- Email: [email protected]
Freeze tolerance evolution among anurans: Frequency and timing of appearance
Cryobiology 58 (2009) 241–247
Contents lists available at ScienceDirect
Cryobiology journal homepage: www.elsevier.com/locate/ycryo
Freeze tolerance evolution among anurans: Frequency and timing of appearance q Yann Voituron a,*, Hervé Barré b,1, Hans Ramløv c,2, Christophe J. Douady a a
Ecologie des Hydrosystème fluviaux (U.M.R. 5023), Université de Lyon, Université Claude Bernard Lyon1, Bât. Darwin C, 43 bd du 11 novembre 1918, 69622 Villeurbanne Cedex, France Laboratoire de Physiologie Intégrative, Cellulaire et Moléculaire (U.M.R. 5123), Université de Lyon, Université Claude Bernard Lyon 1 (Bât. Dubois), 43 bd du 11 novembre 1918, 69622 Villeurbanne Cedex, France c Department of Life sciences and Chemistry, Roskilde University, Universitetvej 1, P.O. Box 260, DK-4000 Roskilde, Denmark b
a r t i c l e
i n f o
Article history: Received 30 September 2008 Accepted 14 January 2009 Available online 22 January 2009 Keywords: Evolution rate Cold hardiness Frog Rana temporaria Rana dalmatina
a b s t r a c t Despite numerous mechanistic studies on physiological responses supporting freeze tolerance in anurans, few have addressed the evolutionary significance of this trait. We thus investigated the phylogenetic relationships among anuran species whose freeze tolerance has been assessed and in combination with new data on freezing tolerance of two closely related species of the European brown frogs (Rana temporaria and Rana dalmatina). The species we studied exhibited short survival times in frozen state (around 8 h for both species). Phylogenetic analysis suggests that freeze tolerance evolved at least two times among Ranidae and one or two times among Hylidae and never in Bufonidae. Furthermore, in order to assess the timing of divergence of this character we used a relaxed molecular clock created, and found that the most recent separation between a freeze tolerant species and a freeze intolerant species dates from 15.9 ± 7.6 Myr (Rana arvalis and R. temporaria). The comparison between these two species thus represents the best current model to understand freeze tolerance evolution. Addressing the evolution of this trait with such large-scale approaches will not only improve our understanding of cold hardiness strategies, but might also create a framework guiding future comparative studies. Ó 2009 Elsevier Inc. All rights reserved.
Introduction Water is essential for life but frequently becomes lethal in the case of freezing since the structures and activities of macromolecules and biological membranes may be severely damaged by ice formation [28]. Organisms living in temperate, alpine and Polar Regions routinely face sub-zero temperatures at least during winter. Such conditions are especially dangerous for ectotherms that cannot regulate their body temperature. In order to survive this stress, these animals have developed two main survival strategies: freeze tolerance, where freezing of some of the body fluid is endured, and freeze avoidance where the body fluids are maintained in a supercooled state during the cold period. Freeze avoidance can now be divided into two sub-categories that imply either (i) the accumulation of low molecular weight cryoprotectant and/or the production of antifreeze proteins or (ii) passive dehydration induced because
q Statement of funding: This work was supported by a research grant from the C.N.R.S. France for Y.V. and a grant from the Danish Science Research Council (# 21-01-0518) for H.R. * Corresponding author. Fax: +33 4 72 43 11 41. E-mail addresses: [email protected], [email protected] (Y. Voituron), [email protected] (H. Barré), [email protected] (H. Ramløv). 1 Fax: +33 (0)4 72 43 11 72. 2 Fax: +45 46 74 30 11.
0011-2240/$ - see front matter Ó 2009 Elsevier Inc. All rights reserved. doi:10.1016/j.cryobiol.2009.01.001
the water vapor pressure of supercooled water is higher than that of ice at the same temperature [20,33]. These two mechanisms lead to a significant decrease of the crystallization temperature (Tc) that allows the animals to keep their body fluids in a metastable supercooled state. Among ectothermic vertebrates, both strategies can be found such as high supercooling capacities for the hatchling painted turtle (Chrysemys picta) that can survive for weeks at sub-zero temperatures as low as 12 °C [29] and high freeze tolerance in the wood frog (Rana sylvatica) that can tolerate 65% of its body water transformed into ice for at least 2 weeks [41]. Currently, freezing tolerance in vertebrates has been detected in only three species of chelonian, one squamate, five anurans and one caudata [3,52] with a large number of studies focussing on North American anurans. Despite numerous mechanistic studies of freeze tolerance in vertebrates, the evolutionary significance of this physiological trait has been much less studied. While Voituron et al. [54] proposed a theoretical model for the evolution of freeze tolerance and freeze avoidance integrating the costs and benefits of each strategy, other works tackled freeze tolerance evolution by comparing physiological responses to freezing between freeze tolerant and freeze intolerant species [8,12,45] mainly considering the potential link between dehydration tolerance and freeze tolerance. However, these comparisons are weakened by fairly large genetic distances
242
Y. Voituron et al. / Cryobiology 58 (2009) 241–247
between the species studied [15]. As far as we know, only Irwin and Lee [23] took individuals from different populations of closely related species (among the gray-treefrog species complex) in a common garden experiment to assess adaptive variation in freeze tolerance. In this paper, these authors also raised, without investigating it, the crucial question of the frequency of appearance of this phenomenon across time. The purpose of this study was two-fold. As a first step towards understanding the evolutionary history of this physiological trait among vertebrates, we reinvestigated the phylogenetic relationships among anuran species (the most studied group of vertebrates regarding freeze tolerance) whose freeze tolerance has been assessed. We also addressed the tempo of evolution of freeze tolerance by developing a relaxed molecular clock. Furthermore, physiological experiments were developed to analyze the freezing tolerance of two European frog species: the common frog, Rana temporaria Linnaeus, 1758 and the agile frog, Rana dalmatina Bonaparte, 1840. These two species are widely distributed in Eurasia and inhabit an extensive range of habitats—from sea level to nearly 1000 m for R. dalmatina and 2500 m in some areas for R. temporaria [18,27]. The choice of these species was guided by their phylogenetic proximity [49] together with their different winter behavior. While R. dalmatina overwinters preferentially on land [1], R. temporaria is known to overwinter underwater in harsh environments [27,31]. We assumed that such behavioral differences could be linked to different physiological performances regarding freezing tolerance [36].
Materials and methods Phylogenetic analysis Data set and alignment Data sets were assembled in order to optimize both taxonomic and genetic representation. Phylogenetic reconstructions were performed on data composed of 20 of the 22 species that have been tested for freeze response. It lacks available sequences for the freeze intolerant toads Bufo paracnemis [37] and Scaphiopus bombifrons [44]. This data includes mitochondrial fragments of cytochrome b and of the two ribosomal RNA subunits, 12S and 16S. A second dataset was constructed for assessing time divergence between organisms. This data was based on mitochondrial ribosomal RNA 12S and 16S and includes five additional taxa for calibrations purpose. Sequences were extracted from the NCBI genetic database. Data set are available upon request. Phylogenetic reconstruction For both data sets, phylogenetic relationships were investigated under Bayesian and maximum likelihood frameworks. Most likely topologies were computed using PAUP 4.0b1.0 [46] with an heristic search, simple sequence addition, TBR branch-swapping algorithm and under the best fitted model of evolution as selected by AIC criterion implemented in Modeltest 3.6 [32]. Bayesian inferences performed using MrBayes 3.0B4 [21], employed a GTR model of sequence evolution allowing a C distribution of rates (4 categories) and a proportion of invariant sites (I). We explored the tree spaces and estimated posterior probability (PP) distributions using 1000,000 generations, sampled every 100 generations. The numbers of generations to obtain convergence of the likelihood value for each data sets, and thus the level at which the ‘‘burn-in” was set, was empirically measured by plotting likelihood values. Phylogenetic reliabilities were measured using non parametric bootstraping under a maximum likelihood framework and using posterior probabilities.
