Genetica DOI 10.1007/s10709-015-9856-z
Integrative taxonomy detects cryptic and overlooked fish species in a neotropical river basin Laı´s Carvalho Gomes1 • Tiago Casarim Pessali2 • Naiara Guimara˜es Sales1 Paulo Santos Pompeu3 • Daniel Cardoso Carvalho1
•
Received: 26 March 2015 / Accepted: 1 July 2015 Ó Springer International Publishing Switzerland 2015
Abstract The great freshwater fish diversity found in the neotropical region makes management and conservation actions challenging. Due to shortage of taxonomists and insufficient infrastructure to deal with such great biodiversity (i.e. taxonomic impediment), proposed remedies to accelerate species identification and descriptions include techniques that combine DNA-based identification and concise morphological description. The building of a DNA barcode reference database correlating meristic and genetic data was developed for 75 % of the Mucuri River basin’s freshwater fish. We obtained a total of 141 DNA barcode sequences from 37 species belonging to 30 genera, 19 families, and 5 orders. Genetic distances within species, genera, and families were 0.74, 9.5, and 18.86 %, respectively. All species could be clearly identified by the DNA
Electronic supplementary material The online version of this article (doi:10.1007/s10709-015-9856-z) contains supplementary material, which is available to authorized users. & Daniel Cardoso Carvalho [email protected]; [email protected] 1
Programa de Po´s-graduac¸a˜o em Zoologia de Vertebrados, Laborato´rio de Gene´tica da Conservac¸a˜o, Pontifı´cia Universidade Cato´lica de Minas Gerais, Rua Dom Jose´ Gaspar, 500,Corac¸a˜o Eucarı´stico, Belo Horizonte, MG 30535-901, Brazil
2
Museu de Cieˆncias Naturais, Laborato´rio de Ictiologia, Pontifı´cia Universidade Cato´lica de Minas Gerais, Rua Dom Jose´ Gaspar, 290,Corac¸a˜o Eucarı´stico, Belo Horizonte, MG 30535-610, Brazil
3
Laborato´rio de Ecologia de peixes, Departamento de Biologia, Universidade Federal de Lavras, Avenida Doutor Sylvio Menicucci, 1001, Kennedy, Lavras, MG 37200-000, Brazil
barcodes. Divergences between meristic morphological characteristics and DNA barcodes revealed two cryptic species among the Cyphocharax gilbert and Astyanax gr. bimaculatus specimens, and helped to identify two overlooked species within the Gymnotus and Astyanax taxa. Therefore, using a simplified model of neotropical biodiversity, we tested the efficiency of an integrative taxonomy approach for species discovery, identification of cryptic diversity, and accelerating biodiversity descriptions. Keywords Integrative taxonomy Neotropical Biodiversity Ichthyofauna
Introduction The precise identification and description of species is essential for biodiversity conservation. Considering the prediction of approximately 8.7 million species distributed worldwide, it is estimated that 91 % of aquatic species have not been formally described (Mora et al. 2011). Among the 13,000 freshwater species estimated for the planet, the described neotropical freshwater fishes comprise more than 5600 species, but estimates of the predicted number of species exceed 7000 (Albert and Reis 2011). However, a shortage of taxonomists, insufficient infrastructure for successful sampling, the need for taxonomic revision as well as difficulties accessing information (Agostinho et al. 2005) have delayed advances in biodiversity description, leading to what is known as taxonomic impediment (Taylor 1983). Proposed remedies include initiatives of an integrative taxonomy (Dayrat 2005) and ‘‘turbo-taxonomy’’ (Riedel et al. 2013), approaches that combine DNA-based identification, concise morphological description, and high-resolution images.
123
Genetica
Molecular identification systems through DNA barcodes, i.e. the use of 650 base pairs (bp) of the mitochondrial gene COI (cytochrome oxidase subunit I), was proposed by Hebert et al. (2003) as a standard molecular method for species identification. The efficiency of DNA barcode identification relies on a lower intraspecific COI genetic distance when compared with interspecific genetic distance estimations (Hebert et al. 2003, 2004a, b; Waugh 2007). A threshold of 2 % for species delimitation was proposed after 95 % of 1088 fish barcodes showed intraspecific genetic distance below 2 % (Ward 2009). However, more complex algorithms such as RESL, ABGD, CROP, and jMOTU were subsequently tested for their speed, and effectiveness in recovering species boundaries, with RESL being selected for the Barcode Index Number (BIN) system (Ratnasingham and Hebert 2007). These alternative procedures using DNA barcoding provide a substitute approach to the traditional tree-based methods for preliminary biodiversity screening and species identification. Developments in biodiversity discovery and species identification can benefit from an integrative taxonomy method, combining morphological characteristics and a standard DNA identification system such as the DNA barcoding (Ratnasingham and Hebert 2007). Initiatives using the methodology of DNA barcoding have been carried out to identify the fish fauna in a global context. Studies using this methodology were conducted in Australia (Ward et al. 2005), Canada (Hubert et al. 2008), Africa (Nwani et al. 2011), as well as in Brazilian river basins such as the Sa˜o Francisco (Carvalho et al. 2011), Paraı´ba do Sul (Pereira et al. 2011), and Parana´ (Pereira et al. 2013) basins, where the rate of species discrimination in these studies ranged from 95.0 to 100 %. Moreover, integrative approaches using DNA barcodes and meristic analysis may be useful, not only to flag potential overlooked species or operational taxonomic units (OTUs), but also to indicate potential new species (i.e. candidate species). Therefore, pipelines for delineation of putative species have been proposed as an efficient start for taxonomic work flows (Kekkonen and Hebert 2014). The Mucuri River basin is located in a set of independent basins that drain the eastern region of Brazil in the northeastern Atlantic Forest freshwater ecoregion, which possesses a high level of fish endemism and great biogeographic significance. Its ichthyofauna is composed of 61 fish, including 12 marine species, distributed in 45 genera and 26 families (Pompeu 2010), and seven nonindigenous species (Leporinus macrocephalus, Prochilodus costatus, Clarias gariepinus, Cichla sp., Oreochromis niloticus, Cyprinus carpio, and Poecilia reticulata). Impacts on fish populations and fishing activities in the Mucuri River have been reported due to intense anthropogenic changes, reflected in the high number of
123
introduced species, as well as deforestation, dumping of domestic and industrial sewage, and damming for hydropower generation (Pompeu 2010). Here we conducted a comprehensive inventory and developed a barcode library integrated with a detailed morphological identification of 75 % of the freshwater ichthyofauna from the Mucuri River basin. The low species richness of the Mucuri ichthyofauna provides a simplified model of neotropical biodiversity, enabling the ability to test the efficiency of an integrative taxonomy approach for species discovery, identification of cryptic diversity, and accelerating biodiversity descriptions.
