Contribution of metabolomics to the taxonomy and systematics of octocorals from the Tropical Eastern Pacific

Introduction

‘Animal forests’, settled on rocky reefs and formed by dense communities of sessile marine invertebrates, especially members of the Cnidaria group, represent one of the most important marine ecosystems with a high diversity of microhabits (Jones et al., 2004; Sánchez, 2016; Reverter et al., 2018). Octocorals (Cnidaria: Anthozoa: Octocorallia: Malacalcyonacea) are widespread and abundant in different marine habitats. In addition to stony corals, this group of marine invertebrates has been the focus of several studies (Idjadi & Edmunds, 2006; Plaisance et al., 2011; Sánchez et al., 2019; Stella, Jones & Pratchett, 2010). They are the most abundant and diverse macrofaunal sessile group in temperate and tropical waters (Bayer, 1961; Sánchez et al., 2003; Soler-Hurtado, Megina & López-González, 2018; Velásquez & Sánchez, 2015). Inventories performed in the Tropical Eastern Pacific (TEP) realm reported the order Alcyonacea, now known as Malacalcyonacea, as the most emblematic group in this region (Breedy & Guzman, 2004; Breedy, Guzman & Vargas, 2009), with the presence of endemic genera and species widely distributed in the underwater substrates of mainland coasts, but also around oceanic islands (Abeytia, Guzman & Breedy, 2013; Breedy & Guzman, 2007; Soler-Hurtado et al., 2016).

Recently, several taxonomic studies have provided valuable information to fill gaps in our knowledge of this group of corals across the TEP realm. Approximately 190 species have been reported on the coasts of Ecuador, Colombia, Panama, Costa Rica, and Mexico (Breedy & Cortes, 2008, 2011; Breedy & Guzman, 2004, 2005a, 2005b, 2003a, 2003b; Soler-Hurtado et al., 2016; Steiner et al., 2018; Vergara-Florez et al., 2022). The coast of Ecuador has been recognized as a hotspot of marine biodiversity within the TEP region (Myers et al., 2000), even though its coastline is relatively shorter than those of Chile, Peru, and Colombia. Eight new species of the order Malacalcyonacea have been identified in this region over the past decade (Soler-Hurtado & López-González, 2012; Soler-Hurtado et al., 2016, 2017; Vergara-Florez et al., 2022). Leptogorgia, Muricea, and Pacifigorgia were the most abundant octocoral genera along the coast of Ecuador. They are widely distributed along the coastline and have been described as excellent builders of marine ecosystems (Abad et al., 2022; Soler-Hurtado et al., 2016; Steiner et al., 2018).

Species descriptions of octocorals present in shallow waters (<50 m) of the Eastern Pacific have focused on morphology, particularly sclerite characterization (Breedy & Guzman, 2004, 2005b, 2007, 2003a; Breedy, Guzman & Vargas, 2009; Zeas Valarezo, 2015). Nevertheless, recent studies have highlighted the potential of DNA barcoding for the identification and classification of octocoral species (Sánchez et al., 2003; Soler-Hurtado, López-González & Machordom, 2017; Vargas et al., 2010; Wirshing & Baker, 2015). This combination of morphological features and molecular barcoding sequences has been recognized by the scientific community as an integrative taxonomic method (Dayrat, 2005; Will, Mishler & Wheeler, 2005). Furthermore, other scientific areas, including biochemistry, histology, and ecology, have contributed to the description and classification of species, enabling a better understanding of their origins, geographical boundaries, and historical evolution (Jaramillo, 2019; Ankita et al., 2016; Padial et al., 2010; Singh, 2012). While molecular barcoding sequence proved to be useful in elucidating the phylogenetic relationships of octocoral, for example, by using mt-COI and mt-MutS, the data results could be inconclusive for the species-level discrimination (Vargas et al., 2014). Therefore, the aim of this study was to assess the potential of a more integrative approach, incorporating chemical data, for the species-level separation of octocorals.

The utilization of chemical data for the identification of species has a long historical precedent within the discipline of taxonomy. This has facilitated the generation of invaluable insights into the prospective use of synapomorphic chemical markers for the classification of different taxa, a process currently recognized as chemotaxonomy (Jaramillo, 2019). Secondary or specialized metabolites play a vital role in the marine environment, with some acting as defense mechanisms. These include the production of mucus, stinging substances, and radiant colors to camouflage themselves. The phenotypic traits expressed by these metabolites are vital for the survival of living organisms and display high levels of specificity between species of the same taxon (Croteau, Kutchan & Lewis, 2000; Pichersky & Gang, 2000). In recent decades, metabolomics has emerged as a novel approach to address the limitations of classical chemistry. This is corroborated by sophisticated analytical techniques, including high-resolution mass spectrometry (HR-MS1), computational and statistical analysis of extensive datasets, which are employed to investigate the entire metabolome of a single organism (Patti, Yanes & Siuzdak, 2012; Smith et al., 2006).

