Guanine nucleotide transport by atractyloside-sensitive and -insensitive carriers in isolated heart mitochondria

Transcription

1 Am J Physiol Cell Physiol 279: C1870 C1879, Guanine nucleotide transport by atractyloside-sensitive and -insensitive carriers in isolated heart mitochondria EDWARD E. MCKEE, ALICE T. BENTLEY, RONALD M. SMITH, JR., JONATHAN R. KRAAS, AND CHRISTINA E. CIACCIO South Bend Center for Medical Education, Indiana University School of Medicine, University of Notre Dame, Notre Dame, Indiana Received 17 February 2000; accepted in final form 17 July 2000 McKee, Edward E., Alice T. Bentley, Ronald M. Smith, Jr., Jonathan R. Kraas, and Christina E. Ciaccio. Guanine nucleotide transport by atractyloside-sensitive and -insensitive carriers in isolated heart mitochondria. Am J Physiol Cell Physiol 279: C1870 C1879, In previous work (McKee EE, Bentley AT, Smith RM Jr, and Ciaccio CE, Biochem Biophys Res Commun 257: , 1999), the transport of guanine nucleotides into the matrix of intact isolated heart mitochondria was demonstrated. In this study, the time course and mechanisms of guanine nucleotide transport are characterized. Two distinct mechanisms of transport were found to be capable of moving guanine nucleotides across the inner membrane. The first carrier was saturable, displayed temperature dependence, preferred GDP to GTP, and did not transport GMP or IMP. When incubated in the absence of exogenous ATP, this carrier had a V max of pmol mg 1 min 1 with a K m of mm for GDP. However, transport of GTP and GDP on this carrier was completely inhibited by physiological concentrations of ATP, suggesting that this carrier was not involved with guanine nucleotide transport in vivo. Because transport on this carrier was also inhibited by atractyloside, this carrier was consistent with the well-characterized ATP/ADP translocase. The second mechanism of guanine nucleotide uptake was insensitive to atractyloside, displayed temperature dependence, and was capable of transporting GMP, GDP, and GTP at approximately equal rates but did not transport IMP, guanine, or guanosine. GTP transport via this mechanism was slow, with a V max of pmol mg 1 min 1 and a K m mm. However, because the requirement for guanine nucleotide transport is low in nondividing tissues such as the heart, this transport process is nevertheless sufficient to account for the matrix uptake of guanine nucleotides and may represent the physiological mechanism of transport. bioenergetics; nucleotide metabolism; biogenesis THE PRESENCE OF GUANINE NUCLEOTIDES in the matrix of mitochondria is uncontested, although the mechanisms that account for their presence in the matrix are not known. In previous work (8), we demonstrated that radioactivity in the matrix pool of guanine nucleotides in heart mitochondria arose exclusively from transport of radioactive GMP, GDP, and/or GTP across the inner membrane, rather than arising via conversion from inosine or adenine compounds, or via uptake and synthesis from the salvage pathways using guanine or guanosine. The addition of a single labeled form of GMP, GDP, or GTP to isolated heart mitochondria resulted in the redistribution and equilibration of the label among all three forms of the nucleotide. Thus the specific form or forms of the nucleotide actually transported have not been determined. To date, only two mechanisms for the transport of nucleotides across the inner mitochondrial membrane have been characterized. Both are involved in the transport of ATP and ADP (1, 2, 6). The first is the ADP/ATP translocase, a major protein of the inner membrane, which catalyzes the electrogenic, antiport exchange of ADP and ATP across the membrane (for review, see Ref. 6). In active mitochondria, the net activity of this carrier is to promote the exchange of medium ADP for matrix ATP. This exchange is extraordinarily rapid (ms) and has been well characterized. It is completely inhibited from the medium side by atractyloside and from the matrix side by bongkrekic acid (6). The second adenine nucleotide carrier, at a much lower activity in the membrane than the first, promotes the net uptake or loss of matrix ADP/ATP, typically in exchange for inorganic phosphate (for review, see Refs. 1 and 2). This carrier, called the MgATP/phosphate exchange carrier, has been characterized in liver and kidney mitochondria and is insensitive to atractyloside and bongkrekic acid (1, 2). This latter carrier is important in maintaining and regulating matrix adenine pools, particularly during growth, and as such might be a prototype for a carrier functioning in guanine nucleotide transport. However, neither of the characterized adenine carriers is thought to have any role in the transport of other nucleotides (1, 2, 6). Prior research (8) demonstrated that the matrix pool of guanine nucleotides arose from the uptake of one or more of the phosphorylated forms of guanine nucleotide. The first goal of this work was to quantitate the distribution of label among the three phosphorylated Address for reprint requests and other correspondence: E. E. McKee, South Bend Center for Medical Education, B15 Haggar Hall, Univ. of Notre Dame, Notre Dame, IN ( The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. C /00 $5.00 Copyright 2000 the American Physiological Society

2 C1871 forms of guanine nucleotide under several conditions and to correlate the label distribution to rates of label uptake. The second goal of this work was to characterize the mechanisms of this transport process. These goals were accomplished by using a vacuum filtration method to measure the rate of uptake of labeled guanine nucleotides into the matrix of isolated intact rat heart mitochondria under a variety of conditions. EXPERIMENTAL PROCEDURES Isolation and incubation of heart mitochondria. Tightly coupled mitochondria were isolated from hearts of control rats (Harlan Sprague Dawley) exactly as described previously (9). Intactness was monitored with an oxygen electrode by determining the respiratory control ratio using glutamate as the substrate (9). Only mitochondria with respiratory control values greater than six were used in the transport experiments. Isolated heart mitochondria were incubated at 30 C in a medium originally defined for mitochondrial translation experiments (9) containing mitochondria at 4 mg protein/ml in (in mm) 25 MOPS buffer, ph 7.2, 90 KCl, 2 magnesium sulfate, 5 potassium phosphate, 0.4 EGTA, 44 mannitol, 14 sucrose, BSA (1 mg/ml), 2 4 ADP or ATP, 20 glutamate, and 0.1 each of the 20 amino acids. Additions and changes in this standard medium are outlined where appropriate. HPLC analysis of purine nucleotides. Purine nucleotides from isolated heart mitochondria were extracted, identified, and quantitated from samples by reverse-phase HPLC over an Alltech nucleotide/nucleoside column as described previously (8). All guanine and adenine nucleotides are separated and identified by this method (8). Guanine nucleotide transport: vacuum filtration. In earlier work, the method of medium and matrix separation via centrifugation through cold silicone oil worked well to obtain sufficient samples for nucleotide analysis by HPLC (8). However, this approach was not appropriate for rapid or smallvolume sampling when HPLC analysis was not necessary. To study transport specifically, a vacuum filtration method was adopted as described by Winkler et al. (13), which is similar