The mitochondrial permeability transition remains an incompletely understood phenomenon, in spite of the fact that it was first observed in the 1950s and has been studied extensively, particularly during the last 15 years (see (1, 2, 3) for review). Recent work utilizing the patch clamp technique applied to isolated mitochondria indicates that the transition is caused by the opening of large pores in the inner membrane(4, 5, 6) . At present there is no known physiological function of this phenomenon; however, it is clear that it occurs in situ, under conditions of oxidative stress, and is an event that can be pivotal in the mechanisms leading to cell death(7, 8, 9, 10) . Association of the transition with cell injury mechanisms and the identification of cyclosporin A (CSA) ()as a potent inhibitor of the phenomenon (11, 12, 13, 14) has created a high level of interest in this aspect of mitochondrial research.

Many of the agents which induce the transition are toxins and pharmacological agents not normally encountered in vivo(1) . The same is true for the known inhibitors(1) . Both inducing agents and inhibitors are chemically diverse but display a common activity in favoring an open or closed state of the permeability transition pore (PTP), respectively. It has proven difficult to explain the analogous actions of diverse compounds on the PTP. Bernardi and co-workers (15, 16) have shown that membrane potential and matrix pH are central PTP regulators. Membrane surface potential may also be a central regulator (3, 17) . It is thus possible that the chemical diversity/common activity problem associated with the transition may arise, in part, from the effects of transition regulators on bioenergetic parameters and membrane physical properties.

Although many known transition effectors would not normally be encountered in cells, a number of activators (e.g. Ca, P, acyl-CoA) and inhibitors (e.g. ADP, Mg, polyamines) are physiological cell constituents. In seeking to understand regulation of the transition in an intracellular setting, it then seems reasonable to consider these agents as primary regulators and to investigate how other cellular components and conditions affect their actions on the PTP. The present report demonstrates that the 14-amino acid peptide mastoparan has a potent stimulatory effect on PTP opening when mitochondria are incubated in the presence of Ca and P. At somewhat higher concentrations, mastoparan appears to induce pore opening in a Ca independent and CSA insensitive manner. This is the first report of a peptide exerting a regulatory influence on the transition and, as such, draws attention to a new class of potential regulators of this phenomenon within cells. In this report, the actions of mastoparan on the transition are also compared to those of an analog, MP14. Differences observed between the potency of these two peptides are consistent with the hypothesis that components of cell signaling mechanisms, apart from Ca, are involved in regulation of the PTP. Aspects of these data have been presented in abstract form(18) .

EXPERIMENTAL PROCEDURES

Reagents

Mitochondrial Preparations and Incubation Conditions

Preparation of Phospholipid Vesicles and Determination of Vesicle Permeability

The permeability of POPC vesicles to Ca/Quin-2 was monitored by incubating them in 10 mM Hepes (Na), pH 7.0, containing 20 mM NaCl and 50 µM CaCl. The development of a permeable membrane was indicated by formation of the CaQuin-2 complex, which was determined by difference absorbance measurements, using the Aminco DW2a spectrophotometer operated in the dual wavelength mode (wavelength pair 264 versus 338 nm). The temperature was 25 °C, and the nominal POPC concentration was 300 µM. At this concentration, the entrapped volume is 0.6 µl/ml(24) . For some experiments, a membrane potential (inside negative) was imposed across the vesicle membrane by adding valinomycin at 0.1 µM. Parallel experiments employing the TPP electrode and application of the Nernst equation showed that the magnitude of the membrane potential produced in this way was approximately 150 mV.

Other Methods

When examining mitochondrial ultrastructure or determining the release of enzymes and matrix space solutes, aliquots taken from incubations were initially centrifuged to sediment the mitochondria (Eppendorf microcentrifuge, 13,000 g, 3 min). For ultrastructural studies, pellets were then fixed, stained, and processed for examination by electron microscopy as described before(28) . Fractions of total enzyme activities and Mg released during incubation were determined by assaying the supernatants, also using established methods(29, 30) . The release of nucleotides and other coenzymes from the matrix space were determined by HPLC analysis of the sedimented mitochondria. Pellets were extracted as described by Stocchi et al.(31) , whereas HPLC analysis was carried out as described by Novgorodov et al.(17) .

RESULTS

CSA-sensitive and Insensitive Actions of Mastoparan on the Permeability Transition

Figure 1: Stimulation of the permeability transition by mastoparan; inhibition by CSA. Mitochondria were incubated at 0.5 mg of protein/ml and at 25 °C. The medium contained 250 mM sucrose, 10 mM Hepes (Tris), 5 mM succinate (Tris), 5 mM P (Tris), pH 7.4, plus 3.3 µM TPPCl. TPP accumulation and release (upper panel) and swelling (lower panel) were determined simultaneously as described under ``Experimental Procedures.'' CaCl (dashed lines) or mastoparan (solid lines) were added where indicated at 25 nmol/mg of protein and 1.0 µM, respectively. For traces labeled a, the medium contained CSA at 0.5 µM from the beginning of the experiment. For traces labeled b, CSA was absent.

