DNA Barcodes for the FIshes of the Narmada, One of India’s Longest Rivers

Please download to get full document.

View again

All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
  DNA Barcodes for the FIshes of the Narmada, One of India’s Longest Rivers
  DNA Barcodes for the FIshes of the Narmada, One of India’s Longest Rivers Gulab Dattarao Khedkar 1 * , Rahul Jamdade 1 , Suresh Naik  2 , Lior David 3 , David Haymer 4 1 Paul Hebert Centre for DNA Barcoding and Biodiversity Studies, Dr. Babasaheb Ambedkar Marathwada University, Aurangabad, India,  2 Biodiversity Institute of Ontario,University of Guelph, Guelph, Ontario, Canada,  3 Department of Animal Sciences, R.H. Smith Faculty of Agriculture, Food and Environment, The Hebrew University of Jerusalem, Rehovot, Israel,  4 Department of Cell and Molecular Biology, University of Hawaii, Honolulu, Hawaii, United States of America Abstract This study describes the species diversity of fishes of the Narmada River in India. A total of 820 fish specimens were collectedfrom 17 sampling locations across the whole river basin. Fish were taxonomically classified into one of 90 possible speciesbased on morphological characters, and then DNA barcoding was employed using COI gene sequences as a supplementalidentification method. A total of 314 different COI sequences were generated, and specimens were confirmed to belong to85 species representing 63 genera, 34 families and 10 orders. Findings of this study include the identification of five putativecryptic or sibling species and 43 species not previously known from the Narmada River basin. Five species are endemic toIndia and three are introduced species that had not been previously reported to occur in the Narmada River. Conversely, 43species previously reported to occur in the Narmada were not found. Genetic diversity and distance values were generatedfor all of the species within genera, families and orders using Kimura’s 2 parameter distance model followed by theconstruction of a Neighbor Joining tree. High resolution clusters generated in NJ trees aided the groupings of speciescorresponding to their genera and families which are in confirmation to the values generated by Automatic Barcode GapDiscovery bioinformatics platform. This aided to decide a threshold value for the discrimination of species boundary fromthe Narmada River. This study provides an important validation of the use of DNA barcode sequences for monitoringspecies diversity and changes within complex ecosystems such as the Narmada River. Citation:  Khedkar GD, Jamdade R, Naik S, David L, Haymer D (2014) DNA Barcodes for the FIshes of the Narmada, One of India’s Longest Rivers. PLoS ONE 9(July03): e101460. doi:10.1371/journal.pone.0101460 Editor:  Gyaneshwer Chaubey, Estonian Biocentre, Estonia Received  March 17, 2014;  Accepted  June 6, 2014;  Published  July 3, 2014 Copyright:    2014 Khedkar et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permitsunrestricted use, distribution, and reproduction in any medium, provided the srcinal author and source are credited. Funding:  This work is supported by Ministry of Food Processing Industries, Government of India, New Delhi under grant code 24/MFPI/R&D/2007 dated Dec 28,2011.(www.mofpi.nic.in). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing Interests:  The authors have declared that no competing interests exist.