Mitogenomic phylogenetic analyses of Leptogorgia virgulata and Leptogorgia hebes (Anthozoa: Octocorallia) from the Gulf of Mexico provides insight on Gorgoniidae divergence between Pacific and Atlantic lineages.

The use of genetics in recent years has brought to light the need to reevaluate the classification of many gorgonian octocorals. This study focuses on two Leptogorgia species-Leptogorgia virgulata and Leptogorgia hebes-from the northwestern Gulf of Mexico (GOM). We target complete mitochondrial genomes and mtMutS sequences, and integrate this data with previous genetic research of gorgonian corals to resolve phylogenetic relationships and estimate divergence times. This study contributes the first complete mitochondrial genomes for L. ptogorgia virgulata and L. hebes. Our resulting phylogenies stress the need to redefine the taxonomy of the genus Leptogorgia in its entirety. The fossil-calibrated divergence times for Eastern Pacific and Western Atlantic Leptogorgia species based on complete mitochondrial genomes shows that the use of multiple genes results in estimates of more recent speciation events than previous research based on single genes. These more recent divergence times are in agreement with geologic data pertaining to the formation of the Isthmus of Panama.

INTRODUCTION

There are 54 valid species in the genus Leptogorgia belonging to the family Gorgoniidae (Milne‐Edwards & Haime, 1857). They are classified as soft corals due to their lack of a protective calcium carbonate exoskeleton. Instead, for support and protection, they rely on small, calcitic structures called sclerites (O'Neal & Pawlik, 2002), from which their white, translucent polyps protrude, and they range in color from yellow to orange to red to purple (White & Strychar, 2010). Leptogorgia and other octocorals provide habitat heterogeneity and therefore allow for large aggregations of diverse fauna (Quattrini et al., 2014). Greater habitat complexity has been shown to be significantly correlated with higher red snapper abundance, an economically important fish species in the Gulf of Mexico and Western Atlantic Ocean (Szedlmayer, 2007). There are nine species of Leptogorgia in the Gulf of Mexico, including Leptogorgia hebes and Leptogorgia virgulata (Cairns & Bayer, 2009). They are found at depths ranging from 2 to 309 m, with the depth range of L. hebes ranging from 9 to 37 m and that of L. virgulata from 3 to 82 m (Cairns & Bayer, 2009; Williamson, Strychar, & Withers, 2011). Leptogorgia hebes and L. virgulata reach reproductive maturity within 2 years, and both are broadcast spawners, releasing eggs, and sperm into the water column (Beasley, Dardeau, & Schroeder, 2003; Gotelli, 1991). The larvae in L. virgulata can spend 3–20 days in the water column before settlement (Gotelli, 1991). The duration of the larval stage for L. hebes is unknown. Both L. hebes and L. virgulata have been successful at colonizing artificial structures in the Gulf of Mexico, including jetties within the subtidal zone (Williamson et al., 2011). A strong holdfast and a rigid, yet flexible skeleton, allows these two species to colonize habitats with swift currents and wave action such as that found in jetties (Williamson et al., 2011). These life history characteristics of L. hebes and L. virgulata, which include relatively fast maturation, broadcast spawning, long survival of larval stages, and adaptations for successful establishment in high energy environments, demonstrate the high potential for dispersal and colonization of new regions of these species.

Old, incomplete, or damaged records in addition to a lack of easily identifiable morphological traits among species make gorgonians particularly difficult to classify (Sánchez, 2007). For example, the genus Leptogorgia was initially split into two genera—Leptogorgia and Lophogorgia—by Milne‐Edwards and Haime (1857). Species in the Leptogorgia genus are described as having disk‐spindles in the outer coenenchyme, while Lophogorgia species have spindles more closely resembling flat rods and were described mostly in the Eastern Pacific, Western Atlantic, Caribbean, and along the eastern and southern coasts of Africa (Bayer, 1961). However, in 1988 these morphological distinctions were questioned and the two genera were united into one as Leptogorgia (Grasshoff, 1988). In 2017, Poliseno et al. conducted a phylogenetic study of Leptogorgia, using specimens from a wide geographical area, including the eastern and Western Atlantic, the Eastern Pacific, and the Mediterranean. They reconstructed two phylogenies, one based on complete mitochondrial genomes and the other based on a partial fragment of the mitochondrial MutS gene (mtMutS). While their phylogeny based on complete mitochondrial genomes only has eleven species of the family Gorgoniidae, including six species of Leptogorgia, the one based on the single mtMutS gene includes 109 species, providing greater taxonomic resolution. In their study, Poliseno et al. (2017) also estimate divergence times with a fossil calibration based on the oldest known fossil of Eunicella, dating back to 28.4 Ma (Kocurko & Kocurko, 1992) using the partial mtMutS gene. Based on their results, Poliseno et al. (2017) call for a global taxonomic revision of the present‐day Leptogorgia genus. They conclude from the mtMutS phylogeny that the genus Lophogorgia should be resurrected for all South African Leptogorgia species, which form an old clade within the Gorgoniidae, sister to Leptogorgia species from the eastern coast of Africa and the Mediterranean. They show that these Eastern Atlantic Leptogorgia species diverged from Western Atlantic species in the late Cretaceous, about 65 Ma, while the divergence between the Western Atlantic and Eastern Pacific species occurs more recently, between 28 and 23 Ma. These observations not only raise the question of taxonomic placement and nomenclature for Eastern Pacific and Western Atlantic species, but also suggest a divergence time between these lineages that dates back to the very early stages of emergence of the Isthmus of Panama (Bacon et al., 2015). This scenario is unexpected since Leptogorgia are shallow water species and significant exchange of seawater between the two basins likely occurred until ~10–15 Ma when the final stages of the closure of the Central American Seaway (CAS) started, with shallow water still connecting these two oceans until 3.5–4.2 Ma with the final rise of the Isthmus of Panama (e.g., Bacon et al., 2015; Montes et al., 2015; O'Dea et al., 2016).

Our study focuses on two species of Leptogorgia from the Gulf of Mexico, L. hebes and L. virgulata. We have two main goals. The first is to determine the taxonomic position of L. hebes (formerly classified in the genus Lophogorgia by Bayer, 1961) and of L. virgulata. The analyses by Poliseno et al. (2017) did not include complete mitochondrial genomes for these two species and their phylogeny based on the partial mtMutS gene leaves the phylogenetic position of both L. hebes and L. virgulata weakly supported. Therefore, in our study, we analyze both complete mitochondrial genomes and the mtMutS gene. Although mitochondrial genomes have been shown to be problematic for phylogenetic reconstruction of scleractinian corals due to the presence of substitution saturation and long branch attraction (i.e., Kitahara et al., 2014), it is only an issue within the Hexacorallia and it does not affect the Octocorallia, such as the gorgonian corals in our study (Figueroa & Baco, 2015). Complete mitochondrial genomes have been demonstrated to provide robust and well‐supported phylogenies for Octocorallia (e.g., Figueroa & Baco, 2014; Figueroa & Baco, 2015; Kayal, Roure, Philippe, Collins, & Lavrov, 2013; Poliseno et al., 2017), while the use of single mitochondrial genes has been demonstrated to result in incongruent largely unresolved trees across a wide range of taxa (Havird & Santos, 2014; Knaus, Cronn, Liston, Pilgrim, & Schwartz, 2011; Luo et al., 2011; Nadimi, Daubois, & Hijri, 2016; Pacheco et al., 2011; Rohland et al., 2007; Urantowka, Kroczak, & Mackiewicz, 2017; Wang et al., 2017; Willerslev et al., 2009). Therefore, we expect that the taxonomic position of L. hebes and L. virgulata will be fully resolved by reconstructing their phylogeny using mitochondrial genomes.

