Marine Ecology. ISSN 0173-9565

ORIGINAL ARTICLE Hierarchical analysis of the population genetic structure in concholepas, a marine mollusk with a long-lived dispersive Leyla Cardenas 1, Juan Carlos Castilla2 &Fred erique Viard3,4

1 Instituto de Ciencias Ambientales y Evolutivas, Facultad de Ciencias, Universidad Austral de , Valdivia, Chile 2 Facultad de Ciencias Biologicas, Pontificia Universidad Catolica de Chile, Santiago, Chile 3 CNRS, team diversity and connectivity of coastal marine landscapes, Station Biologique de Roscoff, Roscoff, France 4 Sorbonne Universites, University Pierre and Marie Curie Paris 06, UMR 7144, Station Biologique de Roscoff, Roscoff, France

Keywords Abstract Concholepas concholepas; long larval dispersal; population genetics. For most marine species, dispersal is achieved mainly or exclu- sively by pelagic larvae. When the duration of the pelagic larval stage is long, Correspondence genetic homogeneity over large geographic scales is expected. However, genetic Leyla Cardenas, Instituto de Ciencias structure has often been reported over small spatial scales, suggesting that more Ambientales y Evolutivas, Facultad de complex processes occur than a simple positive relationship between pelagic Ciencias, Universidad Austral de Chile, larval duration and gene flow. Concholepas concholepas has a larval stage that Casilla 567, Valdivia, Chile. can last up to 3 months in the water column with a wide distributional range E-mail: [email protected] covering from 6°Sto56°S. We used a hierarchical sampling technique to test Accepted: 8 February 2015 if the genetic homogeneity of this highly dispersive species is maintained throughout its total geographic range in spite of environmental heterogeneity. doi: 10.1111/maec.12286 In the three studied regions (Antofagasta Bay, Valdivia and Patagonia), a spa- tial pattern of isolation by distance in conjunction with a spatial genetic struc- ture was observed. Within each region, different spatial genetic patterns were detected. In Antofagasta Bay and Valdivia there was evidence of substantial gene flow among populations, whereas in Patagonia, populations showed genetic structure and a unique, genetically isolated location was identified. These results revealed the existence of spatial differences in the genetic patterns among regions with different coastal topographies in C. concholepas, and give us new insights into the inter-relationships of larval dispersal potential, actual larval dispersal and physical processes. Regarding the sustainable management of C. concholepas, two important issues are derived from this study: (i) to high- light the need for a regional context in the management of C. concholepas, (ii) to determine the distinctiveness of the most austral population and to focus on the conservation efforts due to the relevance of this area.

ies have also provided evidence of complex patterns of Introduction genetic structure, with genetic heterogeneity over rela- Several meta-analyses in marine species have shown a tively small spatial scales (Tatarenkov & Avise 2010). For lack of correlation between dispersal potential (measured example, Becker et al. (2007) showed an unexpected level as the duration of the pelagic stage) and genetic structure of spatial heterogeneity in the connectivity patterns of estimates (often used as a proxy of gene flow and effec- species, which are assumed to be highly dispersive tive dispersal (Weersing & Toonen 2009). Empirical stud- organisms. Several hypotheses have been proposed to

