Unveiling the unknown phylogenetic position of the scallop Austrochlamys natans and its implications for marine stewardship in the Magallanes Province

Two species of scallop, Austrochlamys natans (“Ostión del Sur”) and Zygochlamys patagonica (“Ostión patagonico”) are presently exploited in the southern part of the Magallanes Province (MP). The lack of clarity in taxonomic identification and ecological aspects is generating both erroneous extraction statistics and an unperceived harvesting pressure on A. natans and Z. patagonica. We aim to discriminate these Magallanes scallops accurately, improve our understanding of their complex natural history and discuss possible implications for their management and conservation status, given the current fisheries statistics. To achieve these goals, we present a complete review of the historical identification of the Magallanes scallop and a multi-locus molecular phylogeny which allowed us to recover the phylogenetic position of A. natans. We sampled 54 individuals from five localities across the southern Pacific coast of the MP. We calculated the depth of the byssal notch (BND) and shell height (VH) ratio from morphological characters and conducted phylogenetic reconstructions with mitochondrial (12S and 16S) and nuclear markers (28S) using Bayesian and maximum likelihood analyses. Both morphology and molecular phylogeny identified two distinct entities, Z. patagonica and a distinct, highly divergent lineage that corresponds to A. natans. Our study provides integrative evidence to alert the current fishery management and the need for further conservation studies.


Results
Identification and morphological analyses. The individuals identified in the five locations ( Fig. 1) were mostly adult specimens over 40 mm in length, except for individuals from Parker Island that ranged between 20 and 38 mm length. The individuals identified morphologically as Austrochlamys natans presented a moderately long shell, with an average shell height (VH) of 77 ± 9.75 mm. The shell coloration was quite variable, from white to dull brown or purple-brown to purple (Fig. 2a,e,i). The individuals identified as Zygochlamys patagonica presented a medium-sized shell with an average shell height (VH) of 41 ± 6.59 mm. The shell coloration was mainly white or pale and brick-red (Fig. 2b,c,d,f,h). Individuals of A. natans presented a greater depth of the byssal notch (BND), with an average depth of 5.60 ± 0.74 mm (Fig. 3) and an arcuate (Fig. 4B) to acute byssal notch (BN) (Fig. 4C,D). The individuals of Z. patagonica presented a lower BND (Fig. 3) and a more arcuate BN (Fig. 4F-H). The analysis of the BND/VH relationship discriminates two very different groups (Fig. 3), one composed of all specimens of Z. patagonica (red color), and the second with all specimens of A. natans (blue color). Individuals of Z. patagonica had an average BND/VH ratio of 0.05 ± 0.01, while those of A. natans had an average BND/VH ratio of 0.08 ± 0.01. This difference between the species was statistically significant (U = 2, p value = 0.0000).

Phylogenetic analysis of pectinids. The length of fragments amplified in Zygochlamys patagonica and
Austrochlamys natans for the 12S, 16S and 28S loci are indicated in Table S2. The levels of interspecific divergence (p-distance) between haplotypes of the species were rather high: 5.3% (28S), 13.1% (16S) and 18.7% (12S), reaching 10.8% for the concatenated sequences. The sequences of the specimens previously identified as Z. patagonica by Jonkers' (2003) method resulted highly similar to the sequences of Zygochlamys patagonica in Balmaceda Glacier (BG). All samples identified by Jonkers measurement as A. natans shared the same haplotype, new to Genbank and to scallop genetic diversity. No intraspecific variation was observed for either species at any locus in the Magallanes Region (between 50 and 56°S).  www.nature.com/scientificreports/ Due to the appearance of insertions/deletions and gaps during the alignment with the other 90 pectinid sequences and the 2 propeamussiids, the datasets consisted of 338, 492, 899 base pairs (bp) for 12S, 16S and 28S, respectively. Some ambiguously aligned hypervariable portions were removed from the 16S dataset. Single-locus trees ( Figure S3,S4,S5) were congruent and mainly differ in the low support of some nodes, as the 3 loci have their own resolutive power at different depth of the tree. The concatenation allowed a more substantial node resolution in the multi-loci tree. The concatenated sequence set had 1659 bp, 726 polymorphic sites and 550 parsimoniously informative sites. No substitution saturation was detected in the combined dataset.
