Multiloci analyses suggest synonymy among Rhomboplites , Ocyurus and Lutjanus and reveal the phylogenetic position of Lutjanus alexandrei ( Lutjanidae : Perciformes )

Lutjanidae comprises 21 genera and 135 species widespread throughout Atlantic, Indian and Pacific oceans. Nonetheless, the phylogenetic relationships of Lutjaninae remain uncertain. Furthermore, phylogenetic hypotheses for Lutjanus alexandrei, an endemic species from northeastern Brazilian coast, in Lutjanidae are absent so far. Therefore, we carried out multiloci analyses, combining both mitochondrial and nuclear DNA sequences in Lutjaninae species from Western Atlantic focusing on the controversial relationships among Lutjanus, Rhomboplites, and Ocyurus. Besides, we determined the phylogenetic position and dated the origin of L. alexandrei. The phylogenetics trees based on the 4.4 kb for 11 species corroborated the synonym among Lutjanus and the putative monotypic genera. For the dating of L. alexandrei, another nucleotide dataset (3.0 kb; 40 species) validated the genetic identity of this species that diverged from the sister taxon L. apodus between 2.5 – 6.5 Mya, probably as a result of the barrier caused by the muddy outflow from Orinoco and Amazon rivers along the coastal zone. This report is the most robust multiloci analysis to confirm the synonymy of the three genera of Lutjaninae from Western Atlantic and the first reliable inference about the phylogenetic relationships and origin of L. alexandrei.


Introduction
Lutjanidae is a fish family mainly composed of marine demersal species that can be found up to 550 m deep, even though each species seems to present specific depth preferences.In general, the lutjanids (or snappers) inhabit coastal tropical and subtropical waters in association with rocky bottom and coral reefs, even though some species depend on estuaries during early life cycle (Cervigón, 1993;Nelson et al., 2016).The snappers show remarkable phenotypic variation and medium-to large-sized bodies (up to 1 m in total length), representing important fishery resources along their range (Cervigón, 1993;Nanami, Yamada, 2008;MPA, 2009MPA, , 2010MPA, , 2011)).
Within Lutjanus, the most species-rich genus of Lutjanidae, the Brazilian snapper (L.alexandrei Moura, Lindeman, 2007) stands out as an endemic species of the Brazilian coast.In the description report of this taxon, the authors highlighted that L. alexandrei is a common species from estuarine and reef environments that has been traditionally misidentified as L. griseus (Linnaeus, 1758) or L. apodus (Walbaum, 1792), suggesting that both taxa are not found in Brazil (Moura, Lindeman, 2007).Nonetheless, molecular studies that attempted to validate the phenotypic description of L. alexandrei or to infer the phylogenetic position of this taxon are absent so far.
Indeed, the available phylogenetic hypotheses of snappers, particularly in relation to Lutjaninae species from WA, are conflicting.In this context, a long debate refers to the validation of Rhomboplites and Ocyurus as monotypic genera.Both genera were defined according to morphological traits (Evermann, Marsh, 1900;Vergara, 1980), but several studies have shown their close relationship with Lutjanus representatives.While protein analyses reinforced the separation of Rhomboplites as a distinct genus, they provided no evidence for placing O. chrysurus apart from Lutjanus (Chow, Walsh, 1992).
Furthermore, some molecular phylogenies support the inclusion of O. chrysurus within Lutjanus and cast doubt on the taxonomic status of Rhomboplites (Sarver et al., 1996).The taxonomic classification of these genera has also called into question based on cytogenetic data since chromosomal differences were observed between Rhomboplites and Lutjanus but not between the latter and Ocyurus (Nirchio et al., 2009).
Based on mitochondrial DNA (mtDNA), Gold et al. (2011) inferred the evolutionary relationships of 20 species of Lutjaninae (12 taxa from WA, one species from Pacific Ocean and seven from Indo-Pacific).According to their results, the authors recommended that the monotypic genera should be included within Lutjanus.
