Responses to Freezing Exposure of Hatchling Turtles Trachemys Scripta Elegans: Factors Influencing the Development of Freeze Tolerance by Reptiles

The ability to endure the freezing of extracellular body fluids is an important component of winter cold-hardiness for many ectothermic animals. Freeze tolerance has been reported for numerous species of insects, some other invertebrates, and also for several terrestrially hibernating amphibians and reptiles (Storey and Storey, 1988; Storey, 1990). Among vertebrates, the species with well-developed natural freeze tolerance include several frogs [Rana sylvatica, Hyla versicolor, H. chrysoscelis, Pseudacris triseriata, Pseudacris (formerly Hyla) crucifer], the Siberian salamander (Hynobius keyserlingi), box turtles (Terrapene Carolina) and hatchlings of the painted turtle (Chrysemys picta) (Schmid, 1982; Berman et al. 1984; Storey and Storey, 1984, 1985, 1986; Storey et al. 1988; Costanzo and Claussen, 1990; Costanzo et al. 1990; Churchill and Storey, 1992a; K. B. Storey, J. R. Layne, M. M. Cutwa, T. A. Churchill and J. M. Storey, in preparation). All of these species tolerate high amounts of body ice (> 50% of total body water) and long durations of freezing (days or weeks) at temperatures sometimes encountered in their hibernacula (lows may reach –6 to –8°C). Much more limited freezing survival, which probably has little or no physiological relevance for long-term hibernation, has also been reported for various aquatic frogs (Duméril, 1849; Cameron and Brownlee, 1913; Holmes, 1927), wall lizards (Weigmann, 1929; Claussen et al. 1990) and garter snakes (Costanzo et al. 1988; Churchill and Storey, 1992b). For example, garter snakes can recover after very brief or very mild freezing exposures that keep the percentage of body ice low, but their primary hibernation strategy is to avoid freezing exposure by hibernating in thermally buffered dens under water or deep underground (Macartney et al. 1989; Costanzo, 1989).

Previous studies show considerable variation and gradation in the ability of different vertebrate species to endure freezing. Some have perfected the strategy to the point where it is an integral part of natural hibernation, others can tolerate short freezing exposures, and many are killed by even the briefest freezing exposures. To help to determine which physiological and biochemical traits are important for freezing survival by reptiles, we chose to examine the freezing abilities of a turtle species that has little or no natural need for freeze tolerance and yet is closely related to another species with highly developed freeze tolerance.

Pond turtles of the Chrysemys and Trachemys (=Pseudemys) genera are widely distributed throughout North America. Adults hibernate under water and have extremely well-developed anoxia tolerance, making them the premier facultative anaerobes among vertebrates (Ultsch, 1985). Hatchlings of both share the strategy of delayed emergence from the nest; when eggs hatch late in the season, the young remain within the terrestrial nest cavity over the winter months and emerge in the spring when conditions are optimal for rapid juvenile growth (Gibbons and Nelson, 1978). In the north, this strategy has necessitated the development of freeze tolerance by C. picta hatchlings, both by the midland (C. p. marginata) and by the western (C. p. bellii) subspecies (Storey et al. 1988; Churchill and Storey, 1992a). Both the placement of nests on exposed banks and their shallow depth (less than 10 cm) are such that temperatures within the nest cavities frequently fall to subzero values (Breitenbach et al. 1984; Storey et al. 1988; Packard et al. 1989). In the south, however, hatchlings of the red-eared pond slider Trachemys scripta elegans do not experience subzero temperatures while hibernating in the nest. Our study analyzes the extent of freezing survival by T. s. elegans hatchlings from a Louisiana population, comparing and contrasting the freezing limits, the amount of ice accumulated and the biochemical responses to freezing of these animals with the equivalent parameters displayed by C. picta hatchlings.