The species included in the family Pectinidae cluster in a single clade, suggesting the monophyly of the family. This clade has full support in phylogenies reconstructed by both approaches. BI and ML gave congruent topologies, however ML generally gave lower bootstrap support (bs) than BI posterior probabilities (pp). The base of the pectinid tree is rather four-fold. Only the Chlamydinae appear paraphyletic among subfamilies; Palliolinae and Pectininae are monophyletic and highly supported (Fig. 5). Lineage H is a highly supported monophyletic group, www.nature.com/scientificreports/ composed of species of the tribe Chlamydini (Veprichlamys spp., Zygochlamys spp., Talochlamys gemmulata, T. dichroa) which is sister to the clade grouping Palliolinae-Pectininae, also very well supported. The Magallanes samples of Z. patagonica were tightly associated with its conspecific in a clade (pp: 0.99; bs: 89). The Magallanes specimen of Austrochlamys natans (Austrochlamydini) clustered with Adamussium colbecki (Adamussini) in a fully supported clade. Both monospecific tribes were nested with the tribe Palliolini in the subfamily Palliolinae (Fig. 5).

Discussion
This is the first comparative study of commercial scallop species in the Pacific coast of the MP combining morphological and molecular characters. Our phylogenetic analyses highlight the association between A. natans and Ad. colbecki; two members of monospecific tribes and last extant representatives of their Southern Oceanrestricted genera.  www.nature.com/scientificreports/ These results confirm the presence of both Magallanes scallops in the MP, as well as the so-far unsuspected presence of mixed "banks" where both species occur in sympatry. The BND/VH ratio helps discriminate between two distinct entities that belong to the genetic lineage of Z. patagonica and to a different lineage, highly divergent from the former, which corresponds to A. natans. A. natans is the only species of a whole lineage with a particular phylogenetic value, therefore having developed and tested an accurate identification criterion for both scallops will allow efficient fishery management in the future.
Here we discuss the phylogenetic position and the taxonomic status of both Magallanes scallops, as well as the implications of these results for the future management and conservation of Z. patagonica and A. natans in the Magallanes Region. Despite the numerous classifications built on morphological, ecological or molecular data, the relationships among pectinids are still under constant modification depending on the number of taxa, loci, length of the sequence and the selected outgroups 1,4 . The work of Alejandrino et al. 7 is the most inclusive so far in terms of taxon sampling, with 81 species. Although Scherrat et al. 25 included 143 species, the node supports of the phylogenetic trees are not provided, making it difficult to assess the robustness of this large phylogeny. In order to define the phylogenetic position of Zygochlamys patagonica and Austrochlamys natans, we included 93 pectinid taxa (43 genera) representative of tribes Chlamydini, Crassadomini, Fortipectini, Palliolini, Aequipectinini, Pectinini and Amussini. Comparing to Waller's 5 and Dijkstra's 15 classifications, only the subfamily Camptonectinae and the tribe Mesoplepini are missing. We used three ribosomal regions (one nuclear and two mitochondrial). Compared to Alejandrino et al. 7 , histone H3 is missing here, however this locus is among the least informative 4 . The family Pectinidae appears to be monophyletic with high support values (Fig. 5, S2), as previously demonstrated 4,7,[26][27][28] . According to Dijkstra 15 there are currently five subfamilies of Pectinidae, two of which are absent from our analysis: Camptonectinae and Pedinae. This topology supports the classifications of Waller 5 and Dijkstra 15 , except for the position of the tribe Austrochlamydini.
Our Magallanes scallops separated into two very divergent clades: Z. patagonica is associated with its conspecifics and congenerics in a single lineage (Fig. 5), which also contains species of Veprichlamys and Talochlamys. This lineage already appeared well supported as the sister clade to Palliolinae and Pectininae in Alejandrino 7 . For the first time, Talochlamys dichroa and T. gemmulata are nested with high support values into the Zygochlamys clade, making this latter genus paraphyletic (Fig. 5). These taxa are all restricted to high latitudes of the Southern Ocean. Due to phylogenetic and geographic affinities, we suggest that these three genera may constitute a tribe separate from Chlamydini. Since Dijkstra 15 moved the two Atlantic 'Crassadoma' into the genus Talochlamys, the affinities among Talochlamys spp. had not been explored until now. Talochlamys species rather associate according to geographic affinities, splitting the genus into two highly divergent entities corresponding to European and New Zealand Talochlamys. A systematic revision of these four species would be useful.
