A New Species of Frog (Anura: Dicroglossidae) Discovered from the Mega City of Dhaka

We describe a new species of frog of the genus Zakerana discovered from the urban core of Dhaka, Bangladesh, one of the most densely populated cities in the world. Although the new species is morphologically similar to the geographically proximate congeners in the Bangladeshi cricket frog group, we show that it can be distinguished from all congeners on the basis of morphological characters, advertisement calls and variation in two mitochondrial DNA genes (12S rRNA and 16S rRNA). Apart from several diagnostic differences in body proportions, the new species differs from other Zakerana species in having a flattened snout (from ventral view) projecting over the lower jaw, and diagnostic trapezoid-shaped red markings on the vocal sac in males. Molecular genetic analyses show that the new species is highly divergent (3.1–20.1% sequence divergence) from all congeneric species, and forms a well-supported clade with its sister species, Zakerana asmati. The discovery of a new amphibian species from the urban core of Dhaka together with several recent descriptions of new amphibian species from Bangladesh may indicate that more amphibian species remain to be discovered from this country.

In the present paper we describe a new species of Zakerana from the urban core of Dhaka (the capital of Bangladesh), one of the most densely populated mega cities in the world [35].While the new species is morphologically similar to the above-mentioned five congeners, we show that it can be distinguished on the basis of significant differences in morphology, bioacoustics, and genetic variability in two mitochondrial genes (16S and 12S rRNA).Given that Dhaka continues to grow and develop seemingly without any particular attention to the loss and degradation of its wetland ecosystems [36,37], we hope that the discovery and description of a new frog species from its urban core would evoke public interest and help to facilitate survey and conservation of amphibian biodiversity in Bangladesh.

Ethics statement
This study was conducted with appropriate permissions (CCF letter no.22.01.0000.101.23.2012.681for collecting specimens, CF memo no.22.01.0000.101.23.2012 for transport) and guidelines from the responsible authority, the Forest Department, Ministry of Forest and Environment, the People's Republic of Bangladesh.The protocols for our collection and research were approved by the committee of the Wildlife Section of the Forest Department, Bangladesh.Collected specimens were not recognized as belonging to threatened species and they are not listed in IUCN Redlist or by CITES.

Morphological measurements and analyses
Measurements were taken with digital calipers to the nearest 0.02 mm.The 21 characters that were measured followed the definitions of Islam et al. [38] and Howlader et al. [39,40]; their landmark definitions are depicted graphically in S1 Fig.The measurements included: SVL (distance from tip of snout to vent); HL (head length; distance from tip of snout to the back of mandible); HW (head width; maximum width of the head at the posterior margin of mandible); MN (distance from back of mandible to nostril); SL (snout length; distance from anterior corner of eye to the tip of snout); MFE (distance from back of mandible to front of the eye); MBE (distance from back of mandible to back of the eye); IN (internarial distance); IOD (interorbital distance); EN (distance from front of eyes to the nostril); NS (nostril-snout length); EL (eye width); UEW (maximum width of upper eyelid); HAL (hand length; distance from proximal end of outer palmar metacarpal tubercle to tip of third finger); FAL (forearm length; distance from corner of elbow to proximal end of outer palmar metacarpal tubercle); THIGHL (thigh length; distance from vent to knee); TL (tibia length; distance from knee to heel); TFOL (length of tarsus and foot; distance from heel to tip of fourth finger); FOL (foot length; distance from proximal end of inner metatarsal tubercle to tip of fourth finger).Webbing formula followed that of Glaw and Vences [41].
Apart from comparing ratios of different morphological traits between the new and other Zakerana species, we also performed discriminant function (DFA) and principal component analyses (PCA) to compare the new species (Zakerana sp.nov.) to its five (viz.Z. asmati, Z. nepalensis, pierrei, Z. syhadrensis, and Z. teraiensis) congeners for the 14 measurements (viz.SVL, HW, HL, SL, MBE, EN, NS, EL, IN, IOD, FAL, HAL, FOL, and TL) for which we had sufficient data.In the PCA, we extracted the two first axes, as they were the only ones with Eigenvalue > 1.Their scores were subjected to one-way ANOVA to formally compare if the mean scores differed among species.Tukey's HSD post-hoc tests were used to examine if the mean values for the new species were significantly different from those of the other species.
Examined specimens are deposited at the Finnish Museum of Natural History, Finland (MZH), and Zoology Department, University of Chittagong, Bangladesh (MZD).Additional specimens used for morphological comparisons were examined from the collections of the Zoology Department, University of Chittagong, Bangladesh (MZD).Their accession numbers are listed in S1 Table.