Marine Ecology 37 (2016) 359–369 ª 2015 Blackwell Verlag GmbH 359 Spatial genetic structure in loco Cardenas, Castilla & Viard explain complex spatial genetic structures such as (i) larval duration of C. concholepas. The larvae can stay in demographic and reproductive features that may enhance the water column for more than 3 months (DiSalvo the effect of genetic drift at a local scale (Hedgecock 1988), facilitating larval exchanges over very large dis- 1994), (ii) limitations to estimating effective dispersal tances. However, genetic structure was observed between (e.g. Weersing & Toonen 2009) and (iii) local oceano- certain population’s pairs that were difficult to explain, graphic patterns (White et al. 2010; Selkoe et al. 2010). suggesting more complex and subtle structures operating For species located in coastal habitats, the topography of at a lower spatial scale. Unfortunately, Cardenas et al. the coastline may play a significant role in influencing (2009) study was carried out at a very broad scale and the local oceanographic properties and determining larval with a molecular marker of insufficient sensitivity to dispersal (Banks et al. 2007; Nicastro et al. 2008). Thus, reveal small-scale hydrodynamic and topographic effects. studies on population genetics should consider not only In fact, in many cases where complex patterns of genetic the characteristics of the life-history strategy of species structure were found in their study, the populations were (e.g. presence or absence of a larval stage), but also the located in estuaries, embayments and islands, sites that topographic and oceanographic conditions at local and may favor the disruption of gene flow (Arnaud-Haond intermediate spatial scales (Fernandez et al. 2014). Inte- et al. 2008). grative studies to understand spatial genetic diversity pat- Here, we analysed the spatial patterns in genetic diver- terns of marine species should include multiple sets of sity of a sample of 424 individuals of C. concholepas using populations located in different regions that display vari- six microsatellite loci (Cardenas et al. 2007). The aim was able topography and oceanographic conditions. to study the effects of variation in coastal morphology on The marine gastropod Concholepas concholepas (Bru- the genetic structure of this species by performing a hier- guiere, 1789), common name ‘loco’, is a bentho-pelagic archical design throughout the three main biogeographic species endemic to the Southeastern Pacific with a wide regions along the Chilean coast (Camus 2001): the Peru- distributional range from tropical (Lobos Afuera Island, vian province (PP), Intermediate Area (IA) and Magellan 6°S) to subantartic habitats (Cape Horn, 56°S). Conchole- province (MP). Due to the environmental heterogeneity pas concholepas is a keystone species of rocky inter-tidal as well as the diverse oceanographic patterns and coast- and subtidal communities (Castilla 1999; Manriquez et al. line morphology, we expected to find genetic differentia- 2008) and has high economic value, particularly for tion among regions. Then, in the PP we studied the bay small-scale fisheries and coastal communities (Bustamante of Antofagasta (at around 23°S), a semi-enclosed, south- & Castilla 1987; Leiva & Castilla 2001; Castilla et al. ward-facing bay that shows an upwelling shadow. Evi- 2007). The current approach to management of C. conc- dence from previous studies suggests that this bay may be holepas populations relies heavily on individually man- acting as a larval retention zone (Castilla et al. 2002; aged territorial-user-rights areas referred to as Lagos et al. 2008). We predicted significant genetic differ- ‘Management and Exploitation Areas for Benthic entiation among samples from locations inside and out- Resources’ (MEABR). In 2011, there were 535 MEABR side the bay. Second, in the IA, we examined the Valdivia along the Chilean coast, 392 of which included loco as a coast (at around 39°S), a mainly linearly exposed coast main target resource (Sernapesca, National Service of where upwelling centers exhibit strong seasonal variation Agriculture and Fisheries in Chile, 2012). The geographic (Escribano & Morales 2012) and where wind stress limits of each MEABR are defined by fishers, and initially appears to affect larval movement in C. concholepas coincided primarily with locations of high loco abun- (Moreno et al. 1993, 1998). Here, we hypothesized that dance (Gonzalez et al. 2006). In fact, a MEABR did not the populations within the area are quite well connected consider biological data of the target species to define its through larval transport and therefore exhibit genetic geographic delimitation; thus, it could be expected that a homogeneity among locations. Third, in the MP, we C. concholepas population will include several MEABRs. examined the Chilean Patagonia coast (45° to 56°S), a From a spatial point of view, a phylogeographic study highly fragmented area forming a large number of islands was carried out by Cardenas et al. (2009) using a mito- and protected channels (Ahumada et al. 2000; Ojeda chondrial gene to examine 14 spatially isolated popula- et al. 2000) with a well-defined pycnocline and strong tions covering approximately 4000 km of coastline and tidal flows (Pantoja et al. 2011) where the upwelling sys- three biogeographic regions with different topographic tem does not show any influence. Here, the complex configurations (such as bays and fjords). Their results coastal morphology dictates the exchange of waters showed that the majority of the genetic variation was between inshore coastal regions and the open , cre- explained at the local level and that the biogeographic ating micro-environments with unique oceanographic regions were not associated with phylogeographic breaks. conditions (Aracena et al. 2011), thus suggesting a per- The authors explained these results by the long planktonic manent larval retention zone (Hinojosa et al. 2011). Here,

360 Marine Ecology 37 (2016) 359–369 ª 2015 Blackwell Verlag GmbH Cardenas, Castilla & Viard Spatial genetic structure in loco we hypothesized that the populations are genetically dif- fragment analyzer following the protocols detailed in ferentiated. Cardenas et al. (2007). Although amplification patterns at the six loci were unambiguous, we checked the microsatellite data for Material and Methods evidence of null alleles and technical artefacts such as Sample collection stuttering and large allele dropout using the MICRO- CHECKER 2.2.3 software (Van Oosterhout et al. 2004). Several populations were sampled in each of the study The statistical independence between loci was then areas. In the Bay of Antofagasta (Fig. 1), adults were col- assessed using GENEPOP 4.0 (Rousset 2008). Genotypic lected from two locations inside (Jul, Tra) and four out- linkage disequilibrium between each pair of loci within side the bay, two locations to the north (Lob, Con) and populations and between each pair of loci over the whole two to the south (Bol, Trr). In Valdivia, (Fig. 1) samples data set was tested for using Fisher’s exact tests with a were collected from three sites (Meh, Lom and Puc) char- Markov chain. To adjust for multiple comparisons, Bon- acterized by exposed rocky shores. On the Chilean Pata- ferroni-adjusted P-values were examined (Rice 1989). gonia coast (Fig. 1), we collected two samples along the Tests for deviation from the genotypic proportions Jacaff channel (Gav and Cis) and two along the Aysen expected under HardyWeinberg equilibrium were car- fjord (Ays and Agui). Additionally, to test for spatial ried out using GENEPOP 4.0. The probability of identity structure inside Patagonia, a more exposed location at (PI), i.e. the probability that two individuals share the Wellington Island (at around 49°S) was also sampled same genotype, was calculated in GeneAlex v. 6 (Peakall (Ped). & Smouse 2006). Genetic diversity was analysed for each sampling loca- Microsatellite analysis tion by computing the mean number of alleles (Nall), genetic diversity (He) and the fixation index on an indi- Total DNA was extracted using a NucleoSpin Tissue Kit vidual relative to the subpopulation Fis (ƒ) using FSTAT according to the manufacturer’s protocol (Macherey- v. 2.9.3 (Goudet 2001). Allelic (Rall) and private allelic Nagel). All the individuals were genotyped at six loci: (Prall) richness among regions and within each location Cc1A2, Cc2A5, Cc2A11-1, Cc1H2, CcE5 and Cc2F5. Loci were estimated using a rarefaction method as imple- were amplified by PCR and analysed on a Li-Cor DNA mented in HP-rare (Kalinowski 2005). To test for regio-

Fig. 1. Sampling sites along the Chilean coast. Details and full names for each location are given in Table 1.

