Calcium Stimulates ATP-Mg / Pi Carrier Activity in Rat Liver Mitochondria ”

Adenine nucleotide transport over the carboxyatractyloside-insensitive ATP-Mg/P, carrier was assayed in isolated rat liver mitochondria with the aim of investigating a possible regulatory role for Ca2+ on carrier activity. Net changes in the matrix adenine nucleotide content (ATP + ADP + AMP) occur when ATP-Mg exchanges for Pi over this carrier. The rates of net accumulation and net loss of adenine nucleotides were inhibited when free Ca’+ was chelated with EGTA and stimulated when buffered [Ca’+& was increased from 1.0 to 4.0 FM. The unidirectional components of net change were similarly dependent on Ca’+; ATP influx and efflux were inhibited by EGTA in a concentrationdependent manner and stimulated by buffered free Ca2+ in the range 0.6-2.0 PM. For ATP influx, increasing the medium [Caz+]r,ee from 1.0 to 2.0 PM lowered the apparent K,,, for ATP from 4.44 to 2.44 mM with no effect on the apparent V,,,., (3.55 and 3.76 nmol/ min/mg with 1.0 and 2.0 NM [Ca2+lfree, respectively). Stimulation of influx and efflux by [Ca2’lP,,, was unaffected by either ruthenium red or the Ca2+ ionophore A23 187. Calmodulin antagonists inhibited transport activity. In isolated hepatocytes, glucagon or vasopressin promoted an increased mitochondrial adenine nucleotide content. The effect of both hormones was blocked by EGTA, and for vasopressin, the effect was blocked also by neomycin. The results suggest that the increase in mitochondrial adenine nucleotide content that follows hormonal stimulation of hepatocytes is mediated by an increase in cytosolic [Ca2+Le that activates the ATP-Mg/Pi carrier.

and hibernation (15). It has been suggested that the ATP-Mg/Pi carrier of the mitochondrial inner membrane is responsible for these net changes in adenine nucleotide content because adenine nucleotide transport over this carrier can occur as a counterexchange with Pi resulting in a net increase or a net decrease in the matrix adenine nucleotide pool size (1, 16).
The ATP-Mg/P, carrier activity that facilitates net accu-* This work was supported by National Institutes of Health Grant HD16936.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. $ Submitted this work in partial fulfillment of the requirements for the Ph.D. in Biology at Tufts University. § To whom correspondence should be sent. mulation and net loss of mitochondrial adenine nucleotides in vitro has been characterized in detail (1,(16)(17)(18)(19).
The direction and magnitude of net transport are determined primarily by the ATP-Mg concentration gradient across the inner membrane and its deviation from equilibrium with the Pi concentration gradient (1). A recent report has indicated that in isolated mitochondria, net changes in adenine nucleotide content are stimulated by calcium (20). prevented net accumulation and net loss of adenine nucleotides (Fig. 1). In the presence of 1.0 mM ATP, which is close to a steady-state condition (19), very little change in adenine nucleotide content occurred, and this result was unaffected with either 1 or 100 PM EGTA (Fig. 1). Mitochondrial NAD(H) was determined for the experiment shown in Fig. 1  by Ca'+ has been reported previously (20) but over a [Ca'l+]trt,c range of 0.4-1.0 pM. We were unable to detect net changes when [Ca'+],,,.,. was less than 0.6 pM. However, it is interesting to note that in our initial experiments with Ca-EGTA buf'f'ers (1, 29), the pH of' the assay medium was adjusted before the addition of Ca')+ and thus the pH was accidentally lowered to an extent that was dependent on the total [Ca'+]. As a result there was an initial shock to the mitochondria in which the matrix adenine nucleotides were suddenly lost, dropping to about 4 nmol/mg protein within 30 s (a normal init.ial content is approximately 14-15 nmol/mg prot,ein). This loss was preventable by including ruthenium red. Following the loss there was a recovery of the adenine nucleotide content by net accumulation over several minutes in a manner that was dependent on the concentration of Ca" which. if pH ivere 7.4, was supposed to be in the range of 0.1 to 0.8 PM. The actual concentrations of f'ree Ca'-under these conditions. however, were not known because the binding of Ca'-with EGTA is sensitive to PH. These unusual conditions were the only circumstances under which we observed any net uptake that was dependent on [Ca"],ccL less than 0.6 PM.

