INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

WESTERN PAINTED TURTLES (Chrysemys picta bellii), subjected to prolonged anoxic submergence, utilize their shells in two ways to counteract the resultant lactic acidosis. First, carbonate buffers are released from the shell to supplement extracellular buffering capacity; second, lactic acid enters the shell and is buffered there. Together, these mechanisms may be responsible for as much as 75% of the total body lactic acid buffering under extreme anaerobic conditions (11).

The first mechanism, the release of buffers, is a well-known response to acidosis in numerous organisms with calcified skeletons or exoskeletons. In the crab, Cancer productus, for example, CaCO₃ is released in response to emersion acidosis (6). In mammals, bone demineralization occurs in response to chronic acidosis and is associated with the loss of carbonate (1, 10), calcium (3), and sodium (2). Clinically, skeletal buffering can be very effective in long-term acid-base disturbances caused, for example, by renal insufficiency (20), but extensive demineralization can occur leading to bone pathology (7). The response of the turtle shell to long-term acidosis is similar to mammalian bone and appears to involve the release of calcium and magnesium carbonates (13, 26). The role of sodium is unclear, however. In addition, the anoxic submerged turtle is a relatively closed system with no pulmonary gas exchange and minimal renal function or excretion (16, 26); consequently, ions released from shell remain in the body and accumulate within the extracellular fluid. Plasma total Ca²⁺ and Mg²⁺ concentrations can reach high concentrations after months of anoxic submergence at 3°C (16).

The second mechanism, in which lactate enters shell during anoxic acidosis and is buffered and stored there, has not previously been described in other organisms. The turtle, because of its large shell mass and its high anoxic lactate levels persisting over many months at low temperature, is an ideal animal for observing this phenomenon. It is uncertain, however, whether or to what extent normal bone participates in this response.

These shell buffering mechanisms, together with a capacity for extreme metabolic depression, contribute to making C. picta bellii the most anoxia-tolerant freshwater turtle thus far studied (23). This well-documented physiological capacity of this species conforms with its natural history. Within its normal range that extends into southern Canada, C. picta bellii must endure long periods of winter submergence beneath the ice possibly under severely hypoxic conditions (23).

Field studies, however, reveal that turtles trapped beneath the ice may be exposed to water that is well aerated, and for some species sufficient oxygen may be obtained directly from the water to sustain aerobic metabolism (4). Whether this is the case for C. picta bellii is uncertain. An earlier laboratory study directed at this question produced equivocal results, because many of the turtles developed a fungal infection that may have disturbed their extrapulmonary gas exchange (25).

The present investigation was initiated with several goals in mind that can be stated as hypotheses. First, we hypothesize that C. picta bellii submerged in aerated water at 3°C can sustain aerobic metabolism and maintain a normal blood acid-base state. Second, we hypothesize that turtle shell, like mammalian bone, releases sodium carbonates, in addition to calcium and magnesium carbonates, into the extracellular fluid during anoxia. Finally, we hypothesize that the skeleton of the turtle, as well as its shell, participates in lactic acid buffering, particularly by sequestering lactate during the period of anoxia.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Animals

Experimental Protocol

The sampling procedure was identical for each animal, except that bone and shell samples were taken from all control and anoxic turtles but only from the 125-day normoxic submerged turtles. The tank was opened, the turtle was selected and euthanized, and its body mass was recorded. A hole was then trephined in the plastron, exposing the heart, and an anaerobic blood sample (~0.5 ml) was taken by cardiac puncture into a heparinized tuberculin syringe and immediately taken to the blood-gas analyzer for measurement of pH, PO₂, and PCO₂. A second blood sample (~1 ml) was promptly collected into a dry syringe for determination of hematocrit and, following centrifugation, for determination of plasma osmolality and plasma concentrations of lactate, glucose, Na⁺, K⁺, Cl, total Ca²⁺, total Mg²⁺, and inorganic phosphorus (P_i). The blood did not clot during this procedure. This entire sampling procedure took place in the 3°C cold room and lasted 5-10 min. Because of the extremely slow rate of metabolism in these animals at this temperature, particularly in the anoxic animals, it is unlikely that detectable changes occurred in any measured variables during this time.

Shell samples were taken using a hand punch (Roper Whitney, model no. 5 Jr.) from the plastron disk produced by the drill and from the margins of the carapace. A portion of the collected shell was wrapped in preweighed foil for immediate determination of water content and for later elemental analysis; the remainder was wrapped in foil and stored at 80°C for later analysis of lactate concentration. The right hind limb was removed and was also stored at 80°C. At a later time, the limb bones were isolated and processed similarly to the shell. The choice of the right limb was arbitrary but was taken from each animal for consistency. We assume that values measured on this limb were representative of the skeleton as a whole.