Our second goal is to estimate divergence times of Eastern Pacific and Western Atlantic Leptogorgia species. Since previous research has shown that fossil‐calibrated phylogenetic reconstruction based on single mitochondrial genes results in an overestimation of divergence times (Duchêne, Archer, Vilstrup, Caballero, & Morin, 2011; McCormack, Heled, Delaney, Peterson, & Knowles, 2011), we will base our estimates of diversification times between Eastern Pacific and Western Atlantic lineages of Leptogorgia by targeting complete mitochondrial genomes. We reconstruct a fossil‐calibrated phylogenetic tree for Leptogorgia species based con complete mitochondrial genomes and using Eunicella as an outgroup. We use a fossil calibration point of 28.4 Ma based on the stratigraphy and dating of the Red Bluff Formation in Mississippi where the oldest fossils of Eunicella have been recovered (Cushing, Boswell, & Hosman, 1964; Demchuk & Gary, 2009; Kocurko & Kocurko, 1992; Prothero, Ivany, & Nesbitt, 2003; Tew, 1992). Among Octocorallia, skeletal diversity, such as morphology of sclerites, is a key character for taxonomic identification (Goffredo & Dubinsky, 2016). Sclerites with a balloon club shape are a distinguishing characteristic that is unique to the genus Eunicella (Goffredo & Dubinsky, 2016; Kocurko & Kocurko, 1992). Fossil sclerites with balloon club shape have been found in the Red Bluff Formation in Mississippi and have been clearly attributed to Eunicella (Kocurko & Kocurko, 1992). Stratigraphy of the Red Bluff Formation and dating of this layer within the Oligocene (23–34 Ma) has been intensely studied (i.e., Cushing et al., 1964; Demchuk & Gary, 2009; Hosman, 1996; Prothero et al., 2003; Tew, 1992).

The timeline proposed by Poliseno et al. (2017) for the divergence between Eastern Pacific and Western Atlantic Leptogorgia species coincides with evidence that a land bridge between North and South America began to emerge between 23 and 25 Ma when the Panama Arc collided with South America (Bacon et al., 2015). However, despite this initial emergence and given the life history characteristics of shallow water Leptogorgia species such as L. hebes and L. virgulata that enhance dispersal and colonization (Beasley et al., 2003; Cairns & Bayer, 2009; Gotelli, 1988, 1991; Williamson et al., 2011), gene flow is likely to have continued between the Western Atlantic and Eastern Pacific until full closure of the Central American Seaway (Bacon et al., 2015; Cowman & Bellwood, 2013; Lessios, 2008; Thacker, 2017).. Therefore, we hypothesize that the divergence times of Eastern Pacific and Western Atlantic Leptogorgia lineages to be younger than previously suggested (Poliseno et al., 2017) with the majority of speciation events occurring after 10 Ma when significant seawater exchange between the Pacific and Atlantic Ocean ceased (i.e., Bacon et al., 2015; Montes et al., 2015; O'Dea et al., 2016).

MATERIALS AND METHODS

Study sites and sample collection

Six sites in the Gulf of Mexico off the coast of the United States in South Padre Island, Texas, were sampled for 24 Leptogorgia specimens (seven L. hebes and 17 L. virgulata) by divers collecting coral fragments between June 2014 and July 2017 (Table 1). Once collected, samples were preserved in ethanol and stored at 0°C. Voucher specimens are deposited and curated at the University of Texas Rio Grande Valley's Coastal Studies Laboratory and are available upon request under GenBank accession numbers MK0301586–MK0301592 for specimens of L. virgulata and MN052675–MN052677 for specimens of L. hebes.

Table 1

Leptogorgia sp. found at all sites and site types with respective dates, coordinates, and depths

Location	Date	Lat	Lon	Depth (m)	# of samples/species collected
Port Isabel Reef	6/3/2014	25.9684	−97.0669	22	1 L. hebes; 1 L. virgulata
Port Mansfield Liberty Ship	12/12/2014	26.4296	−97.0241	24	1 L. hebes; 1 L. virgulata
Jack up Rigs/East Bank	7/7/2016	26.1021	−96.9377	32	1 L. virgulata
Port Mansfield Liberty Ship	9/2/2016	26.4296	−97.0241	24	1 L. virgulata
Port Isabel Reef	9/15/2016	25.9684	−97.0669	22	1 L. hebes; 1 L. virgulata
Port Mansfield Liberty Ship	3/20/2017	26.4296	−97.0241	24	1 L. hebes
MU 726 A	6/8/2017	27.8146	−96.7622	24	4 L. virgulata
Texas Clipper	7/8/2017	26.1903	−96.8614	15–41	2 L. virgulata
SPI Jetty	7/30/2017	26.0674	−97.1504	5	3 L. hebes; 6 L. virgulata

DNA Extraction and PCR

Three to five individual polyps were picked off from each coral sample, depending on the size and quality of preservation of the coral fragment. Polyps were visually inspected under a stereo microscope and picked off the coral stalk using forceps. Forceps were sterilized in between each sample using 100% bleach and 100% ethanol. If individual polyps were difficult to distinguish, an ~0.5 cm long piece was broken off of the coral fragment. The PureLink Genomic DNA Mini Kit (Thermo Fisher Scientific) was used to extract DNA from each sample following the manufacturer's standard protocol. Prior to extraction, coral polyps were rehydrated for 1–2 hr in molecular grade water and then digested for at least 5 hr. The final DNA product was eluted two times for maximum yield. The elution buffer was heated to 55°C prior to use, and 60 µl of were used for both elutions. The concentration of the extracted DNA was measured using a Qubit fluorometer (Life Technologies Inc.).

