New species of the Corumbataia cuestae group (Siluriformes: Loricariidae) from the Rio Tocantins basin, with comments on its phylogenetic relationships

A new species of Corumbataia is described from Rio Maranhão, Rio Tocantins basin, central Brazil. The new species is distinguished from all congeners by the presence of a small, naked area on snout tip; by having the abdomen covered with small platelets forming a shield which reaches the lateral mid-ventral plates; by the anterior profile of the head rounded in dorsal view; by the lower lip not reaching the transversal line of the pectoral girdle; and by the presence of 28 or 29 vertebrae. High genetic divergence in mitochondrial cytochrome c oxidase subunit I (COI) further supports the validity of this new species. Our phylogenetic analysis shows a derived subclade in Corumbataia , herein named as the Corumbataia cuestae group, composed of the new species plus C. cuestae , C. tocantinensis , C. britskii , C. liliai , and C. lucianoi . This group is defined by having a conspicuous crest of hypertrophied odontodes on head; absence of the adipose fin or a single series of platelets at adipose-fin position; and anastomosis of the infraorbital and otic sensory canals over the pterotic-supracleithrum. Here we also restrict the distribution of C. tocantinensis to the Rio Araguaia basin.


INTRODUCTION
Loricariidae is the most diverse family of Siluriformes with almost 1000 valid species (Fricke et al., 2020) recognized mainly by having the body covered by bony plates, external tooth-like structures (odontodes), and a ventral mouth with lips forming an oral disk used to adhere to surfaces and to soft substrate for foraging (Greerinckx et al., 2007;Garg et al., 2010). Hypoptopomatinae is a subfamily of loricariids popularly known as cascudinhos, which contains 247 valid species (Fricke et al., 2020) arranged in six major clades: Hisonotini, Neoplecostomini, Otothyrini, Corumbataia clade, Otocinclus clade, and Hypoptopomatini (Roxo et al., 2019).
Detailed examination of specimens of Corumbataia from the Rio Maranhão in the Rio Tocantins basin indicated that these represent a new species of this genus, which is 3/16 ni.bio.br | scielo.br/ni formally described herein. We also provide and discuss a hypothesis for its phylogenetic position.

MATERIAL AND METHODS
Morphology. Measurements and counts were taken from the left side of 27 specimens and were made point to point to the nearest 0.1 mm with digital calipers. Nomenclature of body plates and osteology follow Schaefer (1997). Measurements and abbreviations follow Carvalho, Reis (2009), except for the measurement of body depth at dorsal-fin origin that were not included. We additionally included the following measurements: preanal length (from tip of snout until anal-fin insertion), base of dorsal-fin length (from anterior margin of dorsal-fin spinelet until insertion of last dorsal-fin ray), lower caudalfin spine length (from posterior margin of last lateral plate of ventral plate series to the tip of the lower caudal-fin spine), body width at anal-fin insertion (from left side of the body until right side of the body at anal-fin insertion), snout-opercle length (from tip of snout until opercle opening), and head width (from opercle opening of the left side of the body until opercle opening of the right side of the body). Morphometrics are given as percentages of standard length (SL), except for subunits of head expressed as percentages of head length (HL). Four specimens were cleared and stained (c&s) according to Taylor, Van Dyke (1985). Vertebral counts include the five vertebrae of the Weberian apparatus and the compound caudal centrum (PU1+U1) that was counted as one element. Counts of dorsal-fin rays include the spinelet as the first unbranched ray. Institutional abbreviations follow Sabaj (2019). Specimens were deposited at the Laboratório de Biologia e Genética de Peixes, Universidade Estadual Paulista, Botucatu, Brazil (LBP); Museu de Zoologia, Universidade de São Paulo, São Paulo, Brazil (MZUSP). Zoological nomenclature follows the International Code of Zoological Nomenclature (ICZN, 1999). Molecular analysis. Three specimens were used for the molecular analysis: LBP 19095, 3, Rio Tocantins, tissues 77006-08. Total DNA extraction was performed using the Wizard Genomic DNA Purification Kit (Promega) from ethanol-preserved muscle, fin, or liver. Partial sequences of the cytochrome c oxidase subunit I (COI) gene were amplified in a total reaction volume of 12.5 uL. Each reaction included 1.25 uL of 10 X Buffer, 0.25 uL of MgCL2 (50 mM), 0.2 uL dNTPs (2mM), 0.5 uL of each primer (5 mM), 0.1 uL of PHT Taq DNA polymerase (Phoneutria, Belo Horizonte, Brazil), 1.0 uL of genomic DNA (20 ng) and 8.7 mL ddH2O. The conditions for the PCR reaction consisted of an initial denaturation (5 min at 94ºC), followed by 30 cycles of chain denaturation (40s at 94ºC), primer hybridization (30 s at 50 -54º), nucleotide extension (1 min at 68ºC) and final extension (8 min at 68ºC). The amplified products were checked on 1% agarose gels and purified using ExoSap-IT (USB Corporation). DNA sequencing was conducted in an automatic sequencer ABI 3130 DNA Analyzer (Applied Biosystems).
All individual sequences for each species were initially assembled using the software Geneious 7. 1.4 (Kearse et al., 2012), and aligned by Muscle (Edgar, 2004) under default parameters. Alignments include the newly-generated sequences and the sequences available from the phylogenetic study of Roxo et al. (2014. To evaluate the 4/16 ni.bio.br | scielo.br/ni occurrence of substitution saturation in our molecular data, we estimated whether the Iss (index of substitution saturation) was significantly lower than Iss.cAsym (assuming asymmetrical topology) using the method described by Xia et al. (2003) and Xia, Lemey (2009) with the software DAMBE 5.3.38 (Xia, 2013). Nucleotide variation, substitution patterns, and genetic distances were examined using MEGA v.6.06 (Tamura et al., 2013).
The best-fit nucleotide substitution model for the entire data set was selected under the Akaike information criterion (AICc) and Maximum Likelihood (ML) analyses were performed using the software MEGA v.6.06 (Tamura et al., 2013). Bootstrap (BS) resampling (Felsenstein, 1985) was applied to assess support for individual nodes using 1,000 pseudoreplicates. Random starting trees were used for each independent ML tree search and all other parameters were set to default values. Bayesian inference (BI) (Huelsenbeck, Ronquist, 2001) was performed in MrBayes v.3.2 (Ronquist, Huelsenbeck, 2003), based on the model (GTR+I). Two independent runs of 10 million steps sampling tree every 1000th generation a tree was sampled. Genetic variation within and among species groups under the best fit nucleotide model was also calculated in the MEGA v.6.06 software (Tamura et al., 2013).

