Mechanism of loss of thermodynamic control in mitochondria due to

Val. 267, No . 22, Issue of August 5, pp. 1534615355,1992 Printed in U.S.A.

THEJOURNAL OF BIOLOGICAL CHEMISTRY 0 1992 by The American Society for Biochemistry and Molecular Biology, Inc.

Mechanism of Loss of Thermodynamic Controlin Mitochondria Due to Hyperthyroidism and Temperature* (Received for publication, January 30, 1992)

Sir0 LuvisettoS, Ibolya Schmehl, Elena Intravaia, Elena Conti, and Giovanni Felice Azzone From the Consiglio Nazionale delle Ricerche Unit for the Studyof the Physiology of Mitochondria and Instituteof General Pathology, University of Padoua, 35121 Padoua, Italy

Incubation of normal mitochondria at 45 O C results 3). A basic argument against an exclusive protonophoretic in increases of respiration andof total apparent proton mechanism is that uncoupling, whether measured by stimuconductance (TAPC, respiration/proton motive force) lation of respiration, depression of proton extrusion, or phosand in an upward shiftof the flow-force relationships. phorylation, is not linked by a unique relationship to the Similar effects are observed during operation of the depression ofA,iiH+,li.e. multiple relationships have been obredox proton pumps at different sitesof the respiratory served between the rates of respiration and phosphorylation chain. These effects are accompaniedby an almost and thelevel of A;,+. Furthermore, in some cases uncoupling equivalent increase of the passive proton conductance effects could be obtained with only a little increase of passive (PPC, proton leakage/proton motive force). proton conductance (3, 4).Additional mechanisms of uncouIn mitochondria from 3,3,5-triiodo-~-thyronine (T3)-treated rats there are also increases of respiration pling have been then proposed. These have been denoted, to and of TAPC and an upward shiftof flow-force rela- indicate their localizations, as decoupling (5)or uncoupling tionships, more pronounced at the level of the cyto- within coupling units (6) for a short circuit between the pumps, chrome oxidase proton pump. However, at variance and asslipping (7) for a short circuit within the pumps. The observation that mitochondria from hyperthyroid and from theincubation at 45 OC, in mitochondria from Tshypothyroid rats show a modified rate of state 4 respiration treated rats there is only a slight increase of PPC. Addition of bovine serum albumin to normal mito- has stimulated much interest on the relationship between the chondria incubated at 45 "C resultsin a marked mechanism of action of thyroid hormone and uncoupling. The depression of TAPCin thenonlinear range of the flow- concept that thyroid hormones may act as uncouplers or as regulators of the physiological uncoupling is not new but has force relationships. An equivalent effect is notobserved in mitochondria from Tdreated rats. gone largely in disuse (8), also in view of the evidence of an The experimental results have been compared with increased content of respiratory enzymes after T3administracomputer simulations obtained on the basis of a chem- tion and of the observation of the lack of effect of thyroid iosmotic modelof energy transduction. The increase of hormones on mitochondria i n uitro. Shears and Bronk (9) TAPC following incubation at high temperature is ap- reported that mitochondria from T3-treated rats had both an parently due to changes of the proton conductance increase of respiration and of A;,+ and postulated that the mainly at the level of PPC, while the increaseof TAPC increase of AjiH+ was due to the inhibition of the linear, and following T3 administration is rather due to changes to the parallel increase of the nonlinear proton conductance presumably at thelevel of the redox or ATPase proton pathway. In contrast with these results, Brand et al. (10-12) pumps. found that the increase of respiration in mitochondria from T3-treated rats was accompanied by a decrease of A;,+ and an opposite effect in mitochondria from hypothyroid rats. They concluded that the action of thyroid hormones is not The concept of uncoupling is fundamental to the mecha- concerned primarily with the amountof respiratory chain per nism of energy conservation. Once a dominantrole to protons mitochondria but, rather,with the leak/slip characteristics of in energy coupling is established the same dominant role must mitochondria. In subsequent studies Hafner and Brand (13) be attributed to protons also in uncoupling. Mitchell and have negated the existence of slips as means of physiological Moyle (1)have reported that classical uncoupling agents act energy regulation and their role in hormonal effects. as protonophores and calculated that the respiratory rate of In a preceding paper (14) we have shown that both T3 uncoupler-treated mitochondria could be accounted for by administration and high temperature induces a stimulation proton cycling. Notwithstanding the wide support tothe of respiration in mitochondria. These increases of respiration concept of the protonophoric action of uncouplers, a multi- are accompanied by changes of the extent of nonlinearity in plicity of uncoupling effects induced by other agents or conditions cannotbe accounted for by simple protonophoresis (2, The abbreviations used are: A j i p , transmembrane proton electro-

* This work was supported by fellowships from the Associazione Sviluppo Scienze Neurologiche (A.S.S.NE.) (to E. C. and E. I.) and FIDIA (Abano Terme, Italy)(to I. S.) 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. $ To whom correspondence shouldbe addressed Consiglio Nazionale delle Ricerche Mitochondrial Physiology, c/o Institute of General Pathology, Padova University, via Trieste 75, 35121 Padova, Italy. Fax: 39-49-8286576.

chemical potential gradient; A", transmembrane electrical potential gradient; J., rate of electron transfer; J A T p , rate of ATP hydrolysis; JHP-P rate , of proton pump extrusion;J H l e a k , rate of proton flux through leaks; JK,rate of K+ efflux; LH,membrane proton-leak conductance; n, H+/e-stoichiometry; %, H+/ATP stoichiometry; TAPC,total apparent proton conductance; PPC, passive membrane proton conductance;APC,activemembraneproton conductance; TB, 3,3,5triiodo-L-thyronine; TMPD, N,N,N',N'-tetramethyl-p-phenylenediamine; BSA, bovine serum albumin; EGTA, [ethylenebis(oxyethylenenitrilo)]tetraacetic acid; MOPS,4-morpholinepropanesulfonic acid.

