulfides or reduced metals (Tryon et al. 2001;Levin et al. 2003). Because of the high toxicity of sulfide (i.e., levels greater than 1 mmol/L are toxic to most metazoans; Grieshaber and Völkel 1998) and unavailability of dissolved oxygen, methane seeps are among the most physiologically challenging environments for aerobic animals. Typically, species diversity is low at methane seeps (Levin 2005;Cordes et al. 2010;Levin et al. 2010), but several taxa may have radiated within seeps, including dorvilleid, ampharetid, hesionid, siboglinid, and polynoid annelids as well as vesicomyid clams (reviewed in Sibuet and Olu 1998;Levin 2005).
Previous studies characterizing diversity of methane seep fauna have, understandably, given considerable attention to large symbiotic taxa including tube worms, vesycomid clams, and Bathymodiolus spp. mussels as well as archeal and eubacterial communities that are critical to ecosystem function (reviewed in Sibuet and Olu 1998;Levin 2005). At methane seeps (500-880 m deep) off northern California and Oregon, dorvilleid polychaetes are the dominant macrofauna in microbial-mat-covered sediments, and are abundant in vesicomyid clam beds, ampharetid beds, and on authigenic carbonates rocks ( Fig. 1; Sahling et al. 2002;Levin et al. 2003Levin et al. , 2010Thurber et al. 2009Thurber et al. , 2012. These animals are most concentrated in sediments with sulfide concentrations of 1-5 mmol/L, where they achieve remarkably high densities (reaching greater than 11,000 individuals per square meter; Levin et al. 2003). Furthermore, the majority of seep-dwelling dovilleids are new to science (Levin et al. 2003(Levin et al. , 2010. Three factors make this system unusual: (1) many different species coexist in the same sediments, (2) a single annelid family comprises most of the macrofauna, and (3) high densities of animals thrive at very high-sulfide concentrations. Given these factors, dorvilleids at the Cascadian margin methane seeps provide a suitable system to address questions about evolution at physiologically challenging environments. Hypothetically, the exceptional tolerance to low-oxygen and high-sulfide concentrations of dorvelleid annelids has allowed this group to exploit ecological niches that are unavailable to most organisms. Over evolutionary time, absence of predators and competitors at western North American methane seeps could function as an evolutionary release, facilitating diversification (Levin et al. 2003). Whether this diversification occurred following colonization by a single lineage or multiple-independent colonization events is a key question considered here.
"Dorvilleidae" is an old and diverse polychaete assemblage within Eunicida, comprising at least 33 genera, including Exallopus, Parougia, Pinniphitime, Pseudophryotrocha, and the speciose Ophryotrocha (Struck et al. 2006(Struck et al. , 2007. Dorvilleids occupy a diverse range of habitats and are often opportunistic infauna that are abundant in eutrophic and early-successional environments (Thornhill et al. 2009). These worms are also found in highly reduced and sulfidic extreme environments, including hydrothermal vents, whale-fall sediments, and cold methane seeps in the deep sea (Bernardino et al. 2010). Despite the group's diversity and abundance, only four phylogenetic studies have been conducted within dorvilleids (i.e., Pleijel and Eide 1996;Dahlgren et al. 2001;Heggøy et al. 2007;Wiklund et al. 2009), all of which focus on the numerous Ophryotrocha species from nonseep environments. Inferred relationships among shallow-water and whale-fall Ophryotrocha species were by and large congruent between molecular phylogenetic studies (Dahlgren et al. 2001;Heggøy et al. 2007;Wiklund et al. 2009). Conversely, Pleijel and Eide's (1996) morphological analysis suggested a markedly different ophryotrochan phylogeny (reviewed in Thornhill et al. 2009). Genetic data generally supported a gonochoristic labronica group and a second clade consisting of the hermaphroditic Ophryotrocha species (reviewed in Thornhill et al. 2009; see also Wiklund et al. 2009 where additional clades of hermaphroditic Ophryotrocha were reported). Furthermore, Heggøy et al. (2007) noted that Ophryotrocha was paraphyletic as Iphitime paguri fell within Ophryotrocha. None of these studies included species from hydrothermal-vent or methane-seep settings, where sulfide levels are higher and taxa are more ubiquitous.