Molecular clock Ages of divergence were estimated by the Bayesian relaxed molecular clock approach developed by Thorne et al. [48], Kishino et al. [24], and Thorne and Kishino [47]. This method of relaxed molecular clock was chosen for its ability to integrate both molecular and paleontological/biogeographical uncertainty. It measures the variance–covariance structure of the branch length estimates of the phylogenetic tree and allows for the definition of simultaneous calibration intervals. Considering two previous studies that both aimed at providing a molecular timescale for amphibian evolution [34,56] we assumed that the split between Bufonidae and Hylidae occurred prior to the KT boundary and that the split between Ranidae and Bufonidae + Hylidae occurred less than 195 million years ago (mya) but more that 117 mya. The divergences between (1) Rana bedriagae and R. cerigensis prior 3 mya [2]; (2) between R. cretensis and R. bedriagae + R. cerigensi + Rana ridibunda) prior 5 mya [2]; (3) between R. pipien and the western palearctic brown frog prior 19.5 mya (as suggested by the presence of a R. cf temporaria in the early miocene—[4]); (4) between the green and brown frog prior to mya (supported by the presence of a true western palearctic green frog in the lower oligocene—[35]) provided additional calibration points. Computation: Divergence times were inferred via Estbranches and Multidivtime (http://statgen.ncsu.edu/thorne/multidivtime. html) from the most probable topology inferred using our taxonomically most divers data set. First Estbranches estimated branch lengths and their variance–covariance matrix from the data set using the xenop sequences as an outgroup. An F84 + C5 model was used, with maximum likelihood parameters estimated under PAML [55], version 3.15. The rate parameter was 2.89 and the shape parameter of the gamma parameter was 0.22. Second, Multidivtime estimated the mean ages of divergence between taxa, with the associated standard-deviation and 95% credibility intervals. Markov chains were sampled 10,000 times with 100 cycles between each sample, and burn-in after 100,000 cycles. We used the empirical guidelines of Multidivtime documentation to define prior distributions on the substitution rate at the base of the tree, and how this rate is likely to change along nodes (i.e. rttm = 1,56; rttmsd = 1.56; rtrate = 0.147606; rtratesd = 0.147606; brownmean = 0.961538; brownsd = 0.961538; time unit = 100 Myr). Unless otherwise stated in respective sections default parameters were applied. Animals Rana temporaria individuals (n = 37) were terrestrially trapped from late September to mid October around Roskilde, Denmark (latitude 55°0 N; longitude 12°0 E, altitude 20 m). The R. dalmatina frogs (n = 41) were also terrestrially trapped in late September near Lyon, France (latitude 45°500 N; longitude 4°92E, altitude 150 m). The R. temporaria, R. dalmatina used in these experiments showed a mean body mass of 3.49 ± 1.68 g and 5.82 ± 2.083 g, respectively (mean ± SD). Freezing protocol and survival Before the experiments, frogs were removed, weighed and placed in small plastic containers, the bottom of which was covered with moist filter paper, for 5–6 weeks. All preparations for the freezing experiments were carried out in the cold room. The animals were placed in moist rubber foam tubes with a thermocouple in contact with the ventral side of the individuals. To ensure that the frogs froze in a natural position the legs were tucked in against the body and tape was wrapped around the rubber foam
Y. Voituron et al. / Cryobiology 58 (2009) 241–247
tubes to keep the frogs in place. The rubber foam tubes containing animals were individually placed in 50 ml plastic tubes. These tubes were plugged with cotton wool and immersed in a cooling bath (HetoFrig baths filled with an ethylene glycol/water mixture) maintained at 2 °C. The temperature of the thermocouples was logged to a computer every minute by a DigiSense Scanning Thermocouple thermometer (Barnant Co., Barrington IL, USA). Inoculation of the moist filter paper was done with spray freeze at 0.2 °C allowing the inoculation of the frogs slightly below the melting point of their body fluids. The initiation of freezing was detected by an increase in body temperature due to the release of the heat of crystallization. Freezing exposure was timed from the appearance of exotherm, and frogs were sampled after different intervals of freezing ranging from 1 to 24 h. After the freezing exposure, the thermocouple and the rubber foam was carefully removed and the frogs were placed in small plastic boxes with moist filter paper in the bottom, in a dark temperature-controlled room at 3 °C ± 1 °C. The survival of all frogs was assessed daily for the next 14 days. Results Phylogenetic relationships Both bayesian and maximum likelihood approaches converged towards the same topology (Fig. 1). Unsurprisingly, this topology supports a tripartite structure segregating ranoids, hyloids and bufonoids taxa, hyloids and bufonoids being clearly more similar in their gene sequences. Within each of these three clades, relationships appear fairly well resolved especially with bayesian posterior probability (but see [11] for a discussion of the phylogenetic properties of this index). Only two nodes did not receive maximal posterior probabilities and maximum likelihood support ranged
243
from 45 to 100% for all nodes (73–100% when discarding the two nodes that were not recovered with a posterior probability of one). Relationships between hyloids genera are not conclusively resolved nor are those between Pseudacris species. Among Ranoids and more precisely among Rana genera, relationships were more conclusively resolved with a deep split between the western European green frog and remaining species. New world water frogs appeared polyphyletic with R. pipiens related to the European brown frog and the sister species R. septentrionalis and R. catesbeiana related to R. sylvatica. Divergence times Most probable divergence time as inferred by a relaxed molecular clock constrained by seven calibration points are displayed in Fig. 2. With the exception of the in-group root that had upper and lower constraints, estimations on constrained nodes were substantially older than the calibration. For example, the split between R. cerigensis and R. bedriagae set to be older than 3 Myr (due to biogeographic events [2]) and was inferred to be 7.4 Myr old (S.D. = 4.9). Likewise the post K/T split between hyloids and bufonoids that we constrained to occur prior to the K/T boundary was in fact estimated to have occurred in the Early Cretaceous (123.038 ± 0.22). Various sub-sets of our calibration points lead to fairly consistent divergence time estimations (results not shown). According to these estimations, R. temporaria and Rana arvalis diverged during the Miocene near the Langhian/Burdigalian boundary (15.9 ± 7.6 Myr ago). Other clades mixing freeze tolerant and susceptible organisms diverged even earlier. R. sylvatica split from R. septentrionalis/R. catesbeiana during the Eocene towards the Lutetian/Ypresian boundary (48.0 ± 15.0 Myr ago) and Pseudacris species diverged during the upper Cretaceous around 77.7 Myr ago (S.D. = 19.7).
Fig. 1. Most probable phylogenetic relationship among selected Eurasian and American frogs based on bayesian inference and partial 12S, 16S rDNA, valine and cytochrome b. Numbers indicate robustness of the branches in percentage of bayesian bootstrap (above the branch) and in posterior probabilities (below). Branch lengths are proportional to the amount of evolutionary changes between taxa (in number of substitution per site). Asterix indicate freeze tolerant species. Rana temporaria (this study), R. dalmatina [this study], R. arvalis [51], R. pipiens [25], R. septentrionalis [36], R. catesbeiana [26], R. sylvatica [36], R. lessonae [53], R. esculenta [53], R. ridibunda [53], Bufo americanus [40], B. woodhousei [45], B. cognatus [45], B. bufo [50], Pseudacris triseriata [36], P. crucifer [36], P. regilla [10], Hyla versicolor [36], H. chrysoscelis [36], Acris crepitans [22].
244
Y. Voituron et al. / Cryobiology 58 (2009) 241–247
Fig. 2. Chronogram obtained from the 12S + 16S concatenation, with ages inferred from the Bayesian relaxed clock method with six simultaneous calibration time constraints. Note that this is not a global molecular clock tree. Branch lengths are proportional to the posterior divergence times of the nodes subtending them. Horizontal rectangles stand for plus or minus one standard-deviation on divergence ages. White circles indicate the nodes under paleontological/biogeographic constraint. Triangles indicate calibration enforced. Horizontal scale represents time in Mya. Geological time scale: L. Jur., M. Jur., U. Jur, L. Cret. U. Cret., Pal., Eoc., Olig., Mioc. and Pli. are respectively for the Lower, Middle, Upper Jurassic, the Lower and Upper Cretaceous, Paleocene, Eocene, Oligocene, Miocene, and Plio-Pleistocene epochs. Tight dashed lines represent divisions between epochs.