Figure 2: Stimulation of the permeability transition by mastoparan; inhibition by EGTA and oligomycin. Mitochondria were incubated as described in the legend to Fig. 1, and TPP accumulation (upper panel) and swelling (lower panel) were determined simultaneously as described under ``Experimental Procedures.'' For all traces (a-c), 1.0 µM mastoparan was added where indicated. For traces labeled b and c, the medium contained EGTA (0.5 mM) or oligomycin (1 µg/mg protein), respectively, from the beginning of the incubation.

Mitochondria depleted of endogenous Ca(21) were utilized to investigate the effect of Ca content and P concentration on mastoparan-dependent induction of the transition. When P is present at a concentration of 5 mM, 1.0 µM mastoparan markedly stimulates the transition across a range of Ca loads (Fig. 3A), although little or no activity is seen in the absence of exogenous P (Fig. 3B). In a similar way, when an intermediate Ca load is employed, the action of mastoparan is dependent on P concentration (Fig. 3C) with little or no activity seen in the absence of exogenous Ca (Fig. 3D). These findings distinguish mastoparan from better known inducers of the transition such as hydroperoxides, sulfhydryl reagents, and P. Rather than inducing a rapid transition in Ca-loaded mitochondria, as do these other agents(1) , or acting as a substitute for Ca per se, mastoparan enhances the action of Ca and P which are both still required to obtain the transition.

Figure 3: Effects of Ca content and P concentration on the mastoparan-dependent opening of the permeability transition pore. Ca-depleted mitochondria were incubated at 0.5 mg of protein/ml and at 25 °C. The medium contained 250 mM sucrose, 10 mM Hepes (Tris), and 5 mM succinate (Tris), pH 7.4. Pore opening was monitored by swelling as described under ``Experimental Procedures.'' A, the medium also contained 5 mM P (Tris) and CaCl as shown. P was present from the beginning, whereas CaCl was added at 1 min following the addition of mitochondria. , 1.0 µM mastoparan was added 1 min after the addition of CaCl. , mastoparan was not added. Values plotted are the percent of maximal swelling which had occurred at 10 min following the addition of mitochondria. B, same as A, except that P was deleted from the medium. C, same as A, except that the mitochondrial Ca load was 25 nmol/mg protein, and the medium P concentration was varied as shown. D, same as C, except that Ca was not added.

Mastoparan also differs from other inducers of the transition with respect to the inhibitory activity of CSA. While CSA prevents the action of P, hydroperoxides, and other agents for extended periods, and apparently without a sharp dependence on inducing agent concentration(32, 33) , increasing the concentration of mastoparan from 1 to 3 µM eliminates the inhibitory action of CSA (Fig. 4). A comparison of Fig. 1and Fig. 2with Fig. 4illustrates another point of interest. When a low concentration of mastoparan is employed and the resulting transition is sensitive to CSA ( Fig. 1and Fig. 2), TPP release and swelling proceed after a lag period which is characteristic of the transition induced by most agents. At higher mastoparan concentrations, where sensitivity to CSA is lacking or incomplete (Fig. 4), TPP release and, to some extent, swelling proceed without a significant lag period following mastoparan addition.

Figure 4: CSA insensitive actions of mastoparan on mitochondria. Incubations were conducted as described in the legend to Fig. 1with 0.5 µM CSA present from the beginning. TTP accumulation (upper panel) and swelling (lower panel) were determined simultaneously as described under ``Experimental Procedures.'' For traces labeled a-d, mastoparan was added where indicated at the following concentrations: a, 0 µM; b, 1.0 µM; c, 2.0 µM; d, 3.0 µM.

The mastoparan analog MP14 (Table 1) facilitates opening of the PTP in a manner analogous to the parent peptide. At relatively low concentrations, the actions of MP14 are dependent upon the Ca and P concentrations and are sensitive to CSA, whereas at higher concentrations, MP14 actions are seen in the absence of free Ca and are CSA insensitive (data not shown). Fig. 5shows the CSA-sensitive and insensitive effects of the two peptides on PTP opening (swelling response) as a function of their concentrations. With mastoparan, it is seen that the CSA insensitive activity occurs over a concentration range which is approximately twice that required for the CSA-sensitive activity. MP14 is less effective than mastoparan on a concentration basis, regardless of whether the CSA-sensitive or insensitive activities are considered. However, with the latter peptide, the CSA-sensitive and insensitive concentration curves are offset by 4-5-fold (Fig. 5).

Figure 5: The concentration dependence of mastoparan and MP14 effects on mitochondria in the presence and absence of CSA. Data were obtained from experiments like those shown in Fig. 4. , mastoparan was employed in the absence of CSA. , same as except that 0.5 µM CSA was present from the beginning of the incubations. , MP14 was employed in the absence of CSA. , same as except that 0.5 µM CSA was present from the beginning of the incubations.