to the technique used by Aprille (1) to study atractylosideinsensitive [ 14 C]ATP transport in isolated liver mitochondria. In this method, the labeled compound to be studied was added to the incubation medium (described above but minus BSA) at 0 C. Isolated heart mitochondria were then added and mixed rapidly, and a time 0 aliquot (0.1 ml) was removed. The sample was then incubated at 30 C, and aliquots were removed at other time points. Each aliquot was filtered under vacuum through a m Millipore filter prechilled by washing with 5 ml of 4 C saline immediately before addition of the sample. The filter was then rapidly washed with 10 ml of cold saline, removed, dissolved in Formula 963 or Aquassure scintillation fluid (Packard), and counted in a liquid scintillation counter. Uptake of radioactivity was plotted against time, and, where necessary, the initial rate of uptake was calculated from the early time points. The final equilibrated levels were determined for some experiments from later time points. The saline wash cleared the intermembrane space so that only matrix contents remained on the filter. In most experiments, [ 14 C]sucrose was included in the incubation to assess the washing procedure. The amount of sucrose left on the filter was consistently quite low and indicated that 95% of compounds in the intermembrane space was removed. This eliminated the need to subtract the nucleotide contribution of the intermembrane space. The concentration, form, and specific radioactivity of the nucleotides studied in the vacuum filtration experiments were monitored by taking aliquots of the sample directly to an equal volume of cold 10% TCA at appropriate times. This sample was neutralized using AG 11 A8 resin (Bio-Rad) and analyzed by HPLC as described previously (8). Note that this sample gives the total (medium matrix) concentration for the given nucleotide; however, the matrix space represents only 0.4% of the total [1 l matrix space/mg mitochondrial protein (8) and 4 mg mitochondrial protein/ml], so these samples effectively show changes in the medium only. RESULTS Vacuum filtration. The uptake of labeled guanine nucleotides into the matrix was determined by separating the matrix contents from the medium using rapid vacuum filtration. Two controls were used to verify that vacuum filtration measured the uptake of soluble nucleotides into the matrix. First, experiments were conducted with [ 3 H]guanine nucleotide phosphates in which the incubations were carried out entirely on ice, blocking protein-mediated uptake but not simple diffusion. No uptake was observed (data not shown). An additional concern was that uptake might represent incorporation into mitochondrial RNA. Because CTP and UTP were not present in the medium, RNA synthesis did not seem likely but could occur using endogenous levels of these nucleotides. To prove that the radioactivity taken into the matrix was soluble, filter disks were washed with 5% TCA instead of saline, breaking open the matrix and washing out the acid-soluble components while precipitating the RNA onto the filters. Only background radioactivity was observed on the disks in these experiments, demonstrating that incorporation of guanine nucleotide into RNA did not account for radioactivity on the disk. Time course of equilibration, redistribution, and uptake of [ 3 H]GTP, [ 3 H]GDP, and [ 3 H]GMP into isolated heart mitochondria. Previous work (8) showed that the addition of guanine nucleotide (labeled with 3 Honthe base) to isolated mitochondria resulted in the distribution of that label among all three phosphorylated forms of guanine nucleotide both within the mitochondrial matrix and in the medium. Depending on the rate of interconversion relative to the rate of transport, the uptake of label may have been limited to a single form or may have been the sum of all three forms of guanine nucleotide, each with its own kinetic constants. Thus the first goal of this study was to try to distinguish the individual role of each phosphorylated species of guanine nucleotide in the transport process. This was accomplished by incubating isolated heart mitochondria with [ 3 H]GTP, [ 3 H]GDP, or [ 3 H]GMP and determining the time course of label interconversion to other forms (Fig. 1) together with the time course of label uptake into the matrix (Fig. 2). These experiments were done in the presence (Fig. 1, A C, and Fig. 2A) and absence of 2 mm ATP (Fig. 1, D F, and Fig. 2B) to dramatically alter the interconversion of label. When 0.1 mm [ 3 H]GTP was added to isolated heart mitochondria in the presence of ATP, the label rapidly equili-

3 C1872 Fig. 1. Interconversion of labeled guanine nucleotides in the medium. Isolated heart mitochondria were incubated as described in EXPERIMENTAL PROCEDURES with the indicated labeled nucleotide (0.1 mm, 5 Ci/ml) as follows: A and D: [ 3 H]GTP in the presence (A) and absence (D) of 2 mm ATP. B and E: [ 3 H]GDP in the presence (B) and absence (E) of 2 mm ATP. C and F: [ 3 H]GMP in the presence (C) and absence (F) of 2 mm ATP. At the indicated times, an aliquot of each sample was removed and acidified with an equal volume of 10% TCA. Each sample was centrifuged, and the supernatant was decanted and neutralized as described previously (8). The labeled compounds were separated by HPLC, identified by comparison to standards, and quantitated by both ultraviolet absorption and detection of radioactivity by an in-line scintillation counter (8). The radioactivity in each peak was converted to a concentration by dividing by the specific radioactivity of the added nucleotide. The sum of the 3 forms at each time point in each panel totaled 0.1 mm as expected. Results are from a representative experiment. brated such that the medium concentration of [ 3 H]GTP fell to 0.07 mm with the concomitant rise of [ 3 H]GDP to 0.03 mm. Only a trace amount of [ 3 H]GMP was formed. The uptake of matrix label under this condition was linear at a rate of 1.9 pmol mg protein 1 min 1 (Fig. 2A). If the label was added as 0.1 mm [ 3 H]GDP (Fig. 1B), 5 min was required for equilibration of the label, during which time the labeled GDP medium level fell from 0.1 to 0.03 mm as it was converted to 0.07 mm [ 3 H]GTP. Again, only a trace amount of [ 3 H]GMP was formed. The matrix uptake of label after addition of [ 3 H]GDP was hyperbolic with an initial rate of 12 Fig. 2. Time course of uptake of [ 3 H]GTP, [ 3 H]GDP, and [ 3 H]GMP in the presence and absence of ATP. Mitochondria were incubated as described in EXPERIMENTAL PROCEDURES with the labeled compound indicated (0.1 mm, 5 Ci/ml) in the presence (A) and absence (B) of 2 mm ATP. Samples were removed at various times from 0 to 30 min, and matrix uptake was quantitated by vacuum filtration as described in EXPERIMENTAL PROCEDURES. The radioactivity on the filter was corrected for background and was converted to picomoles of nucleotide by dividing by the specific radioactivity of the labeled nucleotide pool. Results are presented as means and SE of nucleotide uptake per milligram of mitochondrial protein from 5 experiments. In A, the data are presented using two y-axis scales. The solid lines are plotted against the scale on the left and allow differences in uptake to be observed. The dashed lines represent the same data plotted against the scale on the right, which is the same scale as shown in B and allows the data in A and B to be directly compared.