Y. Voituron et al. / Cryobiology 58 (2009) 241–247 Table 1 Survival under freezing in Rana arvalis, Rana temporaria and Rana dalmatina with an air temperature maintained at 2 °C. Duration of freezing (h)
% Survival of Rana temporaria
% Survival of Rana dalmatina
1 5 8 10 24
100 (n = 8) 80 (n = 8) 50 (n = 8) 0 (n = 8) 0 (n = 5)
100 (n = 8) 60 (n = 10) 50 (n = 10) 0 (n = 8) 0 (n = 5)
Freezing survival of R. temporaria and R. dalmatina Upon removal from the cooling chamber, the frogs showed clear evidence of internal freezing such as dark skin, rigid limbs, opaque eyes and lack of responsiveness to mechanical stimuli. Frogs were scored as having survived the freezing if they showed normal positioning movements, reacted towards physical stimulation and were able to show spontaneous movement. No individual temporaria and dalmatina frogs survived more than 10 h freezing (Table 1) which was not sufficient to complete the exotherm. Discussion Evolution of freeze tolerance among anurans Our study presents the first combination of anuran phylogeny with known freeze tolerance, allowing us to infer the evolutionary history of this physiological trait. The present view based on several molecular markers is in general agreement with previously published Neobatrachian phylogenies [14,19]. The present analysis also includes all far north distributed Anuran species probably giving us a nearly complete picture of the evolution of freeze tolerance in anurans. Only the Common tree frog (Hyla arborea), the Green toad (Bufo viridis) in Europe and the western toad (Bufo boreas) in North America, not included in this study, may have developed an ecologically relevant freezing tolerance (mainly because of their superficial terrestrial hibernation sites [43,1]). Based on the phylogenetic distribution of freeze tolerance, it is clear that it has evolved several times in anurans and this at different periods of time. While two apparitions of freezing tolerance is the most plausible scenario among Ranidae, R. arvalis and R. sylvatica being very distantly related species, the situation within Hylidae appears to be more complex. Indeed, because the relationship between Hyla, Acris and Pseudacris species are unsettled, we can hypothesize that freezing tolerance evolved twice, once for Pseudacris species and once for Hyla species while, freeze tolerance was lost in Pseudacris regilla. However, with the present taxonomic sampling it is equally parsimonious to assume that freezing tolerance evolved in the Hyloids’ common ancestor and was subsequently lost in P. regilla and Acris crepitans. As previously stated, only further experimentation will enable the second and less likely scenario to be refuted. However, it must be noticed that the lack of an exhaustive knowledge of the strategies used in all temperate and polar anurans somewhat restraint the breath of our conclusions. Indeed, with the present data, we are only inferring the minimal number of freeze tolerance acquisitions in these taxa. Understanding the evolution of physiological parameters requires comparison between closely related species [16] as is the case for the evolutionary physiology of freeze tolerance [23]. Fig. 2 shows that the most recent separation between a freeze tolerant species and a freeze intolerant species dates back 15.9 ± 7.6 Myr (R. arvalis and R. temporaria). The comparison between these two species thus represents the best model to date
245
to understand the physiological mechanism responsible for the evolution of the freeze tolerance. Our estimation of R. arvalis and R. temporaria separation is fairly different from that of Veith et al. [49]. Indeed their estimate places the R. arvalis and R. temporaria split during the Miocene at about 3 Myr and would have been triggered by the glaciations. Such discrepancy will have to be clarified but it may come from radically different analytical approaches. While our approach assumes a weak autocorrelation in molecular substitution rate between the ancestor and its descendant, Veith et al. [49] assumed a strict molecular clock. Furthermore, since they transposed the global clock defined by Beerli [2] on allozyme distance for water frogs, they assumed that a global clock within Western Paleartic brown frogs exists and that it ticked with the same rhythm in both water and brown frogs. This assumption was based on a non-significant relative rate test that sometimes lacks power in detecting clock irregularities [5]. Finally, it should be also noted that fossil evidence reported by Boehme [4] of a R. cf temporaria in the early miocene (i.e. 19–20 Myr ago) is incompatible with the Veith et al. [49] estimation of the Pelophylax/Rana split between 6.19 and 12.85 Myr ago. In the few papers examining the evolutionary aspects of freeze tolerance in vertebrates, R. pipiens is treated as a ‘‘control” freeze intolerant species in comparison with other highly freeze tolerant species such as R. sylvatica [19,45] which diverged 54.4 ± 15.9 Myr ago (see Fig. 2). The comparison based on partial 12S and 16S rDNA sequences between freeze tolerant/intolerant New World ranidae reveals that R. sylvatica is closer to R. catesbeiana or R. septentrionalis, suggesting that the comparison between R. sylvatica and R. septentrionalis or R. catesbeiana frogs could be more informative than the usual comparison R. sylvatica/R. pipiens. However, because the bull frog is very large, the comparison between the wood frog and the mink frog seems more appropriate for American frogs because it avoids the body mass effect that may affect cold tolerance capacities [6]. Mechanisms of freeze tolerance in anurans Even if the evolution of freeze tolerance among anurans occurred multiple times, the physiological mechanisms developed by these animals have at least some common characteristics. Indeed, every freeze tolerant anuran accumulates low molecular weight carbohydrates (glucose and/or glycerol) as cryoprotectant in response to internal ice formation. These molecules act in a colligative manner to minimize cell volume reduction during the growth of extracellular ice crystals and avoid lethal cellular dehydration [41]. Several studies have demonstrated that tissue carbohydrate levels are a critical determinant of freeze tolerance [7]. As soon as ice formation begins in body extremities, a signal is transmitted to the liver, which stimulates glycogenolysis. Within 2 min, blood glucose levels begin to rise and in just a few hours, glucose is distributed to all organs of the body, with concentrations reaching 150–300 mM, compared with normal values of 1–5 mM [39]. This extreme hyperglycemia is supported by large reserves of liver glycogen and by high hepatic activities of glycogen phosphorylase in the most species (except in Pseudacris triseriata [13]). Because both dehydration and somatic freezing initiate glucose accumulation in amphibians, the cryoprotectant systems may have derived from rudimentary mechanisms of water conservation [42]. Such a hypothesis is supported by recent work demonstrating an important cryoprotective role of urea [9], long known as a balancing osmolyte, in R. sylvatica. However, this discovery together with some freeze-induced up-regulation genes for still unidentified proteins [38] clearly demonstrates our incomplete knowledge of amphibian physiology under freezing. Thus secondary differences will be undoubtedly come to light in the future showing family
246
Y. Voituron et al. / Cryobiology 58 (2009) 241–247
or genus level similarities. For instance, Voituron et al. [53] investigated the freezing tolerance of the Rana esculenta hybridization complex which is characterized by a widespread and abundant natural occurrence of hybrid frogs (review in [17]). The R. esculenta complex, involving the parental species R. ridibunda Pallas 1971 and Rana lessonae Camerano 1882, and the hybridogen R. esculenta Linnaeus 1758, occurs in Central and Eastern Europe. These closely related species differ in their wintering behavior. Whereas R. ridibunda usually hibernates under water, R. lessonae prefers terrestrial wintering sites and R. esculenta can exhibit both wintering behaviors. Even if none of these species can be considered as truly freeze tolerant, the ice mass accumulation rate varied markedly among species (from 0.55 ± 0.12 for Rana lessonae to 1.58 ± 0.32 g ice h1 for R. ridibunda) and leads to the conclusion that physiological responses during freezing vary between these two water frogs. Lower ice accumulation in this system cannot be explained by the accumulation of a cryoprotectant. The mechanism of freeze tolerance developed by water frogs R. lessonae, while not yet determined, is thus different from the one developed by the brown frogs. Closely related but not equal before the cold: R. arvalis, R. temporaria and R. dalmatina The R. arvalis frogs, like some other freeze tolerant frogs, hibernate beneath leaf litter, a few centimeters below the soil surface where subfreezing temperatures are frequently encountered. The well-developed capacity for freeze tolerance (100% survival after 48 h freezing) probably plays a key role for its winter survival [51]. In contrast, R. temporaria and R. dalmatina frogs did not survive more than 8 h of freezing suggesting at least for the latter ones the use of deep overwintering sites allowing highly buffered temperature during winter. Surprisingly, Pasanen and Karhapää [30] argued that R. temporaria exhibited a 100% survival rate after a 24 h freezing trial. This difference may be due to geographical and climatic differences but more probably come from the difference in experimental protocol. Indeed, these authors placed several frogs in plastic containers filled with moss or water and placed them in a freezing cabinet at a temperature of 2 °C [30], without following the frogs’ body temperatures. This protocol did not ensure freezing of the frogs and thus renders conclusions about freezing survival unreliable. In conclusion, this first attempt at understanding the evolutionary history of the freezing tolerance in ectotherms vertebrates reveals multiple evolutionary events at different periods of times, the most recent dating from, no later than 15.9 ± 7.6 Myr (between R. arvalis and R. temporaria). However, several questions arise, especially for Hyloids which will require further investigation to enlighten the evolutionary scenario. Addressing the evolution of freeze tolerance with such a large-scale approach will not only improve our current understanding of cold hardiness strategies, but might also create of a research framework guiding future comparative studies [15]. Acknowledgments We express our gratitude to Thomas Sorensen for assistance with collecting frogs and P. Joly for critically reading the manuscript. This work was supported by a research grant from the C.N.R.S. France for Y.V. and a grant from the Danish Science Research Council (# 21-01-0518) for H.R. The present investigation was carried out according to the ethical principles laid down by the European Convention for the protection of Vertebrate Animals used for Experimental and Scientific purposes (Council of Europe N° 123, Strasbourg, 1985).
