Stability Of Nucleoside Triphosphate Levels In The Red Cells Of The Snake Thamnophis Elegans

The red cells of many lower vertebrates contain high concentrations of nucleoside triphosphates (NTPs: usually in the form of adenosine triphosphate, ATP, and guanosine triphosphate, GTP) which serve, in large part, to regulate hemoglobin allosterically and therefore to modulate cell function (see, for example, reviews by Bartlett, 1980; Isaacks and Harkness, 1980). Concentrations of NTP are responsive to environmental perturbations (such as changes in oxygen availability or temperature) in numerous animals (see reviews by Wood, 1980; Weber and Jensen, 1988). In previous experiments to assess the effect of environmental oxygen tension on the oxygen-binding properties of the red cells of the garter snake (Thamnophis elegans), we exposed snakes to 5 %, 21 % and 100 % oxygen. Although the red cells showed decreases in NTP content with hypoxia and increases with hyperoxia, the relatively modest maximal effect required approximately 2 weeks to develop (Herman and Ingermann, 1996). In contrast to this rather slow responsiveness of red cell NTP levels to an environmental change, we found that the low levels of NTP in the red cells of neonatal garter snakes reach adult levels within 6 h of birth (Ingermann et al. 1991b). Although this event suggests a more rapid turnover of NTP (6 h versus 2 weeks), it is, nonetheless, markedly slower than is the case in typical respiring eukaryotic cells in which ATP turns over with a half-life of several minutes (e.g. Lehninger, 1975). Koszegi et al. (1987) reported that the majority of ATP in two types of cultured human cells is associated with proteins and is thus compartmentalized within the cell. Such an association is likely to be ionic and sensitive to pH, and conceivably this may contribute to the relatively slow change in snake red cell NTP levels in response to environmental perturbation. To assess the stability of the NTP pool in the red cells of the garter snake, we examined the effect of in vitro incubation, in the presence and absence of metabolic inhibitors, on NTP content as a function of time. In addition, since preliminary observations suggested that extremes of extracellular pH have a profound influence on intracellular NTP levels in snake red cells, we examined in more detail the effect of pH on intracellular NTP concentrations of red cells maintained in vitro.

Pregnancy in T. elegans is associated with an approximately 50 % rise in red cell NTP concentration and a concomitant decrease in red cell oxygen affinity (Ingermann et al. 1991a). Pregnancy in this species is associated with a rise in plasma progesterone concentrations (Highfill and Mead, 1975), and a more direct connection between red cell NTP levels and blood progesterone is indicated by the results of Ragsdale et al. (1993), who showed that implanting adult male T. elegans with progesterone-releasing capsules resulted in a rise in red cell NTP levels. The change in red cell NTP levels as pregnancy progresses or following implantation of progesterone capsules is relatively slow and requires weeks to reach maximal levels. It is possible, however, that the red cell response to progesterone is rapid, but that progesterone levels change slowly in these situations. To test this possibility, snakes were injected with progesterone and the levels of NTP in red cells were assayed 1 and 3 days later. Red cells were also incubated in vitro with progesterone to assess the possibility of an acute response leading to a change in the intracellular NTP level.

Garter snakes, Thamnophis elegans (Baird and Girard), were captured, identified and maintained as previously described (Ingermann et al. 1991a). Fetal snakes were at Zehr stages 35–37, where stage 37 is the last prenatal stage (Zehr, 1962). Blood was obtained from adults either by heart puncture under metofane anesthesia or by decapitation of animals rendered unconscious by a blow to the head. Blood was collected into ice-cold heparinized buffer A (in mmol l^{− 1}): 143 NaCl, 3 KCl, 1.5 MgCl₂, 1.5 CaCl₂, 20 Tris, adjusted to pH 7.4 with HCl (except where indicated). Fetal snakes were excised over ice, rinsed in buffer A, sectioned in ice-cold heparinized buffer A and the cell suspension was passed through glass wool. Fetal and adult red cell suspensions were washed three times by centrifugation at 1200 g for 10 min at 4 °C. The third wash was with cold buffer A containing 50 i.u. ml^{− 1} penicillin and 50 μg ml^{− 1} streptomycin, and the final red cell pellet was resuspended in this buffer with antibiotics. 10 % by volume of potassium cyanide (KCN) and/or iodoacetic acid (IAC) stock solution (in buffer A containing penicillin and streptomycin) was added to samples of the resuspended red cells to give final concentrations of 0.1 mmol l^{− 1} KCN and/or 0.5 mmol l^{− 1} IAC. Owing to limited availability, fetal red cells were not exposed to IAC plus KCN. Final red cell suspensions with and without metabolic inhibitors were then placed into glass test tubes, 0.3 ml for fetal cells and 0.4 ml for adult cells. These samples were kept for varying periods of up to 96 h in a humidified chamber at 20 °C. For analyses, 1 ml of buffer A (without antibiotics) was added to each sample, these were centrifuged in a Fisher model 235B microcentrifuge for 2 min, buffer was drawn off, and fresh buffer A was added. The cells were resuspended and the suspension was analyzed for hematocrit by centrifuging for 4 min in an IEC microhematocrit centrifuge. Tetrameric hemoglobin (Hb) concentration was determined spectrophotometrically by adding a sample of the suspension to a larger volume of Drabkin’s solution and calculating the concentration using a millimolar extinction coefficient of 44.0 at 540 nm. Finally, a sample of suspension was extracted with an equal volume of ice-cold 12 % (w/v) trichloroacetic acid. The nucleoside triphosphate concentration of the extract was then analyzed with an enzymatic assay kit (number 366, Sigma Chemical Co., St Louis, MO), which does not distinguish among the different NTPs. Total NTP level was expressed as a molar concentration within the red cells or as mole NTP per mole Hb. From the suspension hematocrit and Hb concentration, the mean corpuscular hemoglobin concentration (MCHC) was also calculated. Based on the measurement of free Hb in the suspension fluid, there was minimal lysis (⩽0.5 %) by the end of the 96 h incubation.