DNA analyses and phylogeny
Whole genomic DNA was extracted from muscle tissue (n = 3) using a silica-based method (Ivanova et al. [42]) and stored at -20°C.PCR amplification and sequencing of all samples was done with two pairs of primers (listed in S2 Table ).PCR conditions for amplification of both genes consisted of 5.72 μl of dH 2 O, 2 μl of 5x buffer, 0.08 μl of dNTP, 0.2 μl of Phire enzyme (Thermo Fisher) and 0.5 μl of each primer, in a total reaction volume of 10 μl.The PCR program was comprised of a preliminary denaturation step at 98°C for 30 s, followed by 34 cycles of 98°C for 10 s, 55°C for 10 s, 72°C for 30 s and final extension at 72°C for 1 min.PCR products were purified by using ExoSap IT (USB Corporation, Cleveland, OH, USA) and sequenced at the Institute for Molecular Medicine, Finland (FIMM).Sequence ambiguities were edited by eye by aligning forward and reverse reads using the Geneious 5.6.5 program [43].Obtained sequences were deposited in GenBank; accession numbers are provided in S3 Table .The nucleotide sequences of the 16S and 12S rRNA genes were separately aligned with available sequences of Zakerana obtained from GenBank (n = 20, S3 Table) using the default parameters in ClustalW (built into BIOEDIT [44,45]).From these two alignment datasets, sequence divergences (uncorrected p-values) were calculated using Mega v 5.5.6 [46] with the pairwise-deletion option, in which all aligned sites except indels were used for calibration.Phylogenetic analyses were conducted both with each gene fragment separately and also combined.The combined dataset was comprised of partial 16S (407 bp) and 12S (349 bp) gene sequences, resulting in a 756 bp sequence.Phylogenetic analyses were performed using Maximum likelihood (ML) and Bayesian inference methods.The most fitting nucleotide substitution model for the combined dataset was found to be the GTR + I + G substitution model.For the ML analysis, branch support was evaluated by using 1000 bootstrap replicates [47] as implemented in Mega v 5.5.6 [46].For the Bayesian analyses, Markov chain Monte Carlo runs were performed for one million generations with a sampling frequency of 100, as implemented in MrBayes 3.1.2[48].Convergence of the runs was assessed from the average split frequency of standard deviations (<0.01) and by checking the potential scale reduction factors (~1.0) for all model parameters.25% of the trees were discarded as burn-in and the remaining ones were used to generate a 50% majority rule consensus tree, as well as to estimate the Bayesian posterior probabilities.

