(Received for publication, February 3, 1997, and in revised form, May 12, 1997)
From the Department of Medical Biochemistry, The Ohio State University, Columbus, Ohio 43210-1218
Yeast mitochondria (Saccharomyces
cerevisiae) contain a permeability transition pore which is
regulated differently than the pore in mammalian mitochondria. In a
mannitol medium containing 10 mM Pi and ethanol
(oxidizable substrate), yeast mitochondria accumulate large amounts of
Ca2+ (>400 nmol/mg of protein) upon the addition of an
electrophoretic Ca2+ ionophore (ETH129). Pore opening does
not occur following Ca2+ uptake, even though ruthenium
red-inhibited rat liver mitochondria undergo rapid pore opening under
analogous conditions. However, a pore does arise in yeast mitochondria
when Ca2+ and Pi are not present, as monitored
by swelling, ultrastructure, and matrix solute release. Pore opening is
slow unless a respiratory substrate is provided (ethanol or NADH) but
also occurs rapidly in response to ATP (2 mM) when
oligomycin is present. Pi and ADP inhibit pore opening
(EC50 ~1 and 4 mM, respectively), however, cyclosporin A (7 µg/ml), oligomycin (20 µg/ml), or
carboxyatractyloside (25 µM) have no effect. The pore
arising during respiration is also inhibited by nigericin or uncoupler,
indicating that an acidic matrix pH antagonizes the process.
Pi also inhibits pore opening by lowering the matrix pH
(Pi/OH antiport). However, inhibition of the
ATP-induced pore by Pi is seen in the presence of mersalyl,
suggesting a second mechanism of action. Since pore induction by ATP is
not sensitive to carboxyatractyloside, ATP appears to act at an
external site and Pi may antagonize the interaction.
Isoosmotic polyethylene glycol-induced contraction of yeast
mitochondria swollen during respiration, or in the presence of ATP, is
50% effective at a solute size of 1.0-1.1 kDa. This suggests that the
same pore is induced in both cases and is comparable in size with the
permeability transition pore of heart and liver mitochondria.
Mammalian mitochondria
(mMt)1 contain a permeability
transition pore (PTP) that, when opened, allows solutes of molecular
mass 1.5 kDa to equilibrate across the inner membrane (1-3). An
elevated matrix Ca2+ concentration in combination with a
second agent (e.g. Pi or an oxidant of matrix
space components) is required for pore opening in most cases.
Cyclosporin A (CsA) is a potent inhibitor of the PTP in mMt (mPTP) (2)
and may function by binding to a mitochondrial cyclophilin that
normally regulates or catalyzes pore opening (4, 5). mPTP opening is
also inhibited by adenine nucleotides, Mg2+,
H+, relatively positive membrane surface potentials, and
high transmembrane potentials (1-4, 6-8). The mPTP has been
postulated to form from the opening of a unique and latent channel in
the inner membrane, by transformation of a normally selective
transporter, or as a highly complex structure formed from the adenine
nucleotide translocase, the peripheral benzodiazepine receptor, and the
outer membrane voltage-dependent anion channel (1-3,
9-11). Whether or not the mPTP opens in cells under physiological
conditions is not known, and evidence for and against this possibility
has been reported (12-14). Thus, the function of the mPTP is also not
known although it has been postulated to have a role in cellular
Ca2+ homeostasis, mMt protein import, thermal regulation,
and apoptosis, to name some of the possibilities. (3, 15-19).
Several studies have sought to determine if a PTP is present in yeast mitochondria (yMt), with conflicting results. Manon and Guerin (20), investigating electrophoretic pathways of ion transport, reported that Ca2+ is without an effect on the coupling state or swelling and concluded that the existence of a PTP in yMt (yPTP) is uncertain. Szabo et al. (21) applied the patch-clamp technique to isolated yMt and observed a channel with multiple conductance levels (100-600 pS) that is not inhibited by CsA, ADP, or H+. These authors also did not observe Ca2+ related swelling of yMt and were inconclusive regarding the existence of a PTP (21). In contrast, Lohret and Kinnally (15, 22) described a channel with multiple conductance levels, having a peak conductance of 1-1.5 nS, in patch-clamped yeast mitoplasts. They concluded that this channel represents the yPTP, or the megachannel to use their nomenclature. In addition, Ballarin and Sorgato (23), also using the patch clamp technique on yeast mitoplasts, described a channel of lower conductance (400-800 pS), which remains open in the presence of ATP.