References [1] ACEMAV coll, R. Duguet, F. Melki, Les amphibiens de France, Belgique et Luxembourg, éds Biotope, Méze (Fance), 2003. [2] P. Beerli, Genetic isolation and calibration of an average protein clock in Western Palearctic water frog of the Aegean region, Thesis, Zürich, 1994. [3] D.I. Berman, A.N. Leirikh, E.I. Mikhailova, Winter hibernation of the Siberian salamander Hynobius keyserlingi, Journal of Evolutionary Biochemistry and Physiology (1984) 323–327. [4] M. Boehme, The oldest representative of a brown frog (Ranidae) from the early miocene of Germany, Acta Palaeontologica Polonica 46 (2001) 19–124. [5] L. Bromham, D. Penny, A. Rambaut, M.D. Hendy, The power of relative rates tests depends on the data, Journal of Molecular Evolution 50 (2000) 296– 301. [6] D. Claussen, P.A. Zani, Allometry of cooling, supercooling, and freezing in the freeze-tolerant turtle Chrysemys picta, American Journal of Physiology 261 (1991) R626–R632. [7] J.P. Costanzo, R.E. Lee, P.H. Lortz, Glucose concentration regulates freeze tolerance in the wood frog Rana sylvatica, Journal of Experimental Biology 181 (1993) 254–255. [8] J.P. Costanzo, R.E. Lee, P.H. Lortz, Physiological responses of freeze-tolerant and intolerant frogs: clues to evolution of anouran freeze tolerance, American Journal of Physiology 265 (1993) R721–R725. [9] J.P. Costanzo, R.E. Lee, Cryoprotection by urea in a terrestrially hibernating frog, Journal of Experimental Biology 208 (2005) 4079–4089. [10] S.A. Croes, R.E. Thomas, Freeze tolerance and cryoprotectant synthesis of the Pacific Tree Frog Hyla regilla, Copeia (2000) 863–868. [11] C.J. Douady, E.J.P. Douzery, Molecular estimation of eulipotyphlan divergence times and the evolution of ‘‘Insectivora”, Molecular Phylogenetics and Evolution 28 (2003) 285–296. [12] J.R. Edwards, J.L. Jenkins, D.L. Swanson, Seasonal effects of dehydration on glucose mobilization in freeze-tolerant chorus frogs (Pseudacris triseriata) and freeze-intolerant toads (Bufo woodhousei and B. cognatus), Journal of Experimental Zoology A 301 (2004) 521–531. [13] J.R. Edwards, K.L. Koster, D.L. Swanson, Time course for cryoprotectant synthesis in the freeze-tolerant chorus, frog, Pseudacris triseriata, Comparative Biochemistry and Physiology A 125 (2000) 367–375. [14] D.R. Frost, T. Grant, J. Faivovich, R.H. Bain, A. Haas, C.F.B. Haddad, R.O. De Sa, A. Channing, M. Wilkinson, S.C. Donnellan, C.J. Raxworthy, J.A. Campbell, B.L. Blotto, P. Moler, R.C. Drewes, R.A. Nussbaum, J.D. Lynch, D.M. Green, W.C. Wheeler, The amphibian tree of life, Bulletin of the American Museum of Natural History (2006) 297–370. [15] J.T. Garland, A.F. Bennett, E.L. Rezende, Phylogenetic approaches in comparative physiology, Journal of Experimental Biology 208 (2005) 3015– 3035. [16] J.T. Garland, S.C. Adolph, Physiological differentiation of vertebrate populations, Annual Review of Ecology and Systematics 22 (1991) 193–228. [17] J.D. Graf, M. Polls Pelaz, Evolutionary genetics of the Rana esculenta complex, in: R.M. Dawley, J.P. Bogart (Eds.), Evolution and Ecology of Unisexual Vertebrates, New York State Museum, New York, 1989, pp. 289–301. [18] K. Grossenbacher, Frosclurch. Handbuch der reptilien und Amphibien Europas, AULA-Verlag, 2003. [19] D.M. Hillis, T.P. Wilcox, Phylogeny of the New World true frogs (Rana), Molecular Phylogenetics and Evolution 34 (2005) 299–314. [20] M. Holmstrup, M. Bayley, H. Ramlov, Supercool or dehydrate? An experimental analysis of overwintering strategies in small permeable arctic invertebrates, Proceedings of the National Academy of Sciences USA 99 (2002) 5716–5720. [21] J.P. Huelsenbeck, F.R. Ronquist, MRBAYES: bayesian inference of phylogeny, Biometrics 17 (2001) 754–755. [22] J.T. Irwin, J.P. Costanzo, R.E. Lee, Terrestrial hibernation in the northern cricket frog, Acris crepitans, Canadian Journal of Zoology 77 (1999) 1240–1246. [23] J.T. Irwin, R.E. Lee, Geographic variation in energy storage and physiological responses to freezing in the gray treefrogs Hyla versicolor and H. chrysoscelis, Journal of Experimental Biology 206 (2003) 2859–2867. [24] H. Kishino, J.L. Thorne, W.J. Bruno, Performance of a divergence time estimation method under a probabilistic model of rate evolution, Molecular Biology and Evolution 18 (2001) 352–361. [25] J.R. Layne, Postfreeze survival and muscle function in the leopard frog (Rana pipiens) and the wood frog (Rana sylvatica), Journal of Thermal Biology 17 (1992) 121–124. [26] D.P. Lotshaw, Temperature adaptation and effects of thermal acclimation in Rana sylvatica and Rana catesbeiana, Comparative Biochemistry and Physiology 56A (1977) 287–294. [27] G. Matz, D. Weber, Guide des amphibiens er reptiles d’Europe, Delachaux Niestlé SA, Lausanne (Switzerland), 1999. [28] P. Mazur, Freezing of living cells: mechanisms and implications, American Journal of Physiology 247 (1984) C125–C142. [29] G.C. Packard, M.J. Packard, To freeze or not to freeze: adaptations for overwintering by hatchlings of the North American painted turtle, Journal of Experimental Biology 207 (2004) 2897–2906. [30] S. Pasanen, M. Karhapää, Can boreal common frog (Rana temporaria L.) survive in frost?, Annales Zoologici Fennici 34 (1997) 247–250 [31] S. Pasanen, J. Sorjonen, Partial terrestrial wintering in a northern common frog population (Rana temporaria L.), Annales Zoologici Fennici 31 (1994) 275–278.
Y. Voituron et al. / Cryobiology 58 (2009) 241–247 [32] D. Posada, K.A. Crandall, MODELTEST: testing the model of DNA substitution, Bioinformatics 14 (1998) 817–818. [33] H. Ramlov, Aspects of natural cold tolerance in ectothermic animals, Journal of Human Reproduction 15 (2000) 26–46. [34] D. San Mauro, M. Vences, M. Alcobendas, R. Zardoya, A. Meyer, Initial diversification of living amphibians predated the breakup of Pangaea, American Naturalist 165 (2005) 590–599. [35] Z.B. Sanchíz, Salientia, in: P. Wellnhofer (Ed.), Handbook of Palaeoherpetology, Verlag Dr. Friedrich Pfeil, München, 1998, pp. 1–275. [36] W. Schmid, Survival of frogs in low temperature, Science 215 (1982) 697–698. [37] A.A. Steiner, S.O. Petenusci, L.G. Brentegani, L.G.S. Branco, The importance of glucose for the freezing tolerance/intolerance of the anuran amphibians Rana catesbeiana and Bufo paracnemis, Revista Brasileira de Biologia 60 (2000) 321–328. [38] K.B. Storey, Strategies for exploration of freeze responsive gene expression: advances in vertebrate freeze tolerance, Cryobiology 48 (2004) 134–145. [39] K.B. Storey, J.M. Storey, Biochemical adaptation for freezing tolerance in the wood frog, Rana sylvatica, Journal of Comparative Physiology B 155 (1984) 29–36. [40] K.B. Storey, J.M. Storey, Freeze tolerance and intolerance as strategies of winter survival in terrestrially-hibernating amphibians, Comparative Biochemistry and Physiology 83A (1986) 613–617. [41] K.B. Storey, J.M. Storey, Freeze tolerance in animals, Physiological Reviews 68 (1988) 27–84. [42] K.B. Storey, J.M. Storey, Physiology, biochemistry, and molecular biology of vertebrate freeze tolerance: the wood frog, in: J. Fuller (Ed.), Life in the Frozen State, CRC Press, Washington, DC, 2004, pp. 243–274. [43] A.H.P. Stumpel, On hibernation sites in the tree frog Hyla arborea, Amphibia– Reptilia 11 (1990) 304–306. [44] D.L. Swanson, B.M. Graves, Supercooling and freeze intolerance in overwintering juvenile spadefoot toads (Scaphiopus bombifrons), Journal of Herpetology 29 (1995) 280–285.
247
[45] D.L. Swanson, B.M. Graves, K.L. Koster, Freezing tolerance/intolerance and cryoprotectant synthesis in terrestrially overwintering anurans in the great plains, USA, Journal of Comparative Physiology B 166 (1996) 110–119. [46] D.L. Swofford, Phylogenetic analysis using parsimony (*and other methods). PAUP* (2002). [47] J.L. Thorne, H. Kishino, Divergence time and evolutionary rate estimation with multilocus data, Systematic Biology 51 (2002) 689–702. [48] J.L. Thorne, H. Kishino, I.S. Painter, Estimating the rate of evolution of the rate of molecular evolution, Molecular Biology and Evolution 15 (1998) 1647–1657. [49] M. Veith, J. Kosuch, M. Vences, Climatic oscillations triggered post-Messinian speciation of Western Palearctic brown frogs (Amphibia, Ranidae), Molecular Phylogenetics and Evolution 26 (2003) 310–327. [50] Y. Voituron, T. Lengagne, Freezing before mating severely impacts fitness of a capital breeder, Physiological and Biochemical Zoology 81 (2) (2008) 204–210. [51] Y. Voituron, L. Paaschburg, H. Barré, M. Holmstrup, H. Ramløv, Survival and metabolism of Rana arvalis during freezing, Journal of Comparative Physiology B 179 (2009) 223–230. [52] Y. Voituron, M. Eugène, H. Barré, Survival and metabolic responses to freezing by the water frog (Rana ridibunda), Journal of Experimental Zoology 299A (2003) 118–126. [53] Y. Voituron, P. Joly, M. Eugene, H. Barre, Freezing tolerance of the European water frogs: the good, the bad and the ugly, American Journal of Physiology 288 (2005) R1563–R1570. [54] Y. Voituron, N. Mouquet, C. deMazancourt, J. Clobert, To freeze or not to freeze? An evolutionary perspective on the cold hardiness strategies of overwintering ectotherms, American Naturalist 160 (2002) 255–270. [55] Z.H. Yang, PAML: a program package for phylogenetic analysis by maximum likelihood, Computational Applied Bioscience 13 (1997) 555–556. [56] P. Zhang, H. Zhou, Y.Q. Chen, Y.F. Liu, L.H. Qu, Mitogenomic perspectives on the origin and phylogeny of living amphibians, Systematic Biology 54 (2005) 391– 400.
