Functional Studies with the Octameric and Dimerie Form of Mitochondri~l Creatine Kinase DIFFERENTIAL pH-DEPENDENT ASSOCIATION OF THE TWO OLIGOMERIC FORMS WITH THE INNER MITOCHONDRIAL MEMBRANE*

Phosphate extraction of mitochondrial creatine kinase (Mi-CK, EC 2.7.3.2) from freshly isolated intact mitochondria of chicken cardiac muscle, after short swelling in hypotonic medium, yielded more than 90% of octameric and only small amounts of dimeric Mi-CK as judged by fast protein liquid chromatography-gel permeation analysis of the supernatants immediately after extraction of the enzyme. In extraction buffer, octameric Mi-CK displayed a tendency to dissociate, albeit at a slow rate with a half-life of approximately 3-5 days, into stable dimers. Experiments with purified Mi-CK octamers or dimers, or defined mixtures thereof, incubated under identical conditions with Mi-CK-depleted mitoplasts revealed that both oligomeric forms of Mi-CK can rebind to mitoplasts. However, the association of Mi-CK was strongly pH-dependent and, in addition, octameric and dimeric Mi-CK showed different pH dependences of rebinding. Therefore, it was possible under certain pH conditions to rebind either both oligomeric forms or selectively the octamers only. Furthermore, evidence is presented that Mi-CK dimers partially form octamers upon rebinding to the inner membrane. The differential association of the two oligomeric Mi-CK forms with the inner mitochondrial membrane together with the dynamic equilibrium between octameric and dimeric Mi-CK (Schlegel, J., Zurbriggen, B., Wegmann, G., Wyss, M., Eppenberger, H.M., and Wallimann, T. (1988) J. Biol. Chem., 263, 16942-16953) suggest that both oligomeric forms are physiologically relevant. A change in the octamer to dimer ratio may influence the association behavior of Mi-CK in general and thus modulate mitochondrial energy flux as discussed in the phosphoryl creatine circuit model (Wallimann, T., Schnyder, T., Schlegel, J., Wyss, M., Wegmann, G., Rossi, A.-M., Hemmer, W., Eppenberger, H.M., and Quest, A.F.G. (1989) Prog. Clin. Biol. Res. 315, 159-176.

Phosphate extraction of mitochondrial creatine kinase (Mi-CK, EC 2.7.3.2) from freshly isolated intact mitochondria of chicken cardiac muscle, after short swelling in bypotonic medium, yielded more than 90% of octameric and only small amounts of dimeric MI-CK as judged by fast protein liquid chromato~aphy-gel permeation analysis of the supernatants immediately after extraction of the enzyme, In extraction buffer, octameric MbCK displayed a tendency to dissociates albeit at a slow rate with a half-life of approximately 3-6 days, into stable dimers.
Experiments with purified Mi-CK octamers ox dimers, or defined mixtures thereof, incubated under identical conditions with Mi-CK-deplete mitopl~ts revealed that both oligomeric forms of Mi-GK can rebind to mitoplasts. However, the association of Mi-CK was strongly pa-dependent and, in addition, octameric and dimeric Mi-CK showed different pW dependences of rebinding. Therefore, it was possible under certain pW conditions to rebind either both oligomerie forms or selectively the octamers only. Furthermore, evidence is presented that &Ii-CK dimers partially form oetamers upon robindi~g to the inner membrane.
MgADP-f PCr2-+ H' g MgATP2-+ Cr Three cytopiasmic isoenzymes, composed of two different subunits, have been described (1,2): MM-&K (M standing for the muscle-type CK subunit); BB-CK (B standing for the brain-type CK subunit); and the heterodimer~c form MB-CK found transiently during differentiation of skeletal muscle and permanently in adult mammalian heart (3). These isoforms of CK are expressed tissue specifically and localized subcellularly in an isoenzyme-sp~i~c manner (for review see Refs,4 and 5). CR activity was also found in the mit~ho~drial fraction where mitochondrial CK or Mi-CK (6) is localized on the outer surface of the inner mitochondrial membrane (7). Recently, two Mi-CK isoenzymes were described which are expressed tissue specifically as well (8)(9)(10). The more basic Mib-CK and the more acidic MI,-CK isoforms are accumulated in mitochondria of cardiac muscle and brain, respectively (9). The different cytosolic and mit~hon~ia1 CK isoenzymes are thought to be involved in energy buffering and transport of "energy-rich" phosphoryl compounds as described in the phosphocreat~~e shuttle model (11)(12)(13)(14) as well as in the regulation of local subcellular ATP levels as discussed within the PCr circuit model (15).
