Carbonic anhydrase activity in mitochondria from rat liver.

An 18O exchange method has been used to determine the location of carbonic anhydrase in mitochondria from rat liver and to examine the role of this enzyme in the kinetics of CO2 in resting and respiring mitochondria. Using digitonin subfractionation, we have determined that a substantial fraction, 40 to 60%, of the carbonic anhydrase activity in the mitochondrion from rat liver is located in the space between the inner and outer membranes; the remaining activity was found in the matrix with no detectable activity in the sedimented membranes. The total catalytic CO2 hydration activity measured in intact mitochondria from rat liver was about 1% of that found in an equal volume of rat erythrocytes. The apparent permeability constant representing the barrier for the diffusion of HCO3(-) from external solution to intramitochondrial carbonic anhydrase, 9 X 10(-5) cm s-1, is near in magnitude to the permeability constant for the diffusion of HCO3(-) across the rat erythrocyte membrane, 4 X 10(-4) cm s-2. Calcium-induced respiratory jumps were shown to cause changes in the rate of 18O exchange between CO2 and H2O that were consistent with a net uptake of CO2 by the mitochondria.

The presence of carbonic anhydrase in mitochondria from various tissues has been the subject of several investigations (1)(2)(3)(4)(5)(6)(7)(8)(9). Rat liver mitochondria have received the most attention (1-7) but the reports have been mixed. Some investigators have failed to detect any carbonic anhydrase activity in the mitochondrial fraction of rat liver homogenates (7). Others detected a small amount of activity but concluded that this was due to contamination from the soluble fraction and erythrocytes (1). Most investigators, however, agree that the small but significant amount of carbonic anhydrase activity (2-10% of total activity of rat liver homogenates) recovered in the mitochondrial fractions is of mitochondrial origin (2, 5, 6). It is important to note that the carbonic anhydrase from the mitochondrid fraction of rat liver is more susceptible to inhibition by certain sulfonamides than is the cytosolic enzyme; these are most likely different isozymes (3). Carbonic anhydrase has also been detected in mitochondrial fractions from other tissues including dog liver (3, 4), guinea pig liver, and skeletal muscle (8,9), rat kidney, and cerebral cortex (2). The mitochondria from guinea pig liver have been found to contain a considerable amount of activity, about %o that of erythrocytes (8).
The exact location within the mitochondrion and the role of the mitochondrial carbonic anhydrase require further This work was supported by Grant GM 25154 from the National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. study. It has been postulated that this is a membrane-bound enzyme (2,5,6), whose physiological role is to provide a rapidly reacting buffer in the matrix space (5). Others suggested that it is an enzyme found in the matrix which could catalyze the hydration of CO, and provide a counterion, bicarbonate, or carbonate, for the accumulation of calcium by respiring mitochondria (10). The activity of mitochondria from guinea pig liver was found to be soluble, not bound to sedimented membranes (9).
We have used an " 0 exchange method to determine the location of carbonic anhydrase in mitochondria from rat liver and to examine the role of this enzyme in the kinetics of CO, in both resting and respiring mitochondria. The main advantages of the "0 method in studying these properties are that it is a sensitive method for the detection of carbonic anhydrase activity and that it may be used to determine intracellular carbonic anhydrase activity and the membrane barriers to diffusion of bicarbonate (9,11,12). Using digitonin subfractionation, we have determined that a substantial fraction (40 to 60%) of the carbonic anhydrase activity in the mitochondrion of rat liver is located in the space between the inner and outer membranes; the remaining activity was found in the matrix with no detectable activity in the sedimented membranes. The total catalytic, COZ hydration activity measured in intact mitochondria from rat liver was about 1% of that found in an equal volume of rat erythrocytes. Moreover, the apparent permeability constant representing the barrier for diffusion of HC03-into a mitochondrion is near in magnitude to the permeability constant for the diffusion of HC03-across the red cell membrane.