The present study was undertaken to clarify whether or not a PTP is present in yMt, and subsequently to investigate its regulation. The results show that a PTP is induced in the isolated organelle by respiration or the presence of external ATP. It is inhibited by ADP, is comparable in size to the homologous structure in mMt, but is Ca2+ independent and not subject to inhibition by CsA. Pi, which promotes opening of the mPTP, is an inhibitor of the yPTP through an action exerted via the matrix pH, and through a second action exerted at a site that is probably accessible from the extramitochondrial volume. Possible functions of the yPTP and other implications of these findings are considered.
The yeast
strain YPH250 (MATa, ura3-52,
lys2-801, ade2-101, trp1-1,
his3-
200, leu2-
1) (ATCC 96519) or W303-1A
(MATa, ade2-1, his3-11 & 3-15,
trp1-1, leu2-3 & 2-112, ura3-1,
can1-100) were grown at pH 5 in 1% yeast extract, 2%
bacto-peptone, 0.05% dextrose, 0.01% adenine, and either 3% glycerol
or 2% lactate. One 4-liter or two 2-liter baffled flasks were employed
per liter of culture medium. The culture was inoculated with 50 ml of a
preculture (36-48 h, A600 = 1.5-2) and
incubated in an orbital shaker (180 rpm, 30 °C). The cells (15-30
g) were harvested after 24 h.
yMt were isolated essentially as described by Daum et al. (24) using 2 mg of Zymolyase 20T (ICN) per g of cells to form spheroplasts. The isolated yMt (typically 50 mg of protein) were suspended in 0.6 M mannitol, 20 mM HEPES (K+), pH 6.8, 0.5 mg/ml BSA, 0.1 mM EGTA (K+), and stored on ice. Rat liver mMt were prepared as described previously (25) and suspended in 0.23 M mannitol, 0.07 M sucrose, and 3 mM HEPES (Na+), pH 7.4.
Incubation Conditions and Methods for Monitoring the Permeability TransitionMedia composition and the additions made are described
in the figure legends. Incubations were performed at 25 °C in a 3-ml cuvette. In most cases, the permeability transition was monitored by
swelling (A540), using an SLM Aminco DW-2C
spectrophotometer operated in the split beam mode (26). Some
variability in this parameter was encountered between different yMt
preparations and the two yeast strains that were used, similar to the
differences that are seen when separate preparations of mMt are
compared. The data shown are typical experiments.
The uptake and release of Ca2+ by mitochondria were
followed using antipyrylazo III as an indicator of the
extramitochondrial Ca2+ concentration (27). Changes in
A720-790 were recorded in the DW-2C operated
in the dual wavelength mode and converted to Ca2+
concentration values using a calibration curve that was generated with
a standard solution of Ca2+. Estimates of membrane
potential (
) were made using safranine (18 µM) as
an indicator. Safranine accumulation was also followed spectrophotometrically, as a change in
A511-533 (28).
The solute exclusion method (26, 29-31) was used to estimate the size of the yPTP. After the swelling which follows opening was complete, polyethylene glycol (PEG) of varying molecular weight was added and shrinkage was observed. Shrinkage, under these conditions, indicates that the yPTP remains open with respect to mannitol permeation while the added PEG is excluded. Varying the molecular weight of PEG reveals a size exclusion property since no shrinkage is seen when the PEG can also penetrate (29). PEGs were added from stock solutions that were 300 mosM with respect to the osmotic pressure arising from PEG and that were buffered with 10 mM HEPES (Na+), pH 7.35. These solutions were isoosmotic with respect to the incubation medium and were added to 10% of the final volume. Approximately 30 mosM of the total osmotic pressure was then derived from PEG during shrinkage (26). Since PEGs are highly non-ideal solutes with respect to their osmotic characteristics, the concentration required to produce a particular osmotic pressure is a function of the PEG molecular weight (26). For the 300 mosM stock solutions, the following values were used, in units of mM; 199 (0.4 kDa), 172 (0.6 kDa), 136 (1.0 kDa), 111 (1.5 kDa), 93 (2.0 kDa), 58 (3.4 kDa), and 50 (4.0 kDa). The osmotic pressure of these solutions was confirmed using a vapor pressure osmometer.
