Unveiling Two Rhinogobius Species, Rhinogobius chiengmaiensis and Rhinogobius mekongianus, in Thailand Using Morphological and Molecular Identification Methods

1. Introduction

The family Oxudercidae (or Gobionellidae) belongs to the order Gobiiformes, and comprises approximately 86 genera and 598 species that are found worldwide from freshwater to marine habitats [1]. Fish species in this family were previously classified in the subfamily Oxudercinae of the family Gobiidae according to their taxonomic arrangement following Nelson [2]. Later, when the taxonomy of fish in this taxa was verified, it was revised that the subfamily Oxudercinae was separated from the family Gobiidae with the status of the family Oxudercidae [1]. One distinguishing feature of the family Oxudercidae, as compared to the family Gobiidae, is the presence of elongated and slender suspensoria structures [1,3].

Rhinogobius, a genus within the family Oxudercidae, contains the most species of freshwater gobies found in the lakes and streams across East and Southeast Asia, including Thailand [4,5,6]. Rhinogobius species exhibit characteristics that are commonly observed in gobies, such as their small body size, the fusion of their pelvic fins to form a disc-like structure on the ventral side, and an elongated and rounded body [7,8]. Moreover, this genus exhibits an elongated form of the head, characterized by a long snout. Presently, the genus Rhinogobius is recognized to encompass more than 80 distinct species, and ongoing taxonomic research is consistently discovering and documenting new species such as R. sangenloensis in Southern China [5], R. maxillivirgatus in Eastern China [9], R. yangminshanensis in Taiwan [10], and R. aonumai aonumai and R. aonumai ishigakiensis in Japan [11]. In addition, two new species, R. rong and R. nami, have recently been described from central Vietnam [12]. However, Panitvong [13] notes that only three Rhinogobius species are currently recognized in Thailand: R. chiengmaiensis, R. mekongianus, and R. giurinus.

Rhinogobius chiengmaiensis exhibits a broad distribution in the upper region of the Chao Phraya river, whereas R. mekongianus is found inhabiting the Mekong river basin [14]. In 2013, Panitvong [13] found R. giurinus within a stream located in Chiang Rai province, Northern Thailand. Nevertheless, there remains limited documentation of the presence of this species in Thailand. Therefore, this study focused on two species, R. chiengmaiensis, and R. mekongianus. Although these two species are found in different natural habitats, they share some similarities in appearance, such as small size, elongated body shape, and the pattern of colors. They are currently becoming popular as ornamental fish, causing the natural population trend to decrease, especially in the case of R. chiengmaiensis. The conservation statuses of R. chiengmaiensis and R. mekongianus were assessed by the International Union for Conservation of Nature (IUCN) as vulnerable (VU) and least concern (LC), respectively (as of February 2025). Hence, both the morphological and molecular identification methods were required to distinguish between these two species.

The morphological identification of fish species involves the assessment of meristic and morphometric characteristics. Meanwhile, the molecular identification, DNA barcoding or the standard sequences of the cytochrome c oxidase I (COI) gene have been commonly employed for global species bioidentification [15], including the case of fishes [16,17,18,19,20]. Both morphological identification and DNA barcoding have distinct advantages and disadvantages. Morphological identification is typically conducted by experienced fish taxonomists, while DNA barcoding is employed for identifying fish species even in cases of incomplete samples such as fish fillets [21], or the various developmental stages such as the eggs [22] or larval stage [23]. Collaboration between morphological identification and DNA barcoding enhances the accuracy of fish species identification and contributes to the expansion of the COI gene sequences in public databases for further analysis.

The objective of this research was to differentiate between two species of the genus Rhinogobius, namely R. chiengmaiensis and R. mekongianus, through the application of both morphological identification and DNA barcoding techniques. Furthermore, the research also focused on studying two other species within the family Oxudercidae, namely Eugnathogobius siamensis and Pseudogobiopsis oligactis. These two species were analyzed in comparison to two species of the genus Rhinogobius. This research provided the first report of the COI sequences of R. chiengmaiensis, R. mekongianus, and E. siamensis. The COI sequences of all four species were expanded and deposited into the GenBank database to serve as reference sequences for aiding in the identification of unknown species and establishing a database for specifying of new species within the family Oxudercidae, especially the genus Rhinogobius.

