On the Relationship between the Mitochondrial Inner Membrane ...

THEJOURNAL OF BIOUXICAL CHEMISTRY 0 1994 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 269, No. 14, Issue of April 8 , pp. 10614-10620, 1994 Printed in U S A .

On the Relationship betweenthe Mitochondrial Inner Membrane Anion Channel and theAdenine Nucleotide Translocase” (Received for publication, November 12,

1993, and in revised form, January 31,

1994)

Mary F. Powers, Lisa L. Smith, and Andrew D. BeavisS From the Department of Pharmacology, Medical College Ohio, of Toledo, Ohio 43699-0008

The mitochondrial inner membrane anion channel brane anion channel) (7, 8). IMAC can be inhibited by a wide (IMAC) is a transport pathway which is believed to be variety of agents (seeRef. 1for a review) including mercurials involved in mitochondrial volume homeostasis. The prowhich block many of the anion carriers in the inner membrane. tein, however, has not been identified.In this paper, we One of the most potent and selective inhibitors of IMAC idenexamine the relationship between IMAC and the ad- tified to date is tributyltin which completely blocks IMAC in rat enine nucleotide translocator.Many inhibitors of the ad- liver mitochondria at doses of 1 nmoVmg (9).The only other enine nucleotide translocase are shown to block IMAC, effect on mitochondria reported at thisdose is inhibitionof the including Cibacron blue 3GA, bromcresol green,alizarin F,F,-ATPase (10).At doses severalfold higher, TBT is also rered S, agaric acid, palmitoyl-CoA, and the fluorescein ported to block the adenine nucleotide translocase (11).An anderivatives erythrosin B, erythrosin isothiocyanate, ion channel, which can also be blocked by TBT, has been obrose bengal, and eosin Y. The following evidence sugserved by patch-clamping the inner mitochondrial membrane gests that Cibacron blue,agaric acid, and palmitoyl-CoA (12). inhibit by binding to a common site. 1) They all only Selwyn and co-workers have shown that several other inhibias C1-, partially block the transportof small anions such tors of the adenine nucleotide translocase, i.e. Cibacron blue, NO,, and HCO,, but completely block the transport of larger anions such as malonate. 2) They decrease the agaric acid (131, and palmitoyl-CoA (€9, alsoinhibit IMAC. IC, values of each other in a manner consistent with Based on this and otherevidence, Selwyn and co-workers (13) have proposed that IMAC may be a member of a family of competitivebinding. 3) N-Ethylmaleimidedecreases their IC,, values by a similar extent. 4) Inhibition by all proteins whichincludes the adenine nucleotide translocase, shows no dependence on matrix pH and only a small phosphate carrier, and brown adipose tissue uncoupling prodependence on medium pH. It is suggested that these tein. More recently, the oxoglutarate carrier(14)and the citrate agents may selectively bind to an open state of IMAC carrier (15)have been added to thisfamily. These proteins all and inhibit by decreasing its conductance. The physi- transport anions and they have many significant structural ological nucleotides CoA, NAD+, NADH,NADP’,NADH, similarities, andAquila et al. (16)suggest they may have arisen and ATP do notinhibit; in fact, IMAC is shown to trans- from triplication of a common gene which codes for approxiport ATP. Despite these similarities between IMAC and mately 100 amino acids. The activities catalyzed by these carthe adenine nucleotide translocase, IMAC appears to be riers include electroneutral uniport, electroneutral exchange, a separate entity, sincesome of the IC,, values differby electrophoretic exchange, and electrophoretic uniport. In this up to &fold, and carboxyatractyloside, the most selec- paper, we characterize the inhibitionof IMAC by a number of tive inhibitor of the adenine nucleotide translocase,has inhibitors of the adeninenucleotide translocase, manyof which no effect on IMAC. In addition, IMAC is also able to are categorized as nucleotide analogs. Furthermore, we demtransport AMP, while the adenine nucleotide translo- onstrate thatIMAC is able to transport ATP and AMP.The data case does not. presented suggest thatIMAC is not identical with the adenine nucleotide translocase, but since IMAC is able to transport the substrates of all the members of the carrier family and the The inner membrane of freshly isolated mitochondria nor- modes of transport catalyzed by this family are so varied, it is mally possesses very low electrophoretic permeability to most not difficult to accept that IMAC may also be a member. anions. However, a pathway exists which allows rapid transport of a wide variety of anions if the mitochondria are depleted EXPERIMENTALPROCEDURES of divalent cations (1)and/or the matrix pH is alkaline (2-7). Assay ofAnionZkansport-Anion transport was assayed by following This pathwayhas been given the acronym IMAC’ (inner mem- swelling which accompanies net salt transport, using the light scattering techniqueas described in detail elsewhere (17,18). Using this tech* This work was supported by National Institutes of Health Grant HL nique, we generate a light scattering variable, p, whichnormalizes P 47735 awarded by the National Institute of General Medical Sciences reciprocal absorbance for mitochondrialproteinconcentration, and the National Heart, Lung, andBlood Institute, United States Pub- (milligramdml),according to the formula lic Health Service, Department of Health and Human Services and also by a grantfrom the NorthwesternOhio Chapter of the American Heart part Association. Thecosts of publication of this article were defrayed in by the payment of page charges. This article must thereforebe hereby marked “advertisement” in accordancewith 18 U.S.C.Section 1734 where a is a machine constant and E‘, (equals 1 mg/ml) is a constant introduced to make p dimensionless. solely to indicate this fact. The rate of salt transport is calculated from the rate of change of /3 $ To whom correspondence shouldbe addressed: Dept.of Pharmacology, Medical College of Ohio, P.O. Box 10008,Toledo, OH 43699-0008. according to the formula (18) Tel.: 419-381-4125. The abbreviations used are: IIvlAC, inner membrane anionchannel; J,=pCMS, p-chloromercuribenzenesulfonate; NEM, N-ethylmaleimide; nb ( d t ) TES, N-tris(hydroxymethyl)methyl-2-aminoethanesulfonicacid; TBT, where + is the medium osmolality (110 mosmolal in most studies retributyltin. ~~~