EXPERIMENTAL PROCEDURES
Solutions for Preparation of Mitochondria-Solution 1 contained 220 mM d-mannitol, 70 m sucrose, 2 m M Hepes,' 0.5 mg/ml of bovine serum albumin (solution la was solution 1 without bovine serum albumin, solution Ib was solution 1 with 1% Triton X-100 (octyl phenoxy polyethoxyethanol)). Preparation of Mitochondria-Mitochondria were isolated from livers of male, albino Sprague-Dawley rats using the procedure of Schnaitman and Greenawalt (13). Mitochondria intended for digitonin subfractionation studies were isolated using solution la, as suggested by Kun et al. (14). All other mitochondria were prepared using solution 1. EDTA (1 IIMI) was included in the homogenization step when the mitochondria were intended for respiration studies. The mitochondrial pellet was washed three times, each time resuspending the pellet with a glass rod by gentle swirling. Mitochondrial preparations were occasionally checked for possible microsomal contamination by comparing their NADPH-cytochrome c reductase content (15) to that of microsomes prepared from the same livers. The mitochondrial preparations had a reductase activity (per mg of total protein) that was less than 10% of the reductase activity of preparations of purified microsomes.
Lysis of Mitochondria and Separation of Mitochondrial Membranes-Lysis was accomplished by resuspending the mitochondrial pellet in solution l b and shaking well. Mitochondrial membranes were separated by centrifuging detergent-lysed mitochondria at I00,cwH) x g for 1 h. The pellet (outer and inner membrane fragments) and the supernatant fluid (matrix and intermembrane space) were iested separately for carbonic anhydrase activity.
Digitonin Subfractionation-Freshly prepared mitochondria were resuspended in solution la at 80 to 100 mg of protein/d and divided into several portions. These were treated with various amounts of digitonin as described by Schnaitman and Greenawalt (13) and centrifuged at 9OOO X g for 15 min. The pellets were washed once and the supernatant fluids from both centrifugations (original and washing) were pooled together. The fluffy layer was included with the supernatant fraction. The pellets, resuspended in solution lb, and supernatant fluids from each treatment were assayed for carbonic anhydrase, protein, malate dehydrogenase, and monoamine oxidase content. On one occasion, the samples were also assayed for succinatecytochrome c reductase content. The recovered amount of enzyme activities and total protein after each digitonin treatment was usually in the range of 90-110%. Because of this variability in total recovery, the activity found in the pellet from each treatment was expressed as the fraction of the activity recovered in the pellet and supernatant fluid, rather than the fraction of the original total activity. At the concentrations used (see Fig. 2), digitonin did not interfere with any of the assays. Most digitonin treatments were repeated with four different mitochondrial preparations.
Protein Determination-Mitochondrial protein was determined by the method of h w r y et al. (16) using bovine serum albumin as the standard. The samples were diluted to the appropriate concentration range with 0.5 N NaOH. A spectrophotometric difference method (17) was also used periodically, especially for quick determinations. The two methods gave similar results (within 10%). Determinations were made in triplicate.
Enzyme Assays-Monoamine oxidase was assayed by following the oxidation of benzylamine spectrophotometrically at 250 nm at 37 "C as described by Schnaitman et al. (18). Each determination was repeated four times. Malate dehydrogenase was assayed by determining the rate of NADH formation at 340 nm at 30 "C as described by Ochoa (19). Each determination was repeated four times. Succinatecytochrome c reductase was assayed as described by Sottocasa et al. (20). Each determination was repeated four times.
Carbonic anhydrase activity was measured by the l80 exchange technique described by Silverman et al. (11). Because of the hydration-dehydration cycle, oxygen 18 in COz and HC03-is exchanged with oxygen 16 of water. This method involves the measurement of the isotopic content of 180-labeled COP with a mass spectrometer equipped with a COZ inlet. The COZ inlet is described in a previous publication (11); it is a vessel, the bottom of which is a membrane permeable to COz. Gas passing across the membrane enters a mass spectrometer, providing a continuous monitor of isotopic content of COz in solution. The uncatalyzed rate of "0 exchange was measured by adding "0-enriched bicarbonate (final concentration 25 mM) to solution 2 and adjusting the pH to 7.4. Four or five half-times for the chemical reaction elapsed before measurements were made; this provided COZ and HC03-concentrations close to their equilibrium values. The volume of the reaction mixture was, at this point, 8.0 ml.
