Phylogenetic relationships of Octopus maya revealed by mtDNA sequences

Las relaciones filogeneticas entre las especies del genero Octopus no han sido establecidas por completo. En este estudio se compararon las secuencias de tres genes mitocondriales (ARN ribosomal 16S, citocromo oxidasa I y citocromo oxidasa III) de Octopus maya, una especie endemica de la peninsula de Yucatan, con las secuencias de genes homologos de otros taxones del genero Octopus, con el objetivo de elucidar el estatus filogenetico de esta especie. Los arboles de maxima verosimilitud agrupan a O. maya con pulpos que habitan en el Pacifico oriental. Debido a que la distribucion de O. maya se limita a la peninsula de Yucatan en el Atlantico occidental y que las secuencias analizadas presentaron tasas evolutivas iguales en los taxones analizados, se propone que O. maya tuvo origen durante la formacion del istmo de Panama, cuando las cuencas del Pacifico y el Atlantico se separaron. Las distancias geneticas que presenta O. maya se encuentran en el intervalo de distancias que presentan las demas especies del genero; por lo tanto, estas secuencias son caracteres que apoyan la inclusion de O. maya dentro del genero Octopus


INTRODUCTION
To date, the phylogenic relationships among the 200 odd species of the genus Octopus have not been fully established (Hochberg et al. 1992, in Barriga-Sosa et al. 1995), with less than 50 of the total number of identified species described in detail (Barriga-Sosa et al. 1995, Warnke et al. 2004).A major difficulty has been that octopuses exhibit great phenotypic plasticity and lack hard structures and so do not lend themselves to classification by traditional phenotypic methods (Pérez-Losada et al. 2002).Molecular genetic data have greatly enhanced knowledge of phylogenetic relationships within the genus (Barriga-Sosa et al. 1995, Söller et al. 2000, Oosthuizen et al. 2004, Guzik et al. 2005).
Octopus maya, a benthic octopus with direct development, is endemic to the Yucatan Peninsula (Solís et al. 1997) and is the subject of much taxonomic conjecture.In a recent allozyme-based study (30 loci), Pérez-Losada et al. (2002) reported that O. maya genotypes clustered outside the Octopus genus.Furthermore, they suggested that its origin
Octopus maya, un pulpo bentónico con desarrollo directo, es endémico de la península de Yucatán (Solís et al. 1997) y objeto de mucha conjetura taxonómica.En un trabajo reciente en el que se utilizaron aloenzimas (30 loci), Pérez-Losada et al. (2002) documentaron que los genotipos de resulted from a speciation event that followed the formation of the Isthmus of Panama, which separated the Atlantic and Pacific Oceans, as O. mimus was its closest taxon.However, the study of Pérez-Losada et al. (2002) only included samples of O. vulgaris, O. mimus, and O. maya.Furthermore, the role of selection effects of genotypes could have potentially obscured true underlying evolutionary relationships.It is noteworthy that O. maya lacks a larval stage, unlike other geographically close species such as O. vulgaris and O. mimus that have larval and paralarval stages (Warnke 1999).This might be interpreted as supporting the placement of O. maya outside the Octopus genus.The question of whether the differences in life history traits reflect the phylogenetic relationships within these octopuses also emerges.Robust phylogenetic and taxonomic conclusions for O. maya require comparisons of both additional taxa and genetic loci.
The objective of the present study was to combine all the 16S ribosomal RNA (rRNA), cytochrome oxydase I (COI), and cytochrome oxydase III (COIII) sequence information available for Octopus species with novel data for O. maya.As a result, this study is the first to adopt a global approach and consequently provide a potentially more complete picture of the evolutionary relationships within the genus Octopus, with particular attention to O. maya.

DNA extraction
Specimens of O. maya were caught off the coast of Sisal (19º40.45, 90º43.00W), Yucatan, Mexico.Sampled tissues were frozen with liquid nitrogen and stored at -80 ºC until use.For the DNA genome extraction, 20 mg of frozen muscle tissue were macerated in an extraction buffer of DNA:100 mM Tris-HCl (pH 8), 50 mM EDTA (pH 8), 500 mM NaCl, 2% SDS, and 2% β-mercaptoethanol.Next, the macerate was incubated with Proteinase K (Promega) and RNAse (Promega).A variation of the phenol:chloroform extraction procedure (Lin et al. 2001) was used for the separation and precipitation of the DNA.

DNA sequencing
Polymerase chain reactions (PCR) were carried out in a Bio-Rad iQ5 Thermocycler, following Guzik et al. (2005), modifying annealing temperature to 55 ºC and adding a final extension step at 72 ºC for 10 min.Amplicons were recovered and purified from an agarose gel with the help of purification columns (Freeze 'N Squeeze DNA Gel Extraction Spin Columns) from Bio-Rad, following the manufacturer's specifications.Cleaned amplicons were ligated into pGEM-T Easy clonation vectors made by Promega (Cat.No. A1360).The ligated vector was used to transform Escherichia coli DH5' bacteria through heat shock (42 ºC 50 to 4 ºC 2) and plated out in a selective medium.Following the selection of positive colonies for each gene, they were planted in tubes with ψ medium for 12 h at 37 ºC, 250 rpm, and the cloned plasmids were extracted following the slow alkalization method with recovery via silica matrix (Lakshmi et al. 1998).The presence of DNA inserts in the plasmids (of each gene on each distinct plasmid) was corroborated via digestion with the restriction enzyme EcoRI (Hedpeth et al. 1972).Plasmids containing DNA inserts were sent to the Institute of Biotechnology, Universidad Nacional Autónoma de México (UNAM), in Cuernavaca, Morelos, where they were sequenced.