The Solute Size Exclusion Properties of Mastoparan and MP14-dependent Pores

In the methods used, the effects of PEGs of various molecular weights on the large amplitude swelling associated with the transition were compared to their effects on matrix solute release and on mitochondrial ultrastructure. Prior to conducting these studies, the effects of PEGs on the osmotic pressure of mitochondrial incubation media were examined. This work was conducted because PEGs increase the osmotic pressure of solutions in disproportion to their concentration(40) , a property which would complicate the use of PEGs to characterize the pores in question. Fig. 6A shows that small, intermediate, and relatively large PEGs increase the osmotic pressure of a dilute buffer solution in marked excess of their molal concentration. The extent of this nonideal behavior is a nonlinear function of the weight of PEG per volume of solution and is more extreme for the larger PEGs. Indeed, as the molecular weight increases, the weight per volume required to establish a given osmotic pressure becomes almost independent of the PEG molecular weight (Fig. 6B). ()To maintain a constant osmotic pressure in mitochondrial incubations containing PEGs (Fig. 7Fig. 8Fig. 9), the media were prepared by mixing the normal medium with 300 milliosmolal solutions of PEG dissolved in a dilute buffer, to yield the desired concentration of the polymer. The table contained within Fig. 6B shows the solution concentrations which yield a 300 milliosmolal osmotic pressure for all PEGs which were used. Fig. 6C shows that when the mitochondrial medium is mixed with isoosmotic solutions of PEGs, the resulting osmotic pressures remain relatively constant when the fraction of osmotic pressure arising from PEG is varied from 0-100%.

Figure 6: The osmotic pressure properties of PEG-containing solutions. A, measured osmotic pressure values of 3 mM Hepes buffer, pH 7.4, containing increasing amounts of 0.6 ()-, 3.4 ()-, or 8.0 ()-kDa PEG. The dashed lines in this panel show how osmotic pressure would vary if solutions of 0.6 (a)-, 3.4 (b)-, or 8.0 (c)-kDa PEG behaved ideally. These lines were located by preparing single solutions of each PEG utilizing weighed amounts of PEG and solvent. The molal concentration of the solution could then be calculated, providing a y axis coordinate for the solution. The x axis coordinate was located by noting the volume of each solution and dividing the weight of PEG used by the volume to yield milligrams of PEG/ml of solution. Straight lines were then drawn as defined by these points and the origin. B, the concentration of PEG in milligrams per ml of solution required to give an osmotic pressure of 300 mosM, as a function of PEG molecular weight. These values were determined from experiments like those shown in A. The table embedded in B shows the same data expressed numerically in concentration units of mM. C, measured osmotic pressure values of 0.6 ()-, 3.4 ()-, and 8.0 ()-kDa PEG solutions progressively diluted with a mitochondrial incubation medium. As seen in the panel, both the initial PEG solutions and the mitochondrial incubation medium had osmotic pressures near 300 milliosmolal when measured individually.

Figure 7: Inhibition of PTP-dependent swelling by selected concentrations of 3.4-kDa PEG. Mitochondria were incubated at 1.0 mg of protein/ml and at 25 °C, in isoosmotic media which did not contain an exogenous respiratory substrate. The media were prepared by mixing 300 milliosmolal mannitol/sucrose/Hepes (mitochondrial isolation medium) with a medium comprised of 64 mM 3.4-kDa PEG in 3 mM Hepes, pH 7.4 (also 300 milliosmolal) to yield the indicated concentrations of PEG, while maintaining a constant osmotic pressure (see Fig. 6). P was present at a final concentration of 2 mM. CaCl was added where shown, at 50 nmol/mg protein, and swelling was subsequently monitored as described under ``Experimental Procedures.''

Figure 8: The fraction of mitochondria which undergo PTP opening as a function of 3.4-kDa PEG concentration. Incubations were conducted as described in the legend to Fig. 7. For all parameters, samples were taken 8 min after the addition of Ca, and mitochondria were sedimented using a microcentrifuge (13,000 g for 3 min). For A-D, the supernatants were removed and the pellets were fixed, embedded, and examined by electron microscopy as described previously (28) (total magnification = 3,000). The medium concentrations of 3.4-kDa PEG were as follows: A, 0 mM; B, 3 mM; C, 8 mM; D, 20 mM. For E, the PEG concentration was as shown. The extents of swelling (), malate dehydrogenase release (), adenylate kinase release (), and Mg release () were determined using the supernatants obtained after sedimenting mitochondria, as described under ``Experimental Procedures.''

Figure 9: Solute size exclusion properties of pores in mitochondria. Mitochondria were incubated at 0.5 mg of protein/ml and at 25 °C in 300 milliosmolal media containing 5 mM concentration each of succinate (Tris) and P (Tris), pH 7.4. 40% of the total osmotic pressure was derived from PEG and the remainder from mannitol/sucrose (3:1 mole ratio) plus the other solutes described above. PEGs ranging from 0.6 kDa to 8 kDa were employed, as illustrated in the figure, and these were mixed with mannitol/sucrose to give the desired fraction of osmotic pressure derived from PEG, while keeping the total value near 300 milliosmolal (see Fig. 6). Swelling (pore formation) was initiated after a 2-min preincubation and allowed to proceed until the apparent absorbance of the suspensions became constant. These final values, presented as a percent of the value obtained in the absence of PEG, are plotted as a function of the PEG molecular weight. Each value is the mean ± S.E. of three determinations, with each replicate value obtained using a separate mitochondrial preparation. Pore formation was induced as follows: A, CaCl (50 nmol/mg protein); B, alamethicin (3.5 µg/ml); C, mastoparan (1.0 µM); D, mastoparan (3.0 µM) and the medium contained 1 µM CSA; E, MP14 (3.0 µM); F, MP14 (11 µM) and the medium contained 1 µM CSA.