After the uncatalyzed rate was measured, the mitochondrial sample, prepared in solution 2a (0.3 to 1.0 ml; 50 to 100 mg of protein), was added. Samples from the digitonin experiments were also assayed with the "0 exchange technique; they were prepared in solution la with 0.1% Triton X-100 and 25 m~ potassium bicarbonate. The solution in which mitochondria were prepared (Pa) was identical with that in the COZ inlet vessel; thus when intact mitochondria were added to the CO? inlet vessel, the chemical composition of the suspending solution was not altered, but the isotopic composition was altered. Due to the relatively large volume of the suspension of mitochondria (up to 1.0 m l ) added to the COz inlet vessel (containing 8.0 d) there was considerable dilution in the "0 content of the COz species in the vessel. Measurement of the catalyzed rate of exchange was therefore begun well after mixing effects had subsided. This prevented detection of any kinetic effects immediately (10 to 20 S) after addition of mitochondria. The pH of 7.4 was monitored throughout the experiment and found to vary within the range of k0.02 pH unit. The temperature of the reaction medium was maintained at 25 * 0.5 "C. In order to determine the extent of lysis in

Mitochondria from
Rat Liver 685 1 experiments with intact mitochondria, the reaction mixture was retrieved at the end of each run, centrifuged, and the supernatant tested for carbonic anhydrase activity. The supernatant was usually found to be contaminated with carbonic anhydrase due to the lysis of about 5-6% of the mitochondria. Carbonic anhydrase activity of respiring mitochondria was measured in the following manner. The uncatalyzed rate was recorded with solution 3 containing 25 mM "0-labeled bicarbonate. Freshly prepared mitochondria resuspended in solution 3a were then added and the catalyzed rate was recorded. Respiratory jumps were induced by the addition of small volumes (0.05 to 0.15 d) of 2.0 M CaC12 or SrClz solutions to give final cation concentrations of 10 to 33 EIM (approximately 2 to 10 pmol/mg of protein). Control experiments were carried out as above by adding bovine carbonic anhydrase, rat erythrocytes, or detergent-lysed mitochondria instead of intact mitochondria. (The concentrations of these preparations were chosen to give oxygen 18 exchange rates which were comparable to those obtained with intact mitochondria.) After 30 min in solutions containing 33 mM Ca2' or S8+, the suspensions of mitochondria showed no more lysis than suspensions maintained in the absence of Ca2+ and Sr2+. Inhibition of carbonic anhydrase was achieved using M methazolamide (KI -2 X lo-@ M).
The "0 content of HC03-was determined by periodically withdrawing small samples (50.1 m l ) from the reaction medium during measurements of the catalyzed rate of "0 exchange. The samples were immediately acidified and the isotopic content of the COZ generated was measured on the m a s spectrometer. The "0 content of HCOP-was calculated from the l80 content of COZ generated by acidification, the "0 content of COZ alone as determined from the COZ inlet, and the fraction of total COZ that is bicarbonate calculated from the Henderson-Hasselbalch equation.
Volume of Mitochondria-The total volume of all mitochondria in suspension was correlated with protein content and estimated in the following manner. The weight of mitochondria was estimated from protein content of the suspension by assuming that rat liver mitochondria contain 70% protein (20, 21) and that the density of mitochondria is 1.0 mg/ml. This gives a value of 1.4 yl/mg of protein. This can be compared with the value 1.0 pl/mg of protein which has been assumed for the mitochondrial matrix The first order rate constants B and y can be expressed as the sum of catalyzed and uncatalyzed components: The effect on the I8O exchange processes of Equation 1 and on the kinetic expressions of Equation 3 caused by the addition of cells or organelles containing carbonic anhydrase is presented elsewhere (9, 11, 12, 23). The various chemical and diffusion steps involved are shown in Fig. 1. Except for the case of respiring mitochondria, chemical and diffusion equilibrium is assumed; that is, chemical equilibrium is assumed inside and outside the mitochondrion and the flux into the mitochondrion of both COZ and HCO3-is equal to the flux out. In the case of red cells (11, 12, 24) and mitochondria from the livers of guinea pigs (9), the large intracellular concentration of carbonic anhydrase and the rapid access of COZ to that carbonic anhydrase causes a biphasic decrease in Tand cca. In the initial rapid phase ( t1,2 < 5 s), the COZ in external solution enters cells and is rapidly depleted of l8O inside the cells. In the subsequent slower phase ( t t / z -10 to 100 s) the "0 content of external HC03is depleted both by the slower diffusion of HCO3into cells and by the uncatalyzed conversion of HCOJinto COZ outside the cells.