Effects of Extranzitochondrial
Ca" on I'n~d~wct~onal A7?'P Influx and A7'P E:ff/lus-Net transport of adenine nucleotides across the inner membrane via the ATP-Mg/P, carrier has been shown to be equal in magnitude to the dif'f'erence between ATP influx and ATP efflux (16). In order to understand the role of [Ca-'+],,,,. on net changes of'adenine nucleotide content it was necessary to examine its ef'fects on the unidirectional fluxes. For these experiments, incubations contained 1.0 m&f ATP so that influx and ef'flux would be approximately equal. producing a steady state in which no net change would occur (16,19). In the absence of EGTA, addition of'extra CaCIL up to 10 pM had no ef'f'ect on the initial rates of ATP influx or ATP ef'flux (data not shown), but when medium [Ca'-lIr'< (normally 3.28 PM, see ahove) was chelated with ECTA influx and efflux were inhibited (Fig. 3). Inhibition was equivalent f'or hoth processes and was dependent on the concentration of EGTA between 2.5 and 10 pM (Fig. 3).   with Ca-EGTA buffers in the range of O-4.0 pM. ATP influx and ATP efflux were stimulated in a concentration-dependent manner between 0 and 2.0 pM. Influx, but not efflux, was further stimulated by 4.0 ELM [Ca2+laee (Fig. 4) (Fig. 5). The values for apparent K, and V,,, obtained here in Ca-EGTA buffers are in the range of those reported previously (16).
Other Characteristics of Ca'+-stimulated ATP Fluxes-Addition of 1 FM ruthenium red had little or no effect on the rates of ATP influx or efflux when [ATP] was 1.0 mM and [Ca2+]rree was buffered at 1.0 or 4.0 pM (Table I). When [Cal+] rree was buffered at 0.6 or 4.0 FM the addition of A23187 (2 pg/ml) to promote Ca'+ entry into the matrix did not stimulate ATP influx or efflux (data not shown). Dibutyryl cyclic AMP (250 ELM) in the presence of low extramitochondrial [Caa+]rree Procedures." In Experiment I, pretreatment with EGTA (3.1 mM) was for 30 s, and in Experiment II, 10 IIlM neomycin pretreatment was for 5 min prior to addition of hormone (either glucagon (1 nM) or vasopressin (10 nM)). After 5 min aliquots were subjected to digitonin fractionation, and the adenine nucleotides were measured in the mitochondrial fraction. Controls were samples from the untreated cell suspension and represent the average (+S.E.) of replicates for each experiment. (0.6 pM) had no effect on unidirectional ATP influx (data not shown).
Stimulation of the ATP-Mg/Pi carrier may be mediated by a Ca'+-binding site on the carrier, by a mitochondrial calmodulin, or by a calmodulin-like protein associated with the carrier. Calmodulin is too large to cross the outer mitochondrial membrane, and whether or not there is calmodulin associated with mitochondria is still being debated. In any case, Ca'+-stimulated processes often can be inhibited by calmodulin antagonists whether or not the stimulation is mediated by calmodulin.
Several calmodulin antagonists that we tested in ATP influx and efflux assays were found to inhibit calcium-dependent ATP-Mg/P, carrier activity (Table  II). The percent inhibition was variable for trifluoperazine, chlorpromazine, or calmidazolium, but each antagonist inhibited both ATP influx and efflux to a similar extent. Trifluoperazine was studied in more detail and found to inhibit ATP influx as a linear function of concentration between 10 and 100 jtM when [Ca2+]rree was unbuffered at 3.28 pM (Fig. 6). The percent inhibition was similar at all inhibitor concentrations whether or not oligomycin was present (Fig. 6) showing that the effect of trifluoperazine was not secondary to a lower matrix ATP/ADP ratio (see below). Chlorpromazine inhibited ATP influx in a similar concentration-dependent manner (data not shown) but was less potent than trifluoperazine at all concentrations tested. The inhibitory effect of trifluoperazine was further tested in both ATP influx and ATP efflux assays over a range of [Ca2+lrree (Fig. 7). Trifluoperazine inhibited both fluxes similarly; in general, the percent inhibition was independent of [Ca2+lfree.