~~~~

w*

10614

Mitochondriain Anion Uniport

0

0.5

1

1.5

t ( m d i

10615 ported here); So is the solute contentof the stock preparation of mitochondria (190nosmoVmg) used to determineb ( 15 mosmolal), the slope of the equilibrium absorbance osmotic curve; and n is the number of moles of osmotically active particles which make up 1mol of the transported salt. At C#J = 110 mosmolal, OS$b is about 1400 nmoVmg. To determine rates of solute transport, we use a Brinkmann Probe Colorimeter (Model PC7001 with a 1-cm probe (2-cm light path). With this probe, for optimum sensitivity, we normally use mitochondria at a concentration between 0.1 and 0.2 mg/ml. For most studies, we use a 520 nm filter for determination of both dpldt and b; consequently, the wavelength dependence of these parameters can be ignored. In the present studies, useof highly colored compounds suchas Cibacron blue and erythrosin B made it necessary to use filters of different wavelength. Since erythrosin isBred and has a peak absorption a t 525.5 nm, we chose t o use a 670 nm filter, while for Cibacron blue, which has an absorption minimum a t about 450 nm, we used a 470 nm filter. Since the magnitudeof light scattering is dependent on wavelength, we chose to normalize our light scattering traces to those obtained at 520 nm. This wasaccomplished by constructing light scattering “protein curves” using aliquots of milk as described in Ref. 17 with each of the three filters. From the slopes (d(A”)/d(pl of m i k - l ) , normalization factors were determined. The ratio of slopes (f, = (slope a t A)/(slope a t 520 nm)) obtained were 1.5 at 670 nm and 0.9 at450 nm. Thus, Equation1was modified as follows.

p 520

n

0

0.5

1

1.5

t ( m d i

Log [ I n h i b i t o r ]

PM)

FIG.1. Cibacron blue inhibits malonate. uniport. Light scattering kinetics of mitochondria (0.12 mg/ml) suspended in K’ malonate assay medium are shown. The assay medium is described under“Experimental Procedures.”A, nigericin(0.5 nmoVmg), rotenone (2 pg/mg), and A23187 (10 nmol/mg) were added at zero time. Valinomycin (0.5 nmoVmg) was added a t 0.25 min. Thefollowing doses of Cibacron blue (p) were added at zero time: a , 0; b, 0.65; c, 1.3; d , 2.6; e, 6.5. Rates of malonate uniport calculatedfrom these traces,as described under“Experimental Procedures,” are: a , 198; b, 133; c, 92; d, 59; e, 28 nmol of malonate/min/mg. B , ascorbate(2.5 m),N,N,N’,N‘-tetramethyl-pphenylenediamine (0.25 m),cytochrome c (10 PM), and rotenone (2 pg/mg) were addeda t zero time. Valinomycin (0.5 nmoVmg) was added at 0.25 min. The following doses of Cibacron blue (p”) were added at zero time: a , 0; b, 0.65; c, 1.3; d , 2.6; e, 6.5.Rates of malonate uniport calculated from these traces as described under “ExperimentalProcedures” are:a, 315; b, 188; c, 119; d , 64; e, 40 nmol of malonate/min/mg.

-A=-(&’ P P

-A f

-

a,)

(Eq. 3)

fA“s

The applicability of this correction was confirmed by comparing standard swelling traces similar to those shown in Fig. 1, obtained at the three wavelengths, and replotting according to Equation 3. After insertion of the appropriate values off, and a,, the traces were indistinguishable. Pretreatment of Mitochondria with Mercurials-The normal mitochondrial stock suspensions (50 mg of proteidml) were diluted 1:l in 0.25 M sucrose containingK+ salts of TES (12 m ~ and ) EGTA (0.5 mM) adjusted to pH 6.7 (at 25 “C) and maintained at0 “C. The desired dose of mercurial was then added, and at least 1min was allowed to elapse after mixing before the mitochondria were transferred to the various assay media. Pretreatment of Mitochondria with NEM-Mitochondria (50 mg/ml) were treated with the desired amount of NEM as described above for mercurials; however, the suspension was kept on ice for 10 min after addition of NEM, and then thioglycolate (10 m ~ was ) added t o terminate the reaction. Assay Media for Anion Dansport-The potassium chloride, malonate, and nitrate media for light scattering studies contained the K’ salts of C1- (55 m)or malonate(36.7 mM) or nitrate (55mM) and EDTA (0.1 m ~ ) EGTA , (0.1 m),and TES(5 m)plus rotenone (2 pg/ml). The pH was adjusted to 7.4 unless indicated otherwise, and the temperature was maintained at 25 “C. Mitochondria (0.1 mg/ml),A23187 (10 nmoV mg), nigericin (1nmoVmg), and valinomycin (0.5 nmol/mg) were added separately to each assay. Binding Studies for erythrosin B-Mitochondria (0.029 mg/ml) were incubated with varying concentrationsof erythrosin B (0-25 PM)for 10 min. After 10 min, the suspension was centrifuged in an Eppendorf bench top centrifuge, and the absorbance of the supernatant was measured at 525.5 nm. Drugs a n d Reagents-Most drugs were obtained from Sigma. The ionophores and rotenone were dissolved in ethanol.Thioglycolate, mersalyl, and pCMS were dissolved in water. Cibacron blue 3GA, alizarin red S, bromcresol green, and rose bengal were dissolved in water. Erythrosin B was dissolved in dimethyl sulfoxide. Rat liver mitochondria were prepared as described previously (17). RESULTS Cibacron blue and Agaric A c i d I n h i b i t IMAC-Selwyn and co-workers (13)h a v e reported that Cibacron blue and agaric C , Hill plots for inhibitionof malonate uniport by Cibacron blue generated from the experiments shown in A ( 0 )and B (0) are shown (J, represents the maximum rate of transport observed in the absence of inhibitor). Fitting the data atostraight line using simple linear regression gives IC,, values and Hill coefficients of 1.23 1.1~and 1.1 and 0.89 p and 1.2 for the Mg2‘ depletion assay and theBrierley assay, respectively. Also shown are data obtained for agaric acid in Mg2‘ depletion assay (A)and using the Brierley assay(A) giving IC,, values and Hill coefficients of 0.56 p~ and 1.4 and 0.52 PM and 1.3, respectively.

