Abstract
The water frogs of western Tuscany (Italy) belong to the L-E hybridogenetic system, and comprise one parental species and its hybrid. A stringent morphological approach for discriminating the Italian hybrids from non-hybrids has yet to be established. In this work, using the molecular marker RrS1, we have determined the hybrid versus non-hybrid status of 43 water frogs collected from two sampling sites (“Fiume Morto Vecchio” and “Padule di Bientina”). At “Fiume Morto Vecchio” we determined 25 non-hybrids and nine hybrids and in “Padule di Bientina” we determined eight hybrids and one non-hybrid individual. All individuals of these two frog populations were analyzed morphologically. We used the derived residuals from regression analysis of all normally distributed morphological parameters on callus internus length (snout to vent length, tibia length, head width, distance nostril eye, first toe length and body mass) to build sex independent variables in discriminant analysis providing a valid contribution to morphologically distinguish hybrids from non-hybrid green frogs in Italy.
Introduction
Italian water frogs belong to the L-E hybridogenetic system, comprising one parental species and its hybrid arising from hybridogenetic reproduction (reviewed by Graf and Polls Pelaz, 1989). The taxonomic status of Italian green frogs has been the subject of several studies. Different authors (Uzzell and Hotz, 1979; Hotz and Bruno, 1980; Uzzell, 1983; Günther and Plötner, 1994; Santucci, Nascetti and Bullini, 1996; Santucci et al., 2000), based on morphological, genetic, immunological and bioacoustic data, have recognized two synklepta (one parental species coexists with its hybrid) of green frogs in Italy: the first one in the Po Valley (lessonae-esculentus) and the second one in peninsular Italy, including the islands (bergeri-hispanicus). Crochet and Dubois (2004), Lanza et al. (2007), Canestrelli and Nascetti (2008) distinguished these two lineages lessonae-like (lessonae and bergeri) as subspecies of Pelophylax lessonae, naming them P. (= Rana) lessonae lessonae and P. (= Rana) lessonae bergeri. Canestrelli and Nascetti (2008) also proposed assigning the status of separate subspecies to the Sicilian pool frogs.
Delimiting valid morphological characteristics to discriminate parental species and hybrid forms is, however, rather challenging. The green frog complex is undoubtedly very variable from a morphological and a genetic point of view (Kierzkowski et al., 2011; Hauswaldt et al., 2012). Recent papers have shown that the L (= lessonae) haplotype has a greater influence on morphology than the R (= ridibundus) haplotype, all the hybrid individuals being morphologically more similar to P. lessonae than to P. ridibundus (Kierzkowski et al., 2011, 2013). Interestingly, it has also been found that environmental parameters affect the genotype composition of all-hybrid populations less than populations of the parental species (Jakob et al., 2010). Specific Mate Recognition System (SMRS) analysis on several species of green frogs has underlined that call and morphology do not distinguish with certainty parental and hybrid species. This is due to the high variability and the large overlap of variables between pure and hybrid forms (Lode and Pagano, 2000).
Therefore, and to furthermore contribute to the understanding of the hybridogenetic LL-LR system variability pattern, here we analyze the genetic composition and morphological variability of individuals from two Italian water frog populations, in order to identify a set of morphological variables that could help to discriminate between hybrid and parental forms.
Materials and methods
Sampling sites
Green frogs were sampled in two natural areas of central western Tuscany: “Fiume MortoVecchio” (FMV: 43°44′51.12″N; 10°16′44.45″E) in the Regional Park of Migliarino San Rossore Massaciuccoli, and “Padule di Bientina” (PB: 43°48′34.71″N; 10°36′10.54″E), in the SIR (Sito Importanza Regionale, Regione Toscana) of the same name.
FMV is an abandoned river, whose mouth was naturally closed some centuries ago. It now has limited water exchange, natural river vegetation and is registered as a Natural Reserve known as “Paduletto”. PB is marshland, close to a hilly area, of relatively small surface area, characterized by typical wetland vegetation, and is among the few remnants of marshland in internal Tuscany.
Methodology
We used two different methodological approaches to distinguish between the parental species and the hybrids at two sites in western Tuscany, central Italy.