Reagents and Other ProceduresPEGs were obtained from Aldrich or Sigma. Oligomycin, CsA, FCCP, antimycin A, and alamethicin were dissolved in dimethyl sulfoxide, rather than ethanol, since the latter solvent is a respiratory substrate for yMt. The procedures employed for electron microscopy were as described previously (32) except that the yMt were sedimented prior to fixation. The nucleotide profile of yMt was obtained by HPLC (8) of alkaline extracts.
Intramitochondrial
Ca2+ has a central role in regulating the mPTP (1, 3);
however, yMt lack a high activity Ca2+ uniporter (20, 33).
These circumstances complicate the interpretation of earlier studies
that reported the absence of an obvious permeability transition in yMt
exposed to Ca2+ (20, 21). We used the electrophoretic
Ca2+ ionophore ETH129 (34, 35) to determine if high
internal Ca2+ loads trigger a permeability transition in
yMt. A typical experiment is shown in Fig.
1A. When oxidizing ethanol in
a medium containing 10 mM Pi, yMt fail to
accumulate significant Ca2+ when 160 nmol/mg of protein is
added externally (Fig. 1A, trace a). However, a
rapid accumulation occurs upon the addition of ETH129, which continues
until a steady state is reached at approximately 10 µM
external Ca2+ (trace b). Significant swelling
does not occur under these conditions, indicating that the inner
membrane remains impermeant to low molecular weight solutes (data not
shown, and see Fig. 2). Indeed,
Ca2+ is retained for extensive periods (Fig. 1A,
trace b) and is released only upon uncoupling (trace
c) or when oxygen or the oxidizable substrate is depleted. Similar
data are obtained at Ca2+ loads of 400 nmol/mg of protein
or higher (data not shown).
Under comparable conditions, rat liver mMt, oxidizing succinate instead
of ethanol (Fig. 1B), do not retain Ca2+
regardless of whether initial uptake occurs via the
endogenous uniporter (trace a, ruthenium red absent
(RRed)), or via ETH129 (trace b,
ruthenium red present (+RRed)). The release of
Ca2+ from rat liver mMt under either condition is
accompanied by swelling and is inhibited by CsA (data not shown),
indicating that the mPTP is responsible. These contrasts indicate that
yMt do not contain a PTP that is induced by Ca2+ plus
Pi, unlike mMt.
However, the capacity of yMt to carry out ETH129 mediated
Ca2+ accumulation is dependent upon the medium
Pi concentration as shown in Fig. 2. In the presence of 10 mM Pi, Ca2+ is accumulated and
retained for over 50 min, but when Pi is reduced to 1 mM significant accumulation is not seen (Fig.
2A). Parallel experiments using safranine to estimate
showed that in the presence of 1 mM Pi, yMt
respiring on ethanol develop a small
which decays to no
measurable value within 60 s (detection limit ~80 mV). In
contrast, a large
is generated and maintained when the medium
contains 10 mM Pi (data not shown). The fact
that
is not maintained explains the absence of Ca2+
uptake in the medium containing 1 mM Pi, but it
is not clear immediately why yMt depolarize under that condition. CsA
does not substitute for the high Pi concentration,
suggesting that depolarization and the failure to accumulate
Ca2+ at 1 mM Pi is unrelated to a
yPTP (Fig. 2A). Nevertheless, parallel swelling and related
experiments yield the opposite conclusion, as follows.
While no swelling is associated with Ca2+ accumulation at
10 mM Pi, a CsA-insensitive swelling is
apparent at 1 mM Pi (Fig. 2B). The
magnitude of swelling, expressed as A540, is
relatively small, compared with mPTP-dependent swelling of
mMt (e.g. Ref. 26). However, it is comparable with the
swelling induced by alamethicin (compare Fig. 2B,
trace a with traces b and c), which is
an antibiotic that forms large pores in the mMt inner membrane (26,
36). These data are consistent with the presence of a CsA-insensitive
PTP in yMt that allows mannitol to enter the matrix space, with
swelling arising from an accumulation of mannitol/water driven by the
Donnan potential and the oncotic pressure differential. Fig.