Analytical Procedures

Blood and plasma. Blood pH was measured using a semi-micro-combination pH electrode (Orion no. 3203BN) and meter (Orion model 330). The calibrated electrode was inserted into a small test tube at 3°C, and the blood sample (ca 0.3 ml) was injected into the bottom of the tube around the electrode. Although the procedure is not strictly anaerobic, the reading was stable and was consistent with earlier measurements using the Radiometer capillary method. Blood PO₂ and PCO₂ were measured at 3°C using Radiometer electrodes, E5046 and E5036, respectively, in combination with a Radiometer Mk2 Blood Micro System. Gas electrodes were calibrated with a gas-mixing pump (Wösthoff, Bochum, Germany). Blood bicarbonate concentration was calculated using the Henderson-Hasselbalch equation with  = 0.0808 mmol · l¹ · Torr¹ (19) and pK' = 6.293 (18). Hematocrit was measured following centrifugation for 3 min in an IEC microhematocrit centrifuge.

Shell and bone. Water content of fresh shell samples was determined by oven drying (24 h at 85°C). The dried shell was then stored frozen in a capped vial. For processing, the dried shell was ground to powder with a Freezer Mill (SPEX Certiprep, model 6700), redried in the oven, and then a weighed fraction of the dry powder was ashed in preweighed porcelain cups at 450°C using a muffle furnace (Thermolyne, model 1300). The weight of residual ash enabled us to calculate ash content of the fresh shell, and the balance of the shell mass was assumed to be organic matter.

Statistics

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Plasma Composition

View larger version (21K): [in this window] [in a new window]   Fig. 1.   Plasma, shell, and bone lactate concentrations during submergence in N2-equilibrated water (anoxia) and air-equilibrated water (normoxia). Plasma concentrations are in mmol/l, shell and bone concentrations are in µmol/g wet wt. Note the progressive increase during anoxia but the transient rise in plasma lactate concentration during normoxia that then plateaued for the final l00 days of submergence.

The observed changes in the submerged normoxic turtles were generally smaller than in the anoxic animals, and no significant changes were observed for plasma [K⁺], total [Ca²⁺], and phosphate concentration. The changes in all other variables were significant (one-way ANOVA), although in several instances (total [Mg²⁺], glucose concentration, and PCO₂) significance was due to transient changes at intermediate sampling periods; comparison of control and 125-day values (t-test) revealed no difference in these three variables. Plasma lactate concentration rose during the first 25 days of submergence but then remained relatively unchanged for the remainder of the sampling periods (Table 2; Fig. 1). Blood pH, although more variable, fell during the first 25 days and then fluctuated at or above this level (Table 2). Plasma osmolality of the normoxic submerged turtles fell significantly, in contrast to the significant rise in the anoxic animals, and this agreed with an overall decrease in measured solute concentration. We observed an increase in body mass, presumably due to uptake of water, in both groups, and this increase was greater in the normoxic group (Table 1).

Shell and Bone Composition

Shell and bone elemental compositions were similar when expressed as micromoles per gram ash (Table 3); however, because of the somewhat higher ash content of shell, mineral concentrations were higher in shell when expressed on a per gram wet weight basis. In shell, [Na⁺] was the only measured element to change significantly (P < 0.01) during anoxic submergence (Table 3; Fig. 2). Shell [Na⁺] decreased by about 19% over the course of anoxic submergence but was unchanged from control in the 125-day normoxic submergence turtles. In bone, [Na⁺] also fell significantly (P < 0.001) by about 17%, but a significant fall occurred as well in bone [Ca²⁺], and an increase was observed in bone [K⁺]. The rise in [K⁺], but not the fall in [Ca²⁺], appeared to be progressive with time of submergence.

View larger version (19K): [in this window] [in a new window]   Fig. 2.   Shell and bone sodium and potassium concentrations during submergence at 3°C in the western painted turtle Chrysemys picta bellii. Shell values are depicted by open symbols and bone values by solid symbols. Shell and bone values for sodium concentration are also shown for group submerged in aerated water (normoxic) for 125 days. Values are means ± SE. In some cases the SE is so small it is hidden in the symbol.

Shell and bone lactate increased parallel to plasma lactate during anoxic submergence (Fig. 1). The changes in shell and bone were closely correlated with the plasma values, although the slope of the relationship of bone lactate vs. plasma lactate was significantly steeper (Fig. 3).

View larger version (21K): [in this window] [in a new window]   Fig. 3.   Shell and bone lactate concentration in µmol/g wet wt vs. plasma lactate concentration in mmol/l. Linear regression equations for the two relationships are shown.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The results from blood and plasma analysis conform to previous observations and illustrate the pattern of change characteristic of this turtle's response to submergence. In this study, in contrast to our earlier study with a generally similar design (17, 25), the turtles were not catheterized. This approach had the disadvantage that repeated measurements were not made on the same individuals, but it had the advantage that a fungal condition that adversely affected the turtles submerged in aerated water in the earlier study was not a factor here, and the turtles in the present group experienced less severe acid-base and ionic disturbances than in the earlier study. In addition, the design we employed in this study was also required to collect terminal shell and bone samples for analysis.

