(Received for publication, February 20, 1997, and in revised form, April 16, 1997)
From the Division of Biochemistry, Department of Molecular and Experimental Medicine, The Scripps Research Institute, La Jolla, California 92037
The redox reactions of the bis-heme cytochrome b of the ubiquinol:cytochrome c oxidoreductase complex (complex III, bc1 complex) were studied in bovine heart submitochondrial particles (SMP). It was shown that (i) when SMP were treated with the complex III inhibitor myxothiazol (or MOA-stilbene or stigmatellin) or with KCN and ascorbate to reduce the high potential centers of complex III (iron-sulfur protein and cytochromes c + c1), NADH or succinate reduced heme bL slowly and incompletely. In contrast, heme bH was reduced by these substrates completely and much more rapidly. Only when the complex III inhibitor was antimycin, and the high potential centers were in the oxidized state, NADH or succinate was able to reduce both bH and bL rapidly and completely. (ii) When NADH or succinate was added to SMP inhibited at complex III by antimycin and energized by ATP, the bis-heme cytochrome b was reduced only partially. Prereduction of the high potential centers was not necessary for this partial b reduction, but slowed down the reduction rate. Deenergization of SMP by uncoupling (or addition of oligomycin to inhibit ATP hydrolysis) resulted in further b reduction. Addition of ATP after b was reduced by substrate resulted in partial b oxidation, and the heme remaining reduced appeared to be mainly bL. Other experiments suggested that the redox changes of cytochrome b effected by energization and deenergization of SMP occurred via electronic communication with the ubiquinone pool. These results have been discussed in relation to current concepts regarding the mechanism of electron transfer by complex III.
Our recent studies (1) have revealed features of the electron transfer system of bovine mitochondrial ubiquinol:cytochrome c oxidoreductase (complex III, bc1 complex) that are incompatible with the Q cycle hypothesis (2-5). (i) It was shown that in SMP1 the reoxidation of the bis-heme cytochrome b of complex III could be inhibited by either antimycin or myxothiazol. The inhibition by either reagent was incomplete, and heme bL (b566) was oxidized through the leak allowed by either inhibitor at least 10 times faster than heme bH (b562). (ii) Cytochrome b of complex III could be partially reduced via cytochrome c1 and the Rieske iron-sulfur protein (ISP) by ascorbate, or faster and to a greater extent by ascorbate plus TMPD. This reaction was inhibited more strongly by antimycin than by myxothiazol (see also Ref. 6). (iii) Ascorbate or ascorbate plus TMPD could also partially reduce b in ubiquinone-depleted bovine heart SMP, in which the molar ratio of Q10 to complex III monomer had been reduced 200-fold from 12.5 to 0.06, or in SMP from a Q-deficient yeast mutant. These results agreed with the finding of others that in bovine or yeast complex III the removal of Q by extraction did not impair the oxidation of b via ISP/c1 by ferricyanide (7, 8). (iv) Our results also showed that antimycin and myxothiazol, which exert their maximal inhibition at concentrations stoichiometric to complex III monomer, each inhibited three reactions of the bis-heme cytochrome b, all incompletely. The strongest effect by either reagent was inhibition of electron transfer from heme bH to heme bL; next was inhibition of the reoxidation of bL via ISP/c1; and least was inhibition of substrate (NADH or succinate) reduction of b, which is known to require the combined actions of both antimycin and myxothiazol (1, 2, 4). In these regards, HQNO (at 20-40 µM) behaved similarly to antimycin, and MOA-stilbene and stigmatellin similarly to myxothiazol (1). It should be added here that the oxidation of reduced bH and bL or their reduction by reverse electron transfer via ISP/c1 was not inhibited to a greater extent when SMP were treated with both antimycin and myxothiazol (1).
The above results were summarized in a scheme, which is reproduced in
Fig. 1A. It was emphasized that this scheme serves only to
present our new results in graphic form and should not be considered as
a new hypothesis for the mechanism of electron transfer by complex III.