Herein, we investigate the adaptive radiation of animals in extreme environments using methane-seep dorvilleid annelids in the northeast Pacific as a study system. The nature of dorvilleid diversification in these habitats provides insight into colonization of, and adaptation to, extreme environments. Specifically, if dorvilleids radiated only after moving into seep environments, adaptations required for life in extreme environments would be assumed to rarely evolve, because dorvilleids overcame physiological challenges of extreme environments only once during their evolutionary history. The single invasion of seeps by dorvilleids is our null hypothesis. By contrast, if dorvilleids radiated prior to (as well as after) colonizing seep habitats, the ability to adapt to such environments would be inferred to have occurred numerous times and with relative ease over evolutionary time. To determine evolutionary origins of methane seep dorvellids, we examined 16S rRNA (16S) and cytochrome b (Cyt b) mitochondrial gene sequence data. We also examined dorvilleid diversification and the coexistence of multiple species in relation to substrate depth, habitat type, food source data as inferred by d 13 C, and geographic location.

Sample collection
Dorvilleid annelids were collected from depths of 590-900 m on the northern California continental slope off shore of the Eel River mouth and on the Oregon margin at Hydrate Ridge (Table 1) or scoop bags using the remotely operated vehicle Tiburon (July 2005) or deep-sea submersible vehicle Alvin (July and October 2006). Methane-seep habitats sampled included vesicomyid-clam aggregations and microbial mats. Habitats, some with active venting of methane bubbles, were identified following Levin et al. (2003Levin et al. ( , 2010. Additional habitats, including tube fields and carbonate deposits, were also sampled when available. Once returned to the surface, samples were stored at 5-6°C (ambient bottom-water temperature) and processed immediately by sectioning the tube cores vertically at 0-1, 1-2, 2-5, and 5-10 cm depths. Fauna was subsequently sorted following Levin et al. (2003). Photographs depicting a representative sub-set of NE Pacific margin dorvilleid diversity are provided in Figure 1. Morphologically identified dorvilleids were either (1) frozen or preserved in 85% ethanol for molecular analyses (n = 131), (2) frozen for analyses of stable isotopic signatures, or (3) preserved in formalin as voucher specimens based on morphological assessments made on board the ship. Sediment position and habitat type were documented for each specimen. Because not all collected dorvilleid worms could be placed in a recognized species, some were assigned temporary epitaphs based on morphological and molecular characterizations. Novel lineages mentioned here will be assigned official names as part of a larger ongoing effort in Dorvilleidae taxonomy.
Purified PCR products were bi-directionally sequenced using a Beckman CEQ 8000 Genetic Analysis System (Beckman Coulter, Brea, CA). Cyt b sequences were translated (Drosophila-mitochondrial code) into MacClade Version 4.06 (Maddison and Maddison 2000) to ensure that stop codons were not present. Each unique dorvilleid mitochondrial haplotype sequence was designated by a number for Cyt b, a lowercase letter for 16S, and alphanumeric combined name for the concatenated data (Table  S1).
For all analyses conducted herein, Cyt b and 16S data were examined both separately and as a concatenated dataset. Topologies were constructed under Bayesian inference (BI) using MrBayes Version 3.12 (Huelsenbeck and Ronquist 2001) implementing the Hasegawa-Kishino-Yano (HKY) + Γ (Cyt b) or General-Time-Reversible (GTR)+I+ Γ (16S and concatenated data) models of substitution, as suggested by the hierarchical Likelihood Ratio Test and the Akaike Information Criterion by MrModeltest V2 (Nylander 2004). For each analysis, two sets of four chains (three hot, one cold) were run for 2.0 9 10 6 generations and sampled every 100 generations. Due to convergence of chains within 1.2 9 10 5 (Cyt b), 9.0 9 10 4 (16S), and 1.5 9 10 5 generations (concatenated), the first 1,200 (Cyt b), 900 (16S), or 1,500 (concatenated) trees were discarded as burn-in, and a 50% majority-rule consensus tree was calculated from remaining trees. Posterior probabilities (PP) were recorded to assess reliability of recovered nodes.