3 confirms this interpretation by showing
that the swelling seen in Fig. 2B reflects ultrastructural
changes characteristic of the permeability transition in mMt (32, 37,
38). In addition, marked release of matrix nucleotides is also seen, as
expected if swelling results from the opening of a large and
solute-nonselective pore (e.g. Ref. 8). We interpret these
findings to indicate that yMt contain an endogenous PTP.
Respiration and ATP Induce Opening of the yPTP
Additional
experiments revealed that neither exogenous Ca2+ nor ETH129
are required to induce swelling of yMt and that the depletion of
endogenous Ca2+ using ionomycin plus an external chelator
is without effect (data not shown). Thus, yPTP opening is apparently a
Ca2+-independent process. However, respiration and
Pi play central roles in regulating the yPTP. Swelling
occurs slowly in the absence of Pi and an exogenous
respiratory substrate but accelerates dramatically upon the addition of
ethanol (Fig. 4A). The same
result is obtained when exogenous NADH is used in place of ethanol
(data not shown). Antimycin A prevents the accelerated swelling
produced by either substrate (data not shown) indicating that it is
respiration/energization that is responsible. On the other hand, ATP
can be substituted for ethanol or NADH, even in the presence of excess
oligomycin. Thus, the action of ATP on yPTP opening is unrelated to the
energetic state (Fig. 4A). Regardless of whether swelling is
spontaneous or promoted by respiration or exogenous ATP, it is strongly
antagonized by the presence of 10 mM Pi (Fig.
4A). The inhibitory action of Pi is
concentration-dependent, with a half-maximal effect
obtained near 1 mM, regardless of whether ethanol or ATP is
employed. (Fig. 4B).
In the absence of Pi, spontaneous and ATP-induced swelling
are pH-dependent, occurring faster as the medium pH is
increased. This is less apparent in the case of ethanol-induced
swelling, which occurs at similar rates as the medium pH is varied
(Fig. 5).
PEG-induced Contraction of Swollen yMt
The size of the mPTP
has been estimated by observing the ability of different size PEG to
inhibit swelling (26) or to induce contraction of previously swollen
mMt, (e.g. Ref. 29). We used the latter approach to further
examine the solute size-exclusion properties of the yPTP.
Ethanol-supported respiration or exogenous ATP was used to open the
yPTP initially. After swelling reached completion, PEG of various
molecular weights were added under isoosmotic conditions to determine
their effectiveness in causing contraction. Small PEG (0.4 and 0.6 kDa)
readily penetrate through the open yPTP since little or no contraction
is observed upon their addition (Fig.
6A). Contraction is seen with
larger PEG, indicating that they are excluded or pass through the yPTP
at a much slower rate. Although the rate of contraction declines for
PEG larger than 1.5 kDa (Fig. 6A), the extent of contraction that is ultimately attained is similar for PEG of 2 kDa and larger. Plotting the maximum extent of contraction versus PEG size
indicates a half-maximal effect at ~1.1 kDa when yMt have swollen
during respiration on ethanol (Fig. 6B) or in response to
ATP (data not shown). This value is virtually identical to that
reported for the PEG-induced contraction of beef heart mMt following
mPTP opening induced by Ca2+ (29). It is somewhat greater
than the values obtained by examining PEG-dependent
inhibition of swelling in rat liver mMt when mPTP opening is induced by
Ca2+ plus phosphate, or by mastoparan. In these cases, a
half-maximal effect is obtained at 0.65 and 0.95 kDa PEG, respectively
(26).
It is not apparent why the rate of contraction decreases progressively with PEG larger than 1.5 kDa, but the same effect is seen with mMt (31). A reduced rate indicates that the membrane has become less permeable to internal solutes. Thus, fewer PTPs may be open at a given time when the larger PEGs are present. Since pretreatment in the present experiments was the same for PEG of all sizes, the same number of yPTPs should be open initially. Accordingly, the larger PEG may reduce the yPTP open probability through a physical-chemical mechanism (39). Alternatively, the reduced rate of contraction might result from a decreased penetration of the larger PEG through channels in the outer membrane since the relationship between swelling and outer membrane rupture has not been determined for yMt.