Measurements
Frogs were captured with fishing nets and with a fishing rod, measured (as in Balletto et al., 1986) and photographed. For all the individuals the following parameters were recorded, according to Hotz and Uzzell (1982), Gubányi and Korsós (1992), Günther and Plötner (1994): tibia length, and callus internus length (on the first toe of the hind limb foot, left one), which represent the highest diagnostic features for water frog species (Günther and Plötner, 1994), snout-vent length (SVL), first toe length, distance between nostrils and the posterior margin of the eye (DNE), head width and body mass. We furthermore considered overall hind limb foot length, recorded shape of the callus internus, pyriform or not pyriform, and the colour of vocal sacs (white, light grey). Vocal sac colour was recorded using Adobe Photoshop.CS 8.0 (2003), measuring the amount of sampled white colour (with the Magic Wand Tool). Body measurements were taken with an analogical calliper (± 0.02 mm accuracy) and body mass with an electronic balance (± 0.1 g accuracy). Furthermore, we cut 5 mm of phalanx of the hind foot fourth finger for molecular analyses. Animals were then released at the capture point within 24 hours. We followed the guidelines for animal care established by the University of Pisa, and according to Regional Law no 56, 6 April 2000, and had permits issued by Parco Migliarino San Rossore Massaciuccoli (prot. 780/7-2-1, 2012 to MALZ).
Molecular analyses
Genomic DNA was prepared from somatic tissues using the DNeasy Tissue Kit (Qiagen). For Southern blot analyses, fragments of completely StuI-digested genomic DNA of all specimens (under the conditions recommended by the manufacturer) were separated electrophoretically and transferred to Hybond-N filters (Amersham). Southern hybridization was carried out, using the digoxigenin (DIG)-labeled pRr300StuI probe as described by Ragghianti et al. (1995). After hybridization, the filters were washed twice in 2xSSC, 0.1% SDS at room temperature and chemoluminescent detection was performed using the DIG-DNA luminescent detection kit (Roche).
Biometrical analyses
We used continuous and normally distributed variables (e.g., SVL) as well as discrete (e.g. callus internus shape, as pyriform – not pyriform) variables. We first analyzed those body features considered important according to Hotz and Uzzell (1982), Gubányi and Korsós (1992), Günther and Plötner (1994), controlling for sex and taxonomic status.
To further investigate relationships between body features and taxonomic position, a non-parametric bivariate correlation (Spearman’s ρ) analysis was carried out on all the body size characteristics considered, selecting those significant at or lower. Given these bivariate relationships, we plotted the two most correlated variables with taxonomic position to obtain a visual perception of the pattern. We then performed a multiway ANCOVA analysis in order to investigate pattern properties and differences in i) sex, and ii) taxonomic status as regards head width and we controlled for covariation of body mass, looking for any possible combination suitable for recognition of hybrid vs not hybrid forms. The original variables were log-transformed and checked again for normality.
Differently from previous classic papers on green frog morphology, we did not build standard indices, but we calculated the derived residuals from regression analysis on callus internus, considered to be among the most important morphological features to discriminate ridibundus, esculentus and lessonae. Discriminant Analysis was carried out on all residuals without sex effect. We tested the null hypothesis of equal covariance matrices; we used eigenvalue > 1. We also performed an assignment test to corroborate the discrimination test. SPSS 13.0 statistical software was used for the statistical analyses.
Results
Molecular approach results
In the present study we used the molecular marker RrS1 to reliably distinguish between the parental species and the hybrids. The autoradiographic pattern of hybrids was characterized by a ladder of hybridizing bands, regularly spaced at a distance of about 100 bp, beginning with the 200 bp (fig. 1, lanes 1, 3, 5, 6, and 8). The pattern of non-hybrids had a few weak bands starting from 800 bp (fig. 1, lanes 2, 4, 7, and 9). Using the above described molecular approach, we determined the taxonomic status of the 43 water frogs collected from the two sampling stations. In FMV we determined 25 non-hybrids and nine hybrids, whereas in PB we determined eight hybrids and one non-hybrid.
Southern blot hybridization of StuI-digested genomic DNA with the DIG-labelled pRr300StuI probe; lanes 1, 3, 5, 6, hybrids and lanes 2, 4, 7, non-hybrids from FMV; lane 8, hybrid and lane 9, non-hybrid from PB. Molecular sizes are expressed in base pairs.