Genetic analysis.
We used 18 specimens of Corumbataia representing a total of nine species and used Curculionichthys paresi (Roxo, Zawadzki & Troy, 2014), C. oliveirai (Roxo, Zawadzki & Troy, 2014) (Regan, 1908) to root the tree (Tab. 1). The combined sequence data resulted in a matrix with 580 bp. All sequences are deposited in GenBank (Tab. 1). The matrix does not contain any insertions, deletions or stop codons. The best nucleotide substitution model selected for the matrix was TN93+I+G (AICc = 6339.934). The nucleotide frequencies under TN93+I+G model were 0.23 (A), 0.27 (T), 0.30 (C), and 0.17 (G). Saturation was not observed, considering that the Iss < Iss.c for all NumOTU in the Xia et al. (2003) and Xia, Lemey (2009) tests. Genetic distances in Corumbataia species and among each nominal species are shown in Tab. 2.
The gene tree represents a 50% majority-rule consensus obtained by maximum likelihood analysis (LogL = -3132.01, Fig. 1). Bayesian analysis resulted in 10001 trees of which the first 2500 were discarded and the remaining 7501 were used to perform the consensus tree. In our analysis the C. canoeiro, C. veadeiros, and C. anosteos are the first groups to diverge, respectively. Corumbataia acanthodela, C. tocantinensis, C. cuestae, C. britskii, C. liliai, and C. lucianoi grouped in a clade herein named Corumbataia cuestae group. Corumbataia acanthodela is sister of the remainder members of Corumbataiacuestae group. The phylogenetic analysis using the COI marker did not resolve the relationships among Corumbataia tocantinensis, C. cuestae, C. britskii, C. liliai, and C. lucianoi, which appeared as a polytomy.    fig. 2A, MCN 13462).