The pH dependence of contraction induced by 1.5 kDa PEG in the absence
of Pi is shown in Fig. 7.
Although ethanol-induced swelling is relatively independent of pH (Fig.
5), the rate of subsequent PEG-induced contraction displays a
significant dependence (initial rate), increasing as the pH becomes
more basic (Fig. 7). This is also true for contraction following
spontaneous swelling (ethanol and ATP absent), however, contraction of
ATP-swollen yMt is relatively pH independent (Fig. 6). The modest
effect of pH on contraction in the presence of ATP is in contrast to
ATP-induced swelling, which is pH-dependent (Fig. 5). At
the highest pH examined (7.75), contraction is rapid regardless of how
an open PTP was initially attained.
Actions of Pi on the yPTP Are Related to Matrix pH
The effects of medium pH on the swelling and contraction of
yMt raise the possibility that the inhibitory action of Pi
results from acidification of the matrix space. As in mMt,
Pi is transported into yMt via a
Pi/OH antiport, resulting in matrix
acidification (40). To determine if a decreased matrix pH plays a role
in the inhibitory action of Pi, the actions of nigericin or
an uncoupler (FCCP) on yPTP opening were examined. Nigericin causes
matrix acidification by exchanging internal K+ for external
H+. Uncouplers produce acidification by collapsing the
H+ electrochemical gradient, which is otherwise present in
respiring mitochondria. As shown in Fig.
8A, uncoupler inhibits yPTP
opening resulting from respiration in the presence of ethanol
(trace c+FCCP versus trace c) or under the
spontaneous condition (trace a+FCCP versus trace a) but has
no effect on yPTP opening induced by ATP (trace b+FCCP versus
trace b). The pattern of uncoupler effects on the PEG-induced
contraction of swollen yMt is similar (Fig. 8B). Contraction
is inhibited when the initial swelling occurred in response to
ethanol-dependent respiration, regardless of whether uncoupler was present during swelling (trace c
+FCCP) or
absent during swelling but added shortly before PEG (trace c+FCCP
versus trace c). Uncoupler also inhibits contraction when swelling
occurred under the spontaneous condition (trace a+FCCP versus
trace a) but is without an effect when swelling was induced by ATP
(trace b+FCCP versus trace b). The effects of nigericin are
analogous to those of uncoupler in all cases (data not shown).
Mersalyl, an inhibitor of Pi/OH
and
Pi/dicarboxylate antiport in yMt (41, 42), does not prevent inhibition of ATP-induced swelling by Pi but diminishes the
inhibition that is seen when swelling is promoted by respiration or
occurs under the spontaneous condition (data not shown).
These results support a role for matrix acidification, via
Pi/OH antiport activity, in the mechanism by
which Pi antagonizes yPTP opening during respiration or
under the spontaneous condition. However, the failure of mersalyl to
diminish the inhibitory effect of Pi when the yPTP is
induced by ATP suggests that an external site is also be
involved.
Several agents that influence the mPTP were tested on yMt by monitoring swelling and subsequent PEG-induced contraction. As indicated above, CsA (up to 7 µg/ml) does not inhibit the swelling of yMt induced by ethanol or ATP or PEG-induced contraction in the presence of either agent. Neither ATP- nor respiration-induced swelling is affected by oligomycin (20 µg/mg of protein), which inhibits the yMt ATP synthase (43) and antagonizes induction of the mPTP by Ca2+ plus Pi (1). Carboxyatractyloside (25 µM), an inhibitor of the adenine nucleotide translocase (44) that favors opening of the mPTP (1, 8, 45), has little or no effect on ATP-induced swelling of yMt but inhibits ethanol-induced swelling by 10-15% (Fig. 8, A and B). DCCD (50 nmol/mg of protein), which can inhibit or activate the mPTP depending upon conditions (46, 47), decreased swelling and contraction after a relatively long preincubation period (60 min) but not following a shorter period (25 min). Mg2+ (2 mM), which favors the closed form of the mPTP (1, 8, 45), prevents the spontaneous swelling of yMt but not ATP-induced swelling (data not shown).
