Epinephelus tankahkeei, a new species of grouper (Teleostei, Perciformes, Epinephelidae) from the South China Sea

Abstract A new species of grouper, Epinephelus tankahkeeisp. nov. is described from the South China Sea based on examination of morphological and molecular characteristics. This new species has been treated as, and is similar to, its congener E. chlorostigma. Epinephelus tankahkeeisp. nov. can be distinguished from E. chlorostigma by the following combination of characters: a convex anal fin; closer dark spots on the body; a lack of dark spots on the abdomen, cheek, and pectoral fin; the absence of a clear posterior white margin on the caudal fin. Molecular analyses of the mitochondrial COI sequence variation, genetic distances, and a phylogeny, all highly support E. tankahkeeisp. nov. as a distinct species. A key to E. tankahkeeisp. nov. and its most closely related species is provided.


Introduction
The groupers are an assemblage of reef fishes in the perciform family Epinephelidae (Smith and Craig 2007;Craig et al. 2011;Zhuang et al. 2013), comprising more than 160 species in 16 genera (Heemstra and Randall 1993;Craig et al. 2011). The genus Epinephelus Bloch, 1793 (type species: Epinephelus marginalis Bloch, 1793 = Epinephelus fasciatus) is the most biologically diverse of all grouper genera (Heemstra and Randall 1993) and contains more than 90 valid species (Frable et al. 2018). These species are characterized by an elongate, robust (subcylindrical), oblong or deep and compressed body; a dorsal fin usually with XI spines (X spines in some species) and 12 to 19 rays; and an anal fin with III distinct spines and 7 to 10 (very rarely 7 or 10) rays. Epinephelus spp. are widespread in the rocky and reef shores of tropical and subtropical oceans, and are usually apex predators in their habitats. They are also commercially important and constitute a significant component of coastal fisheries (Dalzell et al. 1996). Due to the ecological and economic importance of these species, their alpha taxonomy and phylogenetic relationships have been well reviewed (Craig and Hastings 2007;Ma et al. 2016). However, groupers appear to have undergone rapid sympatric speciation and usually show fewer differences in morphology between closely related species, thus some cryptic species might still be undiscovered. Therefore, the use of genetic data is of considerable importance in grouper taxonomic and diversity research (Gilles et al. 2000;Han et al. 2011).
In recent years, we collected a new form of grouper from the South China Sea that had been previously regarded as Epinephelus chlorostigma. Further investigation based on morphometric and molecular characteristics shows that this new form should be a new species of the genus Epinephelus. Herein, we describe this new species as Epinephelus tankahkeei. In addition, a key to E. tankahkeei sp. nov. and its most closely related species is provided.