* Email: gdkhedkar@gmail.com Introduction Many questions in evolutionary biology, ecology, conservationbiology, and biogeography depend on knowledge of species as abiological unit. This makes it essential to critically evaluatemethods for determining both the identification of species andspecies boundaries. Also, in practice, many conservation programsdo not adequately address issues relating to intraspecific diversity,in part because of the difficulty in discriminating such variationthrough morphological analysis [1]. Increasingly, however, bothgenetic and DNA based tools are making it possible to obtain moredetailed and accurate assessments of biodiversity levels both withinand between species and to resolve cryptic species complexes. Thisinformation will also be essential for identifying conservation unitswithin species [2,3].The natural ecology of many river systems makes them an idealsetting for biodiversity studies [4]. Also from an ecologicalperspective, many of the world’s major rivers are under pressuredue to human activities. Asian rivers, in particular those in India,have been heavily impacted in this way. In addition, the impact of climate change and increasing human population density have ledto urgent calls for comprehensive biodiversity assessments toprovide baseline data on species distributions.Rivers in India are known to harbor a very diverse fauna. Thisincludes 868 species of freshwater fishes. Of these, 192 areendemic species and 327 species are listed as threatened by theIUCN [5]. This diversity of fishes reflects in part the presence of great riverine systems such as the Narmada, the third longest riverin India. Studies on the fish fauna of the Narmada River basinhave been conducted by researchers [6–8] and by governmentagencies (CIFRI) during the years 1985 to 1991 [9] using traditional methods of identification based on morphologicaltraits. The first published checklist of fish species by the CICFRIunit from Hoshangabad (1958–66) contained 77 species and thesecond, conducted by the department of fisheries, MadhyaPradesh, India (1967–71), recorded 46 species. Other studies [6]and [7] recorded totals of 76 species, and a third survey of theWestern zone of Narmada fish [8] reported 84 species. Finally, theCICFRI Barrackpore (1991) desk report of the Narmada Riverlisted 95 fish species [10].However, species identification using these methods can resultin misidentification due to high degree of phenotypic plasticity insuch characters leading to enlist different species and fluctuationsin species numbers. In these cases, alternative tools such as geneticand DNA based markers could help taxonomists to resolveambiguities to a great extent.One of these methods, DNA barcoding  [11], relies on thesequencing and comparison of a standardized portion of thegenome to aid in specimen identification and species discovery.The DNA barcoding method now represents the largest effort to PLOS ONE | www.plosone.org 1 July 2014 | Volume 9 | Issue 7 | e101460  catalogue biodiversity using molecular approaches. Althoughinitially regarded as controversial [12], numerous cases have beenreported where the analysis of DNA sequence variation in thecytochrome  c   oxidase subunit 1 (COI) region of mtDNA hasproven highly effective for the delineation and identification of animal species in general (see [13] for a review) and fish inparticular [14].Some of the controversies reflect the fact that early barcodestudies often examined only a few individuals of each species andwere limited in terms of geographic representation [15 – 17].  Although this approach did extend the inclusion of species indatabases, it often left gaps in understanding the extent of regional variation in barcode sequences within species, and deciding speciesboundary [18]. In addition, phylogeographic studies have shownthat past geological and climatic events have resulted in populationdifferentiation for freshwater organisms such as fishes because of their limited dispersal ability [19,20]. Thus, sampling schemes and reference databases must account for these phenomena to permitreliable delineation of species or major lineages. In such situations,new tools such as Automatic Barcode Gap Discovery (ABGD)algorithm have been developed to allow the partitioning of DNAsequence dataset into clusters of like taxa, i.e. candidate or‘primary’ species by utilizing a range of potential barcode gapthresholds [21]. This approach has been applied to the