Later, Gold et al. (2015) proposed a phylogenetic hypothesis for snappers from Eastern Pacific that also included species from WA, Indo-Pacific, and Western Indian oceans.Again, the molecular data (mtDNA sequences) clustered together Rhomboplites, Ocyurus, and Lutjanus.
Even though there are several pieces of evidence for the synonymy among Ocyurus, Rhomboplites and Lutjanus, no conclusive study has already been reported, possibly because the available morphological analyses were based on limited sampling with a few species.Moreover, there is only one phylogenetic study using a single nuclear gene in Lutjanidae (Frédérich, Santini, 2017).
Therefore, the present study provided a thorough phylogenetic inference of Lutjaninae from WA, based on the most extensive molecular data so far, comprising seven loci from mitochondrial and nuclear genome regions.We carried out an integrative analysis focusing on the controversial relationships among Lutjanus, Rhomboplites and Ocyurus, aiming to reinforce the synonimization.Furthermore, we wanted to propose the phylogenetic position and origin period of L. alexandrei.

e180109[3]
Tab. 1. Markers and their evolutionary rates included in the present study, with respective Genbank access codes of sequence data.The species in bold represent specimens samples in the present study selected for the phylogenetic analyses (BI and ML) in Western Atlantic species of Lutjaninae, based on all markers.The total set of species was used to infer the evolutionary relationships and the origin of Lutjanus alexandrei within Lutjanidae, based on the highlighted markers.* Haemulidae representatives used as outgroups used in phylogenetic trees.In both sets, Conodon nobilis was used as outgroup since Haemulidae has been regarded as the sister group of Lutjanidae (Betancur-R et al., 2013;Near et al., 2013).
A great portion of the sampling derived from collection expeditions from the present study, while the remaining samples included sequences available in GenBank.Muscle tissues were collected from C. nobilis and all species of snappers from WA, except for L. apodus; L. griseus; L. mahogoni (Cuvier, 1828); Pristipomoides and Apsilus.These samples are originated from local fish markets, trawl net expeditions and fishing landing ports along the American coast, as detailed in Tab. 1.The fish from which samples were taken, were not collected exclusively for this study.They were captured with fishermen support who available only the samples and later destined the whole fish for markets, without some structures like gut, stomach.In this way, not possible the deposit of individuals in a museum.
The samples were stored in microtubes, preserved in 90% ethanol and transported to the laboratory and kept at -20°C up to their utilization.
Isolation, amplification, and DNA sequencing.Total genomic DNA was isolated using commercial kits (Wizard Genomic®, PROMEGA), according to the manufacturer's instructions.The extracted DNA was visualized after electrophoresis in 1% agarose gel stained with Gel Red TM (BIOTIUM) under ultraviolet light.The selected genome regions were amplified via PCR, comprising four mitochondrial genes (16S rRNA -16S, cytochrome c oxidade subunit 1 -COI, cytochrome b -cyt b, and NADHdehydrogenase subunit 4 -ND-4) and three nuclear genes (rhodopsin -Rhod, TMO-4C4, and RAG-1).Each PCR comprised a mix of deoxynucleotides (dNTPs) at 200 μM, 1x buffer, MgCl 2 at 2 mM, 0.4 μL of each primer (10 μM), 0.06 U/μL of Taq DNA polymerase, about 50 ng of template DNA and ultrapure water to a final volume of 15 μL.The primers used to amplify each gene are described in Tab. 2.
The amplification conditions were: first denaturation step at 95°C for 5 min., followed by 38 cycles of denaturation at 95°C for 30 sec., hybridization for 30 sec.(see temperature details for each primer pair in Tab. 2) and extension at 72°C for 2 min., plus a final extension step at 72°C for 5 min.The RAG-1 fragment was amplified using nested PCR, in which the first reaction used the primers RAG1-2510F (Li, Ortì, 2007) and RAG1-4090R (Lopez et al., 2004), followed by a second reaction with the primers RAG1-2533F and RAG1-4078F (Lopez et al., 2004).