4 C1873 pmol mg protein 1 min 1 in the first 2 min, decreasing after 5 min of incubation to 1.5 pmol mg protein 1 min 1 (Fig. 2A). The initial rapid rate of uptake was 6 times higher than the rate observed during the first 2 min of incubation with [ 3 H]GTP, indicating that GDP was a much better substrate for transport than GTP. Furthermore, as the level of [ 3 H]GDP decreased and the level of [ 3 H]GTP increased, the rate of label uptake into the matrix decreased, again suggesting that GDP was a much better substrate for transport than GTP. In fact, using the value of 12 pmol mg protein 1 min 1 as representing the rate of transport from the 0.1 mm [ 3 H]GDP pool, we could predict the rate of transport from the [ 3 H]GDP pool when both [ 3 H]GTP and [ 3 H]GDP pools were present. The predicted rate of uptake from the 0.03 mm [ 3 H]GDP pool that rapidly formed after [ 3 H]GTP addition (Fig. 1A) was 3.6 pmol mg protein 1 min 1. Because the observed rate of uptake after [ 3 H]GTP addition was only 1.9 pmol mg protein 1 min 1 (Fig. 2A), the data indicated that the 0.07 mm [ 3 H]GTP pool inhibited uptake from the smaller [ 3 H]GDP pool. One interpretation of these data was that both [ 3 H]GTP and [ 3 H]GDP were transported into the matrix by this process but that transport of GTP was much slower. Thus, when GTP was present, some of the carriers were occupied in a slower transport process. The equilibrated levels of [ 3 H]GTP and [ 3 H]GDP observed in Fig. 1, A and B, reflected the energy charge of the isolated mitochondrial preparation. The medium ratio of GTP to GDP in Fig. 1, A and B, was 2.3, and the guanine energy charge was These data agreed exactly with the ATP-to-ADP ratio and adenine energy charge determined for these same samples (data not shown). In other studies from our laboratory in which the medium ATP-to-ADP ratio varied over a wide range, the medium GTP-to-GDP ratio was always nearly identical. These data demonstrated that the adenine and guanine pools in the medium were in rapid equilibrium, presumably mediated by an active nucleoside-diphosphokinase known to be present in the intermembrane space (14). In normal respiring mitochondria, the energy charge of the system is maintained by ATP, which is in turn maintained by oxidative phosphorylation. Thus omission of ATP from the incubation medium should drastically lower the predicted energy charge. When [ 3 H]GTP was added in the absence of ATP (Fig. 1D), it was steadily converted to [ 3 H]GDP, yielding final levels of 0.01 mm [ 3 H]GTP, mm [ 3 H]GDP, and mm [ 3 H]GMP, for a [ 3 H]GTP-to-[ 3 H]GDP ratio of 0.14 and an energy charge of 0.5, dramatically lowered from Fig. 1, A and B. The matrix uptake of label under this condition was slightly hyperbolic, rising to 14 pmol mg protein 1 min 1 at 5 min after addition and falling to a fairly linear rate of 9.7 pmol mg protein 1 min 1 from min. This rate was over five times faster than that observed in the presence of ATP (Fig. 2, compare A, dashed line, with B). When label was added as [ 3 H]GDP in the absence of ATP (Fig. 1E), it was only slowly phosphorylated to [ 3 H]GTP, presumably by substrate level phosphorylation via the succinate thiokinase step of the Krebs cycle. After 30 min of incubation, the level of [ 3 H]GDP fell gradually from 0.1 to 0.08 mm with concomitant increases in [ 3 H]GTP and [ 3 H]GMP of 0.01 mm levels each, the same values observed for [ 3 H]GTP addition in Fig. 1D. As above, in the presence of ATP, the guanine nucleotide levels distributed to reflect the energy charge of the isolated mitochondrial system. The rate of matrix uptake of label after addition of [ 3 H]GDP in the absence of ATP was hyperbolic with an initial 2-min rate of 78 pmol mg protein 1 min 1. This rate was 6.5 times faster than the rate of label uptake observed when [ 3 H]GDP was added in the presence of 2 mm ATP (Fig. 2, compare A, dashed line, with B). Because [ 3 H]GDP was essentially the only labeled form present during the rapid uptake phase (0 5 min), the actual species of nucleotide transported across the inner membrane must have been [ 3 H]GDP and the uptake rate represented uptake solely from the [ 3 H]GDP pool. As noted earlier, this value could be used to calculate uptake rates from the [ 3 H]GDP pool when both labeled GDP and GTP pools were present, allowing an estimation of the potential role of the [ 3 H]GTP pool. The data in Fig. 1D showed that, after addition of 0.1 mm [ 3 H]GTP, the [ 3 H]GDP concentration was already at mm immediately after addition (0 time point), rising to mm by 5 min. At the initial concentration of [ 3 H]GDP, the predicted rate of label uptake from the [ 3 H]GDP pool alone would be 9.4 rising to as much as 38 pmol mg protein 1 min 1 at 5 min. The observed rate of uptake after [ 3 H]GTP addition noted above (Fig. 2B) was 7.2, rising to 14 pmol mg protein 1 min 1 at 5 min. Because the observed rate of uptake potentially from both the [ 3 H]GTP and [ 3 H]GDP pools was less than that expected from the [ 3 H]GDP pool alone, these data suggested that [ 3 H]GTP inhibited the transport of [ 3 H]GDP. Thus the data obtained here in the absence of ATP, as well as the data obtained above in the presence of ATP, suggested that GDP was markedly preferred by this transport system, whereas the presence of GTP inhibited uptake of GDP. As noted above, this inhibition was probably caused by the much slower transport of GTP occupying the carrier. The hyperbolic