For analysis of the influence of pH on red cell NTP levels, red cells from nonpregnant females were first washed three times in buffer A (without antibiotics). They were subsequently washed three additional times in the Tris-based buffer A (titrated to the desired pH within the pH range 7.2–9.0) or in modified buffer A where Tris had been replaced by BisTris or Capso: 20 mmol l^{− 1} BisTris for pH ⩽7.2 or 20 mmol l^{− 1} Capso for pH ⩾9.0. Osmotic concentrations (in mosmol kg^{− 1}) of the buffers were 286±1 for pH 6.0, 288±4 for pH 7.4 and 284±5 for pH 10.0 (mean ± S.D., N=3–5) and were measured using a 3WII Advanced Laboratory wide-range osmometer (Advanced Instruments, Needham Heights, MA). One series of experiments at low and high pH was conducted in the presence of 1.0, 3.3 and 10.0 mmol l^{− 1} ouabain. Samples of the final suspension were incubated for 24 h at 20 °C, the pH of the suspension was determined, and the cells were then washed in buffer and analyzed for NTP and Hb concentrations. Lysis never exceeded 5 %. Incubations at both low and high pH appeared to result in some cell swelling and, therefore, red cells were incubated in hypotonic buffer to establish whether swelling alone could account for the effects of low or high pH on intracellular NTP levels. For these experiments, red cells were incubated for 24 h in media which were 80 %, 60 % or 40 % buffer A (titrated to pH 7.4) in distilled H₂O. Lysis under these conditions never exceeded 7 %.

For analysis of ATPase activity, red cells were collected from nonpregnant females, washed three times in cold buffer A, suspended in diluted buffer A (15 % buffer A, 85 % water), frozen, thawed and centrifuged at 4 °C for 20 min at 20 000 g. The pellet was resuspended in diluted buffer A and washed twice more at 20 000 g. The final pink pellet was resuspended in a small volume of dilute buffer A and samples were taken for protein analysis by the method of Markwell et al. (1976) using commercial bovine serum albumin as a standard. Samples were also taken to measure nonspecific ATPase activity (assayed by a modification of the method of Zaugg and McLain, 1970). To a volume of cell membrane suspension in diluted buffer A was added an equal volume of a buffer containing (in mmol l^{− 1}): 300 sucrose, 20 EDTA, 24 sodium deoxycholate and 100 Tris, adjusted to pH 7.4 with HCl. This mixture was drawn through a 30 gauge needle six times, kept on ice, and assayed within 3 h. Samples were not centrifuged. Approximately 0.03 ml of the mixture was added to a tube in an ice bath containing (1) 0.65 ml of a buffer consisting of (in mmol l^{− 1}): 23 MgCl₂, 86 NaCl, 75 KCl and 115 Tris (for pH 7.4), 115 BisTris (for pH 6.0) or 115 Capso (for pH 10.0) and (2) 0.1 ml of 30 mmol l^{− 1} ATP (titrated to a pH of approximately 7.0). The tube was placed into a water bath at 30 °C for 30 min and then returned to the ice bath, where 2.25 ml of distilled water, 1.0 ml of 15.5 mmol l^{− 1} ammonium molybdate in 3 mol l^{− 1} HCl, 0.85 ml of 7.5 % (w/v) sodium dodecyl sulfate and 0.15 ml of 0.25 % (w/v) Fisk Subbarrow reducer were added. The mixture was then warmed to 20 °C and the absorbance of the solution read 40 min later at 700 nm. Absorbance of the unknown samples was related to phosphate standards run simultaneously.