Metabolomic datasets have recently been added to integrative systematic studies on different groups of marine invertebrates, such as sponges (Ivanišević et al., 2011), zoantharians (Cachet et al., 2015; Jaramillo et al., 2018), cyanobacteria (Duperron et al., 2020), and soft corals (Aratake et al., 2012; Coll, 1992; Gerhart, 1983; Kessel et al., 2023; Vohsen, Fisher & Baums, 2019), providing valuable insights into their classification. Octocorals are good candidates for the use of metabolomic data, as some species of this group are known to be rich sources of structurally diverse specialized metabolites, also known as natural products, such as sesquiterpenoids, sterols, steroidal glycosides, and saponins, among others (Bandurraga & Fenical, 1985; Berrue & Kerr, 2009; Carroll et al., 2019; Popov, Carlson & Djerassi, 1983). In the context of this study, the octocoral genera Muricea and Leptogorgia have been the subject of the most extensive research in terms of their chemical composition, in comparison to other genera such as Psammogorgia, Heterogorgia, and Pacifigorgia. In fact, chemical studies on Muricea species revealed the presence of sterols (Popov, Carlson & Djerassi, 1983), saponins (Bandurraga & Fenical, 1985; Gutiérrez et al., 2004), steroids and derivatives (Gutiérrez et al., 2006; Lorenzo et al., 2006; Murillo-Álvarez & Encarnación-Dimayuga, 2003; Ortega et al., 2002) while studies on Leptogorgia species reported the presence of polyoxygenated steroids (Cimino et al., 1984; Garrido et al., 2000; Moritz et al., 2014) diterpenes such as cembranolides and furanocembranolides (Dorta et al., 2007; Gutiérrez et al., 2005; Ksebati, Ciereszko & Schmitz, 1984; Missakian, Burreson & Scheuer, 1975; Ortega et al., 2008) halogenated furanocembranolides (Gallardo et al., 2018) sesquiterpenes (Kapustina et al., 2020) and tetraprenylated alkaloids (Keyzers et al., 2006). Chemical studies on Pacifigorgia species have revealed the presence of sesquiterpenes and steroids (Izac et al., 1982a, 1982b; Chen et al., 2016), whereas sesquiterpenes have only been reported in Heterogorgia uatumani (Maia et al., 1999). No chemical studies have been conducted on Psammogorgia species.

Metabolomic studies have been performed on octocoral species to identify bioactive metabolites (Farag et al., 2017; Santacruz et al., 2019; Tanaka, Yoshida & Benayahu, 2005). Santacruz et al. (2019) identified some diterpenoids as specific markers to discriminate the cytotoxic extracts of 28 Caribbean soft corals using a metabolomic approach. Aratake et al. (2012) investigated the relationship between soft coral diversity and cembranoid diterpene production across multiple species of the genus Sarcophyton in Okinawa, Japan using an integrative approach (morphology, phylogeny, and chemotaxonomy). The second metabolomic study focused on characterizing the metabolomic diversity of deep-sea corals in an ecological context by investigating patterns across space and phylogeny (Vohsen, Fisher & Baums, 2019).

In this study, we hypothesized that the metabolomic data of octocorals collected from the Marine Reserve El Pelado (REMAPE) could help in the systematics of this group (Berrue & Kerr, 2009; Kobayashi et al., 1990). Therefore, the aim of our study was to assess the potential of an untargeted metabolomics approach to explore and characterize the different genera of octocorals collected in REMAPE and between the different species of Muricea.

Materials and Methods

Biological material

Octocoral samples were obtained under the research permit N°005-17 IC-FAU-DPSE/Maas, which was granted as part of the framework contract for access to genetic resources “MAE-DNB-CM-2015-0021” by the Ministry of Environment, Water and Ecological Transition of Ecuador (MAATE). In addition, the ESPOL University’s Research Review Board has formally approved the ethical framework for conducting the study in its facilities under the following projects: “Characterization of the invertebrates biodiversity at El Pelado Marine Reserve at Metabolomic Metagenomic Taxonomic scales for use in human and animal health”, ethical application reference: CENAIM-470-3-2014 and “Valorization of the biodiversity through biodiscovery for its sustainable use in human health and Aquaculture”, ethical application reference: CENAIM-400-2019).

