Distribution and Diversity of Libellulidae (Odonata: Anisoptera) from Indo-Burma Biodiversity Hotspot Region and their Phylogenetic Organization

Abstract

Mizoram represents an important part of the Indo-Myanmar biodiversity hotspot situated in the southernmost part of Northeast India. Direct searching and observation method with opportunistic sample collection was employed to survey the distribution of dragonfly family Libellulidae. The distribution according to altitudinal range and forest types has been established and correlated with humidity, rainfall and temperature. Twenty eight species of Libellulids are recorded in the present study; of these ten species are new records from the Northeastern Indian region. Higher diversity was confined to the lower altitudes of the Tropical Wet Evergreen forests. The phylogenetic relationship among the Libellulids was established using two mitochondrial genes CO1 and ND1. The molecular phylogeny was inferred by using maximum parsimony, maximum likelihood and the Bayesian methods. Among the eighteen genera analyzed, Trithemis, Neurothemis, Tramea and Orthetrum were recovered as monophyletic. The nucleotide distance between Tramea limbata and Tramea basilaris was found to be lowest and the highest was found between Potamarcha congener and Neurothemis tullia in the combined CO1 and ND1 gene.

Keywords

Altitudinal range, Forest types, Mitochondrial gene, New record and nucleotide distance

Introduction

According to the checklist of Odonata of India there are 87 species of Libellulidae under 39 genera [1]. Lahiri recorded eight species of Libellulids under seven genera from Mizoram [2]. Mitra recorded 18 species in 13 genera of Libellulids from Mizoram [3]. The Zoological Survey of India recorded 24 species in 15 genera from Libellulidae family [4].

Libellulidae are one of the diverse groups of Odonata with a controversial phylogeny, which have been resolved to a certain extent by molecular dataset [5,6]. Eleven monophyletic subdivisions of Libellulidae are tentatively recognized as subfamilies [7]. The mitochondrial gene Cytochrome c Oxidase 1 (CO1) can serve as the core of a global bio-identification system for all animal phyla [8]. NADH Dehydrogenase Oxidase Subunit 1 (ND1) sequence analysis in dragonfly studies revealed strong interspecific and intraspecific differences in the population structures of all species and has been shown to be highly informative at different taxonomic levels in dragonflies [9,10].

Mizoram is located at latitude 21°58’ and 24°35’ N; 92°15’ and 93°29’E longitude covering an area of 21,081 sq. km. Mizoram represents an important part of the Indo-Myanmar biodiversity hotspot situated in the southernmost part of North-East India [11]. The forest can be classified into three broad types Tropical Wet Evergreen Forest, Tropical Semi Evergreen Forest and Mountain Sub Tropical Forest [12]. The genetic relationships of the Libellulidae species found in Mizoram are not known yet and analysis of the phylogenetic work within Libellulidae may provide a framework for understanding the relationship between taxa.

Hence, the present study was carried out to understand the diversity and distributional pattern of Libellulidae in Mizoram and to study the phylogenetic relationship among the Libellulids using two mitochondrial gene markers CO1 and ND1.

Materials and Methods

Sample Collection

Opportunistic sample collection of the Libellulids was used and target species or group of species were observed visually [13]. All the specimens were identified with the help of identification keys provided by Fraser, Prasad and Subramanian [14,15].

The altitudinal range in the study area was divided into 4 categories [<100 m; 100-500 m; 500-1000 m; and >1000 m above sea level (asl)] in the state of Mizoram, Northeast India (Figure 1). The number of species sighted during the survey between 2010 and 2012 were counted in the sampling areas and the Libellulids were categorized based on their abundance in the different altitudinal range. The distribution of each species in all the four altitudinal ranges have been generated from the data collected from all sampling sites. Shannon and Simpson diversity indices were calculated as a measure of diversity in each altitudinal range. Berger-Parker dominance has also been evaluated for each altitudinal range. For the data analysis Biodiversity Pro software was used [16].

Figure 1: Mizoram map showing surveyed sites.

Sampling was done in altitudinal range below 100 m asl which was confined within Tropical Wet Evergreen Forest. In the Tropical Semi Evergreen Forest, the altitude ranged between 100 to 1000 m and the altitudinal ranges above 1000 m asl were found to be confined within Mountain Sub Tropical Forest. During the survey period (2010 - 2012), the monthly average rainfall, temperature and humidity in each forest types were recorded. The diversity of Libellulidae found in the three forests type was also generated using the same three diversity indices. The diversity of dragonfly was statistically correlated with the temperature, humidity, rainfall during the survey period.