15348

HyperthyroidismTemperature-induced and Uncoupling the flow-force relationships between apparent proton pump conductance (ratio between respiration and proton motive force) and proton motive force. A distinctive feature, however, isthatthe changes of respiration are accompanied by a marked increase of membrane proton leakage in mitochondria incubated at high temperature but not in mitochondria from T3-treated rats. In the present paper we have extended the analysis of the flow-force relationships, in mitochondria incubated at high temperature as well as inthose preparedfrom T3-treated rats,during operationof the proton pumpsboth at site I1 and site111, separately, and atsites II+III in the absence and presence of bovine serum albumin (BSA). Furthermore, the experimental results have been compared with simulations of the flow-force relationships obtained on the basis of a kinetic model of chemiosmotic free energy coupling. The comparison between experimental and simulated behavior favors the identification of different uncoupling processes, based either on membrane leaks, whether linear (ohmic) or nonlinear (non-ohmic), or on other membrane processes possessing the character of futile cycles at thelevel of the pumps. It appears that the uncoupling effect is often due to the overlapping of various mechanisms. EXPERIMENTALPROCEDURES

Materials-Hyperthyroidism was induced in male Wistar rats (approximately 300 g) byoral administration of 15 pg of TI (solution 5% BSA)/100 gof body mass for 10 days (15).Rats were deprived of food 18 h prior the killing. Liver mitochondria were prepared simultaneously from euthyroid and hyperthyroid rats in a medium containing 0.25 M sucrose, 10 mM Tris, 0.1 mM EGTA, according to standard procedure ( X ) , and all the experiments were performed within 4 hof preparation. The mitochondrial protein was assayed with the biuret method using BSA as a standard. The standard incubation medium contained 0.2 M sucrose, 30 mM MOPS/Tris, 5 mM PJTris, 5 mM succinate/Tris, 0.2 mM EGTA/Tris, 5 PM rotenone, 1 rg/mg oligomycin, pH 7.4, T 25 "C. In the measurements referred to as site 11, the standard medium was supplemented with 2 mM KCN and 2 mM Fe(CN)e(Tris)3,while in the measurements referred to as site 111, succinate was omitted and themedium was supplemented with 2mM ascorbate, 50 ng/mg antimycin, and variable amounts of TMPD. All reagents were of maximal purity commercial grade. TB, enzymes, inhibitors, and valinomycin were purchased from Sigma. Determinations of the Rates of Respiration-The respiratory rates of the redox chain at sites II+III and a t site 111 were estimated from the rate of oxygen consumption, whose concentration in themedium was measured polarographically withaClark electrode (Yellow Spring) equipped with a Teflon membrane in a closed thermostated and stirred vessel. The calibration of the oxygen electrode as well as the determination of the oxygen concentration in thereaction media for each different temperature was carried out by oxidizing a known amount of spectrophotometrically standardized NADH in the presence of beef heart submitochondrial particles (17). The zero oxygen point was determined with an excess of dithionite. The respiratory rate at siteI1 was estimated from the rate of reduction of Fe(CN)imeasured spectrophotometrically on an Aminco DW2000 dual-wavelength spectrophotometer equipped with magnetic stirring and thermostatic control, following continuously the decrease of absorbance a t 420 minus 480 nm. Determination of the Rate of ATP Hydrolysis-The rate of ATP hydrolysis, J A p , was measured spectrophotometrically by following continuously the decrease in absorbance at 340 minus 374 nm due to NADHoxidation in the presence of excess phosphoenolpyruvate, pyruvate kinase, and lactate dehydrogenase. The rates of ATP hydrolysis were corrected for the contribution of extramitochondrial ATPases, as determined in control samples in the presence of oligomycin (1 pg/mg) and atractyloside (200 p ~ )The . operative conditions and time of incubation with reagents are described in the legend of Fig. 6. Determinution of Membrane Potentials and Passive Proton Fluxes-The transmembrane electrical potential, A*, was evaluated from the distribution of the lipophilic ion triphenylmethylphosphonium and the concentration of triphenylmethylphosphonium in the incubation medium was followed continuously by using a triphenyl-

15349 methylphosphonium-sensitive electrode (response time about 10 s) as described in Ref. 3. In our conditions, namely in the presence of 5 mM Pi in the incubation medium, the variation can be considered to reflect a good approximation of A;* (3). The operative conditions and times of incubations with reagents are described in the figure legends. The rates of passive proton flux, J H ~through , the mitochondrial inner membrane at variable membrane potential levels were estimated by measuring the initial rate of K' efflux, J K ,upon addition of valinomycin (0.15 pg/mg) to antimycin-inhibited (0.05 pg/mg) mitochondria, either from euthyroid or hyperthyroid rats, a t variable K' diffusion potentials. The measurement relies on the principle that the rate of K+ efflux down the K+ electrochemical gradient in respiratoryinhibitedmitochondriamust, for electroneutrality reasons, equal the sum of the rates of anion efflux and cation influx other than K'. If, as under the experimental conditions used, essentially only the permeable ion besides K' is H+ (or OH-), J K is equal to the rate of passive influx of protons. If other ions move, J K represents only an overestimation of the rate of passive influx of protons. The rates of K+ efflux were experimentally determined as described in Ref. 18. The size of valinomycin-induced K+ diffusion potentials were calculated from theNernst equation: A I = (Z.BRT/nF)log[K;]/ [ K 3 ,where [KLJwas determined by using the null point technique as described in Zoratti et al. (18), while [ K f , ]was taken as equal to the variable amount of K' added (0-500 p M ) plus the contribution due to the medium contamination (20-30 pM), The operative conditions and times of incubation with the reagents are described in the legend to Fig. 7. RESULTS