analysis of specimens from widely dispersed locales [22,23].This study aims to first develop a comprehensive DNA barcodelibrary for the fish fauna of the Narmada River. This can improvethe quality of future monitoring programs by linking barcodesequences with carefully identified voucher specimens. This studywill also provide a better understanding of the genetic variation infish fauna and the impact of ecological aspects of the river toprovide baseline information for creating improved conservationstrategies for the Narmada River ecosystem. Furthermore, theinformation should be more readily available to non- taxonomists,researchers and policy makers to aid in their efforts to establisheffective management of this ecosystem. Materials and Methods Ethical statement We declare that, the fish under study are not protected underwildlife conservation act and are routinely caught by professionalfisherman and sold as a food fish in Indian markets. No specificpermit is required for obtaining these fish in India, and noexperimentation was conducted on live specimens in thelaboratory. Sample collection This study examines fish species within the portion of theNarmada River basin that lies between Vindya and Satpuraranges (Figure 1, Table 1). The River has its source near Amarkantak (22 u 40 9 0 0 N to 81 u 45 9 0 0 E) in Madhya Pradesh, andtravels 1312 km before it discharges into the Gulf of Cambay inthe Arabian Sea (21 u 39 9 3.77 0 N to 72 u 48 9 42.8 0 E). The River iscomparatively straight with deep water and hard rocky substratesupporting a rich benthic fauna. Fishes were collected between July 2009 to December 2012 at 17 sites along the main river andits tributaries with  , 100–200 km distance between successivestations (Figure 1; Table S1). Most of the fish specimens weredigitally photographed, in case of multiple specimens, represen-tative images were used. Four species that were lacking imagesinclude  Acanthophagus latus, Mystus gulio, Hyporamphus dussumieri   and Parachaeturichthys ocellatus  . (For detailed methodology refer MethodsS1) Data analysis Sequence alignment and assembly was carried out using Codoncode Aligner v.3.0.1 (CodonCode Corporation) and MEGA 5[24]. Sequence divergence values within and among species wereemployed the Kimura two parameter (K2P) model [25] using analytical functions on BOLD v3.1 ( www.boldsystems.org  ). Aneighbor joining (NJ) tree based on K2P distance, nearestneighbor analysis (NN), and nucleotide composition values werealso obtained using BOLD. The analysis of genetic distances wascomplemented by downloading of related sequences fromGenBank for comparison with specimens of   Labeo dyocheilus, Puntius sarana, Liza subviridis, Nematalosa japonica and Mystus spp . Thesespecies have deep divergence values that can lead to puzzling identifications. In these and other cases, we have used the ABGD(automated barcode gap discovery) interface web tool available at:http://wwwabi.snv.jussieu.fr/public/abgd/abgdweb.html [21].For the analysis the ABGD method was first implemented using default parameters and Kimura 2-parameter (K2P) distances tocorrect for transition rate bias (relative to transversions) in thesubstitutions [25]. The default for the minimum relative gap widthwas set to different values between 0 and 1.2. Sequences werealigned and submitted to BOLD project code DBFN andGenbank with accession numbers JX983210–JX983514 (TableS2) (  dx.doi.org/10.5883/DS-NFDB  ). Results Taxon Diversity  A total of 820 fishes belonging to 90 species, 63 genera, and 34families (Table S3) were collected at the 17 sites. We generated atotal of 314 COI sequences for 83 species (attempts to extract goodquality DNA from two species were not successful and did notproduce barcodes). The collections included 43 (50%, SE=0.02)fish species that were not previously known from the NarmadaRiver basin. Three of these taxa could only be identified to ageneric level. Also five species endemic to India (  Esomus danricus,Glyptothorax lonah, Mystus montanus, Salmophasia boopis, Scatophagus argus   ), and three introduced species (  Cyprinus carpio, Hypophthal-michthys nobilis, Oreochromis mossambicus   ) which have not beenpreviously reported from the Narmada River were included inthis total. Conversely, 35 (29%, SE=0.001) species previouslyreported from the Narmada [26,6] were not encountered (Figure 2;Table S4). All amplified sequences were . 