Polymerase Chain Reaction (PCR) amplification was performed on 0.1–5.0 ng template DNA from 24 samples in order to target the mtMutS gene with forward primer ND42599F (GCCATTATGGTTAACTATTAC; France & Hoover, 2002) and reverse primer Mut3458R (TSGAGCAAAAGCCACTCC; Sanchez, McFadden, France, & Lasker, 2003). The PCR mix consisted of the following in 25 μl total volume: 16.05 µl nuclease free water, 2.5 Invitrogen's 10X PCR Rxn Buffer, 1.25 µl Invitrogen's 50 mM MgCl, 2.0 µl of 10 mM dNTP, 1.0 µl of 10 mM forward primer (ND42599F), 1.0 µl of 10 mM reverse primer (Mut3458R), 0.2 µl Thermo Fisher's Invitrogen Platinum TAQ DNA Polymerase, and 1.0 µl DNA. Samples were then amplified in an Eppendorf Mastercycler pro thermocycler using the following parameters: 94°C for 2 min, 35 cycles at 94°C for 1 min, 50°C for 1 min, 72°C for 1 min, and a final step at 72°C for 5 min. The resulting product was visualized by gel electrophoresis on an ultraviolet light transilluminator to assess DNA length and quality. Once all samples yielded successful amplification, the final PCR product was then purified with the Invitrogen PureLink PCR Purification Kit, following the manufacturer's procedure. The primers and purified PCR products were then sent to Eurofins Genomics for sequencing of forward and reverse strands.

The genomic DNA extraction from each specimen was visualized after gel electrophoresis with an ultraviolet light transilluminator. Genomic DNA of high molecular weight with minimal degradation was identified by looking for high concentrations above 5,000 bp, with minimal streaking below this size. Based on these observations, the mitochondrial genome of ten specimens with the highest quality of genomic DNA was targeted using next generation sequencing technology. The genomic DNA extraction of these ten specimens was used to prepare an indexed library following standard procedures with the Nextera X2 kit. These 10 libraries, along with 86 libraries from other projects, were multiplexed and sequenced on a 100 bp paired‐end lane of Illumina HiSeq 2500 at Harvard's Biopolymers facility. The sequences were de‐multiplexed according to their indices.

Sequence assembly and alignment

For each specimen, the sequences for the forward and reverse strands were assembled with the software CLC Workbench 7.9.1 (CLC Bio) using the settings: minimum aligned read length = 500 bp, alignment stringency = high, conflicts = ambiguity nucleotides, trim sequence ends and trim using quality scores limit = 0.05. A cutoff was used were only bases with Phred scores of 20 or more were kept. A consensus sequence was generated from each assembly. Qiagen's CLC Workbench 7.9.1 was used to align the mtMutS sequences. The mtMutS sequences were aligned using Qiagen's CLC Main Workbench 7 software and include 24 sequences from this study, the 114 sequences examined in Poliseno et al. (2017) and 43 novel sequences available in GenBank for a total of 182 sequences (Table 2). The alignment was visually inspected for errors and inconsistencies. The final mtMutS alignment was 766 bp in length.

Table 2

All 182 Leptogorgia mtMutS sequences incorporated into the mtMutS phylogeny with corresponding GenBank accession numbers