Citation: Amphibia-Reptilia 35, 1 (2014) ; 10.1163/15685381-00002931
Biometric analyses
In table 1 we report body measurements (SVL, first toe length, callus internus, tibia length, head width, DNE and body mass) of Italian water frogs, divided into Italian hybrids and Italian non-hybrids, and sex. Sexual differences from linear untransformed parameters were found only for tibia length (two-way ANCOVA, , 1 df, ), while taxonomic position (e.g. hybrid vs non-hybrid) differed only for callus internus (, 1 df, ) and DNE (, 1 df, ). The interaction sex ∗ taxon did not differ in any of the considered parameters. All parameters varied according to body mass (all of them with P value < 0.0001).
Morphometric measurements of Italian hybrid vs non-hybrid frogs. Each column shows sample size (n), range (mm or mg) and mean ± SD.
Body size features that significantly correlated (bivariate correlation, Spearman’s ρ) with taxonomic position (hybrid = 1, not-hybrid = 2) were: SVL, body mass, hind foot length, head width, vocal sac colour (grey = 1; white = 2), callus internus shape and sex. The highest correlation values were i) taxon/vocal sac colouration (, , ), ii) body mass (, , ), and iii) head width (, ). SVL, hind foot length, callus internus shape and sex had significant, but lower, correlation values (results not shown).
ANCOVA of head width-body mass (fig. 2) was significant (ANCOVA, , 6 df, ) overall, with no difference due to sex (, ), or to taxon (, ), nor for the interactions taxon ∗ sex, taxon ∗ body mass or sex ∗ body mass (, ; , ; , , respectively). Body mass varied significantly (, 1 df, ).
Head width-body mass relationship in hybrid versus non-hybrid Italian water frogs (hybrid: dark grey circles; non-hybrid: light grey circles).
Citation: Amphibia-Reptilia 35, 1 (2014) ; 10.1163/15685381-00002931
Regression analysis to derive standardized residuals was significant (ANOVAregression, , 6 df, ). In Discriminant Analysis, the null hypothesis of equal covariance matrices was rejected (BoxM = 50.666, , ), and first function (Eigenvalue = 1.009) was used (Canonical Correlation = 0.709; Wilk’s Lambda = 0.498, , 6 df, ; functions at group centroids, hybrid = 1.308, non-hybrid = −0.732). In addition, the null hypothesis of equal population covariance matrices of canonical discriminant functions was rejected (BoxM test = 5.313, , ), and Classification Function Coefficients (see table 2) on residuals of selected variables showed a strong, significant, separation between non-hybrid and hybrid frogs (fig. 3).
Classification function coefficients of Italian hybrid vs non-hybrid frogs. Variables are residuals of the regression on Callus internus as an independent variable.
Furthermore, DA correctly assigned 78.6% of hybrid frogs and 96% of non-hybrid frogs (table 3).
Discriminant histograms of Italian hybrid and Italian non-hybrid green frogs.
Citation: Amphibia-Reptilia 35, 1 (2014) ; 10.1163/15685381-00002931
Assignment test (predicted group membership).
Discussion
A firm assignment of the taxonomic position of Italian and European green frogs using external morphology has never been provided, as far as we are aware. The difficulty of establishing a close correspondence between genotypic composition and external morphology of green frogs is likely due to the high morphological variability of these species (Hotz and Bruno, 1980; Gubányi and Korsós, 1992; Günther and Plötner, 1994; Lode and Pagano, 2000). A number of papers have focused on the identification and characterization of genetic variability, DNA contents, and genomic composition of Rana/Pelophylax hybrids and their parental forms (Uzzell and Hotz, 1979; Ragghianti et al., 1995, 2007; Bucci et al., 2000; Ogielska, Kierzkowski and Rybacki, 2004; Kierzkowski et al., 2011, 2013; Marracci et al., 2011). Also, studies that address ecological and habitat features have provided interesting results. Environmental correlates play a key role in the morphological variability of green frogs: Jakob, Arioli and Reyer (2010), investigating the role of pond size and disturbance rate on the hybridogenetic occurrence of diploid vs triploid frogs among all the hybrid frogs in Sweden, found a predominance of LLR in small isolated ponds and of LRR in large wetland ponds. Behavioural contexts gave, on the contrary, contrasting results. Lode and Pagano (2000) confirmed that call characteristics and morphology in hybridogenetic frogs were significantly different among species, but also greatly overlapped among individuals of separate taxa, thus making recognition quite difficult. On the other hand, it has been shown that climate change, over long time series, can affect morphology (e.g., body size) of green frogs. Males of parental species significantly increased their body size, while hybridogenetic form is unrelated to climatic stimulus (Tryjanowski et al., 2006). The overall body structure of green frogs, as described by Nauwelaerts, Ramsay and Aerts (2007), shows neither direct evidence for a trade-off between jumping and swimming performance nor for a coupled optimization. Therefore, anuran shape is conservative and scales mostly isometrically.