Materials and methods
Between 2011 and 2019, nine specimens of the new species were collected from fish markets and fishing boats in Xiamen, Shenzhen, Sansha, and Haikou, China. ODV v5.1.5 software was used to generate a collection site map (Schlitzer 2002). The sampling localities are listed in Suppl. material 1: Table S1. The holotype and paratypes were fixed and preserved in anhydrous ethanol. The specimens were stored in the Fish Collection of the College of Ocean and Earth Sciences, Xiamen University. Institutional codes followed Sabaj (2016).
The methods of counting and measurement followed Randall and Heemstra (1993) and include: total length; standard length (as SL); head length; snout length; body depth; body width; orbit diameter; interorbital width; preorbital depth; maxilla width; upper jaw length; lower jaw length; length of pelvic-fin and anal-fin spines; lengths of the dorsal, anal, pectoral, pelvic and caudal fins; caudal-peduncle depth; caudal-peduncle length; predorsal length; preanal length; prepelvic length; dorsal-fin base; longest hard dorsal spine; longest soft dorsal ray; anal-fin base; and length of the third anal spine, longest anal soft ray, and pelvic-fin spine. The following counts were made: gill rakers, lateral-line scales, lateral scale series, pectoral-fin rays, anal-fin rays, dorsal-fin rays, pelvic-fin rays, caudal-fin rays, and vertebras.
The procedures for DNA isolation, PCR amplification and sequencing followed Qu et al. (2018). DNA was extracted using the standard phenol-chloroform protocol and the ethanol precipitation method and then stored at -20 °C. Polymerase chain reaction (PCR) was performed to amplify the partial fragment of the mitochondrial COI locus using a pair of primers (Fish F1, 5'-TCAACCAACCACAAAGACATTG-GCAC-3' and Fish R1, 5'-TAGACTTCTGGGTGGCCAAAGAATCA-3') (Ward et al. 2005). The thermal cycler program for PCR was 95 °C for 5 min, followed by 35 cycles of 94 °C for 30 s, 52 °C for 30 s and 72 °C for 45 s and a final extension at 72 °C for 10 min. The products were checked by electrophoresis on a 1% agarose gel to confirm the predicted fragment size and were then sequenced. The sequencing results were trimmed and manually proofread using SEQUENCHER 5.4.6 (http://www. genecodes.com) software. All sequences in this study were deposited in GenBank, and the accession numbers are shown in Suppl. material 1: Table S1.
Due to the availability of data for other related species in GenBank, we chose the mitochondrial COI gene sequence to calculate intraspecific and interspecific genetic distances and perform maximum likelihood (ML) and Bayesian analyses in this study. The intraspecific and interspecific genetic distances were generated using the Kimura two-parameter (K2P) distance model with MEGA 7 (Kumar et al. 2016). For the phylogenetic analyses, Epinephelus akaara (GenBank No. MF185437) and Epinephelus awoara (GenBank No. MF185456) were used as the outgroups because they are located in a clade sister to the one containing the E. chlorostigma species-complex (Ma et al. 2016). jModelTest 2.1.9 was used to infer the best evolutionary model, and the TrN+I+G model was selected based on both the Akaike information criterion (AIC) and the Bayesian information criterion (BIC) (Darriba et al. 2012). ML phylogenetic analysis was performed with the PhyML 3.1 program with 1000 bootstrap replicates (Guindon and Gascuel 2003), and the Bayesian phylogenetic analysis was performed by using MrBayes 3.2.6 (Ronquist et al. 2012).
Mouth large and lower jaw slightly projecting and oblique. Lower jaw 3.4 (3.4-3.9) in head length; upper jaw 2.4 (2.3-2.5) in head. Maxilla slightly extending to rear edge of eye and posterior edge of maxilla slightly rounded. Maxilla width 8.1 (7.9-9.2) in head. One or two pairs of canine teeth at anterior part of the upper and lower jaw. Teeth of lower jaw form two rows and expand anteriorly into three rows; teeth in the outer side are larger than the inner side. Villiform teeth present on vomer and palatines. Tongue slender and sharp at tip. Longest gill raker was greater in length than longest gill filament. Nostrils round and posterior nostril larger than anterior nostril. Anterior nostril with a membranous flap. Three spines on operculum, topmost and undermost small, the middle the largest. Tip of middle spine extending farther towards tail than tip of lower spine. Upper edge of opercular membrane slightly convex coming to a rounded point posteriorly. Preopercle rounded with four to five prominent spines at angle and with numerous fine serrae while increasing in size downward. Lateral line starting from posterior opercle and slightly arched over pectoral region. Scales on head, thorax, abdomen, anterodorsal part of body and fin membranes weakly ctenoid. Auxiliary scales absent. Small scales present on inner margins of dorsal, pectoral, pelvic, and caudal fins and not extending to the rear margin area.
Coloration in life (based on photographs of the fresh holotype and paratypes). Head (except chest), body (except abdomen), and fins (pectoral fin only basally) with numerous, irregular, close-set, dark brown spots becoming more widely spaced on the lower part and with the ground color forming a pale network (Fig. 1a); dorsal fin, caudal fin and anal fin dark brown; pectoral fin translucent with reddish brown to light yellowish-brown; body sometimes with four faint, irregular, discrete dark bars; rear margin of the caudal fin without a narrow white line.
Coloration in preservative. Body yellowish-brown to tan with close-set spots remaining prominent or faded (Fig. 1b, c). Dorsal, caudal, and anal fins dark brown. Pectoral fin pale and opaque.
Genetic analyses. Mitochondrial COI gene sequences were obtained from nine specimens of E. tankahkeei. Several sequences of related species were also sequenced in this study or obtained from GenBank. E. tankahkeei has 13 speciesspecific mutations at nucleotide positions 126, 216, 222, 249, 276, 372, 414, 519, 525, 528, 558, 567, and 576 (Table 2). The intraspecific mean distance of E. tankahkeei was 0.0028. The interspecific mean distances indicated that E. tankahkeei differs from E. chlorostigma by 0.0621, from E. polylepis by 0.0771, from E. gabriellae by 0.1263, from E. miliaris by 0.0904, from E. geoffroyi by 0.1219, and from E. areolatus by 0.0855 (Table 3). Phylogenetic trees using both maximum likelihood and Bayesian inference showed almost complete agreement, with E. tankahkeei forming a monophyletic clade that excluded all other closely related species (Fig. 3).
Distribution and habitat. The new species was recently observed in the South China Sea and Taiwan Strait. Similar to other congeners, E. tankahkeei is a reef-associated species that feeds on fishes and invertebrates.