Species	Accession #	Species	Accession #	Species	Accession #	Species	Accession #
Leptogorgia virgulata a	MN159153	Leptogorgia alba b	KX721205	Leptogorgia obscura b	KX721210	Leptogorgia taboguilla c	LT221102
Leptogorgia virgulata a	MN159154	Leptogorgia alba b	KX721203	Leptogorgia piccola c	AY268444	Leptogorgia tricorata c	LT221111
Leptogorgia virgulata a	MN159155	Leptogorgia alba b	KX721202	Leptogorgia pulcherrima c	KY683795	Leptogorgia tricorata c	LT221109
Leptogorgia virgulata a	MN159156	Leptogorgia alba b	KX721201	Leptogorgia pulcherrima c	KY683794	Leptogorgia viminalis c	KY683797
Leptogorgia virgulata a	MN159157	Leptogorgia alba b	KX721195	Leptogorgia pulcherrima c	KY683793	Leptogorgia viminalis c	KY683796
Leptogorgia virgulata a	MN159158	Leptogorgia alba b	KY559410	Leptogorgia pulcherrima c	AY268443	Leptogorgia violacea c	AY268448
Leptogorgia virgulata a	MN159159	Leptogorgia alba b	KX767314	Leptogorgia pumila b	KX767312	Leptogorgia violetta c	AY268446
Leptogorgia virgulata a	MN159160	Leptogorgia alba c	HG917036	Leptogorgia pumila c	LT221116	Leptogorgia virgulata c	AY268458
Leptogorgia virgulata a	MN159161	Leptogorgia alba c	HG917035	Leptogorgia punicea c	AY268449	Leptogorgia virgulata c	AY126418
Leptogorgia virgulata a	MN159162	Leptogorgia alba c	HG917034	Leptogorgia ramulus c	AY268451	Pacifigorgia bayeri c	HG917044
Leptogorgia virgulata a	MN159163	Leptogorgia alba c	AY268452	Leptogorgia ramulus b	KX767322	Pacifigorgia cairnsi c	KY559409
Leptogorgia virgulata a	MN159164	Leptogorgia alba c	LT221108	Leptogorgia regis c	LT221101	Pacifigorgia cairnsi c	HG917041
Leptogorgia virgulata a	MN159165	Leptogorgia alba c	LT221113	Leptogorgia regis c	LT221100	Pacifigorgia catedralensis c	HG917019
Leptogorgia virgulata a	MN159166	Leptogorgia barnardi c	KY236043	Leptogorgia regis c	LT221099	Pacifigorgia cf. cairnsi c	HG917046
Leptogorgia virgulata a	MN159167	Leptogorgia capverdensis c	KY553145	Leptogorgia regis c	LT221098	Pacifigorgia cf. cairnsi c	HG917021
Leptogorgia virgulata a	MN159168	Leptogorgia cf.palma c	KY559406	Leptogorgia rigida c	GQ342496	Pacifigorgia exilis b	KX351871
Leptogorgia virgulata a	MN159169	Leptogorgia cf. gilchristi c	KY236042	Leptogorgia rubra b	KX767323	Pacifigorgia firma b	KX351872
Leptogorgia hebes a	MN159170	Leptogorgia cf. palma c	KY236030	Leptogorgia sarmentosa c	KY559411	Pacifigorgia firma c	HG917022
Leptogorgia hebes a	MN159171	Leptogorgia cf. palma c	KY236031	Leptogorgia sp.b	KX767315	Pacifigorgia irene b	KX351873
Leptogorgia hebes a	MN159172	Leptogorgia chilensis c	AY268460	Leptogorgia sp.b	KX721204	Pacifigorgia irene c	HG917024
Leptogorgia hebes a	MN159173	Leptogorgia chilensis c	JN866554	Leptogorgia sp.b	KY559412	Pacifigorgia machalilla b	KX351876
Leptogorgia hebes a	MN159174	Leptogorgia cofrini c	HG917037	Leptogorgia sp. 2 c	KY236033	Pacifigorgia machalilla b	KX351874
Leptogorgia hebes a	MN159175	Leptogorgia cofrini c	HG917040	Leptogorgia sp. 2 c	KY236032	Pacifigorgia machalilla b	KX351875
Leptogorgia hebes a	MN159176	Leptogorgia cofrini c	HG917039	Leptogorgia sp.c	KY683791	Pacifigorgia media c	GQ342497
Acanthogorgia spc	AY268461	Leptogorgia cofrini c	HG917038	Leptogorgia sp.c	LT221114	Pacifigorgia rubicunda c	HG917027
Antillogorgia acerosa b	JX152763	Leptogorgia cortesi c	LT221105	Leptogorgia sp.c	LT221115	Pacifigorgia sculpta b	KX351877
Antillogorgia sp. 1 b	JX152764	Leptogorgia cuspidata b	KX767318	Leptogorgia sp.c	LT221106	Pacifigorgia senta c	LT221107
Antillogorgia sp. 2 b	JX152765	Leptogorgia cuspidata c	AY268450	Leptogorgia sp. 3 c	KY236035	Pacifigorgia smithsoniana c	HG917023
Eugorgia ampla b	KX767316	Leptogorgia cuspidata c	HG917047	Leptogorgia sp. 3 c	KY236034	Pacifigorgia stenobrochis c	HG917018
Eugorgia daniana c	HG917048	Leptogorgia dichotoma c	AY268445	Leptogorgia sp. 4 c	KY236037	Pacifigorgia stenobrochis c	AY126420
Eugorgia daniana c	LT221110	Leptogorgia diffusa b	KX767319	Leptogorgia sp. 4 c	KY236036	Phyllogorgia dilatata c	AY126428
Eugorgia multifida c	GQ342494	Leptogorgia diffusa b	KX721209	Leptogorgia sp. 5 c	KY236039	Pseudopterogorgia acerosa c	AY126421
Eugorgia mutabilis c	KY559405	Leptogorgia flexilis b	KX767326	Leptogorgia sp. 5 c	KY236038	Pseudopterogorgia americana c	AY126423
Eugorgia mutabilis c	LT221112	Leptogorgia flexilis b	KX767329	Leptogorgia sp. 6 c	KY236041	Pseudopterogorgia australiensis c	AY268442
Eugorgia rubens c	JN866557	Leptogorgia flexilis b	KX767328	Leptogorgia sp. 6 c	KY236040	Pseudopterogorgia bipinnata c	DQ640646
Eugorgia siedenburgae c	LT221097	Leptogorgia flexilis b	KX767327	Leptogorgia sp.b	KX721200	Pseudopterogorgia elisabethae c	AY126422
Eugorgia siedenburgae c	LT221096	Leptogorgia flexilis b	KX767325	Leptogorgia sp.b	KX721199	Pseudopterogorgia fredericki b	JX152766
Eugorgia siedenburgae c	LT221094	Leptogorgia flexilis b	KX767473	Leptogorgia sp.b	KX721197	Pseudopterogorgia rubrotincta b	JX152768
Eunicea mammosa b	JX152767	Leptogorgia flexilis b	KX767474	Leptogorgia sp.b	KX721196	Eunicella cavolinii c	JQ397290
Eunicella albicans c	KY559407	Leptogorgia gaini c	KY559404	Leptogorgia sp.b	KX721198	Eunicella singularis c	JQ397296
Eunicella cavolinii c	KY559408	Leptogorgia gracilis c	AY268454	Leptogorgia styx c	AY268453	Eunicella sp.c	JQ397310
Gorgonia flabellum c	AY126427	Leptogorgia hebes c	AY268459	Leptogorgia sylvanae c	KY683792	Eunicella verrucosa c	JQ397300
Gorgonia mariae c	AY126426	Leptogorgia labiata c	AY268447	Leptogorgia sylvanae c	KY683798
Gorgonia ventalina c	AY126425	Leptogorgia mariarosae b	KX721193	Leptogorgia taboguilla c	LT221103
Leptogorgia alba b	KX767324	Leptogorgia obscura b	KX767321	Leptogorgia taboguilla c	LT221093
Leptogorgia alba b	KX721206	Leptogorgia obscura b	KX767320	Leptogorgia taboguilla c	LT221104

Sequences from this study.

New sequences from GenBank.

Sequences used by Poliseno et al. (2017).

The Illumina sequence reads were assembled using the software CLC Genomics Workbench 11. Default settings were used with reads mapped back to contigs (mismatch cost = 2, insertion cost = 3, deletion cost = 3, length fraction = 0.5, similarity fraction = 0.8). The sequences obtained from the assemblies included the full mitochondrial genome for each specimen with an average read coverage of over 100 and a minimum coverage of 35. The assembled genomes were annotated using Qiagen CLC Genomics Workbench 11 software using previously published Leptogorgia mt genomes as references (Table 3). The ten mitochondrial genomes obtained were analyzed along with eleven mitochondrial genomes available in GenBank (Table 3). Individual genes and RNAs were extracted and aligned separately using MUSCLE v3.8 (Edgar, 2004) with default parameters. The alignments were visually inspected for consistency. The resulting alignments were then concatenated for phylogenetic analyses and deposited in the online database http://figshare.com under https://doi.org/10.6084/m9.figshare.10052030.

Table 3

All 21 gorgonian mitochondrial genomes and their corresponding GenBank accession number

Species	Size (bp)	GenBank Accession #
Leptogorgia virgulata a	18,845	MK301586
Leptogorgia virgulata a	18,845	MK301587
Leptogorgia virgulata a	18,824	MK301588
Leptogorgia virgulata a	18,845	MK301589
Leptogorgia virgulata a	18,845	MK301590
Leptogorgia virgulata a	18,824	MK301591
Leptogorgia virgulata a	18,845	MK301592
Leptogorgia hebes a	19,247	MN052675
Leptogorgia hebes a	19,247	MN052676
Leptogorgia hebes a	19,247	MN052677
Pseudopterogorgia bipinnata	18,733	DQ640646
Leptogorgia capverdensis	18,722	KY553145
Leptogorgia gaini	19,682	KY559404
Eugorgia mutabilis	19,157	KY559405
Leptogorgia cf. palma	18,731	KY559406
Eunicella albicans	19,175	KY559407
Eunicella cavolinii	19,316	KY559408
Pacifigorgia cairnsi	19,156	KY559409
Leptogorgia alba	18,848	KY559410
Leptogorgia sarmentosa	18,722	KY559411
Leptogorgia sp.	18,849	KY559412

The 10 novel mt genomes sequenced in this study.