The amplicons were purified in PEG 8000 (polyethylene glycol) according to Paithankar, Prasad (1991) and submitted to dideoxi sequencing reaction (Sanger et al., 1977) using the Big Dye kit (ABI Prism TM Dye Terminator Cycle Sequencing Reading Reaction -Applied Biosystems, USA).The fragments were bidirectionally amplified using specific forward and reverse primers.The final product was precipitated and sequenced in ABI 3500 automatic sequencer (Applied Biosystems).
Tab. 2. Primers, references and hybridization temperature used for mitochondrial and nuclear markers.Phylogenetic Analyses.The sequences from each gene were organized into individual datasets and edited in the software BioEdit (Hall, 1999).Visual inspection was carried out to correct the sequence position in cases of doubtful or misidentified nucleotides.
The alignment of the corrected sequence datasets was performed automatically in the online version of MAFFT (Katoh, Standley, 2013).The 3' and 5' ends were removed to discard unreadable sequence parts and to decrease the amount of missing data.
The selection of the best scheme of partitions and the evolutionary models to explain the variation in sequences from each gene was performed in the PartitionFinder2 (Guindon et al., 2010;Lanfear et al., 2012;Lanfear et al., 2016) (Tab. 3).
Based on the multilocus dataset, we estimated the phylogenetic relationships among taxa based on maximum likelihood (ML) and Bayesian inference (BI).The ML tree was built using RAxML version 8.2 (Stamatakis, 2014) with the GTR+G evolutionary model for each partition (see Tab. 3).The support of each branch was calculated using bootstrap with 1000 pseudoreplicates.
The BI was carried out in MrBayes, version 3.2 (Ronquist et al., 2011) using the previously selected evolutionary model for each partition (see Tab. 3).
Two independent runs with four MC 3 chains were performed based on 10 million generations, with tree sampling at each 10,000 generations and 10% of burn-in.The chain convergence was evaluated in Tracer, version 1.6 (Rambaut et al., 2014) and the consensus tree was visualized and edited in FigTree, version 1.4.3 (Rambaut, 2016).We also used a multi-species coalescent approach to estimate a species tree using the *BEAST option, available in BEAST v. 1.8.1 (Drummond et al., 2012).The runs were made using a relaxed lognormal clock, Yule Process as tree prior, 400 million of steps, with sampling at 10,000 and using 20% of burn-in.The choice of the optimal partition scheme and evolutionary models were made in PartitionFinder2 (Guindon et al., 2010;Lanfear et al., 2012;Lanfear et al., 2016) (see Tab. 3).The *BEAST runs were conducted in the CIPRES Science Gateway V. 3.3 (Miller et al., 2010).The mixing and convergence of runs were checked in Tracer, version 1.6 (Rambaut et al., 2014), and the consensus tree was made in the Beast v. 1.8.1 application "treeannotator", visualized and edited in FigTree, version 1.4.3 (Rambaut, 2016).
Divergence time of lineages.To infer the time of diversification among lineages that gave rise to L. alexandrei, as well as its phylogenetic position in Lutjanidae, we analyzed a more extensive dataset, including the collected samples of snappers from this study (previously used in ML and BI analyses) as well as the available sequences in Genbank analyzed by Frédérich, Santini (2017), and the sequences provided by Gold et al. (2011) from a non-identified species (L.cf.apodus).Therefore, the data matrix encompassed 48 specimens and 3,037 characters represented by 504 bp of 16S; 1,133 bp of COI; 750 bp of Cyt b, and 650 bp of RAG-1 (see Tab. 1).
The calibration of the divergence period was based on the divergence time between Lutjanidae and Haemulidae (67 Mya ± 10 S. D; Frédérich, Santini, 2017) and on the data estimated for the diversification of "Lutjanus", according to the minimum age of the most ancient fossil reported for this family (Hypsocephalus atlanticus Swift, Ellwood, 1972) data between 33.9 and 50 Mya (Frédérich, Santini, 2017).Since this fossil is considered a close relative of Hoplopagrus guntherii (Gill, 1862), a taxon that has been recovered within Lutjanus, this fossil dating was employed to determine the minimum age of Lutjanus (Frédérich, Santini, 2017).