Materials and methods Study area The Mucuri River basin (MRB) is part of a set of independent basins that drain the eastern region of Brazil, and it possesses a total drainage area of 15,413 km2 (MMA 2006). It is bordered in the northwest by the Jequitinhonha River basin and in the south by the Sa˜o Mateus and Itau´nas River basins (MMA 2006). The temperature of the MRB waters range from 23 to 27 °C; the climate is considered semi-humid and has a dry period lasting 4–5 months per year (IGAM 2008). Sampling Two campaigns for fish collection were held in 2013 and 2014, which included 14 sampling sites along the upper and middle MRB (Fig. 1). Specimens were captured using gill nets, cast nets, and sieves. Fin fragments were obtained from freshly captured fish and fixed in 100 % ethanol. Subsequently, samples were taken to the laboratory and kept at -20 °C. Voucher specimens were fixed in 10 % formalin and later preserved in 70 % ethanol and are held in the Ichthyology Collection of the PUC Minas Natural History Museum. All specimens were photographed, geo-referenced, and identified to the lowest taxonomic level using meristicbased identification keys and taxonomic papers (Vari 1992; Albert et al. 1999; Castro and Vari 2004; Zanata and Camelier 2009) (Supplemental Materials). DNA barcoding Whenever possible, a minimum of five specimens per species were analyzed. DNA extraction was achieved by salting out following an adapted protocol from Aljanabi and Martinez (1997). Partial sequences of the cytochrome oxidase I gene (COI) were amplified by polymerase chain
Genetica
Fig. 1 Map of the Mucuri River basin depicting sampling sites (red dots). The two dark blue dots represent two hydroelectric dams, Mucuri in the upper and Santa Clara in the lower portion of the river. (Color figure online)
reaction (PCR) using primers described by Ward et al. (2005). The PCR contained a final volume of 10 ll, including 1.0 ll 109 buffer containing MgCl2; 0.3 ll of dNTP (10 mM); 0.25 ll of each the two primers (10 lM); 0.2 ll of Taq DNA polymerase (5 U/ll); 7.0 ll of ultra pure water and 1.0 ll of DNA template. Amplifications were performed using a thermocycler (VeritiÒ 96—Well Thermal Cycler, Applied Biosystems), and they consisted of an initial denaturation step of 2 min at 95 °C followed by 35 cycles of denaturation for 30 s at 94 °C; 54 °C for 30 s for primer annealing, 1 min at 72 °C followed by 10 min at 72 °C. The amplified DNA products were visualized on 1 % agarose gel, and positive amplifications were selected for DNA sequencing. Sequencing reactions were performed bi-directionally according to the manufacturer’s instructions using the Big DyeÒ Terminator Cycle Sequencing Kit v3.1. The sequencing reaction had a final volume of 10.0 ll, containing 2.0 ll of 59 Buffer, 0.5 ll of Big Dye, 1.0 ll of primer (10 lM), 5.5 ll of ultrapure water, and 1.0 ll of the amplified DNA product. DNA sequencing was conducted in an automated DNA analyzer ABI 3500 (Life Technologies). Data analysis DNA sequences were edited using the software DNA BaserÒ v.3.5.4 and aligned using the MUSCLE alignment tool (Edgar 2004). The intra and interspecific genetic distances and nearest neighbor distance (NND) were obtained from the BOLD database platform (Ratnasingham and Hebert 2007). Genetic distances were estimated using the
Kimura-2-parameter (K2P) nucleotide evolution model (Kimura 1980). The Mega 5.1 program was used to construct neighbor joining (NJ) (Saitou and Nei 1987) dendrograms and to estimate the K2P genetic divergence between clades. All data including the electropherograms, fish photos, GPS coordinates of each sample site, vouchers numbers, and detailed taxonomic identifications were deposited in The Barcode of Life Database platform (BOLD) under the Barcoding of Neotropical Fish Species Project from the MRB. On-line bioinformatics tools implemented in the BOLD platform were used to estimate intra and inter genetic distances between species, genera, and families. Delimitation of candidate species A diverse range of statistical methods was used for species delimitation. First, traditional indexes indices such as nearest neighbor distance (NND) were used to estimate the minimum genetic distance between pairs of species, and the barcode gap was calculated on BOLD Workbench (http://www.boldsystems.org). Second, two clustering methods, namely the Barcode Index Number System (BIN; Ratnasingham and Hebert 2013) and Automatic Barcode Gap Discovery (ABGD; Puillandre et al. 2012), were applied using the barcode threshold of approximately 2 % as suggested by Ward (2009) who analyzed barcode sequences of 1088 fish species. ABGD and BIN apply clustering algorithms to distinguish partitions in the genetic distances among a group of individuals, using a two-phased procedure to sort sequences into hypothetical species. ABGD first uses a range of prior intraspecific divergences
123
Genetica
to divide the data into groups based on a statistically inferred barcode gap, and then recursively applies the same procedure to the groups obtained in the first step. Whereas, the approach of BIN firstly employs single linkage clustering coupled with a 2.2 % threshold to establish preliminary species boundaries followed by Markov clustering. This biphasic process has the goal of both improving and, if needed, redefining groups recovered in the first phase of both methods. ABGD analyses were performed using the web interface (http://wwwabi.snv.jussieu.fr/public/abgd/, web version ‘October 28 2014’) with a relative gap width value of X = 1.0 and two available distance metrics [JC69 (Jukes and Cantor 1969) and K2P (Kimura 1980)], while the other parameter values employed defaults. Assignments for intraspecific divergence (P) values between 0.001 and 0.100 were recorded. The Barcode Index Numbers (BIN) was automatically estimated in BOLD Workbench. We followed the approach presented in Padial et al. (2010) and the congruence among BIN, ABGD and genetic distance methods to designate candidate species. When referring to a particular candidate species, we used the genus or a combination of the binomial species name of the most similar or closely related nominal species followed (in square brackets) by ‘‘Ca’’ (for candidate) plus a concordant attached Barcode Index Number (BIN), where the number refers to a highly divergent barcode sequence of the species.
Results We obtained 141 sequences of the COI gene for 37 species, representing 75 % of the described freshwater fish fauna from the Mucurı´ River basin, a large neotropical freshwater basin heavily impacted by anthropogenic activities. DNA barcodes had an average of 658 bp, and no insertions, deletions, and stop codons were observed. A mean of 3.81 specimens were analyzed per species, belonging to 30 genera, 19 families, and 5 orders: Characiformes (38 %), Siluriformes (38 %), Perciformes (14 %), Cyprinodontiformes (5 %), and Gymnotiformes (5 %). Only one specimen per species was obtained for Gymnotus sylvius, Hoplosternum littorale, Imparfinis minutus, Pachyurus adspersus, and P. costatus, and therefore, they were not included in the estimation of intraspecific divergences. The NJ tree encompassing all analyzed fish species showed species-specific clades and an absence of shared haplotypes among species (Fig. 2). The estimated K2P genetic distances within species, genera, and families were 0.74, 9.5, and 18.86 %, respectively (Table 1). The mean intraspecific value observed was 0.74 %, ranging from 0 to 3.24 %. Of the 37 species analyzed, 28 had a mean intraspecific genetic divergence smaller than 1 %.
123
The nearest neighbor distance analysis (NND) showed that all species had a K2P divergence higher than 2 % to their closest neighbor. When considering the 2 % intraspecific genetic distance threshold for species delimitation, 35 (95 %) fish species were correctly identified. Only two species Cyphocharax gilbert and Astyanax gr. bimaculatus had intraspecific genetic divergences higher than 2 %, with mean values of 1.61 % (ranging from 0 to 3.24 %) and 0.92 % (ranging from 0 to 2.9 %), respectively. Reflecting the deep intraspecific divergent lineages, the NJ tree clustered C. gilbert and A. gr. bimaculatus specimens in two distinct clades (Fig. 2). The BIN analysis identified 39 clusters, including 22 taxonomically discordant, 16 concordant, and one singleton (Pachyurus adspersus). Cyphocharax gilbert and Astyanax gr. bimaculatus were separated in two distinct BINs (Fig. 2). ABGD analysis detected 61–33 candidate species when varying the prior maximal distance from P = 0.001 to P = 0.1000, respectively, using both the K2P and JC69 nucleotide evolution methods. We chose the partition that recovered 38 groups (intraspecific distance P = 0.0129 and P = 0.0215), since this partition was more consistent with our a priori morphological taxonomic analysis. ABGD species delimitation was in agreement with all the BIN clusters with one exception; Astyanax gr. bimaculatus was considered as a single cluster by ABGD (Fig. 2). Cyphocharax gilbert specimens (n = 6) showed a mean intraspecific value of 1.61 % and were grouped in two clades in the NJ tree, with an average K2P genetic divergence between clades of 2.9 %, ranging from 2.8 to 3.24 %. Within the clades, the divergences ranged between 0.3 and 0.08 % (Fig. 2). Exclusive meristic morphological characteristics frequently used for Curimatidae species identification could not be found for each candidate species (Table 2). Moreover, the BIN analysis recovered 2 distinct BINs, one grouping C. gilbert specimens from MRB and the neighboring Doce River basin (ACK1539) and the other with fish exclusive to the MRB (ACN3641). Astyanax gr. bimaculatus specimens (n = 8) had a mean intraspecific value of 0.92 % and were split in two major clades (Fig. 2) with an average K2P genetic divergence between clades of 2.7 %, ranging from 2.5 to 2.9 %. One clade included only one specimen identified as Astyanax gr. bimaculatus (BIN ABY8634) and despite showing a divergence of 2.7 % when compared to other individuals (n = 7) of the same species, no morphological differences were found. All specimens presented a horizontally oval black humeral spot, a lozenge-shaped caudal-peduncle spot, continuing to the tip of the middle caudal-fin rays, similar to the meristic morphological characteristics described for Astyanax identification and presented in Garutti (1998). Therefore, no significant morphological