ADP favors the closed form of the mPTP (1, 8, 45) and was found to
inhibit both ethanol- and ATP-induced swelling of yMt, with a
half-maximal effect obtained at ~4 mM (Fig.
9). Carboxyatractyloside (25 µM) does not alter the inhibitory action of ADP when yPTP
opening is induced by ATP (Fig. 9B) but enhances the
effectiveness when opening is induced by respiration (Fig. 9A). In addition, ADP (4 mM) is as effective as
FCCP at inhibiting the spontaneous swelling that occurs in the absence
of ethanol or ATP (Fig. 9A). The rate of PEG-induced
contraction of ethanol-swollen yMt is somewhat inhibited by ADP, but
there is little effect on the contraction of yMt that have swollen in
response to ATP (Fig. 9D).
The present study demonstrates that a permeability transition can be induced in yMt, as indicated by swelling, PEG-induced contraction, ultrastructural changes, and the release of matrix space solutes of substantial size (nucleotides). Thus, these data resolve the controversy regarding the existence of a yPTP. The yPTP opens when yMt are allowed to respire, or are incubated in the presence of ATP, when Pi is absent (Fig. 4). The magnitude of the swelling response that follows is 20-25% of that typically seen with rat liver mMt under analogous conditions but coincides with other reports on the optical properties of swollen yMt (48). Reduced swelling is not due to a failure of a PTP to open in a large fraction of yMt as shown by ultrastructural changes, the marked release of matrix nucleotides (Fig. 3), and by the swelling response to alamethicin, which is the same as that produced by opening the yPTP (Fig. 2). We attribute the reduced swelling to ultrastructural differences between yMt and mMt. The former have relatively few cristae (e.g. Ref. 43 and Fig. 3), which limits the increase in matrix space volume, and the resultant change in A540, which is possible when the yPTP is opened.
The absence of a Ca2+ requirement and insensitivity to CsA appear to have been factors causing uncertainty in previous studies that sought to determine if yMt contain a PTP. These properties are analogous to the mPTP induced by mastoparan, which is also Ca2+-independent and CsA-insensitive under some conditions (26). Caution is warranted in such cases when ascribing swelling to a PTP because smaller channels could be responsible. Accordingly in this study, as with mastoparan (26), the solute size exclusion method was also used to characterize the PTP. The data obtained further support the existence of a yPTP, comparable in size with the mPTP and of the same size when induced by ethanol oxidation or ATP (Fig. 6).
Regulation of the yPTPThe present results show that at a minimum, the yPTP is regulated by adenine nucleotides, pH, and Pi. Regarding nucleotides, the site at which ATP acts to promote opening of the yPTP is directly accessible from the external volume, based upon the failure of carboxyatractyloside to influence the process. ATP binding may promote yPTP opening through an allosteric mechanism since there is no requirement for Mg2+, which is normally necessary for protein kinase activity and for the formation of regulatory compounds that arise from ATP. The inhibition of ATP-dependent yPTP opening by ADP can be explained as a competitive binding interaction at the putative allosteric site, with ADP being an ineffective promoter. This interpretation is supported by the failure of carboxyatractyloside to alter the potency of ADP as an inhibitor of yPTP opening induced by ATP (Fig. 9B). However, since ADP also inhibits yPTP opening induced by respiration (Fig. 9A), it is probable that ADP is an authentic inhibitor, as opposed to having a neutral effect when associated with the external site. In the case of yPTP induction by respiration, carboxyatractyloside enhances the inhibitory action of ADP (Fig. 9A). Thus, further investigation may show that adenine nucleotides influence the yPTP through both internal and external sites, as with the mPTP (8).
The influence of respiration and the actions of Pi and pH on the yPTP appear to be closely related, with matrix space pH representing the actual regulatory parameter. Both swelling and contraction data support this interpretation. As regards swelling, the acceleration that is seen as the medium pH rises (Fig. 5) could result from an action exerted on either side of the membrane. However, proton pumping during respiration increases matrix pH while the external pH is unaffected, and respiration promotes opening of the yPTP (Fig. 4A). In addition, nigericin, uncoupler, and external Pi inhibit yPTP opening induced by respiration, and these agents have in common the action of reducing the rise in matrix pH that results from respiration.