The diversification analysis was carried out in BEAST v. 1.8.4 (Drummond et al., 2012) using the CIPRES Science Gateway V. 3.3 (Miller et al., 2010) based on 200 million generations, relaxed log-normal clock, and the evolutionary models and partitions selected by PartitionFinder2 using the "greedy algorithm" (Guindon et al., 2010;Lanfear et al., 2012;Lanfear et al., 2016) (see Tab. 3).Genealogies were sampled at each 10,000 generations using 10% of burn-in.The chain convergence was inspected using the software Tracer, version 1.6 (Rambaut et al., 2014) and the consensus tree was visualized and edited in FigTree, version 1.4.3 (Rambaut, 2016).

Results
The multiloci phylogenetic inferences herein performed for Lutjaninae species from WA comprised the highest number of loci in Lutjanidae phylogenies so far (4.4 kb).The topologies of ML, BI, and species trees were very similar and thereby we only show the BI topology, including the support values of the three analyses (Fig. 1).
The clade A supports the monophyly of studied Lutjaninae, that is, for Lutjaninae species from WA included in this analysis.The clade B comprises two major lineages: one of them diverges giving rise to the clade D, composed of L. cyanopterus and clade C (L. jocu and L. alexandrei), while the other diversifies into clade E, encompassing the remaining taxa from this study.Within this group, the L. synagris + L. analis branch (clade F) is a sister group of clade that contains R. aurorubens and O. chrysurus, which is further related to and L. purpureus, L. campechanus, L. vivanus, and L. buccanella that appear to be closely related being named as clade G.All these clades were characterized by high probability values (0.99 or 1), except for clade D (0.96), and the branch that groups O. chrysurus and R. aurorubens (0.50).However, the bootstrap values in ML inference were more strict (see Fig. 1).
With regard to the phylogenetic positioning of L. alexandrei in Lutjanidae, according to the BI based on a dataset of 3.0 kb and 40 species of Lutjanidae, L. alexandrei is the sister group of L. apodus, forming a well supported clade.This clade is reciprocally monophyletic to L. griseus and L. jocu that, altogether, are closely related to L. cyanopterus + L. argentimaculatus (Forsskal, 1775) (Fig. 2).
Based on the present results, the diversification of the lineage that gave rise to the Brazilian snapper has begun between 4.5 -9.2 Mya in Pliocene, while the L. alexandrei + L. apodus pair diverged between 2.5 -6.5 Mya.The diversification of L. griseus and L. jocu occurred simultaneously between 2.9 -7.3 Mya.

Discussion
Phylogenetic relationships of Lutjaninae species from Western Atlantic.The monophyly recovered for Lutjaninae species from WA included in the study confirms the findings reported by Gold et al. (2011) based on mtDNA (cyt b, COI and ND-4).On the other hand, if we consider any subfamily, and not only WA species, other molecular studies considered that the status of this subfamily as a natural group is questionable.For instance, analyses using 16S rRNA, cyt b, ND-4 and COI mitochondrial genes placed Caesionidae within Lutjaninae (Miller, Cribb, 2007;Chu et al., 2013;Gold et al., 2015).Similarly, Frédérich, Santini (2017) indicated that Lutjaninae is a non-monophyletic group inasmuch as all genera of fusiliers (regarded as a subfamily of Lutjanidae by these authors) were closely related to the former.In fact, the inclusion of Caesionidae within Lutjanidae has already been evidenced by morphological analyses (Johnson, 1993;Nelson, 1994) and corroborated by molecular data (Betancur-R et al., 2013), which in the present study is also reported (see Fig. 2).Lutjanus alexandrei, placed as a sister group of L. jocu, is a recently described species (Moura, Lindeman, 2007), endemic to the Brazilian coast that lack previous phylogenetic inferences.The clade C herein reported (L.jocu + L. alexandrei) agrees with the results from Gold et al. (2011) and Gold et al. (2015) that recovered a close relationship among L. jocu, L. griseus, L. apodus and L. cf.apodus, since the latter is likely to represent L. alexandrei, as suggested by Gold et al. (2011).