A total of 71 octocoral specimens belonging to 12 species were collected during August and September 2017 through SCUBA diving in rocky reef habitats at depths between 10 and 26 m from three different sites (Cuarenta, Pared and Acuario) at the Marine Protected Area “El Pelado” (REMAPE) in the province of Santa Elena, mainland coast of Ecuador as described in Jaramillo et al. (2018) (Fig. 1 and Table 1). For nine species (OC-3, 63, 43, 64, 70, 67, 55, 74, and 20), the same specimen was processed for all three types of approaches: A, a dry subsample was kept for external morphological characterization; B, a subsample was fixed in 95% ethanol for sclerite and molecular analyses and C, a subsample was frozen at –80 °C for metabolomic studies. For the three species, Muricea plantaginea, Muricea squarrosa and Psammogorgia arbuscula, no molecular barcoding sequence could be obtained. However, from three to 12 specimens of each species could be analyzed by metabolomics. Portions of this text have previously been published as part of a thesis (https://researchrepository.universityofgalway.ie/entities/publication/00612b54-607e-411d-bd69-9b09f3cee85f).

Table 1:

Octocorals samples.

Species, specimens, localities, voucher numbers and Genbank accession numbers analysed in this study. The sampling sites coordinates can be found in Jaramillo et al. (2018).

Species name	Sample code	Locality	Depth (m)	Morpho	Metabo	COI	MutS1	Voucher
Muricea fruticosa	OC-1	Acuario	10	_	_	_	_	170816EP02-06
	OC-2	Acuario	25	_	✓	_	_	170816EP02-07
	OC-3	Acuario	10	✓	✓	MN318307	XXXXXXX	170816EP02-08
	OC-4	Acuario	12	_	✓	_	_	170816EP02-09
	OC-33	Cuarenta	22	_	✓	_	_	170816EP07-12
	OC-34	Cuarenta	20	_	✓	_	_	170816EP07-13
	OC-47	Pared	10	_	✓	_	_	170810EP01-12
Muricea plantaginea	OC-48	Pared	26	_	✓	_	_	170810EP01-13
	OC-50	Pared	24	_	✓	_	_	170810EP01-15
	OC-51	Pared	25	_	✓	_	_	170810EP01-16
	OC-52	Pared	26	_	✓	_	_	170810EP01-17
	OC-53	Pared	28	✓	✓	_	_	170810EP01-18
	OC-5	Acuario	20	_	✓	_	_	170816EP02-10
	OC-6	Acuario	31	_	✓	_	_	170816EP02-11
	OC-7	Acuario	20	_	✓	_	_	170816EP02-12
	OC-8	Acuario	25	_	✓	_	_	170816EP02-13
	OC-69	Cuarenta	15	_	✓	_	_	170907EP07-14
	OC-25	Cuarenta	10		✓	_	_	170816EP07-04
Muricea squarrosa	OC-27	Cuarenta	20	_	✓	_	_	170816EP07-06
	OC-28	Cuarenta	31	_	✓	_	_	170816EP07-07
	OC-30	Cuarenta	15	_	✓	_	_	170816EP07-09
	OC-31	Cuarenta	25	✓	✓	_	_	170816EP07-10
	OC-60	Cuarenta	10	_	✓	_	_	170907EP07-05
	OC-61	Cuarenta	10	_	✓	_	_	170907EP07-06
	OC-62	Cuarenta	12	_	✓	_	_	170907EP07-07
	OC-29	Cuarenta	8	_	✓	_	_	170816EP07-08
	OC-57	Cuarenta	10	_	✓	_	_	170907EP07-02
	OC-12	Acuario	10	_	✓	_	_	170816EP02-17
	OC-15	Acuario	15	_	✓	_	_	170816EP02-20
	OC-16	Acuario	12	_	✓	_	_	170816EP02-21
Muricea crassa	OC-63	Cuarenta	10	✓	✓	MN318310	XXXXXXX	170907EP07-08
	OC-22	Cuarenta	15	_	✓	_	_	170816EP07-01
	OC-56	Cuarenta	12	_	✓	MN318309	XXXXXXX	170816EP07-03
	OC-23	Cuarenta	12	_	✓	_	_	170816EP07-02
	OC-73	Acuario	8	_	✓	_	_	170816EP02-15
	OC-24	Acuario	10	_	✓	_	_	170816EP02-18
Muricea purpurea	OC-14	Acuario	18	_	✓	_	_	170816EP02-19
	OC-13	Acuario	5	_	✓	_	_	170816EP02-18
	OC-71	Acuario	15	_	✓	_	_	170816EP02-19
	OC-42	Pared	7	_	✓	_	_	170810EP01-07
	OC-43	Pared	12	✓	✓	MN318308	XXXXXXX	170810EP01-08
	OC-44	Pared	9	_	✓	_	_	170810EP01-09
	OC-46	Pared	13	_	✓	_	_	170810EP01-11
Muricea austera	OC-10	Acuario	7	_	✓	_	_	170816EP02-15
	OC-11	Acuario	12	_	✓	_	_	170816EP02-16
	OC-64	Cuarenta	9	✓	✓	_	XXXXXXX	170907EP07-09
	OC-35	Cuarenta	13	_	✓	_	_	170816EP07-14
	OC-36	Cuarenta	7	_	✓	_	_	170816EP07-15
	OC-59	Cuarenta	5	_	✓	_	XXXXXXX	170907EP07-04
	OC-9	Cuarenta	15	_	✓	_	_	170816EP02-14
Heterogorgia hickmani	OC-18	Acuario	20	✓	✓	_	_	170816EP02-23
	OC-26	Cuarenta	12	_	✓	_	_	170816EP07-05
	OC-32	Cuarenta	10	_	✓	_	_	170816EP07-11
	OC-58	Cuarenta	18	_	✓	_	_	170907EP07-03
	OC-70	Cuarenta	15	✓	✓	MN318313	XXXXXXX	170907EP07-15
Psammogorgia arbuscula	OC-37	Pared	10	_	✓	_	_	170810EP01-02
	OC-38	Pared	17	_	✓	_	_	170810EP01-03
	OC-39	Pared	12	_	✓	_	_	170810EP01-04
	OC-40	Pared	18	_	✓	_	_	170810EP01-05
	OC-41	Pared	14	_	✓	_	_	170810EP01-06
	OC-49	Pared	20	_	✓	_	_	170810EP01-01
Leptogorgia alba	OC-55	Pared	8	✓	✓	MN318312	XXXXXXX	170810EP01-19B
	OC-65	Cuarenta	15	_	✓	_	_	170907EP07-10
	OC-54	Pared	12	_	✓	_	_	170810EP01-19A
	OC-66	Cuarenta	22	_	✓	_	_	170907EP07-11
Lptogorgia cf. alba	OC-67	Cuarenta	18		✓	MN318311	XXXXXXX	170907EP07-12
Leptogorgia obscura	OC-72	Acuario	12	_	✓	_	_	170816EP02-20
	OC-74	Acuario	10	✓	✓	MN318314	XXXXXXX	170816EP02-16
	OC-75	Acuario	15	_	✓	_	_	170816EP02-20
Pacifigorgia rubicunda	OC-20	Acuario	12	✓	✓	XXXXXXX	XXXXXXX	170816EP02-25
	OC-21	Acuario	20	_	✓	_	_	170816EP02-26
	OC-68	Cuarenta	16	_	✓	_	_	170907EP07-13