Under these near isoosmotic conditions, low concentrations of 3.4-kDa PEG markedly reduce the magnitude of swelling response which occurs when the PTP is opened by the action of Ca plus P (Fig. 7). Based upon existing data, a 3.4-kDa PEG would be expected to pass through the outer membrane voltage-dependent anion channel (VDAC) (41) but not through the PTP (42, 43, 44, 45) . Several features of these swelling curves are noteworthy. For all but the highest PEG concentrations employed, the curves are monophasic and produce a final absorbance change which is relatively stable. The latter characteristic is taken to indicate that 3.4-kDa PEG does not slowly permeate through the PTP. Near the high end of the concentration range, some multiphasic behavior is seen. This might indicate that movement of 3.4-kDa PEG through VDAC is somewhat restricted, as is discussed further below. Finally, the initial rate of swelling is seen to be essentially constant as the PEG concentration increases. This suggests that PEG does not impede the access of low molecular weight solutes to the PTP or otherwise slow their rate of permeation.

The above interpretations are based upon the assumption that the decreased swelling produced by increasing PEG concentrations results primarily from a decrease in the maximal matrix space volume which is attained following opening of PTP and that the PTP opens in all mitochondria. By an alternative interpretation, decreased swelling might represent PTP opening in a decreasing fraction of the mitochondria. The latter interpretation must be considered because it is well known that the transition occurs heterogeneously in mitochondrial populations(1) . Furthermore, it is not clear if the PTP is a single molecular entity or the same entity in all mitochondria (17) . Fig. 8distinguishes between the alternative interpretations. In this figure, panels A-D, it is seen that mitochondrial ultrastructure is relatively uniform when observed in samples which were fixed after the swelling response reached completion in the presence of 0, 3, 8, or 20 mM PEG. This structural characteristic is not consistent with PTP opening in a variable subfraction of the mitochondria (see (1) ). In addition, decreased swelling produced by increasing concentrations of 3.4-kDa PEG is not accompanied by a decreased release of matrix Mg (Fig. 8E) nor that of matrix space nucleotides and other cofactors, as determined by HPLC analysis (data not shown). Finally, malate dehydrogenase release was minimal or absent across the full range of PEG concentration investigated, although a marked release of adenylate kinase activity was observed which was diminished by PEG. Taken together, these data strongly indicate that the concentration-dependent inhibition of swelling produced by 3.4-kDa PEG results from a reduced matrix space expansion in all, or nearly all, of the mitochondria and not from PTP opening in a decreasing fraction of the mitochondria. In addition, the enzyme release data show that the limited swelling which is observed in the presence of PEG does not reflect rupture of the inner membrane in some of the mitochondria (absence of malate dehydrogenase release), although rupture of the outer membrane occurs and is subject to inhibition by 3.4-kDa PEG (adenylate kinase release).

With the above information as a background, the size exclusion properties of ``permeability transition pores'' induced in several ways were examined by observing the maximal extents of swelling which occur when the medium contains PEG of various molecular weights. The concentrations of PEGs were selected and the solutions were prepared such that 40% of the osmotic pressure was obtained from PEG, while the total osmotic pressure remained near 300 milliosmolal (Fig. 6). The PTP induced by Ca plus P can be easily distinguished from the pore formed by alamethicin using this method. A comparison of A and B of Fig. 9illustrates this, together with the approach taken when analyzing these data. The Ca plus P-induced PTP produces a swelling response which is one-half inhibited by PEG of a molecular weight near 650 (Fig. 9A). The alamethicin pore is clearly larger, with PEG of molecular weight near 1700 being required to produce a comparable reduction in the extent of swelling (Fig. 9B). The remaining panels in Fig. 9show the size exclusion properties of the pores induced by mastoparan and MP14 in the presence or absence of CSA. No marked differences are seen, although all of these pores may be slightly larger than the PTP as induced by Ca plus P alone.

Mastoparan and MP14 Permeabilize Phospholipid Vesicles: the Effect of Membrane Potential

Figure 10: Membrane potential-dependent and independent actions of mastoparan and MP14 on the permeability of POPC vesicles. Quin 2/Hepes/K load POPC vesicles were prepared and incubated as described under ``Experimental Procedures.'' The nominal POPC concentration was 300 µM, and the external medium contained 10 mM Hepes (Na), pH 7.0, 20 mM NaCl, and 50 µM CaCl. Increased membrane permeability was indicated by formation of the CaQuin 2 complex, which was monitored by dual wavelength absorbance measurements made at 294 versus 338 nm(19) . A (left panel), valinomycin (0.1 µM) was added or not added, where indicated. B (left panel), same as A except that mastoparan (2.0 µM) was added where indicated. Right panel, a summary plot of the initial rates of absorbance change as a function of mastoparan or MP14 concentration, as determined by fitting the data to the equation given under ``Results.'' and , mastoparan and MP14, respectively, membrane potential present. and , mastoparan and MP14, respectively, membrane potential absent.