500 bp (mean, 625 bp) with noinsertions, deletions, stop codons and NUMTs. The overall GCcontent was 45.04% (SE=0.18) and highest in perches (46.27%;SE=0.02), followed by cyprinids (44.85%; SE=0.01) andcatfishes (44.27%; SE=0.02). The mean GC content at codonpositions 1–3 was 56.74% (SE=0.08), 42.9% (SE=0.03) and35.17% (SE=0.29) respectively. Nearly all species exhibitedunique barcode haplotypes or cohesive clusters of very closelyrelated haplotypes, which permitted the differentiation of 94%(SE=0.01) of species. All sequences were submitted to the BOLDproject DBFN (  dx.doi.org/10.5883/DS-NFDB  ). Four of theserepresent new records for NCBI Genebank and 12 species forBOLD Systems. COI sequence divergence analysis Out of the 85 species, 83 were well differentiated by COIbarcoding with average within species variability of 0.36%(SE=0.008) compared with 12.29% (SE=0.06) for species withingenera (Table 2 and Figure 3). Values of 17.87% (SE=0.02) and22.47% (SE=0.02) within families and orders, respectively, werealso obtained. We were not able to generate barcodes for two Fishes of the Narmada River in IndiaPLOS ONE | www.plosone.org 2 July 2014 | Volume 9 | Issue 7 | e101460  species, (  Parachaeturichthys ocellatus   and  Terapon jarbua   ). From the values obtained, a steady increase in genetic diversity was observedwith increasing taxonomic levels, supporting a marked change ingenetic divergence at species boundaries. The average congeneric variability is almost 40 fold higher than the conspecific values, andthis also produces a high level of resolution between clusters in theNJ tree to group the species to their corresponding genera andfamilies with sufficient bootstrap support (Table 2, Figure 3 & 4). Pairwise distances and Automatic Barcoding GapDiscovery (ABGD) The analysis using the ABGD tool with standard settings at firstdid not return any results. After lowering the X value (X=relativewidth of the barcoding gap) to 1.2, the ABGD analysis showed aclustering of the sequences into 8 molecularly defined operationaltaxonomic units (MOTUs) for the COI (Figure 5). Here, we used aprior intraspecific divergence value of (P=0.0215, SE=0.02)which is congruent with the primary species concept. The ABGDresults were confirmed independently of the chosen model (Jukes-Cantor and Kimura) and were unaffected by changes of priorlimits for intraspecific variation and threshold. The prior maximal Figure 1. Map Showing sampling sites within the Narmada River basin and its tributaries. doi:10.1371/journal.pone.0101460.g001 Table 1.  Sampling stations on the Narmada River basin. Sr. no. Sampling site Latitude Longitude Elevation (m) 1 Dindori 22 u 56 9 52.65 0  81 u 04 9 35.09 0  6602 Rusa 22 u 32 9 59.31 0  80 u 44 9 44.66 0  5403 Bamhani 22 u 28 9 49.31 0  80 u 22 9 59.00 0  4484 Bargi 22 u 55 9 10.23 0  79 u 55 9 9.70 0  4145 Sultanpur 23 u 06 9 43.23 0  77 u 57 9 36.00 0  3496 Tawa 22 u 33 9 32.06 0  77 u 57 9 46.82 0  3507 Hoshangabad 22 u 45 9 52.09 0  77 u 42 9 55.22 0  2878 Kolar 22 u 57 9 39.63 0  77 u 20 9 32.04 0  4619 Harda 22 u 20 9 8.59 0  77 u 05 9 7.07 0  28410 Indirasagar 22 u 12 9 46.51 0  76 u 37 9 46.06 0  23911 Choral 22 u 14 9 21.80 0  76 u 03 9 42.64 0  17212 Mortakka 22 u 13 9 29.82 0  76 u 02 9 59.03 0  18013 Maheshwar 22 u 10 9 8.67 0  75 u 35 9 13.59 0  14514 Pati 21 u 56 9 36.54 0  74 u 44 9 43.66 0  19915 Sardar sarovar 21 u 52 9 26.58 0  73 u 41 9 23.42 0  1316 Rajpipla 21 u 55 9 26.80 