Hatchling red-eared pond sliders T. s. elegans were purchased in September 1989 from a commercial breeder in Louisiana. The animals were placed in plastic boxes containing damp sphagnum moss and were held without feeding in an incubator at 5°C (range±0.5°C) for up to 3 months before use. Mean hatchling mass was 6.84±0.13g; animals were of indeterminate sex. All biochemicals were purchased from Sigma Chemical Co., St Louis, MO, or Boehringer Mannheim, Montreal, PQ.

To monitor cooling and freezing of the turtles, individual turtles were placed on a pad of paper towelling, centred over a thermistor that was in contact with the plastron. A band of masking tape was used to secure the turtle and thermistor in place. Thermistors were connected to a YSI telethermometer with output to a linear recorder. Turtles were then placed in an incubator at either –2.5 or –4°C (±0.1°C) and were allowed to cool over 30–40 min until the exotherm resulting from ice nucleation within the body fluids was recorded. The length of freezing was timed from the exotherm, with sampling at intervals up to 3 days at –2.5°C or 8h at –4°C for analysis of percentage ice and survival. If turtles supercooled to –2.5°C without freezing, the incubator temperature was further lowered to –3.5°C and cooling was continued until an exotherm was recorded. Within 5 min of the appearance of an exotherm, incubator temperature was raised to –2.5°C and maintained at –2.5°C for the duration of the experiment.

After a timed interval of freezing, turtles were thawed in a calorimeter, to determine the percentage of body water as ice, and were then returned to the 5°C incubator for assessment of survival after 24 h. If the turtles were alive after 24 h, survival was again checked after 2 weeks. Calorimetry was performed as outlined by Lee and Lewis (1985) in an insulated vessel containing 30 ml of water at 21°C and using experimentally determined factors: body water content=79.5±0.8% of total mass, specific heat of the dry mass=0.336±0.019 cal g⁻¹ degree⁻¹, and F factor for the calorimeter=1.03±0.01. Survival was assessed by several physical responses including breathing, retraction of head or legs when touched, and walking.

Cooling and freezing were carried out as described above but turtles were sampled over a period of up to 4h at –4°C; a survivable stress. After freezing, animals were rapidly removed from the incubator, killed by decapitation, and organs were quickly excised and frozen in liquid nitrogen. Tissues were transferred to –80°C for storage until processed. Control turtles were treated similarly to the experimental animals, with the thermistor attached and cooling in the incubator, but these animals were removed and killed when plastron surface temperature reached 0°C.

Perchloric acid extracts of turtle organs were prepared as outlined by Storey and Storey (1984), except that Tris buffer, instead of imidazole, was included in the neutralization solution. Samples of the neutralized extracts were immediately removed and assayed for pyruvate and phosphoenolpyruvate (PEP); the remainder of the extracts was frozen for subsequent use. Metabolite levels were measured using either fluorometric or spectrophotometric coupled enzyme assays (Lowry and Passonneau, 1972; Bergmeyer, 1984) as described previously (Storey and Storey, 1984; Kelly and Storey, 1988). For analysis of free amino acid levels, frozen tissue samples were homogenized in 0.5% w/v sulphosalicylic acid and then centrifuged at 6000g in an IEC benchtop centrifuge. Amino acids were analyzed in samples of the supernatant using a Waters HPLC after precolumn derivatization with orthophthalaldehyde.

Data are presented as means±s.E.M. Statistical testing used the two-tailed Student’s t-test.

When cooled slowly in an incubator with an air temperature of –4°C, T. s. elegans hatchlings showed a moderate ability to supercool. The mean (±S.E.M.) whole-animal supercooling point was –3.2±0.08°C (N=11). Upon nucleation, plastron surface temperature rebounded to –1.13±0.04°C.

To test the freeze tolerance of T. s. elegans hatchlings, animals were held, after nucleation had occurred, at either –2.5 or –4°C. Turtles readily survived 2h of freezing at –2.5°C with a low ice content, 22.9±2.4%, of total body water (Fig. 1). However, as body ice content rose above 40%, survival dropped; 50% of animals recovered after 6h of freezing (mean ice content 41.5±4.8%) but only 25% survived 12 h of freezing. None of the turtles revived after longer periods of freezing. Maximal ice content in turtles frozen for 3 days at –2.5°C was 54.7±1.4% (N=12).