The benefit to the hibernating animal of oxygen in the water was made more evident by our current results. These animals exhibited a transient progressive metabolic acidosis with associated plasma ionic changes (elevated concentrations of calcium, magnesium, and decreased concentration of chloride), but after 25 days the condition of the animals stabilized and did not worsen for the balance of the submergence period. This indicates that the overall metabolism of the turtles was aerobic during the final 100 days even though a moderate lactic acidosis persisted. This supports our first hypothesis stated in the introduction. We postulate that a downregulation in metabolism occurred by about 25 days that permitted oxygen supply to match oxygen demand. A gradual metabolic reduction has been observed in overwintering frogs (Rana temporaria) by Donohoe et al. (5). It is also possible that lactate production continued to occur in some tissues of the normoxic turtles but that this was counteracted by lactate metabolism elsewhere in the animal.

As discussed in detail earlier (14, 17), the observed plasma changes of the anoxic submerged turtles in the concentrations of K⁺, Cl, total Ca²⁺, and total Mg²⁺ can be interpreted as compensatory to the increase in lactate concentration and contribute to the maintenance of a positive strong ion difference (21). Quantitatively, the most important changes are the increases in total Ca²⁺ and Mg²⁺ and the decrease in Cl. It is now almost certain that the Ca²⁺ and Mg²⁺ derive from the shell and bone of the turtle and that these are released in association with carbonate (CO²₃). The evidence supporting this is as follows: 1) over 99% of the total body stores of Ca²⁺ and Mg²⁺ reside in shell and bone (12), and 2) in vitro studies of incubated shell powder reveal a release of Ca²⁺ and Mg²⁺ as a direct function of the acidity of the incubating solution (13). The evidence for CO²₃ as the associated buffer anion released from shell and bone is that a significant fall occurred in total shell CO₂ during anoxia (26) and that the release of CO₂ from incubating shell powder was stoichiometrically consistent with CO²₃ as the source (13). The other candidate anion to accompany Ca²⁺ and Mg²⁺ is phosphate, but we observed no change in plasma phosphate concentration during anoxic submergence (Table 1), and phosphate was not released significantly from incubating shell powder even at solution pH as low as 6.0 (13). A small (0.2 mM) but significant increase in plasma phosphate concentration was previously observed during anoxic submergence at 10°C (26).

This study also explored the possible contribution of the skeleton of the turtle to this buffering process. The mass of bone outside the shell is relatively small compared with the shell. The shell is about 32% of body mass, and the nonshell skeleton is about 5.5% (11). In addition, the smaller ash content of bone (35-41% in bone vs. 44-46% in shell) means that available mineral buffer in bone is even less than the comparative mass figures would suggest. Although bone and shell Ca²⁺ concentrations expressed per gram ash are similar, [Ca²⁺] expressed as micromoles per gram wet weight is about 4,000 in bone, whereas in shell the comparable value is about 4,400. At present, however, we do not know whether the two bony tissues behave identically in vivo with regard to Ca²⁺ release. Suzuki (22) asserted that the shell, in contrast to the limb bones, does not release Ca²⁺ during egg production. Interestingly, we observed a significant decrease in bone but not in shell [Ca²⁺], although this may have been artifactual, due to sampling from different groups of animals. We have usually assumed that the large background concentration of Ca²⁺ would make it impossible to detect a change in concentration in the shell, even with the large increases in plasma [Ca²⁺].

We observed a significant fall during anoxic submergence in both shell and bone sodium concentration, a response not previously reported in this animal but known to occur in mammalian bone under conditions of acidosis or hyponatremia (8). This confirms the second of our hypotheses stated in the introduction. The mineral release from turtle shell and bone during lactic acidosis is therefore more complex than previously described. Based on this study and the earlier in vitro study (13), we now believe that Ca²⁺, Mg²⁺, and Na⁺ carbonates are mobilized to supplement extracellular buffering. The decrease we observed in shell and bone Na⁺ was substantial, amounting to over 15% of total Na in both shell and bone. This observation, although consistent with the behavior of mammalian bone, is paradoxical in the present context because of the concurrent modest decrease in plasma [Na⁺]. It is not clear what has become of the released Na⁺. This can be illustrated by a simple calculation that is restricted to the dominant shell component. The shell lost 19% of its control Na⁺ as the concentration fell from 159 µmol/g wet wt in the control animals to 130 µmol/g wet wt after 125 days of anoxic submergence. For a 1-kg turtle with 0.32 kg of shell, this represents a total loss of 9.3 mmol of Na⁺. If this Na⁺ distributed throughout the nonshell extracellular fluid (0.24 liters; see Ref. 16), then the [Na⁺] in this compartment should have increased by about 39 mmol/l, assuming a constant volume. If we included the bone contribution, then the projected increase would be even greater, but at most about 43 mmol/l. Instead of an increase in plasma [Na⁺], however, we observed a fall in concentration of 26 mmol/l. As noted above, in earlier studies this fall in plasma [Na⁺] has been smaller and usually insignificant, but an increase in [Na⁺] has never been reported.