Marine Ecology 37 (2016) 359–369 ª 2015 Blackwell Verlag GmbH 361 Spatial genetic structure in loco Cardenas, Castilla & Viard nal differences in genetic diversity among Antofagasta set) to test for spatial genetic structure among regions Bay (n = 6 populations), Valdivia (n = 3) and Patagonia and also inside each region. Each run used 106 iterations (n = 5), we compared the allelic richness (Rall) and after a burn-in of 4 9 104 length using an admixture genetic diversity (He) through a permutation procedure model and the option of correlated allele frequencies as implemented in FSTAT (10,000 permutations were between populations and the individuals’ sampling loca- used). tions to inform cluster assignment, as this configuration Genetic structure was measured by estimating the fixa- was considered best by Hubisz et al. (2009) for cases with tion index among populations, Fst (Wright 1951) using lower levels of divergence or less data. To check for the the algorithm implemented in the software GENEPOP convergence of the Markov Chain Monte Carlo (MCMC) 4.0 based on Weir & Cockerham (1984) method. To test we performed 10 replicates for each value of K. The for regional and local spatial genetic structure, both glo- analysis defined the most probable K value with a peak bal Fst and pairwise Fst were estimated over the whole likelihood of the logarithm of K (LK); however, this value data set and within each region. Hierarchical analysis of varied among different runs of MCMC. To eliminate var- the genetic variance (analysis of molecular variance, AM- iation, the correction proposed by Evanno et al. (2005) OVA) was carried out using the software ARLEQUIN v. was used, which was implemented in the software 3.5.1.3 (Excoffier & Lischer 2010). Locations were clus- STRUCTURE HARVESTER 0.6.7 (Earl 2012). tered in three groups corresponding to the three regions: (i) Antofagasta, (ii) Valdivia and (iii) Patagonia. Esti- Results mates of the genetic structure among regions (Fct), among locations within regions (Fsc) and among loca- In total, 424 individuals of Concholepas concholepas were tions (Fst) were computed. Significance of the F indices genotyped with six microsatellite loci. Across all 14 locali- was assessed by the non-parametric permutational ties studied we found that none of the loci showed pref- method. erential amplification of short alleles or any evidence of Due to the fact that Concholepas concholepas has a dis- scoring errors or linkage disequilibrium. However, we tribution range extending over an almost linear coastline, identified that null alleles were potentially implicated in a pattern of isolation by distance (IBD) was used as the six locationloci combinations out of 84. Considering null hypothesis. To estimate the proportion of genetic these six combinations, the locus Cc2F5 displayed signifi- divergence along the Chilean coast as a function of geo- cant homozygote excess at five locations. Consequently, graphic distance between sampling locations, we analysed we carried out all of the analyses twice, i.e. including and the multilocus data set under a model of isolation by dis- excluding the Cc2F5 locus. tance (Rousset 1997). We calculated a matrix of geo- The average PI for each population is given in Table 1, graphic distances between sites by measuring the shortest when five loci were used the values of PI ranged from shoreline distances between each pairwise combination of 8.9E-09 in Con to 5.6E-07 in Ped. High genetic diversity locations. As a measure of genetic differentiation, we cal- (He) was found in every studied population (Table 1). culated the ratio of Fst/(1Fst) for each pairwise combi- Comparisons among regions showed no significant differ- nation of sites. To assess the significance of the IBD ences in either the level of genetic diversity (He = 0.85; model, we estimated the correlation between matrices of 0.84 and 0.85, for Antofagasta, Valdivia and Patagonia, genetic and geographic distances using a Mantel test respectively) or the number of alleles (Nall = 34.3; 26.3 implemented in GENEPOP. The significance of the and 36; for Antofagasta, Valdivia and Patagonia, respec- matrix correlation was evaluated using a permutation test tively). By contrast, significant differences were detected based on 10,000 randomizations. in allelic richness (Rall) among regions, with the lowest To test for spatial patterns of genetic structure without value in Patagonia (Rall: 13.9; 13.4 and 12.8; for Antofag- any prior assumptions, we used the approach imple- asta, Valdivia and Patagonia, respectively; P = 0.020). mented in STRUCTURE v. 2.3.4 (Pritchard et al. 2000). Comparison among populations showed that Rall ranged This program uses individual multilocus genotype data to from 18.8 in Trr (Antofagasta) to 13.2 in Ped (Patagonia) cluster individuals into K groups while minimizing while Rall showed the highest value (12.7) in Con (Anto- HardyWeinberg and disequilibrium between loci within fagasta) and the lowest (8.5) in Lom (Valdivia). The pri- groups. The estimation procedure consists of running vate allelic richness (Prall) was highest in Patagonia (4.8), trial values of the number of populations, K, and then with Ped showing the highest value (3.2) and Ays the comparing the estimated log probability of data under lowest (0.2). each K, Ln [Pr(X|K)]. We conducted a series of indepen- The genetic differentiation computed over all data sets 5 dent runs with different values of K (using a maximum was statistically significant (Fst = 0.035, P < 10 ). The equal to the total sampled locations available in each data AMOVA analysis (Table 2) revealed that the three regions

362 Marine Ecology 37 (2016) 359–369 ª 2015 Blackwell Verlag GmbH Cardenas, Castilla & Viard Spatial genetic structure in loco

Table 1. Sampled locations and average genetic diversity over loci. The values for the fixation index FIS were calculated in 14 location sampled to Concholepas concholepas using six or five loci are shown (Weir & Cockerham 1984).