Austrochlamys natans associated with the Palliolinae, which was elevated to a subfamily rank by Waller 5 . Of the three extant tribes that compose this group, Mesopleplini are missing from our phylogenetic analyses. We included 4 genera (8 species) of the remaining two tribes: Adamussium (Adamussini) and Palliolum, Pseudamussium, Placopecten (Palliolini). The present sampling of Palliolini is the most inclusive to date and led to the monophyly and full support of the tribe Palliolini. Our phylogenetic results do not support any of the previous classifications of the tribe Austrochlamydini 1,5,9,13,15 , and introduce this monospecific tribe as a new member of the subfamily Palliolinae. Indeed, Austrochlamys natans clusters together with Adamussium colbecki, both in a sister clade to Palliolini. The first molecular characterization of Ad. colbecki did not lead to a clear classification due to the low polymorphism of the 18S 26 . Later, Ad. colbecki appears either as sister species to Chlamydinae or to Palliolini, depending on tribe sampling and the choice of outgroup and loci 4,10,11 . However, in the most recent and inclusive studies of taxon sampling 7 (present study) or genomic cover 29 , Ad. colbecki is the sister group of the tribe Palliolini, as in the present phylogeny.
The subfamily Palliolinae originated from a Chlamydinine ancestor in the Cretaceous and subsequently underwent diversification in the Northern Hemisphere 1 and in the Southern Hemisphere, where the extinct genus Lentipecten spread in the Paleocene-Eocene Thermal Maximum 30 . The genus Adamussium derived from Lentipecten and appeared in the early Oligocene; it comprises 5 endemic Antarctic species; Ad. colbecki is the only one extant 13,31,32 . The genus Austrochlamys also appeared in the Oligocene and was first restricted to King George Island (South Shetlands), then spread around the north of the Antarctic Peninsula and achieved a circum-Antarctic distribution until the Pliocene 13,33,34 . Austrochlamys persisted during the progressive cooling of the Antarctic Continent from the Paleocene to the Pliocene, dominating the coastal areas, while Adamussium occupied the deep seas and continental platform 33 . The opening and deepening of the Drake Passage and the intensification of the Antarctic Circumpolar Current during the Pliocene provoked a drastic cooling and the extension of sea ice over the coastal habitat, which caused the northward movement of Austrochlamys and its subsequent disappearance from Antarctica, along with the circumpolar expansion of Ad. colbecki in Antarctic shallow waters 33 . The colonization of the coastal habitat has been related to the sea ice extent that provided a more stable environment and low-energy fine-grained sediment with which Adamussium was associated in the deep waters. Austrochlamys fossils appear in the Subantarctic Heard Island in late Pliocene layers (3.62-2.5 Ma 35 ). Today Ad. colbecki is a circum-Antarctic and eurybathic species that reaches high local density in protected locations 13,36 , while all Austrochlamys became extinct except for A. natans, which is restricted to southern South America 33 . The phylogenetic affinity highlighted here between A. natans and Ad. colbecki has its origins in the Southern Ocean; the deep divergence between the lineages of these monospecific tribes attests to the long time since their common origin in the Paleogene. These results point out both species as relevant biogeographic models to address longstanding questions regarding the origin of marine biota from Southern Ocean.