Genetica
Fig. 2 NJ tree based on K2P distances, BIN, and ABGD of all species analyzed. Photos and detailed NJ tree of species with intraspecific K2P genetic distances greater than 2 % are shown: a Astyanax spp.; b Cyphocharax gilbert; and c Gymnotus sylvius/G. carapo Table 1 Genetic distances (K2P) between species, genera, and families
Minimum (%)
Mean (%)
Maximum (%)
Standard Error (%)
Within species
0
0.74
3.24
Within genus
4.29
9.5
18.44
0.07
Within family
3.04
18.86
26.02
0.01
divergence was registered for the two recovered BINs (ABZ1711 and ABY8634). One specimen of Astyanax sp. (BIN ACJ1244) had a genetic divergence of 18.5 % to other congeneric individuals and a similarity of 99.8 % with Astyanax vermilion previously deposited in the BOLD database. The morphological identification of Astyanax sp. was divergent from the A. vermilion description presented in Zanata and Camelier (2009). Astyanax vermilion possessed a dark distal portion of the distal pelvic fin, 32–34 lateral line scales, 20–23 branched anal-fin rays, five series of scales between dorsal-fin origin and lateral line and no more than five cusps on the dentary teeth. The undescribed species
0
Astyanax sp. presented homogeneous pelvic fin color, 35 lateral line scales, 21 branched anal-fin rays, six series of scales between dorsal-fin origin and lateral line and seven cusps on dentary teeth. Incongruences between morphological and barcode identifications were observed within the Prochilodus and Gymnotus species. One specimen identified as P. costatus according to the meristic characteristics presented in Castro and Vari (2004) had a genetic similarity of 100 % with the species Prochilodus argenteus (BOLD number BSB39210) previously deposited in BOLD. Gymnotus carapo had an average intraspecific genetic distance of 2.42 % (0–5.05 %), but it was possible to correctly identify them
123
Genetica Table 2 Morphological identification of Cyphocharax gilbert specimens based on diagnostic characteristics Specimen
BIN
Number of lateral line scales
Number of scales above lateral line to origin of dorsal fin
Number of scales below lateral line to origin of anal fin
LGC 4091
ACK1539
37
6
5
Mu014
ACK1539
39
7
5
LGC 4112
ACN3641
38
7
5
LGC 4105
ACN3641
38
6
5
LGC 4114
ACN3641
37
7
5
LGC 4113
ACN3641
39
6
5
as G. sylvius and Gymnotus aff. carapo based on a morphogenetic approach. Gymnotus sylvius (BIN AAB6215) exhibited wider white interbands than dark band regions, and the dark bands were divided into blotches and spots above the lateral line (Albert et al. 1999), which were not observed for the other specimens identified as Gymnotus aff. carapo (BIN AAB6216) (Fig. 2).
Discussion Herein, we implemented an integrative approach that aimed to test the pipeline accuracy in the identification of the freshwater fish fauna from the MRB. Barcode analysis recovered 38–39 putative species from the 37 morphologically identified species analyzed, which represents around 75 % of freshwater ichthyofauna of the MRB. The mean K2P genetic intraspecific divergence was 0.74 %, 13 times smaller when comparing to the mean interspecific distance (9.5 %). Therefore, a barcode gap was successfully recovered, since no overlap of intra and interspecific divergence was observed. Neotropical freshwater fish barcode surveys have reported similar mean intra and interspecific genetic distance values, ranging from 0.50 to 10.61 % for the Sa˜o Francisco River (Carvalho et al. 2011), 0.13 to 10.36 % for the Paraı´ba do Sul River (Pereira et al. 2011), and 1.3 to 7.1 % for the Upper Parana´ River basin (Pereira et al. 2013). By applying the barcode threshold of 2 % defined by Ward (2009), we were able to differentiate all analyzed species from the MRB. The NND reinforced this result by showing that all species had a higher than 2 % K2P divergence to their closest neighbor. Overlooked cryptic species were flagged within the C. gilbert specimens due to the intraspecific divergence above 2 %, and corroborated by distinct BIN and ABGD clusters (Fig. 2b). However, only one species of Cyphocharax is reported for the MRB (Vari 1992). The absence of morphological differences between distinct clades, supported by both BIN and ABGD analysis, clearly indicates a case for a new candidate cryptic species. Moreover, since one BIN was shared with a Cyphocharax species from the
123
neighboring Doce River while the other one was exclusive to the MRB, we suggest the occurrence of a new candidate species, namely Cyphocharax aff. gilbert [ACN3641], endemic to the MRB. With the growth of the barcode database, including all other neighboring river basins, we may be able to expose endemic cryptic species using BIN analysis as well as decreasing the risk of synonymies. Astyanax gr. bimaculatus had a 2.7 % mean intraspecific divergence between NJ tree clades and recovered BINs (Fig. 2a) but with no ABGD or morphological support. The genus Astyanax is a well know complex of species and one of the most specious neotropical fish groups with over 140 described species (Froese and Paulay 2010; Eschmeyer 2015), occurring in practically all river basins of the neotropics (Marinho and Lima 2009). Eigenmann (1921) redescribed Astyanax bimaculatus as being comprised of several subspecies, and Garutti (1995), reviewing nominal species, recognized 30 valid species, suggested the need for more systematic studies to objectively determine the actual levels of diversity within the bimaculatus complex. Interestingly, a deep divergence of 18.5 % was found when comparing one unidentified specimen of Astyanax sp. (ACJ1244) to other congeneric specimens sampled from the Mucuri basin. This highly divergent haplotype had high similarity (100 %) to an A. vermilion barcode sequence previously deposited in the BOLD database. However, A. vermilion is not described for the Mucuri basin, but from a coastal northeastern Brazilian river 300 km distant, and morphological analysis did not corroborate the barcode identification. Morphologic variation wider than described for A. vermilion or a mistake in the morphological identification of the species deposited as A. vermilion in BOLD may explain this discordance. The native species Prochilodus vimboides could be clearly differentiated from the introduced species P. costatus by DNA barcoding, since interspecific divergence of 9 % was recovered. Nonetheless, the non-indigenous species identified morphologically as P. costatus had a barcode similarity of 100 % with another non-indigenous species P. argenteus (BOLD number: BSB392-10). The divergence between the morphological and molecular identification of P. costatus may indicate the presence of
Genetica
Prochilodus hybrids in the Mucuri basin between the two species of the Sa˜o Francisco River. Hybrids of Prochilodus are commonly produced in aquaculture farms in Brazil and may escape into the wild. In the case of the electric knifefishes, the integrative approach based on molecular taxonomy and morphological characteristics allowed the precise identification of the overlooked species G. sylvius, not yet reported for the MRB. All Gymnotus specimens were first identified as G. aff. carapo and due to their high intraspecific genetic divergence and the presence of two clades in the NJ tree (Fig. 2c), the specimens were subjected to a more detailed morphological analysis, allowing their correct taxonomic identification. However, G. sylvius is described only for the Ribeira de Iguape, Paraı´ba do Sul, and Upper Parana´ River basins, but not for the MRB (Albert et al. 1999; Pompeu 2010). Therefore, its distribution range should be expanded to the MRB based on the results of our integrative approach. This integrative approach using a comprehensive inventory through the development of a barcode library combined with a detailed morphological identification was a robust methodology for species delimitation. For instance, our results indicated a new candidate species (Cyphocharax aff. gilbert [Ca ACN3641]) and detected overlooked species within the Astyanax (A. vermilion) and G. sylvius specimens. Therefore, our integrative pipeline using a simplified biodiversity model (i.e. Mucuri ichthyofauna) showed a possible remedy to the taxonomic impediment, and may be used to adequately describe and conserve the biodiversity of other complex neotropical ichthyofaunas. Acknowledgments We thank Angelo Monteiro, Danilo Neto, Iago Penido and Ivo for helping with the fieldwork and laboratory analysis. DCC is grateful to Conselho Nacional de Desenvolvimento Cientı´fico e Tecnolo´gico (CNPq) for his productivity fellowship (308537/20149). This work was funded by the Brazilian Barcode of Life Network BrBol/CNPq (564953/2010-5), Coordenac¸a˜o de Aperfeic¸oamento de ´ EQUIPAMENTOS Grant Pessoal do Ensino Superior (CAPES—PRO Number 783380/2013), Conselho Nacional de Desenvolvimento Cientı´fico e Tecnolo´gico—CNPq (Process Numbers INCT 573899/2008-8 and 482852/2011-9) and Fundac¸a˜o de Amparo a` Pesquisa do Estado de Minas Gerais—FAPEMIG (Process Number INCT APQ-0084/08).