The reason for emphasizing this point was that Fig. 1A does
not explain an important feature of complex III, namely the
oxidant-induced extra reduction of cytochrome b, which was a
major consideration in the design of the Q-cycle (Fig. 1B)
(9-11). This report addresses this feature of complex III.
NADH was obtained from Calbiochem. Ascorbic acid and TMPD were from Fluka. ATP was from Pharmacia Biotech Inc. Potassium ferricyanide was from Fisher Chemicals. Antimycin A and FCCP were from Sigma. Myxothiazol was from Boehringer Mannheim. MOA-stilbene (12) was the generous gift of Dr. G. von Jagow, Universitatsklinikum, Frankfurt, Germany. The sources of other chemicals were as indicated elsewhere (1, 13, 14).
Preparation of Bovine SMPSMP were prepared from bovine heart mitochondria in the presence of 1.5 mM ATP during sonication as described previously (13). Protein concentration was determined by the method of Lowry et al. (15).
AssaysReduction of cytochrome b was monitored spectrophotometrically at 563 minus 575 nm in a buffer containing 0.25 M sucrose, 5 mM MgCl2, and 50 mM Tris-HCl, pH 8.0. SMP concentration was 1.2-1.3 mg/ml. Results were essentially the same at 565 minus 575 nm. Reduction of cytochromes c + c1 was monitored at 550 minus 540 nm. At this wavelength pair, interference due to the absorbance changes of the b hemes was negligible. KCN, ascorbic acid (neutralized with NaOH), TMPD, deoxycholate (neutralized with KOH) were added where indicated at 10 mM, 0.4 mM, 5 µM, and 0.05%, respectively. Inhibitors of complex III were added from an ethanolic solution at the concentrations indicated in the figure legends. Ethanol concentration never exceeded 1%. Absorbance changes and spectra were recorded using an SLM DW2000 dual wavelength spectrophotometer. The data shown were collected and stored in a computer on line to the spectrophotometer. Assay temperature was 30 °C unless otherwise stated.
According to the Q-cycle
hypothesis (Fig. 1B), QH2 at the
QO site delivers one electron to ISP, 2H+ to
the outside, and one electron to bL. Electron
transfer to ISP must occur first, so that Q· ,
which is considered to be the electron donor to
bL, is generated. Therefore, if ISP is
prereduced and antimycin is added to prevent b reduction via
the Qi site, QH2 or a respiratory substrate
should be unable to rapidly reduce the two b hemes (2, 4).
However, as will be seen, our results are difficult to reconcile with
this expectation. The experiments were carried out with bovine heart SMP for the reasons stated previously (1), and the reduction of
cytochrome b was monitored at 563 minus 575 nm. The
particles were treated with KCN to inhibit cytochrome oxidase, and 0.4 mM neutralized sodium ascorbate was added to reduce
cytochromes c and c1. In control
experiments, the reduction of c + c1
by ascorbate was monitored spectrophotometrically, and the time
required after ascorbate addition to achieve complete reduction of
c + c1 was determined before
proceeding with the experiments reported here.
Fig. 2, traces A and B, show the
results with antimycin as the complex III inhibitor. It is seen in Fig.
2A that addition of succinate to SMP, pretreated with KCN,
ascorbate and antimycin, resulted in rapid reduction of cytochrome
b followed by a slow further reduction. When shortly after
the start of the slow phase, ferricyanide was added to oxidize
c + c1, the slow reduction phase was
changed to a rapid one (Fig. 2B). This ferricyanide-induced effect is what is known as the oxidant-induced extra reduction of
cytochrome b, which was observed in the early studies on
complex III (16, 17) and subsequently rationalized in the design of the
Q cycle (2, 4, 9-11). It may be noted, however, that the initial rapid
reduction of cytochrome b by succinate does not agree with
the Q cycle.
