The marine brooder Excirolana braziliensis (Crustacea: Isopoda) is also a complex of cryptic species on the coast of Chile

Speciation is a direct consequence of isolated populations in taxa with low dispersal potential. The brooding crustacean Excirolana braziliensis, with a presumably wide geographic range of distribution (~16o N-41o S in the Pacifi c and ~19o N-35o S in the Atlantic), has been detected to correspond to cryptic species on the coast of Panama. Latitudinal variations in reproductive features in E. braziliensis have been attributed to phenotypic plasticity, however, the differences may be the result of divergent populations. Considering that the taxon has been reported to be a complex of cryptic species in other geographic areas and given the phenotypic differences detected along its distribution range, we hypothesized that E. braziliensis is a complex of species in the coast of Chile. We used partial cytochrome c oxidase subunit I (COI) sequences from 132 individuals with the diagnostic morphology of E. braziliensis collected along ~2200 km of coast to determine the genetic structure of E. braziliensis. Phylogenetic and phylogeographic analyses showed three distinct clades with 14 to 19 % of genetic divergence and high values of genetic differentiation. Intra and inter-clade divergence revealed the existence of a species complex of E. braziliensis on the coast of Chile, supporting growing evidence of the high abundance of cryptic species in marine invertebrate taxa.


INTRODUCTION
Cr yptic species are species that are diffi cult to recognize based in morphology (Knowlton 1993, Bickford et al. 2007).While 'sibling species' may refer specifically to cr yptic species that share a common ancestor, the term is usually considered a synonymous of 'cr yptic species' (reviewed by Bickford et al. 2007).The abundance of cr yptic species discoveries reflects a lack of morphological divergence or an inadequate knowledge of

RESEARCH ARTICLE
the morphological dif ferences of divergent groups.Information provided by morphology is limited by our poor understanding of functional morphology, which is strongly infl uenced by natural selection (Knowlton 1993, Bickford et al. 2007).Moreover, morphological stasis can be promoted by stabilizing or convergent selection under extreme environmental conditions (Bickford et al. 2007).Research to uncover cr yptic species complexes can have serious implications for evolutionar y biology and for management and conser vation plans; for example, the discover y of cr yptic species in a particular habitat might reveal underscored levels of diversity and/or endemism upgrading the priority level for conser vation for that habitat (reviewed by Bickford et al. 2007).Complexes of sibling species and cr yptic species have been widely identified using DNA markers (e.g., Gouws et al. 2004, Moura et al. 2008, Jordaens et al. 2010, Azuma et al. 2011), and most cr yptic species are marine invertebrates (Knowlton 1993).Cryptic species have been repor ted along the geographic range that was assumed for one benthic marine species with low dispersal potential (Held & Wägele 2005, Raupach & Wägele 2006, Remerie et al. 2006, Linse et al. 2007, Raupach et al. 2007, Boissin et al. 2008, Baird et al. 2011, Carr et al. 2011, Doellman et al. 2011).Under a low dispersal scenario, isolation by distance and outbreeding depression are enough to account for a high speciation rate (Hoelzer et al. 2008).Many marine brooders may have an enhanced speciation potential given their small body size and low vagility, resulting in low potential for active dispersal (Teske et al. 2007).
Peracarids are a group of direct developing crustaceans that brood their lar vae up to a crawl-away stage.For some peracarids longdistance dispersal may occur through rafting on fl oating substrata such as detached macroalgae (Thiel & Haye 2006), or via anthropogenic transfer (e.g., Wittmann & Ariani 2009).Other peracarid species, however, live in habitats that are not suitable for alternative dispersal mechanisms, for example, the abundant amount of species that live in soft bottom sandy beaches, cannot sur vive on floating objects (Thiel & Gutow 2005).Occasional dispersal events between beaches may occur if individuals remain attached to its prey (e.g., live fi sh) for enough time to complete a passive transport (Sponer & Lessios 2009).Similarly, marine birds may occasionally transpor t individuals if they casually remain attached to their plumage or feet, as reported for aquatic non-marine invertebrates (Frisch et al. 2007).These stochastic long-distance dispersal events are likely to be infrequent not allowing for suf ficient gene flow to counterbalance the effects of genetic drift.Therefore, the wide range of geographic distribution reported for many soft-bottom peracarids may be the result of undetected species complexes.
The beach dwelling isopod Excirolana braziliensis Richardson has a wide geographic range in America and Chile.It has been found in the Pacific Ocean from the Gulf of Califor nia in Mexico (Dexter 1976) to Chiloé Island in southern Chile (Jaramillo et al. 2000).In the Atlantic Ocean it has been repor ted from the Gulf of Mexico (Dexter 1976) to Uruguay (Defeo et al. 1997).For the coast of Panama Lessios & Weinberg (1994) and Sponer & Lessios (2009) reported three divergent lineages in E. braziliensis, indicating the presence of a species complex.Along the coast of Chile E. braziliensis has been scarcely studied.A highlight is the study of Cardoso & Defeo (2003); they found strong latitudinal variations in reproductive features in populations of E. braziliensis from the Pacifi c and Atlantic Oceans.Even though the authors attributed the variations to phenotypic plasticity, the phenotypic divergence could also be a consequence of genetic differentiation between local populations.
Given the likelihood of speciation in taxa with low active and passive dispersal potential, the presence of a species complex of E. braziliensis on the coast of Panama and the phenotypic variation detected among populations, we hypothesized that there is a species complex in E. braziliensis in the Chilean coast.For this, the genetic structure between local populations of E. braziliensis along the nor th-central coast of Chile was analyzed using cytochrome c oxidase subunit I (COI) sequences.