Isotopic diet analyses
Based on the morphological and molecular identifications of dorvilleid specimens, 4-62 individuals of each clade were analyzed for tissue d 13 C. Specimens were rinsed in MilliQ water, dried, powdered, and homogenized (when necessary), placed in tin boats, and acidified with 10% PtCl 2 to remove carbonate. Specimens were analyzed on a Costech elemental analyzer with a "zero-blank" autosampler interfaced with a continuous-flow Micromass Isoprime isotope-ratio mass spectrometer at Washington State University or on a Finnigan Conflow 2 continuousflow system and a Fisons NA 1500 elemental analyzer coupled to a Finnegan Delta S isotope-ratio mass spectrometer at Boston University. Isotope ratios are expressed as d 13 C in per mil units (&). Standards for 13 C were PeeDee Belemnite.

Results
The molecular data set consisted of 131 total methaneseep dorvilleid samples (n = 130 Cyt b, n = 128 16S, n = 127 samples sequenced for both genes). These were grouped into 41 unique haplotypes for Cyt b and 18 haplotypes for 16S. The number of representatives per haplotype ranged from 1 to 24 for Cyt b and from 1 to 38 for 16S (Table S1,  . When these data were concatenated into a combined dataset, a total of 43 unique haplotypes were encountered. This higher number relative to the individual genes reflects the fact that the unambiguously aligned region of 16S is more conserved than Cyt b (Mueller 2006); many samples of the same 16S haplotype exhibited Cyt b nucleotide differences. Concatenation was not possible for the "Pinniphitime Seep" and "Pseudophryotrocha Seep" samples, which were successfully sequenced for only one gene each (Table S1). The number of representatives per haplotype ranged from 1 to 23 for the concatenated data.
Topologies of methane seep dorvilleids, as estimated by BI and ML, are shown in Figs. 2-4. Overall, the two mitochondrial genes yielded similar estimations of evolutionary relationships among taxa. However, the Cyt b topology exhibited relatively longer branch lengths (reflecting more substitutions per site in this gene between taxa) than was observed for 16S. Specific patterns within the three topologies are highlighted below.

16S ribosomal mtDNA
Similar to Cyt b results, approximately 11 lineages of methane-seep dorvilleids were detected in the 16S dataset (Fig. 2). These terminal clades were either well supported or represented by only a single haplotype in the dataset. Notably, several clades that were differentiated by Cyt b were only distinguished by short branch lengths in the 16S topology (i.e., P. oregonensis Clades 1 and 2; Parougia Clades CA and OR; Fig. 3). In most cases, similar relation-ships between clades were inferred based on 16S versus Cyt b. Within the "Mixed Genera" group, associations were still observed between Ophryotrocha clades (Seep 3, 4, and 5) and specimens from other genera, including Exallopus and Pseudophryotrocha (PP = 1.00; BP = 100). The BI and ML topologies were incongruent only in the placement of the latter two genera within "Mixed Genera" (Fig. 3). In contrast to the affiliation between the "Parougia" and "Mixed Genera" groups in the Cyt b topology, the "Ophryotrocha" and "Mixed Genera" groups were sister to one another, but poorly supported (PP = 0.76; BP >50).
Addition of 16S data from nonseep Ophryotrocha species (Dahlgren et al. 2001) provided additional insight into the evolution of methane-seep dorvilleids. Three major ophryotrochan groups were recovered including a gonochoristic labronica group (PP = 0.98; BP = 91), a second group of hermaphroditic species plus two Ophryotrocha clades from seeps (PP = 0.80; BP <50), and "Mixed Genera" as the third group. Moreover, Ophryotrocha Seep  Table S1) and number of replicates (designated as 'n=') are provided. For described species, the species name is provided to the right of the phylogeny. Undescribed species are each labeled by their putative genus (identified based on morphological characters) and a tentative cladal designation (e.g., Ophryotrocha Seep 1). Major groupings on the phylogeny (i.e. Ophryotrocha, Mixed Genera, and Parougia) are also labeled.

