Record of the rare Caribbean mud eel, Pythonichthys sanguineus (Heterenchelyidae, Anguilliformes), in the region of the Amazon Reef

As they spend most of their time buried in the substrate and are not a fishery resource, heterenchelyids are seldom seen. These eels are characterized by their greatly reduced eyes, which are covered by semi-transparent skin, the absence of a pectoral fin or lateral line, and no pores on the head or body. Pythonichthys sanguineus is a particularly poorly-known species, with only eight scientific records from Cuba, Puerto Rico, Colombia, Venezuela, Guyana, and Suriname . The present study is based on six adult specimens of P. sanguineus captured by vessels of the shrimp trawling fleet along the northern Brazilian coast, in the vicinity of the Amazon Reef. We provide meristic, morphometric and DNA barcoding data. These findings provide insights into the distribution of the species off the northern Brazilian coast and contribute to the discussion about the southern limit of the Greater Caribbean fauna.


INTRODUCTION
The order Anguilliformes includes 938 species in 19 families and 159 genera (Nelson et al. 2016). The marine mud eels of the family Heterenchelyidae, a sister group to the Myrocongridae and Muraenidae (Inoue et al. 2010), are a small group of two genera, Panturichthys Pellegrin 1913 and Pythonichthys Poey 1868, found in the Mediterranean, tropical Atlantic, and eastern Pacific (Nelson et al. 2016). As they spend most of their time buried in the substrate, heterenchelyids are seldom seen, not representing a fishery resource (Smith 2015). These eels, which can reach a total length of 1 m, are characterized by their greatly reduced eyes, which are covered by semi-transparent skin, the absence of a pectoral fin or lateral line and no pores on the head or body (Smith 1989).
At least four species of Pythonichthys are known to exist. Pythonichthys microphthalmus (Regan 1912) and Pythonichthys macrurus (Regan 1912) are found in the Eastern Atlantic (EA), while Pythonichthys asodes Rosenblatt & Rubinoff 1972 occurs in the Eastern Pacific (EP), and Pythonichthys sanguineus Poey 1868 is the only species found in the Western ACTA AMAZONICA Atlantic (WA) (Figure 1a). Smith (1989) described the history and diversity of the heterenchelyids and redescribed the osteology of P. asodes and P. macrurus. Eagderi and Adriaens (2010) also described the osteology of these species, Blache (1977) described the leptocephalus larvae from the Gulf of Guinea. Pythonichthys sanguineus, a senior synonym of Heterenchelys biaggii Böhlke (1956) (sensu Smith 1989Smith et al. 2012) is the least well known species of the genus, and nothing is known of its biology (Smith 1989(Smith , 2002Smith et al. 2012). Only eight records of adult specimens of P. sanguineus are found in zoological collections (in 10 collection lots), representing six different localities. Two specimens are known from Cuba (MCZ 32805 and USNM 111340) and three from Puerto Rico, including the type specimens of Heterenchelys biaggii (FMNH 61775, ANSP 73872 and FMNH 61774). From South America, five specimens are available from Venezuela (four specimens, MBUCV 293 and one specimen from the Fernandez Yepez's collection), and one each from Colombia (UMML 222150), Guyana (BMNH 1961.8.31.99) and Suriname (NSMT 40060), while the 11 specimens examined by Poey (1868) have been lost (Faloh-Gandarilla et al. 2016).
Considering the general lack of data on the fish diversity of the northern coast of Brazil (Marceniuk et al. 2013;Marceniuk et al. 2017), the Centro de Pesquisa e Gestão de Recursos Pesqueiros do Litoral Norte (CEPNOR), together with taxonomists from the Museu Paraense Emílio Goeldi (MPEG), in Belém, Brazil, has obtained specimens from the industrial shrimp otter trawling fleet based in the Brazilian state of Pará, in a concentrated effort to expand the scientific inventory of the marine fish diversity of the region. This sampling effort has resulted in the collection of six specimens of P. sanguineus. Meristic, morphometric and DNA barcoding data are presented (cataloged specimens), and provide important insights into the distribution and diversity of the genus in the western Atlantic.

MATERIAL AND METHODS
The present report on Pythonichthys sanguineus is based on three adult specimens collected in Amapá state on April 28, 2017 andMarch 2-3, 2018 (cataloged, Figure1b,c) (Smith 1989, 2002, Smith et al. 2012, and measured according to Smith (1989). All measurements were taken using either a digital calliper (precision 0.01 mm) or a ruler (precision 1 mm) for dimensions over 200 mm. The data on the occurrence of P. sanguineus were obtained from the zoological collection records available at http://www.fishnet2.net and in Smith (1989).