Contents lists available at ScienceDirect
Cryobiology journal homepage: www.elsevier.com/locate/ycryo
Freeze tolerance evolution among anurans: Frequency and timing of appearance q Yann Voituron a,*, Hervé Barré b,1, Hans Ramløv c,2, Christophe J. Douady a a
Ecologie des Hydrosystème fluviaux (U.M.R. 5023), Université de Lyon, Université Claude Bernard Lyon1, Bât. Darwin C, 43 bd du 11 novembre 1918, 69622 Villeurbanne Cedex, France Laboratoire de Physiologie Intégrative, Cellulaire et Moléculaire (U.M.R. 5123), Université de Lyon, Université Claude Bernard Lyon 1 (Bât. Dubois), 43 bd du 11 novembre 1918, 69622 Villeurbanne Cedex, France c Department of Life sciences and Chemistry, Roskilde University, Universitetvej 1, P.O. Box 260, DK-4000 Roskilde, Denmark b
a r t i c l e
i n f o
Article history: Received 30 September 2008 Accepted 14 January 2009 Available online 22 January 2009 Keywords: Evolution rate Cold hardiness Frog Rana temporaria Rana dalmatina
a b s t r a c t Despite numerous mechanistic studies on physiological responses supporting freeze tolerance in anurans, few have addressed the evolutionary significance of this trait. We thus investigated the phylogenetic relationships among anuran species whose freeze tolerance has been assessed and in combination with new data on freezing tolerance of two closely related species of the European brown frogs (Rana temporaria and Rana dalmatina). The species we studied exhibited short survival times in frozen state (around 8 h for both species). Phylogenetic analysis suggests that freeze tolerance evolved at least two times among Ranidae and one or two times among Hylidae and never in Bufonidae. Furthermore, in order to assess the timing of divergence of this character we used a relaxed molecular clock created, and found that the most recent separation between a freeze tolerant species and a freeze intolerant species dates from 15.9 ± 7.6 Myr (Rana arvalis and R. temporaria). The comparison between these two species thus represents the best current model to understand freeze tolerance evolution. Addressing the evolution of this trait with such large-scale approaches will not only improve our understanding of cold hardiness strategies, but might also create a framework guiding future comparative studies. Ó 2009 Elsevier Inc. All rights reserved.
Introduction Water is essential for life but frequently becomes lethal in the case of freezing since the structures and activities of macromolecules and biological membranes may be severely damaged by ice formation [28]. Organisms living in temperate, alpine and Polar Regions routinely face sub-zero temperatures at least during winter. Such conditions are especially dangerous for ectotherms that cannot regulate their body temperature. In order to survive this stress, these animals have developed two main survival strategies: freeze tolerance, where freezing of some of the body fluid is endured, and freeze avoidance where the body fluids are maintained in a supercooled state during the cold period. Freeze avoidance can now be divided into two sub-categories that imply either (i) the accumulation of low molecular weight cryoprotectant and/or the production of antifreeze proteins or (ii) passive dehydration induced because
q Statement of funding: This work was supported by a research grant from the C.N.R.S. France for Y.V. and a grant from the Danish Science Research Council (# 21-01-0518) for H.R. * Corresponding author. Fax: +33 4 72 43 11 41. E-mail addresses: [email protected], [email protected] (Y. Voituron), [email protected] (H. Barré), [email protected] (H. Ramløv). 1 Fax: +33 (0)4 72 43 11 72. 2 Fax: +45 46 74 30 11.
0011-2240/$ - see front matter Ó 2009 Elsevier Inc. All rights reserved. doi:10.1016/j.cryobiol.2009.01.001
the water vapor pressure of supercooled water is higher than that of ice at the same temperature [20,33]. These two mechanisms lead to a significant decrease of the crystallization temperature (Tc) that allows the animals to keep their body fluids in a metastable supercooled state. Among ectothermic vertebrates, both strategies can be found such as high supercooling capacities for the hatchling painted turtle (Chrysemys picta) that can survive for weeks at sub-zero temperatures as low as 12 °C [29] and high freeze tolerance in the wood frog (Rana sylvatica) that can tolerate 65% of its body water transformed into ice for at least 2 weeks [41]. Currently, freezing tolerance in vertebrates has been detected in only three species of chelonian, one squamate, five anurans and one caudata [3,52] with a large number of studies focussing on North American anurans. Despite numerous mechanistic studies of freeze tolerance in vertebrates, the evolutionary significance of this physiological trait has been much less studied. While Voituron et al. [54] proposed a theoretical model for the evolution of freeze tolerance and freeze avoidance integrating the costs and benefits of each strategy, other works tackled freeze tolerance evolution by comparing physiological responses to freezing between freeze tolerant and freeze intolerant species [8,12,45] mainly considering the potential link between dehydration tolerance and freeze tolerance. However, these comparisons are weakened by fairly large genetic distances
242
Y. Voituron et al. / Cryobiology 58 (2009) 241–247
between the species studied [15]. As far as we know, only Irwin and Lee [23] took individuals from different populations of closely related species (among the gray-treefrog species complex) in a common garden experiment to assess adaptive variation in freeze tolerance. In this paper, these authors also raised, without investigating it, the crucial question of the frequency of appearance of this phenomenon across time. The purpose of this study was two-fold. As a first step towards understanding the evolutionary history of this physiological trait among vertebrates, we reinvestigated the phylogenetic relationships among anuran species (the most studied group of vertebrates regarding freeze tolerance) whose freeze tolerance has been assessed. We also addressed the tempo of evolution of freeze tolerance by developing a relaxed molecular clock. Furthermore, physiological experiments were developed to analyze the freezing tolerance of two European frog species: the common frog, Rana temporaria Linnaeus, 1758 and the agile frog, Rana dalmatina Bonaparte, 1840. These two species are widely distributed in Eurasia and inhabit an extensive range of habitats—from sea level to nearly 1000 m for R. dalmatina and 2500 m in some areas for R. temporaria [18,27]. The choice of these species was guided by their phylogenetic proximity [49] together with their different winter behavior. While R. dalmatina overwinters preferentially on land [1], R. temporaria is known to overwinter underwater in harsh environments [27,31]. We assumed that such behavioral differences could be linked to different physiological performances regarding freezing tolerance [36].
Materials and methods Phylogenetic analysis Data set and alignment Data sets were assembled in order to optimize both taxonomic and genetic representation. Phylogenetic reconstructions were performed on data composed of 20 of the 22 species that have been tested for freeze response. It lacks available sequences for the freeze intolerant toads Bufo paracnemis [37] and Scaphiopus bombifrons [44]. This data includes mitochondrial fragments of cytochrome b and of the two ribosomal RNA subunits, 12S and 16S. A second dataset was constructed for assessing time divergence between organisms. This data was based on mitochondrial ribosomal RNA 12S and 16S and includes five additional taxa for calibrations purpose. Sequences were extracted from the NCBI genetic database. Data set are available upon request. Phylogenetic reconstruction For both data sets, phylogenetic relationships were investigated under Bayesian and maximum likelihood frameworks. Most likely topologies were computed using PAUP 4.0b1.0 [46] with an heristic search, simple sequence addition, TBR branch-swapping algorithm and under the best fitted model of evolution as selected by AIC criterion implemented in Modeltest 3.6 [32]. Bayesian inferences performed using MrBayes 3.0B4 [21], employed a GTR model of sequence evolution allowing a C distribution of rates (4 categories) and a proportion of invariant sites (I). We explored the tree spaces and estimated posterior probability (PP) distributions using 1000,000 generations, sampled every 100 generations. The numbers of generations to obtain convergence of the likelihood value for each data sets, and thus the level at which the ‘‘burn-in” was set, was empirically measured by plotting likelihood values. Phylogenetic reliabilities were measured using non parametric bootstraping under a maximum likelihood framework and using posterior probabilities.