Diagnosis. The new species differs from all congeners, except for the members of
Corumbataia cuestae group by having a conspicuous pair of tufts with enlarged odontodes on the tip of the supraoccipital (vs. absence of crest of enlarged odontodes in C. anosteos, C. canoeiro, and C. veadeiros); by the absence of an adipose fin or single series of platelets at adipose fin position (vs. presence of an adipose fin in C. canoeiro, or single series of platelets in C. anosteos and C. veadeiros); by anastomosis of the infraorbital and otic sensory canals over the pterotic-supracleithrum (Fig. 3A) (vs. anastomosis of infraorbital and otic sensory canals over the sphenotic in C. anosteos, C. canoeiro, and C. veadeiros) (Fig. 3B). Additionally, the new species differs from all species of Corumbataia cuestae group by having a small naked area on snout tip (vs. large naked area on snout tip, Fig. 4). Moreover, Corumbataia acanthodela differs from C. cuestae, C. liliai, C. lucianoi, and C. britskii by having abdominal platelets reaching the mid-ventral lateral plates (vs. abdominal platelets far from reaching mid-ventral lateral plates); from C. cuestae, C. liliai, and C. tocantinensis by having two rounded and more anteriorly positioned hyaline areas on the caudal-fin (Fig. 2)      Lower surface of head naked. Head lacking ridges. Parieto-supraoccipital process elevated and with conspicuous pair of tufts of hypertrophied odontodes in specimens of all examined sizes. Predorsal region without ridges. Body elongate and compressed at caudal peduncle. Greatest body width at cleithral region, progressively narrowing anteriorly towards snout tip and posteriorly towards caudal-fin. Head and trunk covered by dermal plates, except for naked area around dorsal-fin insertion. Body dorsoventrally compressed. Dorsal profile convex from snout tip to posterior margin of parietosupraoccipital and slightly concave from that point to dorsal-fin origin. Dorsal profile slightly concave and descending from dorsal-fin origin to first upper procurrent caudalfin ray, rising posteriorly to insertion of caudal fin. Greatest body depth at unbranched dorsal-fin ray insertion. Cleithrum and coracoid exposed in ventral view, covered by odontodes. Arrector fossae completely enclosed by ventral lamina of coracoid.
Ventral profile straight and descending from snout tip to opercular region, slightly convex from opercular region to anal-fin origin and slightly concave from that point to lower procurrent caudal-fin ray origin. Lateral surface of body entirely covered by plates; dorsal series with 23-24 plates. Mid-dorsal plate series truncated with 18-19 plates not reaching end of caudal peduncle. Lateral plate series with 22-23 plates. Lateral line incomplete, with gaps in line of pores along mid-length of body. Midventral series of lateral plates well-developed, reaching middle of caudal peduncle (17-19 plates). Ventral plates series with 20-21 plates. Body plates covered with minute odontodes.
Dorsal fin ii,7; its origin slightly posterior to vertical with pelvic-fin origin. Unbranched rays of dorsal fin slightly convex. Tip of adpressed dorsal-fin rays surpassing anal-fin origin. Dorsal-fin spinelet small and rounded (4 c&s). Pectoral-fin rays i,6; tip of adpressed pectoral fin surpassing pelvic-fin origin. Pectoral-fin axillary slit small. Pelvic-fin rays i,5. Distal margin of pelvic fin straight to slightly convex; tip of adpressed pelvic-fin ray reaching anal-fin origin in mature males, but not in females. Adipose fin absent. Anal-fin rays i,5; distal margin slightly convex. Caudal-fin rays i,7-i,7. Slightly emarginated; unbranched rays of same size. Rays of all fins covered with pointed odontodes. Lowest body depth at caudal peduncle. Caudal peduncle ellipsoid in cross section, rounded dorsally and ventrally. Hypurals elements fused in a unique hypural plate. Total vertebrae 28 or 29 (4 c&s).

Color in alcohol.
Background color of dorsal region of head and trunk brown. Four dark brown saddles along dorsal portion of body: first at dorsal-fin origin, second at end of dorsal-fin base, third at middle of caudal peduncle, and fourth reaching anteriormost caudal procurrent ray. Unpigmented portion of snout appears as two parallel hyaline stripes from rostral plate to nares. Mid-lateral dark brown stripe extending from tip of snout to caudal peduncle. Ventral portion of body almost entirely yellowish, except for dark, randomly-distributed chromatophores and for a concentration of chromatophores at anal-fin origin. Dorsal, pectoral, and pelvic fins with dark, irregularly distributed chromatophores. Caudal-fin brown, with two smaller rounded hyaline areas anteriorly, and two large rounded hyaline areas posteriorly (Fig.1).
(2) Absence of an adipose fin or platelets at the adipose-fin region. The plesiomorphic species Corumbataia canoeiro possesses a developed adipose fin, whereas in C. veadeiros and C. anosteos, the next lineages to diverge in Corumbataia, there is a single series of platelets at the adipose-fin position. This suggests that the adipose fin, present in C. canoeiro, was reduced in C. veadeiros and C. anosteos and totally lost in the species of the Corumbataia-cuestae group. The primitive condition in Corumbataia clade is the presence of adipose fin or platelets at adipose-fin region, present in the first lineages to diverge in the Corumbataia clade: Nannoplecostomus eleonorae and Microplecostomus foresti (Roxo et al., 2019). Curculionichthys, the sister to Corumbataia, exhibits the derived condition (absence of platelets or adipose fin), suggesting that this derived condition occurs homoplastically in Curculionichthys and in the species of Corumbataia cuestae group.
(3) Anastomosis of the Infraorbital and otic canals over the pterotic-supracleithrum (Fig. 3a). This derived condition was proposed by Britski (1997) to Corumbataia cuestae and C. tocantinensis, and here was also observed in all members of Corumbatia cuestae group. The plesiomorphic condition is the anostomoids of the infraorbital and otic canals over the sphenotic (See fig. 3A in Ribeiro et al., 2012), and is found in C. canoeiro, C. veadeiros, and C. anosteos, and in the other members of Corumbataia clade.  LBP 19311, 2, 29.4-27.4 mm SL;LBP 19302, 3, 28.5-33.9 mm SL.

ACKNOWLEDGMENTS
We thank Bruno F. Melo, Renato Devidé, Jefferson H. Martinez, and Martin I. Taylor for their help during field expeditions. We are also grateful to Lais Reia for helping with sequencing and figure preparation. Maxwell J. Bernt significantly contributed with comments and revision of the English on drafts of the manuscript. This research received financial support from Programa Institucional de Bolsas de Iniciação Científica e Tecnológica Unesp (MGRT), a PNPD Capes grant (GSCS), and from Conselho