10616

Anion Uniport in Mitochondria

TABLE I Inhibition of IMAC and the adenine nucleotide translocase by nucleotide analogs IC,, values and Hill coefficients for inhibition of IMAC were obtained as described in Fig. 1 and under "Experimental Procedures." IC,, (n = Hill coefficient) Dye

ANT

IMAC PM

Cibacron blue 3GA

1.5 ( n = 1.1) 2.8 ( n = 1.9) 0.7 ( n = 2.9) 20.8 ( n = 1.8) 0.6 ( n = 1.7) 63.4 ( n = 1.96) 10.7 ( n = 1.5)

Erythrosin B Erythrosin isothiocyanate Eosin Y bengalRose Alizarin red S Bromcresol green Data taken from Ref. 34. * Spontaneous swellingis observed at doses >4 p

~

3.95" 20.4" 34.7" 2.7b 53.5" 88.8" .

-1.2'

.5

' '0 I

I

I

I

0.5

1

I .5

acid inhibit C1- uniport in mitochondria suspended in ammoLog [ E r y t h r o s i n B] ( M) nium chloride medium at pH 8.0. In this assay, IMAC is activated by alkalinization of the matrix (1) without depletion of FIG.2. Inhibition of IMAC by e r y t k s i n . Hill plots for inhibition matrix M e . When we examined theeffect of Cibacron blue on of malonate (0)and C1- (A)transport are shown. J , represents the was C1- transport in KC1 medium using Mg2' depletion (1) to acti- maximum rate observed in the absence of inhibitor. The experiment show the malvate IMAC, we observed someinhibition at pH 8.4; however, a t carried out as described in Fig. L4.The open circles (0) onate data plotted as a function of the free erythrosin concentration pH 7.4, we foundvery little inhibition and even saw some determined as described in the text. The IC, values and Hill coef€istimulation at low doses. Since we had observed similar behav- cients determined for thesedata are: malonate, 2.8 w, 1.9 (2.0 p ~ 1.7 , , ior with mercurials (19), and for these the transport of larger for free erythrosin); C1- 6.4 p ~ 1.5. anions is inhibited more completely, we went on to investigate the effect of Cibacron blue on malonate uniport. Typical light the 5"position of the benzoic acid moiety increases thepotency scattering traces areshown in Fig. l A , and the corresponding 4-fold. The IC,, for this compound is very close to that of rose Hill plot obtained for the maximum transport rates determinedbengal which has four chloro groups attached to the benzoic in each traceis shown in Fig. 1C (closed circles). Note that the acid moiety. plot is linear suggesting that,for this anion, inhibition iscomSince inhibition by these nucleotide analogs could reflect the plete and reversible. existence of a regulatory nucleotide binding site, we investiIMAC can alsobe assayed using the"Brierley assay" (1,201, gated theeffect of a number of physiologically important nuclein which the channel is activated by alkalinization of the ma- otides, including NAD+,NADP, NADH, NADPH, coenzyme A, trix upon addition of valinomycin to respiring mitochondria in ATP, and ADP. None of these wasfound to inhibitat millimolar Fig. 1B concentrations; moreover, 0.6 mM ATP has no effect on the IC,, the absenceof permeant acids. The traces contained in show that Cibacron blue blocks IMAC also under these condi- for Cibacron bluesuggesting that it does not compete for bindtions. Note, however, that the small increase involume which ing at the inhibitory site (data not shown). immediately follows addition of valinomycin is unaffected by The Hill coefficient for Cibacron blue is close to unity; howCibacron blue. As we have shown previously, this represents ever, for most of the other dyes, as with agaric acid, the Hill uptake of K+ driven by the proton pumps of the respiratory coefficient was found to be greater thanunity. A typical Hill plot chain and concomitant matrix alkalinization (see Ref. 1). The for inhibition by erythrosin B is shown in Fig. 2 (closed circles) Hill plot obtained from these data is shown in Fig. 1C (open and has a slope of 1.9. Since erythrosin B was themost potent circles). inhibitor and is able to bind atovariety of proteins, we examSimilar results