Total genomic DNA was extracted from all individuals using modified protocol of [17]. The COI and ND1 gene were amplified using previously reported primer sequences [8,10]. The mitochondrial genes were amplified using PCR performed in a final volume of 25 μl. The PCR reaction mixtures contained 1X amplification buffer, 2.5 mM MgCl₂, 0.25 mM dNTP, 0.2 pM each primer, 0.8 BSA and 0.5 U Taq DNA polymerase. The PCR thermal regime for amplification was 5 min at 95°C, followed by 30 cycles of 30 s at 95°C, 40 s at 43° - 58°C, 30 s at 72°C and a final elongation for 6 min at 72°C.

The PCR products were sequenced using Sanger’s di-deoxy method (GCC Biotech, Kolkata). All the sequences were checked for contaminations using BLAST. Sequences were aligned and checked using Pairwise Sequence Alignment (EMBOSS-water, EBI) and FinchTV version 1.4.0 followed by manual adjustments. All the protein coding sequences were translated into amino acids and their ORFs checked [Sequence Manipulation Suite (Bioinformatics.org) and ORF Finder (NCBI)] [18].

Phylogenetic Analysis

Phylogenetic and molecular evolutionary analyses were conducted using MEGA version 6 [19]. The sequences of CO1 and ND1 from each species were combined and aligned using ClustalW implemented in the program MEGA6 [19]. Relationships between individual taxa were assessed by maximum likelihood (ML) method with nucleotides distances (p-distance), transition/ trans version rate ratios and nucleotide diversity.

Maximum parsimony (MP) trees were obtained using PAUP* by heuristic search option with tree-bisection-reconnection (TBR) branch-swapping [20]. The number of bootstrap replicates was set at 100. Starting tree was obtained via stepwise addition and the number of trees held at each step during stepwise addition equals one. Steepest descent option was not in effect and the initial MaxTrees setting was 100 wherein gaps were treated as missing data. All characters were equally weighted, and zero length branches were collapsed to polytomies. Multistate taxa were interpreted as uncertainty, topological constraint was not enforced and the generated 50% consensus trees were saved.

Maximum likelihood (ML) analysis was performed with RaxML [21]. Phylip file was generated for RaxML analysis using ALTER Alignment Transformation Environment [22]. For ML analysis, the bootstrap was set at 100 and the model was set at GTR-GAMMAI.

Bayesian inference was performed with MrBayes v3.1.2 [23]. Bayesian posterior probabilities for the clades were obtained using Metropolis-coupled Markov chain Monte Carlo (MCMC) analysis as implemented in MrBayes. Trees were sampled every 100 generations and burn-in was assessed with Tracer v1.5 [24].

Results

A total of 28 species under 18 genera Libellulids were found to be present in Mizoram and 10 species were record for the first time during the present study period of 2010 - 2012 (Table 1). Shannon H’ and Simpson’s result showed that diversity was highest (1.315 and 21.474) below 100 m asl and in the altitudinal ranges between 100 - 500 m diversity was slightly higher (1.172 and 14.895) than 500 - 1000 m asl range (1.163 and 14.875). The least diversity (1.041 and 11.262) was found at an altitude above 1000 m asl. High dominance percentage (12.5) was found in the altitudinal range above 1000 m asl, in the altitudinal ranges between 100 - 500 m and 500 - 1000 m asl the species dominance percentage were 9.375 and 9.74 respectively and the lowest dominance percentage (8) was found below 100 m range (Table 2).

Table 1. List of Taxa recorded for first time in Mizoram, Northeast India.

Table 2. Diversity indices and Berger-Parker dominance among the different altitudinal zones.

Altitudinal ranges below 100 m asl confined within Tropical Wet Evergreen Forest, was found to have the highest number of species diversity (1.32 in Shannon and 21.47 in Simpson), highest average humidity (78.18%) and highest average temperature (23.59°C) during the study period. Altitudinal ranges between 100 m and 1000 m asl confined within Tropical Semi Evergreen Forest was found to received highest average rainfall (337.42 mm) with an average temperature of 22.56°C and average humidity 75.41% during the surveyed period, the diversity was 1.19 in Shannon and 14.64 in Simpson. The altitudinal ranges above 1000 m asl confined within Mountain Sub Tropical Forest was found to have the lowest species diversity (1.04 in Shannon and 11.26 in Simpson), lowest average temperature (20.64°C) and lowest average humidity (70.12%). In Tropical Wet Evergreen forest, the average rainfall received during the surveyed period was 238.81 mm and in the Mountain Sub Tropical Forest, it was 228.76 mm (Table 3). The correlation between species richness, temperature and humidity was statistically significant but there was no significant correlation between species richness and rainfall.

Table 3. Diversity indices and physical factors in the three forest types during 2010-2012.