Molecular clock Ages of divergence were estimated by the Bayesian relaxed molecular clock approach developed by Thorne et al. [48], Kishino et al. [24], and Thorne and Kishino [47]. This method of relaxed molecular clock was chosen for its ability to integrate both molecular and paleontological/biogeographical uncertainty. It measures the variance–covariance structure of the branch length estimates of the phylogenetic tree and allows for the definition of simultaneous calibration intervals. Considering two previous studies that both aimed at providing a molecular timescale for amphibian evolution [34,56] we assumed that the split between Bufonidae and Hylidae occurred prior to the KT boundary and that the split between Ranidae and Bufonidae + Hylidae occurred less than 195 million years ago (mya) but more that 117 mya. The divergences between (1) Rana bedriagae and R. cerigensis prior 3 mya [2]; (2) between R. cretensis and R. bedriagae + R. cerigensi + Rana ridibunda) prior 5 mya [2]; (3) between R. pipien and the western palearctic brown frog prior 19.5 mya (as suggested by the presence of a R. cf temporaria in the early miocene—[4]); (4) between the green and brown frog prior to mya (supported by the presence of a true western palearctic green frog in the lower oligocene—[35]) provided additional calibration points. Computation: Divergence times were inferred via Estbranches and Multidivtime (http://statgen.ncsu.edu/thorne/multidivtime. html) from the most probable topology inferred using our taxonomically most divers data set. First Estbranches estimated branch lengths and their variance–covariance matrix from the data set using the xenop sequences as an outgroup. An F84 + C5 model was used, with maximum likelihood parameters estimated under PAML [55], version 3.15. The rate parameter was 2.89 and the shape parameter of the gamma parameter was 0.22. Second, Multidivtime estimated the mean ages of divergence between taxa, with the associated standard-deviation and 95% credibility intervals. Markov chains were sampled 10,000 times with 100 cycles between each sample, and burn-in after 100,000 cycles. We used the empirical guidelines of Multidivtime documentation to define prior distributions on the substitution rate at the base of the tree, and how this rate is likely to change along nodes (i.e. rttm = 1,56; rttmsd = 1.56; rtrate = 0.147606; rtratesd = 0.147606; brownmean = 0.961538; brownsd = 0.961538; time unit = 100 Myr). Unless otherwise stated in respective sections default parameters were applied. Animals Rana temporaria individuals (n = 37) were terrestrially trapped from late September to mid October around Roskilde, Denmark (latitude 55°0 N; longitude 12°0 E, altitude 20 m). The R. dalmatina frogs (n = 41) were also terrestrially trapped in late September near Lyon, France (latitude 45°500 N; longitude 4°92E, altitude 150 m). The R. temporaria, R. dalmatina used in these experiments showed a mean body mass of 3.49 ± 1.68 g and 5.82 ± 2.083 g, respectively (mean ± SD). Freezing protocol and survival Before the experiments, frogs were removed, weighed and placed in small plastic containers, the bottom of which was covered with moist filter paper, for 5–6 weeks. All preparations for the freezing experiments were carried out in the cold room. The animals were placed in moist rubber foam tubes with a thermocouple in contact with the ventral side of the individuals. To ensure that the frogs froze in a natural position the legs were tucked in against the body and tape was wrapped around the rubber foam
Y. Voituron et al. / Cryobiology 58 (2009) 241–247
tubes to keep the frogs in place. The rubber foam tubes containing animals were individually placed in 50 ml plastic tubes. These tubes were plugged with cotton wool and immersed in a cooling bath (HetoFrig baths filled with an ethylene glycol/water mixture) maintained at 2 °C. The temperature of the thermocouples was logged to a computer every minute by a DigiSense Scanning Thermocouple thermometer (Barnant Co., Barrington IL, USA). Inoculation of the moist filter paper was done with spray freeze at 0.2 °C allowing the inoculation of the frogs slightly below the melting point of their body fluids. The initiation of freezing was detected by an increase in body temperature due to the release of the heat of crystallization. Freezing exposure was timed from the appearance of exotherm, and frogs were sampled after different intervals of freezing ranging from 1 to 24 h. After the freezing exposure, the thermocouple and the rubber foam was carefully removed and the frogs were placed in small plastic boxes with moist filter paper in the bottom, in a dark temperature-controlled room at 3 °C ± 1 °C. The survival of all frogs was assessed daily for the next 14 days. Results Phylogenetic relationships Both bayesian and maximum likelihood approaches converged towards the same topology (Fig. 1). Unsurprisingly, this topology supports a tripartite structure segregating ranoids, hyloids and bufonoids taxa, hyloids and bufonoids being clearly more similar in their gene sequences. Within each of these three clades, relationships appear fairly well resolved especially with bayesian posterior probability (but see [11] for a discussion of the phylogenetic properties of this index). Only two nodes did not receive maximal posterior probabilities and maximum likelihood support ranged
243
from 45 to 100% for all nodes (73–100% when discarding the two nodes that were not recovered with a posterior probability of one). Relationships between hyloids genera are not conclusively resolved nor are those between Pseudacris species. Among Ranoids and more precisely among Rana genera, relationships were more conclusively resolved with a deep split between the western European green frog and remaining species. New world water frogs appeared polyphyletic with R. pipiens related to the European brown frog and the sister species R. septentrionalis and R. catesbeiana related to R. sylvatica. Divergence times Most probable divergence time as inferred by a relaxed molecular clock constrained by seven calibration points are displayed in Fig. 2. With the exception of the in-group root that had upper and lower constraints, estimations on constrained nodes were substantially older than the calibration. For example, the split between R. cerigensis and R. bedriagae set to be older than 3 Myr (due to biogeographic events [2]) and was inferred to be 7.4 Myr old (S.D. = 4.9). Likewise the post K/T split between hyloids and bufonoids that we constrained to occur prior to the K/T boundary was in fact estimated to have occurred in the Early Cretaceous (123.038 ± 0.22). Various sub-sets of our calibration points lead to fairly consistent divergence time estimations (results not shown). According to these estimations, R. temporaria and Rana arvalis diverged during the Miocene near the Langhian/Burdigalian boundary (15.9 ± 7.6 Myr ago). Other clades mixing freeze tolerant and susceptible organisms diverged even earlier. R. sylvatica split from R. septentrionalis/R. catesbeiana during the Eocene towards the Lutetian/Ypresian boundary (48.0 ± 15.0 Myr ago) and Pseudacris species diverged during the upper Cretaceous around 77.7 Myr ago (S.D. = 19.7).
Fig. 1. Most probable phylogenetic relationship among selected Eurasian and American frogs based on bayesian inference and partial 12S, 16S rDNA, valine and cytochrome b. Numbers indicate robustness of the branches in percentage of bayesian bootstrap (above the branch) and in posterior probabilities (below). Branch lengths are proportional to the amount of evolutionary changes between taxa (in number of substitution per site). Asterix indicate freeze tolerant species. Rana temporaria (this study), R. dalmatina [this study], R. arvalis [51], R. pipiens [25], R. septentrionalis [36], R. catesbeiana [26], R. sylvatica [36], R. lessonae [53], R. esculenta [53], R. ridibunda [53], Bufo americanus [40], B. woodhousei [45], B. cognatus [45], B. bufo [50], Pseudacris triseriata [36], P. crucifer [36], P. regilla [10], Hyla versicolor [36], H. chrysoscelis [36], Acris crepitans [22].
244
Y. Voituron et al. / Cryobiology 58 (2009) 241–247
Fig. 2. Chronogram obtained from the 12S + 16S concatenation, with ages inferred from the Bayesian relaxed clock method with six simultaneous calibration time constraints. Note that this is not a global molecular clock tree. Branch lengths are proportional to the posterior divergence times of the nodes subtending them. Horizontal rectangles stand for plus or minus one standard-deviation on divergence ages. White circles indicate the nodes under paleontological/biogeographic constraint. Triangles indicate calibration enforced. Horizontal scale represents time in Mya. Geological time scale: L. Jur., M. Jur., U. Jur, L. Cret. U. Cret., Pal., Eoc., Olig., Mioc. and Pli. are respectively for the Lower, Middle, Upper Jurassic, the Lower and Upper Cretaceous, Paleocene, Eocene, Oligocene, Miocene, and Plio-Pleistocene epochs. Tight dashed lines represent divisions between epochs.