were obtained with agaric acid and palmi- ined the bindingof erythrosin B to mitochondria to determine toyl-CoA using both types of assay, and theHill plotsfor agaric whether this hada significant effect on the IC,, and Hillcoefacid are shown in Fig. 1C (triangles). The Hill coefficient of ficient. Threeexperimentswerecarriedout in which the both of these inhibitors was found to be higher than that for amount of bound erythrosin B was determined by measuring Cibacron blue. For all three inhibitors, both assaysyield very the free concentration spectrophotometrically following censimilar IC,, values. This suggests that, unlike inhibition by trifugation of the suspension topellet the mitochondria. In the M e (21) and drugs like propranolol (22), inhibition by Ciba- first experiment, the time course of binding was examined, and cron blue, agaric acid, and palmitoyl-CoA is not very sensitive the absorbance of the supernatantceased to decline after about 10 min. In the second experiment, the concentrationof erythto matrix pH. Inhibition by Other Nucleotide Analogs-Cibacron blue is a rosin B was varied at a constant mitochondrial concentration member of a class of agents frequently referred to nucleotide as (0.6 mg/ml) while in the third, themitochondrial concentration analogs, which contain moieties which resemble AMP (23).We was varied (0.3-1.9 mg/ml) at a constant erythrosin B concenhave examined theeffect of a number of these agents and find tration (10 VM). The results of these latter experiments, conthat most of them inhibit IMAC. The most potent of these tained in Fig. 3, reveal that, over the range examined, erythagents are the tetraiodofluorescein derivatives erythrosin B, rosin B binding to mitochondria is a saturable process which erythrosin 5'-isothiocyanate, and rose bengal. Fluorescein it- may be describedby a Kd of 7.75 PM and a B,, of 42.6 nmol/mg. self, however, has no effect, while eosin Y (tetrabromofluores- The high number that most of the sites of binding sites suggests cein), alizarin red S, and bromcresol green are relatively weak must be nonspecific. Since both sets of data maybe adequately inhibitors (see Table I). The potency of the fluorescein deriva- described by the same curve, binding appears to be independtives is modulatedby the substituents of the rings. Thus, sub- ent of absolute erythrosinB and mitochondrial concentrations; stituting thefour iodo groups of erythrosin B for Br (eosin Y) consequently, we were able to use these values of Kd and B,,, to decreases the potency 7-fold, while adding a n -NCS group to calculate the concentration of free erythrosin B in our dose-

P

Uniport Anion

in Mitochondria

10617

c5

m E

\ I+

0 E

c

Y

U

c 53

0

m m W

‘0

I

2

3

[ E r y t h r o s i n B]

4

r

5

( M)

[ C i b a c r o nb l u e ]

( M)

/tl

FIG.3. Binding of erythrosin B to mitochondria.Bound erythrosin B (nmoWmg) is plotted uersus free erythrosin B (pd. Free erythrosin concentration was varied in two ways: 1) by varying total erythrosin concentration at a constant mitochondrial concentration (0.6 mg/ml (0))and 2) by varying the mitochondrial concentration at a constant total erythrosin concentration (10p,0). Both sets of data may be described by a curve with a B,, = 42.6 nmoWmg and aKd = 7.75 p and aHill coefficient= 1. See “Experimental Procedures”for further details.

FIG.4. Dose-response curves for Cibacron blue inhibition of nitrate and phosphate transport. The rate of transport for nitrate (0)and phosphate (0) (pmol/mg/min)are plotted uersus the concentration of Cibacron blue(w).The experiment was carried out as described in Fig. 1 except the pH of the assay media was7.8 and the concentration of mitochondria was 0.08 mg/ml. Uniport of Pi was measured in mitochondria which werepretreated withN-ethylmaleimide (30 nmollmg)to inhibit the classical electroneutral phosphate carrier as described under “Experimental Procedures.”