0  73 u 26‘13.57 0  1017 Bharuch 21 u 40‘57.74 0  72 u 59‘37.91 0  07doi:10.1371/journal.pone.0101460.t001 Fishes of the Narmada River in IndiaPLOS ONE | www.plosone.org 3 July 2014 | Volume 9 | Issue 7 | e101460  distance of P=0.0215; SE=0.02 is sufficient to distinguish the fishspecies in this study (Figure 5). Here, the values below thethreshold are treated as false positives since they split real speciesinto two or more partitions. On the other hand greater values (  . P=0.0215, SE=0.02) are treated as false negatives since thesedrop the species to a no barcode gap. For example, at a priormaximal distance of P=0.0215 the  L. dussumier   (NF236) resultsshow congruence with the remaining individuals of this species,but when the prior maximal distance values is lowered (P=1.29,SE=0.02), it splits into separate partitions. Considering individualNF236 as a different species to genus Labeo when analyzed, theK2P distance values show a clear overlap between intraspecific(2.19%) and intrageneric (2%) divergences (Table 2). Thisconfirming that, individual (NF236) does not represent a speciesdistinct from  L. dussumier  . This supports the robustness of barcodebased delineation of fish species in this study as well as theappropriate use of threshold value.The nearest-neighbor distance (NND) analysis revealed theclosest conspecific individuals to be at an average distance of 0.36% (SE=0.008) based on a range from 0 to 1% for 93% of theindividuals and , 3% for the remaining 7% of individuals(Figure 3). The lowest interspecific divergence was observedamong   Labeo species   (2.19%; SE=0.008) and highest in  Channa species   (24%; SE=0.03). Intraspecific divergence and possible hidden taxa i. Labeo dyocheilus.  Genetic divergence (K2P) among individuals of   Labeo dyocheilus   occurring in the Banjar tributary of the Narmada River (Figure 6), was the highest (2.98%) of anyregion sampled here, indicating the possible presence of sibling species or recently diverged and geographically subdividedpopulations (voucher ids NF136). Relatively little conspecific variation (0.30% to 0.33%; SE=0.03) within lineages wasobserved (Figure 6). When analyzed with AGBD we foundoptimal threshold level (P=0.0215, SE=0.03) to infer NF136 as aputative new species of genus Labeo from the Narmada River. NJtree analysis showing higher boot strap values are also inconfirmation with this new lineage (Figure 6). ii. Puntius sarana.  Puntius sarana   at the Dindori sampling station also exhibited extensive divergence (2.35%; SE=0.07),forming subclusters in the NJ tree (Figure 6) with intercluster values ranging from 0 to 1.89% (SE=0.04). This also suggests thepresence of possible sibling species or recently diverged geograph-ically subdivided populations (voucher ids NF115 & NF98). Toconfirm this, we have analyzed all the individuals of this speciesusing ABGD threshold values for partitioning (P=0.0215; SE Figure 2. Comparison of fish species with species records fromearlier studies. (*specific species to the study; **commonspecies for two studies; ***common species for all studies). doi:10.1371/journal.pone.0101460.g002     T   a    b    l   e    2 .     G   e   n   e    t    i   c     d    i   v   e   r   g   e   n   c   e    (    K    2    P    )   w    i    t     h    i   n   v   a   r    i   o   u   s    t   a   x   o   n   o   m    i   c     l   e   v   e     l   s .     D    i   s    t   a   n   c   e    (    %    )   n    T   a   x   a    C   o   m   p   a   r    i   s   o   n   s    M    i   n    M   e   a   n    M   a   x    S    E     W    i    t     h    i   n    S   p   e   c    i   e   s    2    9    0    6    0    8    5    6    0    0 .    3    6    2 .    9    9    0 .    0    0    8    W    i    t     h    i   n    G   e   n   u   s    1    4    2    1    0    1    0    0    3    4 .    4    8    1    2 .    2    9    2    3 .    1    2    0 .    