latitude longitude region location ID (S) (W) n nall rall prall He SD PI fis (6 loci) fis (5 loci)

Antofagasta Loberıas Lob 23°500 70°250 24 14.3 13.5 0.4 0.82 0.10 2.3E–06 0.028* –0.003* Constitucion Con 23°240 70°350 36 17.5 14.6 0.9 0.82 0.12 9.9E–08 –0.032* –0.050 Juan Lopez Jul 23°310 70°320 34 16.8 13.3 0.5 0.83 0.08 1.4E–06 0.036* 0.0248 Triple A Tra 23°420 70°250 30 16.0 13.7 0.8 0.84 0.08 5.9E–07 –0.036 –0.033 Bolfin Bol 23°510 70°300 32 17.2 14.3 0.4 0.82 0.11 5.6E–07 –0.029 –0.029 Tres Reyes Trr 24°200 70°320 34 18.8 15.3 1.0 0.85 0.09 1.8E–06 –0.002 –0.025 Total 190 34.3 14.6 4.6 0.85 0.09 1.0E–08 –0.006 –0.018 Valdivia Mehuın Meh 39°250 73°120 28 15.5 14.1 1.0 0.84 0.11 2.8E–08 0.058 0.062 Los Molinos Lom 39°400 73°200 30 15.7 13.6 0.7 0.82 0.11 1.1E–07 –0.008* –0.059 Pucatrihue Puc 40°320 73°430 30 16.5 14.6 1.0 0.81 0.10 7.3E–08 0.087** 0.092 Total 88 26.3 15.0 4.5 0.84 0.11 1.5E–08 0.45 0.03 Patagonia Puerto Gaviota Gav 44° 50 73°200 29 15.5 13.1 1.4 0.83 0.08 1.2E–07 0.048** –0.023 Puerto Cisnes Cis 44°200 72°570 34 16.7 13.4 0.7 0.82 0.09 6.2E–08 0.025** –0.003 Puerto Aysen Ays 45°160 73°170 28 14.2 12.0 0.2 0.79 0.12 4.8E–07 –0.017 –0.041 Puerto Aguirre Agui 45°080 73°300 30 15.0 13.0 0.3 0.82 0.10 8.0E–08 0.045** 0.016* Puerto Eden Ped 49°080 74°270 25 13.2 10.4 3.2 0.82 0.09 5.6E–07 0.132** 0.128** Total 146 36 15.2 4.8 0.85 0.11 1.6E–08 0.042 0.014 n, number of individuals per locality; Nall, number of alleles per locality; Rall, allelic richness; Prall, private allelic richness; He, genetic diversity (SD); Pi, probability of identity. Exact test for deviation from HardyWeinberg equilibrium: *P < 0.05, **P < 0.005.

(i.e. Antofagasta, Valdivia and Patagonia) are genetically and Patagonia except for individuals from Ped, which differentiated (Fct = 0.036, P < 0.0001). The genetic formed an independent genetic cluster (Fig. 3A). The differentiation within regions was also significant Bayesian analysis also revealed spatial genetic heterogene-

(Fsc = 0.011, P < 0.0001) and a similar result was ity when the analysis was carried out within each region observed among locations (Fst = 0.047, P < 0.0001). Out (Fig. 3B–D). In Antofagasta, two clusters (K = 2, LnP of the 90 pairwise Fst tests performed, only four were (K) = 4841.4) were identified, with Lob and Trr found to be non-significant at a 5% nominal level; after belonging to different clusters, and with evidence of Bonferroni correction this number increased to 14 admixture between the two groups based on the assign- (Table 3). Within each region, different levels of genetic ment of individuals to clusters (Fig. 3B). In Valdivia, two differentiation were detected. In Antofagasta, Tra and clusters (K = 2, LnP(K) = 2233.5) were identified and

Lob showed a pairwise Fst of 0.01 (P = 0.02) and Tra and no evidence of any spatial structure across locations was Bol a pairwise Fst of 0.01 (P = 0.003). Similarly, the pair- observed (Fig. 3C). In Patagonia, three clusters were wise Fst for the comparison between Jul and Trr was sig- identified (K = 3, LnP(K) = 3587.8), with Ped remain- nificant (Fst = 0.008, P = 0.035). In Valdivia, genetic ing as an independent genetic group and with evidence differentiation was significant for the comparisons of admixture between the other two clusters among the 5 MehLom (Fst = 0.005, P < 10 ) and MehPuc locations. Interestingly, individuals from Ays were mainly (Fst = 0.006, P = 0.0001). In Patagonia, all pairwise Fst assigned to one cluster, suggesting asymmetric admixture comparisons were significant after Bonferroni correction in the region (Fig. 3D). (Table 3). A strong and significant IBD pattern was found (r = 0.77, P < 10 5), showing that there is a Discussion decrease in the genetic similarity among populations as geographic distance increases (Fig. 2). Distinct genetic groups of the marine gastropod Conc- The STRUCTURE analysis provided consistent results holepas concholepas and a clear pattern of isolation by dis- irrespective of the model used (Fig. 3). At the regional tance were identified in this study on the basis of a scale, three well-differentiated genetic clusters were hierarchical analysis covering more than 3000 km of the revealed (K = 3, LnP(K) = 10743.7). The first cluster Chilean coast. These results suggest that population grouped together exclusively individuals from Antofagas- boundaries and genetic exchange in this species may be ta, the second cluster grouped individuals from Valdivia more restricted than previously supposed, thus contrast-