shape of the label uptake curve after addition of [ 3 H]GDP was attributed to approaching equilibration of [ 3 H]GDP transport. The 0.1 mm medium level of [ 3 H]GDP was equivalent to a matrix level of 100 pmol [ 3 H]GDP/mg mitochondrial protein [assuming a 1 l/mg protein space (8)]. Thus the 400 pmol/mg protein value reached after 30 min of transport represented a fourfold concentration of label in the matrix. This fourfold increase in concentration determined by label was also confirmed by HPLC chemical measurement of GDP in the medium and matrix (data not shown). These data were very similar to the active compartmentalization of adenine nucleotides observed in isolated heart mitochondria (McKee and Bentley, unpublished observations) and suggested that, in the absence of ATP, heart mitochondria concentrate guanine nucleotides in the matrix in much the same way as they do adenine nucleotides. However, in the pres-

5 C1874 ence of ATP (Fig. 2A), the matrix uptake of guanine nucleotides (0.1 mm) after 30 min of incubation was still significantly below the medium concentration. When label was added as [ 3 H]GMP (Fig. 1, C and F), the label did not readily equilibrate with the other phosphorylated forms and showed only a slow conversion to GDP and GTP. Although slow, this conversion nevertheless occurred at a constant rate of 13% in 30 min (0.11 nmol min 1 mg 1 mitochondrial protein) in the presence of ATP (C) and fell to just 1.4% in 30 min in the absence of ATP (F). This steady conversion suggested the presence of a low-activity nucleoside monophosphate kinase capable of using GMP as a substrate. To our knowledge, this activity has not been previously shown in mitochondria. The subcellular localization of the three known members of the mammalian guanylate kinase family has not been completely examined, but there is no evidence that any of them reside in the mitochondria (3, 13). Alternatively, GMP may be a weak substrate for the mitochondrial matrix form of adenylate kinase, which is a GTP:AMP phosphotransferase (4). The 10-fold decrease in conversion observed on omitting ATP (compare C and F, Fig. 1) points out the expected role of ATP or GTP as the donor phosphate for this reaction. The rate of uptake of [ 3 H]GMP in the presence or absence of ATP was very similar (Fig. 2, compare A, dashed line, and B) and analogous to the rate observed when [ 3 H]GTP was added in the presence of ATP (Fig. 2A, circles vs. squares). Because label was present almost solely as [ 3 H]GMP (Fig. 1, C and F), particularly during the first 10 min of incubation, these data clearly demonstrated that [ 3 H]GMP was transported across the inner membrane. Furthermore, HPLC analysis of the matrix contents after a 60-min incubation with 0.1 mm [ 3 H]GMP indicated that [ 3 H]GMP was also the predominant form of guanine nucleotide in the matrix ( 90%, data not shown). Effect of atractyloside on guanine nucleotide uptake. The data above described a transport process that preferred GDP, was stimulated by the omission of ATP, and was capable of concentrating guanine nucleotide in the matrix in the absence of ATP. A good candidate for such transport was the ADP/ATP translocase, an exchange carrier that typically transports ADP into the matrix in exchange for ATP. GDP, and perhaps GTP, might serve as secondary substrates for this carrier in the heart, in which case their transport would be competitively inhibited by the presence of ATP. Because this carrier is specifically inhibited by atractyloside (6), we determined if guanine nucleotide served as a substrate for this translocase by testing the effects of atractyloside on guanine nucleotide uptake. This was done under two conditions. The first used a very low exogenous concentration of [ 3 H]GTP in which 5 Ci/ml label was added directly from the commercially prepared isotope (10 Ci/mmol mm GTP) (Fig. 3A). The second condition (Fig. 3B) used a more physiological concentration of [ 3 H]GTP (0.5 mm, 0.02 mci/ ml). These two conditions were chosen because preliminary evidence suggested that the characteristics of transport at these two widely different concentrations were significantly different, a fact confirmed by the data in Fig. 3. At low exogenous concentrations of Fig. 3. Effect of atractyloside (Atr) and ATP on uptake of [ 3 H]GTP. Mitochondria were incubated as described in EXPERIMENTAL PROCEDURES with 2 different concentrations of exogenous [ 3 H]GTP: 5 Ci/ml at 10 Ci/mmol ( mm) in A and 0.5 mm, 5 Ci/ml in B. A: uptake of [ 3 H]GTP was measured as described in the legend of Fig. 2 except that the uptake of label was not divided by the specific radioactivity (see text) and is expressed as dpm taken up by the matrix per milligram of mitochondrial protein. The level of radioactive GTP in the matrix that would match the medium radioactive concentration is indicated (11,000 dpm/ l, dashed line), which assumes a matrix space of 1 l matrix water/mg protein (8). B: uptake of GTP is expressed as described in EXPERIMENTAL PROCEDURES as nmol of GTP taken up per milligram of mitochondrial protein. The matrix concentration of GTP equivalent to the 0.5 mm medium concentration is 0.5 nmol GTP/mg and is shown as a dashed line. Uptake of GTP in both panels was measured under 4 conditions: F, 2 mm ATP and 1 M atractylate; Œ, 2 mm ATP and no atractylate;, no ATP and 1 M atractylate;, no ATP and no atractylate. Results represent means and SE of 3 6 independent determinations.