Phylogenetic analyses

Both mtMutS and complete mt genome alignments were used in phylogenetic analyses using maximum likelihood (ML) and Bayesian methods. The model of evolution and partitioning scheme was determined by PartitionFinder v1.1.1 (Lanfear, Calcott, Kainer, Mayer, & Stamatakis, 2014) using linked branches and the Akaike information criterion (AIC). The RAxML v8.0.0 program (Stamatakis, 2017) was used to conduct the ML analyses and Mr. Bayes 3.1 (Ronquist & Huelsenbeck, 2003) was used for the Bayesian analyses. Data blocks were created for mtMutS based on codon position (Table 4). PartitionFinder selected GTR + G as the best evolutionary model for three partitions: (a) mtMutS1; (b) mtMutS2; and (c) mtMutS3.

Table 4

Data block definitions for partition analysis

Region	Codon positions
	1	2	3
(A)
Atp6	1–708	2–708	3–708
Atp8	709–924	710–924	711–924
Cox1	925–2,550	926–2,550	927–2,550
Cox2	2,551–3,312	2,552–3,312	2,553–3,312
Cox3	3,313–4,098	3,314–4,098	3,315–4,098
Cytb	4,099–5,273	4,100–5,273	4,101–5,273
MutS	5,274–8,231	5,275–8,231	5,276–8,231
Nad1	8,232–9,203	8,233–9,203	8,234–9,203
Nad2	9,204–10,361	9,205–10,361	9,206–10,361
Nad3	10,362–10,734	10,363–10,734	10,364–10,734
Nad4	10,735–12,183	10,736–12,183	10,737–12,183
Nad4L	12,184–12,477	12,185–12,477	12,186–12,477
Nad5	12,478–14,320	12,479–14,320	12,480–14,320
Nad6	14,321–14,878	14,322–14,878	14,323–14,878
rRNA (12s)	14,879–15,807
rRNA (16s)	15,808–17,999
(B)
MutS	1–766	2–766	3–766

(A) Mitochondrial genome concatenated alignment including 14 protein‐coding genes and 2 RNAS. (B) mtMutS alignment.

For the mitochondrial genome analyses, data blocks were created based on codon positions for all 14 protein‐coding genes (Cox1, Nad1, CytB, Nad6, Nad3, Nad4L,mtMutS, Nad2, Nad5, Nad4, Cox3, Atp6, Atp8, and Cox2) and two ribosomal RNAs (Table 4). For the ML analysis, PartitionFinder selected General Time Reversible plus Gamma (GTR + G) as the best evolutionary model for 11 partition subsets and GTR + I+G for three subsets (Table 5). For the Bayesian analysis, the data were partitioned into 16 subsets. PartitionFinder selected GTR + I as the best model for two subsets, F81 for one subset, GTR + G for four subsets, GTR + I+G for two subsets, GTR for one subset, HKY for two subsets, HKY + G for two subsets, and HKY + I+G for one subset (Table 5).

Table 5

Partition scheme for the concatenated mitochondrial genome alignment for ML and Bayesian analyses

Subset	Best model	# of Sites	Maximum likelihood partitions
1	GTR + G	808	Nad21, Atp61, Nad61
2	GTR + I+G	2,333	Nad62, Nad42, Cytb2, Nad52, Atp62, Nad4L2, Nad12
3	GTR + G	1,423	Cox13, Cytb3, Cox23, Atp63
4	GTR + I+G	3,193	Atp81, 16s rRNA, 12s rRNA
5	GTR + G	582	Nad22, Atp82, Nad32
6	GTR + G	782	Nad23, Nad13, Atp83
7	GTR + G	1,605	Cox21, Nad4L1, Nad31, Cox31, Cox11, Nad11
8	GTR + G	1,058	Cox22, Cox32, Cox12
9	GTR + G	546	Nad63, Nad4L3, Cox33
10	GTR + I+G	1,490	Nad51, Cytb1, Nad41
11	GTR + G	1,110	MutS1, Nad33
12	GTR + G	986	MutS2
13	GTR + G	1,600	Nad53, MutS3
14	GTR + G	483	Nad43

Superscript numbers indicate codon position 1, 2, or 3.

The best maximum likelihood tree was reconstructed with RAxML for both, the mtMutS alignment and the mt genome alignment, using bootstrap values from 10,000 replicates. Note that in RAxML partitions cannot be analyzed with different evolutionary models and one model must be used for all partitions. Therefore, the mt genome alignment was analyzed under a GTR + G model given that PartitionFinder selected this as the best model for 11 of the 14 partitions. Phylogenetic trees were also reconstructed for both by Bayesian methods. Using Mr. Bayes, four chains were carried out for 1,100,000 generations, sampling every 200th generation. After inspecting the trace files generated by the Bayesian Markov Chain Monte Carlo (MCMC) runs, the initial 100,000 of sampled generations were omitted prior to building the consensus tree. Both ML and Bayesian phylogenies were rooted with sequences of species of Eunicella downloaded from Genbank (Tables 2 and 3).

Divergence time estimates were performed by Bayesian analyses using full mitochondrial genomes only, with the software BEAST 2.3.2 (Bouckaert et al., 2019). The alignment was partitioned as specified above for the Bayesian phylogenetic reconstruction (Table 3). An uncorrelated log‐normal relaxed clock model was used along with the calibrated yule speciation model. The tree was calibrated based on the earliest fossil evidence for Eunicella (Kocurko & Kocurko, 1992) with a date of origination set to 28.4 Ma (mean = 1 and standard deviation = 1). One chain was carried out for 10,000,000 generations, sampling every 1,000th generation. After inspecting the trace files generated by the Bayesian Markov Chain Monte Carlo (MCMC) runs, the initial 25% of sampled generations were omitted prior to building the tree. Mean divergence times were summarized with TreeAnnotator.