Marine Ecology 37 (2016) 359–369 ª 2015 Blackwell Verlag GmbH 363 Spatial genetic structure in loco Cardenas, Castilla & Viard

ing with the simple idea that species with an intermediate larval stage and long residence in the plankton show high connectivity and genetic homogeneity. Gene flow among populations is the ultimate conse- quence of a series of processes that influence the effective dispersal of propagules in natural environments (Cowen & Sponaugle 2009). In marine , larval trans- port is expected to play a prominent role in facilitating gene flow and determining population structure, including over large spatial scales, in species with a long dispersive stage. Here, we detected evidence of a distance- dependent pattern of dispersal (IBD pattern). This pattern could be explained by stepping-stone migration models and it can be expected in a ‘continuous’ population dis- tributed in a one-dimensional habitat such as the South- eastern Pacific coast. Interestingly, this result is in accordance with the recent findings of Garavelli et al. (2014) who, based on a biophysical individual-based larval dispersal model, reported a mean dispersal distance for C. concholepas ranging from 170 to 220 km, depending upon the depth of larvae release and planktonic larval duration. Several authors have suggested that the IBD slope could be considered in assessing the spacing of mar- ine protected areas (Durrant et al. 2014). In this context, the IBD distance could be used as a tool to determine the geographic limits of a regional unit that allows MEARB to unify efforts for the conservation of C. concholepas. Moreover, northsouth gradients in reproduction pat- tern and dispersal may be important in the connectivity of C. concholepas. First, Fernandez et al. (2007) showed values between populations within regions. that the number of embryos of C. concholepas per capsule st . Details of the sampled locations are given in Fig. 1 and Table 1. Significant values after 1000 permutations test was lower in Northern than in Southern Chile. Similarly, planktonic larval duration (PLD) has been found to vary with latitude. In laboratory studies, Disalvo (1988) deter- mined the PLD to be around 2–3 months whereas in field studies it was determined to be between 3 and 4 months in South-central Chile (Moreno et al. 1993) and up to 6–12 months in Patagonia (Molinet et al. Concholepas concholepas 2005). In addition, the finding of significant genetic differenti- ation among regions and locations (Tables 2, 3 and Fig. 3) supports the hypothesis that environmental heter- 0.0006). The boxes indicate the pairwise F ogeneity affects the effectiveness of larval dispersal by =

a generating a heterogeneous connectivity pattern. A recent study in C. concholepas using larval microchemistry as a tool to identify natal sources studied larvae and recruit

analysis among 14 locations of individuals from three distant areas (including Antofagas- st ta and Valdivia), finding evidence of spatial segregation among areas and reduced inter-change of larvae. Accord- ing to these authors, these differences in elemental signa- Pairwise F tures might be influenced by factors such as different ambient elemental concentrations, upwelling regimes, are in bold (Bonferroni correction Table 2. nutrient-rich rainwater runoff and temperature (Zacherl

364 Marine Ecology 37 (2016) 359–369 ª 2015 Blackwell Verlag GmbH Cardenas, Castilla & Viard Spatial genetic structure in loco

Table 3. Hierarchical analysis of genetic variance (analysis of molecular variance) of three regions (Antofagasta, Valdivia and Patagonia) and 14 locations of Concholepas concholepas.

five loci six loci

sum of variance percentage sum of variance percentage source of variation squares components of variation squares components of variation

Among regions 45.16 0.071 3.29 60.18 0.096 3.63 Among pop/within 36.93 0.022 1.00 46.73 0.029 1.09 region Within pop 1505.74 2.06 95.71 2080.24 2.51 95.28 Total 1787.83 2.16 2187.15 2.64

five loci six loci

Fct 0.033 (<0.0001) 0.036 (<0.0001)

Fsc 0.010 (<0.0001) 0.011 (<0.0001)

Fst 0.042 (<0.0001) 0.047 (<0.0001)

Fig. 2. Relationship between genetic distance (Fst/1Fst) and geographic distance (km) (r = 0.77, P < 105)inConcholepas concholepas.

2005; Warner et al. 2009). These results highlight the larvae are capable of moving regardless of the local need to consider the regional context in the management coastal flow movements and thermal temperature front in of C. concholepas, suggesting that regional management the bay (Avendano~ & Cantillanez 1997; Castilla et al. units should be implemented in order to co-ordinate the 2002). Lob and Trr are the most isolated locations specific efforts of MEABR in each area. (141 km apart) within this region, and given the spatial In the Antofagasta region the tested hypothesis (i.e. sig- pattern of isolation by distance detected here, it may be nificant genetic structure among locations outside and possible that the geographic distance that the larvae can inside the bay) was not supported by our results. The move in this area would be not higher than the distance pairwise Fst analysis showed genetic differences among between these populations. populations inside and outside the bay (Tra was different In Valdivia, the tested hypothesis (i.e. connectivity and from Lob and Bol, and Jul was different from Trr). How- genetic homogeneity) was supported by the results. In the ever, the STRUCTURE analysis did not reveal a similar STRUCTURE analysis, when the whole data set was used pattern, instead showing that the most probable number to generate the regional spatial pattern (Fig. 3A ), they of genetic groups in the area is K = 2, with Lob and Trr appear to be connected with populations from Patagonia. belonging to different clusters and evidence of admixture However, when the analysis was performed using only between the two groups. This suggests that C. concholepas regional data, the most probable genetic clusters were

Marine Ecology 37 (2016) 359–369 ª 2015 Blackwell Verlag GmbH 365 Spatial genetic structure in loco Cardenas, Castilla & Viard

A

BCD

Fig. 3. Clustering analysis based on (A): the complete data set, (B): Antofagasta Bay, (C): Valdivia and (D): Patagonia. A thin vertical line divided into K colored segments represents the individual’s estimated membership fractions. See Table 1 for full names of each location.