To analyze data like those shown in Fig. 10, trace B, the relative initial rate of CaQuin-2 complex formation was estimated by fitting early portions of the progress curve (nonlinear least square techniques, see (19) ) to the expression: A = A + Bt + Ct. In this expression, A and A are the total and initial difference absorbance values, respectively, B is the initial rate, C is a correction factor for nonlinearity, and t is time. The right side panel of Fig. 10shows the effect of mastoparan and MP14 concentration on vesicle permeability as determined by this method. No significant differences between the two peptides are apparent in the absence of a membrane potential. When a membrane potential is present, however, mastoparan is markedly more effective than its analog. The increment produced by membrane potential, compared to no membrane potential was 5.8- and 2.3-fold for mastoparan and MP14, respectively. This panel also shows that for all cases examined, the relationship between peptide concentration and vesicle permeability are linear.

DISCUSSION

Mastoparan is a 14-amino acid amphipathic peptide obtained from wasp venom(47, 48, 49, 50, 51, 52, 53) . It possesses a wide spectrum of pharmacological activities including mast cell degranulation(50) , activation of G-protein-mediated mechanisms(54, 55, 56, 57, 58) , inhibition of calmodulin-mediated mechanisms(59, 60, 61) , stimulation of phospholipases A and C(55, 62, 63) , stimulation or inhibition of cation-specific channels(64) , and others (see (51) and (65) ). Facilitation of the mitochondrial permeability transition can now be added to this broad spectrum of activities. It should be noted that this newly recognized action of mastoparan is marked at concentrations <1 µM (Fig. 5), whereas the other activities are typically seen over a 5-100 µM range. Thus, the actions of mastoparan on mitochondria are the most potent described to date and may well be involved in the toxicological mechanism of this peptide, given the relationship between the transition and the death of injured cells (see introduction to the text). It should also be noted that as a modulator of the transition, mastoparan is more potent than most of the known agents(1) .

Whether or not the inner membrane pore(s) induced by both low (1 µM) and high (2 µM) levels of mastoparan are in fact the PTP is an important question because all other inducers require Ca for activity, with the apparent exception of phenylarsine oxide(66) , and are poorly active in the presence of CSA, even when they are used over a broad concentration range(1, 3) . In the lower concentration range, pore induction by mastoparan also requires Ca, is facilitated by P, and is inhibited by CSA and by oligomycin (Fig. 1Fig. 2Fig. 3). Given these characteristics, it is very likely that this pore is in fact the PTP. To test the identity of the pore formed in the higher mastoparan concentration range, the solute size exclusion properties of several pores were compared (Fig. 9). By this criterion, the pore induced by mastoparan in the presence of CSA is also the PTP as are the CSA-sensitive and insensitive pores induced by MP14. In considering these interpretations it must also be noted that mastoparan and MP14 perturb phospholipid bilayers by a mechanism which is membrane potential-dependent (Fig. 10) but otherwise unknown. It is thus possible that the peptide molecules per se form pores in the inner mitochondrial membrane and that these pores, rather than the PTP, are responsible for the CSA independent activities. While this interpretation remains viable, it seems improbable because it would be highly fortuitous if the pores formed by both peptides were so similar to each other, and to the PTP, by the criterion of solute size exclusion.

There is a second aspect of the solute size exclusion data presented here which is of interest and which becomes apparent when these data are compared to earlier studies in which PEGs were utilized to characterize the size exclusion properties of the PTP(42, 44, 45) . There are disagreements between the early studies which may reflect pore induction by differing agents, and/or technical considerations arising from unrecognized osmotic pressure properties of PEGs, the use of a limited number of PEG molecular weights, and other factors. Haworth and Hunter (induction by arsenate) (42) reported that 1.5- and 4-kDa PEGs are excluded, 0.6-kDa or smaller PEGs are permeant, whereas an intermediate condition exists with 1-kDa PEG. Vercesi (induction by oxalacetate) (45) reported that swelling is eliminated by 1-kDa PEG. Lê-Quôc and Lê-Quôc (induction by N-butylmaleimide) (44) found no permeation of 6-kDa PEG but progressively faster permeation of 4-, 1.5-, and 0.6-kDa PEG. Large fractions of matrix space enzyme activities were released under the conditions of their study and the authors concluded that an association of VDAC with the adenine nucleotide translocase forms the PTP(44) .