Data analysis
All COI, COIII, and 16S rRNA sequences available for the genus Octopus in GenBank, along with those obtained from O. maya, were aligned (each gene separately) using the ClustalW program provided in the software package MEGA4 (Tamura et al. 2007).The alignment was then used as character matrix in order to construct the phylogenetic trees.Phylogenetic analysis was performed with PAUP v4.0b10 (Swofford 2003), using the maximum likelihood (ML) method.The software ModelTest v3.06 (Posada and Crandall 1998) was used to determine the best-fit model of DNA evolution among 56 different possible evolutionary models, using the Akaike Information Criterion, which was used for the ML analysis.
The ML analysis was done using heuristic searches, including gaps in sequence alignment.Evaluation of statistical confidence in nodes was based on 100 non-parametric bootstrap replicates in the ML analysis (Felsenstein 1985).
Independent trees were constructed for each gene and also for concatenated sequences for the three genes (for those taxa for which such data were available; see appendix).Opisthoteuthis species were used as outgroup and Enteroctopus dofleini (previously Octopus dofleini) was also AAGAGTTGGGCCTGCTCGGT (hacia atrás) para ARNr 16S, y CTACGTCTACAAAATGTCAGTATCA (hacia adelante) y ATTCAATGATGACGTGATATTATTC (hacia atrás) para COIII.
The equality of evolutionary rate between the analyzed species in each tree was tested using Opisthoteuthis massyae as an outgroup in Tajima's relative rate test (Tajima 1993) implemented in MEGA4 software.The divergence time for the concatenated gene tree was estimated and P-values for the chi-square test on pairs of sequences were used in the construction of linearized branches (Takezaki et al. 2004).For P-values < 0.05 the null hypothesis of equal evolutionary rates between taxa was rejected.For the molecular clock calibration, the common mutation rate for mitochondrial DNA of 2% per million years (Avise 2004) was used to identify possible selective influences.
Across all sequences the GTR+I+G model was identified as the most suitable mutation model.Table 1 summarizes the features of the phylogenetic analyses performed.Due to the respective availability of sequence data, there were different numbers of operational taxonomic units (OTUs) in the trees constructed for the various single gene regions.
The 16S rRNA gene tree had the highest number of OTUs (fig.1).The sequence of O. maya was grouped with the sequences obtained from octopuses that inhabit the East Pacific.The phylogenetic groupings were concordant with se editaron con MrEnt v2.1 (© 2005-2010, Alessandro Zuccon y Dario Zuccon).
Para todas las secuencias, el modelo GTR+I+G resultó ser el modelo de mutación más adecuado.En la tabla 1  2) illustrates that the O. maya sequence is related to octopus' sequences from the East Pacific.The COI gene tree shows that the taxa are split into two major clades.The biogeographical agreement of the COI tree groupings seems to be biased because there are only two taxa (O.maya and O. joubini) that do not correspond to the Pacific basin.Interestingly, although both octopuses inhabit the Caribbean and the Gulf of Mexico, they were not grouped together in the tree.Both species are more closely related to octopuses from different major clades.
The COIII gene tree (fig.3) includes more species from the East Pacific than the 16S rRNA and COI gene trees, and clearly supports the evolutionary relationship of O. maya with East Pacific octopuses.It also supports the split of taxa into two major clades with similar sorting to that of the COI gene tree.The COIII gene tree shows groups with clear biogeographical agreement, regarding the monophyletic clade composed of Australian octopuses as well as American octopuses.
Table 2 shows the genetic distances between taxa included in the concatenated gene tree, as well as the relative rate test (Tajima 1993) for each pair of sequences.Octopus californicus and E. dofleini show very short genetic distances between them (0.098) but large distances with respect to the rest of the taxa (22.6-28.3).Also, these two North Pacific octopuses have similar evolutionary rates for the genes analyzed, but both show significant rate difference with O. vulgaris (P-values = 0.02259 and 0.01669), so they were excluded from the linearized tree, where the divergence time between taxa was estimated.
Figure 4a shows the concatenated gene tree with an early divergence of the O. californicus-E.dofleini pair with respect to the rest of the taxa.The bootstrap support values in that tree are remarkably high due to the highest number of informative sites, so it is the best supported tree.
Figure 4b shows that the diversification of the octopuses that present equal evolutionary rates across all three genes occurred between 5 and 2.5 million years ago.