In static head, the efflux of protons, extruded by the redox proton pumps, J H p u w , is equal to thepassive influx of protons through the proton leaks, J p a k . Assuming a tight coupling in the redox pumps the two processes behave as a stoichiometric process. Hence, in static head, JHpvmp

=

JHIeat

=

nJ,

(1)

where n represents the H'/e- stoichiometry of the proton pumps. Equation 1 indicates that the relationship between deand the membrane potential should in principle reflect the dependence of the rate of the passive proton diffusion through the membrane on the membrane potential. Hence the ratio between deand membrane potential at each level of membrane potentials should reflect the proton conductance of the inner mitochondrial membrane. The same should also be true for the ATPase proton pumps, i.e. the ratio between the rate of ATP hydrolysis multiplied by the H'/ATP stoichiometry, and membrane potentials should also reflect the proton conductance through the inner membrane. However, it has been argued (2) that thisparameter, measured through respiration or ATPase activity, is not equivalent to the real passive proton conductance of the inner mitochondrial membrane, since the respiration or the ATPase activity in static head reflects the proton conductance througha structure comprising the lipids and a multiplicity of inner membrane integral proteins,including the proton pumps and theion and metabolite carriers. In an attempt todistinguish and characterize these processes, in the following we operatively define: ( a ) an overall conductance, measured from the rates of respiration or of ATP hydrolysis in static head (both multiplied by the pump stoichiometries) divided by the proton motive force, as total apparent proton conductance, TAPC. ( b ) A real passive proton conductance (presumably largely through lipids), measured from the efflux of K" in respiratory inhibited mitochondria divided by the K' diffusion potential, as passive proton conductance, PPC. (c) A conductance due to slip in the proton pumps calculated from the difference between the values of the total apparent proton conductance, TAPC, and of passive proton conductance, PPC, as active proton conductance, APC.

15350

Hyperthyroidism-and Temperature-induced Uncoupling

From the above definitions the total apparent proton conductance is taken to reflect the sum of both the passive and the active conductances.

was similar to that observed in normal mitochondria incubated at 45 "C. However, in thecase of T3administration the flow-force relationship was shifted less to theleft with respect to what was observed at 45 "C, i.e. the increase of TAPC TAPC = PPC + APC (2) occurred mostly at higher potentials and only slightly at lower Strictly speaking the termflow-force relationship should be potentials. Furthermore, the increase of TAPC at high memused only to indicate the relationship between a flow, the brane potentials closely resembled that caused by nonprotonrespiratory rate, and a force, the proton motive force. Follow- ophoric uncouplers, such as chloroform (results not shown; ing a common use, however, the relationship between a con- however, see Ref. 20 for comparison). Finally, T, administraductance (the respiratory rate divided by the potential) and tion caused a change in the flow-force relationship more the membrane potential have been denoted below as flow- marked at site I11 (panel C) than at siteI1 (panel B ) . The interpretation of the nonlinearity in the relationship force relationship. Fig. 1,A-C, shows the effect of the temperature increase on between total apparent proton pump conductance and A;,+ the relationship between TAPC ( d e / A \ k ) and membrane (or between APC and Ai&) has been attributed toa nonlinear potentials, for the redox proton pumpsat sites I1 + I11 together conductance for protons of the inner membrane (21), to an ( A ) , the pump at site I1 ( B ) , and the pump at site 111 (C). intrinsic uncoupling in the proton pumps (slips) (22), or to The different respiratory rates andmembrane potentials were inhomogeneity of the mitochondrial preparation (23, 24), i.e. obtained by titrations with respiratoryinhibitors (at sites presence of a fraction of leaky vesicles. Recently, Garlid et al. II+III and11) or by variations of the substrateconcentrations (25) have suggested thesenonlinearrelationships to be a (at site 111). Values of n at different proton pumping sites natural property of all biological membranes. According to were taken according to Beavis (19). Different values of the Nicholls (21) and Garlid et al. (25), an exponential relationstoichiometries cause only an upward or a downward shift of ship between the rate of passive proton influx and the memthe curves without affecting the interpretationof the results. branepotential should lead to alinear plot between the Fig. lA shows that with the proton pumps at sites II+III, the logarithm of de and membrane potential. relationship between TAPCand membrane potentials, at during titrations with redox In Fig. 3 the values of de, 25 "C, had the usual shape (cf. Luvisetto et al. (20)); there was inhibitors in euthyroid mitochondria at low and high tempera linear behavior (characterized by a constant value of TAPC) ature (25 and 45 "C) and in mitochondria from T3-treated at low and a nonlinear behavior at high membrane potentials. rats, have been plotted as l n ( d , ) as a function of membrane At 45 "C two phenomena became apparent. First, in the potential. The linearity so obtained suggests that theincrease absence of respiratory inhibitor (condition corresponding to of TAPC observed at 45 "C, at all values of membrane potenthe first points of the curves obtained at high membrane tial, can be completely accounted for by assuming an expopotential), the increase of temperature resulted in a marked nential relationship between the rateof passive proton influx increase of TAPC which was accompanied only by a slight through the lipid bilayer and the membrane potential, as decrease of the membrane potential. Second, the relationship predicted by Garlid et al. (25). On the otherhand, the lack of was so modified that it was impossible to see the two ranges linearity observed at 25 "C in the semilog plot suggests that of linear and nonlinear behaviors observed at 25 "C. Similar the increase of TAPC in the high potential range cannot be effects were measured with the proton pumps at site I1 (Fig. accounted for by such a simple relationship. An intermediate 1B) and site I11 (Fig. 1C). The distinction between the two pattern was obtained with mitochondria from T3-treated rats. ranges of TAPC, linearand nonlinear, vanished progressively This will be discussed later. with the increase of temperature (results not shown). In a previous paper (ZO), we have reached the conclusions Fig. 2 shows the relationship between TAPC andmembrane that BSA causes a decrease of TAPC in the linear but not in potential in mitochondria from T3-treated rats. The three the nonlinear range. Furthermore, the effect ofBSAwas panels represent the effects of T3 treatment on the redox accompanied by a shift of the flow-force relationship to the proton pumps at sites II+III ( A ) together or on the isolated right, i.e. toward higher values of membrane potential. The proton pumps at sites I1 ( B ) or I11 (C). At sites II+III (panel effect of BSA was apparently that of a proton leak inhibitor A ) the relationship in mitochondria from hyperthyroid rats presumably due to itsknown properties of binding fatty acids. FIG. 1. Effect of temperature on the relationship between respiration-linked TAPC (nJ./A*) and membrane potential with redox pumps at sites II+III, 11, and 111. The temperature was 25 (0) and 45 "C (O), respectively. Panel A, sites II+III, n = 3.5; panel B, site 11, n = 1; panel C, site 111, n = 2.5. Rat liver mitochondria (1 mg/ml) were incubated in the various media (see text) for 2 min. In panels A and B, succinate (5 mM) was then added, followed after 2 min by increasing concentrations of malonate (0-5 mM), and the rates of oxygen consumption or ferrocyanide reduction and A* were measured. In panel C, instead of succinate, increasing concentrations of TMPD (0300 sM) were added, and after 2 min, the rate of oxygen consumption were measured. The dimensions of the ordinate axis are nanomole of mg" min" mV".