The stability of ATP in buffer A at pH 5.8, 7.4 and 10.2 was assessed by keeping 1 mmol l^{− 1} ATP in buffer at 20 °C for 24 h. Initial and final concentrations were measured using the NTP enzymatic assay kit.

To assess whether red cell NTP levels could respond acutely to a high concentration of progesterone in vivo, males were injected with 0.1 mg of progesterone (or a comparable molar amount of cholesterol) in 10 ml of 5 % ethanol per kilogram animal mass. This dose was calculated to yield approximately 20 times the maximal progesterone levels found in T. elegans by Highfill and Mead (1975). Eight cholesterol-injected and eight progesterone-injected animals were maintained on a 12 h:12 h light:dark cycle with an electrically heated rock for 24 or 72 h. They were subsequently anesthetized and blood samples taken by heart puncture. In addition, to establish whether red cell NTP levels were acutely sensitive to a high progesterone concentration in vitro, washed red cells from three males were incubated in pH 7.4 buffer containing 5 mmol l^{− 1} glucose, 5 mmol l^{− 1} pyruvate and 5 mmol l^{− 1} glutamine with 1 μmol l^{− 1} cholesterol or progesterone (and 0.5 % ethanol). After 24 h at 20 °C, cells were washed once in pH 7.4 buffer and red cell [NTP]/[Hb] ratios determined.

All data are presented as mean ± S.D. Slopes of [NTP]/[Hb] ratios versus time in the in vitro incubation studies were analyzed by one-way ANOVA followed by Scheffe’s F-test. The influence of metabolic inhibitors on fetal [NTP]/[Hb] ratios was analyzed with an unbalanced randomized-block ANOVA with a 3×3 factorial. To test the influence of pH on ATPase activity, randomized-block ANOVAs were followed by either protected LSD or one-sample t-test analyses. P<0.05 was considered significant.

All reagents were obtained from Sigma Chemical Co.

Incubation of red cells from adult snakes in vitro at 20 °C resulted in a decrease in the [NTP]/[Hb] ratio over a period of 4 days (Figs 1A–C, 2). However, red cells incubated in vitro in the absence of metabolic inhibitors showed particularly low rates of decline in intracellular [NTP]/[Hb] ratio. For example, after 96 h, [NTP]/[Hb] ratio in red cells of nonpregnant females had decreased to 89±9 % of initial values (mean ± S.D., N=5), while in pregnant females the level had fallen to 81±8 % (mean ± S.D., N=5). Metabolic inhibitors prompted a more pronounced reduction in [NTP]/[Hb] values over the same period so that, in the presence of IAC plus KCN, the 96 h value for nonpregnant females was 67±6 % and for pregnant individuals 57±12 % of initial values (mean ± S.D., N=5 for each group). As indicated by the rates (slopes) of the [NTP]/[Hb] decline, red cells from the pregnant female and male were sensitive to IAC but not to KCN (Fig. 2). In contrast, the red cells of the nonpregnant female were more sensitive to KCN than to IAC (Fig. 2).

There were no significant differences in [NTP]/[Hb] ratios for fetal red cells incubated for 6 or 24 h in the absence of inhibitors or with KCN or IAC (Table 1). Nonetheless, there did appear to be a trend suggesting that these cells were sensitive to KCN (and less so to IAC) prior to 6 h; however, by 24 h, they appeared to be insensitive to KCN but sensitive to IAC. From 48 to 96 h (72 and 96 h data not shown), for reasons that are not clear, scatter in the data precluded any possible interpretation.

Levels of NTP were appreciably more labile when red cells of adults were maintained in vitro at either low or high pH (Fig. 3A,B). Maintenance of cells in buffer at pH 6.0 for 24 h resulted in a decrease in the NTP level of approximately 50 % and at pH 10.0 the decrease was approximately 80 %. In contrast, maintaining commercial ATP in low-and high-pH buffers for 24 h caused no detectable decrease in the ATP concentration: at 24 h, ATP kept at pH 5.8 was 101.6±2.5 % of the initial concentration, while at pH 7.4 it was 102.9±1.7 % and at pH 10.2 it was 106.6±5.8 % (mean ± S.D., N=3 for each group). Total, nonspecific, membrane ATPase activity at pH 6.0 was approximately 55 % greater than that at pH 7.4, while this activity at pH 10.0 was reduced to approximately 6 % of that at pH 7.4 (Table 2).