Fig. 2C shows an experiment similar to that of Fig.
2B, except that the complex III inhibitor used here was
myxothiazol, instead of antimycin. It is seen that upon addition of
succinate a biphasic reduction of cytochrome b was observed,
again with an initial rapid phase followed by a slow phase. When
ferricyanide was added at the plateau of the slow reduction phase, a
rapid partial oxidation of cytochrome b took place,
corresponding in extent to the slow reduction phase. This was followed
by a slow re-reduction (Fig. 2C). Spectral analyses showed
that the succinate-induced rapid initial reduction phase was due mainly
to the reduction of bH (Fig. 3
dashed trace, max at ~563 nm), and that the
ferricyanide-induced oxidized heme, in addition to c + c1, was bL (Fig. 3, solid
trace, cytochrome b
max at ~566 nm).
When myxothiazol was omitted in an experiment similar to that of Fig.
2C, addition of succinate to KCN + ascorbate treated SMP
resulted in a biphasic b reduction essentially identical to
the biphasic b reduction phase of Fig. 2C (data
not shown). In addition, data similar to those of Fig. 2 were obtained
when, instead of succinate, NADH was used as the reductant, except that
in an experiment similar to that of Fig. 2B it was necessary
for the molar equivalent of NADH to be greater than that of
ferricyanide. Otherwise NADH was rapidly oxidized by ferricyanide at
the level of complex I, resulting in b oxidation.
Another point of clarification regarding the effects of antimycin and
myxothiazol on the redox reactions of the b hemes of complex
III is afforded by the data of Fig. 4. In these
experiments, no KCN and ascorbate were added. The oxidation of
cytochrome b was blocked by antimycin in Fig. 4, trace
A, and by myxothiazol in Fig. 4, trace B. It is seen
that addition of succinate resulted in a rapid and complete reduction
of the b hemes in the presence of antimycin, and in a
biphasic and incomplete reduction of these hemes in the presence of
myxothiazol. The difference in the spectra of the plateau regions of
traces A and B, shown in Fig. 4, trace C, indicated that the heme incompletely reduced in the presence of
myxothiazol was bL (Fig. 4, trace C,
max at ~566 nm).
Previous results of others have shown a slow reduction of both
bL and bH in preparations
pretreated with antimycin and ascorbate (±TMPD). In one case, a
succinate-cytochrome c reductase preparation was used (18)
(see also Ref. 19), which had been obtained by fractionating the
mitochondria in the presence of 10 mg of cholate per ml (20). In
another case, CO-treated mitochondria were used in the absence or
presence of 6.7 mM ascorbate, with 0.3 mM
succinate as the electron donor and 3-30 mM malonate,
which greatly inhibited succinate oxidation (17). Respiration was then
initiated by flash photolysis of the CO-inhibited cytochrome oxidase.
We have also found that 5-10 times higher ascorbate concentrations than used in our experiments diminished the rate of
bH reduction in experiments as in Fig. 2.
However, as mentioned earlier, ascorbate concentrations higher than 0.4 mM were not necessary to achieve complete reduction of
c + c1 in the above experiments.
Another point that requires further explanation here concerns the data of Fig. 2C, where it is seen that addition of ferricyanide
to SMP, pretreated with KCN, ascorbate, myxothiazol, and succinate, resulted in reoxidation of the bL component of
the reduced cytochrome b. In other experiments, it was found
that this partial b reoxidation by ferricyanide took place
also with either MOA-stilbene or stigmatellin as the complex III
inhibitor and regardless of whether the SMP were pretreated with KCN
and ascorbate (data not shown). Furthermore, as seen in Fig.