The present results are not subject to some of the uncertainties in the earlier data, and, as a consequence, they show that when the pore is induced by Ca plus P, increasing the molecular weight of PEG in the medium from 0.4 to 4 kDa produces a progressive decrease in swelling. This behavior is unexpected because each point was obtained at a (essentially) constant osmotic pressure, with PEG providing the same fraction of that pressure, and represents an apparent equilibrium condition (i.e. the extents of swelling were not increasing significantly at the time the values were taken). If the PTP is a rigid structure, one would expect a sharper cutoff in the molecular weight of PEG which is fully excluded. If it is a flexible structure which can impede yet pass solutes of variable size, one would expect the smaller PEGs to decrease the rate of swelling instead of producing an intermediate and stable value. The presence of molecular weight heterogeneity in the various samples of PEG, together with a sieving action by the PTP, cannot easily explain the observed behavior. According to the manufacturer, this heterogeneity does not exceed ±10% of the stated value, and filtration experiments conducted with some of the samples were consistent with this specification (data not shown). At the higher end of the molecular mass range (4 and possibly 3.4 kDa), PEG may act, in part, because it is partially excluded by VDAC(41) . This could relax the oncotic pressure gradient across the outer membrane, contract the intermembrane space volume, and raise the protein concentration, so that reduced swelling occurs when the PTP is opened. This possible explanation is not applicable in the case of the smaller PEGs, however, since they clearly permeate VDAC(41) . The existence of different size pores in individual mitochondria could explain the observed behavior, and differences in the mechanisms which gate solute entry from opposite sides of the membrane might also explain this behavior. No choice between such explanations seems possible without further studies, which are now in progress.

Induction of the transition by mastoparan and MP14 may occur through PTP regulatory mechanisms which are already recognized(1, 2, 3) . However, since these peptides have a wide spectrum of pharmacological activities (see above), new potential regulatory mechanisms can also be considered. Regarding the established mechanisms, it does not appear that reducing membrane potential is the primary mode by which the peptides act under CSA-sensitive conditions. This is indicated by the TPP accumulation data in Fig. 1and Fig. 2which show that mastoparan produces only minor changes in membrane potential when CSA, EGTA, or oligomycin are present to inhibit pore opening. The same is true for MP14 (data not shown). Under CSA insensitive conditions, an early and extensive release of TPP is seen and this seems to occur somewhat faster than swelling (Fig. 4). This depolarization could be brought about by peptide molecules acting upon the membrane lipid phase as illustrated in Fig. 10. It is thought that no PTP effector dominates in the interactive system of pore regulation through allosteric and membrane/bioenergetic mechanisms. Instead, the open/closed probability appears to be established as a sum of positive and negative actions exerted at a number of sites(17) . Thus, the peptides may act at an allosteric site, synergistically with Ca and P acting their sites, under CSA-sensitive conditions. It could then be the additional influence of depolarization produced by higher peptide levels acting upon the membrane lipid phase which allows the pore to open in a Ca independent and CSA insensitive manner. The present data cannot identify the allosteric site at which the peptides may act, although it is noted that mastoparan carries three positive charges, whereas MP14 carries two (Table 1). Thus, the relative potency of these peptides correlates with positive charge and they might act at a site which normally binds a cation. The Ca binding site is one possibility and an action at this site could explain why pore induction can occur in the absence of Ca when the peptide concentration is high. Possible displacement of CSA from its site of action by high concentrations of mastoparan is another possibility to consider.

Analogies between regulation of the PTP and the N-methyl-D-aspartic acid receptor channel suggest that the PTP belongs to the super family of ligand gated ion channels and thus may be regulated by covalent modification (see (3) ). Such systems often involve G-proteins, and a mitochondrial G-protein has recently been identified and isolated(67, 68) . According to a recent report, MP14 retains the G-protein-mediated activities of mastoparan but has a diminished capacity to perturb membranes(69) . It is for these reasons that the actions of MP14 and mastoparan on the PTP were compared in this study. MP14 is only slightly less active than mastoparan when inducing pore opening in a CSA-sensitive manner, whereas there is a larger activity differential when the pore is induced in the presence of CSA (Fig. 5). If the CSA insensitive induction involves membrane perturbation (depolarization) as suggested above, then these data suggest regulation of the PTP through a G-protein. The relative effectiveness of mastoparan and MP14 have not yet been compared in a variety of systems, however, and so it is not clear how completely the G-protein-mediated and membrane perturbation-mediated activities of MP14 are distinguished. In particular, it appears that a membrane perturbation activity differential between these two peptides is only seen in the presence of a membrane potential (Fig. 10) and the reason for this deserves further investigation.

The actions of mastoparan and MP14 on the PTP suggest that low concentrations of other amphipathic peptides could regulate the PTP in cells. Such peptides are a product of processing imported proteins and may accumulate in the mitochondrial matrix under some conditions. Extra- and intramatrix space amphipathic peptides may also accumulate in injured cells due to proteolysis. These peptides may promote pore opening, even in presence of CSA, and become a factor in maintaining the protection of injured cells afforded by CSA.

FOOTNOTES

*

This research was supported by United States Public Health Service Grants HL 49182, HL 49181, and HL 36124 from the National Institutes of Health, NHLBI. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore by hereby marked ``advertisement'' in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

§

To whom correspondence and reprint requests should be addressed: The Ohio State University, Dept. of Medical Biochemistry, 310A Hamilton Hall, 1645 Neil Ave., Columbus, OH 43210-1218. Tel.: 614-292-8774; Fax: 614-292-4118.