Concatenated mtDNA genes
With the exception of the missing samples representing the Pinniphitime (Cyt b), Pseudophryotrocha (16S), nonseep Ophryotrocha (16S), and P. albomaculata (16S) lineages, the topology based on concatenated data was highly consistent with topologies produced by individual genes (Figs. 2-4). Furthermore, results of BI and ML analyses were congruent. Therefore, we present only the ML tree (Fig. 4). Within three major groups, "Ophryotrocha," "Mixed Genera," and "Parougia," approximately 10 terminal clades of methane seep dorvilleids were detected. Support values for these groups were high for 7 of 10 lineages, with posterior probabilities and bootstrap values above 0.95 and 95, respectively. The remaining three cladesincluding Ophroytrocha Seep 4, Parougia Seep Clade CA and Clade ORhad moderate-to-high support, with posterior probabilities ! 0.85 and bootstrap values ! 77.

Topology testing
We tested alternative hypotheses that were not recovered by the best tree using the AU test. Monophyly of a group comprising all Ophryotrocha taxa and no other taxa were significantly rejected by all three datasets (P 0.010) ( Table 3). For the 16S dataset, monophyly of Ophryotrocha Seep taxa was also significantly rejected (P < 0.001). Additionally, monophyly of Ophryotrocha Seep 1 and 2 resembling O. maciolekae is significantly different from the best tree in the 16S dataset (P < 0.001). The other two datasets are not appropriate to test this hypothesis due to the lack of hermaphroditic ophryotrochans. In contrast, monophyly of Ophryotrocha Seep 3-5 resembling O. platykephale to the exclusion of Pinniphitime Seep in the Cyt b dataset cannot be rejected (P = 0.229). This monophyly was given in the 16S and concatenated datasets, but these datasets lacked Pinniphitime sp. Finally, the three datasets recover different placements of the "Mixed Genera" group. 16S analyses grouped this clade with "Ophryotrocha," whereas other analyses grouped it with "Parougia." However, no dataset was able to reject the alternative scenarios for the placement of this "mixed" group (Table 3).  (Table 2). Similarly, several clades resembling Ophryotrocha displayed limited distributions (Table 2). Ophryotrocha Seep 1, 2, 4, and 5 were found solely at Eel River. Conversely, Ophryotrocha Seep 3 occurred only at Hydrate Ridge East. However, the sample size is low for certain clades and more exhaustive sampling could uncover broader distributions. Within each methane seep, several different habitat types were observed, including clam beds, bacterial mats, ampharetid-tube fields, and carbonate deposits. Although virtually no oxygen penetrated into bacterial-mat sediments, clam-bed sediments were penetrated by oxygen in the first few millimeters (Levin et al. 2003). Sulfide concentration also varied by location and habitat type. The sulfide concentrations were highest in bacterial mats at Hydrate Ridge (Sahling et al. 2002;Levin et al. 2003). By comparison, clam beds at Hydrate Ridge and bacterial mats at Eel River exhibited approximately one order of magnitude lower sulfide concentrations (Sahling et al. 2002;Levin et al. 2003;Ziebis and Haese 2005). The lowest sulfide levels occurred in Eel River clam beds (Sahling et al. 2002;Levin et al. 2003;Ziebis and Haese 2005).
Despite occurrence of different habitats, there was little absolute partitioning of dorvilleid clades by habitat ( Table 2). Note that ampharetid-tube fields were poorly sampled and less common than other habitats at these seeps. Some seep dorvilleids are commonly found in several different habitats and therefore appear to be methane-seep habitat generalists (e.g., Exallopus Seep, Pseudophryotrocha Seep, Pinniphitime Seep). Nevertheless, certain clades/species were more abundant in one habitat. Specifically, P. oregonensis and Parougia Seep OR were most abundant in clam beds relative to other habitat types. By contrast, Parougia Seep CA and all Ophryotrocha clades were dominant in the bacterial-mat habitats of Eel River or Hydrate Ridge East, respectively. Whether these differences represent actual habitat affiliations, as opposed to differences between sites, differential sulfide tolerance between taxa, or geographic partitioning of these species, remains to be determined.
At finer spatial scales, most dorvilleid clades were concentrated in uppermost sediment layers at methane seeps (approximately 0.96-1.75 cm; Table 2). A notable  Table S1) and number of replicates (designated as 'n=') are provided. For described methane seep species, the species name is provided to the right of the phylogeny. Undescribed seep species are each labeled by their putative genus (identified based on morphological characters) and a tentative cladal designation (e.g., Ophryotrocha Seep 1). Nonseep species are labeled following Dahlgren et al. (2001). Major groupings on the phylogeny are also labeled.