5, when cytochrome b was reduced in
myxothiazol-treated SMP, then antimycin was added to block electron
transfer from succinate to b, subsequent addition of
ferricyanide resulted in the reoxidation of both
bL and bH. Similar
results were obtained with intact SMP in the absence of deoxycholate
(0.05%, see legend to Fig. 2) in the reaction mixture. It is possible
that modification of cytochrome b by myxothiazol,
MOA-stilbene, or stigmatellin makes it susceptible to oxidation by
ferricyanide. In addition, the data of Figs. 2C, 3, and 5
indicate that bH is in rapid electronic communication with complexes I and II regardless of whether the complex
III inhibitor is antimycin or myxothiazol, even when c + c1 are prereduced. In other words, the rate of
bH reduction by succinate appears to supersede
that of its oxidation (directly or indirectly) by ferricyanide. These
data are in complete agreement with our previous results, which showed
that the step most strongly inhibited by antimycin or myxothiazol is
electron transfer from bH to
bL (1). Together, the above results also
indicate that only in the presence of antimycin is
bL rapidly and completely reducible by
substrates, and that this rapid substrate reduction of
bL in the presence of antimycin can be inhibited
by prereduction of the high potential centers of complex III.
Effect of SMP Energization on the Reduction of bH and bL by NADH or Succinate
In an experiment such as that of Fig. 2B, it was necessary to treat SMP with a detergent to allow ferricyanide access to c + c1 on the interior surface of the SMP vesicles. Therefore, for uniformity, all the experiments of Figs. 2, 3, 4, 5 were carried out in the presence of the minimal necessary amount (0.05%) of deoxycholate. Under these conditions, 0.4 mM ascorbate was sufficient to completely reduce c + c1, and the addition of TMPD was not necessary. In the experiments to be reported in this section, intact, well coupled SMP were used. Under these conditions, it was necessary to add TMPD (5 µM) together with 0.4 mM ascorbate to achieve complete reduction of c + c1.
As seen in Fig. 6, SMP were treated with KCN, antimycin,
ascorbate, and TMPD. Then ATP was added to energize the system, and NADH was added to reduce the b hemes of complex III.
Cytochrome b reduction was monitored, as before, at 563 minus 575 nm. It is seen that the addition of NADH resulted in partial
reduction of cytochrome b, and subsequent addition of the
uncoupler FCCP resulted in greater b reduction. As shown in
Fig. 7A, the order of addition of ATP and
NADH could be changed, with NADH added before ATP. A greater reduction
of b was achieved when NADH was added in the absence of ATP,
but subsequent ATP addition caused partial b oxidation,
which was reversed upon addition of FCCP. Similar results were obtained
when the respiratory substrate was succinate instead of NADH (Fig.
7B), or when the complex III inhibitor was myxothiazol
instead of antimycin (data not shown), except that in these cases the
extent of reoxidation of b induced by ATP addition was less.
Spectral analysis of the plateau regions of Fig. 7A
suggested that the component remaining reduced after ATP addition is
mainly bL (max at 564.5 nm). Fig.
8, A and B, shows experiments
similar to Fig. 7, A and B, except that
myxothiazol was added where indicated before (Fig. 8A) or
after (Fig. 8B) the addition of ATP. It is seen that the
combination of antimycin and myxothiazol prevented the ATP-induced
reoxidation (Fig. 8A) and the FCCP-induced re-reduction
(Fig. 8B) of cytochrome b. In view of the known
fact that the combination of antimycin and myxothiazol inhibits the
rapid reduction of bH and
bL by QH2 (1, 2, 4), the data of
Fig. 8 suggest that in the absence of myxothiazol ATP addition results
in reverse electron transfer from the b hemes to the Q pool,
and uncoupling in the reversal of this process.
The experiment of Fig. 6 is somewhat analogous to that of Fig.