K = 2 and a total admixture pattern among locations was presence of the Pena Gulf (Aracena et al. 2011). From a observed. This genetic pattern could be related to the genetic point of view, Patagonia remains largely unex- presence of the West Wind Drift (WWD) current that plored. There is a great need for more studies in order to runs against the Chilean coast and splits at approximately compare the present results with other marine inverte- 43°S into an equator-ward Peru (or Humboldt) Current brate species. For now, based on the results of this study, and a pole-ward Cape Horn Current (Strub et al. 1998; Ped can be considered a genetically unique population of Acha et al. 2004). The WWD has an enormous influence C. concholepas, thus suggesting that this population on the geographic distribution of biota in this area requires special treatment in order to conserve the diver- (Camus 2001). Additionally, previous work reported vari- sity of this species. ability in the recruitment of larva of C. concholepas asso- Although a genetic approach is an indirect way of ciated with the variation in the upwelling system near to determining larval dispersal, the present work is a step Valdivia (Moreno et al. 1998). Thus, we suggest that the towards understanding the connectivity of marine species C. concholepas population around Valdivia may constitute across the Southeastern Pacific coast. Here, we identified a dynamic entity located in the main biogeographic a clear IBD pattern with genetic groups of C. conchole- boundary region and is maintained by the WWD and its pas, suggesting that population ranges and genetic variability. Temporal studies will be necessary to sub- exchange could be more restricted than previously sup- stantiate this hypothesis and to test for the stability of posed. These findings have important implications (given gene flow in this area. that more than 86% of MEABRs in Chile have C. conc- In Patagonia, the tested hypothesis (i.e. populations holepas as their main target species Jerez & Figueroa genetically differentiated) was supported by the results 2009). For the sustainable management of C. concholepas: and genetic structure was detected among locations, in they highlight the necessity of (i) a regional context in particular when Ped, the most austral location, was con- the management of C. concholepas, and (ii) the determi- sidered. Here, salinity has been reported to affect C. conc- nation of the distinctiveness of the most austral popula- holepas larval distribution by acting as a barrier that tion (Ped) and the relevance of this area to focus the prevents competent larvae from reaching the surface conservation efforts. (Molinet et al. 2008). In addition, several studies have shown that variability in current flows is related to the Acknowledgements morphological characteristics of fjords, such as sill depth across the channel, variations in coastal morphology and We would like to thank the people who helped us to col- wind intensity (Caceres et al. 2002; Caceres 2004; Caceres lect samples along the Chilean coast: Andres Caro, Patri- et al. 2006). The most southern location in the present cio Manriquez, and especially the fishermen who helped study (Ped) differs from the others in terms of its exten- us in the remote region of Patagonia. This study was sive ice coverage, major rivers and, to the north, in the funded by grants awarded by ECOS Program no. C03B04