6 C1875 [ 3 H]GTP (Fig. 3A), the label equilibrated between the medium and the matrix within 30 min to a level that was concentrated two- to threefold in the matrix [assumes 1 l/mg volume in the matrix (8)]. Because this did not occur at 4 C, carrier-mediated uptake of the label was clearly demonstrated. Furthermore, the presence or absence of ATP or of atractyloside had only a minimal effect on uptake rates and equilibration levels, although equilibrium levels of guanine nucleotides in the absence of ATP and atractyloside did show a modest statistical difference from levels in the presence of both (P 0.001). Note that at this low added level of guanine nucleotide, the endogenous amount of guanine nucleotide [ 4 nmol guanine nucleotide/ml (8)] was eightfold higher than the amount of GTP added. Because the extent to which this initial matrix pool enters and dilutes the medium pool of guanine nucleotides was not known, the concentration and specific radioactivity of GTP in the medium were not known. As a result, the data in Fig. 3A are expressed simply in units of radioactivity taken up into the matrix. In contrast, at the more physiological concentration of [ 3 H]GTP, the endogenous GTP was negligible and the data are expressed as nanomoles of GTP uptake per milligram of mitochondrial protein (Fig. 3B). At a physiological level of GTP, the effect of ATP and atractyloside on the uptake of label was dramatic (Fig. 3B). In the absence of ATP and atractyloside, uptake of label added as [ 3 H]GTP was rapid and hyperbolic. The rate at the first time point (15 min) was pmol mg protein 1 min 1, and the level as equilibration approached was 1,500 pmol/mg protein, a level threefold above the medium level (Fig. 3B, inverted triangles). The addition of 2 mm ATP to the incubation inhibited uptake by 90% (Fig. 3B, triangles, P 0.001), whereas the addition of atractyloside inhibited uptake by 98% (Fig. 3B, squares, P 0.001). Dose-response experiments with atractyloside have shown that the concentration of atractyloside used in these studies (1 M) provided maximal inhibition of uptake of both labeled guanine and adenine nucleotides (data not shown). Increasing the concentration of atractyloside had no further affect on uptake. These data clearly implicated the ADP/ATP translocase in the transport of guanine nucleotide but raised the question as to the physiological significance of this transport, given the competitive inhibition of guanine nucleotide transport by ATP. The data in Fig. 3B suggested that 2 mm ATP was not quite sufficient to completely inhibit guanine nucleotide uptake over this carrier because addition of atractyloside to the 2 mm ATP resulted in a further modest inhibition ( to pmol uptake mg protein 1 min 1 for results with ATP and without atractyloside vs. with both ATP and atractyloside, P 0.001, Fig. 3B). To better understand the competitive nature of ATP in the uptake of guanine nucleotides, rates of label uptake at a fixed addition of 0.5 mm [ 3 H]GTP were measured as a function of increasing medium ATP concentrations (Fig. 4). Half-maximal inhibition of label uptake was observed at mm ATP, and Fig. 4. Dose-response curve of ATP concentration ([ATP]) on [ 3 H]GTP uptake in the presence and absence of atractyloside. Mitochondria were incubated with [ 3 H]GTP (0.5 mm, 10 Ci/ml) as described in the legend to Fig. 2, in the presence and absence of 1 M atractylate at various concentrations of ATP as indicated. The initial rate of GTP uptake per minute per milligram of mitochondrial protein (velocity) was calculated for each concentration of ATP by obtaining the best-fit slope to the linear portion of the uptake time course by linear regression (SigmaPlot 5.0, SPSS). The velocity obtained was plotted against the ATP concentration. Results represent means and SE of 3 independent determinations. maximum inhibition was observed at ATP levels 4 mm. As expected, at maximal inhibition, the rate of uptake became the same as the rate observed in the presence of atractyloside. The level of maximum inhibition of [ 3 H]GTP uptake was reached at an 8-to-1 ratio of adenine nucleotide to guanine nucleotide. These data explained the lack of affect of atractyloside on [ 3 H]GTP uptake at very low exogenous [ 3 H]GTP levels (Fig. 3A) because, in the absence of addition of exogenous nucleotides, the adenine nucleotide level was about eight times greater than the guanine nucleotide level (8). Thus uptake of label when [ 3 H]GTP was added at the low concentration (Fig. 3A) was already restricted to the atractyloside-insensitive carrier. Because the ratio of adenine nucleotide to guanine nucleotide under typical conditions in the cell is 14:1 (5, 10), it seems unlikely that guanine nucleotides are carried on the ATP/ADP translocase under physiological conditions, suggesting a role for the slow atractyloside-insensitive transport process. Interestingly, at physiological levels of GTP, the addition of 2 mm ATP to atractyloside stimulated the uptake of label added as [ 3 H]GTP over twofold ( to pmol uptake mg protein 1 min 1 for experiments without ATP and with atractyloside vs. those with both ATP and atractyloside, respectively, P 0.001, Fig. 3B). This suggested that ATP did not compete for transport on the slower atractyloside-insensitive carrier.

7 C1876 Specificity of transport on the atractyloside-insensitive carrier. The slow atractyloside-insensitive pathway of uptake was detected in Figs. 3 and 4 using [ 3 H]GTP as the added label. However, as noted in Fig. 1, A and D, label added as [ 3 H]GTP was converted partially to [ 3 H]GDP, making it difficult to know which species (or both) was actually transported across the inner membrane on this slower carrier. Furthermore, because the uptake of [ 3 H]GMP was not reduced by the addition of ATP (Fig. 2), it seemed likely that uptake of [ 3 H]GMP occurred via the atractyloside-insensitive pathway. To obtain a better understanding of the specificity of guanine nucleotide uptake via the atractyloside-insensitive pathway, the uptake of [ 3 H]GMP, [ 3 H]GDP, [ 3 H]GTP, and [ 3 H]IMP were determined in the presence of 1 M atractyloside and 2 mm ATP and compared (Fig. 5). The distribution of label observed for each addition of guanine nucleotide was essentially the same as shown in Fig. 1, A C, except that atractyloside somewhat slowed the interconversion process (data not shown). Label added as [ 3 H]IMP remained solely in IMP. Because the rate of label redistribution was slower in the presence of atractyloside, a preference for a specific species of guanine nucleotide for this transport process should be readily apparent at the Fig. 5. Time course of uptake of [ 3 H]GTP, [ 3 H]GDP, [ 3 H]GMP, and [ 3 H]IMP on the atractyloside-insensitive carrier. Mitochondria were incubated as described in the legend of Fig. 2 except that 2 mm ATP was present in all samples, and uptake was measured in the presence of 1 M atractylate (solid lines). Samples were removed at various times from 0 to 60 min, and matrix uptake was quantitated by vacuum filtration and expressed as described in the legend of Fig. 2. Guanine nucleotide results are presented as means and SE of nucleotide uptake per milligram of mitochondrial protein from at least 4 experiments; IMP results are the means of 3 experiments. The uptake curves of [ 3 H]GMP, [ 3 H]GDP, and [ 3 H]GTP from Fig. 2 (no atractylate) are shown as dashed lines for comparison. early time points. However, the