-A

B

3-

i

2-

5. 3

7

C

1-

-120

2

a-

-,"A-

160

200 120

160

Ay,rnV

2C


15351

2.01 A

FIG. 2. Effect of Ts-induced hyperthyroidism (0)on the relationship between respiration-linked TAPC and membrane potential with redox pumps at sites II+IIL, 11, and 111. Panel A , sites II+III, n = 3.5; panel B, site 11, n = 1;panel C, site 111, n = 2.5. Open symbols represent the relationship a s obtained from normal mitochondria incubated at 25 “C. Experimentalprocedures and incubation media were the same as reported in the legend of Fig. 1.

120

I

6-

, i

5-

8

PP

L-

-

110

150

190 Ay,mV

FIG. 3. Log-linear plots of the respiratory rates and membrane potentials at sites II+III. The logarithms of the respiratory rates multiplied by the H’/e- stoichiometry (sites II+III, n = 3.5) are plotted as a function of the membrane potentials. Different curves were obtained from mitochondria of euthyroid rats incubated at 25 (0) or 45 “C (0)and T3-treated rats(B).

Fig. 4 shows the effect of increased concentrations of BSA on the relationship between TAPC and membrane potential as obtained in euthyroid mitochondria incubated at 45 “C. The inset shows the same data plotted as ln(nJ,) as a function of the membrane potential. Alow concentration of BSA, namely 0.05%, caused a depression of TAPC at each value of membrane potential without a marked change in the exponential shape of the flow-force relationship. In accordance with this, and membrane the relationship between the logarithm of de potential was downward shifted. At higher BSA concentrations, namely 0.5%, the exponential shape of the curve was modified with a more distinct change in slope moving from the range of lower to that of higher potentials. This was also evidenced in the log-linear plot, i.e. at higher BSA concentrations the log-linear plot was modified becoming almost nonlinear as it was the case for the relationships at lower temperatures. Fig. 5 shows the effect of increasing BSA concentrations on therelationship between either TAPCor logarithm of de

A y.mV FIG. 4. Effect of BSA on the relationship between respiration-linked TAPC and membrane potential at high temperature. Rat liver mitochondria ( 1 mg/ml) of euthyroid rats were incubated at 45 “C for 2 min. Succinate (5 mM) was then added followed after 2 min by increasing concentrations of malonate (0-5 mM). At each malonate, concentration respiratoryrate andA* were measured. Inset, log-linear plots of the ln(nJ.) uersus membranepotentials. Different curves were obtained in the absence (0, 0) or in the presence (0.05%,B, 0,0.5%, A,A) of BSA in the incubation medium.

(inset of Fig. 5) and membrane potential as obtainedin mitochondria isolated from hyperthyroid rats. The comparison with the curve obtained in the absence of BSA indicates that the shape of the flow-force relationship was markedly modified already at low BSA concentrations (0.05%),becoming steeper and shifting toward higher membrane potentials. At higher BSA concentrations (0.5%) the flow-force relationship became more similar to that observed with chloroform (see Ref. 20 for comparison). The log-linear plot show that and membrane potential was the relationship between In(de) nonlinear already in the absence of BSA, and thenonlinearity was more pronounced at increased concentrations of BSA. The increased nonlinearity was evident also at the lowest BSA concentrations. Although the extentof nonlinearity was not constant, identical patterns were obtained in all different mitochondrial preparations. Furthermore, the extent of non-

15352

Hyperthyroidism- and Temperature-induced Uncoupling 4-

33

% $2"

7

/. i.! t T$

n

C

1-

- -b

" "