Mitochondrial respiration was measured under similar assay conditions with the concentrations of trifluoperazine and chlorpromazine reported in Table II The inhibition of state 3 that we observed may have been due in part to inhibition of electron transport (uncoupled respiration also was inhibited), but trifluoperazine is known also to inhibit FOF1-ATPase activity (30). This raised the possibility that inhibition of transport might be secondary to a decrease in the matrix ATP/ADP ratio. To clarify this issue unidirectional ATP influx was measured in the presence of oligomycin. Oligomycin maximally lowers the matrix ATP/ ADP ratio so that possible inhibition of ATP flux by calmodulin antagonists could be assessed independently of any effect on the matrix ATP content. In the presence of oligomycin ATP influx was inhibited by trifluoperazine, chlorpromazine, and W-7 to approximately the same extent as with no oligomycin; with calmidazolium, inhibition was greater when oligomycin was included ( Fig. 6 and  for cellular regulation of the mitochondrial adenine nucleotide content, we studied the effects of glucagon and vasopressin in isolated hepatocytes. Treatment of hepatocytes with glucagon has previously been shown to cause an increase in intramitochondrial adenine nucleotides (9,11,13), whereas a similar effect of vasopressin has not as yet been reported. Incubation of isolated hepatocytes with glucagon or vasopressin increased the adenine nucleotide content in the mitochondrial fraction (Table  III). The total adenine nucleotide content of the cells did not change (not shown). In the presence of an amount of EGTA (3.1 mM) sufficient to lower extracellular [Ca2+]rree to approximately 1 pM, the movement of adenine nucleotides from the cytoplasm into the mitochondrial fraction that was caused by either hormone was prevented (Table III). A 5-min incubation of hepatocytes with dibutyryl cyclic AMP (250 pM) caused a 28% increase in the mitochondrial adenine nucleotide content relative to vehicle treated hepatocytes (11.26 f 2.48 uersus 8.82 f 1.74 nmol/106 cells; p < 0.05, paired t test). Pretreatment of hepatocytes with 10 mM neomycin (a potent inhibitor of the inositol phosphate pathway) for 5 min prevented the increase in mitochondrial adenine nucleotides observed with vasopressin (Table III) (20) for net changes except that the effective concentration ranges of Ca*+ stimulation are somewhat different. We were unable to detect net changes when [Ca2+lrree was less than 0.6 pM. This may reflect differences in methods used to set up the incubations (see our comments under "Results") or in the method used to determine [Ca2+lrree; we relied on Fura-2, whereas Haynes et al. (20) (32), caused an increase in the mitochondrial adenine nucleotide content.
For vasopressin, an increase in intracellular [CaP+lfree is mediated by inositol polyphosphates (34) but not by cyclic AMP (35). Neomycin, shown to decrease the intracellular [CaP+lfree caused by vasopressin (321, also completely prevented vasopressin-induced adenine nucleotide accumulation by mitochondria in hepatoeytes. These results are consistent with the hypothesis that the net increase in mitochondrial adenine nucleotide content that follows glucagon or vasopressin administration may occur because of a change in cytosolic [Caz+lfree which regulates the ATP-Mg/Pi carrier. An increase in cytosolic [Caz+lfee may translate into an increase in matrix [Ca2c]free as well (36-38), but for ATP-Mg/ Pi carrier stimulation an increase in cytosolic [Ca2+lrree is probably sufficient.
Adenine nucleotide recompartmentation between the cytosol and mitochondria occurs as an adaptive response to changing physiological conditions, with important effects on metabolic activity (2-8, 12, 13). The shift in adenine nucleotides from the cytosol to the mitochondria observed at parturition in the rat and rabbit has already been shown to be affected by the changing hormonal status (increasing glucagon/insulin ratio) of the newborn animal (7,39,40). This is probably related to an increase in cytosolic [Ca2+lme which may now be presumed to stimulate the ATP-Mg/Pi carrier. In the normal newborn and in normoxic adults, activation of the carrier normally results in net uptake of adenine nucleotides into the mitochondria, because the normal cytoplasmic and matrix ATP concentrations favor net movement in that direction (1). If the cytoplasmic ATP concentration falls to very low levels, as it does in hypoxia, less uptake or even net loss of adenine nucleotides from mitochondria is predicted to occur whenever Ca*+ is present to activate the carrier. This prediction is consistent with published observations (7,14,40). Further investigations of cellular mechanisms of adenine nucleotide recompartmentation that occur via the ATP-Mg/ Pi carrier will require careful evaluation of both cytosolic [Ca2+lfr.. and ATP concentration gradients across the inner mitochondrial membrane.