References Agostinho AA, Thomaz SM, Gomes LC (2005) Conservac¸a˜o da biodiversidade em a´guas continentais do Brasil. Megadiversidade 1:70–78 Albert JS, Reis RE (2011) Historical biogeography of Neotropical freshwater fishes. University of California Press, Oakland Albert JS, Fernandes-Matioli FMC, Almeida-Toledo LF (1999) New species of Gymnotus (Gymnotiformes, Teleostei) from
southeastern Brazil: toward the deconstruction of Gymnotus carapo. Copeia 1999:410–421 Aljanabi SM, Martinez I (1997) Universal and rapid salt-extraction of high quality genomic DNA for PCR-based techniques. Nucleic Acids Res 25:4692–4693 Carvalho DC, Oliveira DAA, Pompeu PS, Leal CG, Oliveira C, Hanner R (2011) Deep barcode divergence in Brazilian freshwater fishes: the case of the Sa˜o Francisco River Basin. Mitochondr DNA 22:80–86 Castro RM, Vari RP (2004) Detritivores of the South American fish family Prochilodontidae (Teleostei: Ostariophysi: Characiformes): a phylogenetic and revisionary study. Smithsonian Contributions to Zoology, Washington Dayrat B (2005) Towards integrative taxonomy. Biol J Linn Soc 85:407–415 Edgar RC (2004) MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Res 32:1792–1797 Eigenmann CH (1921) The American characidae. Mem Mus Comp Zool 43:208–310 Eschmeyer WN (2015) Catalog of fishes: world wide web electronic publication. http://researcharchive.calacademy.org/research/icht hyology/catalog/fishcatmain.asp. Accessed 30 March 2015 Froese R, Paulay D (2010) FishBase: World Wide Web electronic publication. http://fishbase.org. Accessed 10 Dec 2014 Garutti V (1995) Revisa˜o taxonoˆmica dos Astyanax (Pisces, Characidae), com mancha umeral ovalada e mancha no pedu´nculo caudal, estendendo-se a` extremidade dos raios caudais medianos, das bacias do Parana´, Sa˜o Francisco e Amazoˆnica. Thesis, Universidade Estadual Paulista Garutti V (1998) Descric¸a˜o de uma espe´cie nova de Astyanax (Teleostei, Characidae) da bacia do Tocantins, Brasil. Iheringia, Se´r Zool 85:115–122 Hebert PDN, Cywinska A, Ball SL, deWaard JR (2003) Biological identifications through DNA barcodes. Proc R Soc Lond B Biol 270:313–321 Hebert PDN, Penton EH, Burns JM, Janzen DH, Hallwachs W (2004a) Ten species in one: DNA barcoding reveals cryptic species in the neotropical skipper butterfly Astrapes fulgerator. Proc Natl Acad Sci USA 101:14812–14817 Hebert PDN, Stoeckle MY, Zemlak TS, Francis CM (2004b) Identification of birds through DNA Barcodes. PLoS Biol 2:e312. doi:10.1371/journal.pbio.0020312 Hubert N, Hanner R, Holm E, Mandrak NE, Taylor E, Burridge M, Watkinson D, Dumont P, Curry A, Bentzen P, Zhang JB, April J, Bernatchez L (2008) Identifying Canadian freshwater fishes through DNA barcodes. PLoS One 3:e2490. doi:10.1371/journal. pone.0002490 ´ guas. Bacias hidrogra´ficas IGAM—Instituto Mineiro de Gesta˜o das A do Leste (2008) Bacia Hidrogra´fica do Rio Mucuri (MU1). www. igam.mg.gov.br/component/content/158?task=view. Accessed 2 August 2014 Jukes TH, Cantor CR (1969) Evolution of protein molecules. In: Munro HN (ed) Mammalian protein metabolism. Academic Press, New York Kekkonen M, Hebert PDN (2014) DNA barcode-based delineation of putative species: efficient start for taxonomic workflows. Mol Ecol Resour 14:706–715 Kimura M (1980) A simple method for estimating evolutionary rate of base substitutions through comparative studies of nucleotide sequences. J Mol Evol 16:111–120 Marinho MM, Lima FC (2009) Astyanax ajuricaba: a new species from the Amazon basin in Brazil (Characiformes: Characidae). Neotrop Ichthyol 7:169–174 MMA—Ministe´rio do Meio Ambiente (2006) Atlaˆntico Leste In Caderno da regia˜o hidrogra´fica Atlaˆntico Leste. Ministe´rio do Meio Ambiente, Secretaria de Recursos Hı´dricos, Brası´lia
123
Genetica Mora C, Tittensor DP, Adl S, Simpson AGB, Worm B (2011) How many species are there on earth and in the ocean? PLoS Biol 9:e1001127. doi:10.1371/journal.pbio.1001127 Nwani CD, Becker S, Braid HE, Ude EF, Okogwu OI, Hanner R (2011) DNA barcoding discriminates freshwater fishes from southeastern Nigeria and provides river system-level phylogeographic resolution within some species. Mitochondr DNA 22:43–51 Padial JM, Miralles A, De la Riva I, Vences M (2010) The integrative future of taxonomy . Front Zool 7(1):16. doi:10.1186/1742-9994-716 Pereira LHG, Maia GMG, Hanner R, Foresti F, Oliveira C (2011) DNA barcodes discriminate freshwater fishes from the Paraı´ba do Sul River Basin, Sa˜o Paulo, Brazil. Mitochondr DNA 22:71–79 Pereira LHG, Hanner R, Foresti F, Oliveira C (2013) Can DNA Barcoding accurately discriminate megadiverse Neotropical freshwater fish fauna? BMC Genet 14:14–20 Pompeu PS (2010) Os peixes do rio Mucuri. MG BIOTA 2:36–48 Puillandre N, Lambert A, Brouillet S, Achaz G (2012) ABGD, automatic barcode gap discovery for primary species delimitation. Mol Ecol 21:1864–1877 Ratnasingham S, Hebert PDN (2007) BOLD: the barcode of life data system (www.barcodinglife.org). Mol Ecol Notes 3:355–364 Ratnasingham S, Hebert PDN (2013) A DNA-based registry for all animal species: the barcode index number (BIN) system. PLoS One. doi:10.1371/journal.pone.0066213
123
Riedel A, Sagata K, Suhardjono YR, Ta¨nzler R, Balke M (2013) Integrative taxonomy on the fast track-towards more sustainability in biodiversity research. Front Zool. doi:10.1186/17429994-10-15 Saitou N, Nei M (1987) The neighbor-joining method: a new method for reconstructing phylogenetic trees. Mol Biol Evol 4:406–425 Taylor RW (1983) Descriptive taxonomy: past, present, and future. In: Highley E, Taylor RW (eds) Australian Systematic entomology: a bicentenary perspective. CSIRO, Melbourne Vari R (1992) Systematics of the neotropical characiform genus Cyphocharax Fowler (Pisces: Ostariophysi). Smithson Contrib Zool 529:1–137 Ward RD (2009) DNA barcode divergence among species and genera of birds and fishes. Mol Ecol Resour 9:1077–1085 Ward RD, Zemlak TS, Innes BH, Last PR, Hebert PDN (2005) DNA barcoding Australia’s fish species. Philos Trans R Soc B 360:1847–1857 Waugh J (2007) DNA barcoding in animal species: progress, potential and pitfalls. BioEssays 29:188–197 Zanata AM, Camelier P (2009) Astyanax vermilion and Astyanax burgerai: new characid fishes (Ostariophysi: Characiformes) from Northeastern Bahia, Brazil. Neotrop Ichthyol 7:175–184
Integrative taxonomy detects cryptic and overlooked fish species in a neotropical river basin Laı´s Carvalho Gomes1 • Tiago Casarim Pessali2 • Naiara Guimara˜es Sales1 Paulo Santos Pompeu3 • Daniel Cardoso Carvalho1
•
Received: 26 March 2015 / Accepted: 1 July 2015 Ó Springer International Publishing Switzerland 2015