The rate constants of Fig. 1 are most completely described in Ref. 23. The pseudo first order rate constant k, describes intramitochondrial hydration of COn and k', describes dehydration of HCOs-;' the rate constants k and k' describe the corresponding reactions outside the mitochondrion. These constants are related by the equilibrium constant K = (HCOJ-)(H+)/(COZ); that is, k, = k', K/(H+)n, where subscript 2 refers to inside the mitochondrion. The rate constants k, and k', describe the rate of entry of Con and HC03into mitochondria, and can be measured when there is a difference in "0 content between external and internal fluids. Thus, the rate of "0 loss from COn in external fluid due to the flux of COO into a mitochondrion is given by k , ( T~ -7,). The complete set of kinetic eouations describing "0 loss iS given by TU et at. i12). - The rate constant k', for diffusion of HC03into mitochondria and k', for diffusion out can be related to an apparent permeability constant of the mitochondrion to HC03-. This represents an overall barrier for the diffusion of HC03from external fluid to intracellular sites of carbonic anhydrase and does not differentiate between diffusion into the intermembrane space or into the matrix. This permeability constant P is given by (23) where V, is the total volume of all mitochondria in suspension and V,/A is the ratio of volume to surface area of a mitochondrion. For this purpose, the mitochondrion was assumed to have a cylindrical shape of radius r = 2.5 X cm (21, 25). The ratio of volume to The velocity of the catalyzed dehydration at chemical equilibrium, which equals the velocity of the catalyzed hydration, is given by carbonic anhydrase 11. Then k', for dehydration is given by the curved surface area of a cylinder is r/2, which we have used in these calculations. VZ was estimated from the mitochondrial protein content as described under "Experimental Procedures." The ratio of volumes also relates k', to k'i and k, to k,: k', = ( Vl/Vz) k', (23), valid for nonrespiring mitochondria at diffusion equilibrium.

RESULTS
Localization of Carbonic Anhydrase- Fig. 2 shows the release of protein, carbonic anhydrase, and several enzymes used as markers for different mitochondrial fractions from the 9000 X g pellet of rat liver mitochondria after treatment with various concentrations of digitonin, which causes the destruction of the outer membrane at lower concentrations (60.1 mg/ mg of protein) and the inner membrane at higher concentrations. Monoamine oxidase serves as a marker for outer membrane, succinate-cytochrome c reductase for inner membrane, malate dehydrogenase for the matrix, and protein for matrix and inner membrane (13).
No carbonic anhydrase activity as determined by "0 exchange rates could be detected in suspensions of mitochondrial membranes (inner and outer) isolated from detergenttreated mitochondria.   8 inhibit by 50% the value of Beat in solutions containing human carbonic anhydrase I1 (the high activity isozyme). In contrast, concentrations of Inear 1 m~ inhibit Beat by 50% in the presence of the lower activity isozyme, human carbonic anhydrase I. In this single respect, the inhibition property of the enzyme from rat mitochondria resembles that of carbonic anhydrase 11. These data were used to obtain solutions to the complete equations describing the processes in Fig. 1 In the experiments with mitochondria, the known quantities are B and y for the single phase of "0 loss observed (this assumes we can make reasonable estimates of intramitochondrial pH, ionic strength, and volume). Our procedure was to assume a reasonable value for k, also, and then calculate k', and k". An apparent permeability constant of the mitochondrion to HCOs-, P, was then found using Equation 5.

Composition of suspending solution (solution 2 under "Experimental
In Table I are listed values of k'c, k'L, and P for HCOs-using the numerical analysis of Tu et al. (12) for an intramitochondrial pH of 7.5, a pK of 6.11 for the C02-HC03-equilibrium, and a radius of 2.5 X cm for a cylindrical mitochondrion. Using this same pH, pK, and radius, calculations were performed using values of B and y from six different mitochondrial preparations. The results (means and standard deviations) are: k', = 1.4 & 0.3 s-', k', = 7.6 +-2.9 s-l, P for HC03-= (9.5 k 3.6) X cm s-l. Table I1 shows the effect on the calculation of uncertainties in assumed values of mitochondrial radius, intramitochondrial pH, pK for the COZ-HCOS-equilibrium, and value of ki. It is apparent from Table I1 that the rate of diffusion of COz out of mitochondria is so much faster than the intramitochondrial hydration rate that COZ diffusion is not a rate-determining event. Thus, a wide variation in the rate constant k , produces no change in the results of the numerical analysis. I 8 0 Exchange in Suspensions of Respiring Mitochondria-Respiratory jumps were induced by addition of CaClz to suspensions of mitochondria. Fig. 3 shows typical results; the 18 TABLE I1  Table I) To assess the effect of uncertainties in this calculation, we have performed the numerical analysis at different values of r, pH2 the intramitochondrial pH, pK for the COz-HC03-equilibrium, and h,.