Bioacoustic analyses
Call recordings were made using a Sony Cyber-shot camera on video mode (model: DSC-W530).During the recordings, the focal male was closer to the camera than to other calling neighbours.After locating an advertising male, advertisement calls were recorded from approx.0.4 m distance.Advertisement calls of holotypes of both the new species and Z. asmati were recorded in air temperature between 23°C and 24°C when the individuals were calling on land in their respective type localities (Fig 1).Calls of Z. asmati were recorded in front of Teacher's Club and Guest House (22°28'13.83"N,91°47'36.06"E),Chittagong University campus, Hathazari, Chittagong, Bangladesh on 15 May 2008.The calls of the new species were recorded in front of the Kabi Kazi Nazrul Islam Hall (23°46'15.25"N,90°22'39.87"E),Sher-e-Bangla Agricultural University campus, Sher-e-Bangla Nagar, Dhaka, Bangladesh on 11 June 2012.
Calls of the adult male (holotype) were analyzed with the acoustic software Adobe Audition 3.0 (following [49] and [50]), and compared to the described vocalizations of Zakerana species available in literature [34,[51][52][53].Recordings were re-sampled at 44.1 Hz and 16 bit resolution in the mono pattern and in "Waveform" extension.Frequency information was obtained through Fast Fourier Transformation (FFT, width 1024 points); the audio spectrogram was obtained at Hamming window function with 256 bands resolution.We measured dominant frequency (the frequency in call containing the greatest energy, determined from Fourier transformation), pulse repetition rate (number of pulses repeated in a defined period of time within a note; the value is provided as pules per second), number of pulses per note (total number of pulses in a note), note repetition rate (number of notes repeated in a defined period of time within a call; the value is provided as notes per second), note duration (duration of temporally uninterrupted sound element composing the call and made up of pulses), inter-note intervals (the time elapsed from the end of a note to the beginning of the next note), as described by Cocroft and Ryan [54], Köhler [55], and Martins and Jim [56].Mean, standard deviation and range (as well as number of analysed units, n), of call parameters with temporal measurements in seconds (s) or milliseconds (ms) are provided.Mann Whitney U-tests were used to determine if there are differences in call parameters between the new species and the phylogenetically closely related Z. asmati.
Statistical analyses were performed using R software 3.0.1 (The R Project for Statistical Computing, Vienna, Austria).

Nomenclatural acts
The electronic edition of this article conforms to the requirements of the amended International Code of Zoological Nomenclature, and hence the new names contained herein are available under that Code from the electronic edition of this article.This published work and the nomenclatural acts it contains have been registered in ZooBank, the online registration system for the ICZN.The ZooBank LSIDs (Life Science Identifiers) can be resolved and the associated information viewed through any standard web browser by appending the LSID to the prefix "http://zoobank.org/".The LSID for this publication is: urn:lsid:zoobank.org:pub:C9DA37D3-EB61-4444-8C57-DB2E35843200.The electronic edition of this work was published in a journal with an ISSN, and has been archived and is available from the following digital repositories: PubMed Central, LOCKSS.
Z. dhaka sp.nov.can be further distinguished from Z. asmati, Z. pierrei, and Z. teraiensis by the following additional criteria: distance from anterior margin of eye to the nostril is 1.20-1.29 times greater than distance from nostril to snout tip (vs.1.66-1.77 in Z. asmati; and 1.03-1.18 in Z. pierrei and Z. teraiensis); snout length 38-40% of head length (vs.snout length 40-51% of length in Z. asmati, Z. pierrei, Z. syhadrensis, and Z. teraiensis); middorsal line is narrow and distinct (vs.distinct but very wide all the way from snout to vent in Z. pierrei, absent in Z. syhadrensis, present or absent in Z. teraiensis); two red spots found on the middorsal line and additional two red spots on each of the forelimbs are usually present in males (none of the other species, except for Z. asmati, has these spots).
Formal multivariate analyses supported the inference based on the comparisons above: DFA of 14 morphometric traits identified Z. dhaka sp.nov.as clearly different from the other species (Fig 3A).Likewise, PCA showed that while Z. dhaka sp.nov.exhibited intermediate values on the first PCA-axis that were not significantly different from Z. asmati and Z. pierrei (Tukey's HSD, p > 0.05), it was clearly distinct from all other species along the second axis (Tukey's HSD, p < 0.05; Fig 3B).Close inspection of the matrix loadings revealed that the first axis was a general size factor (S5 Table) with all traits loading strongly and positively, and the second axis captured shape differences mainly attributable to effects of traits HAL and IOD, but to some degree also to those of TL, MBE, EL and EN (S5 Table ).
Coloration in life.Basic color of dorsal surface olive green with a few dark irregular spots fused into transverse bands; spots are formed by tubercle arrangement on dorsal side; sides of the body marbled brown.A narrow mid-dorsal line present from vent up to the posterior of eyes.Two red spots are found on the mid-dorsal line.Two additional red spots are present on each of the forelimbs.Toe webbings faintly marbled.Forelimbs with dark transverse bands; hind limbs with pigmented spots on thighs and stripes on lower legs.Ventral surface immaculate.
Coloration in preservative.In preservative, general color pattern remained unaffected, but colors faded.Olive green surfaces became whitish brown.Mid-dorsal line and forelimbs lost their red spots.Variation in paratopotypes.Accession numbers and morphometric variability are shown in Table 1 and S4 Table.One specimen lacked a mid-dorsal line (male MZH-3372).Female (MZH-3373) had no vocal sac markings and lacked red spots.In regard to coloration in life and preservative, the descriptions of the holotype apply: there was little variation among individuals in the type series.