Biogeographical support of the trees and the origin of O. maya
All constructed trees show broad-scale agreement between the groupings and the geographical distribution of the taxa, and the evolutionary relationship between O. maya and the East Pacific octopuses was consistent in the whole analysis.
The COI gene tree suggests that allopatric speciation processes have occurred independently but in the same way se resumen las características de los análisis filogenéticos realizados.Debido a la disponibilidad respectiva de los datos de secuencia, hubo diferentes números de unidades taxonómicas operativas (UTO) en los árboles construidos para las diversas regiones de un gen.
La figura 4b muestra que la diversificación de los pulpos que presentan tasas evolutivas iguales para los tres genes sucedió hace 5 a 2.5 milliones de años.Panama Isthmus) triggered several speciation events within the genus.The divergence time estimated for the sequences in the concatenated gene tree (fig.4b) was between 5 and 2.5 million years ago.That time corresponds to the formation of the Isthmus of Panama (Stanley 2005), supporting the possibility that this geological event initiated an important process of speciation of octopuses not only in the American continent but also worldwide.The estimated divergence time also corresponds to that determined by Pérez-Losada et al. (2002).The COIII tree also suggests that O. maya shares ancestry with the Californian octopuses O. bimaculoides and O. bimaculatus and with the South American octopuses O. mimus, which is distributed from Peru to Chile (Pérez-Losada et al. 2002), and O. oculifer, which is endemic to the Galapagos Islands.The common ancestor of these species possibly had a life cycle with a larval phase, although both O. maya and O. bimaculoides follow a life cycle marked by direct development.The presence of larval stages during the life cycle has been determined to be the ancestral state and has not presented regressions along the branches in other octopus phylogenies (Guzik 2004).We propose that all these octopuses (O. maya, O. vulgaris, O. tetricus, O. bimaculoides, O. bimaculatus, O. mimus, and O. oculifer) compose a monophyletic group that was diversified during the formation of the Isthmus of Panama.
The topology obtained from the phylogenetic tree of the COIII gene was identical to that obtained by Barriga-Sosa et al. (1995) in terms of the taxa common to both studies.The monophyletic clades in figure 3 that correspond to the Australian and West Pacific octopuses are very similar and include the same species that were analyzed by Guzik et al.
( (2005), though in that study cytochrome b and a nuclear gene (elongation factor-1) were analyzed along with COIII.

Possible implications of the results in Octopus systematics and evolution
Table 2 shows that the O. maya genetic distances (11-19.9%)are in the same range as those of the Octopus species analyzed, with the exception of O. californicus that shows larger distances (24.1-27.6%).These results support the inclusion of O. maya in the Octopus genus.This conclusion is different to that reached by Pérez-Losada et al. (2002), presentan regresiones en las ramas de otras filogenias de pulpo (Guzik 2004) La topología obtenida del árbol filogenético del gen COIII es idéntica a la obtenido por Barriga-Sosa et al. (1995) en cuanto a los taxones comunes a ambos estudios.Los clados monofiléticos en la figura 3 que corresponden a los pulpos de Australia y el Pacífico occidental son muy similares e incluyen las mismas especies analizadas por Guzik et al.Bootstrap support values changed after taxa with different evolutionary rates were excluded from the analysis.
who suggested that the genetic distances of O. maya could place it out of the Octopus genus.On the other hand, we propose that the current status of O. californicus should be reviewed as it shows genetic distances out of the range of the Octopus genus, showing shorter genetic distances as well as similar evolutionary rates as E. dofleini.
It is necessary to put limitations on the interpretation of real evolutionary history.While trees based on DNA sequences are very useful tools for understanding the evolution of a group given the great phenotypic variability and a limited fossil record (Strugnell et al. 2005), it is first necessary to sequence a great quantity of genes and genomes, generating the greatest quantity of characters available for resolving the phylogeny.
Appendix.GenBank accession numbers for Octopus sequences used.Apéndice.Número de acceso en GenBank de las secuencias de Octopus utilizadas.
and Atlantic basins, as illustrated by the O. maya and O. joubini sequences.Both species have similar geographical distribution and development and their sequences were also related to Pacific octopuses, but were sorted in different clades.According to the COI tree, species with the same development are not necessarily closely related.The evolutionary relationship between Pacific and Atlantic sequences, even for different clades, suggests that a single oceanographic incident (possibly the lifting of theDISCUSIÓN Apoyo biogeográfico de los árboles y el origen de O. mayaTodos los árboles construidos muestran una amplia congruencia entre las agrupaciones y la distribución geográfica de los taxones, y la relación evolutiva entre O. maya y los pulpos del Pacífico oriental fue consistente durante todo el análisis.

Figure 4 .
Figure 4. Evolutionary relationships of the COI, COIII, and 16S rRNA concatenated sequences for the Octopus genus: (a) genetic distance tree and (b) linearized tree.The tree produced by the concatenated sequences provides better support than the individual sequences.The diversification of the concatenated sequences in Octopus species is estimated to have started between 5 and 2.5 million years ago (MYA).Bootstrap support values changed after taxa with different evolutionary rates were excluded from the analysis.