Nonetheless, taking into account the first sample set, the present results differ from other studies (Sarver et al., 1966;Gold et al., 2011) in relation to L. analis, herein grouped with L. synagris (clade F).In addition, L. cyanopterus, here, groups together with L. jocu and L. alexandrei (clade D), corroborating the grey snapper group of the phenetic classification (Rivas, 1966;Vergara, 1980).But, other studies with snappers from WA report that L. cyanopterus is not closely related to any other studied species in this Atlantic region (Sarver et al., 1966;Gold et al., 2011).As commonly observed for many other Lutjaninae species from WA, the putative sister taxon of L. cyanopterus is L. novemfasciatus (Gill, 1862), a native species from Pacific Ocean (Chu et al., 2013;Gold et al., 2015;Frédérich, Santini, 2017).
The relationships in the clade F and in the clade D, apparently discordant with previous studies (Sarver et al., 1966;Gold et al., 2011), are in fact the result of the more limited sampling of the first sample set analyzed, therefore analyzing the second sample set, the relationships between L. synagris, L. mahogoni and L. analis, as well as among gray snapper, resemble results already reported (Chu et al., 2013, Frédérich, Santini, 2017).
The controversial taxonomic status of Rhomboplites and Ocyurus.Rhomboplites aurorubens and Ocyurus chrysurus were included in clade E along with other Lutjanus species, thus reaffirming the paraphyletic status of this genus.This result brings back a long debate about the inclusion of Rhomboplites and Ocyurus in Lutjanus.
These two monotypic genera were recognized by Gill (1862) that differed them from Lutjanus based on the following traits: the forked caudal fin with thin lobes in Ocyurus and the presence of a rhomboid vomerine tooth in Rhomboplites versus more or less truncate caudal fin and triangular vomerine tooth in Lutjanus (Gill, 1862).Nonetheless, more extensive and recent studies revealed that the vomerine teeth in Lutjanus species might range from "V"-shape to rhomboid or anchor-like shape (Cervigón, 1993;Anderson, 2002;Claro, Lindeman, 2008).
Other morphological traits that have been used to differentiate both monotypic genes from Lutjanus include the number of gill rakers (lower in Lutjanus) and the presence of ectopterygoid teeth (absent in Lutjanus) (Cervigón, 1993;Anderson, 2002).Rhomboplites is also diagnosed by the number of spines in dorsal fins (n=12) while Lutjanus species present 10 to 11 spines (Cervigón, 1993;Claro, Lindeman, 2008).
Accordingly, Sarver et al. (1996) proposed the synonymy between the genera Ocyurus and Lutjanus inasmuch as O. chrysurus is clustered along other Lutjanus species a pattern also observed in relation to Rhomboplites, but not mentioned by the authors.Likewise, the phylogenetic trees reported by Miller, Cribb (2007) also recovered R. aurorubens within the Lutjanus lineage.
In a comparative approach, Gold et al. (2011) showed that both R. aurorubens and O. chrysurus are placed along with Lutjanus in the phylogenetic trees, thereby suggesting the reallocation of these species in the genus Lutjanus.These authors stated that most of the distinctive morphological features might represent differential adaptation to feeding thereby representing homoplasies while the chromosomal peculiarities reported by Nirchio et al. (2009) between Rhomboplites and Lutjanus could be autapomorphic since few species have been cytogenetically analyzed.Moreover, they pointed out that the weak support of the branches related to O. chrysurus and R. aurorubens might reflect fast speciation processes that are hardly identified by mtDNA markers only (Zink, Barrowclough, 2008).