At –4°C, 100% of turtles survived up to 4 h of freezing with mean ice contents of 49.6±2.4% of body water (Fig. 2). However, only 25% of animals recovered after 5 h of freezing and none revived after 8 h of freezing. Not unexpectedly, ice content was higher in turtles frozen at –4°C, 64.5±0.7% after 8h, than the maximum at –2.5°C. Fig. 2 also shows that plastron surface temperature remained elevated and fairly constant (–1.3 to –1.4°C) over the first 2h of freezing but then dropped steadily towards the ambient temperature as body ice content neared a maximum.

The effect of freezing exposure on metabolite levels in seven organs was assessed in hatchlings exposed to –4°C for up to 4 h. Levels of glucose in control turtles (chilled only to 0°C) were low in all organs, less than or equal to 1.5 μmolg⁻¹ wet mass (Fig. 3A). During freezing, liver showed a progressive rise in glucose content, culminating in a twofold increase to 3.2μmol g⁻¹ after 4h. Glucose content in muscle also doubled and glucose rose transiently to 1 μmol g⁻¹ in brain after 1h of freezing. Freezing affected lactate levels in some organs (Fig. 3B). Lactate level increased by two-to threefold in heart and brain over the 4h of freezing exposure, and rose slightly in liver but no change in lactate content occurred in muscle, kidney or lung. In gut (stomach+intestines combined), lactate content decreased during freezing. Levels of other possible cryoprotectants, glycerol, sorbitol, fructose and mannose, were assessed in liver samples. Glycerol content was 0.36±0.10μmolg⁻¹ in liver of control turtles and 0.67±0.13μmolg⁻¹ (not significantly different) in frozen animals (both N=4).

Sorbitol, fructose and mannose levels were each less than 0.55μmolg⁻¹ and did not change with freezing. Since freezing leads to an ischaemic state in organs, levels of two alternative products of anaerobic metabolism, alanine and succinate, were also assessed in the seven organs. However, there were no significant changes in the level of either compound during freezing exposure. Mean organ alanine content ranged from 0.20±0.02gmolg⁻¹ wet mass in skeletal muscle to 2.80±0.23 μmol g⁻¹ in gut, whereas succinate levels ranged from 0.45±0.03 μmol g^−J in muscle to 3.17±0.14μmolg⁻¹ in heart (control+freezing-exposed combined, N=15).

Table 1 shows that total amino acid levels rose significantly in T. s. elegans liver, muscle and blood after 4h of freezing at –4°C. In liver, the major peaks were identified as glutamine, valine, taurine and alanine, as also occurs for C. picta hatchlings (Storey et al. 1988; Churchill and Storey, 1992a). The increase in liver amino acid content with freezing was primarily due to a 4.1μmolg⁻¹ rise in taurine. In muscle, aspartate, serine and glycine contents increased during freezing. Levels of eight amino acids increased in T. s. elegans blood during freezing, but about 50% of the total rise in the amino acid pool was due to increased taurine plus alanine concentrations.

Levels of some glycolytic intermediates in liver changed over the course of freezing at –4°C (Fig. 4). Glucose 6-phosphate (G6P) content rose threefold over the first 2h, suggesting an activation of liver glycogenolysis, but subsequently declined. Fructose 1,6-bisphosphate (F1,6P₂) and pyruvate (PYR) contents remained constant over the first 2h, but then increased sharply at 4 h of freezing exposure. These increases in the products of the two regulatory enzymes of glycolysis, phosphofructokinase and pyruvate kinase, suggested an activation of flux through both loci. Levels of fructose 6-phosphate (F6P), dihydroxyacetonephosphate (DHAP), glyceraldehyde 3-phosphate (GAP) and PEP were unchanged over the freezing exposure.