2. Materials and Methods

2.1. Sample Collection

Seven live samples of two species belonging to the genus Rhinogobius, R. chiengmaiensis and R. mekongianus were obtained from collectors of fish from nature with known sources (Table 1). In addition, four samples of two other species of the family Oxudercidae, E. siamensis and P. oligactis, were also received. All samples were photographed and anesthetized with 0.2 g/L of MS-222 (Sigma, Missouri, USA) solution. Whole samples were preserved in absolute ethanol at room temperature for morphology identification. This animal use protocol has been approved by the Institutional Animal Care and Use Committee, University of Phayao (No. UP-AE 1-011-67).

2.2. Morphological Identification

All samples were first identified based on morphological appearances including body shape, mouth and eye size. Counts and measurements following Nagao Natural Environment Foundation [24] were performed on the left side of the samples. Fin rays were counted for the dorsal, pectoral, pelvic, and anal fins. The longitudinal, predorsal, and circumcaudal scales were also recorded. Measurements were taken point to point with digital vernier calipers to the nearest 0.01 mm. Subunits of the head were presented as proportions of head length (HL). Measurements of head and body parts were provided as proportions of standard length (SL).

2.3. Molecular Identification

After morphological identification, a small piece of right pectoral fin from each sample was used for the extraction of the genomic DNA using the GF-1 Nucleic Acid Extraction kit (Vivantis Technologies Sdn. Bhd., Selangor Darul Ehsan, Malaysia). The quantity and quality of the extracted DNA was determined using a NanoDrop Microvolume Spectrophotometer (Thermo Fisher Scientific Inc., MA, USA) and 1.2% agarose gel electrophoresis at 100 V for 30 min, respectively.

A 707-bp fragment of the COI gene was amplified for each sample using the equal volume of each 10 µM primer, namely FishF1 (5´-TCAACCAACCACAAAGACATTGGCAC-3´), Fish F2 (5´-TCGACTAATCATAAAGATATCGGCAC-3´), FishR1 (5´-TAGACTTCTGGGTGGCCAAAGAATCA-3´), and FishR2 (5´-ACTTCAGGGTGACCGAAGAATCAGAA-3´) [20], with the PCR technique. Each PCR reaction mixture included of 1 µl of extracted DNA (approximately 100 ng/µl), 2.25 µl of 10x Taq buffer, 0.8 µl of 50 mM MgCl₂, 1 µl of 2.5 mM dNTP mix, 1 µl of 10 µM mixed primers, 0.2 µl of 1 U Taq DNA polymerase (BIO-HELIX Co., Ltd., New Taipei City, Taiwan), and nuclease-free water to 25 µl. The thermal cycling conditions included an initial denaturation at 95 °C for 2 min followed by 35 cycles of denaturation at 94 °C for 30 s, annealing at 54 °C for 30 s, and extension at 72 °C for 1 min, with a final extension at 72 °C for 10 min. The PCR products were analyzed using 1.2% agarose gel electrophoresis stained with Novel Juice (BIO-HELIX Co., Ltd., New Taipei City, Taiwan). The target bands were then purified to remove unwanted reagents using the HiYield™ Gel/PCR DNA Fragments Extraction kit (RBC Bioscience Corp., New Taipei City, Taiwan), following the manufacturer's instructions.

The purified PCR products were sequenced bidirectionally at ATGC Co., Ltd. (Thailand) using primer FishF1/FishF2 for the 3' end and FishR1/FishR2 for the 5' end. Two sequences of each sample were aligned and assembled using the Clustal Omega program [25]. All sequences covering the primer sequences were searched for the stop codon, deletion, and insertion using the Open Reading Frame Finder website (ORF finder; https://www.ncbi.nlm.nih.gov/orffinder/). The scientific name of each sample was investigated by comparing its sequence with reference sequences in two public databases, namely the GenBank (https://www.ncbi.nlm.nih.gov/) using the Basic Local Alignment Search Tool (BLAST) program [26] and BOLD (https://boldsystems.org/). Each sample was assigned a scientific name based on the top match with a minimum of 99% sequence similarity.