RESULTS

Mitochondrial MutS phylogeny

The 17 sequences of mtMutS of L. virgulata are identical, while the seven sequences of L. hebes range from 99.74% to 100% identity. The phylogenetic reconstruction based on mtMutS included the 17 sequences of L. virgulata and the seven sequences of L. hebes generated by this study (Figure 1). These 24 sequences were combined with 158 additional Leptogorgia mtMutS sequences and two of Eunicella (outgroup), downloaded from GenBank. The topology between the Bayesian and maximum likelihood analyses is relatively similar. There are nine major clades (A–I) that are strongly supported (>70 bootstrap and >95 posterior probability) except for Clade F (<50 bootstrap and <50 posterior probability), each corresponding to taxa from a particular geographic region (Figure 1). Clade A corresponds to species found in the Eastern Atlantic and Mediterranean. Clade A is sister to Clade B, which corresponds to species found in South Africa. Clade A and B, along with Pseudopterogorgia fredericki and Pseudopterogorgia australiensis form a clade that is weakly supported (51 bootstrap and 85 posterior probability) and sister to all other species (Figure 1). This sister clade with the remaining species is moderately supported (77 bootstrap and 86 posterior probability) and contains clades C–I. Clade C (100 bootstrap and posterior probability) consists of 11 different species representing multiple genera (Pseudopterogorgia, Antillogorgia, Gorgonia, and Phyllogorgia). All species in Clade C are found in the Caribbean (Figure 1). Clade C is sister to the remaining species which form a strongly supported group (97 bootstrap and 86 posterior probability) containing clades D–I. Clade D (100 bootstrap and posterior probability) is made up of species of Pacifigorgia along with a few species of Leptogorgia, all of which are from the Eastern Pacific (Figure 1). Clade D is sister to a strongly supported clade (100 bootstrap and 100 posterior probability) that consist of the remaining species within clades E–I. Clade E (89 bootstrap and 100 posterior probability) consists of several species of Leptogorgia and Eugorgia all from the Eastern Pacific (Figure 1). Clade E is sister to a strongly supported group (91 bootstrap and 100 posterior probability) containing clades F–I. Clade F does not have statistical support (<50 bootstrap and <50 posterior probability). Within clade F are Leptogorgia violacea, L. punicea, and L. rubra along with a well‐supported clade (90 bootstrap and 99 posterior probability) containing all specimens of L. hebes (Figure 1). All species in clade F are from the Western Atlantic and Gulf of Mexico. Clade F is sister to a clade with no statistical support (<50 bootstrap and <50 posterior probability) containing the remaining species within clades G‐I. Clade G (100 bootstrap and posterior probability) contains several species of Eugorgia along with Leptogorgia pumila, all from the Eastern Pacific (Figure 1). Clade G is sister to a group that is not statistically supported (<50 bootstrap and <50 posterior probability) and contains clades H and I. Clade H (100 bootstrap and posterior probability) consists of species from the Western Atlantic and Gulf of Mexico, Leptogorgia gracilis and L. virgulata (Figure 1). Specimens of L. virgulata form a clade that is moderately supported in the ML tree (64 bootstrap) and strongly supported in the Bayesian tree (100 posterior probability). This L. virgulata clade is sister to L. gracilis. Clade H is sister to Clade I (100 bootstrap and posterior probability) which contains numerous species of Leptogorgia from the Eastern Pacific (Figure 1).

Figure 1

Maximum likelihood phylogenetic reconstruction of the genus Leptogorgia based on mtMutS. Major clades by geographic regions labeled A–I. Support values shown only for clades A–I and denoted by an * on branch, indicating strong support (>70 bootstrap/>95 posterior probability)

Mitochondrial genomes

A total of ten new Leptogorgia mitochondrial genomes were obtained—seven L. virgulata mt genomes and three L. hebes mt genomes. The L. virgulata mt genomes range in length from 18,824 to 18,845, and all L. hebes mt genomes are 19,247 bp. The L. virgulata mt genomes range from 99.87% to 100% identity while those for L. hebes range from 99.5% to 99.98% identity. All ten mt genomes consist of 14 protein‐coding genes (Cox1, Nad1, CytB, Nad6, Nad3, Nad4L, mtMutS, Nad2, Nad5, Nad4, Cox3, Atp6, Atp8, and Cox2, in respective order) and two ribosomal RNAs (Figure 2). Both species have what is presumed to be the ancestral gene order found in octocorals (Brugler & France, 2008; Figueroa & Baco, 2014, 2015; Medina, Collins, Takaoka, Kuehl, & Boore, 2006; Park et al., 2012; Uda et al., 2011).

Figure 2

Complete mitochondrial genomes for Leptogorgia virgulata and Leptogorgia hebes including all 14 protein‐coding genes, shown in blue, and rRNAs, shown in red

Mitogenomic phylogeny

The phylogenetic reconstruction based on full mitochondrial genomes included 7 mt genomes of L. virgulata and 3 mt genomes of L. hebes generated by this study. These mt genomes were combined with 11 additional mt genomes from the family Gorgoniidae and two mt genomes of Eunicella (outgroup), downloaded from GenBank (Table 3). Maximum likelihood and Bayesian analyses resulted in similar topology (Figure 3). There are nine well‐supported clades (clades I–IX) that roughly match those identified in the mtMuS phylogeny (Figure 3). Clade I (96 bootstrap and 100 posterior probability) is made up of Leptogorgia palma (mtMuS clade B) as sister to clade II (100 bootstrap and posterior probability, mtMuS clade A) which contains Leptogorgia capverdensis and Leptogorgia sarmentosa. Clade I is sister to all other Leptogorgia, but this sister clade is weakly supported (61 bootstrap and 85 posterior probability) and contains Pseudopterogorgia bipinnata (mtMuS clade C) as sister to clade III. Clade III (100 bootstrap and posterior probability) contains Pacifigorgia cairnsi (mtMuS clade D) as sister to clade IV (100 bootstrap and posterior probability). Clade IV contains Leptogorgia sp. (KY559412) as sister to clade V (89 bootstrap and 100 posterior probability). Clade V consists of clade VI (90 bootstrap and 100 posterior probability) as sister to clade VII (65 bootstrap and 100 posterior probability). Clade VI has Leptogorgia alba (mtMuS clade I) as sister to Eugorgia mutabilis (mtMuS clade E). Clade VII contains clade VIII (mtMuS clade F) as sister to clade IX (mtMuS clade H). Clade VIII (100 bootstrap and 100 posterior probability) consists of three specimens of L. hebes. Clade IX (100 bootstrap and posterior probability) consists of seven specimens of L. virgulata. Within the L. hebes clade, two individuals (accession #s MN052676 and MN052675) form a strongly supported clade (97 bootstrap and 81 posterior probability). The L. virgulata clade also has two individuals (accession #s MK0301589 and MK0301591) forming an internal clade, strongly supported by maximum likelihood only (96 bootstrap).

Figure 3

Maximum likelihood phylogenetic reconstruction of the family Gorgoniidae using complete mitogenomes. Major clades labeled I–IX. Branch labels show support values (bootstrap/posterior probability). * on species names indicates mitochondrial genomes generated by this study. Branches labeled A–H correspond to clades defined in the phylogeny based on mtMutS

Mitogenomic divergence time estimation

The phylogenetic reconstruction based on mitochondrial genomes using fossil‐calibrated coalescent methods as implemented by Bayesian analyses in BEAST (Figure 4) resulted in topology similar to the maximum likelihood (ML) and Bayesian analysis with RaxML and Mr. Bayes, with some key differences. The tree is rooted with two Eunicella sp. (outgroup). Emerging from the root are two main clades which diverged from one another 25.96 Ma. Within the first main clade, there are two branches containing a single species each—P. bipinnata and Leptogorgia cf. palma which diverged 25.01 and 21.95 Ma, respectively. Following these two branches is a branch containing Leptogorgia gaini, which diverged 12.75 Ma from a sister subclade consisting of L. sarmentosa and L. capverdensis. Leptogorgia sarmentosa and L. capverdensis diverged from one another 0.46 Ma. However, this subclade is weakly supported. The grouping of the five aforementioned species is consistent between all three mitogenomic trees, with the exception of P. bipinnata. On the ML and Mr. Bayes' Bayesian trees, P. bipinnata does not emerge until after the 4 other species—L. cf. palma, L. gaini, L. sarmentosa, and L. capverdensis—and it forms a basal branch.