()

The abbreviations used are: CSA, cyclosporin A; PEG, polyethylene glycol; POPC, 1-palmitoyl-2-oleoyl-sn-glycerophosphatidylcholine; PTP, permeability transition pore; TPP, tetraphenylphosphonium cation; VDAC, voltage-dependent anion channel; HPLC, high performance liquid chromatography.

()

The binding of a large number of water molecules per molecule of PEG and/or other structural ordering of water, with a resulting reduction in the activity of water, are thought to be responsible for the nonideal osmotic properties of PEGs(40) .

ACKNOWLEDGEMENTS

REFERENCES

1. Gunter, T. E., and Pfeiffer, D. R. (1990) Am. J. Physiol. 258, C755-C786
2. Gunter, T. E., Gunter, K. K., Sheu, S.-S., and Gavin, C. E. (1994) Am. J. Physiol. 267, C313-C339
3. Bernardi, P., Broekemeier, K. M., and Pfeiffer, D. R. (1994) J. Bioenerg. Biomembr. 26, 509-517 [CrossRef][Medline] [Order article via Infotrieve]
4. Szabó, I., and Zoratti, M. (1992) J. Bioenerg. Biomembr. 24, 111-117 [CrossRef][Medline] [Order article via Infotrieve]
5. Kinnally, K. W., Antonenko, Y. N., and Zorov, D. B. (1992) J. Bioenerg. Biomembr. 24, 99-110 [CrossRef][Medline] [Order article via Infotrieve]
6. Zoratti, M., and Szabó, I. (1994) J. Bioenerg. Biomembr. 26, 543-553 [CrossRef][Medline] [Order article via Infotrieve]
7. Nazareth, W., Yafei, N., and Crompton, M. (1991) J. Mol. Cell. Cardiol. 23, 1351-1354 [CrossRef][Medline] [Order article via Infotrieve]
8. Broekemeier, K. M., Carpenter-Deyo, L., Reed, D. J., and Pfeiffer, D. R. (1992) FEBS 304, 192-194 [CrossRef][Medline] [Order article via Infotrieve]
9. Pastorino, J. G., Snyder, J. W., Serroni, A., Hoek, J. B., and Farber, J. L. (1993) J. Biol. Chem. 268, 13791-13798 [Abstract/Free Full Text]
10. Imberti, R., Nieminen, A.-L., Herman, B., and LeMasters, J. J. (1993) J. Pharmacol. Exp. Ther. 265, 392-400 [Abstract/Free Full Text]
11. Fournier, N., Ducet, G., and Crevat, A. (1987) J. Bioenerg. Biomembr. 19, 297-303 [CrossRef][Medline] [Order article via Infotrieve]
12. Crompton, M., Ellinger, H., and Costi, A. (1988) Biochem. J. 255, 357-360 [Medline] [Order article via Infotrieve]
13. Broekemeier, K. M., Dempsey, M. E., and Pfeiffer, D. R. (1989) J. Biol. Chem. 264, 7826-7830 [Abstract/Free Full Text]
14. Halestrap, A. P., and Davidson, A. M. (1990) Biochem. J. 268, 153-160 [Medline] [Order article via Infotrieve]
15. Petronilli, V., Cola, C., and Bernardi, P. (1993) J. Biol. Chem. 268, 1011-1016 [Abstract/Free Full Text]
16. Nicolli, A., Petronilli, V., and Bernardi, P. (1993) Biochemistry 32, 4461-4465 [CrossRef][Medline] [Order article via Infotrieve]
17. Novgorodov, S. A., Gudz, T. I., Brierley, G. P., and Pfeiffer, D. R. (1994) Arch. Biochem. Biophys. 311, 219-228 [CrossRef][Medline] [Order article via Infotrieve]
18. Novgorodov, S. A., Gudz, T. I., and Pfeiffer, D. R. (1993) Biophys. J. 64, A107 (Abstr. M-Pos 492)
19. Erdahl, W. L., Chapman, C. J., Taylor, R. W., and Pfeiffer, D. R. (1994) Biophys. J. 66, 1678-1693 [Abstract/Free Full Text]
20. Broekemeier, K. M., Schmid, P. C., Schmid, H. H. O., and Pfeiffer, D. R. (1985) J. Biol. Chem. 260, 105-113 [Abstract/Free Full Text]
21. Wingrove, D. E., and Gunter, T. E. (1986) J. Biol. Chem. 261, 15166-15171 [Abstract/Free Full Text]
22. Beavis, A. D., Brannan, R. D., and Garlid, K. D. (1985) J. Biol. Chem. 260, 13424-13433 [Abstract/Free Full Text]
23. Kamo, N., Muratsugu, M., Hongoh, R., and Kobatake, Y. (1979) J. Membr. Biol. 49, 105-121 [CrossRef][Medline] [Order article via Infotrieve]