Assumed conditions
Results of the Newton-Raphson iteration" rb k,'  Ca2' was added at the arrows. The addition of 11 mM Ca2+ (about 2 to 4 pmol/mg of protein) to different mitochondrial suspensions caused a decrease by 25-90% in 8, the rate constant for "0 depletion from COZ. This decreased value of B persisted for about 100 s, after which B returned to its original value. At the higher calcium concentration of 33 mM, we observed an increase in the l8O content of CO,. Similar effects, although smaller in magnitude, were obtained with respiratory jumps induced by adding SrC12.
Several control experiments were done to establish that the effect of Ca2+ and Sr2+ on 0 was the result of respirationdependent events characteristic of mitochondria. Ca2+ was shown to have no effect on the uncatalyzed "0 exchange or the exchange catalyzed by soluble bovine carbonic anhydrase, intact rat erythrocytes, or detergent-treated mitochondria. The relatively high concentrations of C1-, about 20 mM, how- Inhibition of the intramitochondrial carbonic anhydrase by high concentrations of methazolamide produced a large, transient rise in T very similar to that observed during the respiratory jumps using 33 mM Ca2+. A smaller increase in T was also produced by adding soluble carbonic anhydrase to the external solution.
By acidifying samples of suspending solution and measuring "0 content of the COz evolved, we found that the "0 content of HC03-was greater than the "0 content of COS in suspensions of mitochondria. This result was expected since COz is known to pass across membranes much more rapidly than HCOJ-; and thus Con has more rapid access to carbonic anhydrase.

DISCUSSION
Localization of Carbonic Anhydrase-Mitochondria from rat liver were subfractionated using digitonin. Fig. 2 shows the release from the sedimented pellet of carbonic anhydrase and several other enzymes which serve as markers. The profde of the outer membrane marker, monoamine oxidase, agreed very closely with that obtained by other researchers (13,18). At concentrations below 0.1 mg/mg of protein, digitonin removes the outer membrane of mitochondria without damaging the inner membrane (13,18). Damage to the inner membrane was detected here by assaying for malate dehydrogenase, a matrix marker, and succinate-cytochrome c reductase, an inner membrane marker. The profiles of these two enzymes taken together indicated that at low digitonin concentrations there was very little or no damage to the inner membrane. The protein profile indicated that there may have been some damage to the inner membrane. At about 0.09 mg of digitonin/ mg of protein, the fraction of protein retained in the pellet was only 0.6. This is considerably lower than 0.85, the fraction of mitochondrial protein supposed to be found in the mitoplasts (matrix and inner membrane) of rat liver (13). However, the fraction obtained here compares favorably with the values reported for mitoplast preparations by other researchers, 0.63 by Schnaitman et al. (18), and 0.54 by Greenawalt (26). It is clear, therefore, that if there were any damage to the inner membrane it was only minimal and it could not account for z Mitochondria from Rat Liver all the release from the pellet of carbonic anhydrase activity seen here at low digitonin concentrations.
At 0.09 mg of digitonin/mg of protein there was a release of carbonic anhydrase activity that was about 50% of the activity released upon lysis of mitochondria. Such release can only be explained by the presence of carbonic anhydrase in a compartment outside the mitoplasts. The mitochondrial outer and inner membranes were found to contain insignificant amounts of carbonic anhydrase activity. Hence, these data strongly suggest that carbonic anhydrase is present not only in the matrix, as it was generally assumed (9, 10) but also in the intermembrane space. That there is enzyme in the matrix is indicated by the residual activity in the pellets after treatment with digitonin concentrations higher than 0.1 mg/mg of protein.