The concurrent and even greater fall of plasma [Cl] may provide an insight into what is occurring. Extracellular volume may have expanded as a result of water uptake, producing the large fall in [Cl] and a potentially similar fall in [Na⁺]. The fall in [Na⁺], however, was counteracted by the release of Na⁺ from bone and shell. The numbers in general support this hypothesis. The decrease in [Cl], from 84.0 to 46.7 mM is 44%; the estimated overall decrease in [Na⁺], from 174 (131 + 43) to 105.5 mM, is 39%. An expansion of the extracellular fluid by about 70-80% would produce this degree of dilution. But now, the issue becomes the source of this fluid. One possibility is the uptake of water into the body, indicated by the increase in body mass. But for the anoxic turtles, this amounted to only about 5-6% of the original body mass, and even if all this water remained within the extracellular fluid (0.24 l/kg), this would produce a dilution of only about 20%, well below the magnitude required for the observed concentration changes. Even this must be an overestimate, however, because water would likely distribute osmotically throughout the body water. A further problem with the dilution hypothesis is that the normoxic submerged turtles experienced an even greater apparent uptake of water from the surroundings (greater increase in body mass) and did not lose Na from their shells or bones, yet both plasma [Na⁺] and [Cl] fell far less than in the anoxic turtles. The resolution of this problem will require further studies involving measurements of fluid compartment volumes during anoxic submergence and a quantitative accounting of the distribution of these ions. Possible excretion of ions cannot be conclusively excluded at this time, although existing evidence indicates that renal function is minimal during submergence anoxia (16, 26).

This study also confirms the third hypothesis stated in the introduction that bone, as well as shell, accumulates lactate in proportion to the increase in plasma lactate concentration. Previous studies reported shell lactate accumulation at 20°C (16) and at 10°C and 3°C (11), but the 3°C measurements were restricted to turtles after 90 days of anoxia. We have also recently found (15) that bone takes up lactate during anoxic submergence at 20°C in both the painted turtle (C. picta bellii) and in the softshell turtle (Apalone spinifera). In addition, in vitro experiments in which shell powder was incubated in solutions with elevated lactate concentration indicated that the uptake of lactate is acid-base relevant and contributes to lactic acid buffering (13). Together, the shell and bone comprise 35-40% of the body mass, and therefore the uptake of lactic acid at a concentration in millimoles per kilogram that is nearly the same as the plasma concentration in millimoles per liter represents a very substantial contribution to whole body lactic acid buffering. In an earlier study (11), it was estimated that 44% of the total body lactate was sequestered (and buffered) within the shell. Utilizing the data from the present study that includes the bone, we raise this estimate to 47% of the total. This does not include the role of the shell and bone in exporting buffer to the extracellular fluid, which would bring the total contribution of the shell and bone to nearly 75% of the total body lactic acid buffering. These calculations illustrate clearly the crucial role of the shell and bone in this animal's long-term tolerance of anoxia.

Our data indicate that bone accumulates lactate even more effectively than shell (Fig. 3). This observation is of interest because of the uncertainty regarding the physical state of lactate within the bone. Because of the low water content of shell (26-28%), the lactate is unlikely to be in simple solution, since the concentration in that small volume of water would have to be extremely high. For example, in the present study in which final shell lactate concentration was 135.6 mmol/kg wet wt, lactate would have had to be at a concentration well over 500 mM, if it were restricted to the shell water. Therefore, it is probable that much of the lactate exists in combined form, perhaps complexed with calcium, as is known to occur in the plasma of these turtles during anoxic acidosis (14). This reasoning would suggest, however, that the more calcium present, the more lactate that could combine and be sequestered. Bone, however, contains more water and less calcium, yet holds more lactate. This observation complicates the interpretation of the lactate uptake phenomenon, and the resolution of this issue clearly requires further work. Another possibility, of course, is that the exchange kinetics between plasma and bone are more favorable and that this explains the higher levels in bone.

Perspectives

This study also reinforces the potential importance of aquatic O₂ to submerged, hibernating turtles. The maintenance of constant blood acid-base and ionic state during 100 days of submergence in aerated water indicates that painted turtles can avoid severe acidotic consequences if they select aerated microenvironments.