response curve. The results contained in Fig. 2 (open circles) show that this correction decreases the IC,, from 2.8 w to 2.0 PM; however, there is only a small decrease in the Hill coefficient; consequently, the difference in Hill coefficients obtained with Cibacron blue and erythrosin B cannot be attributed to nonspecific binding. Moreover, due to the relatively low mitochondrialconcentrations used in the light scattering assay, nonspecific binding has only a small effect on the free concentration of these dyes. Reversibility of inhibition by Cibacron blue, erythrosin B, and erythrosinisothiocyanate was investigated by pretreating the mitochondria with 50 dye prior to transfer to the assay medium. In no case was the inhibition observed greater than that predicted from the final concentration of the dye in the assay medium (0.2 w)(data notshown); moreover, the percent inhibition did not increase withtime. Thus, inhibition by these agents appears tobe freely reversible. Effect of Nucleotide Analogs on Uniport of C1; N O , a n d HCOj-Having established the inhibitory properties of these agents on malonate uniport, we went back to examine their effect on the smaller anions C1-,NO;, and HCO;. A typical dose-response curve for inhibition of C1- transport by erythrosin B is shown in Fig. 2 (closed triangles). The data havebeen plotted assuming that inhibition can reach 100% and yield an IC,, of 6.4 w with a Hill coefficient of 1.5. Of all the inhibitors tested, erythrosinB was found to be among the most efficacious inhibitors of C1- transport. In contrast, Cibacron blue was the least efficacious, being unable to block the transport ofC1-, NO;, or HCO, completely. This phenomenon is illustrated by the datacontained in Fig. 4, in which we compare inhibition of NO; and Pi uniport. The curves, fitted to the data without assuming 100% inhibition, yield IC,, values of 6.8 w (Hill coefficient 1.05) for NO; and 2.5 1.1~(Hill coefficient 1.5) for Pi. The predicted maximum inhibition is 56% for NO,, while it is greater than 96% for Pi. Alizarin red S and rose bengal also inhibited incompletely, while bromcresol green and eosin Y were found to behave similarly to erythrosin B. Another phenomenon associated with these nucleotide analogs is that at low doses they frequently cause a small stimu-

lation of the swelling rate in KC1 a t pH 7.4. Although stimulation of malonate uniportby nucleotide analogs is not evident in rat liver mitochondria, the phenomenon does not appear tobe peculiar to C1-, since in potato mitochondriait is seen with both C1- and malonatefor both Cibacron blue and erythrosinB (24). In these mitochondria, the IC,, values are, however, about 10fold higher, and this may be related to thedifference. Relationship between Nucleotide Analogs and Other Inhibitors of IMAC-The data contained in Fig. 1 suggest thatinhibition by nucleotide analogs is not modulated by matrix pH; however, data presented elsewhere (25) suggest that external pH may play a small role. Raising the pH of the suspending medium from pH 7.5 to pH 8.5 leads to a n increase in the IC,, for Cibacron blue from 1.5 to2.4 1.1~and anincrease in theIC,, for agaric acid from 0.8 PM to 1.1 PM. These effects are much smaller than the effect of pH on inhibition by Mg2‘and cationic amphiphiles. Thus, inhibition by the nucleotide analogs appears to occur by a different mechanism. Like the nucleotide analogs, under certain conditions, mercurials andN-ethylmaleimide are also able to stimulateIMAC (19, 261, and this hasbeen shown to result from a shift in the IC,, for protons. We could find no evidence, however, that the effect of nucleotide analogs is mediated by a similar mechanism. In fact, the data contained in Fig. 5 show that addition of Cibacron blue (1.3J ~ M )and agaricacid (0.5 PM)has theopposite effect increasing the PIC,, for protons from 7.89 to 8.34 and 8.29, respectively. This effect is predicted by the observed pH dependence of the Cibacron blue IC,,. We have also determined the effect of N-ethylmaleimide on inhibition by Cibacron blue, agaric acid, and palmitoyl-CoA. The dataobtained are summarized inTable I1 and reveal that N-ethylmaleimide decreases the IC,, values for all threeinhibitors. This finding suggests that these inhibitors do not bind at the NEM-reactive site. It is also noteworthy that this effect of NEM is in theopposite direction to its effect on the IC,, values for protons, M e , and propranolol (26). Is Inhibition by Nucleotide Analogs Related to Inhibition by Mersalyl?-Of all the inhibitors ofIMAC thus far identified, only mercurials and the nucleotide analogs appear to inhibit

10618 Mitochondriain

Uniport Anion p d and tributyltin (2.8 nmol/mg). This latter dose is not ex-

h

cn E

c -r(

E \ d

0 E

-3 7 \ )

.,

FIG.5. Effect of Cibacron blue and agaric acid on inhibition of IMAC by H+. The reciprocal of the rate of malonate uniport (min/mg/ pmol) is plotted uersus proton concentration (m).0,control; A, plus 1.3 pd Cibacronblue; plus 0.5 pd agaric acid. The experiment was carried out as described in Fig. LA except the mitochondria concentration was 0.07 mg/ml and valinomycin (0.5 nmol/mg) was added at zero time. The PIC,, and slopes determined from these Dixon plots are: control, 7.89, 0.11; plus Cibacron blue, 8.36,0.44; and plus agaricacid, 8.29, 0.31.

pected to block the adenine nucleotide translocase (11).Mersalyl and dicyclohexylcarbodiimide also inhibited 98% when used at doses which block IMAC (data not shown). Only with propranolol was inhibition lower than expected, 300 VM inhibited by only 50%. However, we believe that this probably results from an interactionbetween the cationic amphiphile and the polyanionic ATP, since in other experiments inwhich malonate was the substratefor IMAC (not shown), we found that when Cibacron blue and propranolol were present together, the percent inhibition was less than thatobserved when each was present alone. We have also investigated the effect of pretreating mitochondria with carboxyatractyloside, a specific inhibitor of the adenine nucleotide translocase. Whether added alone or together with oligomycin (Fig. 7, traces d and e, respectively), carboxyatractyloside had no effect on the transport ofATP induced by A23187 and valinomycin. These findings strongly suggest that IMAC and not the adenine nucleotide translocase is responsible for the observed ATP transport. In similar experiments, we have also shown that AMP is transportedat rates of about 30 nmol/midmg at pH 7.4. Interestingly, the ratesof both ATP and AMP transport are significantly faster than some smaller anions such as glucuronate, gluconate, and adipate (27). DISCUSSION