DOI: 10.7717/peerj.19009/table-1

Results

Phylogenetic analyses

The mt-MutS variable region was successfully sequenced for 11 octocoral specimens attributed to nine different species, while the mt-COI was successfully sequenced for eight octocoral specimens assigned to eight different species (Table 1). The BLAST results from the 19 sequences matched with octocorals (families Plexauridae, Acanthogorgiidae and Gorgoniidae) reported for the TEP region. The phylogeny reconstructed using ML for mt-MutS is shown in Fig. 2, and the ML mt-COI tree is shown in Fig. 3A. For the families Plexauridae and Acanthogorgiidae, phylogenies drawn from both sets of data identified OC-70 as a member of Heterogorgia (Vargas et al., 2014). In the mt-MutS tree, our specimen was assigned closer to H. verrucosa with only two different base pairs and four base pairs to H. hickmani as reported by Vargas et al. (2010). A similar situation was observed in the mt-COI tree, with the difference that no COI sequence for H. hickmani was available for comparison. Both mitochondrial trees showed highly supported values (100% bootstrap support for MutS and 98.6% bootstrap support for COI) for Heterogorgia species.

The specimens identified as Muricea crassa (OC-63, OC-56), M. fruticosa (OC-3), M.austera (OC-64, OC-59), and M. purpurea (OC-43) clustered with other Muricea species recently reported in the TEP region (Figs. 2 and 3A) (Soler-Hurtado, López-González & Machordom, 2017; Vargas et al., 2010, 2014). Low resolution was observed for both loci (mt-MutS and mt-COI) for this genus, preventing a clear distinction between Muricea species. However, both markers supported the identification at the genus level, with high bootstrap values of 100% for the Muricea clade. Despite the low resolution of both loci, small groups were observed within the Muricea clade of the mt-MutS tree. The first group included the species M. crassa and M. fruticosa with a small difference (4 bp) between both species, and the second group contained the species M. austera and M. purpurea, and no difference was found between the sequences. The latter subclade reflects the morphological similarities between M. purpurea and M. austera (Table 2). Finally, a third group was formed with other Muricea species, in which only M. muricata was distinct. Furthermore, the mt-COI MLphy tree revealed that all Muricea sequences clustered together without any discrimination at the species level. Only one specimen of M. fruticosa exhibited a minor difference of six base pairs from the other species. These findings provide evidence to support the hypothesis that this locus has lower resolution than the mt-MutS region.