120

l>O

200 180 160

A y ,mv FIG. 5. Effect of BSA on the relationship between respiration-linked TAPC and membrane potential inTs-induced hyperthyroidism. Rat liver mitochondria (1 mg/ml) isolated from T S treated rats. Experimental procedures were the same as described in the legend to Fig. 4. Inset,log-linear plots of ln(nJ.) uersus membrane potentials. Different curves were obtained in the absence ( 0 , O )or in the presence (0.05%,W, 0; 0.5%, A,A) of BSA.

linearity of the different preparations was enhanced by increasing the BSA concentrations. The results in Figs. 4 and 5 suggest that the uncoupling induced by high temperature and hyperthyroidism are due to different modifications occurring at the level either of the membrane bilayer or proton pumps. At 45 "C, the linearity of the log-linear plots inFigs. 3 and 4 and theeffect of increasing BSA concentrations is in accord with the view that most of the uncoupling effect is localized at the level of the bilayer. At high BSA concentrations a minor residual effect remains at the level of the proton pumps. On the other hand, the different pattern observed in mitochondria from hyperthyroid rats is in accord with a localization of the uncoupling effect, only to a minor extent at thelevel of the lipid bilayer and to a major extent at the level of other membrane protein components. The effect at thelevel of the lipid bilayer is removed even by low BSA concentrations. Fig. 6 shows the effect of T3administration and incubation of mitochondria at 45 "C on the relationship between TAPC, measured on the rate of ATP hydrolysis (npJATp/A!P), and AQ. The incubation at high temperature resulted in a very marked increase of nonlinearity in a similar manner as obtained with the redox proton pumps. In contrast, T3 administration resulted in a modest, although constant, increase of nonlinearity. For comparison, the effect of chloroform is also reported. The K+ diffusion potential as a tool to analyze various mitochondrial processes has been used since the early sixties (26, 27). As discussed by Zoratti et al. (2) and by Krishnamoorthy and Hinkle (28), the parameter J K / Aconstitutes ~ a direct expression of the dependence of the passive proton conductance of the membrane on the membrane potential. This conductance has been designated as passive proton conductance of the membrane, PPC, todistinguish it from TAPC. Fig. 7 reports the flow-force relationship between PPC, and as measured in normal mitochondria incubated at 25 or 45 "C, and in mitochondria from hyperthyroid rats. In euthyroid mitochondria at 25 "C, the relationship was approximately

I

80 160

AJ

I

120

I

I

200 Ay,mV

FIG. 6. Effect of temperature (0,45 "C), Ts-induced hyperthyroidism ,).( and CHCls (A, 15 mM) on the relationship between ATPase-linked TAPC and membrane potential. Rat liver mitochondria (1 mg/ml) were incubated in the standard incubation medium supplemented with 2 mM MgC12, 1 mM phosphoenolpyruvate, 0.1 mM NADH, and excess pyruvate kinase and lactate dehydrogenase. After 3 min of incubation with increasing concentrations of oligomycin (0-1 pg/mg), ATP (3 mM) was added, and after 2 min the rate of ATP hydrolysis was measured. Chloroform was added 1 min before ATP. The membrane potentials were measured in parallel samples, under the same conditions. Dashed line: relationship obtained in normal mitochondria at 25 "C. n, was taken as equal to 4.

linear with a smallincrease of PPC athigh membrane potentials. Only in some preparations the increase of PPC at high potentials was more pronounced (results not shown). This is compatible with a variable extent of nonlinear proton leak in different mitochondrial preparations. At 45 "C, there was a marked increase of PPC not only at the lower but also, and particularly, at the higher membrane potentials. In the case of the incubation at high temperature the comparison between the effects on the TAPC, whether redox-linked or ATPaselinked proton pumps, and PPC indicates a parallelism between both increases, i.e. between the increases of basal respiratory rate, or ATP hydrolysis, and of passive proton fluxes through the membrane. On the other hand, it appears that T3caused an increase of the respiratory rate with only a slight increase in the passive proton flux through the membrane. DISCUSSION

The protonophoric effect of classical uncouplers has been extensively studied and attributed tocycling across the membrane of the undissociated acid in one direction and the dissociated anion in the other. Whether the endogenous uncoupling of mitochondria is closely related, or identical to, the mechanism of action of exogenous protonophoric agents constitutes an unsolved bioenergetic problem. In the early seventies Nicholls (21) reported that thisprocess is composed of two types of membrane conductances for protons denoted as ohmic and non-ohmic on the base of the dependence on the membrane potential. In the early eighties the nonlinear behavior has been attributed, at least partly, to a slip in the pump (2-4, 7). The slip concept indicates a cycle of electron transfer not

15353


low dependence may be due to eithercharge neutralization or the presence of special carrier systems lowering the energy barrier (25). In the present study we have compared the effect of temperature and T3 administration on two types of flow-force relationships, namely the TAPC- and PPC-potential relationships. While the differences between the various agents or conditions are marked, the molecular interpretation of the effects is difficult. To provide a more sound kinetic basis, the experimental results reported in Figs. 1-7 have been studied 1.5 by computing the flow-forcerelationships for different degrees 3 and types of uncoupling. The computing simulation was based on a chemiosmotic proton model constituted by two elements 7 in parallel: ( a ) a A j i p generating a proton pump model, as 1.0 described by Pietrobon and Caplan (31), with kinetic parameters able tosimulate the protontranslocationthrougha redox proton pump,' and ( b ) a pathway accounting for the passive diffusion of protons across the membrane. The chemiosmotic proton circuit has been translated into an electric 0.5 network whose properties have been analyzed by the use of SPICE (VMS version implemented on a Digital VAX8650), a sophisticated simulationprogram originally developed to analyze electrical integrated circuits. The same model has already 1 1 I I I been used to simulate the effect of classical protonophoric 110 100 180 220 agents and slip inducers on the relationship between TAPC AymV (nJ,/Aji,+) and A i p (20, 33). FIG. 7. Effect of temperature and hyperthyroidism on the Following Garlid et al. (25) we have assumed the presence relationship between K+ release and K+ diffusion potential. in themembrane of a nonlinear protonleak with an exponenSymbols: normalmitochondriaincubated at 25 (0)or 45 "C (0); mitochondria from hyperthyroid rats (m). Rat liver mitochondria (1 tial dependence on the membrane potential. The following mg/ml) were incubated for 2min inthe presence of increasing equation was taken to represent the passive proton diffusion,