incompletely. Thus, we investigated whether mercurials could block inhibition by Cibacron blue. To prevent any interference from effects of mercurials reacting at the N-ethylmaleimidesensitive site, the mitochondria were first treated with NEM before investigating theeffect of the mercurial. pCMS was chosen since, unlike mersalyl (19), it does not block malonate transport completely, and, therefore, the IC,, for Cibacron blue could be determined in pCMS-treated mitochondria. The results, summarized in Table 11, show that pCMS increases the IC,, for Cibacron blue but does not preventinhibition. Thus, we conclude that although themechanisms of inhibition differ, the binding site of the mercurial may be close enough to that of Cibacron blue to interfere with its binding. This effect may result from an interactionbetween the manynegative charges on Cibacron blue and the negative charge on pCMS, since pCMS has a negligible effect on the IC,, values for erythrosin B, bromcresol green, and agaric acid which all lack the sulfonate groups that Cibacron blue possesses. Interestingly, however, pCMS does appear to decrease to the Hill coefficients of all these inhibitors to valuesclose to 1. Do Nucleotide Analogs Znhibit ZMACby Binding to a Common Site-To furtherinvestigatetherelationship between these inhibitors, we examined the effect of each on the IC,, of the others. The data contained in Fig. 6, A and B, show that addition of agaric acid increases theIC,, for Cibacron blue, and that Cibacron blue increases theIC,, for agaric acid. Note that the Dixon plots are parallel as predicted for inhibitors which bind at a common site. Similarly, palmitoyl-CoA increases the IC,, for Cibacron blue (data not shown). Can ZMAC Dunsport ATP?-The finding that several inhibitors of the adeninenucleotide translocase also inhibit IMAC led us to examine the possibility thqt IMAC may also transport ATP. The data contained in Fig. 7 show that mitochondria suspended in27 mM potassium ATP at pH 7.4 do indeed swell when A23187 and valinomycin are added. Since the rate observed, which corresponds to about 50 nmol of ATPImidmg, is much slower than the rate ofC1- or malonate transport, we examined the effect of inhibitors of IMAC to establish that the swelling did in fact reflect the activity of IMAC. As shown in Fig. 7, this transportis blocked 97% by both Cibacron blue (30

In thispaper, we have demonstrated that many inhibitors of the adeninenucleotide translocase block IMAC and thatIMAC can transportATP. These findings are consistent with the proposal of Selwyn and co-workers (13)that IMAC may be a member of the family of anion portersfound in the innermitochondrial membrane which share a common basic structure. This family includes the phosphate carrier (16, 281, the adenine nucleotide translocase (29,30), the uncoupling protein (16),the oxoglutarate carrier (141, and also the tricarboxylate carrier (15). Interestingly, IMAC is able to transport all the substrates of these carriers, but ina n electrophoretic manner. In the present studies, we have confirmed the inhibitory properties of Cibacron blue,agaricacid, and palmitoyl-CoA first reported by Selwyn and co-workers (8, 13). Moreover, our finding thatthetransport of largeranions is completely blocked suggests that this lack of complete inhibition truly reflects incomplete inhibition and not anincreased permeability due toa detergent effect as suggested by Halle-Smith et al. (8) for palmitoyl-CoA. Incomplete inhibition appears to be a common property of this class of inhibitors; however, the maximum extent of inhibition is dependent not only on the anionic species being transported butalso on the species of inhibitor. Of the nucleotide analogs, erythrosin B appears to be the most efficacious. Incomplete inhibition of IMACis also observed with mercurials, and this has been attributed to a simple steric effect in which the mercurial partially blocks the anion channel or the anion binding site (19). A similar mechanism may be responsible for the nucleotide analogs; however, our data clearly demonstrate that the nucleotide analogs and the mercurials do not bind to the same site. Since all the nucleotide analogs possess one or more negative charges, itis interesting to speculate whether they might inhibit transport by competing with the substrate anionsfor a positively charged binding site or region of the pore. One intriguing property of this group of inhibitors is their ability to stimulate anion uniport at low doses. This stimulaincrease in rate at low doses tion is sometimes seen as an actual and may contribute to the nonlinearity of the Dixon plots. Stimulation can also be induced with mercurials and NEM; however, these agents stimulate IMAC by increasing the IC,,

Mitochondriain

Uniport Anion

10619

TABLEI1 Effect of N-ethylmaleimide and mercurials The data were obtained as described in the legend to Fig. 1 for malonate transport at pH 7.4.Pretreatment protocols for NEM and pCMS are described under “ExDerimental Procedures.” IC,, ( n = Hill coefficient) Experiment

Control

+NEM

NEM

+ pCMS

PM

1.66(n = 0.95)0.82

Cibacron blue Cibacron blue .Agaric acid Agaric acid Palmitoyl-CoA Erythrosin B Bromcresol meen

1.09(n = 1.37) 1.79( n = 1.34)0.90

(n = 0.88(n = 0.61 (n = 0.44(n = (n= 1.90( n = 8.9 ( n =

1.31) 1.46)1.53 1.42) 1.46) 1.37) 1.31) 1.36) I

( n = 1.02)

0.38(n = 1.1) 1.85( n = 0.90) 7.4(n = 1.01) I

I

0

3 I 0,

I

I

I

I

I

2

3

4

[Cibacron blue]

I

5

)AIM)

0.5

I

1.5

t (min) FIG.7. IMAC transports ATP. Light scattering kinetics of oligomycin-treated (2.3nmol/mg) mitochondria(0.12mg/ml) suspended in K+ATP assay medium are shown. Nigericin(0.5 nmol/mg)and rotenone (2 pg/mg) wereadded at zero time. A23187 (10 nmol/mg) wasadded at 0.3 min, and valinomycin (0.5 nmovmg) was added at 0.5 min. Dace a , control. Dace b, TBT (2.9nmol/mg) was added at 0.1 min. Dace c, Cibacron blue (30p)was added at zero time. Dace d , mitochondria were pretreated with carboxyatractyloside(2 nmol/mg). Dace e, mitochondria were pretreated with carboxyatractyloside(2nmol/mg) without oligomycin. The assay medium contained the K+salts ofATP (27 m ~ )TES , (5 m ~ ) 0.1 , mM EGTA, 0.1m~ EDTA and was adjusted to pH 7.4at 25 “C.