Other closely related species in this study C A, G T T, C A A A A, G C G, A A, G, C A A, T
Etymology. Epinephelus tankahkeei is named after Tan Kah Kee (1874-1961), who was a famous overseas Chinese educator, philanthropist, and social activist and the founder of Xiamen University and Jimei School, in honor of his significant contribution to the motherland. Table 3. Analysis of the intraspecific and interspecific (K2P model) distances, interspecific distances (in lower left) and standard errors (in upper right) based on the COI locus between Epinephelus tankahkeei and closely related species; IMD = Intraspecific mean distance; SE = standard error.

Discussion
Epinephelus chlorostigma was formerly reported to have a wide distribution range from the Red Sea and the coast of Africa to the western Pacific Ocean. It was considered a species complex (the E. chlorostigma species complex) (Heemstra and Randall 1993). Since the early 1990s, new species have been successively distinguished from E. chlorostigma and described. Epinephelus gabriellae has a restricted range from Oman to Somalia and differs in having fewer dorsal-fin rays (14-15 vs. 16-18). Epinephelus polylepis is distributed from the western coast of India to the coast of Yemen and has more lateral-line scales and lateral-scale series (65-72 and 126-137 vs. 48-53 and 96-122, respectively) (Randall and Heemstra 1991). Epinephelus geoffroyi is local to the Red Sea and has more gill rakers (25-29 vs. 23-26) (Randall et al. 2013). Interestingly, the three recently described species above are all distributed to the west of the Indo-Australian Archipelago (IAA), even though the type locality of E. chlorostigma is the Seychelle Islands in the Indian Ocean. Currently, E. tankahkeei collected from the China Seas can be morphologically distinguished from E. chlorostigma by its rounder anal fin, closer dark spots on the body, lack of dark spots on the abdomen, and lack of a narrow, pale whitish posterior margin on the caudal fin. Our molecular analyses also corroborated the morphological results. In E. tankahkeei, 13 species-specific mutations were found in the COI gene fragment (Table 2). Genetic distance analysis also revealed high divergence between E. tankahkeei and its closely related species. The interspecific mean distance between E. tankahkeei and E. chlorostigma was 0.0621, which was greater than the distance (0.0545) between E. chlorostigma and E. polylepis (Table 3). The phylogenetic analyses performed with both ML and Bayesian inference also revealed a distinct monophyletic group formed by all samples of E. tankahkeei, and this group was separated from E. chlorostigma, E. polylepis, and E. areolatus. Although the phylogenetic relationships of the four species could not be well resolved by only the COI gene, each of the species formed a monophyletic clade with high support, supporting their validity (the ML bootstrap value was 94% for E. tankahkeei, 99% for E. chlorostigma, 100% for E. polylepis, and 100% for E. areolatus) (Fig. 3).
In the China Seas, the first record of E. chlorostigma was Serranus reevesii Richardson, 1846, type locality Canton, China (based on a painting by John Reeves) (Richardson 1846), which was later treated as a synonym of E. chlorostigma (Heemstra and Randall 1993). However, it is difficult to confirm the validity of S. reevesii due to its unclear description and lack of reliable photos and type specimen. As mentioned by Randall and Heemstra (1991), there are no confirmed records of E. chlorostigma in the continental waters of Asia. Our sampling in the China Seas over the last 20 years also never resulted in any E. chlorostigma specimens. Together with the distribution range, morphological characteristics, and molecular data, we suppose that most, or even all, of the former records of E. chlorostigma in the China Seas might be misidentifications of E. tankahkeei. More samples should be taken in the future to verify the distribution range of E. chlorostigma.