ª 2012 The Authors. Published by Blackwell Publishing Ltd.
exception was Exallopus Seep, which exhibited a broader sediment-depth distribution. Exallopus Seep was found at 10-cm depth, with individuals being most abundant at approximately 4-5 cm below the sediment surface (Table 2). These worms also had high sulfide tolerance (worms occurred at sulfide concentrations >10 mmol/L; data not shown).
Finally, partitioning among lineages is possibly driven by food sources. Thus, diets of methane-seep dorvilleids were inferred via measurement of carbon stable isotope ratios. For d 13 C, values near À20& reflect photosynthetic food sources, whereas much lighter values reflect chemosynthetic food sources. Values between À25 and À40& probably indicate carbon fixed by sulfur oxidation and values of approximately À45& and below reflect methane-derived carbon (Fisher 1990;Summons et al. 1998;Van Dover 2000;Levin and Michener 2002). Based on these considerations, Parougia Seep CA had d 13 C values indicative of photosynthetically derived carbon (Fig. 5). Its congeners had lower d 13 C values; Parougia Seep OR Table 3. Results of topology testing using the AU test of different alternative hypotheses not recovered by the best tree for the three datasets. Significant values (P < 0.05) are in bold.

Hypothesis
Cyt b 16S Concatenated Monophyly of "Mixed Genera" and "Parougia" Not applicable = Recovered by best tree and dataset is not appropriate due to lack of hermaphroditic "Ophryotrocha." 3 Not applicable = the same as the hypothesis "Monophyly of Ophryotrocha. . Alphanumeric names (designated by letters corresponding to Cyt b haplotype and numbers corresponding to 16S haplotype) and number of replicates (designated as 'n = ') are provided for each haplotype. For described species, the species name is provided to the right of the phylogeny. Undescribed species are each labeled by their putative genus (identified based on morphological characters) and a tentative cladal designation (e.g., Ophryotrocha Seep 1). Major groupings on the phylogeny (i.e., Ophryotrocha, Mixed Genera, and Parougia) are also labeled.
appeared to derive its carbon from sulfur oxidation and P. oregonensis had values consistent with methane as a carbon source (Fig. 5). The Exallopus clade also had low d 13 C values intermediate between methane-derived and sulfur-oxidation-derived carbon (but note the deeper sediment distribution (

Discussion
Diversity of dorvilleids at methane seeps Cold methane seeps off of the U.S. Pacific Northwest host highly diverse dorvilleid assemblages, consisting of at least 12 mtDNA species (terminal clades) in 5 different genera. Although two of these clades represented a known nominal species (i.e., P. oregonensis), most of the dorvilleid lineages reported here are new to science. By comparison, the mtDNA sequence divergences observed in this study are equivalent to or greater than the genetic distances reported for different shallow-water dorvilleid species by Dahlgren et al. (2001). Therefore, assuming consistent rates of mtDNA evolution across dorvilleids, each of these methane seep clades probably represents separate and distinct species. In many marine settings, only a single dorvilleid species is present (reviewed in Thornhill et al. 2009). Despite this, instances of multiple co-occurring species have occasionally been previously observed. Smith and Baco (2003) report finding 45 different dorvilleid species on whale falls on the California margin. Wiklund et al. (2009) (Prevedelli et al. 2005). Such cases provide precedents for diverse dorvilleid communities in sulfidic environments. Methane seeps along the NE margin host highly diverse dorvilleid communities, with at least 12 putative sympatric species.
Formal description of the new methane seep taxa is part of a larger project and will be the subject of future reports. However, the lack of morphological variation between certain lineages has stymied traditional taxonomic approaches. Clades, such as Ophryotrocha Seep 1 versus Seep 2, Ophryotrocha Seep 3 versus 4 versus 5, and Parougia Seep CA versus OR were indistinguishable morphologically during shipboard sorting (unpub. data), yet these taxa were well differentiated on both the Cyt b and (to a lesser degree) 16S phylogenies. Such examples of putative cryptic speciation may be common among dorvilleids, including species of Ophryotrocha. For instance, although many Ophryotrocha lineages examined by Dahlgren et al. (2001) are morphologically similar, breeding experiments attempting to cross hybridize these different lineages have failed to yield viable offspring, suggesting that these taxa were reproductively isolated and were therefore different species according to the biological species concept (Å kesson 1978).