The percentage of base composition of the four gobiid species was calculated using the MEGA11 program [27]. The intraspecific and interspecific genetic distances were calculated using the MEGA11 program with the Kimura-2-parameter (K2P) distance model [28]. Because the COI sequences of R. chiengmaiensis, R. mekongianus, and E. siamensis have not been reported in any databases, several sequences of other fish species of the genus Rhinogobius and P. oligactis were downloaded from GenBank database for construction of a maximum likelihood (ML) tree [29] to analyze their evolutionary relationship. The tree was generated using the MEGA11 program, employing 1000 bootstrap replications. A COI sequence from Oxyeleotris marmorata (OK670610), a member of the family Butidae, was used as an outgroup.

3. Results

3.1. Morphological Identification

Although these two species of the genus Rhinogobius showed very similar characteristics, their appearances were clearly different (Table 2), including two other members of the family Oxudercidae, E. siamensis and P. oligactis. A short description of the four gobiid species examined in this study is provided below.

Rhinogobius chiengmaiensis Fowler, 1934

D1. vi-vii; D2. i,7-8; P1. i,14; P2. 12 (total); A. i,6-7; C. branched 10

First dorsal fin with 6-7 single rays; i,7-8 second dorsal fin: i,14 pectoral fin rays: 12 (total) pelvic fin rays: i,6-7 anal fin rays: 10 caudal branched rays; ctenoid scale; longitudinal scales 29-30; predorsal scales 3-4; circumcaudal scales 12; body elongate, moderate slender (15.85 ± 0.31 %SL); pelvic fins origin slightly in front of opercula margin (29.71 ± 2.56 vs. 31.82 ± 1.67 %SL); pelvic fins length 19.37 ± 1.15 %SL; mouth moderate large (35.50 ± 3.86 %HL), maxillary extending to middle of eyes; eyes diameter 20.16 ± 1.39 %HL.

Rhinogobius mekongianus (Pellegrin & Fang, 1940)

D1. vi; D2. i,7-8; P1. i,14; P2. 12 (total); A. i,6-7; C. branched 10

First dorsal fin with 6-7 single rays; i,7-8 second dorsal fin: i,14 pectoral fin rays: 10 (total) pelvic fin rays: i,6 anal fin rays: 12 caudal branched rays; ctenoid scale; longitudinal scales 29-30; predorsal scales 3-4; circumcaudal scales 12; body elongate, very slender (13.28 ± 0.54 %SL); pelvic fins origin slightly in front of opercula margin (29.33 ± 0.18 vs. 30.50 ± 0.38 %SL); pelvic fins length 17.88 ± 1.17 %SL; mouth moderate large (39.19 ± 2.06 %HL), maxillary extending to middle of eyes; eyes diameter 19.74 ± 0.57 %HL.

Eugnathogobius siamensis (Fowler, 1934)

D1. vi; D2. i,6-7; P1. i,16-18; P2. 10 (total); A. i,6; C. branched 12-14

First dorsal fin with 6 single rays; i,6-7 second dorsal fin: i,16-18 pectoral fin rays: 10 (total) pelvic fin rays: i,6 anal fin rays: 12-14 caudal branched rays; large ctenoid scale; longitudinal scales 24; predorsal scales 7; circumcaudal scales 12; body elongate, moderate slender (18.08 ± 2.29 %SL); pelvic fins origin slightly behind opercula margin (33.31 ± 1.54 vs. 32.60 ± 1.24 %SL); pelvic fins length 24.85 ± 0.35 %SL; caudal fin length 22.61 ± 3.17 %SL; mouth large (48.33 ± 12.41 %HL), maxillary extending well beyond posterior margin of eyes in male and extending to middle of eyes in female; small eyes, diameter 13.16 ± 0.72 %HL.