The nomenclature, taxonomy and ecology of both A. natans and Z. patagonica have been problematic for almost 200 years. Since its original description 37 , Z. patagonica, a.k.a. the "Ostión Patagónico" has been named with more than 10 synonyms, probably due to the great intra-specific morphological variability throughout its distribution 19 www.nature.com/scientificreports/ records in the scientific literature and no genetic data on A. natans, a.k.a. the "Ostión del Sur" 13,14,17,19 , and some problems of nomenclature and establishing diagnostic characters persist since its description 13,39 . Many of the current junior synonyms of both species were described from small and juvenile specimens (under 52 mm VH [39][40][41]. Indeed, all deposited type material of A. natans ranges from 23.5 to 52 mm VH; the latter is half of the maximum size 39 . The criteria most commonly used for the identification of both scallops were number of radial primary ribs, maximum size, shell colour and presence of laminated concentric lines (Supplementary Table S1). Specimens with marked primary and secondary radial ribs alternated regularly and more whitish colouring of the right shell were attributed to Z. patagonica, while those with weaker and less markedly coloured radial ribs and the maximum size were considered as A. natans 42 . However, the number of radial ribs overlaps between Z. patagonica (26-42 12,43 ) and A. natans (22-50 17,19 ). These characters also have high variability across different environments and during ontogeny 13,17 . Thus the use of a taxonomy based on environment-sensitive and allometric characters has led to confusion in the morphological identification of these species 13,38 . The criterion used in the present study, the BND/VH ratio established by Jonkers 13 , discriminates the species efficiently. As attested by the narrow dispersal cluster in Fig. 3, this character has low intra-population variability 13 . In some cases a level of intraspecific variation can be detected, and this is mainly due to the environments where the scallop populations inhabit 19 (e.g. exposed, protected, substrate type, fjord, oceanic). However, although there may be some intraspecific variability between populations, this variability does not generate problems for the identification of the two species. Individuals of A. natans generally presented a significantly greater BND/VH ratio than those of Z. patagonica. However, it is important to consider that, given that this character varies during ontogeny, it is more accurate in individuals over 25 mm VH 13 . Only the molecular identification was able to discriminate juvenile scallops of both species accurately. According to the literature, A. natans is restricted to interior waters of channels and is associated with kelp forests of M. pyrifera (Supplementary Table S1). Z. patagonica inhabits a wider range of environments such as bottoms of shells, sand, mud and gravel in protected and exposed areas, between 2 and 300 m depth (Supplementary Table S1), but is also associated with kelp forests in fjords with different degrees of glacial retreat 12,16,44 . The juveniles of both scallops recruit in kelp forests 44,45 . According to the local artisanal fishermen, adults of "Ostión del Sur" (A. natans) occur in fjords with glaciers (orange circles in Fig. 1 23 ). We included two sampling locations near glaciers (in Pia and Montañas fjords), where large individuals (between 46 and 86 mm) of A. natans and Z. patagonica occur in sympatry. This sympatry was previously reported in Silva Palma Fjord between 5 and 25 m depth 16 . In conclusion, scallop banks are not monospecific but rather mixed and Z. patagonica occurs in the interior waters of the channels and fjords. Consequently, these two species have overlapping ecology (recruiting zone and glacial affinity) in the channels and fjords, overturning a long-held view that these scallops have marked habitat segregation. www.nature.com/scientificreports/ The fishery for both species was established in the 1990s in the political-administrative Region of Magallanes 16 , despite the complexity of the morphological recognition of scallops. The distinction between species was based on shell colour and radial ribs 42 , two characters that, given the results of this study, do not have this diagnostic capacity. Consequently, the scallop fisheries in the Magallanes Region are currently based on inaccurately discriminative characters. Scallop banks in MP have always been considered as monospecific 16,47 . A great part of scallop landing has always been attributed to A. natans 47 , about which the scientific literature is scarce (Supplementary Table S1). Conversely, Z. patagonica, which was erroneously considered as the commercial species of southern Chile, has more scientific research (Supplementary Table S1).
The difficulty to discriminate A. natans and Z. patagonica morphologically may lead to incorrect fishery statistics and uncertain conservation status of A. natans. Incorrect fishery statistics could overestimate the abundance of banks of A. natans compared to Z. patagonica. If the minimum catch size is reduced 23 in the context of the fishing overuse of the last decade, A. natans may suffer a reduction of its maximum size 48 . Therefore, an identification criterion between species is a need to improve fishery management. We showcased a quantified criterion that is useful to identify both species. In the short-term, this method can be used, but it is difficult to enforce in practical ways. We suggest to train fishing inspectors, following three guidelines. First, the identification should consider only the right valve (RV) for species identification, since the left valve is not taxonomically informative. Second, for visual classification, check the outline of the BN, mainly because the individuals of Z. patagonica have a more arcute BN. Third, a reliable identification has to measure the depth of the byssal notch (BND) and shell height (VH) ratio. Lastly, future research and fishery monitoring should follow these criteria to carry out a correct identification and subsequently better landings statistics.