1

5

'

concentrations of K' (40-800 PM). Then, antimycin (50 ng/mg) and valinomycin (200 ng/mg) were added and the ratesof K' release were measured. The concentration of internal K' was estimated by null point titrations and was comprised in the range 120-150 mM, under the various conditions considered. The K' diffusion potentials was calculated using the Nernst equation.

coupled to vectorial proton translocation. The operation of the slip has subsequently been denoted as intrinsic uncoupling to differentiate it from classical protonophoric uncoupling. Walz (29) has recently presented a general kinetic scheme accounting for coupled flows both in enzyme and carrier reactions (cf. see also Klingenberg (30)). Inthis scheme, similarly to the six-state proton pump model of Pietrobon and Caplan (31), there is a transition thatcauses the flows to be not tightly coupled. Operation of the transition, whether in an enzyme or in a carrier, leads to an intrinsic uncoupling and can therefore be denoted as a slip. This suggestion has been strengthened by the discovery that some agents cause uncoupling in mitochondria by acting as slip inducers rather than as classical protonophores (3,4, 20). Garlid et al. (25) have recently criticized both the concept of slip and of an innermitochondrial membrane possessing a regulated and nonconstant permeability, ohmic at low and non-ohmic at high potentials, and introduced the concept of a constant nonlinear behavior withexponential ion fluxvoltage relationships. The linear behavior would thus become a limitingcase, partly due to thephysiological need of limiting energy losses at high Gibbs free energy values of the driving forces. Using a membrane barrier model based on the Eyring rate theory (32), a relatively simple relationship between ion fluxes and membrane electrical potential has been developed (25) with the different extents of nonlinearity reflecting a different shape of the energy barrier to ion leaks across the membranes. In thisview, the linear behavior is a special case occurring under conditions either of low potentials or of low dependence of ion fluxes on the membrane potential. This

JHleak

=

LH x eXp(N x &Hf)

(3)

where LHis the proton conductance at A F p = 0, N represents aparameter reflecting the exponential dependence of the passive proton flux across the lipid bilayer on the membrane potential. A similar equationwas used by Garlid et al. (25) to analyze the nonlinear flux of ion across biological membranes, as for tetraethylammonium ions in mitochondria (34). In the analysis of Garlid et al. (25) the value of N is a parameter accounting for different shapes of the energy barriers to ion fluxes within the membranes. A change in N , corresponding to a change in PIRT (25), would correspond to a change in the mechanism of transport through the membrane. P assumes different values (25) depending on whether the leak pathways are localized in the bulk lipid, or in the lipid bilayer, or at theprotein-lipid interfaces. On the otherhand, changes in LH should reflect alterations in the height of the energetic barrier to proton fluxes through the membrane. As we have already discussed, a marked difference becomes apparent (cf. Fig. 3) between the log-linear plots observed at 25 and 45 "C, in that the latterare linear while the former are not. According to the analysis of Garlid et al. (25), a value of /3 in therange between 0.2 and 0.5 is compatible with a proton The six-state proton pump model of Pietrobon and Caplan (31) is a kinetic scheme, adapted from one of the possible reaction sequences of the general cube model proposed by Wikstrom et al. (37). The following transition steps have been considered: 1) binding of protons to an empty site on the inside; 2) binding of substrates; 3) conversion of bound substrates to products and unbinding of products; 4) conformational change-linked translocationof bound protons from the inside to the outside; 5) unbinding of protons; 6) conformational change-linked translocation of the empty proton-binding sites back to the inside. Each of these transitions are characterized by first- or pseudo-first order rate constants in forward as well as in the backward direction (31). The slip transition concerns the shiftof the protonated binding sites from outside to inside. In the present paper, the rate constants of the kinetic transitions have been partially modified, respect on the original value proposed in Ref. 31.