0    6    5    W    i    t     h    i   n    F   a   m    i     l   y    2    5    1    1    3    9    0    3    3    4 .    6    6    6    1    7 .    8    7    3    2 .    1    0    0 .    0    2    W    i    t     h    i   n    O   r     d   e   r    1    2    3    1    0    1    1    0    1    0    1    5 .    8    3    6    2    2 .    4    3    3    2 .    4    7    0 .    0    2    5     d   o    i   :    1    0 .    1    3    7    1    /    j   o   u   r   n   a     l .   p   o   n   e .    0    1    0    1    4    6    0 .    t    0    0    2 Fishes of the Narmada River in IndiaPLOS ONE | www.plosone.org 4 July 2014 | Volume 9 | Issue 7 | e101460  0.02) and bootstrap analysis. This is consistent with the suggestionthat NF115 and NF98 are putative sibling species within the genusPuntius (Figure 6). iii. Liza species.  Similarly  t  he  Liza sp . collected from theBharuch estuary in the Gulf of Cambay shows 6% divergencewhen compared with the sister species  Liza klunzingeri   (avg.diveregence 0.98%; SE=0.1) and the sequences obtained fromGenBank as  Liza sp.  (avg. divergence 0.43%; SE=0.001). The restof species also show extensive divergence (  Liza subviridis,  0.64%  –  0.68%, SE=0.03;  Liza macrolepis,  0.72%  –  0.74%, SE=0.1). The ABGD analysis further clarifies the species partitioning at anoptimum threshold value (P=0.0215; SE=0.02). Here, fourpartitions are formed, and this supports NF550 and NF565 asputative new species (Figure 6). This analysis also clearly indicatesthe species downloaded for analysis from NCBI Genebank (EF607446.0 and EF607447.1) recorded as Liza spp. were notdifferent from  Liza subvirdis   as the threshold values and bootstrapsupport can not partition them separately (Figure 6). iv. Nematalosa species.  The  Nematalosa   sp. (NF257) collect-ed from Hoshangabad shows 10% genetic variation, whereassequences from the sister species  Nematalosa nasus   shows an averagedivergence of 0.173% (SE=0.09). Comparatively, GenBank sequences of   Nematalosa nasus   (HQ231349.1, HQ231350.1) and  Nematalosa erebi   (EF609412.1) showed an average genetic distanceof 0.193% (SE=0.05) with  Nematalosa japonica   (AP009142.1,EF607513.1) 0.181 (SE=0.1) (Figure 6). The ABGD basedanalysis shown for a threshold value (P=0.01  –  1.00, SE=0.02)can partition these four species accurately and suggests that NF257may be a sibling species of the genus Nematolosa.  v. Mystus species.  ThespeciesbelongingtothegenusMystusare native to India. Four species of   Mystus   (   M. bleekeri, M. cavasius, M.vitatus, M. guilio, M. malbaricus   ) were collected in the Narmada River.One specimen procured from the Banjar tributary could only beidentifiedtothegenuslevelbasedonahigherK2Pdivergencevalue.To clarify this a few GenBank records for mystus species (   M.malbaricus,  HQ219109.1-HQ219111.1;  M. vitatus,  JN228952.1, JN228053.1) were included in our anaysis. This result shows 11%(SE=0.03) genetic divergence with an average value of 0.12%(SE=0.001) between both  M. malabaricus   and  M. vittatus   (Figure 6).TheABGDbasedanalysispartitioned,withoutambiguity,thesefivedescribed species (   M. bleekeri, M. cavasius, M. vitatus, M. guilio, M.malbaricus   )andonesuspectedputativenew/siblingspeciesabsolutely Figure 3. Distribution of conspecific and congeneric K2P mean divergence of 83 fish species from the Narmada River (ascendingorder).  The maximum conspecific divergence (2.9%, blue solid circles) and minimum congeneric divergence (4.66%, black hollow circle) representthe threshold level of conspecific and congeneric divergence respectively. Data series were represented by more than one sequence. 93% of the total83 species showed divergence below # 1% and represented true species.doi:10.1371/journal.pone.0101460.g003Fishes of the Narmada River in IndiaPLOS ONE | www.plosone.org 5 July 2014 | Volume 9 | Issue 7 | e101460
Related Search
Similar documents
View more
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!