Figure 4

Fossil‐calibrated phylogenetic reconstruction using Bayesian methods, showing divergence times in millions of years ago, indicated by values to the right of the nodes. The scale below the tree is millions of years, and the scale bar is 3.0 million years. Bar labels indicate posterior probability. Color indicates geographic region of species. Branches labeled I–IX as defined by the noncalibrated phylogenetic reconstruction using mitochondrial genomes and those labeled A–I correspond to clades as defined in the phylogenetic reconstruction based on mtMutS

In the second main clade, P. cairnsi diverges at 20.38 Ma and forms a basal branch to a subclade containing L. hebes, E. mutabilis, Leptogorgia sp. (KY559412), L. alba, and L. virgulata. Leptogorgia hebes is the first species to diverge from this subclade at 10.89 Ma. Following the L. hebes group are two branches containing E. mutabilis and Leptogorgia sp. (KY559412), diverging at 9.88 and 6.75 Ma, respectively. Leptogorgia alba and L. virgulata then diverged from one another at 5.82 Ma. In the ML and Mr. Bayes' Bayesian trees, E. mutabilis and L. alba are sister to one another, but on the BEAST tree L. alba is sister to L. virgulata.

DISCUSSION

The reconstructed mtMutS phylogeny uses 68 new mtMutS sequences (24 from this study and 44 from GenBank) added to the sequences used in the phylogenetic tree by Poliseno et al. (2017). This new mtMutS phylogeny agrees with the phylogeny presented by Poliseno et al. (2017). The Leptogorgia species from South Africa form a sister clade to species from the Eastern Atlantic and Mediterranean (Figure 1). The Caribbean clade from Poliseno et al. (2017) is also recovered (clade C, Figure 1). There are several clades with species exclusively from the Eastern Pacific. Most notably, Eastern Pacific clade I is sister to the Western Atlantic and Gulf of Mexico clade H that contains L. virgulata and L. gracilis. As in Poliseno et al. (2017), the major clades identified (A–I) have species that are exclusive to a particular geographic region. And while all of these clades are strongly supported (except for clade F), relationships between several of these clades is not clear due to low or no statistical support. The South African clade (clade B, Figure 1) contains L. palma, formerly known as Lophogorgia crista, which is the type species for the Lophogorgia genus (Poliseno et al., 2017). Because this South African group is monophyletic and strongly supported, Poliseno et al. (2017) recommend that the genus Lophogorgia be resurrected and assigned to this clade. This complicates matters when it comes to other species formerly classified as Lophogorgia by Bayer (1961) which are not in the South African clade‐specifically, L. dichotoma, L. capverdensis, L. gaini, Lophogorgia viminalis, L. hebes, L. punicea, and L. violacea. The former four all belong to the eastern Atlantic clade, while the latter three are Western Atlantic species. Further morphological and genetic analyses of these species in particular will be necessary in order to more accurately classify them and determine whether resurrecting the genus Lophogorgia would be appropriate. If the South African clade is recognized as its own genus, whether through the resurrection of Lophogorgia or by a new name, it would complicate the taxonomy of the remaining Leptogorgia species. The issue is that the type species for the genus Leptogorgia is L. viminalis, formerly known as Gorgonia viminalis (Breedy & Guzman, 2007) is within a monophyletic group with eastern Atlantic‐Mediterranean species, sister to the South African group. Therefore, if the South African group is granted species status, then any species that are not in the sister clade with L. viminalis (which are the majority of Leptogorgia species) could not be classified as Leptogorgia and would have to be renamed. This supports Poliseno et al.'s (2017) call to reclassify almost all Leptogorgia species and revise the genus in its entirety, which leaves the case of L. hebes all the more ambiguous, as it does not fit in either Leptogorgia or Lophogorgia. It is likely that new genera need to be defined within this group to resolve these taxonomic issues.

The complete mitochondrial genomes of 21 gorgonian specimens were examined to elucidate phylogenetic relationships and to test the efficacy of using complete mt genome over the single mtMutS gene. This is the first study to sequence complete mitochondrial genomes for L. virgulata and L. hebes, and the resulting mitogenomic phylogeny is in agreement with our mtMutS phylogeny and with that of Poliseno et al.'s (2017), albeit with stronger branch support. The tree topology also matches that of the mitogenomic phylogeny presented by Poliseno et al. (2017) while adding L. hebes and L. virgulata from the Gulf of Mexico as a sister clade to E. mutabilis and L. alba from the Eastern Pacific. These observations support Poliseno et al.'s conclusions that Western Atlantic gorgonians are more closely related to Eastern Pacific gorgonians than to eastern Atlantic gorgonians (L. cf. palma, L. gaini, L. sarmentosa and L. capverdensis).

Divergence time estimation

This is the first study to place divergence time estimates on complete mitochondrial genomes of Leptogorgia species. Poliseno et al. (2017) suggested the first divergence event between Eastern Pacific and Western Atlantic species occurred about 28 Ma (with error bars ranging from 12 to 45 Ma). However, the fossil‐calibrated mitogenomic phylogeny presented in this study suggests that this first split between Eastern Pacific and western Atlantic species occurred later between 11 and 20 Ma (Figure 4), which is within Poliseno et al.'s (2017) lower error range. According to O'Dea et al. (2016), the formation of the Isthmus of Panama was not a singular event, but rather a series of geological events that took place over the course of the last 30 million years. Between 20 and 10 Ma, the Panama Arc island chain began to rise, based on O'Dea et al.'s (2016) estimated rates of Arc uplift. Gene flow by the exchange of gametes and larvae through the CAS was likely high up to10 Ma, while there was still significant seawater exchange between the Atlantic and Pacific oceans (O'Dea et al., 2016). Both, L. hebes and L. virgulata are adapted to shallow water habitat ranging from 3 to 82 m (Cairns & Bayer, 2009; Williamson et al., 2011). They mature rapidly (<2 years) and are broadcasts spawners, releasing eggs and sperm into the water column (Beasley et al., 2003; Gotelli, 1991). While larval duration in L. hebes is not known, it can last up to 20 days in L. virgulata (Gotelli, 1991). These characteristics indicate a potential for high dispersal and suggest that gametic and larval connectivity likely occurred between the Pacific and Atlantic oceans through a shallow CAS. Divergence of Leptogorgia between these basins likely increased after 10 Ma as seawater exchange became more constricted. Our data suggests that the L. hebes speciated at about 11 Ma and it forms the first Western Atlantic clade on the mitogenomic tree. This divergence time coincides with the timing of more restricted water flow between the two basins.