Pseudogobiopsis oligactis (Bleeker, 1875)

D1. vi; D2. i,6-7; P1. i,14; P2. 12-14 (total); A. i,7-8; C. branched 12-14

First dorsal fin with 6 single rays; i,6-7 second dorsal fin: i,14 pectoral fin rays: 12-14 (total) pelvic fin rays: i,7-8 anal fin rays: 12-14 caudal branched rays; large ctenoid scale; longitudinal scales 24-25; predorsal scales 7; circumcaudal scales 10; body elongate, moderate slender (15.88 ± 0.51 %SL); pelvic fins origin slightly behind opercula margin (36.06 ± 0.37 vs. 35.86 ± 3.28 %SL); pelvic fins length 22.75 ± 0.36 %SL; mouth large (63.24 ± 0.80 %HL), maxillary extending well beyond posterior margin of eyes; small eyes, diameter 12.63 ± 0.19 %HL. The diagnostic key to all four species gobiid in this study was provided as follows:

Diagnostic key to species of four gobiid in this study

1a

large ctenoid scale, longitudinal scales 24-25, predorsal scales 7; pelvic fins origin slightly behind opercula margin .…………………………………..….…………..…..... 2

1b

moderate ctenoid scale, longitudinal scales 29-30, predorsal scales 3-4; pelvic fins origin slightly behind opercula margin .……………………………….......................... 3

2a

big head, head length 35.86 ± 3.28 %SL, head width 62.30 ± 3.15 %HL; mouth large (63.24 ± 0.80 %HL), maxillary extending well beyond posterior margin of eyes both gender; pelvic fins length 22.75 ± 0.36 %SL …………..……… Pseudogobiopsis oligactis

2b

big head, head length 32.60 ± 1.24 %SL, head width 55.61 ± 3.85 %HL; mouth large (48.33 ± 12.41 %HL), in male maxillary extending well beyond posterior margin of eyes (57.10 %HL), and extending to middle of eyes in female (39.56 %HL), pelvic fins length 24.85 ± 0.35 %SL ……………...………….…..….. Eugnathogobius siamensis

3a

mouth moderate large, maxillary extending to middle of eyes in both gender (35.50 ± 3.86 %HL); body moderate slender, body depth at pelvic fins origin 15.85 ± 0.31 %SL; pelvic fins length 19.37 ± 1.15 %SL; caudal fin length 24.77 ± 0.54 %SL ………………………………………………………………… Rhinogobius chiengmaiensis

3b

mouth moderate large, maxillary extending to middle of eyes in both gender (39.19 ± 2.06 %HL); body very slender, body depth at pelvic fins origin 13.28 ± 0.54 %SL; pelvic fins length 17.88 ± 1.17 %SL; caudal fin length 27.34 ± 2.28 %SL …...……..…..…

………………………….……………………………………….. Rhinogobius mekongianus

3.2. Molecular Identification

The 707-bp fragments of the COI gene were successfully amplified from all samples using the PCR technique. The 235 amino acid residues were translated without stop codon, deletion, and insertion for any sequences. From sequence alignment, two sequences of R. chiengmaiensis showed 100% similarity (1 haplotype). While five sequences of R. mekongianus showed 99.4% similarity (4 haplotypes), four bases were different. Two COI sequences of R. chiengmaiensis were compared to reference sequences in the GenBank and BOLD databases, which top matched with Rhinogobius virgigena at 96.24% and 96.21% identities, respectively. Likewise, five sequences of R. mekongianus also top matched with R. virgigena at 95.92-96.24% and 95.89-96.21% identities in the GenBank and BOLD databases, respectively.

The COI sequences of E. siamensis and P. oligactis demonstrated complete similarity, with each species showing a haplotype. Sixteen bases were different from each species. For comparison with the reference sequences in the GenBank and BOLD databases, two COI sequences of E. siamensis were top matched with P. oligactis at 98.14% and 98.92% identities, respectively. Meanwhile, two COI sequences of P. oligactis were top matched with P. oligactis at 98.14% and 98.92% identities in the GenBank and BOLD databases, respectively.