The rate of contraction also increases as the pH rises (Fig. 7), however, the relative rates in the presence of ethanol versus ATP are reversed at lower pH, compared with what is seen during swelling (Fig. 5). This difference also supports a central importance of matrix pH in regulating the yPTP, as follows. With ATP acting at an external site to cause opening through a mechanism that is relatively pH-insensitive, it is expected that the site would remain occupied after swelling is complete. In that case, PEG-induced contraction should proceed readily, as observed (Fig. 7), since the yPTP would remain open. No effect of uncoupler on the contraction rate would be expected, and none is seen (Fig. 8B). With ethanol promoting opening through an increase in matrix pH, a reduced open probability is expected after the PTP opens initially since the matrix pH would fall toward the medium value. A reduced and medium pH-dependent rate of contraction would then be expected and is observed (Fig. 7). With the yPTP subject to a reduced open probability and with respiration continuing, uncoupler might further inhibit contraction since it would reduce any transient recovery of an elevated matrix pH during periods when the yPTP was closed. This effect of uncoupler is also observed (Fig. 7B).
Thus, all aspects of the present data are consistent with respiration, medium pH, and Pi regulating the yPTP via changes in matrix pH. However, as noted under "Results," the inhibition of ATP-dependent yPTP opening by Pi when mersalyl is present suggests an external site of action for Pi, and further investigation of this point will be necessary.
Attempts to quantitate the relationship between yPTP opening and matrix
pH by fluorescent indicator methods were unsuccessful because neither
BCECF or cSNARF-1 was loaded into the matrix of yMt when presented as
membrane permeant esters. However, previous reports show that
Pi has protective effects on yMt under numerous conditions. For example, Verlours et al. used
electron microscopy to show that yMt oxidizing external NADH swell and
are otherwise altered when Pi is absent (48). Although
unrecognized at the time, the ultrastructural changes were reflecting
the permeability transition. The same study showed that ultrastuctural
changes are not seen in medium containing 10 mM
Pi and attributed that effect to a decreased matrix pH
(48). The acetate distribution method was used subsequently to show
that pH drops from ~0.8 to 0.3 units when 10 mM
Pi is added to respiring yMt (49). Thus, the change in
matrix pH that is required to induce or inhibit opening of the yPTP is
small, on the order of ± 0.5 units when the external pH is near
neutrality.
When regulation of the yPTP by nucleotides, Pi, and pH is
compared with the effects of these parameters on the mPTP, both similarities and differences are apparent. There are no reports of ATP
favoring the open form of the mPTP, although ADP favors the closed form
(8) in common with the yPTP. Basic pH has the same effect on the PTP
from both sources although, with the mMt, other factors are apparently
overriding since there is no requirement to reduce pH to maintain a
closed PTP in respiring mMt. Indeed, Pi favors opening of
the mPTP although it is not clear if this is a direct effect or related
to other regulatory parameters.
Of several other agents known to influence the mPTP by direct or indirect mechanisms, none were found to effect the yPTP in an analogous way. These include Ca2+, Mg2+, oligomycin, carboxyatractyloside, DCCD, and CsA. This suggests that regulation of the yPTP is greatly simplified compared with the mPTP. The lack of inhibition by CsA is of particular interest because with mMt it is thought that the potent inhibitory action of CsA reflects its binding to a matrix cyclophilin (5, 50, 51). This in turn has lead to the proposal that cyclophilin is a structural component of the mPTP (51) or catalyzes mPTP opening via its peptide bond isomerase activity (4, 8). Yeast contain at least five cyclophilins, one of which (cpr3p) is localized in yMt (52). This cyclophilin has been shown to mediate the folding of imported proteins in CsA-sensitive fashion (53). Thus, the failure of CsA to inhibit opening of the yPTP is not related to an absence of this target protein or to a failure of CsA to inhibit its peptide bond isomerase activity. It may indicate that the PTP in yMt and mMt are unrelated, or that the involvement of cyclophilin in forming/regulating the mPTP evolved relatively recently.