366 Marine Ecology 37 (2016) 359–369 ª 2015 Blackwell Verlag GmbH Cardenas, Castilla & Viard Spatial genetic structure in loco and FONDAP-FONDECYT 1501-0001 Program 6 to the Camus P.A. (2001) Biogeografıa marina de Chile continental. CASEB Centro de Estudios Avanzados en Ecologıa y Bio- Revista Chilena Historia Natural, 74, 587–617. diversidad (CASEB). Leyla Cardenas acknowledges the Cardenas L., Daguin C., Castilla J.C., Viard F. (2007) support of the Millennium Nucleus Center for the Study Isolation and characterization of eleven polymorphic of Multipledrivers on Marine Socio-Ecological Systems microsatellite markers for the marine gastropod (MUSELS) by MINECON Project NC120086. Concholepas concholepas (Brugiere, 1789). Molecular Ecology Notes, 7, 464–466. Cardenas L., Castilla J.C., Viard F. (2009) A phylogeographic References analysis across three biogeographic provinces of the Acha E.M., Mianzan H.W., Guerrero R.A., Favero M., Bava J. southeastern Pacific: the case of the marine gastropod 36 – (2004) Marine fronts at the continental shelves of austral Concholepas concholepas. Journal of Biogeography, , 969 South America: physical and ecological processes. Journal of 981. Marine systems, 44,83–105. Castilla J.C. (1999) Coastal marine communities: trends and Ahumada R.B., Pinto L.A., Camus P.A. (2000) The Chilean perspectives from human-exclusion experiments. Trends in 14 – coast. In: Sheppard C.R.C. (Ed.) Seas at the millennium: an Ecology & Evolution, , 280 283. ~ environmental evaluation. Regional Chapters, vol I. The Castilla J.C., Lagos N.A., Guinez R., Largier J.L. (2002) Americas and West Africa Europe, Pergamon, Amsterdam: Embayments and nearshore retention of plankton: The 699–717. Antofagasta Bay and other examples. In: Castilla J.C., Aracena C., Lange C.B., Iriarte J.L., Rebolledo L., Pantoja S. Largier J.L. (Eds.), The Oceanography and Ecology of the (2011) Latitudinal patterns of export production recorded Nearshore and Bays in Chile. Ediciones Universidad Catolica – in surface sediments of the Chilean Patagonian fjords (41– de Chile, Santiago, Chile: 179 203. 55°S) as a response to water column productivity. Castilla J.C., Gelcich S., Defeo O. (2007) Successes, lessons, and Continental Shelf Research, 31, 340–355. projections from experience in marine benthic invertebrate Arnaud-Haond S., Vonau V., Rouxel C., Bonhomme F., Prou artisanal fisheries in Chile. Fisheries Management: Progress J., Goyard E., Boudry P. (2008) Genetic structure at Toward Sustainability. Blackwell Press, London: 2542. different spatial scales in the pearl (Pinctada Cowen R.K., Sponaugle S. (2009) Larval dispersal and marine 1 margaritifera cumingii) in French Polynesian lagoons: population connectivity. Annual Review of Marine Science, , – beware of sampling strategy and genetic patchiness. Marine 443 466. Biology, 155, 147–157. Disalvo L.H. (1988) Observations on the larval and post- Avendano~ M., Cantillanez M. (1997) Necesidad de crear una metamorphic life of Concholepas concholepas (Bruguiere, 30 – reserva marina de ostiones en el banco de la rinconada 1789) in laboratory culture. The , , 358 368. (Antofagasta-II region, Chile). Estudios Oceanologicos, 16, Durrant H., Burridge C.P., Kelaher B.P., Barrett N.S., Edgar 109–113. G.J., Coleman M.A. (2014) Implications of macroalgal Banks S.C., Piggott L.M., Williamson J.E., Bove U., Holbrook isolation by distance for networks of marine protected areas. 28 – N.J., Beheregaray L.B. (2007) Oceanic variability and coastal Conservation Biology, , 438 445. topography shape genetic structure in a long-dispersing sea Earl D.A. (2012) STRUCTURE HARVESTER: a website and urchin. Ecology, 88, 3055–3064. program for visualizing STRUCTURE output and Becker B.J., Levin L.A., Fodrie J.F., McMillan P.A. (2007) implementing the Evanno method. Conservation Genetics 4 – Complex larval connectivity patterns among marine Resources, , 359 361. invertebrate populations. Proceedings of the National Escribano R., Morales C.E. (2012) Spatial and temporal scales Academy of Sciences, 104, 3267–3272. of variability in the coastal upwelling and coastal transition – ° Bustamante R.H., Castilla J.C. (1987) The shellfishery in Chile: zones off central-southern Chile (35 40 S). Progress in 92 – an analysis of 26 years of landings (1960-1985). Biologia Oceanography, ,1 7. Pesquera (Chile), 16,79–97. Evanno G., Regnaut S., Goudet J. (2005) Detecting the Caceres M. (2004) Surface features of the circulation in fjords number of clusters of individuals using the software 14 of southern Chile observed in ERS and LANDSAT images. STRUCTURE: a simulation study. Molecular Ecology, , – Gayana, 68,71–76. 2611 2620. Caceres M., Valle-Levinson A., Sepulveda H., Holderied K. Excoffier L., Lischer H.E. (2010) Arlequin suite ver 3.5: a new (2002) Transverse variability of flow and density series of programs to perform population genetics analyses 10 in a Chilean fjord. Continental Shelf Research, 22, under Linux and Windows. Molecular Ecology Resources, , – 1683–1698. 564 567. Caceres M., Valle-Levinson A., Molinet C., Castillo M., Bello Fernandez M., Calderon R., Cancino J.M., Jeno K. (2007) The M., Moreno C. (2006) Lateral variability of flow over a sill effect of temperature on the development of encapsulated in a channel of southern Chile. Ocean Dynamics, 56, 352– embryos of Concholepas concholepas along a latitudinal cline. 348 – 359. Marine Ecology Progress Series, , 229 237.

Marine Ecology 37 (2016) 359–369 ª 2015 Blackwell Verlag GmbH 367 Spatial genetic structure in loco Cardenas, Castilla & Viard