Two int~rconve~ible oligomeric forms, an octameric and a dimeric form of MGCK, have been isolated and characterized both from cardiac muscle ~Mi~-~K) and brain (Mix-OK (9,10,16)). Since no obvious difference in specific enzyme activity was observed between the two oligomeric forms, the possible functional significance of these two molecular structures with respect to their interaction with the inner m~tochondrial membrane was investigated. A number of authors have studied the solubilization and reassociation of Mi-CK from and with mitoplasts, respectively, long before anything about the existence of the octameric Mi-CK species was known (11,(17)(18)(19)(20)(21)(22)(23). From these results it became clear that the extraction as well as the rebinding of Mi-CK was strongly dependent on the source of mitochondria and on the buffer conditions used, Under certain in vitro conditions a multiple of the quantity of Mi-CK released from a given amount of mitochondria could be rebound to extracted homologous mitoplasts or even to liver mitoplasts that do not show any Mi-CK activity in vivo (18,23). These and further results (22) indicated that mitochondria from a variety of tissues show low as well as high affinity binding sites for Mi-CK. Marcillat et al. (24) presented evidence that only the high A4, form of Mi-CK with a molecular weight of 350,000 did rebind to extracted mitoplasts but not the lower M, or dimeric form. In two independent experiments, these authors tried to rebind either dimeric or octameric Mi-CK from crude extracts to mitoplasts, but the reassociation conditions chosen were different for the two forms. Rebinding experiments with octamers were done at pH 7.4, whereas pH 8.8 and p-chloromercuribenzoate were used for rebinding of dimers. As we know now, pH 7.4 is favoring rebinding of Mi-CK in general, that is octamers as well as dimers. In contrast, alkaline pH (25) and sulfhydryl reagents (19) are conditions known to release chicken or rat Mi-CK from mitoplasts. Thus, conclusive results about the rebinding behavior of the two oligomeric forms of Mi-CK are still missing.
We have shown recently that the octamer/dimer equilibrium of Mi-CK can be influenced in vitro and that at low protein concentrations the addition of substrates and cofactors inducing a transition state analogue complex leads to a complete conversion of octamers into dimers (10). In addition, it is now possible to generate mixtures of octameric and dimeric Mi-CK with relatively stable ratios of the two forms.
In order to investigate the association behavior of Mi-CK octamers and dimers with the inner mitochondrial membrane, we performed a series of experiments with highly purified Mi-CK octamers, dimers, or mixtures thereof with a concomitant quantitative analysis of the octamer to dimer ratio before and after the experiment using FPLC-gel permeation chromatography. With these new tools at hand, one was able to study in detail the reassociation behavior of Mi-CK octamers and dimers to mitoplasts under controlled conditions. Strong evidence for an association of both oligomeric Mi-CK forms with mitoplasts was found, but at the same time clear functional differences in the association behavior between Mi-CK octamers and dimers were discovered. This may be relevant for regulation of energy transfer from mitochondria to cytosolic compartments where part of the cytosolic CK is specifically associated with subcellular structures (26) as discussed within the framework of the PCr circuit model (15).
Parts of this work have been presented as abstract (27) Laemmli (29). The isoelectric points (IEP) of Mi-CK from brain and heart were determined by isoelectric focusing on IEF 3-9 gels with the PhastSystem'" of Pharmacia and with the IEF calibration kit "broad" (Pharmacia, Sweden) as reference.

Extraction of Creatine Kinase from Freshly Prepared Intact
Heart Mitochondria-The analysis of a fresh phosphate extract at pH 8.6 of intact swollen mitochondria on a FPLC-gel filtration column (Superose 12) with subsequent measurement of the CK activity of the different fractions (Fig. IA) clearly showed that Mi-CK analyzed immediately after extraction consists of 90% octameric and only 10% dimeric Mi-CK (Table I). The extraction conditions (phosphate buffer at high pH) did not favor octamer formation, since after 2 days 23% (Table I Table I). at pH 8.6 with a half-life of approximately 2 days (not shown). It is therefore unlikely that under these conditions the high percentage of octamers uersw dimers found immediately after extraction was due to immediate formation of octamers from dimers during the very fast extraction step. This octamerdimer dissociation process was also not due to denaturation of the enzyme or degradation of the protein, since the enzymatic activity of Mi-CK remained constant during this time (Table I). This also indicates that the octamer-dimer dissociation has no effect on the specific activity of the enzyme. On SDS-polyacrylamide gels only one major band, corresponding to the intact polypeptide of Mi-CK with an apparent M, of 43,000, and a very faint band at M, 84,000 corresponding to the Mi-CK dimer (10) were seen which reacted on an immunoblot with specific anti-Mi-CK antibodies (not shown; see also Ref. 10).