2B in the sense that in an antimycin-treated system
substrate-induced reduction of the cytochrome b of complex
III was partially inhibited, then the partial inhibition was reversed
by the addition of ferricyanide in Fig. 2B, and of FCCP in
Fig. 6. However, the fact that prereduction of c + c1 is not required for the ATP/uncoupler effect is
shown in Fig. 9. In this experiment SMP were treated
only with antimycin. Then ATP was added, followed by NADH. It is seen
that NADH addition resulted in partial, but in this case more rapid,
b reduction, and subsequent addition of oligomycin to
inhibit ATP hydrolysis (or uncoupler, data not shown) resulted in
further b reduction. Then, after exhaustion of NADH,
cytochrome b was oxidized in a biphasic manner, as
demonstrated previously (1, 21).
Therefore, by analogy to the phenomenon of oxidant-induced extra reduction of cytochrome b, which is rationalized by the design of the Q cycle, one could term the results of Fig. 6 and 9 as deenergization-induced extra reduction of cytochrome b. However, as will be discussed below, it is also difficult to rationalize the data of Figs. 6, 7, 8, 9 on the basis of the Q cycle hypothesis.
The results of two sets of experiments have been reported here. One set concerns substrate reduction of hemes bH and bL in SMP under conditions that b oxidation was inhibited by antimycin or myxothiazol, and the high potential centers of complex III (ISP and c + c1) were prereduced by ascorbate. In the second set, the effect of SMP energization and deenergization on the reduction of bH and bL was explored under conditions that b oxidation was inhibited by antimycin or myxothiazol, and the high-potential centers of complex III were either prereduced or not.
In the first set of experiments, it was found that when myxothiazol was the complex III inhibitor addition of succinate or NADH resulted in rapid and complete reduction of bH, and slow and incomplete reduction of bL, regardless of whether or not c + c1 were prereduced (Figs. 2C and 4B). Similar results on the substrate reduction of bH and bL were obtained in the absence of myxothiazol when c + c1 were prereduced (data not shown). The ferricyanide reoxidation data of Fig. 2C indicated that when the inhibitor was myxothiazol, ferricyanide oxidized only the partially reduced bL, but not bH. However, when antimycin was added after succinate (Fig. 5), and the combination of myxothiazol and antimycin inhibited electron transfer from succinate to b, then ferricyanide rapidly oxidized both bL and bH. These results suggested that, under the conditions of Fig. 2C, ferricyanide was still capable of oxidizing both b hemes. However, in this case the rate of bH reduction, but not of bL reduction, by succinate superseded its rate of oxidation (directly or indirectly) by ferricyanide. This interpretation of the data of Figs. 2C, 3, and 5 agrees with our previous finding that the strongest inhibition exerted by myxothiazol (or antimycin) is on electron transfer from bH to bL (1).
The biphasic reduction of the b hemes in the experiment of Fig. 2C, including the partial reduction of bL, can also be explained by the Q cycle, because with only myxothiazol as the complex III inhibitor (or when myxothiazol was absent, but c + c1 were prereduced) the Qi site of the Q cycle would still be open for reduction of the b hemes, and the lower EM of bL would be responsible for its slow and partial reduction via bH. However, when SMP were pretreated with KCN, ascorbate and antimycin (the Qi site inhibitor of the Q cycle), addition of succinate or NADH also resulted in a much faster reduction of bH than bL (Fig. 2A). These results are still compatible with those shown in Fig. 1A, but not with the Q cycle (Fig. 1B), because under the conditions of Fig. 2A both the QO and the Qi pathways for b reduction would be inhibited, and neither bL nor bH should have been rapidly reduced.
In summary, the results of the first set of experiments suggest that prereduction of ISP/c1 results mainly in inhibition of the reduction of bL by NADH or succinate. This effect is similar to that of myxothiazol (or MOA-stilbene or stigmatellin). It differs from the effect of myxothiazol in that the combination of myxothiazol and antimycin severely inhibits substrate reduction of bH as well, but the combination of prereduced ISP/c1 and antimycin does not (Fig. 2A). Comparing Fig. 2, A and C, it also seems that the combination of ISP/c1 prereduction and antimycin addition has a greater (additive?) effect on the inhibition of bL reduction than the combination of ISP/c1 prereduction and myxothiazol addition.