Molecular tools allowed evaluating the phylogenetic relationships of scallops globally or regionally and incorporating parameters that can be used for the management and conservation of species of commercial interest 49 . For example, in the last few decades metrics have been developed to address conservation problems that give us a measure of the current state of particular taxa. These conservation priorities are often seen as measures for threatened species categorized by the IUCN Red List (World Conservation Union, 1980), one of the most widely and recognized systems. Although this prioritization metric incorporates phylogenetic distinctiveness (PD), this factor has been updated due to the importance of quantifying the loss of evolutionary diversity that would be implied by the extinction of a species 50 . The magnitude of the PD loss from any species will depend (but not exclusively) on the fate of its close relatives 51 . The "Ostión del Sur", Austrochlamys natans is the last representative of its tribe (Austrochlamydini) in the Southern Ocean. Its phylogenetic position and the long branch length (i.e. the length of the branch from the tip to where it joins the tree), which represents an important amount of evolutionary change, highlights the degree of isolation of A. natans and calls attention to the possible loss of a unique genetic lineage. There is currently no conservation value for this relict species; we sought to alert the current fishery management that the "Ostión del Sur" is a distinct taxon and provide integrative evidence for further conservation studies.
Finally, regarding the overlapping niche of these scallops and the conservation importance of the clade of A. natans, we propose three key recommendations for the future scallop fishery policies in the sub-Antarctic channels. First, it is necessary to assess the proportion of both species per bank and landing to generate a distribution map through the sub-Antarctic channels. For this assessment, the byssal notch depth is the most appropriate morphological character. Second, we recommend reassessments of biological and ecological parameters (e.g. size at first maturity) for A. natans across the glacial fjords, which are the most relevant fishing sites. As a final point, today there is a complete lack of knowledge of the genetic connectivity along the Subantarctic Channels. Thus we should generate more research about spatial population genetics at different temporal scales. The integration of genomic approaches (e.g. SNPs) with macro-and micro-environmental modelling approaches provide enormous opportunity to establish a new regional zoning for fishery management and conservation scallop strategy. Identification and morphological analyses. An exhaustive review of the morphological literature on both species was carried out to gather the best identification criteria (see Supplementary Table S1). The selected character described by Jonkers 13 is the relation between the depth of the byssal notch (BN) and the height of the right valve (RV). To measure these characters, the right valve of all collected specimens was photographed and the valve height (VH) and the depth of the byssal notch (BND) were measured (Fig. 4A). According to Jonkers 13 , individuals with a BN equal to or greater than 4.5 mm and a BND/VH ratio of approximately 0.08 were attributed to A. natans (Fig. 4B-D)  www.nature.com/scientificreports/ ratio of approximately 0.05 were attributed to Z. patagonica (Fig. 4E-H). The non-parametric Mann-Whitney U test was used to assess whether there are significant differences in the relationship (BND/VH) between species, using the R environment 52 .

Methods
Genetic analyses. Genomic DNA was extracted from muscle tissue samples of 50 pectinids using the salting-out method 53  Phylogenetic analyses. Phylogenetic relationships among Pectinidae were reconstructed using Bayesian Inference (BI) and maximum likelihood (ML) on each dataset of 12S, 16S, 28S, separately (data not shown) and on a combined sequence set (12S + 16S + 28S). Given that the use of more divergent outgroup taxa may affect pectinid relationships and decrease resolution 4 , the outgroup selected here includes Parvamussium pourtalesianum and Propeamussium sibogai, two species of the family Propeamussidae that was identified as the closest sister group to Pectinidae 7 . The concatenated dataset was aligned with corresponding sequences of 90 pectinid and 2 propeamussiid species retrieved from Genbank (see Supplementary Table S3). The substitution saturation of the concatenated data set was assessed with DAMBE 55 . The most appropriate model of evolution for each dataset was determined with the program Modeltest 3.7 56 . BI were conducted using the program MrBayes v.3.2.7 57 using Metropolis-coupled Markov Chain Monte Carlo (MC3) priors as follows: 10,000,000 generations, two independent runs, four chains, sampling trees every 1000 generations. The burnin period was identified by tracking the stability of highest likelihood values for each generation to determine whether they reached a plateau. ML analyses were conducted using PhyML 3.0 (http:// www. atgc-montp ellier. fr/ phyml/; 58 ); the program SMS 59 uses the Akaike Information Criterion (AIC) to select the best model of evolution. The analysis started with a BioNJ tree. Nearest neighbour interchange (NNI) was selected as heuristic search. Statistical support for each branch was obtained by 1000 bootstraps. Both SMS and Modeltest programs selected the 'generalised time-reversible' model with a gamma distribution and a proportion of invariant sites (GTR + Γ + I 60 ) as the best model of evolution of each of the three loci. 28S sequences that were lacking in Genbank for Pseudamussium clavatum, P. sulcatum and Semipallium amicum were replaced by missing data in the concatenated dataset. Gaps were treated as missing data. www.nature.com/scientificreports/