Within the family Gorgoniidae, the sequences obtained for both mt-MutS and mt-COI for Leptogorgia obscura (OC-74), L. alba (OC-55, pink morphotype) and L. cf. alba (OC-67, white morphotype) clustered with other species of the same genus. L. obscura (OC-74) was consistent with a previous L. obscura reported by Soler-Hurtado et al. (2016), and this small clade was supported on the COI tree (74.7%) (Fig. 3A). Data from both mitochondrial loci indicated that L. alba (OC-55) and L. cf. alba (OC-67) were distinct. The small clade containing L. alba (OC-55) in the MutS tree was strongly supported by bootstrap analysis, with a value of 88%. In contrast, L. cf. alba (OC-67) was placed in a distinct subclade with bootstrap values greater than 70%, indicating a high degree of confidence in its placement. Furthermore, the species reported as L. cofrini (Vargas et al., 2014) was placed in the same group as L. alba (OC-55) in both trees, and the sequences were identical. All Leptogorgia species from the TEP region form a monophyletic group with strong support (mt-MutS, 99%; mt-COI, 89.6%) (Soler-Hurtado, López-González & Machordom, 2017).

Finally, the mt-MutS sequence from Pacifigorgia rubicunda (OC-20) clustered in a small clade with other Pacifigorgia species, which was highly supported by a bootstrap value of 99%. No significant differences were found between P. rubicunda sequences; only four base pairs differed between P. rubicunda (OC-20) and the single sequence reported for P. rubicunda by Soler-Hurtado, López-González & Machordom (2017). However, the sequence from OC-20 (P. rubicunda) contained several indels compared to other Pacifigorgia species in the TEP realm. Despite these minor sequence differences, the results revealed the low-resolution potential of the mt-MutS region for separating species within the genus Pacifigorgia.

Despite several attempts, we could not amplify the mt-COI gene of the species M. austera (OC-64 and OC-59) and P. rubicunda (OC-20), and neither locus for the species M. plantaginea (OC-50, OC-7), M. squarrosa (OC-15, OC-16), P. arbuscula (OC-38).

Metabolomics analyses

Untargeted metabolomics at the genus level

The quality control samples used in the non-targeted metabolomics approach showed very low variability compared to the variability between groups, thus enabling reliable statistical analysis of the results. With the main aim of assessing the significance of the separation between groups and to flag some variable importance in projection (VIP), we decided to perform supervised PLS-DA (Fig. 4A).

A total of 3,595 molecular ions (m/z and rt) were obtained after performing a non-targeted metabolomic approach using UHPLC-HRMS on 71 octocoral samples (six replicates for each species), including 14 samples for morphological assessment and molecular studies. The metabolomic profiles exhibited consistency between all replicates within a species, while significantly differences were observed between the 12 species under study, and between the five genera (Figs. S4–S15). Supervised PLS-DA of the molecular ions distributed among the five groups showed a high level of similarity within a genus (Fig. 4A). However, using only the two main components, Muricea and Leptogorgia were not well separated in the PLS-DA analysis. Overall, the profiles of the five genera were significantly dissimilar, as demonstrated by the low permutation score (p < 0.01).

HCA was performed to investigate the similarity between the 71 metabolomic profiles from the five genera (Fig. S1). Each genus formed a separate cluster, with Muricea being the genus with the most distinct metabolome, followed by the Leptogorgia and Psammogorgia clusters, which were more similar, and finally, Heterogorgia and Pacificgorgia. All specimens belonging to the same genus were more similar than two specimens belonging to two different genera, confirming the excellent permutation score of PLS-DA and the efficiency of the metabolomics approach in distinguishing genera (Fig. 4A). The cross-validation test of PLS-DA also revealed high scores for R² and Q² with an accuracy higher than 0.7 considering only the principal component (Fig. S2). The three values were clearly significant using the three principal components of the PLS-DA.