The average base compositions of eleven COI sequences were T (29.0 ± 1.3%), C (28.6 ± 0.1%), A (22.9 ± 0.9%), and G (19.5 ± 0.5%), as shown in Table 3. The GC content was 48.2 ± 0.5% at all sites while the AT content was 51.8 ± 0.5%. Pseudogobiopsis oligactis presented the highest GC content (49.1 ± 0.0%), while R. mekongianus had the lowest (47.8 ± 0.1%).

The intraspecific genetic distances were 0.00% for R. chiengmaiensis, E. siamensis and P. oligactis, and 0.28% for R. mekongianus (Table 4), while the average intraspecific distance was 0.07%. In contrast, the interspecific genetic distances ranged from 0.86% to 16.63%. The lowest distance was between R. chiengmaiensis and R. mekongianus (0.86%). The highest distances were between R. chiengmaiensis and P. oligactis (16.63%) and R. chiengmaiensis and E. siamensis (15.38%), followed by R. mekongianus and P. oligactis (12.00%) and R. mekongianus and E. siamensis (11.02%), respectively. Meanwhile, the distance between P. oligactis and E. siamensis was 1.64%. The average interspecific distance was 9.59%. Thus, the interspecific genetic distance was 137-fold greater than the intraspecific genetic distance.

The ML phylogenetic tree presented the relationship between the COI sequences of the four gobiid species of this study and other species retrieved from the GenBank database (Figure 1). Two major clades were clearly delineated, one comprising Rhinogobius species and the other including E. siamensis and P. oligactis. Within these groups, R. chiengmaiensis showed the greatest similarity to R. mekongianus, while E. siamensis and P. oligactis clustered together. Additionally, the COI sequence of Oxyeleotris marmorata was distinctly separated from the other sequences.

4. Discussion

The morphometric characters of R. chiengmaiensis and R. mekongianus provided in this study aligned with findings from previous research [4,14,30,31]. Rhinogobius chiengmaiensis and R. mekongianus exhibited quite similar appearances; however, they differed in certain characteristics such as the number of rays in the pelvic fin, anal fin, and caudal branch, as well as the diameter of the eyes. Additionally, a notable distinguishing characteristic was the body shape, with R. mekongianus being more slender compared to R. chiengmaiensis.

The concise characterizations of E. siamensis and P. oligactis presented in this study were also consistent with the descriptions provided in the previous research [32,33]. Considerable confusion has arisen in accurately distinguishing between these two gobiid fish species due to their similar appearances, particular the coloration of their head and body [24] and greatly enlarged jaws [34]. In addition, these two species have been greatly confused, as evidenced by their synonyms being swapped between genera in FishBase, a global fish species database (https://www.fishbase.se/). At one time, the invalid names Pseudogobiopsis siamensis and Eugnathogobius oligactis were used for E. siamensis and P. oligactis, respectively. In 2009, Larson [32] conducted a review of the gobiid fish genera Eugnathogobius and Pseudogobiopsis, which significantly enhanced the ability to distinguish between these two species.

To accurately identify these four gobiid species, which exhibit highly similar external characteristics, it is essential to employ molecular data and DNA barcoding, alongside traditional taxonomic methods. This integrative approach ensures precise identification of these closely related fish species. Xia et al. [9] examined the COI sequences of the newly discovered species R. maxillivirgatus, revealed that it is closely related to its nearest species, R. wuyanlingensis, while still being distinct.

Due to low identity percentages, the COI sequences of two fish species belonging to the genus Rhinogobius in Thailand have not yet been reported in any databases. Thus, this study was the first investigation to report the COI sequences of R. chiengmaiensis and R. mekongianus. Additionally, the COI sequences of E. siamensis have also been initially provided. However, the top match identity percentages for the COI sequences of P. oligactis were below 99%, likely due to genetic variation influenced by their diverse habitats [35]. A total of eleven COI sequences were deposited in the GenBank database with accession numbers PQ193904-PQ193914 (Table 1).