note, however, that these agentsalso have a dual effect on K+ uniport in intactmitochondria (31). In this case, however, low doses (I,, = 0.13 1.1~Cibacron blue) of the dyes inhibit K+uniport while higher doses (EC,, = 13 1.1~Cibacron blue) stimulate. These findings lead one to question whether there is some functional relationship between the two channels, which have I I I both been proposed to be involved in mitochondrial volume 0.5 I 1.5 homeostasis. Other points for comparison include the findings that mercurials(19) and TBT (9) block IMAC and stimulate the [ A g a r i c Acid] OJM) K+ channel (32-351, while dicyclohexylcarbodiimide and proFIG.6. Cibacron blue and agaric acid inhibit at the same site. pranolol are reported to block both pathways (11,22,36,37).In The reciprocal rate of malonate uniport is plotted uersus the concentraA, inhibition by Cibacron blue. 0, control, slope contrast,ATP,ADP, and AMP inhibit K+uniport (311, while they tion of inhibitor (p). 3.40,IC,, = 1.07p;0, +0.5 p agaric acid, slope 3.48,IC, = 1.96p. have no effect on the rate of anion uniport. Furthermore, proB , inhibition by agaric acid. 0, control, IC,, = 0.53 p;0,+1.3 p tons inhibit IMAC (l),but K+ uniport shows little dependence Cibacron blue, IC, = 0.95 PM. The experiments were carried out as on pH (31). described in Fig. 1 except the protein concentration was 0.17 mg/ml. Quite a large number of inhibitors of IMAC have been deThe compositionof the assay medium is described under “Experimental scribed (see Ref. l for review). One explanation for this which Procedures.” has been proposed is that IMAC exists in open and closed for H+. This does not appear to be the mechanism of action of conformations and that interaction with avariety of agents can the nucleotide analogs, sincethese agentsdecrease the IC,, for lock it into one of the two conformations (38). The effect of NEM protons. This effect is consistent with the pH dependenceof the could then be explained by a shift in the opedclosed equilibIC,, for Cibacron blue, since a decrease in pH would lead toa n rium constant.For example, the effect of NEM on inhibition by increase in the percent inhibition by the dye. Thus, the cause of H+, M P , and propranolol could be explained if these agents the stimulation remains to be determined. It is interesting to bind selectively to the closed state. In contrast,since Cibacron

I

10620

Anion Uniport in Mitochondria

blue cannot block transport completely, it cannot inhibit by observe (60 nmol ofATP/midmg) aremuch faster than the rate binding to the putative closed state, but must inhibit by binding of 8.3 nmol/min/mg reported by Austin and Aprille (43) for the to anddecreasing the conductance of the open state. Consistent V,,, of ATP influx on their Me-activated transporter. with this hypothesis is our finding thatNEM decreases the IC,, AcknowledgmentJoel Shiffler is thanked for expert technical assistfor Cibacron blue, palmitoyl-CoA, and agaricacid. The effect of pH on the IC,, for Cibacron blue and theeffect of Cibacron blue ance. on the PIC,, for protons are both consistent with a pH depenREFERENCES dence of Cibacron blue binding. This pH dependence is much 1. Beavis, A. D. (1992) J. Bioenerg. Biomembr 24, 77-90 smaller than that seenwith M e and propranolol, and, since 2. Azzi, A , , and Azzone, G . F. (1966) Biochim. Biophys. Acta 120, 466-468 the IC,, is the sameor lower when the Brierley assay is used, 3. Azzi, A,, and Azzone, G . F. (1967) Biochim. Biophys. Acta 131, 46-76 4. Brierley, G . P. (1969) Biochem. Biophys. Res. Commun. 35,396-402 the protonation site is probably located on the outersurface of 5. Brierley, G . P. (1970) Biochemistry 9,677-707 the inner membrane. 6. Selwyn, M. J.,Dawson, A. P., and Fulton, D. V.(1979)Biochem. Soc. Duns. 7 , The following common properties lead us to suggest that the 21G219 7. Garlid, K. D.,and Beavis, A. D. (1986)Biochim. Biophys. Acta 853, 187-204 nucleotide analog dyes, agaric acid, and palmitoyl-CoA all in8. Halle-Smith, S. C., Murray, A. G., and Selwyn, M. J. (1988) FEES Lett. 236, hibit by a common mechanism. All only partially inhibit C1155158 9. Powers, M. F., and Beavis, A. D. (1991) J. Biol. Chem. 266, 17250-17256 transport, all have IC,, values which are decreased by NEM, 10. Sone, N., and Hagihara, B. (1964) J. Biochem. fTokyo) 56,151-156 each increases theIC,, values of the others, and all have a very 11. Hams, E. J.,Bangham, J. A,, and Zukovic, B. (1973) FEES Lett. 29,339-344 small dependence on pH. Although these compounds all block 12. Antonenko,Y.N., Kinnally, K. W., Perini, S., and Tedeschi, H. (1991)Biophys. J. 59,216a the adeninenucleotide translocase (39-42), inhibition of anion 13. Murray,A. G., Halle-Smith, S. C., and Selwyn, M. J.