15354


leak localized at thelevel of the lipid membrane bilayer, while a value of /3 equal or higher than 0.5 is indicative of integral protein-induced perturbations in the bilayer. We have calculated the following slopes by linear regression of the log-linear mV" (0.326 mol kcal") plots ofFig. 3: (i) at25 "C, 7.52 x at low membrane potentials, and 2.87 X 10" mV" (12.46 mol kcal") at high membrane potentials; (ii) at 45 "C, 2.53 X lo-* mV" (1.1 mol kcal"). These slope values correspond to values of p equal to 0.193, 7.38, and 0.651 (mol kcal"), at 25 "C, low and high potential, and 45 "C, respectively. These calculations indicate that the very steep slope observed in the high potential range at 25 "C require a value of p too high and apparently incompatible with the p values predicted by the analysis Garlid et al. (25). The alternative is that theincreased respiration and TAPCin the high potential range are accounted for by a mechanism different from that predicted by Equation 3. A similar argument holds for the behavior of mitochondria from T3-treated rats,where in the log-linear plots there is also nonlinearity and, specially in the presence ofBSA, an increased slope which cannot be accounted for only by an increase of p values. Fig. 8 shows the simulations of the relationship between n J e / A q , In(de), or J H I - k / A q , as a functionof A*, as obtained by reducing the fraction of active redox proton pumps in the presence of a different extent of nonlinear proton leak. The variable extent of nonlinear leak was introduced using Equation 3, with a constant value of LH equal to 0.2 nmol mg" min" mV", as taken from the intercept on the ordinate of Figs. 1-2 in normal mitochondria at 25 "C, and values of the parameter N as calculated on the basis of the log-linear plot ofFig. 3. Different extents of slip transition in the redox proton pumps were also introduced. In theleftpanel of Fig. 8, the relationships between TAPC and membrane potential were obtained onthe basis of the following assumptions: curue a,low extents of slip and of nonlinear proton leak; curue b, moderate increase of the extent of slip and only aslight increase of nonlinear proton leak; curue c, a more marked increase of the extent of nonlinear proton leak. In the inset are the simulations of the log-linear plots obtained with the same values of respiration and membrane potential. The right panel of Fig. 8 shows the parallel simulations of the relationship between PPC and membrane potential as obtained on the basis of the same parameters used in the simulations of the left panel. It is seen that the increases of nonlinear leak and slip have different effects on the PPC in that the former

FIG. 8. Simulations of rkJJA* (Zen panel), ln(rkJe) (inset), and of JHl4 A* (rightpanel)as a function of A*. The various curves were obtained in the presence of increasing values of nonlinear proton leak: N = 0.326 mol kcal", N = 0.84 molkcal", and N = 1.10 mol kcal", for curues a, b, and c, respectively. As to the slip transition, the values used in curue a were recalculated on the basis of those originally used in the model of Pietrobon and Caplan (31) to yield the presumably physiological degree of slip in the redox pumps. At all events these values for the slip transition account for a portion of the state 4 respiratory rate in normal mitochondria. This value was left unaltered in curue c and increased by a factor of 2.5 in curue b.

cause a marked increase of PPC while the latter do not. The consequence is that while in curve c the increase of TAPC is completely accounted for by the increase of PPC, in curue a there is no parallelism between the increases of TAPC and PPC. Curve b represents an intermediate case, where the increase of TAPC is only to a minor extent accounted for by the increase of PPC. It is of significance that the values of respiration, membrane potentials, conductances, leaks, and slips used in the simulations are almost identical to those measured in the experiments (compare with the values reported in theexperiments of Figs. 1-3 and 7). In conclusion, the comparison between simulations and experiments indicates that curues a recall rather closely the behavior at 25 "C, while curues c recall the behavior at 45 "C (see Figs. l A , 3, and 7). Furthermore, curues b correspond to the behavior of mitochondria from hyperthyroid rats (see Figs. 2 A , 3, and 7). That the experimental behavior observed at high temperature can be simulated by increasing the value of N (equivalent the analysis of Garlid to /3/RT)in Equation 3 has a bearing on et al. (25). This analysis in fact predicts a decrease of the exponential dependence of the proton flux through the membrane parallel to the increase of temperature. However, this is not observed. This result tends tosupport the view that the temperature increase is accompanied by a change in the value of /3 and thusalso in the mechanism of proton transport. Such change could take place more likely at thelevel of the number or nature of steps in the proton transport process rather than in theshape of the energy barrier. The comparison between simulations and experiments suggest that: 1) temperature increase and hyperthyroidism induce uncoupling by a mechanism different from that of classical protonophores; 2) the effect of temperature may be explained with an increase of nonlinear proton leak, an effect presumably occurring at thelevel of the lipid bilayer; 3) the effect of hyperthyroidism may be explained with a mixed behavior, i.e. at thelevel, to a major extent, of the proton pumps and, only to a minor extent, of the lipid bilayer or the ion and metabolite carriers. From measurements of the rate of swelling of respiring mitochondria in tetramethylammonium acetate and inpotassium acetate during titrations with malonate, Hafner et al. (35) have suggested a decrease of the cation permeability in the mitochondria from hypothyroid compared to mitochondria from euthyroid rats. Assuming then a parallelism between

C

1.5

"

2

2

ZFZ

2

$-1.E

.5

2

-E

I

1

1.5

I

120

I

I

1LO 180160

I

I

200 Ay,rnV

I

120

I

1LO

I

160

I

I

180

200

Hyperthyroidism-and Temperature-induced Uncoupling cation leak and theproton leak, they have also suggested that the differences in respiratory rate linked to thyroid status be uniquely due to differences in the extent of passive proton leak. Unfortunately the lack of quantitative assays in the experiments of Hafner et al. (35) renders difficult the comparison with the present data. In conclusion, experimental resultsand simulations tend to support theconcepts oE ( a ) a multiplicity of proton conductance pathways on one side and of a difference in nature between slips and passive proton leaks on the other and( b ) a multiplicity of mechanisms by which thermodynamic control is released from the bioenergetic organelles. The different uncoupling mechanisms may be distinguished on the basis of their dependence on changes of either the passiue proton conductance at thelevel of the membrane lipid bilayer or the actiue proton conductance at the level of the proton pumps. Moreover, the increase of the passiue conductance of the inner membrane for protons may occur with either a linear or a nonlinear behavior, thus defining at least two different mechanisms of proton transport across the membrane. Classical protonophores, such as 2,4-dinitrophenol or gramicidin in the presence of univalent cations (4),act essentially as extrinsic uncouplers causing an increase of ohmic passiue proton leak conductance. The temperature increase behaves also as a protonophore but at variance from classical protonophores, it causes an increase of the non-ohmicpassiue proton leak conductance. Chloroform (20), or gramicidin in the absence of univalent cations (4), cause mainly an increase of active proton conductance. The uncoupling effects in hyperthyroidism are intermediates between those observed in the presence of chloroform (see, for comparison, Ref.20) and those observed at high temperature. A mixed increase of passive non-ohmic proton leak and active proton conductance has been observed also by Ouhabi et al. (36) in nonphosphorylating yeast mitochondria. Acknowledgments-We t h a n k Prof. K. D. Garlid for helpful discussions on the experimental results and L. Pregnolato for expert technical assistance.