uptake of label when added as [ 3 H]GMP, [ 3 H]GDP, or [ 3 H]GTP was about the same regardless of the form of guanine nucleotide added (Fig. 5, solid lines). Thus the atractyloside-insensitive transport pathway appeared to move all three nucleotides more or less equally. This differed from the uptake results observed in the absence of atractyloside (Fig. 5, dashed lines), which demonstrated that there was a clear preference for uptake of [ 3 H]GDP, a preference now ascribed to the ADP/ATP translocase. The similarity of uptake rates of [ 3 H]GMP in the presence and absence of atractyloside (Fig. 5, dashed vs. solid line) demonstrated that GMP was not a substrate for the ADT/ATP translocase and was transported solely via the insensitive pathway. As an additional test of specificity, the transport of a closely related nucleotide, [ 3 H]IMP, was measured (Fig. 5), and no uptake of label into the matrix was observed, demonstrating that this pathway has specificity for guanine nucleotide. In earlier work, we demonstrated that guanine and guanosine were also not transported into the matrix (8). The failure of transport of IMP, guanine, and guanosine under these conditions provided strong evidence that the slow uptake of guanine nucleotide by this mechanism was not an artifact caused by a small number of damaged mitochondria in the preparation. Effect of nucleoside/nucleotide transport inhibitors on [ 3 H]GTP uptake. The thiol active reagents mersalyl and N-ethylmaleimide have been shown to inhibit the MgATP/phosphate exchange carrier (2), whereas p- hydroxymercuribenzoic acid has been shown to inhibit uptake of deoxyguanosine, all in liver mitochondria (12). The effect of these three reagents on the atractyloside-insensitive transport of GTP was examined. Mersalyl ( mm), N-ethylmaleimide (0.2 1 mm), and p-hydroxymercuribenzoic acid (0.1 mm) were added to isolated heart mitochondria in the presence of 2 mm ATP, atractyloside, and 0.1 mm [ 3 H]GTP, and uptake of label was determined. These reagents, at concentrations that have been reported to maximally inhibit the carriers above (2, 12), had no effect on the uptake of labeled GTP (data not shown). In the presence of a range of mersalyl concentrations, GTP uptake was recorded only during the first 5 min of incubation, as this reagent appeared to damage the mitochondria, as evidenced by their aggregation. These data suggested that an active sulfhydryl group was not involved with the atractyloside-insensitive transport of GTP. Nitrobenzylmercaptopurine ribonucleoside (NBMPR) has been shown to inhibit nucleoside/nucleotide transport in nuclei and across plasma membranes (7). NBMPR, at a concentration reported to maximally inhibit this process (0.01 mm), was tested in the same manner as the reagents above and was also without effect on GTP uptake. Effect of incubation temperature on [ 3 H]GTP uptake. As shown in Fig. 6, the rate of uptake of [ 3 H]GTP in the presence and absence of atractyloside was dependent on the temperature of incubation as expected for a protein-mediated process. These data provide evidence

8 C1877 Fig. 6. Velocity of GTP uptake vs. temperature. Mitochondria were incubated at 4 different temperatures, using [ 3 H]GTP at 0.01 mci/ ml, 0.1 mm in the presence of 2.0 mm ATP, and in the presence or absence of 1 M atractylate. Aliquots were removed at various times from 0 to 60 min, and matrix uptake was quantitated by vacuum filtration as described in the legend of Fig. 2. The rate of uptake (velocity) was calculated for each concentration of GTP as described in the legend of Fig. 4 and plotted as a function of temperature. The symbols represent means and SE of the velocity of uptake of GTP of 4 experiments. that the slow uptake of guanine nucleotide is not mediated by simple diffusion or diffusion through some type of channel. Kinetic constants of the atractyloside-sensitive and -insensitive guanine nucleotide carriers. The data in Figs. 1 and 2 indicate that the atractyloside-sensitive carrier displayed a marked preference for GDP. The kinetic constants of this uptake process were determined by measuring the velocity of label uptake measured at a range of GDP concentrations in the absence of ATP (Fig. 7, top). As shown in Fig. 2B, the concentration of labeled GDP remained relatively constant over the time course (0 15 min) used to measure initial velocity. Kinetic constants of [ 3 H]GDP uptake were determined by computing the best fit of these data (SigmaPlot 5.0, SPSS) to the transport equation for carrier-mediated uptake: V V max [GDP]/(K m [GDP]), where K m is the Michaelis-Menten constant and V is velocity. As shown in Fig. 7, top, an excellent fit of the data was obtained, demonstrating the presence of a saturable carrier protein with a V max of uptake of pmol mg protein 1 min 1 and a K m of mm for GDP. At equilibrium, this carrier was capable of concentrating label in the matrix about fourfold over the medium for GDP additions below 0.25 mm, a fact confirmed by HPLC analysis. However, as Fig. 7. Kinetic values of guanine nucleotide uptake on the atractyloside-sensitive and -insensitive carriers. Top: [ 3 H]GDP uptake on the atractyloside-sensitive carrier. Mitochondria were incubated as described in the legend of Fig. 2 in the absence of ATP and with varying concentrations of [ 3 H]GDP. Aliquots were removed at various times from 0 to 30 min, and matrix uptake was quantitated by vacuum filtration as described in EXPERIMENTAL PROCEDURES. The initial rate of uptake (velocity) was calculated over the first 15 min for each concentration of GDP by obtaining the best-fit slope to the linear portion of the uptake time course by linear regression (SigmaPlot 5.0) Velocities were plotted as a function of concentration. Points with error bars represent the mean and SE of 5 independent experiments; other points represent the average of 2 independent experiments. The best-fit line to the data was calculated by computer using the equation V V max [GDP]/(K m [GDP]) (SigmaPlot 5.0). The best fit was obtained with V max pmol mg 1 min 1, K m mm, and r Middle: [ 3 H]GTP/GDP uptake on the atractyloside-insensitive carrier. Mitochondria were incubated as described in the legend of Fig. 2 with 2 mm ATP and 1 M atractylate present in all samples with varying concentrations of [ 3 H]GTP. The data were obtained and presented as described for top. The symbols represent means and SE of the velocity of uptake of GTP for 5 experiments. The best-fit line to the data points for the carrier-mediated uptake curve was calculated by computer using the equation V V max [GTP]/(K m [GTP]) (SigmaPlot 5.0). A good fit was obtained between to 25 mm GTP with V max pmol mg 1 min 1, K m mm, and r The predicted line underestimates the rates below 0.01 mm. Bottom:[ 3 H]GMP uptake on the atractyloside-insensitive carrier. Mitochondria were incubated as described in the legend of Fig. 2 with 2 mm ATP and 1 M atractylate present in all samples with varying concentrations of [ 3 H]GMP. The data were obtained and presented as described for top. The symbols represent the average of the velocity of uptake of GMP for 2 experiments. The best-fit line to the data points for the carriermediated uptake curve was calculated by computer using the equation V V max [GMP]/(K m [GMP]) (SigmaPlot 5.0). A good fit was obtained between 0.05 and 25 mm GMP with V max pmol mg 1 min 1, K m mm, and r The predicted line underestimates the rates below 0.05 mm.