In skeletal muscle, levels of F6P, DHAP and PYR had increased after 2h of freezing, but declined thereafter (Fig. 5). G6P, GAP and PEP levels showed no change over the course of freezing. The level of F1,6P₂ declined at 30min, but subsequently returned to the control value.

Trachemys s. elegans hatchlings tolerated short-term freezing exposures at either –2.5 or –4°C as long as the percentage of body water as ice remained relatively low. All animals recovered fully after 2h of freezing and with ice contents of less than 30% of total body water. Survival was also 100% for turtles exposed to –4°C for 4h and with a mean ice content of 49.6±2.4%. However, slightly longer exposure times, 5h at –4°C or 6h at –2.5°C, and/or higher ice contents resulted in a sharp drop in survival rate. Thus, both time spent frozen and the amount of ice were significant factors in survival; all animals survived 4h at –4°C with 50% ice but only one-quarter of turtles revived after 12 h freezing at –2.5°C when ice content was also 50% of body total water. It should be noted that the rapid thawing in the calorimeter could be detrimental to the animals and that survival rates might be higher if thawing were done slowly at a lower temperature, as would occur in nature.

Freezing survival by T. s. elegans hatchlings from Louisiana was much poorer than the freeze tolerance exhibited by both midland and western subspecies of painted turtles from Canadian populations (Churchill and Storey, 1992a). Autumn-collected C. p. marginata and C. p. bellii hatchlings, when tested using identical methodologies and with rapid thawing in the calorimeter, showed 100% survival after 3 days of continuous freezing at –2.5°C. Spring-collected C. p. marginata showed even greater tolerance, reviving after 11 days frozen at –2.5°C (Churchill and Storey, 1992a). Maximal ice contents were also substantially lower for autumn C. picta hatchlings, averaging 43.5–46.6% at –2.5°C (Churchill and Storey, 1992a) compared with the mean 55% for T. s. elegans at the same temperature (Fig. 1). Comparable results were seen at –4°C; maximal ice content was 64.5% for T. s. elegans (Fig. 2) and substantially lower, 49.8–52.6% of body water, for autumn C. picta hatchlings (Churchill and Storey, 1992a). Both animal and cryomedical studies have shown that the ability to minimize both the total amount of ice formed and the rate of ice formation are important to freezing survival. The above comparison of two closely related turtle species, C. picta, which routinely employs freeze tolerance for hibernation, and T. s. elegans, which does not, appears to confirm this for reptiles.

Although freeze tolerance is clearly less well developed in Louisiana T. s. elegans hatchlings than it is in northern C. picta hatchlings, freezing survival by T. s. elegans compared favourably with that of garter snakes (Costanzo et al. 1988; Churchill and Storey, 1992b) and far exceeded that of wall lizards (Claussen et al. 1990). Garter snakes frozen at –2.5°C recovered fully after freezing exposures of 3 h or less, which produced ice contents of up to about 40% of total body water; with longer times and greater ice contents, however, survival declined (Churchill and Storey, 1992b). Studies of wall lizards found that only one-third of animals recovered after freezing for 2 h or less and the maximal ice content tolerated was only 28% (Claussen et al. 1990).

The limited freeze tolerance displayed by the Louisiana T. s. elegans hatchlings of the present study probably has little ecological relevance for the natural overwintering of southern populations of the species since the probability of experiencing subzero temperatures within the nest cavity would be very low in their southern homeland. Furthermore, for the hatchlings to survive, the net freezing exposure would have to be very mild (combining a short freezing time and only a few degrees below 0°C) and such circumstances would not often occur, particularly since, once frozen, the hatchlings cannot thaw until the temperature rises above the melting point (–0.5°C) of their body fluids. However, we cannot rule out a natural use for freezing abilities displayed by this species, since the range of the subspecies extends to northern Illinois (Behler and King, 1979), where colder winters would be encountered. Among T. s. elegans populations in colder climates, the freeze tolerance of hatchlings could be much more highly developed. Indeed, Ultsch et al. (1985) have demonstrated a north-south cline in physiological responses that are vital to hibernation success among adult painted turtles. Northern C. p. bellii, which may spend up to 6 months hibernating under water, showed far less metabolic stress during long-term anoxic submergence at 3°C than did southern C. p. dorsalis, which may not hibernate at all. We have previously postulated that anoxia tolerance is an important factor in freezing survival (Storey and Storey, 1988) and turtles of the Trachemys and Chrysemys genera are among the premier vertebrate facultative anaerobes (Ultsch, 1985; Kelly and Storey, 1988). Latitudinal clines in anoxia tolerance among these genera may be a key element of freezing survival by the hatchlings in their winter nests.