Abstract The great freshwater fish diversity found in the neotropical region makes management and conservation actions challenging. Due to shortage of taxonomists and insufficient infrastructure to deal with such great biodiversity (i.e. taxonomic impediment), proposed remedies to accelerate species identification and descriptions include techniques that combine DNA-based identification and concise morphological description. The building of a DNA barcode reference database correlating meristic and genetic data was developed for 75 % of the Mucuri River basin’s freshwater fish. We obtained a total of 141 DNA barcode sequences from 37 species belonging to 30 genera, 19 families, and 5 orders. Genetic distances within species, genera, and families were 0.74, 9.5, and 18.86 %, respectively. All species could be clearly identified by the DNA
Electronic supplementary material The online version of this article (doi:10.1007/s10709-015-9856-z) contains supplementary material, which is available to authorized users. & Daniel Cardoso Carvalho [email protected]; [email protected] 1
Programa de Po´s-graduac¸a˜o em Zoologia de Vertebrados, Laborato´rio de Gene´tica da Conservac¸a˜o, Pontifı´cia Universidade Cato´lica de Minas Gerais, Rua Dom Jose´ Gaspar, 500,Corac¸a˜o Eucarı´stico, Belo Horizonte, MG 30535-901, Brazil
2
Museu de Cieˆncias Naturais, Laborato´rio de Ictiologia, Pontifı´cia Universidade Cato´lica de Minas Gerais, Rua Dom Jose´ Gaspar, 290,Corac¸a˜o Eucarı´stico, Belo Horizonte, MG 30535-610, Brazil
3
Laborato´rio de Ecologia de peixes, Departamento de Biologia, Universidade Federal de Lavras, Avenida Doutor Sylvio Menicucci, 1001, Kennedy, Lavras, MG 37200-000, Brazil
barcodes. Divergences between meristic morphological characteristics and DNA barcodes revealed two cryptic species among the Cyphocharax gilbert and Astyanax gr. bimaculatus specimens, and helped to identify two overlooked species within the Gymnotus and Astyanax taxa. Therefore, using a simplified model of neotropical biodiversity, we tested the efficiency of an integrative taxonomy approach for species discovery, identification of cryptic diversity, and accelerating biodiversity descriptions. Keywords Integrative taxonomy Neotropical Biodiversity Ichthyofauna
Introduction The precise identification and description of species is essential for biodiversity conservation. Considering the prediction of approximately 8.7 million species distributed worldwide, it is estimated that 91 % of aquatic species have not been formally described (Mora et al. 2011). Among the 13,000 freshwater species estimated for the planet, the described neotropical freshwater fishes comprise more than 5600 species, but estimates of the predicted number of species exceed 7000 (Albert and Reis 2011). However, a shortage of taxonomists, insufficient infrastructure for successful sampling, the need for taxonomic revision as well as difficulties accessing information (Agostinho et al. 2005) have delayed advances in biodiversity description, leading to what is known as taxonomic impediment (Taylor 1983). Proposed remedies include initiatives of an integrative taxonomy (Dayrat 2005) and ‘‘turbo-taxonomy’’ (Riedel et al. 2013), approaches that combine DNA-based identification, concise morphological description, and high-resolution images.
123
Genetica
Molecular identification systems through DNA barcodes, i.e. the use of 650 base pairs (bp) of the mitochondrial gene COI (cytochrome oxidase subunit I), was proposed by Hebert et al. (2003) as a standard molecular method for species identification. The efficiency of DNA barcode identification relies on a lower intraspecific COI genetic distance when compared with interspecific genetic distance estimations (Hebert et al. 2003, 2004a, b; Waugh 2007). A threshold of 2 % for species delimitation was proposed after 95 % of 1088 fish barcodes showed intraspecific genetic distance below 2 % (Ward 2009). However, more complex algorithms such as RESL, ABGD, CROP, and jMOTU were subsequently tested for their speed, and effectiveness in recovering species boundaries, with RESL being selected for the Barcode Index Number (BIN) system (Ratnasingham and Hebert 2007). These alternative procedures using DNA barcoding provide a substitute approach to the traditional tree-based methods for preliminary biodiversity screening and species identification. Developments in biodiversity discovery and species identification can benefit from an integrative taxonomy method, combining morphological characteristics and a standard DNA identification system such as the DNA barcoding (Ratnasingham and Hebert 2007). Initiatives using the methodology of DNA barcoding have been carried out to identify the fish fauna in a global context. Studies using this methodology were conducted in Australia (Ward et al. 2005), Canada (Hubert et al. 2008), Africa (Nwani et al. 2011), as well as in Brazilian river basins such as the Sa˜o Francisco (Carvalho et al. 2011), Paraı´ba do Sul (Pereira et al. 2011), and Parana´ (Pereira et al. 2013) basins, where the rate of species discrimination in these studies ranged from 95.0 to 100 %. Moreover, integrative approaches using DNA barcodes and meristic analysis may be useful, not only to flag potential overlooked species or operational taxonomic units (OTUs), but also to indicate potential new species (i.e. candidate species). Therefore, pipelines for delineation of putative species have been proposed as an efficient start for taxonomic work flows (Kekkonen and Hebert 2014). The Mucuri River basin is located in a set of independent basins that drain the eastern region of Brazil in the northeastern Atlantic Forest freshwater ecoregion, which possesses a high level of fish endemism and great biogeographic significance. Its ichthyofauna is composed of 61 fish, including 12 marine species, distributed in 45 genera and 26 families (Pompeu 2010), and seven nonindigenous species (Leporinus macrocephalus, Prochilodus costatus, Clarias gariepinus, Cichla sp., Oreochromis niloticus, Cyprinus carpio, and Poecilia reticulata). Impacts on fish populations and fishing activities in the Mucuri River have been reported due to intense anthropogenic changes, reflected in the high number of
123
introduced species, as well as deforestation, dumping of domestic and industrial sewage, and damming for hydropower generation (Pompeu 2010). Here we conducted a comprehensive inventory and developed a barcode library integrated with a detailed morphological identification of 75 % of the freshwater ichthyofauna from the Mucuri River basin. The low species richness of the Mucuri ichthyofauna provides a simplified model of neotropical biodiversity, enabling the ability to test the efficiency of an integrative taxonomy approach for species discovery, identification of cryptic diversity, and accelerating biodiversity descriptions.