Only recently, Frédérich, Santini (2017) published a study about the time and mode of diversification in Lutjanidae based on fossil calibration that encompassed 70% of their actual diversity and a single nuclear gene (RAG-1).The authors hypothesized that the remarkable distinct ecological traits between Lutjanus and other closely related lineages (e.g., peculiar food preferences, occurrence at differential depths and specific life history strategies) may be the result of fast and recent adaptive radiation, differently from the pattern usually reported for reef-associated fish.Furthermore, this report also placed Rhomboplites and Ocyurus along with Lutjanus species, reinforcing this genus as a non-monophyletic group (Frédérich, Santini, 2017).
Taking into account that our results corroborated several reports that reveal the high genetic affinities between Lutjanus, Rhomboplites and Ocyurus, and that the morphological differences among these putative genera are subtle and mostly derived from ecological adaptation, we consider that there is sufficient evidence to allocate these monotypic genera within Lutjanus.Even though, the monophyly of the genus remains unsolved and should be investigated by integrative studies, since the grouping of Lutjanus with other genera of Lutjaninae and Caesioninae have also been indicated (Chu et al., 2013;Frédérich, Santini, 2017).
The relationships recovered in BI for L. alexandrei, L. apodus, L. jocu and L. griseus are consistent with previous inferences (Sarver et al., 1996;Gold et al., 2011;Chu et al., 2013;Gold et al., 2015;Frédérich, Santini, 2017), the present study stands out by the inclusion of L. alexandrei that was selected for dating inferences.
The phylogenetic analysis by Gold et al. (2011) included a non-identified species of Lutjanus (L.cf.apodus), placed as a sister taxon of L. apodus, and related to both L. jocu and L. griseus.According to these authors, this specimen could represent L. alexandrei since it was collected in northeastern Brazil, has proved to be evolutionary related to the abovementioned species, and presented morphological traits similar to those reported by Moura, Lindeman (2007).The relationships herein presented in this clade agree with the reports by Gold et al. (2011) and Gold et al. (2015), allowing us to confirm that L. cf.apodus cited in both studies actually corresponds to L. alexandrei.
Our estimates of time for the diversification events that led to the L. alexandrei lineage were similar to the ranges recorded by Gold et al. (2011) for the L. apodus + L. cf.apodus (= L. alexandrei) pair as well as for the cluster that comprises both taxa, even though these authors used calibration based on the molecular evolutionary rate of cyt b marker while we relied on fossil dating as carried out by Frédérich, Santini (2017).
The Pliocene was a period of low tectonism preceded by volcanic activities that formed the western arch of Lesser Antilles islands in the Caribbean Sea.In addition, the uplift of the Isthmus of Panama has possibly intensified the Gulf Current, thus leading to a greater flow of warm waters to the regions southern to the Caribbean (Gold et al., 2011).
The second diversification event that has probably been responsible for the separation between L. apodus and L. alexandrei should be related to the 2,300 km muddy outflow from the Amazonas and Orinoco rivers in the coastal region of South America that prevented the formation of coral reefs thus acting as a putative barrier (Gold et al., 2011).Indeed, in their description of L. alexandrei, the authors emphasized the importance of the freshwater discharges from both rivers in population isolation and allopatric speciation between the Caribbean and most of South America coast (Moura, Lindeman, 2007).
For the separation of L. apodus and L. griseus we were also in agreement with Gold et al. (2011).The species that take part of this clade share some ecological traits, including the preference for estuarine environments for reproduction and during juvenile stages (Cervigón, 1993;Moura, Lindeman, 2007;Claro, Lindeman, 2008).