Molecular analysis
Molecular identification of the specimens was done according to the DNA barcoding protocol (Hebert et al. 2003) and deposited in the BOLD system (Ratnasingham and Hebert 2007) and in Genbank (https://www.ncbi.nlm.nih.gov/ genbank). A partial COI sequence (578 bps) was obtained from a single P. sanguineus specimen deposited at Museu Paraense Emílio Goeldi (MPEG 35269), and deposited in the Genbank database (accession number: MG833002) and BOLD (RMRBR001-18). For comparison, COI sequences of additional 57 anguilliforms were obtained from NCBI (see Supplementary Material, Appendix S1). These sequences were derived from a phylogenetic study that analyzed the mitochondrial genomes of all 19 anguilliform families (Inoue et al. 2010) and represent an almost complete sample of Anguilliformes at genus level.
The molecular assays were conducted at the Laboratory of Fish Biology and Genetics at Universidade Estadual Paulista Júlio de Mesquita Filho in Botucatu, São Paulo, Brazil. Total genomic DNA was isolated from muscular  (Ivanova et al. 2008). The fragment at the 5' end of the mitochondrial COI gene was amplified by PCR using the primers FishF1 and FishR1 (Ward et al. 2005). The reactions were carried out in a 12.5 µL reaction volume containing 1.25 µL of 10 × PCR buffer, 0.375 µL MgCl2 (50 mM), 0.5 µL of the dNTPs (2 mM), 0.25 µL of each primer (5 µM), 0.2 µL of 1.25 U Taq platinum DNA polymerase, and 1 µL of the DNA template (200 ng), and 8.675 µL of ultrapure water. The PCR cycle was as follows: 95°C for 3 min, followed by 35 cycles of 95°C for 30 s, 54°C for 45 s, 72°C for 1 min, and then 72°C for 5 min. The PCR products were visualized in a 1% agarose gel and purified using ExoSAP-IT (USB Europe GmbH, Staufen, Germany) at 37°C for 60 min followed by 15 min at 80°C. The sequencing reactions were conducted using a BigDye Terminator Cycle Sequencing kit (Applied Biosystems, California, USA) according to the manufacturer's instructions. The sequences were read in both the forward and reverse directions, and the primers used for sequencing were the same as those used for the PCR reaction. Finally, the labeled sequences were assessed using an ABI PRISM 3130 Genetic Analyzer (Applied Biosystems, California, USA).
Bidirectional sequences were assembled using Geneious v.5.6 (Kearse et al. 2012) to obtain consensus sequences and to check for indels or stop codons. Indels or stop codons, which may indicate the amplification of nuclear mithocondrial DNA segment (NUMTs), were absent from all sequences. Genetic distances between and within taxonomic groups were calculated using the Kimura 2-parameter (K2P) substitution model (Kimura 1980) in MEGA 7 (Kumar et al. 2016). The best model of nucleotide substitution was GTR+G+I as calculated in MEGA 7 (Kumar et al. 2016). The maximum likelihood tree was obtained in MEGA 7 (Kumar et al. 2016) using the GTR+G+I model, with node supports estimated by 1000 nonparametric bootstrap replicates (Felsenstein 1985). While counts of vertebrae and proportional measurements can be used to distinguish P. sanguineus from P. macrurus (EA) and P. asodes (EP), these parameters are highly similar to those observed in P. microphthalmus (EA). Pythonichthys sanguineus can nevertheless be distinguished from P. microphthalmus by the number of vomerine tooth rows (2-3), in contrast with a single row in P. microphthalmus, and 3-5 maxillary and mandibular tooth rows (vs. 2 rows in P. microphthalmus). The two species are also differentiated clearly by the molecular data (see below).

Molecular analysis
Maximum likelihood analysis showed that the specimen analyzed here clusters with Pythonichthys microphthalmus, with a bootstrap support of 60% (Figure 2), corroborating the assignment of our specimens to the genus Pythonichthys. The K2P genetic distance between P. sanguineus and P. microphthalmus was 25.0 ± 1.9%, with 89/578 positions being different between species, corroborating our hypotheses that these are in fact different species.

Morphological measurements
The morphometric and meristic parameters of the specimens analyzed in here did not exceed the amplitude of values presented in the original description or subsequent studies (Table 1). Scharpf and Lazara (2017) refer to the coloration of fresh Pythonichthys sanguineus as blood-red. The living specimens (Figure 1d,e) presented the same coloration described by Scharpf and Lazara (2017), while the fresh specimen was light gray, with a brownish anterior dorsal region extending to approximately the anus, and a pale belly (Figure 1b,c). The distal portions of the dorsal and anal fins were blackened posteriorly from the median region of the body.   (Smith, 1989), expressed in millimeters (mm), percent of total length (TL%) or percent of standard length (SL%

Distribution
Tropical Western Atlantic Ocean, from Cuba to Brazil ( Figure  3). The present study extends the occurrence of the Atlantic mud eels from Suriname to Amapá state in northern Brazil.