Materials and methods Study area The Mucuri River basin (MRB) is part of a set of independent basins that drain the eastern region of Brazil, and it possesses a total drainage area of 15,413 km2 (MMA 2006). It is bordered in the northwest by the Jequitinhonha River basin and in the south by the Sa˜o Mateus and Itau´nas River basins (MMA 2006). The temperature of the MRB waters range from 23 to 27 °C; the climate is considered semi-humid and has a dry period lasting 4–5 months per year (IGAM 2008). Sampling Two campaigns for fish collection were held in 2013 and 2014, which included 14 sampling sites along the upper and middle MRB (Fig. 1). Specimens were captured using gill nets, cast nets, and sieves. Fin fragments were obtained from freshly captured fish and fixed in 100 % ethanol. Subsequently, samples were taken to the laboratory and kept at -20 °C. Voucher specimens were fixed in 10 % formalin and later preserved in 70 % ethanol and are held in the Ichthyology Collection of the PUC Minas Natural History Museum. All specimens were photographed, geo-referenced, and identified to the lowest taxonomic level using meristicbased identification keys and taxonomic papers (Vari 1992; Albert et al. 1999; Castro and Vari 2004; Zanata and Camelier 2009) (Supplemental Materials). DNA barcoding Whenever possible, a minimum of five specimens per species were analyzed. DNA extraction was achieved by salting out following an adapted protocol from Aljanabi and Martinez (1997). Partial sequences of the cytochrome oxidase I gene (COI) were amplified by polymerase chain
Genetica
Fig. 1 Map of the Mucuri River basin depicting sampling sites (red dots). The two dark blue dots represent two hydroelectric dams, Mucuri in the upper and Santa Clara in the lower portion of the river. (Color figure online)
reaction (PCR) using primers described by Ward et al. (2005). The PCR contained a final volume of 10 ll, including 1.0 ll 109 buffer containing MgCl2; 0.3 ll of dNTP (10 mM); 0.25 ll of each the two primers (10 lM); 0.2 ll of Taq DNA polymerase (5 U/ll); 7.0 ll of ultra pure water and 1.0 ll of DNA template. Amplifications were performed using a thermocycler (VeritiÒ 96—Well Thermal Cycler, Applied Biosystems), and they consisted of an initial denaturation step of 2 min at 95 °C followed by 35 cycles of denaturation for 30 s at 94 °C; 54 °C for 30 s for primer annealing, 1 min at 72 °C followed by 10 min at 72 °C. The amplified DNA products were visualized on 1 % agarose gel, and positive amplifications were selected for DNA sequencing. Sequencing reactions were performed bi-directionally according to the manufacturer’s instructions using the Big DyeÒ Terminator Cycle Sequencing Kit v3.1. The sequencing reaction had a final volume of 10.0 ll, containing 2.0 ll of 59 Buffer, 0.5 ll of Big Dye, 1.0 ll of primer (10 lM), 5.5 ll of ultrapure water, and 1.0 ll of the amplified DNA product. DNA sequencing was conducted in an automated DNA analyzer ABI 3500 (Life Technologies). Data analysis DNA sequences were edited using the software DNA BaserÒ v.3.5.4 and aligned using the MUSCLE alignment tool (Edgar 2004). The intra and interspecific genetic distances and nearest neighbor distance (NND) were obtained from the BOLD database platform (Ratnasingham and Hebert 2007). Genetic distances were estimated using the
Kimura-2-parameter (K2P) nucleotide evolution model (Kimura 1980). The Mega 5.1 program was used to construct neighbor joining (NJ) (Saitou and Nei 1987) dendrograms and to estimate the K2P genetic divergence between clades. All data including the electropherograms, fish photos, GPS coordinates of each sample site, vouchers numbers, and detailed taxonomic identifications were deposited in The Barcode of Life Database platform (BOLD) under the Barcoding of Neotropical Fish Species Project from the MRB. On-line bioinformatics tools implemented in the BOLD platform were used to estimate intra and inter genetic distances between species, genera, and families. Delimitation of candidate species A diverse range of statistical methods was used for species delimitation. First, traditional indexes indices such as nearest neighbor distance (NND) were used to estimate the minimum genetic distance between pairs of species, and the barcode gap was calculated on BOLD Workbench (http://www.boldsystems.org). Second, two clustering methods, namely the Barcode Index Number System (BIN; Ratnasingham and Hebert 2013) and Automatic Barcode Gap Discovery (ABGD; Puillandre et al. 2012), were applied using the barcode threshold of approximately 2 % as suggested by Ward (2009) who analyzed barcode sequences of 1088 fish species. ABGD and BIN apply clustering algorithms to distinguish partitions in the genetic distances among a group of individuals, using a two-phased procedure to sort sequences into hypothetical species. ABGD first uses a range of prior intraspecific divergences
123
Genetica
to divide the data into groups based on a statistically inferred barcode gap, and then recursively applies the same procedure to the groups obtained in the first step. Whereas, the approach of BIN firstly employs single linkage clustering coupled with a 2.2 % threshold to establish preliminary species boundaries followed by Markov clustering. This biphasic process has the goal of both improving and, if needed, redefining groups recovered in the first phase of both methods. ABGD analyses were performed using the web interface (http://wwwabi.snv.jussieu.fr/public/abgd/, web version ‘October 28 2014’) with a relative gap width value of X = 1.0 and two available distance metrics [JC69 (Jukes and Cantor 1969) and K2P (Kimura 1980)], while the other parameter values employed defaults. Assignments for intraspecific divergence (P) values between 0.001 and 0.100 were recorded. The Barcode Index Numbers (BIN) was automatically estimated in BOLD Workbench. We followed the approach presented in Padial et al. (2010) and the congruence among BIN, ABGD and genetic distance methods to designate candidate species. When referring to a particular candidate species, we used the genus or a combination of the binomial species name of the most similar or closely related nominal species followed (in square brackets) by ‘‘Ca’’ (for candidate) plus a concordant attached Barcode Index Number (BIN), where the number refers to a highly divergent barcode sequence of the species.
Results We obtained 141 sequences of the COI gene for 37 species, representing 75 % of the described freshwater fish fauna from the Mucurı´ River basin, a large neotropical freshwater basin heavily impacted by anthropogenic activities. DNA barcodes had an average of 658 bp, and no insertions, deletions, and stop codons were observed. A mean of 3.81 specimens were analyzed per species, belonging to 30 genera, 19 families, and 5 orders: Characiformes (38 %), Siluriformes (38 %), Perciformes (14 %), Cyprinodontiformes (5 %), and Gymnotiformes (5 %). Only one specimen per species was obtained for Gymnotus sylvius, Hoplosternum littorale, Imparfinis minutus, Pachyurus adspersus, and P. costatus, and therefore, they were not included in the estimation of intraspecific divergences. The NJ tree encompassing all analyzed fish species showed species-specific clades and an absence of shared haplotypes among species (Fig. 2). The estimated K2P genetic distances within species, genera, and families were 0.74, 9.5, and 18.86 %, respectively (Table 1). The mean intraspecific value observed was 0.74 %, ranging from 0 to 3.24 %. Of the 37 species analyzed, 28 had a mean intraspecific genetic divergence smaller than 1 %.
123
The nearest neighbor distance analysis (NND) showed that all species had a K2P divergence higher than 2 % to their closest neighbor. When considering the 2 % intraspecific genetic distance threshold for species delimitation, 35 (95 %) fish species were correctly identified. Only two species Cyphocharax gilbert and Astyanax gr. bimaculatus had intraspecific genetic divergences higher than 2 %, with mean values of 1.61 % (ranging from 0 to 3.24 %) and 0.92 % (ranging from 0 to 2.9 %), respectively. Reflecting the deep intraspecific divergent lineages, the NJ tree clustered C. gilbert and A. gr. bimaculatus specimens in two distinct clades (Fig. 2). The BIN analysis identified 39 clusters, including 22 taxonomically discordant, 16 concordant, and one singleton (Pachyurus adspersus). Cyphocharax gilbert and Astyanax gr. bimaculatus were separated in two distinct BINs (Fig. 2). ABGD analysis detected 61–33 candidate species when varying the prior maximal distance from P = 0.001 to P = 0.1000, respectively, using both the K2P and JC69 nucleotide evolution methods. We chose the partition that recovered 38 groups (intraspecific distance P = 0.0129 and P = 0.0215), since this partition was more consistent with our a priori morphological taxonomic analysis. ABGD species delimitation was in agreement with all the BIN clusters with one exception; Astyanax gr. bimaculatus was considered as a single cluster by ABGD (Fig. 2). Cyphocharax gilbert specimens (n = 6) showed a mean intraspecific value of 1.61 % and were grouped in two clades in the NJ tree, with an average K2P genetic divergence between clades of 2.9 %, ranging from 2.8 to 3.24 %. Within the clades, the divergences ranged between 0.3 and 0.08 % (Fig. 2). Exclusive meristic morphological characteristics frequently used for Curimatidae species identification could not be found for each candidate species (Table 2). Moreover, the BIN analysis recovered 2 distinct BINs, one grouping C. gilbert specimens from MRB and the neighboring Doce River basin (ACK1539) and the other with fish exclusive to the MRB (ACN3641). Astyanax gr. bimaculatus specimens (n = 8) had a mean intraspecific value of 0.92 % and were split in two major clades (Fig. 2) with an average K2P genetic divergence between clades of 2.7 %, ranging from 2.5 to 2.9 %. One clade included only one specimen identified as Astyanax gr. bimaculatus (BIN ABY8634) and despite showing a divergence of 2.7 % when compared to other individuals (n = 7) of the same species, no morphological differences were found. All specimens presented a horizontally oval black humeral spot, a lozenge-shaped caudal-peduncle spot, continuing to the tip of the middle caudal-fin rays, similar to the meristic morphological characteristics described for Astyanax identification and presented in Garutti (1998). Therefore, no significant morphological