CO1 and ND1 Data Analysis Using MEGA6

CO1 primers amplified approximately 750 bp long fragment of the mitochondrial genome of CO1 gene for all the 28 species of Libellulids identified in the present study. ND1 primer amplified approximately 580 bp long fragment of the mitochondrial genome, which includes fragments of 16S rRNA, the intervening tRNA leu region and the ND1 gene region. The sequences generated for the present study were submitted in GenBank (Table 4).

Table 4. Voucher name and GenBank accession number given for the present study.

The genetic distance using CO1 and ND1 data between the 28 species of Libellulids of Mizoram was generated. The nucleotide maximum distance (p-distance) between Tramea limbata and Tramea basilaris was found to be lowest at 0.07 and the highest p-distance was found between Potamarcha congener and Neurothemis tullia at 0.32 (Table 5).

Table 5. The ND1 genetic distance (p-distance) of the Libellulids of Mizoram.

The estimated Transition/Transversion bias (R) is 1.005. Substitution pattern and rates were estimated under the General Time Reversible model (+G). A discrete Gamma distribution was used to model evolutionary rate differences among sites (five categories, [+G], parameter = 0.2603). The Maximum composite likelihood estimate of transitional substitution matrix between A/G = 13.7, T/C = 22.29, C/T = 8.65 and G/A = 8.66. The nucleotide frequencies are 27.72% (A), 39.45% (T), 15.30% (C), and 17.53% (G). Codon positions included were 1^st, 2^nd, 3^rd and non-coding. There were a total of 947 positions in the final dataset and positions containing gaps and missing data were eliminated.

Phylogenetic Analysis among Libellulidae

The sequences of CO1 and ND1 were combined and in Parsimony analysis a total of 947 characters were included; 475 characters were constant and the number of parsimony informative characters was 367. Bayesian analysis for the combined gene set was set at 10 million generations resulted in trees. The effective sample size was 2057.8059 and LnL score was ‒10988.589. The first 20% trees were considered as the burn-in phase and discarded.

The phylogenetic analysis supported six major clades (Figure 2). Clade 1 contained all the three Trithemis, Clade 2 included A. brevipennis and P. congener as sister clade; B. contaminata and T. tillarga as sister clade and a separate D. trivialis. Clade 3 included D. nebulosa, P. sexmaculata and B. geminata which formed a sister clade with C. servillia and clade 4 having all the four Orthetrum which formed a sister clade with C. lineata. Clade 5 has the two Tramea forming a sister clade with R. variegata and P. flavesence and clade 6 included the three Neurothemis which formed a sister clade with A. paranorpoides and B. chalybea and together they formed a sister clade with B. sobrina and T. Platyptera.

Figure 2: Maximum likelihood tree based on ND1 and CO1 sequence of Libellulidae. Clades with more than 50% support value in the order ML, Bayesian and MP are shown.

Discussion

The present study included 28 species of the Libellulidae which is 34% of the Libellulidae found in India, of these ten species are new record from the area. Higher diversity was confined to the lower altitudes of the Tropical Wet Evergreen forests. A significant correlation between species richness with temperature and humidity was found but rainfall did not significantly affect the species richness.

The distributions of Libellulidae are widespread and are represented by more than 1000 species and attempt has been made to organize the Libellulidae into subfamilies using molecular and morphological characters. In the combined gene of CO1 and ND1, transitional substitution matrix between C/T was highest and the nucleotide frequencies were high in AT which is typical for arthropods [25]. The nucleotide maximum distance (p-distance) between Tramea limbata and Tramea basilaris was found to be lowest and but highest value was between Potamarcha congener and Neurothemis tullia.

Among the 18 genera of the Libellulide of Mizoram, six of them are represented by more than one species. The Neurothemis represented by three species and the Tramea represented by two species were recovered in one clade in all the analysis with high bootstrap value. Trithemis having three species were recovered in one clade in Bayesian analysis but the status of T. pallidinervis was not conclusive since the support value of Maximum Likelihood and Maximum Parsimony analysis was low. The Orthetrum represented by four species were also recovered as monophyly but the bootstrap value of O. sabina was low in ML and MP analysis and C. lineata was included in the clade with low bootstrap value. The Brachydiplax represented by two species were not recovered as monophyly in any of the analysis but remained in one group. The Diplacodes represented by two species were recovered as monophyletic only in Bayesian analysis. A. paranorpoids and B. chalybea seems to form a close relationship with good support value in all the analysis of the 11 subfamilies of Libellulidae recognized by Carle et al., nine subfamilies were included in the present study. Zyxommatinae represented by Brachythemis and Tholymis were in one clade, Palpopleurinae represented by six genera were found in two separate groups one group containing Diplacodes, Palpopleura, Bradinopyga and Crocothemis, another group containing Neurothemis and Acisoma. The Libellulinae represented by Orthetrum and Cratilla remained in one group but Potamarcha was in a different clade. Pantalinae represented by Trithemis and Pantala were not recovered in one clade [7,26].