9 C1878 GDP levels increased, the ratio of matrix to medium GDP concentrations decreased to 2 at 1 mm GDP and to 1 at 5 mm GDP. The atractyloside-insensitive uptake of guanine nucleotide was more complex, having the ability to transport all three phosphorylated forms of guanine nucleotide. This uptake process could represent a single carrier capable of transporting all three forms of guanine nucleotide with similar kinetics. Alternatively, the process could represent more than one carrier, each with specificity for a single nucleotide form. We chose to evaluate the kinetic constants of this carrier by first using [ 3 H]GTP. This was accomplished by measuring the velocity of uptake of label as a function of [ 3 H]GTP medium concentration in the presence of atractyloside and 2 mm ATP (Fig. 7, middle). The initial velocity was determined by uptake over the first 15 min of incubation. As above, these data were fit to the carrier-mediated transport equation by computer, yielding a curve that fit the data well between and 25 mm GTP. This curve described the presence of a potential carrier with a V max pmol GTP mg protein 1 min 1 and a K m mm (Fig. 7, middle). Below 0.01 mm GTP, the measured rate of uptake significantly exceeded the rate predicted from this curve. This discrepancy may be attributed to another carrier that has a very low V max ( 0.1 pmol GTP mg 1 min 1 ) and a high-affinity K m ( 0.01 mm). HPLC analysis of the matrix guanine nucleotide pools under these conditions showed that the vast majority of the label remained in guanine nucleotides (8), making it unlikely that the faster uptake at low GTP concentrations was caused by uptake of a small amount of contaminating radioactivity in the commercially prepared product. In any event, at physiological levels of guanine nucleotide [0.42 mm (8)], this discrepancy was inconsequential. As noted in Fig. 1A, when label was added as [ 3 H]GTP, 30% of the label was fairly quickly converted to [ 3 H]GDP; thus uptake probably represented transport of both species of nucleotide. To further evaluate atractyloside-insensitive transport, the uptake of label as a function of [ 3 H]GMP concentration in the medium was determined (Fig. 7, bottom). As noted earlier, under this condition, the label remained predominantly as [ 3 H]GMP (Fig. 1C). The results were very similar to the results observed for [ 3 H]GTP addition. The best-fit curve of [ 3 H]GMP uptake fit the data well between 0.05 and 25 mm GMP and described the presence of a potential carrier with a V max pmol GMP mg 1 min 1 and a K m mm (Fig. 7, bottom). As was noted for [ 3 H]GTP addition, below 0.05 mm GMP, the measured rate of uptake significantly exceeded the rate predicted from this curve and suggested another carrier with low V max ( 1 pmol GMP mg 1 min 1 ) and high-affinity K m ( 0.05 mm). DISCUSSION From the results presented in this paper, we propose that guanine nucleotides can be transported into the matrix of isolated heart mitochondria by two mechanisms. The first mechanism is carrier mediated, with a V max of pmol mg 1 min 1 and a K m of mm for GDP. This carrier is atractyloside sensitive, temperature dependent, capable of active transport, prefers GDP to GTP, does not transport GMP or IMP, and only transports GDP when adenine nucleotide levels are limiting (Fig. 4). This carrier is likely to represent the ATP/ADP translocase, a major protein of the inner membrane that catalyzes the electrogenic exchange of ADP and ATP across the membrane (for review, see Ref. 6). In active mitochondria, the net activity of this carrier is to promote the rapid (ms) exchange of medium ADP for matrix ATP. The ADP/ ATP translocase is encoded by a family of genes, with the liver and heart expressing different forms of the protein (6). In mitochondria isolated from the heart, but not the liver, this carrier has been reported to catalyze atractyloside-sensitive transport of adenine nucleotide from the mitochondria to the medium in exchange for phosphate and thus may be capable of net transport (11). Transport of guanine nucleotides on the ATP/ADP translocase has not been previously reported and was only detected in this study by measuring uptake of labeled guanine nucleotide when adenine nucleotide was omitted from the medium. The V max of GDP uptake on this carrier is several orders of magnitude below the V max of ATP/ADP exchange. Because the heart expresses a specific form of the translocase, it is not possible to extrapolate this result to other tissues. In any event, under physiological concentrations of adenine and guanine nucleotide, our results suggest that the translocase may not be involved in guanine nucleotide transport in vivo. The second mechanism of guanine nucleotide transport is consistent with a carrier-mediated process that is atractyloside-insensitive, temperature dependent, and appears to transport all three phosphorylated forms of guanine nucleotide equally. However, the maximum velocity of this carrier(s) is low, 48.7 pmol mg 1 min 1 with a K m of 4.4 mm for GTP/GDP and 36.8 pmol mg 1 min 1 with a K m of 4.1 mm for GMP, far lower than any other known carrier. This may suggest a limited number of carriers in the membrane or a very slow rate of transport after binding nucleotide. Nevertheless, we believe we have demonstrated that this is a bona fide physiological mechanism of uptake of guanine nucleotide and is not caused by potential artifacts of the system. The atractylosideinsensitive uptake is unlikely to be caused by residual uptake via incomplete inhibition of the ATP/ADP translocase because the amount of atractylate used (1 M) maximally inhibited labeled ATP uptake (Bentley and McKee, personal observation) and because increasing the concentration of atractylate had no further effect on labeled GTP uptake. Furthermore, as shown in Fig. 4, the addition of increasing amounts of ATP maximally reduced the uptake of labeled GTP to the same level as that observed in the presence of 1 M atractylate, again suggesting that the remaining uptake of GTP is not via the adenine nucleotide translocase. The temperature dependence of uptake together with the failure to transport closely related molecules