In their description of L. alexandrei, Moura, Lindeman (2007) discuss the validation of the occurrence of L. apodus and L. griseus in Brazil, once the specimens formerly recognized as both taxa available in museums, photographic records, voucher material and collections along estuarine and reef areas of Brazilian coast invariably corresponded to the new species identified by these authors.Therefore, the native range of L. apodus and L. griseus would be restricted from Massachusetts to the Caribbean Sea, while the Brazilian snapper should be endemic to northeastern Brazil, from Maranhão to Bahia coast, being sympatric to L. jocu, thus leading to their misidentification (Moura, Lindeman, 2007).Nonetheless, L. jocu is widespread from the USA coast to northeastern Brazil (Allen, 1985;Cervigón, 1993).
The apparent absence of L. alexandrei in the estuaries from the northern Brazilian coast might be related to ecological features such as the low tolerance of this species to salinity variation when compared to L. jocu, a species that can be adapted to riverine habitats with salinity levels above 40‰ (Claro, Lindeman, 2008).Therefore, even though both species depend on the estuarine environments during early life cycles, L. jocu is able to inhabit estuaries from northern Brazil which are characterized by remarkable salinity fluctuations influences by the outflow of large rivers from Amazon, with salinity levels close to zero during rainy seasons (Barletta-Bergan et al., 2002;Barletta et al., 2005;Asp et al., 2016).On the other hand, the estuaries from northeastern Brazil are characterized by high salinity levels (Schwamborn et al., 2001;Araujo, 2006;Silva et al., 2009).
Similar ecological features are reported between L. jocu and L. griseus and L. alexandrei and L. apodus, respectively, since juveniles of L. griseus are commonly found in brackish and nearly freshwater estuarine waters as well as in hypersaline waters, an adaptation that is not observed in L. apodus (Cervigón, 1993;Claro, Lindeman, 2008).
In relation to O. chrysurus and R. aurorubens, several reports have already indicated that both taxa should be included within Lutjanus.Nevertheless, some authors considered that the morphological differences among them would be enough to their allocation into distinct genera, even though today we currently acknowledge that that such differences may reflect adaptation processes that have probably arisen several times throughout their evolution, thus representing homoplasies.
On the other hand, molecular studies have refuted the validation of both monotypic genera of Lutjanidae, even though they were based only on mitochondrial markers, which jeopardize this inference.However, a recent phylogenetic study included a single nuclear gene also demonstrated the overestimated status of Rhomboplites and Ocyurus as distinct genera.In the present study, we evaluated a higher number of nuclear and mitochondrial genes, thus providing additional support to the synonymy of these genera and Lutjanus, besides reinforcing the apparent non-monophyletic status of the latter.
Therefore, the adaptive radiation of snapper and fisiliers should be further investigated, since a single report focused this aspect (Frédérich, Santini, 2017).Studies like this could be helpful to explain and organize the diversity of Lutjanidae.
As for L. alexandrei, this is the first report, comprising the largest and most reliable molecular dataset so far, that inferred, unequivocally, the phylogenetic relationships 11 e180109 [11] of this species in relation to other lutjanids.The Brazilian snapper, endemic from the northeastern coast, and L. apodus represent sister taxa that diverged between 2.5 -6.5 Mya, as a result of the vicariance caused by the outflow of Amazon rivers between the actual range of both species.

Fig. 1 .
Fig. 1.Phylogenetic relationships in Lutjaninae species from the Western Atlantic estimated by Bayesian (BI), maximum likelihood (ML) and a Species Tre analyses inferences based on a data matrix (4.4 kb) comprising mitochondrial (16S rRNA, COI, cyt b, and ND-4) and nuclear (Rhodopsin, TMO4C4 and RAG-1) markers.The BI topology is presented with the posterior probability (BI and Species Tree) and bootstrap (ML) values.The letters in the nodes refer to the clades discussed in the text.

Fig. 2 .
Fig. 2. Bayesian inference based on the 3.0 kb dataset of mtDNA (16S, COI, and cyt b) and nuclear (RAG-1) sequences in 40 Lutjanidae species.The dating was based on the estimated minimum age for the oldest fossil reported to the family.The branch mainly discussed in the text is highlighted, revealing the origin of Lutjanus alexandrei between 2.5 -6.5 Mya.