To identify some metabolites responsible for the separation between the different genera, we focused on the 15 molecular ions with the highest scores (>3.2) among the VIPs (Fig. 4B). Among the 15 VIPs, the13 molecular ions exhibiting the highest intensity (in red) were detected in the Heterogorgia (5) and Psammogorgia (8) samples. Unfortunately, the lack of chemical studies on both genera have prevented accurate identification of these molecular ions. However, two of the high-intensity molecular ions detected in Heterogorgia, as well as in Leptogorgia species (M388T165 and M344T129), were putatively identified as furanocembranoids by comparison with previously reported compounds from Leptogorgia setacea (Ksebati, Ciereszko & Schmitz, 1984) and Leptogorgia sp. (Gallardo et al., 2018). Furthermore, the mass spectra (HR-MS1) of octocoral samples were manually inspected to tentatively identify muricin 4 from M. fruticosa, leptodienone B from L. alba and leptolide from L. cf. alba, Table 3 following the general accepted guidance described by Alseekh et al. (2021). The MS data of these compounds were then compared with those reported in the literature (Bandurraga & Fenical, 1985; Ortega et al., 2008; Gutiérrez et al., 2005), as well as in databases such as Chemspider and the metabolomic work bench database.

Table 3:

Molecular ions identified from octocoral species.

Summary of the molecular ions identified from the genera Muricea, Leptogorgia.

Organism	Rt (min)	Putative metabolite	Molecular formula	Ion Type	ES (+) Measured m/z	ES (+) Theoretical m/z	m/z error (ppm)	Reference
Muricea fruticosa	10.57	muricin 4	C39H59NO9	[M + H]+	686.4254	686.4190	9.6151	Bandurraga & Fenical (1985)
Leptogorgia alba	8.58	leptodienone B	C20H30O2	[M + H]+	303.2266	303.2245	6.9255	Ortega et al. (2008)
Leptogorgia cf. alba	6.28	leptolide	C20H22O6	[M + H]+	359.1439	359.1416	6.4041	Gutiérrez et al. (2005)

DOI: 10.7717/peerj.19009/table-3

Additionally, we aimed to assess the potential of the metabolomics approach to classify the genera and to compare this metabolomic classification with a phylogenetic tree. Overall, the classification of the genera according to metabolomic data did not agree with the phylogenetic relationships proposed in this study (Figs. 2, 3A). For example, in metabolomic HCA, the genus Leptogorgia was closely related Psammogorgia rather than Pacifigorgia (Fig. S1) while, on both phylogenies (mt-COI and mt-MutS) presented above, Leptogorgia appeared as a sister clade to the genus Pacifigorgia. Likewise, specimens from the genus Muricea were very distinct from those four genera via metabolomics, whereas phylogenetic analyses using molecular barcoding sequence demonstrated that they should be closer to Heterogorgia. However, differences in metabolomic profiles among the three species from the genus Leptogorgia; L. alba (OC-55, 65, 66), L. cf. alba (OC-67), and L. obscura (OC-74, 75, 72) were evident (Fig. S1). This outcome was supported by both phylogenies (mt-COI and mt-MutS). Combining these data with morphological analysis, we suggest that L. cf. alba (OC-67) and L. alba are distinct species.

Untargeted metabolomics for Muricea species

A second metabolomic analysis was conducted to separate Muricea species to evaluate the potential of untargeted metabolomics at the species level. Therefore, only 51 samples (with a minimum of three and a maximum of 12 replicates for each species) were used; 48 specimens belonging to the six species of the genus Muricea (M. fruticosa, M. squarrosa, M. plantaginea, M. purpurea, M. crassa, M. austera) and three specimens of L. obscura were included as an outgroup. The PLS-DA analysis did not provide a supported separation between the six Muricea species (Fig. S3). However, the HCA revealed that the chemical profiles of all replicates within a species were similar, while significant dissimilarities were observed between the six Muricea species and the outgroup species L. obscura (p < 0.01) (Fig. 3B). In this case, the accuracy was lower than 0.7, indicating a lower predictability of the model. Consequently, the VIPs scores for the Muricea species were considered not representative, as these molecular ions are assumed to represent minor metabolites. Therefore, no chemical markers were identified in this study for the purpose of species-level separation.

Although muricin 4 (steroidal glycoside) was putatively identified from M. fruticosa (Table 3), the presence of fragmentation patterns resulting from the loss of H₂O (M⁺-18), which is characteristic of hydroxylated sterols and steroidal glycosides, suggests that these compounds are present in Muricea species. It has been previously reported that certain sterols, including 19-nor-cholest-4-en-3-one, dinosterol, and muriceanol, as well as several saponins, are present in Muricea species (Popov, Carlson & Djerassi, 1983; Bandurraga & Fenical, 1985; Murillo-Álvarez & Encarnación-Dimayuga, 2003; Lorenzo et al., 2006; Gutiérrez et al., 2006).