Conclusion

The present study revealed the diversity and distributional patterns of the Libellulidae in Mizoram and the phylogenetic analysis from this area may contribute to the classification of the family Libellulidae into a smaller group and offered a better understanding of the diversity within the mitochondrial genes.

Acknowledgments

The authors thank Department of Biotechnology, New Delhi, India for assistance through Bioinformatics Infrastructure Facility (No. BT/BI/12/060/2012(NERBIF-MUA). We thank Ministry of Social Justice and Empowerment and Ministry of Tribal Affairs for funding UGC’s Rajiv Gandhi National Fellowship Scheme for SC/ST [F-14-265(ST)/2007(SA-III), March 2007]. The authors acknowledge Praveen Karanth, IIsc, Bangalore for the phylogenetic analysis.

References

1. Subramanian KA. A Checklist of Odonata of India (Version 2.0.2014). Zoological Survey of India, 2014.
2. Lahiri AR. Odonata (Insecta) from different states of northeastern India. Oriental Insects 1979;13:119-132.
3. Mitra TR. Geographical Distribution of Odonata (Insect) of Eastern India. Zoological Survey of India, Kolkata. Memoirs of the Zoological Survey of India 2002;19:1-208.
4. Prasad M. Fauna of Mizoram, State fauna series. Zoological Survey of India, Kolkata, India 2007;14:143-186.
5. Dijkstra KDB. A review of the taxonomy of African Odonata - finding ways to better identification and biogeographic insight. Cimbebasia 2003;18:191-206.
6. Ware JL, et al. Phylogeny of the higher Libelluloidea (Anisoptera: Odonata): an exploration of the most speciose superfamily of dragonflies. Molecular Phylogenetics and Evolution 2007;45:289-310.
7. Carle FL, et al. A molecular phylogeny and classification of Anisoptera (Odonata). Arthropod Systematics and Phylogeny 2015;73:281-301.
8. Hebert PD, et al. Biological identifications through DNA barcodes. Proceedings of the Royal Society B Biological Sciences 2003;270:313-321.
9. Hadrys H, et al. The present role and future promise of conservation genetics for forest odonates.Forest and Dragonflies. 4th WDA International symposium of odonatology, Pontevedra (Spain) 2006;279-299.
10.  Rach J, et al. Character-based DNA barcoding allows discrimination of genera, species and populations in Odonata. Proceedings of the Royal Society B Biological Sciences 2008;275:237-247.
11.  Pai M. 2008 Fauna of Northeast India. Accessed on November 2012.
12.  Pachuau R. Geography of Mizoram. Northern Book Centre, New Delhi, 2009.
13.  Sutherland WJ. Census of fauna in Ecological Census Techniques. University Press, New York, 1996.
14.  Fraser FC. Fauna of British India including Ceylon and Burma Vol. 3. Taylor and Francis LTD., London, 1936.
15.  Subramanian KA. Dragonflies of India, A field guide. Vigyan Prasar, Department of Science and Technology, Noida, India, 2009.
16.  McAleece N, et al. BioDiversity Professional statistics analysis software. Jointly developed by the Scottish Association for Marine Science and the Natural History Museum, London, 1997.
17.  Zimmermann M, et al. Phylogeny of Euphydryas Checkerspot Butterflies (Lepidoptera: Nymphalidae) Based on Mitochondrial sequence data. Annals of the Entomological Society of America 2000;93:347-355.
18.  Patterson J, et al. 2004-2006. FinchTV version 1.4.0. Geospiza Inc.   
19.  Tamura K, et al. MEGA6: Molecular Evolutionary Genetics Analysis version 6.0. Molecular Biology and Evolution 2013;30:2725-2729.
20.  Swofford DL. PAUP* Phylogenetic Analysis Using Parsimony (*and Other Methods), Sinauer Associates, Sunderland, MA, 2002.
21.  Silvestro D and Michalak I. raxmlGUI: A graphical front-end for RAxML. Organisms Diversity and Evolution 2012;12:335-337.
22.  Glez-Peña D, et al. ALTER: program-oriented format conversion of DNA and protein alignments. Nucleic Acids Research. Web Server issue. ISSN, 2010;0305-1048.
23.  Ronquist F and Huelsenbeck JP. MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 2003;19:1572-1574.
24.  Rambaut A and Drummond AJ. Tracer v1.4, 2007.
25.  Simon C, et al. Evolution, weighting, and phylogenetic utility of mitochondrial gene sequences and a compilation of conserved polymerase chain reaction primers, Annals of the Entomological Society of America 1994;87:651-701.
26. Nei M and Kumar S. Molecular Evolution and Phylogenetics. Oxford University Press, New York, 2000.