Genetica
Fig. 2 NJ tree based on K2P distances, BIN, and ABGD of all species analyzed. Photos and detailed NJ tree of species with intraspecific K2P genetic distances greater than 2 % are shown: a Astyanax spp.; b Cyphocharax gilbert; and c Gymnotus sylvius/G. carapo Table 1 Genetic distances (K2P) between species, genera, and families
Minimum (%)
Mean (%)
Maximum (%)
Standard Error (%)
Within species
0
0.74
3.24
Within genus
4.29
9.5
18.44
0.07
Within family
3.04
18.86
26.02
0.01
divergence was registered for the two recovered BINs (ABZ1711 and ABY8634). One specimen of Astyanax sp. (BIN ACJ1244) had a genetic divergence of 18.5 % to other congeneric individuals and a similarity of 99.8 % with Astyanax vermilion previously deposited in the BOLD database. The morphological identification of Astyanax sp. was divergent from the A. vermilion description presented in Zanata and Camelier (2009). Astyanax vermilion possessed a dark distal portion of the distal pelvic fin, 32–34 lateral line scales, 20–23 branched anal-fin rays, five series of scales between dorsal-fin origin and lateral line and no more than five cusps on the dentary teeth. The undescribed species
0
Astyanax sp. presented homogeneous pelvic fin color, 35 lateral line scales, 21 branched anal-fin rays, six series of scales between dorsal-fin origin and lateral line and seven cusps on dentary teeth. Incongruences between morphological and barcode identifications were observed within the Prochilodus and Gymnotus species. One specimen identified as P. costatus according to the meristic characteristics presented in Castro and Vari (2004) had a genetic similarity of 100 % with the species Prochilodus argenteus (BOLD number BSB39210) previously deposited in BOLD. Gymnotus carapo had an average intraspecific genetic distance of 2.42 % (0–5.05 %), but it was possible to correctly identify them
123
Genetica Table 2 Morphological identification of Cyphocharax gilbert specimens based on diagnostic characteristics Specimen
BIN
Number of lateral line scales
Number of scales above lateral line to origin of dorsal fin
Number of scales below lateral line to origin of anal fin
LGC 4091
ACK1539
37
6
5
Mu014
ACK1539
39
7
5
LGC 4112
ACN3641
38
7
5
LGC 4105
ACN3641
38
6
5
LGC 4114
ACN3641
37
7
5
LGC 4113
ACN3641
39
6
5
as G. sylvius and Gymnotus aff. carapo based on a morphogenetic approach. Gymnotus sylvius (BIN AAB6215) exhibited wider white interbands than dark band regions, and the dark bands were divided into blotches and spots above the lateral line (Albert et al. 1999), which were not observed for the other specimens identified as Gymnotus aff. carapo (BIN AAB6216) (Fig. 2).
Discussion Herein, we implemented an integrative approach that aimed to test the pipeline accuracy in the identification of the freshwater fish fauna from the MRB. Barcode analysis recovered 38–39 putative species from the 37 morphologically identified species analyzed, which represents around 75 % of freshwater ichthyofauna of the MRB. The mean K2P genetic intraspecific divergence was 0.74 %, 13 times smaller when comparing to the mean interspecific distance (9.5 %). Therefore, a barcode gap was successfully recovered, since no overlap of intra and interspecific divergence was observed. Neotropical freshwater fish barcode surveys have reported similar mean intra and interspecific genetic distance values, ranging from 0.50 to 10.61 % for the Sa˜o Francisco River (Carvalho et al. 2011), 0.13 to 10.36 % for the Paraı´ba do Sul River (Pereira et al. 2011), and 1.3 to 7.1 % for the Upper Parana´ River basin (Pereira et al. 2013). By applying the barcode threshold of 2 % defined by Ward (2009), we were able to differentiate all analyzed species from the MRB. The NND reinforced this result by showing that all species had a higher than 2 % K2P divergence to their closest neighbor. Overlooked cryptic species were flagged within the C. gilbert specimens due to the intraspecific divergence above 2 %, and corroborated by distinct BIN and ABGD clusters (Fig. 2b). However, only one species of Cyphocharax is reported for the MRB (Vari 1992). The absence of morphological differences between distinct clades, supported by both BIN and ABGD analysis, clearly indicates a case for a new candidate cryptic species. Moreover, since one BIN was shared with a Cyphocharax species from the
123
neighboring Doce River while the other one was exclusive to the MRB, we suggest the occurrence of a new candidate species, namely Cyphocharax aff. gilbert [ACN3641], endemic to the MRB. With the growth of the barcode database, including all other neighboring river basins, we may be able to expose endemic cryptic species using BIN analysis as well as decreasing the risk of synonymies. Astyanax gr. bimaculatus had a 2.7 % mean intraspecific divergence between NJ tree clades and recovered BINs (Fig. 2a) but with no ABGD or morphological support. The genus Astyanax is a well know complex of species and one of the most specious neotropical fish groups with over 140 described species (Froese and Paulay 2010; Eschmeyer 2015), occurring in practically all river basins of the neotropics (Marinho and Lima 2009). Eigenmann (1921) redescribed Astyanax bimaculatus as being comprised of several subspecies, and Garutti (1995), reviewing nominal species, recognized 30 valid species, suggested the need for more systematic studies to objectively determine the actual levels of diversity within the bimaculatus complex. Interestingly, a deep divergence of 18.5 % was found when comparing one unidentified specimen of Astyanax sp. (ACJ1244) to other congeneric specimens sampled from the Mucuri basin. This highly divergent haplotype had high similarity (100 %) to an A. vermilion barcode sequence previously deposited in the BOLD database. However, A. vermilion is not described for the Mucuri basin, but from a coastal northeastern Brazilian river 300 km distant, and morphological analysis did not corroborate the barcode identification. Morphologic variation wider than described for A. vermilion or a mistake in the morphological identification of the species deposited as A. vermilion in BOLD may explain this discordance. The native species Prochilodus vimboides could be clearly differentiated from the introduced species P. costatus by DNA barcoding, since interspecific divergence of 9 % was recovered. Nonetheless, the non-indigenous species identified morphologically as P. costatus had a barcode similarity of 100 % with another non-indigenous species P. argenteus (BOLD number: BSB392-10). The divergence between the morphological and molecular identification of P. costatus may indicate the presence of
Genetica
Prochilodus hybrids in the Mucuri basin between the two species of the Sa˜o Francisco River. Hybrids of Prochilodus are commonly produced in aquaculture farms in Brazil and may escape into the wild. In the case of the electric knifefishes, the integrative approach based on molecular taxonomy and morphological characteristics allowed the precise identification of the overlooked species G. sylvius, not yet reported for the MRB. All Gymnotus specimens were first identified as G. aff. carapo and due to their high intraspecific genetic divergence and the presence of two clades in the NJ tree (Fig. 2c), the specimens were subjected to a more detailed morphological analysis, allowing their correct taxonomic identification. However, G. sylvius is described only for the Ribeira de Iguape, Paraı´ba do Sul, and Upper Parana´ River basins, but not for the MRB (Albert et al. 1999; Pompeu 2010). Therefore, its distribution range should be expanded to the MRB based on the results of our integrative approach. This integrative approach using a comprehensive inventory through the development of a barcode library combined with a detailed morphological identification was a robust methodology for species delimitation. For instance, our results indicated a new candidate species (Cyphocharax aff. gilbert [Ca ACN3641]) and detected overlooked species within the Astyanax (A. vermilion) and G. sylvius specimens. Therefore, our integrative pipeline using a simplified biodiversity model (i.e. Mucuri ichthyofauna) showed a possible remedy to the taxonomic impediment, and may be used to adequately describe and conserve the biodiversity of other complex neotropical ichthyofaunas. Acknowledgments We thank Angelo Monteiro, Danilo Neto, Iago Penido and Ivo for helping with the fieldwork and laboratory analysis. DCC is grateful to Conselho Nacional de Desenvolvimento Cientı´fico e Tecnolo´gico (CNPq) for his productivity fellowship (308537/20149). This work was funded by the Brazilian Barcode of Life Network BrBol/CNPq (564953/2010-5), Coordenac¸a˜o de Aperfeic¸oamento de ´ EQUIPAMENTOS Grant Pessoal do Ensino Superior (CAPES—PRO Number 783380/2013), Conselho Nacional de Desenvolvimento Cientı´fico e Tecnolo´gico—CNPq (Process Numbers INCT 573899/2008-8 and 482852/2011-9) and Fundac¸a˜o de Amparo a` Pesquisa do Estado de Minas Gerais—FAPEMIG (Process Number INCT APQ-0084/08).