Different conclusions can be drawn regarding the potential of the metabolomics approach to classification. The chemical profiles of M. austera, M. crassa, M. fruticosa and M. purpurea were closely related and distant from those of the two other and well-separated species, M. squarrosa and M. plantaginea. It is possible that some molecular ions detected in the latter two Muricea species may have acted as PCR inhibitors, thereby preventing an accurate separation of the species. The presence of PCR inhibitors in octocoral species, including polysaccharides, polyphenols, and other metabolites, has been documented (Wong et al., 2024). This suggests the need for alternative protocols to improve DNA extraction. This provides an explanation for the difficulties encountered during the gene amplification in both species (M. squarrosa and M. plantaginea). The molecular ions detected in other Muricea species appeared to be shared among all species, and thus did not appear to be specific biomarkers. When comparing the chemical feature dendrogram (HCA) with the phylogenetic tree built in Figs. 3A–3B some similarities were observed. For example, M. austera and M. purpurea were identified within the same clade through the DNA barcoding and phylogenetic analysis, and they also exhibited a close relationship with the chemical HCA. This result illustrates the potential of metabolomic analysis to distinguish closely related sister species, which have not been previously revealed through phylogenetic analysis.

Discussion

In the current study, we assessed the diversity of octocoral group Malacalcyonacea, representative invertebrates in the shallow waters of the REMAPE area, using an integrative taxonomy approach. Current taxonomic approaches applied to 71 octocoral specimens allowed us to identify 11 known species and one potentially new species currently identified as Leptogorgia cf. alba. The identified octocorals belong to the three families Gorgoniidae (L. alba, L. obscura, P. arbuscula and Pacifigorgia rubicunda), Plexauridae (M. plantaginea, M. squarrosa, M. purpurea, M. austera, M. crassa, M. fruticosa) and Acanthogorgiidae (H. hickmani). These species have already been reported in nearby areas of this region, such as Panama, Colombia, Costa Rica, and in other locations along the Ecuadorian coast (Abad et al., 2022; Breedy & Guzman, 2005a, 2003b; Breedy, Guzman & Vargas, 2009; Glynn, 2003; Soler-Hurtado, López-González & Machordom, 2017; Williams & Breedy, 2004). However, for the first time, metabolomic data were added to strengthen the identification of the species and improve species delimitation in cases where molecular analysis alone was unable to discriminate. Furthermore, this study contributes new records of these species to our region, including new molecular and chemical data.

The phylogenetic relationships among the octocoral species presented here were consistent with those reported in previous phylogenetic studies in the TEP realm (Soler-Hurtado, López-González & Machordom, 2017; Vargas et al., 2010, 2014). Both the mitochondrial regions were successfully amplified in most octocoral species. The higher resolution of mt-MutS over the mt-COI locus makes it a more suitable marker for identifying octocoral species (order Malacalcyonacea), as demonstrated previously (Soler-Hurtado, López-González & Machordom, 2017; Vargas et al., 2014). The molecular phylogenetic analyses presented here revealed that many inconsistencies remain in the separation of octocorals at the species level, similar to the species in the genus Muricea. This is a consequence of both the low resolution of the mitochondrial markers reported by Vargas et al. (2014) and the lack of accessible sequences in the databases. However, we employed these markers because our objective was to identify and confirm the identity of Ecuadorian specimens. To achieve a more accurate separation of sister species, we decided to combine phenotypic characteristics, such as morphological features and metabolomics, with molecular barcoding sequences. The metabolomics approach proved to be an efficient method for the separation of all genera, as well as the different species of Muricea, demonstrating its value as an additional dataset for species characterization. However, in terms of their suitability for classification, the results are controversial. An untargeted approach did not yield a good match with the phylogenetic trees of the targeted genera. However, the high level of similarities between sister species, such as M. austera and M. purpurea, and L. alba and L. cf. alba, was consistent with morphological data and molecular barcoding sequence.

By applying this integrative systematic methodology, we successfully identified a potentially new species within the Leptogorgia genus. The case of L. cf. alba provides an excellent illustration of the value of an integrative taxonomic approach. Indeed, the morphology did not provide clear separation of the L. alba specimens. Consequently, they were first described as distinct morphotypes based on variations, such as the color of the colony and sclerite sizes. Molecular analyses placed them in the same clade; however, the sequences of L. cf. alba (OC-67) and L. alba (OC-55) differed. The distinct species status was supported by metabolomic analyses, which also demonstrated a clear separation. Although only two compounds were putatively identified in both L. alba and L. cf. alba species (leptodione B and leptolide, respectively), the presence of unknown molecular ions indicates the need for further chemical studies to properly characterize their chemical composition in both species. Dorta et al. (2007) suggest that furanocembranolides could be used as chemotaxonomic markers for octocorals, depending on the degree of oxidation at C-18. In this context, we propose that different types of diterpenes can be considered as biomarkers of octocorals. Based on the aforementioned results, it was concluded that 12 species were correctly identified in this study, including one potential new species (L. cf. alba) described in detail herewith. We believe that a combination of molecular and metabolomic analyses has the potential to reveal new species within closely related morphotypes.