Selective Rebinding to P,-extracted Chicken Heart Mitoplasts of Octameric Mi-CK Out of a Mixture Containing Both Dimeric and Octameric Mi-CK-Purified and enzymatically active Mib-CK consisting of a mixture of both oligomeric forms was bound under a variety of conditions to freshly prepared extracted mitoplasts. The sample of Mi-CK chosen for the rebinding experiment consisted of 85% octameric and 15% dimeric Mi-CK as determined by gel filtration analysis on a FPLC-Superose 12 column ( Fig. 2A). Using this fast gel permeation technique, which was completed in less than 30 min, it was found that almost 100% of the Mi-CK from this sample bound to the mitoplasts at pH 7.0 (Table II), whereas at pH 8.1 and 8.8 only 87 and 19%, respectively, of the total Mi-CK was rebound (Table II).
After rebinding of a mixture of dimeric and octameric Mi-  CK to mitoplasts at pH 8.1, gel permeation analysis of the supernatant showed (Fig. 3A) that all the octameric Mi-CK (85% of the total Mi-CR present in the sample) had bound to the mitoplasts whereas the dimeric Mi-CK (15%) remained in the supernatant. This agrees well with the fact that 13% of the total Mi-CK activity measured did not bind to mitoplasts at pH 8.1 (Table II).
If mitoplasts, to which the 85/15% octamer/dimer mixture of Mi-CK was bound, were re-extracted and the extract thereof analyzed by gel permeation chromatography, mostly octameric Mi-CK was re-extracted as seen by the CK activity profile in Fig. 33. Since by this second extraction of the mitoplasts a number of other proteins are also released in small quantities by alkaline phosphate buffer, the absorbance profile at 280 nm (Fig. 3B, upper part) does not reflect Mi-CK protein only. The absorbance shoulder (OO,,,) at the Mi-CK dimer position is due to extracted proteins other than Mi-CK which eluted at a similar position but did not show any CK activity as shown on the lower panel of Fig. 3B. The finding that these fractions contain only very little MGCK but a number of other proteins was verified by SDS-PAGE and immunoblot analysis (not shown). With these rebinding and re-extraction experiments it was clearly proven that at pH 8.1 Mi-CK, which had bound to the mitoplasts, was of octameric nature (Fig. 3B), whereas the dimeric Mi-CK did not bind under these conditions (Fig. 3A).
The reason why dimeric Mi-CK did not bind to mitoplasts at pH 8.1 was not due to inactivation, denaturation, or degradation of the dimeric Mi-CK fraction for (i) the very same Mi-CK sample (shown in Fig. 2A) which was used for rebinding was readily converted into 295% octameric Mi-CK by simply concentrating the Mi-CK sample (Fig. 2B); (ii) the dimeric fraction present in the mixture (Fig, 24) did bind to mitoplasts at pH 7.0 (Table II), was enzymati~ly active (Table II and Fig. 3A), and had the same specific activity as octameric Mi-CK; and (iii) sodium dodecyl sulfa~-polya~ryIamide gels with samples of Mi-CK shown in Fig. 2.4 displayed one single band only at the &i, position of 43,000 with no indication of proteolytic de~adation (not shown).
The specificity of the rebinding of octamers only from a mixture cont~ning both octamers and dimers, observed at alkaline pH (PH 8.1, Fig. 3), did not depend on the relative proportion of the two forms in the mixture added to extracted mitoplasts (see below).
~~anti~at~a~ of tize Re~i~i~ of separated D~~er~ and ~~ta~r~~ Mi-CK in. De~~e~~e of the pH Value-In order to study the rebinding of the two oligomeric Mi-CK forms, dimers or octamers, separately but under identical eqmimentul conditiotzs, samples which consisted of 990% octameric Mi-CK or of 78% dimeric Mi-CK were used. Samples containing this very high percentage of octamers or dimers were obtained by concentrating the mixture shown in Fig. 24 or by the addition of MgADP, creatine, and nitrate, respectively, as described under "Materials and Methods." However, in order to rebind the two different Mi-CK samples under exactly the same conditions the buffer of the samples was changed within 10 min to rebinding buffer by desalting on a FPLC fast desalting column HR lO/lO (Pharmacia~ directly before rebinding to the mitoplasts (see "Materials and Methods"). As shown in a parallel experiment by gel filtration analysis on FPLC, the ratio of dimeric to octameric Mi-CK in these desalted samples was not altered during the time course of the rebinding experiment. rebinding itself was performed at the same protein concentrations in the very same rebinding buffer containing 10 mM sodium phosphate and 5 mM BME. Rebinding was measured at different pH values.