References Agostinho AA, Thomaz SM, Gomes LC (2005) Conservac¸a˜o da biodiversidade em a´guas continentais do Brasil. Megadiversidade 1:70–78 Albert JS, Reis RE (2011) Historical biogeography of Neotropical freshwater fishes. University of California Press, Oakland Albert JS, Fernandes-Matioli FMC, Almeida-Toledo LF (1999) New species of Gymnotus (Gymnotiformes, Teleostei) from
southeastern Brazil: toward the deconstruction of Gymnotus carapo. Copeia 1999:410–421 Aljanabi SM, Martinez I (1997) Universal and rapid salt-extraction of high quality genomic DNA for PCR-based techniques. Nucleic Acids Res 25:4692–4693 Carvalho DC, Oliveira DAA, Pompeu PS, Leal CG, Oliveira C, Hanner R (2011) Deep barcode divergence in Brazilian freshwater fishes: the case of the Sa˜o Francisco River Basin. Mitochondr DNA 22:80–86 Castro RM, Vari RP (2004) Detritivores of the South American fish family Prochilodontidae (Teleostei: Ostariophysi: Characiformes): a phylogenetic and revisionary study. Smithsonian Contributions to Zoology, Washington Dayrat B (2005) Towards integrative taxonomy. Biol J Linn Soc 85:407–415 Edgar RC (2004) MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Res 32:1792–1797 Eigenmann CH (1921) The American characidae. Mem Mus Comp Zool 43:208–310 Eschmeyer WN (2015) Catalog of fishes: world wide web electronic publication. http://researcharchive.calacademy.org/research/icht hyology/catalog/fishcatmain.asp. Accessed 30 March 2015 Froese R, Paulay D (2010) FishBase: World Wide Web electronic publication. http://fishbase.org. Accessed 10 Dec 2014 Garutti V (1995) Revisa˜o taxonoˆmica dos Astyanax (Pisces, Characidae), com mancha umeral ovalada e mancha no pedu´nculo caudal, estendendo-se a` extremidade dos raios caudais medianos, das bacias do Parana´, Sa˜o Francisco e Amazoˆnica. Thesis, Universidade Estadual Paulista Garutti V (1998) Descric¸a˜o de uma espe´cie nova de Astyanax (Teleostei, Characidae) da bacia do Tocantins, Brasil. Iheringia, Se´r Zool 85:115–122 Hebert PDN, Cywinska A, Ball SL, deWaard JR (2003) Biological identifications through DNA barcodes. Proc R Soc Lond B Biol 270:313–321 Hebert PDN, Penton EH, Burns JM, Janzen DH, Hallwachs W (2004a) Ten species in one: DNA barcoding reveals cryptic species in the neotropical skipper butterfly Astrapes fulgerator. Proc Natl Acad Sci USA 101:14812–14817 Hebert PDN, Stoeckle MY, Zemlak TS, Francis CM (2004b) Identification of birds through DNA Barcodes. PLoS Biol 2:e312. doi:10.1371/journal.pbio.0020312 Hubert N, Hanner R, Holm E, Mandrak NE, Taylor E, Burridge M, Watkinson D, Dumont P, Curry A, Bentzen P, Zhang JB, April J, Bernatchez L (2008) Identifying Canadian freshwater fishes through DNA barcodes. PLoS One 3:e2490. doi:10.1371/journal. pone.0002490 ´ guas. Bacias hidrogra´ficas IGAM—Instituto Mineiro de Gesta˜o das A do Leste (2008) Bacia Hidrogra´fica do Rio Mucuri (MU1). www. igam.mg.gov.br/component/content/158?task=view. Accessed 2 August 2014 Jukes TH, Cantor CR (1969) Evolution of protein molecules. In: Munro HN (ed) Mammalian protein metabolism. Academic Press, New York Kekkonen M, Hebert PDN (2014) DNA barcode-based delineation of putative species: efficient start for taxonomic workflows. Mol Ecol Resour 14:706–715 Kimura M (1980) A simple method for estimating evolutionary rate of base substitutions through comparative studies of nucleotide sequences. J Mol Evol 16:111–120 Marinho MM, Lima FC (2009) Astyanax ajuricaba: a new species from the Amazon basin in Brazil (Characiformes: Characidae). Neotrop Ichthyol 7:169–174 MMA—Ministe´rio do Meio Ambiente (2006) Atlaˆntico Leste In Caderno da regia˜o hidrogra´fica Atlaˆntico Leste. Ministe´rio do Meio Ambiente, Secretaria de Recursos Hı´dricos, Brası´lia
123
Genetica Mora C, Tittensor DP, Adl S, Simpson AGB, Worm B (2011) How many species are there on earth and in the ocean? PLoS Biol 9:e1001127. doi:10.1371/journal.pbio.1001127 Nwani CD, Becker S, Braid HE, Ude EF, Okogwu OI, Hanner R (2011) DNA barcoding discriminates freshwater fishes from southeastern Nigeria and provides river system-level phylogeographic resolution within some species. Mitochondr DNA 22:43–51 Padial JM, Miralles A, De la Riva I, Vences M (2010) The integrative future of taxonomy . Front Zool 7(1):16. doi:10.1186/1742-9994-716 Pereira LHG, Maia GMG, Hanner R, Foresti F, Oliveira C (2011) DNA barcodes discriminate freshwater fishes from the Paraı´ba do Sul River Basin, Sa˜o Paulo, Brazil. Mitochondr DNA 22:71–79 Pereira LHG, Hanner R, Foresti F, Oliveira C (2013) Can DNA Barcoding accurately discriminate megadiverse Neotropical freshwater fish fauna? BMC Genet 14:14–20 Pompeu PS (2010) Os peixes do rio Mucuri. MG BIOTA 2:36–48 Puillandre N, Lambert A, Brouillet S, Achaz G (2012) ABGD, automatic barcode gap discovery for primary species delimitation. Mol Ecol 21:1864–1877 Ratnasingham S, Hebert PDN (2007) BOLD: the barcode of life data system (www.barcodinglife.org). Mol Ecol Notes 3:355–364 Ratnasingham S, Hebert PDN (2013) A DNA-based registry for all animal species: the barcode index number (BIN) system. PLoS One. doi:10.1371/journal.pone.0066213
123
Riedel A, Sagata K, Suhardjono YR, Ta¨nzler R, Balke M (2013) Integrative taxonomy on the fast track-towards more sustainability in biodiversity research. Front Zool. doi:10.1186/17429994-10-15 Saitou N, Nei M (1987) The neighbor-joining method: a new method for reconstructing phylogenetic trees. Mol Biol Evol 4:406–425 Taylor RW (1983) Descriptive taxonomy: past, present, and future. In: Highley E, Taylor RW (eds) Australian Systematic entomology: a bicentenary perspective. CSIRO, Melbourne Vari R (1992) Systematics of the neotropical characiform genus Cyphocharax Fowler (Pisces: Ostariophysi). Smithson Contrib Zool 529:1–137 Ward RD (2009) DNA barcode divergence among species and genera of birds and fishes. Mol Ecol Resour 9:1077–1085 Ward RD, Zemlak TS, Innes BH, Last PR, Hebert PDN (2005) DNA barcoding Australia’s fish species. Philos Trans R Soc B 360:1847–1857 Waugh J (2007) DNA barcoding in animal species: progress, potential and pitfalls. BioEssays 29:188–197 Zanata AM, Camelier P (2009) Astyanax vermilion and Astyanax burgerai: new characid fishes (Ostariophysi: Characiformes) from Northeastern Bahia, Brazil. Neotrop Ichthyol 7:175–184