The metabolomic results for the four Muricea species, including M. crassa, M. fruticosa, M. purpurea and M. austera, revealed a degree of consistent with the molecular and morphological data. M. crassa and M. fruticosa were placed close to each other, and a similar situation was observed between M. purpurea and M. austera (see Figs. 3A, 3B and compare with Fig. 2). In the latter case, the metabolomic data had a higher resolution than the molecular barcoding sequence, and combined with distinct morphological features, we can infer the presence of the two species M. austera and M. purpurea. This case illustrates how the combination of morphological and metabolomic data enables the distinction of sister species that are difficult to separate using the existing molecular barcoding sequence method.

In terms of classification, one of the main problems raised by metabolomic studies is the lack of the chemical marker identification, which makes it very difficult to assess the potential for systematics. For further analysis, we propose the inclusion of molecular network chemometrics to circumvent this key issue. Indeed, it could allow the grouping of families of natural products and mapping of a network by taxonomic groups to quickly visualize the synapomorphic characteristics of the groups. A trait shared by two or more taxonomic groups has been derived through evolution from a common ancestral form. In this case, the specialized metabolites are considered to be the final products of the expression of unique biosynthetic gene clusters (Reverter et al., 2018). This study identified several putative molecular ions from Muricea and Leptogorgia species. The presence of unreported molecular masses, particularly from Muricea species, encourages further investigation of the chemical composition, metabolite characterization, and biological properties of the octocoral samples analysed.

The development of new tools in natural product chemistry, such as untargeted metabolomics using MS data, enables quick assessment of the broad chemical content of an organism owing to its high sensitivity and short working time. Taxonomy is known to be highly challenging in some complex groups, such as Cnidarians and Porifera, and metabolomic methods have already found some applications as an integrative methodologies for the taxonomy of those groups (Aratake et al., 2012; Cachet et al., 2015; Jaramillo et al., 2018; Santacruz et al., 2019; Tribalat et al., 2016; Vohsen, Fisher & Baums, 2019; Reveillaud et al., 2012). For example, an HPLC-MS-targeted metabolomics tool revealed the presence of different chemotypes (cembranoid diterpenes) associated with various soft corals of the genus Sarchophyton (Aratake et al., 2012). Jaramillo et al. (2018) demonstrated that the use of both an untargeted and targeted metabolomics methods with UHPLC-HRMS was an efficient approach not only as a complementary tool for the taxonomy of Zoantharians, as well as for their classification. Indeed, certain specialized metabolites, such as zoanthamines or 2-aminoimidazole alkaloids, have been identified as key biomarkers for specific zoantharian species (Guillen et al., 2019), as demonstrated by other group (Vohsen, Fisher & Baums, 2019). Despite the extensive and diverse applications of untargeted metabolomics, annotation of chemical markers detected by mass spectrometry remains a significant challenge. In some cases, minor metabolites of limited taxonomic relevance may contribute to the separation of the groups (Quinn et al., 2016).

Recently, metabolomic studies have revealed changes in the metabolomic content of marine invertebrates in response to environmental factors, including geographical location and seasonal changes. For instance, Reverter et al. (2018) employed an untargeted metabolomics approach to investigate the temporal and spatial intraspecific variation in two Mediterranean Haliclona sponges. However, intraspecific variation in the metabolome was found to be relatively minor compared to interspecific variation. Another investigation was performed on two zoantharian species from the genus Palythoa along the Brazilian coast, which assessed geographical patterns vs. species-specific metabolomes (Costa-Lotufo et al., 2018). The combined results of the metabolome variability and the molecular network analyses demonstrated that the differences between intraspecific and interspecific metabolome variability were greater than previously assumed. This finding calls into question the use of this approach for the separation of Zoantharian species. It also highlights the need for studies on larger temporal and geographical scales when environmental changes become more significant.

Conclusions

The current study illustrates the potential of untargeted metabolomics for the separation of different octocoral genera from the order Malacalcyonacea at the REMAPE. Furthermore, the method has been shown to be effective in distinguishing between sister species, a finding that was corroborated by an independent dataset. This has been evidenced by the morphological features of the two species of Muricea, as well as the molecular analysis of Leptogorgia cf. alba, revealing the latter to be a new species. The chemical data proved to be a valuable resource for the characterization of species. The detected unknown molecular ions in octocoral species such as Muricea and Leptogorgia suggests that they may represent untapped sources of novel chemical scaffolds, which could be the subject of further investigation.

Supplemental Information