Within this scenario, the identity or, perhaps, the correspondence between body features and genotypic structure of diploid-triploid and hybridogenetic frogs is still difficult to assess if plotting and testing linear measures or even unmodified ratios. On the contrary, positive results are obtained when morphological features are extracted with multivariate methods, as with Discriminant Analysis. In addition, the assignment of genetic framework to morphological system was actually performed, despite the number of sampled frogs was limited. Selected parameters (SVL, callus internus, tibia length, head width, DNE, first toe length and body mass) made it possible to distinguish between Italian hybrids from Italian non-hybrids in the studied areas of western Tuscany. Furthermore, we are confident that the tool we propose could also be tested and applied to water frog LE population system of different geographic areas. The tool is suitable as a rapid and cost-effective approach for identifying the hybrid or the not hybrid status of water frogs and may be useful in deciding whether or not perform additional genetic analyses.
Acknowledgements
We are grateful to G. De Matienzo and M. Fabbri for technical assistance. This work was supported by Pisa University grants. Parco Regionale Migliarino San Rossore Massaciuccoli provided visas to enter protected areas. A special thanks to the fisherman G. Spadoni for his help in capturing frogs. D. Buckley and an anonymous referee commented and highly improved this manuscript. Chris Powell (International House, Pisa) revised the English form and style.
References
Balletto E., Cherchi M.A., Salvidio S., Lattes A., Malacrida A., Gasperi G., Doria G. (1986): Area effect in South Western European green frogs. Boll. Zool. 53: 97-109.
Bucci S., Ragghianti M., Guerrini F., Cerrini V., Mancino G., Morosi A., Mossone M., Pascolini R. (2000): Negative environmental factors and biodiversity: the case of the hybridogenetic green frog system from Lake Trasimeno. Ital. J. Zool. 67: 365-370.
Canestrelli D., Nascetti G. (2008): Phylogeography of the pool frog Rana (Pelophylax) lessonae in Italian peninsula and Sicily: multiple refugia, glacial expansion and nuclear-mitochondrial discordance. J. Biogeogr. 35: 1923-1936.
Crochet P.-A., Dubois A. (2004): Recent changes in the taxonomy of European amphibians and reptiles. In: Atlas of Amphibians and Reptiles in Europe, p. 495-516. Gasc J.-P., Cabela A., Crnobrnja-Isailovic J., Dolmen D., Grossenbacher K., Haffener P., Lescure J., Martens H., Martìnez Rica J.P., Maurin H., Oliveira M.E., Sofianidou T.S., Veith M., Zuiderwijk A., Eds, Reprint edition: Muséum National d’Historie Naturelle, Paris.
Graf J.D., Polls-Pelaz M. (1989): Evolutionary genetics of the Rana esculenta complex. In: Evolution and Ecology of Unisexual Vertebrates, p. 289-302. Dawley R.M., Bogart J.P., Eds, Albany.
Gubányi A., Korsós Z. (1992): Morphological analysis of two Hungarian water frog (Rana lessonae-esculenta) populations. Amphibia-Reptilia 13: 235-243.
Günther R., Plötner J. (1994): Morphometric, enzymological and bioacoustic studies in Italian water frogs (Amphibia, Ranidae). Zool. Pol. 39 (3-4): 387-415.
Hauswaldt J.S., Höer M., Ogielska M., Christiansen D.G., Dziewulska-Szwajkowska D., Czernicka E., Vences M. (2012): A simplified molecular method for distinguishing among species and ploidy levels in European water frogs (Pelophylax). Mol. Ecol. Resour. 2: 797-805.