O'Dea et al.'s (2016) uplift data show that after this uplifting period between 20 and 10 Ma, a deepening event occurred between 10 and 6 Ma, in which the Panama Arc began to drop, resulting in greater connectivity between the ocean basins. This span of time is also characterized by shifts in migration rates of both terrestrial and marine fauna, referred to as migration pulses, by Bacon et al. (2015). They specifically highlight a migration shift among marine organisms at around 7.96 Ma, based on their free model migration estimate. The synchrony of submergence of the Panama Arc and a migration event in marine organisms suggest that gene flow could have increased between the Eastern Pacific and Western Atlantic during this time. Following this period of subsidence, at around 6 Ma the Panama Arc began to emerge again and has continued to rise until the present day (O'Dea et al., 2016). Divergences of marine organisms begin to increase at this time, peaking at about 4 million years ago (O'Dea et al., 2016). This timing of events supports the divergence estimate of the Western Atlantic Leptogorgia sp. (KY559412) at 6.75 Ma and the divergence of the Western Atlantic L. virgulata clade at 5.82 Ma, both diverging from sister clades in the Eastern Pacific. The estimated times of Leptogorgia species divergence obtained from this study are concordant with geologic data and historic migration data (Bacon et al., 2015; O'Dea et al., 2016), supporting an initial divergence between Eastern Pacific and Western Atlantic species at about 20–11 Ma with extant lineages arising in each basin in an alternating pattern at 11 (Western Atlantic), 10 (Eastern Pacific), 7 (Western Atlantic), and 6 (Eastern Pacific) Ma (Figure 4).

The divergence times obtained from this study are more recent than those presented by Poliseno et al. (2017) and with lower error estimates (2–4 million‐year range as opposed to a 12–40 million‐year range). This discrepancy is most likely attributed to our use of complete mitochondrial genomes that include fourteen protein‐coding genes and two RNAs instead of a single, partial gene (mtMutS). There are numerous studies of multiple taxa showing a pattern of incongruent tree topology between single mitochondrial markers and complete mitochondrial genomes despite the fact that they are the same locus and therefore share the same phylogenetic history (Havird & Santos, 2014; Knaus et al., 2011; Luo et al., 2011; Nadimi et al., 2016; Pacheco et al., 2011; Rohland et al., 2007; Urantowka et al., 2017; Wang et al., 2017; Willerslev et al., 2009). For example, Havird, Santos Scott, and Schierwater, (2014) analyze the performance of single and concatenated sets of mitochondrial genes relative to complete mitochondrial genomes for phylogenetic reconstruction of metazoans. Their findings show that single genes are not able to reproduce the topology of a mitogenomic phylogeny (Havird & Santos, 2014). A similar study, but focusing on birds, showed that single mitochondrial genes resulted in incorrect and contradictory phylogenetic relationships, while the use of complete mitochondrial genomes accurately reflected the species tree (Urantowka et al., 2017). The same pattern has been observed in insects, where individual mitochondrial genes can result in different and contradicting tree topologies, while using the complete mitochondrial genome performs well at various taxonomic levels (Wang et al., 2017). In fungi, the phylogenetic signal differs between single mitochondrial genes, subsets of concatenated mitochondrial genes, and complete mitochondrial genomes, despite all being the same locus (Nadimi et al., 2016). In addition to potentially generating different and contradicting tree topologies, there are numerous examples across widespread taxa on how single mitochondrial genes oftentimes result in poorly supported phylogenetic trees that become fully resolved and well supported when using complete mitochondrial genomes (i.e., Arquez, Colgan, & Castro, 2014; Justice, Weese, & Santos, 2016; Perseke, Golombek, Schlegel, & Struck, 2013; Williams, Foster, & Littlewood, 2014; Yu, Li, Ryder, & Zhang, 2007). Phylogenies in Octocorals present a similar issue; previous research shows that it is difficult to distinguish between species when using the single gene mtMutS and that even using a concatenated set of 2–3 different mitochondrial regions only allows to distinguish 70%–80% of morphological species (i.e., Baco & Cairns, 2012; McFadden et al., 2011). The low resolution provided by the use of a single mitochondrial region explains the low support for many clades in the mtMutS phylogeny presented in this study and that of Poliseno et al. (2017). Greater resolution and strong support of clades within the Octocorallia is achieved by using complete mitochondrial genomes, as demonstrated in our present study and in previous research (i.e., Figueroa & Baco, 2015, 2014; Kayal et al., 2013; Poliseno et al., 2017).

In addition to incongruent topologies and weakly supported clades, the use of single genes can result in overestimation of calibrated divergence times (Duchêne et al., 2011; McCormack et al., 2011). McCormack et al. (2011) demonstrates that divergence estimation from single mitochondrial genes results in earlier divergence times when compared to the use several markers from the mitochondrial and nuclear genome. They show that the gene tree reconstructed from single mitochondrial markers is not as robust and differs from the species tree reconstructed by using multiple markers from various loci (McCormack et al., 2011). While McCormack et al. (2011) did not examine if complete mitochondrial genomes alone would yield better results, similar to those obtained when using several mitochondrial and nuclear markers, the research by Duchêne et al. (2011) suggests that this might be the case. In their study, Duchêne et al. (2011) compare phylogenetic divergence estimates for cetaceans based on single mitochondrial genes, different combinations of concatenated genes, and complete mitochondrial genomes. Their results show that tree topology from single genes can differ from each other due to different substitution rates and that single gene divergence time estimates consistently resulted in overestimation of divergence times when compared to the use of complete mitochondrial genomes (Duchêne et al., 2011). These results from previous research are congruent with our observations that in octocorals, such as the gorgonians analyzed in our study, the use of complete mitochondrial genomes as opposed to single mitochondrial genes, results in better resolved, well supported, trees that have earlier and more precise divergence time estimates. Since our divergence time estimates are concordant with regional geological events and divergence patterns of other organisms, it supports our hypothesis that the divergence times of Eastern Pacific and Western Atlantic Leptogorgia lineages is younger than previously suggested (Poliseno et al., 2017) with the majority of speciation events occurring after 10 Ma when significant seawater exchange between the Pacific and Atlantic Ocean ceased (e.g., Bacon et al., 2015; Montes et al., 2015; O'Dea et al., 2016). However, future work that includes multiple nuclear markers in addition to mitochondrial genomes is necessary to fully test this hypothesis.

CONFLICT OF INTEREST

The authors declare they have no conflicts of interest.

AUTHOR CONTRIBUTIONS

DFF conceived the ideas and designed methodology; DH collected the specimens; SS, NJF, and DFF generated the genetic data; SS and NJF analyzed the data; DFF and DH supervised research and analyses; SS and DFF wrote the manuscript; SS, DFF, NJF, and DH contributed to the interpretation of data. All authors contributed critically to the drafts and gave final approval for publication.

OPEN RESEARCH BADGES

This article has earned an https://openscience.com for making publicly available the digitally‐shareable data necessary to reproduce the reported results. The data is available at https://www.ncbi.nlm.nih.gov/genbank/.