Y. Voituron et al. / Cryobiology 58 (2009) 241–247 Table 1 Survival under freezing in Rana arvalis, Rana temporaria and Rana dalmatina with an air temperature maintained at 2 °C. Duration of freezing (h)
% Survival of Rana temporaria
% Survival of Rana dalmatina
1 5 8 10 24
100 (n = 8) 80 (n = 8) 50 (n = 8) 0 (n = 8) 0 (n = 5)
100 (n = 8) 60 (n = 10) 50 (n = 10) 0 (n = 8) 0 (n = 5)
Freezing survival of R. temporaria and R. dalmatina Upon removal from the cooling chamber, the frogs showed clear evidence of internal freezing such as dark skin, rigid limbs, opaque eyes and lack of responsiveness to mechanical stimuli. Frogs were scored as having survived the freezing if they showed normal positioning movements, reacted towards physical stimulation and were able to show spontaneous movement. No individual temporaria and dalmatina frogs survived more than 10 h freezing (Table 1) which was not sufficient to complete the exotherm. Discussion Evolution of freeze tolerance among anurans Our study presents the first combination of anuran phylogeny with known freeze tolerance, allowing us to infer the evolutionary history of this physiological trait. The present view based on several molecular markers is in general agreement with previously published Neobatrachian phylogenies [14,19]. The present analysis also includes all far north distributed Anuran species probably giving us a nearly complete picture of the evolution of freeze tolerance in anurans. Only the Common tree frog (Hyla arborea), the Green toad (Bufo viridis) in Europe and the western toad (Bufo boreas) in North America, not included in this study, may have developed an ecologically relevant freezing tolerance (mainly because of their superficial terrestrial hibernation sites [43,1]). Based on the phylogenetic distribution of freeze tolerance, it is clear that it has evolved several times in anurans and this at different periods of time. While two apparitions of freezing tolerance is the most plausible scenario among Ranidae, R. arvalis and R. sylvatica being very distantly related species, the situation within Hylidae appears to be more complex. Indeed, because the relationship between Hyla, Acris and Pseudacris species are unsettled, we can hypothesize that freezing tolerance evolved twice, once for Pseudacris species and once for Hyla species while, freeze tolerance was lost in Pseudacris regilla. However, with the present taxonomic sampling it is equally parsimonious to assume that freezing tolerance evolved in the Hyloids’ common ancestor and was subsequently lost in P. regilla and Acris crepitans. As previously stated, only further experimentation will enable the second and less likely scenario to be refuted. However, it must be noticed that the lack of an exhaustive knowledge of the strategies used in all temperate and polar anurans somewhat restraint the breath of our conclusions. Indeed, with the present data, we are only inferring the minimal number of freeze tolerance acquisitions in these taxa. Understanding the evolution of physiological parameters requires comparison between closely related species [16] as is the case for the evolutionary physiology of freeze tolerance [23]. Fig. 2 shows that the most recent separation between a freeze tolerant species and a freeze intolerant species dates back 15.9 ± 7.6 Myr (R. arvalis and R. temporaria). The comparison between these two species thus represents the best model to date
245
to understand the physiological mechanism responsible for the evolution of the freeze tolerance. Our estimation of R. arvalis and R. temporaria separation is fairly different from that of Veith et al. [49]. Indeed their estimate places the R. arvalis and R. temporaria split during the Miocene at about 3 Myr and would have been triggered by the glaciations. Such discrepancy will have to be clarified but it may come from radically different analytical approaches. While our approach assumes a weak autocorrelation in molecular substitution rate between the ancestor and its descendant, Veith et al. [49] assumed a strict molecular clock. Furthermore, since they transposed the global clock defined by Beerli [2] on allozyme distance for water frogs, they assumed that a global clock within Western Paleartic brown frogs exists and that it ticked with the same rhythm in both water and brown frogs. This assumption was based on a non-significant relative rate test that sometimes lacks power in detecting clock irregularities [5]. Finally, it should be also noted that fossil evidence reported by Boehme [4] of a R. cf temporaria in the early miocene (i.e. 19–20 Myr ago) is incompatible with the Veith et al. [49] estimation of the Pelophylax/Rana split between 6.19 and 12.85 Myr ago. In the few papers examining the evolutionary aspects of freeze tolerance in vertebrates, R. pipiens is treated as a ‘‘control” freeze intolerant species in comparison with other highly freeze tolerant species such as R. sylvatica [19,45] which diverged 54.4 ± 15.9 Myr ago (see Fig. 2). The comparison based on partial 12S and 16S rDNA sequences between freeze tolerant/intolerant New World ranidae reveals that R. sylvatica is closer to R. catesbeiana or R. septentrionalis, suggesting that the comparison between R. sylvatica and R. septentrionalis or R. catesbeiana frogs could be more informative than the usual comparison R. sylvatica/R. pipiens. However, because the bull frog is very large, the comparison between the wood frog and the mink frog seems more appropriate for American frogs because it avoids the body mass effect that may affect cold tolerance capacities [6]. Mechanisms of freeze tolerance in anurans Even if the evolution of freeze tolerance among anurans occurred multiple times, the physiological mechanisms developed by these animals have at least some common characteristics. Indeed, every freeze tolerant anuran accumulates low molecular weight carbohydrates (glucose and/or glycerol) as cryoprotectant in response to internal ice formation. These molecules act in a colligative manner to minimize cell volume reduction during the growth of extracellular ice crystals and avoid lethal cellular dehydration [41]. Several studies have demonstrated that tissue carbohydrate levels are a critical determinant of freeze tolerance [7]. As soon as ice formation begins in body extremities, a signal is transmitted to the liver, which stimulates glycogenolysis. Within 2 min, blood glucose levels begin to rise and in just a few hours, glucose is distributed to all organs of the body, with concentrations reaching 150–300 mM, compared with normal values of 1–5 mM [39]. This extreme hyperglycemia is supported by large reserves of liver glycogen and by high hepatic activities of glycogen phosphorylase in the most species (except in Pseudacris triseriata [13]). Because both dehydration and somatic freezing initiate glucose accumulation in amphibians, the cryoprotectant systems may have derived from rudimentary mechanisms of water conservation [42]. Such a hypothesis is supported by recent work demonstrating an important cryoprotective role of urea [9], long known as a balancing osmolyte, in R. sylvatica. However, this discovery together with some freeze-induced up-regulation genes for still unidentified proteins [38] clearly demonstrates our incomplete knowledge of amphibian physiology under freezing. Thus secondary differences will be undoubtedly come to light in the future showing family
246
Y. Voituron et al. / Cryobiology 58 (2009) 241–247
or genus level similarities. For instance, Voituron et al. [53] investigated the freezing tolerance of the Rana esculenta hybridization complex which is characterized by a widespread and abundant natural occurrence of hybrid frogs (review in [17]). The R. esculenta complex, involving the parental species R. ridibunda Pallas 1971 and Rana lessonae Camerano 1882, and the hybridogen R. esculenta Linnaeus 1758, occurs in Central and Eastern Europe. These closely related species differ in their wintering behavior. Whereas R. ridibunda usually hibernates under water, R. lessonae prefers terrestrial wintering sites and R. esculenta can exhibit both wintering behaviors. Even if none of these species can be considered as truly freeze tolerant, the ice mass accumulation rate varied markedly among species (from 0.55 ± 0.12 for Rana lessonae to 1.58 ± 0.32 g ice h1 for R. ridibunda) and leads to the conclusion that physiological responses during freezing vary between these two water frogs. Lower ice accumulation in this system cannot be explained by the accumulation of a cryoprotectant. The mechanism of freeze tolerance developed by water frogs R. lessonae, while not yet determined, is thus different from the one developed by the brown frogs. Closely related but not equal before the cold: R. arvalis, R. temporaria and R. dalmatina The R. arvalis frogs, like some other freeze tolerant frogs, hibernate beneath leaf litter, a few centimeters below the soil surface where subfreezing temperatures are frequently encountered. The well-developed capacity for freeze tolerance (100% survival after 48 h freezing) probably plays a key role for its winter survival [51]. In contrast, R. temporaria and R. dalmatina frogs did not survive more than 8 h of freezing suggesting at least for the latter ones the use of deep overwintering sites allowing highly buffered temperature during winter. Surprisingly, Pasanen and Karhapää [30] argued that R. temporaria exhibited a 100% survival rate after a 24 h freezing trial. This difference may be due to geographical and climatic differences but more probably come from the difference in experimental protocol. Indeed, these authors placed several frogs in plastic containers filled with moss or water and placed them in a freezing cabinet at a temperature of 2 °C [30], without following the frogs’ body temperatures. This protocol did not ensure freezing of the frogs and thus renders conclusions about freezing survival unreliable. In conclusion, this first attempt at understanding the evolutionary history of the freezing tolerance in ectotherms vertebrates reveals multiple evolutionary events at different periods of times, the most recent dating from, no later than 15.9 ± 7.6 Myr (between R. arvalis and R. temporaria). However, several questions arise, especially for Hyloids which will require further investigation to enlighten the evolutionary scenario. Addressing the evolution of freeze tolerance with such a large-scale approach will not only improve our current understanding of cold hardiness strategies, but might also create of a research framework guiding future comparative studies [15]. Acknowledgments We express our gratitude to Thomas Sorensen for assistance with collecting frogs and P. Joly for critically reading the manuscript. This work was supported by a research grant from the C.N.R.S. France for Y.V. and a grant from the Danish Science Research Council (# 21-01-0518) for H.R. The present investigation was carried out according to the ethical principles laid down by the European Convention for the protection of Vertebrate Animals used for Experimental and Scientific purposes (Council of Europe N° 123, Strasbourg, 1985).