The rebinding experiments summarized in Fig. 4 were completed after 3-5 min of incubation of extracted mitoplasts with Mi-GK. Control experiments showed that, at all pH values used, no change at all in the octamer-dimer ratios occurred within this time range in the supernatants or with isolated Mi-CK in solution. The rebinding experiments showed that octameric Mi-CK Q----Cl) was quantitatively rebound to extracted chicken heart mitoplasts in rebinding buffer between pH 7.0 and 8.1. However, under identical conditions, as far as protein concentration, buffer composition, and pH were concerned, dimeric Mi-CK (0. . * .o) did rebind quantitatively only at pH 7.0 with the percentage of rebound dimers decreasing drastically with increasing pH, e.g* at pH 7.5 and 8.1 only 80 and 22X, respectively, of the sample containing 78% dimeric Mi-CK were rebound (Fig. 4). The 22% of Mi-CK activity which did rebind at pH 8.1 correspond exactly to the fraction of octameric Mi-CK in the sample, indicating that at pH 8.1 dimeric Mi-CK did not rebind at all to extracted mitoplasts, whereas octameric ML-CK, under the same conditions, bound to nearly 100% (Fig. 4). These results correspond well to the data shown in Fig. 3.
If mitoplasts, to which the mixture of 78% dimeric and 22% octameric Mi-CK was rebound at pH 7.0, were re-extracted ;;I. ., ., . . ., , __ ,>-",I only (0. . .O) to phosphate-extracted washed mitoplasts using either a preparation containing 290% octameric or 78% dimeric Mi-CK, respectively, is indicated as a function of pH and was measured in percent of CK activity recovered in the mitoplast pellet after incubation of the mitoplasts with the two Mi-CK forms (see also Table  II). At the same time, CK activity of the supernatants and total CK activity were measured. In addition, supernatants and extracted pellets of some experiments were quantitated in terms of the octamer/ dimer ratio by gel permeation chromatography as described for the experiments shown in Fig. 3. No change in the octamer/dimer ratios was observed by incubation of Mi-CK at the pH values indicated over the time course of the rebinding experiments.
at pH 8.8 (as above) after 30-60 min and the extract analyzed by FPLC-gel permeation chromatography, the supernatant of the extract contained about 50% dimeric and 50% octameric Mi-CK (not shown here). This finding indicates that some of the dimeric Mi-CK did form octamers while bound to the mitoplasts. This provides evidence that under these experimental conditions some octamerization of dimeric Mi-CK can occur on the inner mitochondrial membrane, probably by frequent collision of dimers on the membrane.

Extraction of Freshly Purified Mitochondria from Chicken
Heart-Phosphate extraction of freshly prepared swollen chicken heart mitochondria yielded 90% octameric Mi-CK in the supernatant (Fig. lA and Table I). Since the extraction was done under conditions favoring dimer formation (see Fig.  1, B and C, and Table I), it is very unlikely that Mi-CK was bound in dimeric form to the mitochondria before extraction and subsequently would have formed octamers during extraction. The same holds true for rabbit heart mitochondria (2.4, 30), whereas from mitochondria of rat brain' and pig heart (30) a significant proportion of dimeric Mi-CK is also extracted. Furthermore, our rebinding and re-extraction experiments with mostly dimeric Mi-CK at pH 7.0 showed that the contact of dimeric Mi-CK with the inner mitochondrial membrane induces the formation of some octameric Mi-CK probably while attached to the inner membrane, because after rebinding of 78% dimeric Mi-CK to the mitoplasts, approximately 50% dimeric Mi-CK was also re-extracted. This new result is a further indication that the extraction process per se is not the reason for the fact that in a phosphate extract of freshly prepared mitochondria mostly octameric Mi-CK was found. From all these experiments it can be concluded that in intact heart mitochondria Mi-CK is present mostly in octamerit form.