Mean corpuscular hemoglobin concentrations (MCHC) appeared to decrease, particularly at low pH, suggesting pH-dependent cell swelling (Fig. 4). To determine to what extent cell swelling might alter the [NTP]/[Hb] value, red cells were incubated for 24 h in dilute buffer A at pH 7.4. Red cells maintained in 80 % buffer A contained 99.4±1.5 % of the [NTP]/[Hb] value of those maintained in 100 % buffer A, while at 60 % buffer A this value was 90.0±10.1 % and at 40 % buffer A it was 78.0±16.1 % (mean ± S.D., N=3 for each group).

Cells maintained at pH 6.3±0.1 in 20 mmol l^{− 1} BisTris buffer showed a 51.5±3.6 % (mean ± S.D., N=4) decline in intracellular [NTP]/[Hb] ratio by 24 h. The presence of ouabain attenuated this decline, with 1.0 mmol l^{− 1} ouabain reducing the effect by 13.2±6.8 %, 3.3 mmol l^{− 1} by 14.8±7.4 % and 10.0 mmol l^{− 1} by 19.9±8.6 % (mean ± S.D., N=3 for each concentration). Cells maintained at pH 9.5±0.1 in 20 mmol l^{− 1} Capso buffer showed a decline in intracellular [NTP]/[Hb] ratio of 49.9±9.2 % (mean ± S.D., N=4). The presence of ouabain had no noticeable effect, 1 mmol l^{− 1} ouabain altered the decline by − 2.6±5.5 %, 3.3 mmol l^{− 1} by 4.5±8.7 % and 10.0 mmol l^{− 1} by 3.0±10.1 % (mean ± S.D., N=4 for each concentration). Ouabain at 10 mmol l^{− 1} had no significant effect on the red cell [NTP]/[Hb] ratios of cells maintained for 24 h at pH 7.4 (data not shown).

Injection of T. elegans with progesterone did not result in a rise in red cell NTP concentration at either 24 or 72 h. The red cell [NTP]/[Hb] value at 24 h for animals injected with 0.1 mg kg^{− 1} of animal progesterone was 2.25±0.18, and for animals injected with the same concentration of cholesterol it was 2.26±0.13. At 72 h, these values were 2.28±0.17 for progesterone-treated snakes and 2.52±0.23 for cholesterol-treated snakes (mean ± S.D., N=4 for each group). Similarly, incubation of red cells in vitro with exogenous nutrients and progesterone for 24 h did not result in an increase in red cell [NTP]/[Hb] ratio relative to cells incubated with nutrients and cholesterol, as the [NTP]/[Hb] ratio of progesterone-treated cells was 101.6±3.7 % (N=3) of that of cholesterol-treated cells. Cells incubated for 24 h at pH 7.4 without or with 5 mmol l^{− 1} glucose plus 5 mmol l^{− 1} pyruvate and 5 mmol l^{− 1} glutamine showed decreases in the [NTP]/[Hb] ratio of 8.2± 5.2 % in the absence of additional energy sources and 5.1±7.3 % when these were present (mean ± S.D., N=3 for each group), respectively. These values were not significantly different as assessed by a Student’s t-test.

The red cells of the killifish Fundulus heteroclitus, when maintained in vitro without exogenous nutrients at room temperature, show no decline in red cell ATP concentration over 4–8 h (Greaney and Powers, 1978). Similarly, red cells of Squalus acanthias show no decline in ATP concentration over 18 h when maintained at 10 °C without nutrients (Wells and Weber, 1983). Consistent with these findings is the very slight (if any) decrease in cellular NTP levels noted in the present study with red cells from the garter snake (Fig. 1A–C). In the presence of KCN or IAC, the level of ATP in F. heteroclitus red cells shows a 40 % decline over 4 h and, under anoxia, red cells of S. acanthias show a 30–35 % decrease in ATP levels over 18 h. In contrast to these findings, the use of a metabolic block resulted in a slower decline in NTP levels in red cells from adult T. elegans; total NTP levels in these cells appeared particularly stable. For unknown reasons, NTP levels in red cells from fetal snakes appeared less stable than those in red cells from adult snakes. Despite the red cells of snakes being nucleated, having mitochondria and generating ATP by oxidative phosphorylation (Ogo et al. 1993), 0.1 mmol l^{− 1} KCN had little or no effect on the red cells from the adult male and pregnant female. It did, however, have a demonstrable effect on the cells from the nonpregnant female. In contrast, 0.5 mmol l^{− 1} IAC did appear to be an effective inhibitor of ATP synthesis, at least in the red cells of the males and pregnant females, suggesting that normal ATP turnover in these cells is primarily glycolysis-based.