15355

REFERENCES 1. Mitchell, P., and Moyle, J. (1967) Biochem. J. 1 0 4 , 58&600 2. Zoratti, M., Favaron, M., Pietrobon, D., and Azzone, G. F. (1986) Biochemistry 2 5 , 760-767 3. Luvisetto, S., Pietrobon, D., and Azzone, G. F. (1987) Biochemistry 2 6 , 7332-7338 4. Luvisetto, S., and Azzone, G. F. (1989) Biochemistry 28,1100-1108 5. Rottenberg, H., and Hashimoto, K. (1986) Biochemistry 25,1747-1755 6. Westerhoff, H. V., Melandri, B. A., Venturoli, G., Azzone, G. F., and Kell, D. B. (1984) Biochim. Biophys. Acta 768,257-292 7. Pietrobon, D.,Zoratti, M., Azzone,G. F., andCaplan, S. R. (1986) Biochemistry 2 5 , 767-775 8. Guernesey, D.L., and Edelman, I. S. (1983) in The Molecular Basis of Thyroid Hormone Action (Oppenheimer, J. H., and Samuels, H. H., eds) 293-324, Academic Press, New York 9. Shears, S. B., and Bronk, J. R. (1979) Biochem. J. 1 7 8 , 505-507 10. Hafner, R., Nobes, C. D., McGown, A. D., and Brand, M. D. (1988) Eur. J. Biochem. 178,511-518 11. Brand, M. D. (1990) J. Theor. Biol. 145,267-286 12. Brand, M. D., Couture, P., Else, P. L., Whithers, K. W., and Hulbert,A. J. (1991) Biochem. J. 276.81-86 13. Hafner,’R. P., and Brand,”. D. (1991) Biochem. J. 2 7 5 , 75-80 14. Luvisetto, S., Schmehl, I., Conti, E., Intravaia, E., and Azzone, G. F. (1991) FEBS Lett. 291.17-20 15. Carr, F. E., Bingham, C., Oppenheimer, J. H., Kistner, C., and Marriash, C. N. (1984) Proc. Natl. Acad. Sci. U. S. A. 81,974-978 16. Massari, S., Frigeri, L., and Azzone, G. F. (1972) J. Membr. Biol. 9 , 57-60 17. Chappell, J. B. (1964) Biochem. J. 90,225-237 18. Zoratti, M., Petronilli, V., and Azzone, G. F. (1986) Biochim. Biophys. Acta 851.123-135 , ~~. .~~ 19. BeavG, A. D. (1987) J. Biol. Chem. 262,6165-6173,6174-6181 20. Luvisetto, S., Conti, E., Buso, M., and Azzone, G. F. (1990) J . Biol. Chem. 2 6 6 , 1034-1042 21. Nicholls, D. G. (1977) Eur. J. Biochem. 77,349-356 22. Pietrobon, D., Azzone, G. F., and Walz, D. (1981) Eur. J. Biochem. 1 1 7 , 389-394 23. Duszynski, J., and Wo’tczak, L (1985) FEBS Lett. 182,243-248 24. Zolkiewska, A,, Zablocia, B., Duszynski, J., and Wojtczak, L. (1989) Arch. Biochem. Bio hys 275,580-590 25. Garlid, K. D., geavis, A. D., and Ratkje, S. K. (1989) Biochim. Biophys. Acta 9 7 6 , 109-121 26. Azzone G. F and AzziA. (1966) in Re ulation o Metabolic Processes in Mitohmd& (Tager b. M., Pa a, S., duagliarieho, E., and Slater, E.C., eds) Vol. 7, pp. 332-946, BBA library, Elsevier, Amsterdam 27. Rossi, E., and Azzone, G. F. (1970) Eur. J. Bcochem. 12,319-327 28. Krishnamoorthy, G., and Hinkle, P.C. (1984) Baochemzstry 23,1640-1645 29. Walz, D. (1990) Bcochim. Bzophys. Acta 1019,171-224 30. Klingenber M (1990) Trends Biol. Sci. 1 5 , 108-112 and Caplan, S. R. (1985) Biochemistry 24,5764-5776 31. Pietrobon 32. Eyring, H:, and Eyring, E. (1963) Modern Chemical Kinetics, Rheinhold, New York 33. Pietrobon, D., Luvisetto, S., and Azzone, G. F. (1987) Biochemistry 2 6 , 7339-7347 34. Toninello, A., Miotto, G., Siliprandi, D., Siliprandi, N., and Garlid, K. D. (1988) J. Biol. Chem. 263,19407-19411 35. Hafner, R., Leake, M. J., and Brand, M. D. (1989) FEBS Lett. 2 4 8 , 175178 36. Ouhabi, R., Rigoulet, M., Lavie, J.-L., and Guerin, B. (1991) Biochim. Biophys. Acta 1060,293-298 37. Wikstrom, M., Krab, K., and Saraste, M. (1981) Annu. Reu. Biochem. 5 0 , 623-655

B.,