The second set of experiments (Figs. 6, 7, 8, 9) have shown that in
antimycin-treated SMP energization by ATP alters the manner in which
NADH or succinate reduces the b hemes of complex III. Only
partial b reduction occurred in ATP-energized SMP, but
subsequent addition of an uncoupler or oligomycin (to inhibit ATP
hydrolysis) resulted in further b reduction (Figs. 6 and 9).
Prereduction of c + c1 was not
necessary for partial b reduction in ATP-energized SMP, nor
for further b reduction upon SMP deenergization. However, prereduction of c + c1 did lower the
rate at which b was reduced by substrates. The effect of SMP
energization and deenergization on the extent of b reduction
could also be seen after the b hemes had been reduced by
NADH or succinate (Fig. 7). Furthermore, addition to the
antimycin-treated SMP of myxothiazol before or after ATP addition
inhibited the subsequent effects of energization and deenergization
(Fig. 8), suggesting that energization results in partial b
oxidation by the Q/QH2 pool and deenergization in further
b reduction by the same source. Analysis of the spectra of
the plateau regions of the experiment of Fig. 7A before and after ATP addition suggested that the b heme remaining
reduced after ATP addition was mainly bL (major
peak
max at 564.5 nm), which agrees with early
reports that energization of mitochondria by ATP raises the apparent
EM of bL to +245 mV at pH
7.0 (22, 23). The altered redox properties of cytochrome b
in energized SMP merit detailed investigation, because this
transmembranous component of complex III may be involved in proton
translocation (Fig. 1A).
The results shown in Figs. 6, 7, 8, 9 are difficult to reconcile with the Q cycle hypothesis. One would expect on the basis of the Q cycle that substrate reduction of b would be inhibited in SMP treated with antimycin to block the Qi site; with KCN, ascorbate and TMPD to reduce ISP/c1 and prevent the single-electron oxidation of QH2; and with ATP which by creating a protonmotive force should inhibit QH2 deprotonation at the QO site. However, our results shown in Fig. 6 (see also Figs. 7 and 8) are not compatible with this expectation. Furthermore, one would expect from the Q cycle that in a system such as above reverse electron transfer from b would also be inhibited. This expectation is again not fulfilled by our results (Figs. 7 and 8). By contrast, the results of Figs. 6, 7, 8 are not inconsistent with Fig. 1A, because in the presence of either antimycin or myxothiazol the path of electrons between the Q pool and the b hemes would be open, and ATP-induced reverse electron transfer would be expected to register on the redox state of cytochrome b.
In conclusion, the data presented here make it difficult to rationalize the phenomenon of oxidant-induced extra reduction of cytochrome b on the basis of the Q cycle hypothesis. This phenomenon appears to result from two conditions: (a) presence of antimycin (or HQNO), because only under this condition can both bH and bL be completely reduced by substrates and (b) oxidized state of ISP/c1 (or absence of myxothiazol, MOA-stilbene, or stigmatellin), because otherwise bL would be slowly and incompletely reduced by substrates. Therefore, when condition (a) is fulfilled and restriction (b) is removed, bL (as well as bH) becomes reduced by substrate faster and to a greater extent. The reason for the partial and slow reduction of bL when c + c1 are prereduced or when complex III is treated with myxothiazol, MOA-stilbene or stigmatellin remains to be investigated. So does the reason for partial b reduction under the energized conditions of Fig. 6. Therefore, although Fig. 1A is more consistent than the Q cycle (Fig. 1B) with the results presented here and previously (1), much remains to be done before a complete picture emerges. The anticipated crystal structure of bovine complex III (24) should pave the way toward this goal.
This is publication 9895-MEM from The Scripps Research Institute, La Jolla, CA.
We thank C. Munoz for the preparation of bovine heart mitochondria.