24. Chapman, C. J., Erdahl, W. L., Taylor, R. W., and Pfeiffer, D. R. (1990) Chem. Phys. Lipids 55, 73-84 [CrossRef][Medline] [Order article via Infotrieve]
25. Chapman, C. J., Erdahl, W. L., Taylor, R. W., and Pfeiffer, D. R. (1991) Chem. Phys. Lipids 60, 201-208 [CrossRef][Medline] [Order article via Infotrieve]
26. Fry, D. W., White, J. C., and Goldman, D. (1978) Anal. Biochem. 90, 809-815 [CrossRef][Medline] [Order article via Infotrieve]
27. Barlett, G. R. (1959) J. Biol. Chem. 234, 466-468 [Free Full Text]
28. Igbavboa, U., and Pfeiffer, D. R. (1988) J. Biol. Chem. 263, 1405-1412 [Abstract/Free Full Text]
29. Riley, W. W., Jr., and Pfeiffer, D. R. (1985) J. Biol. Chem. 260, 12416-12425 [Abstract/Free Full Text]
30. Broekemeier, K. M., Schmid, P. C., Dempsey, M. E., and Pfeiffer, D. R. (1991) J. Biol. Chem. 266, 20700-20708 [Abstract/Free Full Text]
31. Stocchi, V., Cucchiarini, L., Magnani, M., Chiarantini, L., Palma, P., and Crescentini, G. (1985) Anal. Biochem. 146, 118-124 [CrossRef][Medline] [Order article via Infotrieve]
32. Broekemeier, K. M., and Pfeiffer, D. R. (1989) Biochem. Biophys. Res. Commun. 163, 561-566 [CrossRef][Medline] [Order article via Infotrieve]
33. Broekemeier, K. M., Dempsey, M. E., and Pfeiffer, D. R. (1989) J. Biol. Chem. 264, 7826-7830
34. Katsu, T., Kuroko, M., Morikawa, T., Sanchika, K., Yamaraka, H., Shinoda, S., and Fujita, Y. (1990) Biochim. Biophys. Acta 1027, 185-190 [Medline] [Order article via Infotrieve]
35. Tanimura, A., Matsumoto, Y., and Tojyo, Y. (1991) Biochem. Biophys. Res. Commun. 177, 802-808 [CrossRef][Medline] [Order article via Infotrieve]
36. Raynor, R. L., Kim, Y.-S., Zheng, B., Vogler, W. R., and Kuo, J. F. (1992) FEBS Lett. 307, 275-279 [CrossRef][Medline] [Order article via Infotrieve]
37. Danilenko, M., Worland, P., Carlson, B., Sausville, E. A., and Sharoni, Y. (1991) Biochem. Biophys. Res. Commun. 196, 1296-1302
38. Eisenberg, M., Hall, J. E., and Mead, C. A. (1973) J. Membr. Biol. 14, 143-176 [CrossRef][Medline] [Order article via Infotrieve]
39. Latorre, R., and Alvarez, O. (1981) Physiol. Rev. 61, 77-150 [Free Full Text]
40. Schiller, L. R., Emmett, M., Santa Ana, C. A., and Fordfran, J. S. (1988) Gastroenterology 94, 933-941 [Medline] [Order article via Infotrieve]
41. Colombini, M. (1980) J. Membr. Biol. 53, 79-84 [CrossRef]
42. Haworth, R. A., and Hunter, D. R. (1979) Arch. Biochem. Biophys. 195, 460-467 [CrossRef][Medline] [Order article via Infotrieve]
43. Novgorodov, S. A., Gudz, T. I., Milgrom, Y. M., and Brierley, G. P. (1992) J. Biol. Chem. 267, 16274-16282 [Abstract/Free Full Text]
44. Lê-Quôc, K., and Lê-Quôc, D. (1985) J. Biol. Chem. 260, 7422-7428 [Abstract/Free Full Text]
45. Vercesi, A. B. (1984) Arch. Biochem. Biophys. 232, 86-91 [CrossRef][Medline] [Order article via Infotrieve]
46. Igbavboa, U., and Pfeiffer, D. R. (1991) Biochim. Biophys. Acta 1059, 339-347 [Medline] [Order article via Infotrieve]
47. Zimmerberg, J., and Parsegian, V. A. (1986) Nature 323, 36-39 [CrossRef][Medline] [Order article via Infotrieve]
48. Vodyanoy, I., Bexrukov, S. M., and Parsegian, V. A. (1993) Biophys. J. 65, 2097-2105 [Abstract/Free Full Text]
49. Hirai, Y., Yasuhara, T., Yoshida, H., Nakajima, T., Fujino, M., and Kitada, C. (1979) Chem. Pharm. Bull. (Tokyo) 27, 1942-1944 [Medline] [Order article via Infotrieve]
50. Hirai, Y., Ueno, Y., Yasuhara, T., Yoshida, H., and Nakajima, T. (1979) Chem. Pharm. Bull. (Tokyo) 27, 1945-1946 [Medline] [Order article via Infotrieve]
51. Hirai, Y., Ueno, Y., Yasuhara, T., Yoshida, H., and Nakajima, T. (1980) Biomed. Res. 1, 185-197
52. Hirai, Y., Yasuhara, T., Yoshida,