10 C1879 of IMP, guanine, and guanosine clearly indicate that uptake is not caused by a small number of damaged mitochondria or the transient opening of a pore or channel, since neither of these processes would be able to provide the specificity of transport noted. It is also unlikely that uptake is caused by contaminants in the radioisotope, since such uptake should be dependent only on the amount of isotope added and should not be affected by addition of various levels of unlabeled guanine nucleotide. Our data throughout these experiments show a strong and consistent correlation between the concentration and specific radioactivity of guanine nucleotide and the rate of uptake of label that cannot be explained by label contamination. Furthermore, HPLC analysis of matrix contents in the presence of atractyloside demonstrated that the vast majority of the label in the matrix was present in guanine nucleotides (8). Lastly, although guanine nucleotide uptake is slow, the process supports biogenesis, not metabolism, and mitochondria in postmitotic tissue do not require a more rapid system. The only reaction that consumes GTP in the matrix is RNA synthesis, and, under steady-state conditions of RNA synthesis and degradation, this may be replenished by GMP, provided that the matrix contains an enzyme capable of phosphorylating GMP. We demonstrated in this study that isolated heart mitochondria (Fig. 1) contain a low-activity enzyme capable of this phosphorylation; however, whether this enzyme resides in the intermembrane space, in the matrix, or both, is not yet known. Mitochondria, like all cellular components, are degraded and turn over even in nondividing tissues and thus must be capable of continuing growth and division. It is this process that requires the net uptake of guanine nucleotide. At an intracellular guanine nucleotide concentration of 0.42 mm (8), this carrier would be capable of uptake at a rate of 255 pmol mg protein 1 h 1. This rate is sufficient to produce the endogenous matrix concentration of guanine nucleotide (0.96 mm; Table 1 in Ref. 8) in the 1 l space of 1 mg of mitochondrial protein in 4 h. This would seem to be more than adequate to account for mitochondrial turnover. In rapidly growing tissues, it is possible that the number of guanine nucleotide carriers in the inner membrane may be increased. Because the results in this paper only address the influx of guanine nucleotide from the medium to the matrix, it is not yet possible to conclude whether this carrier is a simple uniport carrier or a more complex symport or antiport carrier that depends on a second molecule for transport. The transport system most similar to the atractyloside-insensitive process described here is the MgATP/phosphate exchange carrier, which has been characterized in liver and kidney mitochondria and is insensitive to atractyloside and bongkrekic acid (1, 2). This adenine nucleotide carrier is thought to be important in the net transport of adenine nucleotide across the inner membrane. An analogous role is proposed here for guanine nucleotide transport. However, significant differences exist between these two carrier systems. Mersalyl and N-ethylmaleimide, reagents that attack active sulfhydryl residues, have been reported to inhibit the MgATP/ phosphate exchange carrier (2) but had no effect on guanine nucleotide uptake. The MgATP/phosphate exchange carrier is also reported to have rates (in liver) an order of magnitude above those noted here for the guanine nucleotide carrier. The increased rate may reflect the increased concentration and role of adenine nucleotide in the mitochondria. To our knowledge this is the first report of a transport process that accounts for the movement of guanine nucleotides from the cytoplasm into the mitochondrial matrix to support mitochondrial biogenesis. In recent work, we have found that a similar transport process may function in moving deoxythymidine nucleotides from the cytoplasm into the matrix in support of mitochondrial biogenesis (McKee and Bentley, unpublished observations). This work was supported by a Grant-in-Aid from the American Heart Association, National Chapter. REFERENCES 1. Aprille JR. Regulation of the mitochondrial adenine nucleotide pool size in liver: mechanism and metabolic role. FASEB J 2: , Aprille JR. Mechanism and regulation of the mitochondrial ATP-Mg/P i carrier. J Bioenerg Biomembr 25: , Brady WA, Kokoris MS, Fitzgibbon M, and Black ME. Cloning, characterization, and modeling of mouse and human guanylate kinases. J Biol Chem 271: , Jamil T, Fisher RA, and Harris H. Studies on the properties and tissue distribution of the isozymes of guanylate kinase in man. Hum Hered 25: , Kleineke J, Duls C, and Soling HD. Subcellular compartmentation of guanine nucleotides and functional relationships between the adenine and guanine nucleotide systems in isolated hepatocytes. FEBS Lett 107: , Klingenberg M. Molecular aspects of the adenine nucleotide carrier from mitochondria. Arch Biochem Biophys 270: 1 14, Mani RS, Hammond JR, Marjan JMJ, Graham KA, Young JD, Baldwin SA, and Cass CE. Demonstration of equilibrative nucleoside transporters (hent1 and hent2) in nuclear envelopes of cultured human choriocarcinoma (BeWo) cells by functional reconstitution in proteoliposomes. J Biol Chem 273: , McKee EE, Bentley AT, Smith RM Jr, and Ciacco CE. Origin of guanine nucleotides in isolated heart mitochondria. Biochem Biophys Res Commun 257: , McKee EE, Grier BL, Thompson GS, and McCourt JD. Isolation and incubation conditions to study heart mitochondrial protein synthesis. Am J Physiol Endocrinol Metab 258: E492 E502, Muhonen WW and Lambeth DO. The compartmentation of nucleoside diphosphate kinase in mitochondria. Comp Biochem Physiol 110B: , Sandhu GS and Asimakis GK. Mechanism of loss of adenine nucleotides from mitochondria during myocardial ischemia. J Mol Cell Cardiol 23: , Watkins LF and Lewis RA. The metabolism of deoxyguanosine in mitochondria. Characterization of the uptake process. Mol Cell Biochem 77: 71 77, Winkler HH, Bygrave FL, and Lehninger AL. Characterization of the atractyloside-sensitive adenine nucleotide transport system in rat. J Biol Chem 243: 20 28, Yamada M, Shahjahan M, Tanabe T, Kishi F, and Nakazawa A. Cloning and characterization of cdna for mitochondrial GTP:AMP phosphotransferase of bovine liver. J Biol Chem 264: , 1989.