Sampling, DNA extraction, PCR and sequencing
Samples were collected from seven localities of the Chilean coast influenced by the Humboldt Current System (from 18.4º S to 38.2º S, along ~2200 km) (Table 1, Fig. 1A).During low tide, macroinfaunal species were washed out of bulk sand samples with 2 mm mesh size bags.Excirolana species are easily and clearly distinguishable from other macroinfaunal isopods of sandy beaches (Jaramillo et al. 1998).Samples were were preserved in 95 % ethanol at -20º C until DNA extraction.
Of the three species of the genus Excirolana present along the Chilean coast, E. braziliensis can be easily distinguished from the others (E.hirsuticauda and E. monodi) by the shape of the cephalon and the telson, as well as by the length of the antennae.In E. braziliensis the cephalon width is three times the length, the telson has a rounded apex, antenna 1 reaches half of pereonite 3, and antenna 2 extends until the end of pereonite 5 (Jaramillo 1982).All individuals used in this study had the diagnostic characteristics of E. braziliensis (Jaramillo 1982).
Depending on the availability, seven to 22 individuals of E. braziliensis per locality were analyzed (Table 1).Whole individuals were used for the DNA extraction; except for individuals larger than 5 mm for which only telson muscle was used.Genomic DNA was extracted with Qiagen's QIAamp DNA mini kit following manufacturer's instructions.
We used the COI gene to detect population genetic structure.This marker has been extensively validated as a good species marker for most animal species, and it is the preferred marker for DNA Barcoding in many animal groups including crustaceans (e.g., Bucklin et al. 2011, Havermans et al. 2011). Sponer & Lessios (2009) also used COI and could easily distinguish the clades that comprise the species complex of E. braziliensis in the coast of Panama.
A partial sequence of the COI gene was amplifi ed with PCR using the primers HCO2198 and LCO1490 (Folmer et al. 1994).PCR's were carried out with 1X PCR buffer, 2 mM MgCl 2 , 0.4 µM of each primer, 0.2 mM of each dNTPs, 0.03 U µL -1 Taq and 1.5 mg mL -1 Bovine Serum Albumin (BSA), 1µL of template DNA (~20 ng) and distilled H 2 O to complete fi nal volume.The cycling conditions were: initial denaturing of 94 ºC for 10 min followed by 35 cycles of 1 min denaturing at 94 ºC, 1 min annealing at 51 ºC and 2 min extension at 72 ºC, and ending with a fi nal extension at 72º C by 13 min.Amplicons were purifi ed by adding 28.8 µL Shrimp Alkaline Phosphatase and 7.2 µL of Exonuclease I to 45 µL PCR product followed by incubation at 37ºC for 15 min and 80 ºC for 15 min.Purifi ed products were sequenced in both directions with an ABI 3730XL capillary automated sequencer.
CODONCODE Aligner 1.2.0 (CodonCode) was used to inspect chromatographs for ambiguous sites, obtain consensus sequences, and for sequence alignment.Alignment was verifi ed through translation to amino acids.Sequences were submitted to GenBank (Accession Numbers: FJ541198-FJ541243).