We have determined the M, of isolated octameric chicken ' V. Adams, personal communication.
cardiac Mi-CK in vitro by different methods to be between 320,000 and 360,000 (10,16). This is in good agreement with the recent findings of Quemeneur et al. (31) and of Lipskaya et al. (32) who provided evidence by radiation inactivation measurements of Mi-CK and cross-linking experiments with glutaraldehyde, respectively, that in isolated mitochondria Mi-CK is an oligomer with a M, of about 330,000-380,000. Rebinding of Mi-CK to Extracted Chicken Heart Mitoplasts-At pH 7.0, both oligomeric forms of Mi-CK were rebound to extracted chicken heart mitoplasts, whereas at slightly alkaline pH, octameric Mi-CK had a clearly higher affinity to the inner mitochondrial membrane than dimeric Mi-CK (Fig. 4). Due to this fact, it was possible to selectively rebind octameric Mi-CK out of a mixture of both dimeric and octameric Mi-CK back to extracted mitoplasts (Fig. 3). It was shown that both forms of Mi-CK, dimeric and octameric Mi-CK, were enzymatically active showing identical specific enzyme activities, that they were not degraded, and that both forms did rebind at pH 7.0 to extracted mitoplasts. Therefore, it can be concluded that the observed difference in binding behavior between octameric and dimeric Mi-CK was not a consequence of an experimental artifact due to degradation or aggregation of Mi-CK. Thus, dimeric and octameric Mi-CK differ not only in their size but depending on the pH also in their ability to bind to the inner mitochondrial membrane. In contrast, Marcillat et al. (24) claimed that only octameric Mi-CK could rebind to mitoplasts and Lipskaya and Trofimova (32) argued that the octamer is the only form bound to mitochondrial membranes during cross-linking. The data available on the IEP of dimeric and octameric Mi-CK are somewhat conflicting and may reflect species differences. Whereas by IEF the IEPs of native rabbit heart Mi-CK were shown to be 8.83 and 8.24 for the octamer and dimer, respectively (30), a higher IEP (9.67) was reported for pigeon breast muscle Mi-CK dimers than for octamers (8.93) (33). In accordance with the first report (30) we found IEPs of 8.7-9.0 and 8.4-8.5 for chicken brain octamers and dimers and of 9.4-9.5 and 9.3 for chicken heart octamers and dimers of Mi-CK, respectively, as determined by IEF on the PhastSystem'". Native octameric Mi-CK from rabbit and chicken seems to have a somewhat higher IEP compared with native dimeric Mi-CK. Since the interaction of Mi-CK with the inner mitochondrial membrane is at least in part of ionic character for it is easily extractable by phosphate at alkaline pH (lo), the observed difference in IEPs may to some extent be the reason for the very specific rebinding of octamers over dimers at slightly alkaline pH. At pH 7.0 and below, that is at pH conditions observed in muscle under a heavy work load, the specificity of rebinding is lost so that both oligomeric forms can rebind to the inner mitochondrial membrane.
From the facts mentioned above it may be assumed that dimeric and octameric Mi-CK are two distinct forms of Mi-CK, both with their own physiological role. It seems very likely that the importance of these two forms in uiuo is related to the regulation of the energy transfer from mitochondria to the cytosol, depending on the energy requirement of the cell. However, since both oligomeric forms of Mi-CK have the same specific activity in vitro, the regulation of the energy transfer in uiuo is unlikely to take place simply via the formation of different oligomeric states but perhaps may occur in combination with other relevant physiological parameters, e.g. by interaction of Mi-CK with mitochondrial membranes.