Doubtful records from the Brazilian coast
The occurrence of Pythonichthys sanguineus in Brazilian waters was reported by Miranda-Ribeiro (1919). This record was based on the observation (no specimen collected) of an eel at Trindade Island, off the eastern coast of Brazil, which was identified as Pythonichthys sanguineus based on its intense red coloration. This record was considered doubtful by Simon et al. (2013) and was not accepted by Pinheiro et al. (2015). Hudson Pinheiro (personal information) confirmed that the species does not occur at Trindade Island ( Figure 3).

Biology, habitat, depth, and presumed schooling pattern
There is very little information available for this species, that inhabits muddy substrates of reefs and rocky areas (Smith 2015). Pythonichthys sanguineus has previously been recorded at depths of 12-55 m. The cataloged specimens were captured on a gravel bottom, at a depth of between 44 and 50 m, at 2°19'42"N,48°36'46"W to 1°52'45"N,48°16'32"W (Figure 3), and the non cataloged specimens were captured on a gravel bottom, at a depth of 70 m, at 2°26'73"N, 48°41'52"W ( Figure 3), both sites in the vicinity of the Amazon Reef (Moura et al. 2016;Francini-Filho et al. 2018).

Fisheries and conservation
Otter trawling is practiced worldwide, and is characterized by the largest discarded catches, causing profound impacts on the environment through the destruction of bottom substrates and the elimination of benthic organisms such as coral reefs (Stobutzki et al. 2001). In this context, the capture of the mud eel specimens reported here always ocurred after the end of the closure periods determined by the government to protect shrimp stocks. No other specimens were caught after the beginning of the shrimping season, indicating that the fishing activity may have influence on the occurrence of the species in the fishing zone, as this pattern was observed during the two sampling seasons. This question clearly requires further investigation, given the well-known impact of otter trawling on benthic communities.

DISCUSSION
Pythonichthys sanguineus is the only species of the genus found in the tropical western Atlantic, and is known from localities in the Caribbean and northern South America, being recognized as part of the typical Greater Caribbean fauna (sensu Robertson and Cramer 2014). The present record of P. sanguineus from the vicinity of the Amazon Reef (Moura et al. 2016;Francini-Filho et al. 2018) is relevant to the discussion on the range of the southern limit of the Greater Caribbean fauna, which has been inferred by different authors to reach the region from the Gulf of Paria in Venezuela to Suriname (Petruch 2013) or specific points on the Brazilian coast (Robins 1971), an imprecision that stems from the lack of any prominent physical barriers in the region. In this context, the Amazonian-Orinoco plume, formed by the rearrangements of the hydrographic basins of the Amazon (Shepard et al. 2010) and the Orinoco (Rod 1981), during the middle to late Miocene (7.2 mya) (Figueiredo et al. 2009;Hoorn et al. 2017), has been recognized as an important barrier to the dispersal of coral reef fishes of the Caribbean and Brazil (Bernal and Rocha 2011). Aguilera et al. (2014) described Pythonichthys pirabasensis based on a fossil from the Aquitanian period of the early Miocene (20.4-23.0 mya), from the Pirabas Formation of Atalaia Beach (municipality of Salinópolis, in the northern Brazilian state of Pará) and concluded that the current interruption of the shallow marine bioprovince, found off the Amazon delta cannot be perceived in the fish fauna of the Pirabas Formation, which is closely related to that of the Gatunian/proto-Caribbean bioprovince (Petuch 1982(Petuch , 2013Landau et al. 2008). The shift of the mouth of the Amazon River, close to its present day location, has terminated the carbonate cycle of the Pirabas Formation, pushing the Pirabas fish fauna northwards, including Pythonichthys and the congrid genus Paraconger Kanazawa 1961, also captured with the Pythonichthys specimens described herein (A.P. Marceniuk, unpublished data).
Our record of Pythonichthys sanguineus indicates that the shift in the location of the Amazon plume may not have displaced the fauna of the Pirabas Formation further north, but only modified the distribution of these lineages from the coastal zone further offshore, to an area where environmental conditions are more similar to those of the Pirabas Formation of the early Miocene (Aquitanian). This area coincides with the recently-discovered coral reefs off the mouth of the Amazon. The fauna associated with the reefs is still poorly-known, but is possibly closely related to those of the Pirabas Formation and the present-day Caribbean (Robertson and Cramer 2014).