Measurements (in mm
Distribution.Zakerana dhaka is thus far known only from Dhaka, Bangladesh, from the type locality and Curzon Hall (23°43'34.67"N,90°24'4.91"E)situated 5.6 km from the type locality (Fig 1).However, the species also seems to be present in areas adjacent to Dhaka city, such as Mymenshing district (Fig 1) from where two unidentified specimens with haplotypes similar to that of the new species have been reported [38,57].In comparison, the map (Fig 1) also shows the localities from where the species Z. asmati has reported.
Natural history.The new species was observed only at night in the rain.At the type locality, specimens were found in or around temporary pools located in a grass field.In the Curzon Hall site, the species was found in an unused sewage canal of a building.Hence, the new species seems to be resistant to urban disturbance by occupying small habitat patches in different places that are subject to human commotion.
Advertisement call.Advertisement calls from one male (calling on land) were recorded on June 11, 2012, in Sher-e-Bangla Nagar (Dhaka, Bangladesh) at 22:35 in air temperature of 23°C.They consisted of short chirping notes of variable duration (Fig 6).These note duration lasted between 61 to 144 ms (x = 104.2±22.5, n = 192) and were emitted singly or repeated at irregular intervals, or in short series of 2 to 16 notes (x = 4.8 ± 2.97, n = 37).Notes were arranged in series, and inter-note intervals ranged from 18 to 50 ms (x = 32.4 ± 6.7, n = 140), with intervals between two series ranging from 113 to 350 ms (x = 189.2± 51.5, n = 29).The number of pulses per note varied from 8 to 20 (x = 14.5 ± 3.2, n = 190) and the average pulse repetition rate was 144.5 ± 4.2 pulses/s (n = 8).The average note repetition rate was 5.7 notes/s (range: 4.6-7.7 notes/s).The call frequency ranged from 2500 to 4700 Hz with a dominant frequency lying between 2600 and 3800 Hz (Fig 6).
Molecular phylogenetic positioning of the new species.Maximum-likelihood and Bayesian posterior probability methods resulted in similar phylogenetic trees with a strongly supported clade for Z. dhaka (Fig 7).The new species also showed high genetic divergence from all other congeners of the genus Zakerana.The sequence divergences between Z. dhaka sp.nov.and other species were significant, ranging from 5-20.1% for 12S rRNA, and from 3.1-17.3%for 16S rRNA (Table 3).The intraspecific divergence was low for both genes.A sister relationship between Z. dhaka sp.nov.and Z. asmati was supported by robust bootstrap values and high posterior probabilities ( Fig 7).
Conservation.During the course of this study we observed the destruction of species habitats resulting from construction of buildings and roads.Additionally, we observed water pollution, at least partially related to garbage dumping by the Sher-e-Bangla Agricultural University and Medical Hospital.In particular, expansion of human settlements and road constructions are increasing in Dhaka, and may cause risks to the new species.Further notes on conservation status would require a more through censuses to document the actual distribution range of this species outside of Dhaka city.