References [1] ACEMAV coll, R. Duguet, F. Melki, Les amphibiens de France, Belgique et Luxembourg, éds Biotope, Méze (Fance), 2003. [2] P. Beerli, Genetic isolation and calibration of an average protein clock in Western Palearctic water frog of the Aegean region, Thesis, Zürich, 1994. [3] D.I. Berman, A.N. Leirikh, E.I. Mikhailova, Winter hibernation of the Siberian salamander Hynobius keyserlingi, Journal of Evolutionary Biochemistry and Physiology (1984) 323–327. [4] M. Boehme, The oldest representative of a brown frog (Ranidae) from the early miocene of Germany, Acta Palaeontologica Polonica 46 (2001) 19–124. [5] L. Bromham, D. Penny, A. Rambaut, M.D. Hendy, The power of relative rates tests depends on the data, Journal of Molecular Evolution 50 (2000) 296– 301. [6] D. Claussen, P.A. Zani, Allometry of cooling, supercooling, and freezing in the freeze-tolerant turtle Chrysemys picta, American Journal of Physiology 261 (1991) R626–R632. [7] J.P. Costanzo, R.E. Lee, P.H. Lortz, Glucose concentration regulates freeze tolerance in the wood frog Rana sylvatica, Journal of Experimental Biology 181 (1993) 254–255. [8] J.P. Costanzo, R.E. Lee, P.H. Lortz, Physiological responses of freeze-tolerant and intolerant frogs: clues to evolution of anouran freeze tolerance, American Journal of Physiology 265 (1993) R721–R725. [9] J.P. Costanzo, R.E. Lee, Cryoprotection by urea in a terrestrially hibernating frog, Journal of Experimental Biology 208 (2005) 4079–4089. [10] S.A. Croes, R.E. Thomas, Freeze tolerance and cryoprotectant synthesis of the Pacific Tree Frog Hyla regilla, Copeia (2000) 863–868. [11] C.J. Douady, E.J.P. Douzery, Molecular estimation of eulipotyphlan divergence times and the evolution of ‘‘Insectivora”, Molecular Phylogenetics and Evolution 28 (2003) 285–296. [12] J.R. Edwards, J.L. Jenkins, D.L. Swanson, Seasonal effects of dehydration on glucose mobilization in freeze-tolerant chorus frogs (Pseudacris triseriata) and freeze-intolerant toads (Bufo woodhousei and B. cognatus), Journal of Experimental Zoology A 301 (2004) 521–531. [13] J.R. Edwards, K.L. Koster, D.L. Swanson, Time course for cryoprotectant synthesis in the freeze-tolerant chorus, frog, Pseudacris triseriata, Comparative Biochemistry and Physiology A 125 (2000) 367–375. [14] D.R. Frost, T. Grant, J. Faivovich, R.H. Bain, A. Haas, C.F.B. Haddad, R.O. De Sa, A. Channing, M. Wilkinson, S.C. Donnellan, C.J. Raxworthy, J.A. Campbell, B.L. Blotto, P. Moler, R.C. Drewes, R.A. Nussbaum, J.D. Lynch, D.M. Green, W.C. Wheeler, The amphibian tree of life, Bulletin of the American Museum of Natural History (2006) 297–370. [15] J.T. Garland, A.F. Bennett, E.L. Rezende, Phylogenetic approaches in comparative physiology, Journal of Experimental Biology 208 (2005) 3015– 3035. [16] J.T. Garland, S.C. Adolph, Physiological differentiation of vertebrate populations, Annual Review of Ecology and Systematics 22 (1991) 193–228. [17] J.D. Graf, M. Polls Pelaz, Evolutionary genetics of the Rana esculenta complex, in: R.M. Dawley, J.P. Bogart (Eds.), Evolution and Ecology of Unisexual Vertebrates, New York State Museum, New York, 1989, pp. 289–301. [18] K. Grossenbacher, Frosclurch. Handbuch der reptilien und Amphibien Europas, AULA-Verlag, 2003. [19] D.M. Hillis, T.P. Wilcox, Phylogeny of the New World true frogs (Rana), Molecular Phylogenetics and Evolution 34 (2005) 299–314. [20] M. Holmstrup, M. Bayley, H. Ramlov, Supercool or dehydrate? An experimental analysis of overwintering strategies in small permeable arctic invertebrates, Proceedings of the National Academy of Sciences USA 99 (2002) 5716–5720. [21] J.P. Huelsenbeck, F.R. Ronquist, MRBAYES: bayesian inference of phylogeny, Biometrics 17 (2001) 754–755. [22] J.T. Irwin, J.P. Costanzo, R.E. Lee, Terrestrial hibernation in the northern cricket frog, Acris crepitans, Canadian Journal of Zoology 77 (1999) 1240–1246. [23] J.T. Irwin, R.E. Lee, Geographic variation in energy storage and physiological responses to freezing in the gray treefrogs Hyla versicolor and H. chrysoscelis, Journal of Experimental Biology 206 (2003) 2859–2867. [24] H. Kishino, J.L. Thorne, W.J. Bruno, Performance of a divergence time estimation method under a probabilistic model of rate evolution, Molecular Biology and Evolution 18 (2001) 352–361. [25] J.R. Layne, Postfreeze survival and muscle function in the leopard frog (Rana pipiens) and the wood frog (Rana sylvatica), Journal of Thermal Biology 17 (1992) 121–124. [26] D.P. Lotshaw, Temperature adaptation and effects of thermal acclimation in Rana sylvatica and Rana catesbeiana, Comparative Biochemistry and Physiology 56A (1977) 287–294. [27] G. Matz, D. Weber, Guide des amphibiens er reptiles d’Europe, Delachaux Niestlé SA, Lausanne (Switzerland), 1999. [28] P. Mazur, Freezing of living cells: mechanisms and implications, American Journal of Physiology 247 (1984) C125–C142. [29] G.C. Packard, M.J. Packard, To freeze or not to freeze: adaptations for overwintering by hatchlings of the North American painted turtle, Journal of Experimental Biology 207 (2004) 2897–2906. [30] S. Pasanen, M. Karhapää, Can boreal common frog (Rana temporaria L.) survive in frost?, Annales Zoologici Fennici 34 (1997) 247–250 [31] S. Pasanen, J. Sorjonen, Partial terrestrial wintering in a northern common frog population (Rana temporaria L.), Annales Zoologici Fennici 31 (1994) 275–278.
Y. Voituron et al. / Cryobiology 58 (2009) 241–247 [32] D. Posada, K.A. Crandall, MODELTEST: testing the model of DNA substitution, Bioinformatics 14 (1998) 817–818. [33] H. Ramlov, Aspects of natural cold tolerance in ectothermic animals, Journal of Human Reproduction 15 (2000) 26–46. [34] D. San Mauro, M. Vences, M. Alcobendas, R. Zardoya, A. Meyer, Initial diversification of living amphibians predated the breakup of Pangaea, American Naturalist 165 (2005) 590–599. [35] Z.B. Sanchíz, Salientia, in: P. Wellnhofer (Ed.), Handbook of Palaeoherpetology, Verlag Dr. Friedrich Pfeil, München, 1998, pp. 1–275. [36] W. Schmid, Survival of frogs in low temperature, Science 215 (1982) 697–698. [37] A.A. Steiner, S.O. Petenusci, L.G. Brentegani, L.G.S. Branco, The importance of glucose for the freezing tolerance/intolerance of the anuran amphibians Rana catesbeiana and Bufo paracnemis, Revista Brasileira de Biologia 60 (2000) 321–328. [38] K.B. Storey, Strategies for exploration of freeze responsive gene expression: advances in vertebrate freeze tolerance, Cryobiology 48 (2004) 134–145. [39] K.B. Storey, J.M. Storey, Biochemical adaptation for freezing tolerance in the wood frog, Rana sylvatica, Journal of Comparative Physiology B 155 (1984) 29–36. [40] K.B. Storey, J.M. Storey, Freeze tolerance and intolerance as strategies of winter survival in terrestrially-hibernating amphibians, Comparative Biochemistry and Physiology 83A (1986) 613–617. [41] K.B. Storey, J.M. Storey, Freeze tolerance in animals, Physiological Reviews 68 (1988) 27–84. [42] K.B. Storey, J.M. Storey, Physiology, biochemistry, and molecular biology of vertebrate freeze tolerance: the wood frog, in: J. Fuller (Ed.), Life in the Frozen State, CRC Press, Washington, DC, 2004, pp. 243–274. [43] A.H.P. Stumpel, On hibernation sites in the tree frog Hyla arborea, Amphibia– Reptilia 11 (1990) 304–306. [44] D.L. Swanson, B.M. Graves, Supercooling and freeze intolerance in overwintering juvenile spadefoot toads (Scaphiopus bombifrons), Journal of Herpetology 29 (1995) 280–285.
247
[45] D.L. Swanson, B.M. Graves, K.L. Koster, Freezing tolerance/intolerance and cryoprotectant synthesis in terrestrially overwintering anurans in the great plains, USA, Journal of Comparative Physiology B 166 (1996) 110–119. [46] D.L. Swofford, Phylogenetic analysis using parsimony (*and other methods). PAUP* (2002). [47] J.L. Thorne, H. Kishino, Divergence time and evolutionary rate estimation with multilocus data, Systematic Biology 51 (2002) 689–702. [48] J.L. Thorne, H. Kishino, I.S. Painter, Estimating the rate of evolution of the rate of molecular evolution, Molecular Biology and Evolution 15 (1998) 1647–1657. [49] M. Veith, J. Kosuch, M. Vences, Climatic oscillations triggered post-Messinian speciation of Western Palearctic brown frogs (Amphibia, Ranidae), Molecular Phylogenetics and Evolution 26 (2003) 310–327. [50] Y. Voituron, T. Lengagne, Freezing before mating severely impacts fitness of a capital breeder, Physiological and Biochemical Zoology 81 (2) (2008) 204–210. [51] Y. Voituron, L. Paaschburg, H. Barré, M. Holmstrup, H. Ramløv, Survival and metabolism of Rana arvalis during freezing, Journal of Comparative Physiology B 179 (2009) 223–230. [52] Y. Voituron, M. Eugène, H. Barré, Survival and metabolic responses to freezing by the water frog (Rana ridibunda), Journal of Experimental Zoology 299A (2003) 118–126. [53] Y. Voituron, P. Joly, M. Eugene, H. Barre, Freezing tolerance of the European water frogs: the good, the bad and the ugly, American Journal of Physiology 288 (2005) R1563–R1570. [54] Y. Voituron, N. Mouquet, C. deMazancourt, J. Clobert, To freeze or not to freeze? An evolutionary perspective on the cold hardiness strategies of overwintering ectotherms, American Naturalist 160 (2002) 255–270. [55] Z.H. Yang, PAML: a program package for phylogenetic analysis by maximum likelihood, Computational Applied Bioscience 13 (1997) 555–556. [56] P. Zhang, H. Zhou, Y.Q. Chen, Y.F. Liu, L.H. Qu, Mitogenomic perspectives on the origin and phylogeny of living amphibians, Systematic Biology 54 (2005) 391– 400.