Establishment and evolution of dorvilleids at methane seeps
Based on the proposed 16S phylogeny, including methane-seep and nonseep dorvilleids, the ability to inhabit seeps appears to have evolved independently four or more times in this annelid group. Seep Ophryotrocha and other taxa fall within the larger phylogeny of nonseep dorvilleids (this paper, Struck et al. 2006;Eibye-Jacobsen and Kristensen 1994). This broader phylogenetic perspective indicates that the ancestor of this clade was likely a nonseep dwelling organism and colonization of seeps occurred multiple times during dorvilleid evolution. For instance, Ophryotrocha Seep 1 and Seep 2 are intermingled with various hermaphroditic nonseep Ophryotrocha. This phylogenetic position also suggests that the reproductive mode of Ophryotrocha Seep 1 and 2 is simultaneous hermaphrodism; however, reproductive mode has not been determined for any of the seep species discussed here. The intermingling of seep and nonseep Ophryotrocha indicates that dorvilleids either colonized seep environments multiple times in independent events or have moved in and out of seep environments throughout evolutionary history. Determination of the number of instances where dorvilleid species moved from nonseep habitats into cold seeps requires more exhaustive sampling.
Abundance and diversity of methane-seep dorvilleids suggest that some dorvilleid taxa, such as Ophryotrocha and Parougia, are preadapted to life at seeps. Notably, some of the closest nonseep relatives to seep-dwelling Ophryotrocha, such as O. adherens and O. hartmanni, are able to survive in marginal, sulfidic, and/or organically enriched environments that are inhospitable to most metazoans (reviewed in Thornhill et al. 2009). Success at marginal and polluted habitats presumably includes mechanisms for detoxifying, tolerating, or avoiding toxic chemicals such as sulfides. Life at seeps presents similar physiological challenges to survival in polluted marine environments, including low levels of dissolved oxygen and high concentrations of hydrogen sulfide (see Introduction). As a result, the finding of intermingled seep and nonseep lineages within the 16S phylogeny fits within the context of ophryotrochan biology. Adaptation to life in marginal habitats may have preadapted certain Ophryotrocha spp. to colonize and succeed at methane seeps, as well as in sulfidic sediments from other environments (e.g., Smith et al. 1998;Mullineaux et al. 2003).
A primary underlying question in the diversification of dorvilleids at seeps is: how do so many confamilial taxa coexist in this ecosystem? Here, and in previous studies (Levin et al. 2003), it was hypothesized that stressful conditions (e.g., low dissolved oxygen, high-concentrations toxic sulfide) allowed dorvilleids to exploit an environment that was inhospitable to most taxa. Data presented here are consistent with the evolutionary-release hypothesis of Levin et al. (2003). At the hydrocarbon seeps of Hydrate Ridge and Eel River, a high genetic diversity and abundance of dorvilleids were encountered (Levin et al. 2003(Levin et al. , 2010this study). However, no single ecological factor definitively distinguished all species. Preliminary examinations of geographic, habitat, sediment-depth, and dietary differences between taxa suggested that, in many instances, dorvilleid clades were ecologically differentiated from one another through specialization on different resources. Recent iso-tope and fatty acid analyses of dorvilleds from Eel River and Hydrate Ridge support diet partitioning (Thurber et al. 2012;Levin et al. unpublished). Differences in geographic range (e.g., Parougia Seep CA vs. Seep OR), habitat affiliation, depth of sediment, sulfide tolerance (e.g., P. oregonensis and Exallopus Seep have similar diets, but different sediment distributions), and diet are hypothesized to reduce resource competition between taxa. Such niche partitioning within the environment allows for co-existence of ostensibly similar taxa. High dorvilleid abundance and diversity at whale falls (Smith and Baco 2003) and hydrothermal-vent sediments ) may also relate to release from competition and niche specialization. On the basis of the phylogenetic framework and ecological data presented here, more rigorous investigation of this hypothesis in future studies would be worthwhile.