Possible Relationships between the yPTP, ATP-induced Ion Transport Pathways, and Ion-conducting Channels in yMtATP-induced ion transport pathways for cations and anions in yMt have been described by three different laboratories (54-58). Prieto et al. (56, 58) reported that ATP activates an anion-conducting pathway by binding to a site on the outer surface of the inner membrane. Guerin et al. (54) described an ATP-induced nonspecific pathway that gives rise to swelling in K+ or Na+ salts of gluconate, glutamate, chloride, or acetate when the medium pH is neutral. The ATP-induced swelling was inhibited by ADP, apparently by acting at the ATP-binding site on the external side of the inner membrane (54). Roucou et al. (55, 57) described an ATP-induced electrophoretic K+ transport pathway in respiring yMt. It differs from the one described by Prieto et al. (56, 58) in that it requires KCl. Unlike the nonspecific activity (54), the latter activity is specific for K+ over Na+ while the ATP regulatory site is located in the matrix (57). In similarity to the yPTP described here, each of these ATP-induced transport pathways is inhibited by Pi. Thus, it seems possible that one or more of these is related to the yPTP.
The same is true as regards the ion conducting channels in yMt that are observed by patch clamp techniques since the larger examples display conductance values similar to those of the mPTP (15, 21-23). In addition, ATP favors the open forms although the site is on the matrix side (21, 23), and additional regulatory features do not parallel those of the yPTP reported here. The substantial differences in techniques, preparations, and conditions that were used in these studies, and those involving ATP-regulated ion transport, discourage more specific cross-comparisons at present. However, it is worthwhile to note the mPTP displays substates that are solute selective under some circumstances (4, 59). Thus, further investigation of the yPTP and its relationship to other ion transporters and channels in yMt is warranted.
Potential Physiological FunctionsAs with the mPTP, the physiological function of the yPTP is not clear, but it is useful nevertheless to consider possibilities. The differences in regulatory features of the yPTP and mPTP may indicate that the functions are not equivalent and may also reflect the differing environmental challenges faced by yeast and mammalian cells. If only the regulators identified in this study are considered, the yPTP would be expected to open if the cell attained a high phosphorylation potential. During growth on a nonfermentable carbon source, the resultant uncoupling could provide a mechanism to dispose of excess reducing equivalents generated during carbon flux into the biosynthetic pathways. Rial and co-workers (56, 58) propose this function for their ATP-regulated proton conducting pathway, which may be related to the yPTP (or a substate) as noted above. Lohret and Kinnally (15) propose that the yeast megachannel is involved in protein import since leader peptides decrease the open probability. This possibility also applies to the yPTP since these structures may again be related. Additional regulators of all these structures will probably be identified and such studies should help to test these potential functions and interrelationships.
It is also possible the yPTP remains closed during rapid growth but is present to facilitate a response to changing conditions, such as a change of carbon source or the onset of a nutrient deficiency. Such environmental factors initiate complex signaling cascades (60) that cause the dynamic adjustment of mitochondrial properties, including structure, metabolic capacity, location, and numbers per cell (61, 62). In cells grown on a nonfermentable carbon source, for example, yMt represent 10-13% of cell volume but only ~3% in cells grown on glucose (62). During adaptation, only 1 h of glucose repression produces a fraction of yMt that are swollen and contain few cristae (63). Although the underlying mechanisms are not known, opening the yPTP may well be involved since this produces a collapse of ion gradients and swelling. In this way, the yPTP might function as the target of signals that terminate yMt function and initiate degradation/nutrient recycling.
A role for the yPTP in yeast cell death can also be envisioned since mPTP is implicated, in mammalian cell apoptosis (e.g. Refs. 17-19), where members of the Bcl-2 protein family influence opening of the mPTP (64) and the progression to death (65, 66). The same proteins expressed in yeast can kill cells or rescue them, suggesting that components of mammalian system are present (19, 67). As shown here, these components include a mitochondrial PTP.
The versatility of yeast as an experimental system and the genomic sequence data that are now available should greatly facilitate further studies on the function and structure of the yPTP.
We thank Drs. Tien-Hsien Chang and Bernard Trumpower for providing the yeast strains used in this study; Dr. Kathy Jung for helpful discussions and advice on techniques; Donald Ordaz and Warren Erdahal for technical assistance; and Drs. Kemal Baysal, Gerald Brierley, and Kimberly Broekemeier for helpful discussions and critical reading of the manuscript.