Hotz H., Bruno S. (1980): Il problema delle rane verdi e l’Italia (Amphibia, Salentia). Rend. Acc. Naz. Sc. dei XL Ser. 4, 98: 49-112.
Hotz H., Uzzell T. (1982): Biochemical detected sympatry of two water frog species: Two different cases in the Adriatic Balkans (Amphibia: Ranidae). P. Acad. Nat. Sci. Phila. 134: 50-79.
Jakob C., Arioli M., Reyer H.-U. (2010): Ploidy composition in all-hybrid frog populations in relation to ecological conditions. Evol. Ecol. Res. 12: 633-652.
Kierzkowski P., Kosiba P., Rybacki M., Socha M., Ogielska M. (2013): Genome dosage effect and colouration features in hybridogenetic water frogs of the Pelophylax esculentus complex. Amphibia-Reptilia 34: 493-504.
Kierzkowski P., Pasko L., Rybacki M., Socha M., Ogielska M. (2011): Genome dosage effect and hybrid morphology – the case of the hybridogenetic water frogs of the Pelophylax esculentus complex. Ann. Zool. Fenn. 48: 56-66.
Lanza B., Andreone F., Bologna M.A., Corti C., Razzetti E., Eds (2007): Amphibia. Fauna d’Italia. Ministero dell’Ambiente e della Tutela del Territorio. Calderini.
Lode T., Pagano A. (2000): Variations in call and morphology in male water frogs: taxonomic and evolutionary implications. Compt. Rendus Acad. Sci. Ser. III-Sci Vie-Life Sci. 323: 995-1001. DOI:10.1016/S0764-4469(00)01245-2
Marracci S., Michelotti V., Guex G.-D., Hotz H., Uzzell T., Ragghianti M. (2011): RrS1-like sequences of frogs from central Europe and around the Aegean Sea: chromosomal organization, evolution, possible function. J. Mol. Evol. 72: 368-382.
Nauwelaerts S., Ramsay J., Aerts P. (2007): Morphological correlates of aquatic and terrestrial locomotion in a semi-aquatic frog, Rana esculenta: no evidence for a design conflict. J. Anat. 210: 304-317. DOI:10.1111/j.1469-7580.2007.00691.x
Ogielska M., Kierzkowski P., Rybacki M. (2004): DNA content and genome composition of diploid and triploid water frogs belonging to the Rana esculenta complex (Amphibia, Anura). Can. J. Zool. 82: 1894-1901. DOI:10.1139/Z04-188
Ragghianti M., Bucci S., Marracci S., Casola C., Mancino G., Hotz H., Guex G.-D., Plötner J., Uzzell T. (2007): Gametogenesis of intergroup hybrids of hemiclonal frogs. Genet. Res. 89: 39-45.
Ragghianti M., Guerrini F., Bucci S., Mancino G., Hotz H., Uzzell T., Guex G.-D. (1995): Molecular characterization of a centromeric satellite DNA in the hemiclonal hybrid frog Rana esculenta and its parental species. Chromosome Res. 33: 497-506.
Santucci F., Andreani P., Nascetti G., Bullini L. (2000): Genetic identification and geographic distribution of water frogs Rana lessonae and Rana esculenta in Italy, Sicily and Corsica (Amphibia, Anura). In: Atti I Congresso Nazionale della Societas Herpetologica Italica (Torino 2-6 ottobre 1996), p. 359-367. Giacoma, C., Ed., Museo Regionale di Scienze Naturali, Torino.
Santucci F., Nascetti G., Bullini L. (1996): Hybrid zones between two genetically differentiated forms of the pond frog Rana lessonae in southern Italy. J. Evol. Biol. 9: 429-450.
Tryjanowski P., Sparks T., Rybacki M., Berger L. (2006): Is body size of the water frog Rana esculenta complex responding to climate change? Naturwissenschaften 93: 110-113. DOI:10.1007/s00114-006-0085-2
Uzzell T. (1983): An immunological survey of Italian water frogs (Salientia: Ranidae). Herpetologica 39: 225-234.
Uzzell T., Hotz H. (1979): Electrophoretic and morphological evidence for two forms of green frogs (Rana esculenta complex) in peninsular Italy (Amphibia, Salentia). Mitt. Zool. Mus. Berlin 55: 13-27.
Footnotes
Associated Editor: Matthias Stöck