"0 Exchange in Suspensions of Nonrespiring Mitochondria-The "0 exchange method provides information to describe COZ kinetics in cells containing carbonic anhydrase, particularly which of the steps shown in Fig. 1 is rate-determining. Due to the large concentration of carbonic anhydrase in red cells, diffusion into the red cell is rate-limiting for "0 loss from both COZ and HC03- (11,12,27). The data obtained from mitochondria of rat liver indicate that "0 depletion from HC03-in suspensions of mitochondria is in part controlled by diffusion of HC03-into the mitochondria, and in part controlled by the rate of the catalyzed dehydration in the mitochondria. Evidence is the value of y / B = 1.7 (Table I); an "0 depletion which is completely chemically controlled must have y / B = 2.0 (27). Further evidence in Table I is that the rate constant for dehydration of HCOa-inside the mitochondria k', is of magnitude comparable to the rate constant for diffusion of HC03-out of the mitochondria k'L. Although it can be interpreted in several ways, the small increase in "0 exchange (50-100%) observed upon lysis of mitochondria is consistent with a permeability barrier to diffusion of HC03into mitochondria. It is apparent from the results that diffusion of COS into mitochondria is not a rate-limiting event. This can be concluded from Table I1 which shows that Table I1 can be compared with the value of K, found for red cells of about 200 s-' (12). This conclusion is easily reconciled with the low activity of carbonic anhydrase in a mitochondrion from rat liver, which is about 1% of that of rat erythrocytes.
The values of k'i for diffusion of HC03-out of mitochondria given in Table I are roughly equal to values of k'i = 10 s-l found for rat erythrocytes. However, when differences in size between erythrocytes and mitochondria are taken into consideration using Equation 5 it is seen that P for HCOJ-, the apparent permeability constant of mitochondria to HC03-, at 9 X cm s-' is somewhat smaller than the value of about 4 X cm s-' calculated for red cells (12). Dodgson et al. (9) estimated P for HC03-in mitochondria from guinea pig liver to be in the range of to cm s-'. This permeability constant for HC03-which we have estimated is a composite of the permeability barriers between external solution and intramitochondrial carbonic anhydrase. Since we have detected carbonic anhydrase in both the intermembrane space and matrix, the value of P for HC03-is of limited sigmficance since it may reflect the permeability barrier of two membranes. Moreover, we have not included in the calculations the effect of our experimental conditions on the shape of the mitochondria, which frequently swell when isolated (28). As shown in Table 11, our calculations were sensitive to the radius of the mitochondrion which we have Carbonic Anhydrase Activity in Mitochondria from Rat Liver 6855 assumed to be cylindrical in shape. It is also shown in Table   I1 that the value of intramitochondrial pH assumed in the calculations had a large effect on kc; however, changes in the assumed value of pK for the C02-HC03-equilibrium did not affect the calculations sigrdcantly.

Exchange in Suspensions of Respiring Mitochondria-
Elder and Lehninger (10) showed that Ca2+ accumulates in respiring mitochondria in a 1:1 ratio with C02 and suggested that a role of mitochondrial carbonic anhydrase is to hydrate C02 and provide a counter ion for calcium uptake. This was confirmed when they found that calcium uptake in respiring mitochondria is decreased by 6 8 % in the presence of M acetazolamide. Elder and Lehninger suggested the accumulation in the mitochondria of calcium carbonate. Granules of calcium phosphate have been shown to deposit along the inner membrane and matrix of rat liver and heart mitochondria (29). Presumably, a similar precipitation of Ca2+ may occur when phosphate is replaced by carbonate. Precipitation of CaC03 containing "0 would sequester the labeled oxygen in a form which is not susceptible to exchange with water. This explanation is consistent with the observed decrease in the rate constant B for exchange of "0 between C02 and water observed upon stimulating respiration by addition of Ca2+ (Fig. 3). The value of Bobserved in suspensions oinonrespiring mitochondria results from "0 labeled C02 passing into the mitochondria being replaced by C02 of lower "0 content passing out of the mitochondria. A decrease in 0 upon stimulating respiration is consistent with the unidirectional uptake of labeled C02; the mere disappearance of COn from the external fluid without its replacement by CO2 of lower "0 enrichment will not result in "0 depletion of C02 outside the mitochondria. About 100 s after the respiratory jump the C 0 2 kinetics of the mitochondria return to rates of "0 exchange very similar to those measured before the jump. The transient increase in 7 and c that was observed with the higher Ca2+ concentrations (33 m~ in Fig. 3) arises from the difference in the enrichment of the Con and HC03-species in the external solution. Due to its slow diffusion into and out of the mitochondria, bicarbonate has a larger "0 content than COz in suspensions of intact mitochondria. When the system is disturbed, as when the external reaction rate is increased or the intramitochondrial rate decreased, there is a tendency to equalize the enrichment of the two species (24). 18