Data analyses
Two sequences of E. hirsuticauda (Genbank Accession Numbers: FJ532105 and FJ532142) were used as outgroup taxa for phylogenetic analyses.Phylogenetic hypotheses were searched performing Maximum Likelihood Analysis (MLA), and Bayesian Analysis (BA), implemented in PAUP 4.b10 (Swofford 2002) and MrBayes 3.1.2(Ronquist & Huelsenbeck 2003), respectively.The model of DNA evolution that better fi t the data according to jModelTest 0.1.1 (Posada 2008) was Hasegawa-Kishino-Yano plus gamma (HKY+G), and was used for MLA and BA.A bootstrap analysis with 10000 replicates was used to obtain support values for the nodes for the MLA using PAUP.In the BA, posterior probability values for the nodes were obtained using 5000000 iterations and sampling every 10 generations.Two simultaneous and independent analyses were run until the standard deviation of split frequencies was below 0.01.The fi rst 25 % of saved trees were discarded as burn-in.
The mutational relationships between all haplotypes detected were visualized with a median joining haplotype network generated using NETWORK 4.5 (Bandelt et al. 1999).Ambiguous loops in the network were resolved using the criteria of Crandall & Templeton (1993).The genetic p distance among haplotypes was calculated with PAUP*4.b10.Pairwise F ST values between localities were calculated in ARLEQUIN 3.11 (Excoffi er et al. 2005).The signifi cance of the F ST values was evaluated by 10000 permutations.

RESULTS
Sequences of 583 base pairs were obtained for 132 individuals from seven localities of the Chilean coast (Table 1, Fig. 1A).The MLA Consistent with phylogenetic analyses, the arrangement of the 45 haplotypes in the median joining haplotype network shows three distinct haplogroups, each corresponding to the clades detected on the phylogenetic analyses (Fig. 1C).The Northern Clade was found in Arica (18.4ºS), the Center Clade from Chipana (21.3ºS) to Puerto Aldea (30.3ºS), and the Southern Clade in Pangua (34.5ºS) and Queule (38.2ºS) (Fig. 1A and 1B).
The Nor ther n and Center clades are separated by 110 mutational steps and have 18 % of genetic divergence.The Center and Southern clades are separated by 81 mutational steps and have 14 % of genetic divergence.Northern and Southern clades diverge in 19 %.While the Northern Clade was represented by only one locality, the genetic divergence within the Center clade ranged from 0.45 % to 0.67 % and was 0.66 % for the Southern Clade.
Pairwise F ST values between localities from the three clades were close to 1.0 and highly signifi cant (P < 0.0001) (Table 2), indicating almost complete genetic differentiation between clades.Intra-clade divergence was evaluated for the Center and Southern clades, which had more than one location included in the study; both Center and Southern clades had significant intra-clade population-pair wise F ST values (P < 0.001), however, genetic Mapa parcial de Chile (Pacífi co Sur Este) indicando las localidades donde se recolectaron las muestras de E. braziliensis (A).Filograma de Máxima Verosimilitud de los 45 haplotipos de E. braziliensis (B).Red de haplotipos de E. braziliensis construida con el método "median joining" (C).En el fi lograma, los valores sobre las ramas corresponden a los valores de iteraciones obtenidos con Máxima Verosimilitud y los valores bajo las ramas a probabilidades posteriores Bayesianas.Los colores en el fi lograma y en la red de haplotipos representan las diferentes localidades indicadas en el mapa (A).El símbolo║indica que la línea fue truncada y los números al lado de indican el número de pasos mutacionales entre los grupos de haplotipos.
differentiation between localities within clades was less than half than between clades (Table 2).