The average AT content (51.8 ± 0.5%) was greater than the average GC content (48.2 ± 0.5%) due to the average T being the highest in the base composition of eleven sequences, followed by C and A, respectively. The highest average thymine base composition was exhibited in several fishes including freshwater fishes of Bangladesh [16], marine and coastal fishes of Bangladesh [17], fish species in the Taiwan Strait [18], four fish species in the family Notopteridae [19] and Wallago attu [35]. Furthermore, the average G content was the lowest, representing a clear pattern of anti-G bias [16,17].

Generally, the intraspecific genetic distances based on the COI gene for each animal species are usually less than 2% [36], including fishes [20]. The average intraspecific genetic distances of each species in this study ranged from 0.00% to 0.28%. Conversely, the average interspecific genetic distances were higher, ranging from 0.86% to 16.63%. The lower intraspecific and higher interspecific genetic distances suggested that the COI sequences were effective in distinguishing between the four gobiid species examined in this study. This investigation made use of observations in several studies, including fish species in the Taiwan Strait [18], four fish species in the family Notopteridae [19], and Cyprinidae fish in the midstream of the Yangtze river [37]. Furthermore, the 137-fold differences between the average interspecific and intraspecific genetic distance exhibited the species delimitation of the four species. According to previous studies, the difference was greater than the 214-fold difference observed in four fish species in the family Notopteridae [19]. Hebert et al. [38] suggested that a 10-fold COI sequence difference between the average interspecific and intraspecific differences serve as a criterion for animal species differences. These investigations demonstrated that the COI gene sequences were effective in distinguishing the same species from other species. Nevertheless, the small sample size of this study was one of its limitations.

The evolutionary relationship of the COI sequences among the four fish species and the other species is shown in Figure 1, according to the fish taxonomy of Nelson et al. [1]. The species in the family Oxudercidae, which consisted of Rhinogobius spp., E. siamensis, and P. oligactis, were clearly separated from O. marmorata, that belongs to the family Butidae. One major clade was the Rhinogobius species. The COI sequences of R. chiengmaiensis and R. mekongianus were the closest species, which agreed with their morphological identification, followed by R. virgigena and other Rhinogobius species. The second clade comprised E. siamensis and P. oligactis, indicating a close evolutionary relationship between these two species. The genera Eugnathogobius and Pseudogobiopsis were categorized as a part of the Mugilogobius-lineage, whereas the genus Rhinogobius was classified to the Acanthogobius-lineage, for which the phylogenetic analysis was similar to studies conducted by Agorreta et al. [39].

At present, these four gobiid species are increasingly popular as ornamental fish. They are often collected from the wild, which may lead to a decline in their natural populations. Furthermore, the IUCN has assessed one species, R. chiengmaiensis, as vulnerable (VU), indicating that this species is at considerable risk of becoming endangered in its natural habitat unless effective conservation measures are implemented [40]. The remaining species were assessed as least concern (LC). Therefore, promoting the commercial breeding of these fish can resolve this problem. A few research studies have been conducted on the breeding of these four gobiid fish species, including R. chiengmaiensis [41]. However, these fish have very similar appearances, making accurate species identification necessary before breeding. Furthermore, the results of this study will contribute to species identification and serve as information for future research on new fish species in the genus Rhinogobius.

5. Conclusions

The species R. chiengmaiensis and R. mekongianus have been identified through both morphological and molecular methods. Additionally, E. siamensis and P. oligactis, members of the family Oxudercidae, were also identified. The morphology of the four gobiid species exhibited distinctive characteristics, although some species displayed considerable similarity. For molecular identification, DNA barcoding proves to be an effective approach for distinguishing between these four species, considering both intraspecific and interspecific genetic distances as well as phylogenetic analysis. Moreover, this study is the first to report the COI sequences for R. chiengmaiensis, R. mekongianus, and E. siamensis. These findings can serve as a valuable database for future research, including breeding programs and the discovery of new species.