Two of the specific adaptations typically displayed by freeze-tolerant animals are ice-nucleating proteins and low molecular weight cryoprotectants (Storey and Storey, 1988). Plasma of hatchling C. p. marginata contains proteinaceous icenucleators, whereas that of adult painted turtles does not (Storey et al. 1991). The status of ice nucleators in T. s. elegans is not yet known. The addition of cryoprotectants to body fluids helps to reduce, via colligative effects, the amount of ice that can form at any given temperature. Freeze-tolerant frogs accumulate glucose or glycerol for this purpose in amounts ranging up to several hundred millimoles per litre (Schmid, 1982; Storey and Storey, 1984,1985,1986). However, our studies with C. picta hatchlings indicated only low amounts of cryoprotectants. Levels of glucose, glycerol, lactate, taurine and glutamate increased during freezing exposure in spring-collected C. p. marginata (Storey et al. 1988), but analyses of autumn-collected turtles indicated that only glucose and lactate levels showed sufficiently large changes to have an impact on cryoprotection (Churchill and Storey, 1992a,c). Box turtles also produced only low amounts of these compounds during freezing (K. B. Storey, J. R. Layne, M. M. Cutwa, T. A. Churchill and J. M. Storey, in preparation). Organs of T. s. elegans hatchlings accumulated glucose, lactate and some amino acids during freezing. Net accumulations of glucose and lactate were low (a maximum of 3.2μmolg⁻¹ for glucose; 9.8μmolg⁻¹ for lactate) and restricted to selected organs (glucose in liver and muscle; lactate in brain and heart) with a pattern that was similar to the effects of anoxia on these metabolites in organs of adult T. s. elegans (Kelly and Storey, 1988). The accumulation of these compounds in T. s. elegans organs may have been due, therefore, to the anoxia and ischaemia stresses associated with freezing. By comparison, organs of C. p. bellii hatchlings built up greater amounts of these compounds during equivalent 4-h freezing exposures at –4°C (up to 16μmolg⁻¹ glucose and 25μmolg⁻¹ lactate) (Churchill and Storey, 1992c). This could suggest that the established responses to anoxia have been accentuated by C. p. bellii in order rapidly to build up pools of glucose and lactate for cryoprotection during freezing.

Changes in glycolytic intermediates in liver of T. s. elegans hatchlings revealed increased levels of both F1,6P₂ and PYR after 4h of freezing. Elevated levels of these products of the regulatory enzymes of glycolysis, phosphofructokinase and pyruvate kinase, indicate enzyme activation and suggest that the 4h of freezing exposure caused an energetic stress on liver that resulted in glycolytic activation. For long-term survival while frozen, such a metabolic activation would be detrimental. Not unexpectedly, the opposite response occurred in liver of C. picta hatchlings, with changes in the levels of glycolytic intermediates that were consistent with metabolic rate depression during freezing (Churchill and Storey, 1992c). Thus, another critical element of the development of freeze tolerance by reptiles may be the ability to use metabolic arrest strategies to reduce metabolic rate further during freezing and, thereby, maximize the time that an animal can survive in the frozen state.

We thank W. Bridger, University of Alberta, for providing the succinyl-CoA synthetase for the succinate assay. We also thank J. M. Storey for critical commentary. Supported by operating grants from the NIH (GM43796-01) and the NSERC Canada to K.B.S. and an NSERC postgraduate scholarship to T.A.C.