(1988)EBECRep. 5,206 uniport does not appearto be related to this. Many inhibit both 14. Runswick,M. J., Walker, J. E., Bisaccia, F., Iacobazzi, V., and Palmieri, F. (1990) Biochemistry 29, 11033-11040 processes with similar IC,, values; however, erythrosin B, bro15. Kaplan,R. S., Mayor, J. A., and Wood, D. 0. (1993) J. Biol. Chem. 268, mcresol green, and Rose bengal are significantly more potent 13682-13690 inhibitors of IMAC than of the translocase. Also, the most se- 16. Aquila, H.,Link, T. A,, and Klingenberg,M. (1987) FEES Lett. 212, 1-9 lective inhibitor of the translocase, carboxyatractyloside, has 17. Beavis, A. D.,Brannan, R. D., and Garlid, K. D.(1985) J. Biol. Chem. 260, 13424-13433 no effect on the activity ofIMAC even when ATP is being 18. Garlid, K. D.,and Beavis, A. D. (1985)J. Biol. Chem. 260, 13434-13441 transported. The requirements for inhibition of IMAC at this 19. Beavis, A. D.(1989) Eur. J. Biochem. 185,511-519 site appear to be negative charges and/or a hydrophobic tail. 20. Brierley, G. P. (1970) Biochemistry 9,697-707 21. Beavis, A. D., and Powers, M. F. (1989) J. Biol. Chem. 264, 17148-17155 One could imagine that there ais very weaknucleotide binding 22. Beavis, A. D. (1989) J . Biol. Chem. 264,1508-1515 site, to which the so-called nucleotide analogs bind stronglyand 23. Neslund, G. G., Miara, J. E., Kang, J.J., and Dahms, A. D. (1984) Curr Top. Cell. Regul. 24,447-469 to which other groups such as the CoA moiety of palmitoyl-CoA 24. Beavis, A. D., and Vercesi, A. E. (1992) J. Biol. Chem. 267,30793087 and citrate moiety of agaric acid only bind when anchored to 25. Powers, M. F., and Beavis, A. D. (1990) Biophys. J. 57, 179a the membrane by their hydrophobic tails. Neither palmitoyl- 26. Beavis, A. D.(1991) Biochim. Biophys. Acta 1063, 111-119 27. Beavis, A. D., and Garlid, K D. (1987) J. Biol. Chem. 262,15085-15093 carnitine nor CoA has any effect. In this respect, IMAC is 28. Runswick, M. J., Powell, S. P., Nyren, P., and Walker,J. E. (1987)EMBO J . 6, 1367-1373 probably very similar to the adenine nucleotide translocase. W., Wachter, E., Aquila, H., and Klingenberg, M. (1991) Biochim. BioThe sensitivity of IMACand the adenine nucleotide translocase 29. Babel, phys. Acta 670,176-180 to TBT also appears todiffer. We have shown that doses of less 30. Aquila, H., Misra, D., Eulitz, M., and Klingenberg,M. (1982)Hoppe Seyler's 2. Physiol. Chem. 363,345-349 than 3 nmoVmg block IMACcompletely even in thepresence of 31. Beavis, A. D., Lu, Y.,and Garlid, K. D. (1993) J. B i d . Chem. 268,997-1004 27 mM ATP, while Harris et al. (11)showed that, in the presence 32. Diwan, J. J.,Aronson, D., and Gonsalves, N. V.(1980) J. Bioenerg. Biomembr of ATP, doses as high as 6 nmol/mg have no effect on the ad12, 205-212 33. Diwan, J. J., Srivastava, J., Moore, C., andHaley, T. (1986) J. Bioenerg. enine nucleotide translocase. Biomembr. 18,123-134 ATP- and AMP2- are currently the largest anions which 34. Diwan, J.J. (1982) J. Bioenerg. Biomembr 14, 15-22 have been shown to be transported via IMAC. This finding 35. Diwan, J. J., DeLucia, A., and Rose, P.E. (1983) J . Bioenerg. Biomembr 15, 277-288 again demonstrates that the number of charges and perhaps 36. Beavis, A . D.,and Garlid, K. D. (1988) J . Biol. Chem. 263,7574-7580 location of these charges havea significant effect on the rateof 37. Gauthier, L. M., and Diwan, J.J.(1979) Biochem. Biophys. Res. Commun.87, 1072-1079 transport of anions throughIMAC. It should alsobe noted that 38. Beavis, A. D.(1994) in The Molecular Biology of Mitochondrial l h ~ n s p o r t this uniport pathway for ATP-, when active, is able to transSystems (Forte, M., and Colombini, M., eds) Springer-Verlag, pp. 137-151 port ATP- much faster than the MgATP carrier which has been 39. Boos, K. S., and Schlimme, E. (1981)FEES Lett. 1 2 7 , 4 0 4 4 F., Lauauin, G . , Lunardi, J.. Duszynski, J., and Vignais, P. V. (1974) studied by Aprille's group (43).The fact that the latter pathway40. Morel, FEkS'Lett. 39, 133-138 appears toinvolve net transportof M e and ATP suggests that 41. Chavez, E., and Klapp, M. (1978) Biochem. Biophys. Res. Commun. 67,272278 it is unrelated to the transport process demonstrated in the 42. Petrescu, I., Lascu, I., Porumb, H., Presecan, E., Pop, R., and Banu, 0. (1982) present work which is stimulated by removal of M e . It should FEBS Lett. 141,148-152 also be pointed out, however, that the rates of transport we 43. Austin, J., and Aprille, J.R. (1984) J. Bid. Chem. 259, 156160