Discussion
Most species in the genera Zakerana and Fejervarya are morphologically very similar to one another and until recently, were considered to form a single genus "Fejervarya" ( [16,17,58].However, evidence from morphological and genetic analyses indicates the existence of two genera [16,17,[59][60][61].The new species described in this study belongs to the genus Zakerana, and can be distinguished from all other known species in this genus on the basis of morphological and genetic characteristics.Although sequence divergence as such has been considered to be a reliable marker for amphibian species identification (e.g., [62][63][64]), morphological comparisons were also performed to support the genetic inference.However, consistent with earlier reports [34,65,66], the morphological divergence among species was mainly restricted to differences in body proportions.Hence, this together with the minor interest towards amphibian biology in Bangladesh, may explain why Z. dhaka has gone unrecognized as a distinct species in the middle of a mega city.
Although the genetic divergence between Z. dhaka and Z. asmati was not very high (5% for 12S rRNA, and 3.1% for 16S rRNA), it exceeded the conventional threshold value of 3% considered to be indicative of species level diverge in amphibians (e.g.[67][68][69][70]).In fact, low genetic divergence in conserved 16S rRNA gene sequence; comparable to the present study, have also been reported in many other anuran species which show distinct morphological and bioacoustic differences (e.g.[71][72][73][74][75][76][77][78]).Apart from the genetic differentiation, the fact that we found clear-cut morphometric (both quantitative and qualitative) and acoustic divergence between Z. dhaka and Z. asmati strongly suggest that they are distinct species.However, given that acoustic analyses were based on very limited material (calls of one Z.dhaka male), the strongest evidence for the species status of Z. dhaka comes from genetic and morphological data.
Interestingly, haplotype similar to that of the new species (Z.dhaka) described here has been earlier reported by Islam et al. [38], and Hasan et al. [57] from Mymensingh of Bangladesh (Figs 1 and 8).Both Islam et al. [38] and Hasan et al. [57] identified this haplotype as "medium type Fejervarya".Although Islam et al. [38] performed phylogenetic and morphological comparisons-and even crossing experiments among four haplotypes (viz.large, mangrove, medium, and small types [38])-to other species from the genera Fejervarya and Zakerana, these comparisons did not include the sister species Z. asmati.Islam et al. [38] ended up designating this haplotype (GenBank accession: AB372011; Fig 8 ) as "Bangladesh medium type", and suggesting that only further morphological, acoustic, and genetic investigations can confirm its species status.Later on, Hasan et al. [57] reported that this haplotype ("Fejervarya sp.medium") could be a new species, though no morphological or bioacoustics analyses were reported.Our detailed phylogenetic, morphological and bioacoustics analyses bring clarity to this debate, suggesting that the haplotype first described by Islam et al. [38] is in fact a new species, Z. dhaka.
Amphibians are now recognized to be the most threatened class of vertebrates [79].Habitat loss has been identified as one of the major threats for 45% of rapidly declining amphibian species around the world [80].Nevertheless, in the midst of these declines, the number of new amphibian species described during the past decade has been increasing (e.g.[81,82]), even from areas subject to a long history of studies in amphibian biology and systematics (e.g.[14,15]).Unfortunately, there has been a general lack of interest in amphibian taxonomy, systematics and diversity in Bangladesh during the entire 20 th century, possibly due to political instability, which has not nurtured interest in these types of activities.In fact, the last new amphibian species from Dhaka before the one described in this paper was reported almost 150 years ago [83].To meet the demands of a large and growing human population, 2.1% of the country's forests are disappearing annually [84], and more than half of the green areas and water bodies of Dhaka city have disappeared within the past 50 years [85,86].Although the Zakerana species may be tolerant towards urbanization, it remains unclear how many amphibian species and populations have been lost due to these developments.If the present study is indicative of the unrecognized diversity in Bangladesh, future efforts focused on amphibian systematics and taxonomy in this country are likely to lead to the discovery of many new species, as also indicated by recent descriptions of two other new species from this country [39,40].

Fig 6 .
Fig 6.Advertisement call of Zakerana dhaka sp.nov.(A) Spectrogram and (B) waveform of an advertisement call.(C) A detailed image of the waveform of the second note from (B). doi:10.1371/journal.pone.0149597.g006

Table 1 .
Summary of quantitative and qualitative diagnostic characters in Zakerana dhaka sp.nov.and its closest congeners.Morphological ratios given as mean (± S.D.)

Table 2 .
A comparison of Zakerana dhaka sp.nov.with its congeners found in south Asia (except Bangladesh).