Since pregnancy in T. elegans is associated with a rise in NTP levels of the adult red cell (Ingermann et al. 1991a), we looked at the possibility that the turnover of the additional NTP associated with pregnancy is different from that of the basal levels in the red cells of the male and nonpregnant female. NTP levels of the red cells of the pregnant animal did indeed decrease more rapidly in the absence, as well as in the presence, of KCN+IAC than did those of the cells of the nonpregnant female. This suggests that the pregnancy-associated NTP distributes into intracellular pools differently from the smaller amount of NTP in the nonpregnant female.

Previous findings have suggested that the pregnancy-associated rise in red cell NTP levels was due to, or at least correlated with, a chronic rise in plasma progesterone levels (Ragsdale et al. 1993). To test whether the red cells themselves could respond quickly to elevated progesterone levels, animals were injected with progesterone or red cells were incubated in vitro with progesterone. In neither case was an acute change noted in red cell NTP values. Therefore, it appears that these cells do not respond to progesterone directly and/or that they cannot respond quickly to this potential stimulus.

The pronounced stability of total NTP levels in the red cells of the snake, in the absence or presence of metabolic inhibition, appears inconsistent with ATP having a half-life of minutes, as in other eukaryotic cells (e.g. Lehninger, 1975). However, ATP has been reported to bind to various proteins; it binds to lamin C (Schwartz and Clawson, 1991), serum albumin (Bauer et al. 1992) and α-crystallin (Reddy et al. 1992), and Koszegi et al. (1987) have suggested that in several cell types the bulk of cellular ATP is associated with proteins rather than being in a soluble form that is uniformly distributed in the cytoplasm. Conceivably, this may represent a possible resolution to the paradox of the stability of NTP levels in the snake red cell. The apparently low rate of NTP turnover in the snake red cell may be due to a compartmentalization of NTP with cellular proteins or membranes. Compartmentalization of ATP appears to account for the high ATP content of mammalian platelets (Ugurbil et al. 1979) and has been reported in the human red cell (Gupta et al. 1978; Petersen et al. 1989, 1990). Perhaps such an association in the snake red cell renders NTP relatively unavailable for hydrolysis by nucleoside-triphosphate-consuming processes.

Since NTP is highly charged at physiological pH, any association with intracellular proteins (or membranes) would probably be ionic. Extremes of pH should therefore disrupt these putative interactions and possibly make ATP accessible to intracellular ATPase activity. As indicated in Fig. 3A,B, 24 h of incubation at low or high pH (in the absence of metabolic inhibition) resulted in a pronounced and relatively rapid decrease in red cell NTP levels. That incubation of commercial ATP at low and high pH did not result in any detectable change in concentration suggests that ATP was not itself unusually labile under these conditions. This is consistent with the reported inherent stability of the nucleoside triphosphates (Miller and Westheimer, 1966; Westheimer, 1987).

The decrease in the level of NTP in red cells noted at extreme pH could have been due to osmotic challenges to the cell; however, incubation in hypo-osmotic buffer at pH 7.4 resulted in no comparable decrease in red cell NTP levels after 24 h. The decrease in NTP concentration could have been due to activation of ATPase activity. Indeed, ATPase activity at low pH was greater than that at pH 7.4. Thus, the decrease in red cell NTP levels after 24 h at low pH could have been due to direct stimulation of ATPase activity under these incubation conditions. In contrast, increased ATPase activity could not have accounted for the marked reduction of intracellular NTP levels after incubation at high pH, as nonspecific red cell ATPase activity at high pH was only approximately 6 % of that at pH 7.4. Therefore, the decrease in red cell NTP levels during incubation at high pH does not appear to have been due to an inherent instability in ATP at this pH, to osmotic perturbation or to an enhanced total nonspecific ATPase activity. A net reduction in total NTP at the pH extremes could have resulted from an increased availability of NTP to a variety of intracellular consuming processes and/or to cellular export (Forrester, 1990; Bergfeld and Forrester, 1992).

Overall, it appears that total NTP levels of the red cells of T. elegans are very stable in vitro, relatively insensitive to metabolic inhibition and completely insensitive to short-term progesterone exposure, but are labile under extremes of pH. These results are consistent with the hypothesis that intracellular NTP is compartmentalized into at least two pools within these snake red cells: a labile, cytoplasmic pool and a more stable, protein- or membrane-associated pool.

We are grateful to Dr Kirk Steinhorst and Dr Barbara Sorg for valuable comments and suggestions. This study was supported in part by a grant from the Idaho State Board of Education.