Fernandez R., Lemer S., McIntyre E., Giribet G. (2014) Molinet C., Niklitschek E., Moreno C.A., Arevalo A. (2008) Comparative phylogeography and population genetic Vertical distribution of early and competent larvae of structure of three widespread mollusc species in the Concholepas concholepas in two systems of Chilean inland Mediterranean and near Atlantic. Marine Ecology, 36, seas. Marine Biology, 153, 779–788. 701–715 Moreno C.A., Asencio G., Ibanez~ S. (1993) Patrones de Garavelli L., Kaplan D.M., Colas F., Stotz W., Yannicelli B., asentamiento de Concholepas concholepas (Bruguiere) Lett C. (2014) Identifying appropriate spatial scales for (: ) en la zona intermareal rocosa de marine conservation and management using a larval Valdivia, Chile. Revista Chilena Historia Natural, 66,93– dispersal model: the case of Concholepas concholepas (loco) 101. in Chile. Progress in Oceanography, 124,42–53. Moreno C.A., Asencio G., Duarte W.E., Marın V. (1998) Gonzalez J., Stotz W., Garrido J., Orensanz J.M., Parma A.M., Settlement of the muricid Concholepas concholepas and its Tapia C., Zuleta A. (2006) The Chilean TURF system: how relationship with El Nino~ and coastal upwellings in is it performing in the case of the loco fishery? Bulletin of southern Chile. Marine Ecology Progress Series, 167, 171–175. Marine Science, 78, 499–527. Nicastro K.R., Zardi G.I., McQuaid C.D., Teske P.R., Barker Goudet J. (2001) FSTAT: A Program to Estimate and Test N.P. (2008) Coastal topography drives genetic structure in Gene Diversities and Fixation indices (version 2.9.3). marine . Marine Ecology Progress Series, 368, 189– Available from http://www.unil.ch/izea/softwares/fstat.html. 195. Hedgecock D. (1994). Does variance in reproductive success Ojeda F.P., Labra F.A., Munoz~ A.A. (2000) Biogeographic limit effective population size of marine organisms? In: patterns of Chilean littoral fishes. Revista Chilena de Historia Beaumont A. (Ed.), Genetics and Evolution of Aquatic Natural, 73, 625–641. Organisms. Chapman and Hall, London: 122–134. Pantoja S., Iriarte J.L., Daneri G.D. (2011) Oceanography of Hinojosa I.A., Rivadeneira M.M., Thiel M. (2011) Temporal the Chilean Patagonia. Continental Shelf Research, 31, 149– and spatial distribution of floating objects in coastal waters 153. of central–southern Chile and Patagonian fjords. Continental Peakall R.O.D., Smouse P.E. (2006) GENALEX 6: genetic Shelf Research, 31, 172–186. analysis in Excel. Population genetic software for teaching Hubisz M.J., Falush D., Stephens M., Pritchard J.K. (2009) and research. Molecular Ecology Notes, 6, 288–295. Inferring weak population structure with the assistance of Pritchard J.K., Stephens M., Donnelly P. (2000) Inference of sample group information. Molecular Ecology Resources, 9, population structure using multilocus genotype data. 1322–1332. Genetics, 155, 945–959. Jerez G., Figueroa M. (2009) Desafıos y perspectivas de la Rice W.R. (1989) Analysing tables of statistical tests. Evolution, repoblacion de moluscos bivalvos en Chile. In Lovatelli A., 43, 223–225. Farıas A., Uriarte I. (Eds). Estado actual del cultivo y manejo Rousset F. (1997) Genetic differentiation and estimation of de moluscos bivalvos y su proyeccion futura: factores que gene flow from F-statistics under isolation by distance. afectan su sustentabilidad en America Latina. Actas de Pesca Genetics, 145, 1219–1228. y Acuicultura. No. 12. FAO, Roma: 223–235. Rousset F. (2008) Genepop 007: a complete reimplementation Kalinowski S.T. (2005) hp-rare 1.0: a computer program for of Genepop software for Windows and Linux. Molecular performing rarefaction on measures of allelic richness. Ecology Resources, 8, 103–106. Molecular Ecology Notes, 5, 187–189. Selkoe K.A., Watson J.R., White C., Horin T.B., Iacchei M., Lagos N.A., Castilla J.C., Broitman B.R. (2008) Spatial Mitarai S., Siegel D.A., Gaines S.D., Toonen R.J. (2010) environmental correlates of intertidal recruitment: a test Taking the chaos out of genetic patchiness: seascape genetics using in northern Chile. Ecological Monographs, reveals ecological and oceanographic drivers of genetic 78, 245–261. patterns in three temperate reef species. Molecular Ecology, Leiva G.E., Castilla J.C. (2001) A review of the world marine 19, 3708–3726. gastropod fishery: evolution of catches, management and Strub P.T., Mesias J.M., Montecino V., Rutllant J., Salinas S. the Chilean experience. Reviews in Fish Biology and Fisheries, (1998) Coastal ocean circulation off western South America. 11, 283–300. In: Robinson A.R., Brink K.H. (Eds), The Sea, 11, 273–313. Manriquez P.H., Delgado A.P., Jara M.E., Castilla J.C. (2008) Tatarenkov H., Avise J.C. (2010) Microgeographic population Field and laboratory pilot rearing experiments with early structure of green swordail fish: genetic differentiation ontogenic stages of Concholepas concholepas (: despite abundant migration. Molecular Ecology, 19, 257–268. Muricidae). Aquaculture, 279,99–107. Van Oosterhout C., Hutchinson W.F., Wills D.P.M., Shipley P. Molinet C., Arevalo A., Gonzalez M.T., Moreno C.A., Arata J., (2004) MICRO-CHECKER: software for identifying and Niklitschek E. (2005) Patterns of larval distribution and correcting genotyping errors in microsatellite data. settlement of Concholepas concholepas (Bruguiere, 1789) Molecular Ecology Notes, 4, 535–538. (Gastropoda, Muricidae) in fjords and channels of southern Warner R.R., Hamilton S.L., Sheehy M.S., Zeidberg L.D., Chile. Revista Chilena de Historia Natural, 78, 409–423. Brady B.C., Caselle J.E. (2009) Geographic variation in natal

368 Marine Ecology 37 (2016) 359–369 ª 2015 Blackwell Verlag GmbH Cardenas, Castilla & Viard Spatial genetic structure in loco

and early larval trace-elemental signatures in statoliths of White C., Selkoe K.A., Watson J., Siegel D.A., Zacherl D.C., the market Doryteuthis (formerly Loligo) opalescens. Toonen R.J. (2010) Ocean currents help explain population Marine Ecology Progress Series, 379, 109–121. genetic structure. Proceedings of the Royal Society B Weersing K., Toonen R.J. (2009) Population genetics, larval Biological Sciences, 277, 1685–1694. dispersal, and connectivity in marine systems. Marine Wright S. (1951) The genetical structure of populations. Ecology Progress Series, 393,1–12. Annals of Eugenics, 15, 323–354. Weir B.S., Cockerham C.C. (1984) Estimating F-statistics for Zacherl D.C. (2005) Spatial and temporal variation in statolith the analysis of population structure. Evolution, 38, 1358– and protoconch trace elements as natural tags to track larval 1370. dispersal. Marine Ecology Progress Series, 290, 145–163.

Marine Ecology 37 (2016) 359–369 ª 2015 Blackwell Verlag GmbH 369