DISCUSSION
As hypothesized, COI sequence analyses revealed that the wide range of geographic distribution of the brooder E. braziliensis along the north-central Chilean coast is explained by the presence of a species complex.The low dispersal potential of Excirolana species, together with their small body size and concomitant low intrinsic dispersal have led to three highly divergent lineages between 18 and 38º S on the South-East Pacifi c coast under the infl uence of the Humboldt Current System.
The genetic divergence among clades of E. braziliensis ranged from 14 % to 19 %, while the divergence within clades was ver y low (< 0.67 %).Comparisons of species of the family Cirolanidae revealed a sequence divergence from 13.6 % to 14.7 % (Wetzer 2001).Other studies comparing species within a genus of peracarids have led to similar or slightly higher values.For the genus Idotea, Wares (2001) determined a genetic divergence of 17-25 %; Markow & Pfeiler (2010) repor ted divergences from 14.9 % to 30.3 % among species of Ligia, including ter restrial and marine-ter restrial species; for the amphipod genus Eusirus interclade distances fell between 4.6 to 10.9 % (Baird et al. 2011);and Baratti et al. (2005) suggested that Sphaeroma terebrans was a complex of cryptic species with genetic distances of 15 to 18 %.The high values of genetic divergence among clades detected within E. braziliensis were close to the divergence of the COI gene generally reported between species.When genetic divergence between clades exceeds the typical values for congeners it is reasonable to infer there are sibling/cr yptic species.Therefore, the detected levels of genetic divergence between clades whose individuals have the diagnostic features of E. braziliensis suggest that there is a complex of cryptic species under the formal taxon E. braziliensis on the Chilean coast, as had been shown for the coasts of Panama The results of this study provide insights into a former study of E. braziliensis performed by Cardoso & Defeo (2003).They found that Pair wise F ST values between localities.Values between localities from the different clades are indicated by "A".All the other correspond to the values between localities within Center and Southern Clades.All the values were signifi cant (P < 0.001).Codes as in Table 1.
Valores de F ST entre localidades.Los valores entre localidades de los diferentes clados están indicados por "A".Todos los otros corresponden a los valores entre localidades dentro de los clados Centro y Sur.Todos los valores fueron signifi cativos (P < 0.001).Las siglas representan localidades como se muestra en la Tabla 1.  2003, Padial et al. 2009)).Target taxa should be those with low intrinsic and extrinsic dispersal potential that have a large geographic range reported, as is the case of E. braziliensis and other beach dwelling peracarids.W ithin the detected lineages of E .braziliensis there was a high and signifi cant genetic dif ferentiation.Similarly, other marine taxa with low dispersal potential show high and significant genetic dif ferentiation, such as the Atlantic clade of the mysid Mesopodopsis slabberi (Remerie et al. 2006), and Mediterranean clades of the brooding brittle star, Amphipholis squamata (Boissin et al. 2008).High levels of genetic differentiation within clades or species seem to be a common pattern in marine brooders, and may refl ect an ongoing process of speciation.

Ari
In conclusion, E. braziliensis from the Chilean coast shows a similar genetic pattern to that detected in Panama for this species, for others isopods worldwide, and for few marine taxa from the Chilean coast.There are highly divergent clades that most likely correspond to a complex of cryptic species.Our study supports growing evidence of the high abundance of cr yptic species within marine inver tebrate taxa.Fur ther studies in other geographic areas where E. braziliensis is found (e.g., between Panama and Chile) are required to determine the number of cr yptic species that make up the complex and how are they distributed in the Atlantic and Pacifi c coast.

Fig. 1 :
Fig. 1: Partial map of Chile (South-East Pacifi c coast) showing the localities where E. braziliensis were sampled (A).Maximum Likelihood phylogram of 45 COI haplotypes of E. braziliensis (B).Median joining haplotype network of E. braziliensis (C).In the phylogram, values above the branches correspond to Maximum Likelihood bootstrap and values below to Bayesian posterior probabilities values.Colors in the phylogram and haplotype network represent the different localities shown in the map (A).The symbol║indicates that the line was truncated and the numbers next to show the number of mutational steps between groups of haplotypes.
(Lessios & Weinberg 1994, Sponer & Lessios  2009).So far, few studies have identifi ed marine cr yptic species on the Chilean coast.Using morphological, developmental and genetics features, Véliz et al. (2003) found three cryptic species of Crepidula from northern Chile.Later, Tellier et al. (2009) identifi ed two divergent lineages of the kelp Lessonia nigrescens at 30º S that likely evolved through parapatric speciation.Recently, Montecinos et al. (2012) detected three highly divergent clades along the Chilean coast in the red alga Mazzaella laminarioides, which likely cor respond to cryptic species.Further studies are needed to reveal the possible existence of other cryptic species on the Chilean coast.

TABLE 1
Number of individuals of Excirolana braziliensis analyzed in each locality of the Chilean coast.Número de individuos de Excirolana braziliensis analizados en cada localidad de la costa de Chile.