An interaction of Mi-CK with mitochondrial substructures and proteins of the inner and outer mitochondrial membranes is very likely, because Adams et al. (34) showed that Mi-CK from rat brain and kidney is accumulated within the contact Interaction of Creatine Kinase with Mitochondrial Membranes sites between inner and outer mitochondrial membranes. This was also confirmed by direct immunoelectron microscopy where we found an Immuno-gold labeling with anti-Mi-CK antibodies along the cristae membranes as well as an additional accumulation of Immuno-gold labeling at those sites where inner and outer mitochondrial membranes are close, presumably at the mitochondrial contact sites (10). The localization of CK at the contact sites has recently been demonstrated also by histochemical methods (35). In addition, it was shown by morphometric measurements that the extent of both contact sites and CK activity present at these sites increased upon stimulation of cardiac muscle tissue (35,36). This is in good agreement with recent results showing that the proportion of dimeric to octameric Mi-CK within mitochondria seems indeed to be regulated by the respiratory condition of the mitochondria and as such also by the energy requirement of the ce11.3 Model of Regulation of Mi-CK Activity within Mitochondria-The facts that (i) Mi-CK octamers as well as dimers bind to mitoplasts, but the latter with lower affinity and in a strongly pH-dependent manner, that (ii) dimers, once bound to the inner membrane, seem to be able to form octamers, e.g. by frequent collision while attached to the membrane (see results of the experiment involving re-extraction of rebound Mi-CK dimers), and that (iii)  these results indicate that due to changes in the immediate environment beyond the contact sites, dimers floating freely in the intermembrane space get bound, octamerize on the membrane, and accumulate at the contact sites. These features are depicted in a model shown in Fig. 5. Thus, in this model, Mi-CK is accumulated in respiring mitochondria at the contact sites (10,34,35). There, as some recent evidence points out, Mi-CK is thought to interact with the outer mitochondrial membrane pore protein (34,37), the inner membrane adenine nucleotide translocator (ANT) (11,23,34), and probably also with different lipids of both membranes, e.g. cardiolipin (A), a prominent phospholipid of the inner mitochondrial membrane (38,51). Mi-CK would receive ATP from ANTS, transphosphorylate freely diffusible uncharged creatine that can enter through the pore at the contact sites (39), and release the produced PCr into the cytosol. ADP produced by the creatine kinase reaction would be passed from Mi-CK back to the nearest ANT that would transport it through the inner membrane in exchange for ATP produced by oxidative phosphorylation.
In the model presented in Fig. 5, the functional coupling of Mi-CK and ANT and/or mitochondrial respiration, which is described in many publications (l&13,40-48), could be based on protein-protein interactions where Mi-CK at the contact sites is in close proximity or even in direct contact with the pore protein of the outer and the ANT of the inner membrane. Such a tightly regulated microcompartment would function as an efficient energy channeling unit (10,15,16). The protein-protein interactions within a common multienzyme microcompartment with close proximity or direct contact of the partners within the contact sites make possible an efficient functional coupling between Mi-CK and ANT (11,16,23,45) on one hand and Mi-CK and the pore protein on the other hand (37,49,50).
This would also explain why the functional interaction of these components is abolished by disturbing the integrity of the outer mitochondrial membrane, a phenomenon observed by several investigators (e.g. Refs. 47 and 48). The model not only shows a state fixed in time of PCrproducing mitochondria but also indicates the possibilities of regulation by influencing different equilibrium reactions involved, for example breaking up the contact sites (equilibrium 1, Fig. 5), which are known to be regulated via an unknown pathway depending on the respiratory state of mitochondria (35, 36), could lead to interference of the functional interaction of Mi-CK with ANT and with pore protein, which would then lead to a reduced PCr-production of the mitochondria. The contrary events would take place upon stimulation of mitochondria accompanied by formation of contacts (36) and accumulation of octameric Mi-CK at these sites (16,34). Furthermore, influencing the dimerloctamer equilibrium of Mi-CK (equilibrium 2, Fig. 5) could lead to accumulation of Mi-CK at the contact sites or to the "removal" of Mi-CK from the contact sites or even (see "Results") to the extraction of Mi-CK from the inner mitochondrial membrane (equilibriums 3 and 5, Fig. 5).
Thus, by influencing one of the five equilibrium constants described in Fig. 5 which in concert can generate a complex network of regulation based on the mutual influence of at least five different equilibrium reactions, the regulation of Mi-CK activity and energy transport in vivo could OCCUR by these parameters. The above equilibrium reactions could be influenced by one or more of as yet unknown regulatory substances, by post-translational modification of Mi-CK itself, or simply by changing the concentration of the CK substrates and of the pH value in the mitochondrial intermembrane space. The latter two factors have been shown in vitro to strongly affect the octamer/dimer equilibrium (10) as well as the association of the two oligomeric forms of Mi-CK with the inner mitochondrial membrane (see "Results"). Since lowering of pH, as observed in ischemia or metabolically stressed muscle, favors the association of Mi-CK dimers with the inner mitochondrial membrane, the pH value within the intermembrane compartment of mitochondria which is supposed to change locally as a function of the respiratory state may be an important regulatory factor influencing the proportion of free to bound Mi-CK. Both forms of Mi-CK, dimeric and octameric Mi-CK, each emerge to play an important physiological role within the mitochondrial energy metabolism whereby the octameric form of Mi-CK is likely to be the "active form" at the contact sites in respiring mitochondria. Furthermore, the involvement of contact sites forming multienzyme microcompartments for efficient transfer of energy is also emerging. This complex regulatory network allows mitochondria to adjust the rate of PCr production to the energy requirements of the cell as illustrated by the PCr circuit model (15).