Interaction of mitochondrial phosphate carrier with fatty acids and hydrophobic phosphate analogs

Interaction of mitochondrial phosphate carrier with fattyacids and hydrophobic phosphate analogs

MarkeÂ ta ZÏ aÂ cÏ kovaÂ a, Reinhard KraÈ merb, Petr JezÏ eka,*aInstitute of Physiology, Department of Membrane Transport Biophysics, Academy of Sciences of the Czech Republic, VõÂdenÏskaÂ 1083,

14220, Prague 4, Czech RepublicbInstitut fuÈr Biochemie, UniversitaÈt zu KoÈln, 50674, KoÈln, Germany

Received 3 September 1999; accepted 17 January 2000

Abstract

Mitochondrial transporters, in particular uncoupling proteins and the ADP/ATP carrier, are known to mediateuniport of anionic fatty acids (FAs), allowing FA cycling which is completed by the passive movement of FAs acrossthe membrane in their protonated form. This study investigated the ability of the mitochondrial phosphate carrier tocatalyze such a mechanism and, furthermore, how this putative activity is related to the previously observed

HgCl2-induced uniport mode. The yeast mitochondrial phosphate carrier was expressed in Escherichia coli and thenreconstituted into lipid vesicles. The FA-induced H+ uniport or Clÿ uniport were monitored ¯uorometrically afterHgCl2 addition. These transport activities were further characterized by testing various inhibitors of the two

di�erent transport modes. The phosphate carrier was found to mediate FA cycling, which led to H+ e�ux inproteoliposomes. This activity was insensitive to ATP, mersalyl or N-ethylmaleimide and was inhibited bymethylenediphosphonate and iminodi(methylenephosphonate), which are new inhibitors of mitochondrial phosphate

transport. Also, the HgCl2 induced Clÿ uniport mediated by the reconstituted yeast PIC, was found to be inhibitedby these reagents. Both methylenediphosphonate and iminodi(methylenephosphonate) blocked unidirectional Clÿ

uptake, whereas Clÿ e�ux was inhibited by iminodi(methylenephosphonate) and phosphonoformic acid only. Theseresults suggest that a hydrophobic domain, interacting with FAs, exists in the mitochondrial phosphate carrier,

which is distinct from the phosphate transport pathway. This domain allows for FA anion uniport via thephosphate carrier and consequently, FA cycling that should lead to uncoupling in mitochondria. This might beconsidered as a side function of this carrier. 7 2000 Elsevier Science Ltd. All rights reserved.

Keywords: Yeast phosphate carrier; Fatty acid cycling; H+ uniport; Clÿ uniport; Liposomes

1. Introduction

Uniport of anionic fatty acids (FAs) mediated

by several mitochondrial carrier proteins was

The International Journal of Biochemistry & Cell Biology 32 (2000) 499±508

1357-2725/00/$ - see front matter 7 2000 Elsevier Science Ltd. All rights reserved.

PII: S1357-2725(00 )00006-6

www.elsevier.com/locate/ijbcb

* Corresponding author. Tel.: +011-4202-475-2285; fax:

+011-4202-455-2488.

E-mail address: [email protected] (P. JezÏ ek).

made responsible for the FA cycling mechanismleading to uncoupling [1±4]. After protein-mediated passage to the trans side of the mem-brane, FA anions become protonated and neutralFAs, while physically carrying a proton, canreadily move back across the membrane by a¯ip±¯op mechanism [5]. The carrier protein itselfdoes not physically transport protons, by the FAcycling mechanism. The existence of FA cyclinghas been supported experimentally. Examplesinclude the uncoupling protein of brown adiposetissue mitochondria (UCP1) [2,4,6], the plantuncoupling mitochondrial protein (PUMP) [7], thenewly discovered uncoupling proteins UCP2 andUCP3 [8], and the ADP/ATP carrier [3,9±11]. Allthese proteins belong to the mitochondrial carriergene family [12,13]. The strongest support for theFA cycling mechanism was provided by the ®ndingthat so-called inactive fatty acids, i.e. FA deriva-tives that are unable to undergo ¯ip±¯op move-ment across the lipid bilayer, were also found to beunable to induce H+ uniport in proteoliposomescontaining UCP1 [6], PUMP [7] or other proteins(UCP2, UCP3, ZÏ aÂ cÏ kovaÂ and JezÏ ek, unpublished).FA interaction with two other members of themitochondrial gene family, namely the dicarboxy-late carrier [14] and the aspartate/glutamate car-rier [15] has also been reported.

The mitochondrial phosphate carrier (PIC)[16±21] is a member of the mitochondrial carriergene family. It mediates electroneutral phosphateuptake, which can be interpreted as a stoichio-metric Pi/H

+ symport [22,23], or alternatively asa Pi/OHÿ antiport [23,24]. Phosphate wassuggested to be transported in monovalent form[25], but arsenate and divalent mono¯uoropho-sphate [26] are also translocated by PIC.Recently, we identi®ed methanephosphonate as anew substrate of PIC, while butyl- and decane-phosphonate were found to not be translocated,and neither were the other phosphonates such asphosphoformate and phosphonopyruvate [27].However, we have found that methylenedipho-sphonate and iminodi(methylenephosphonate)are e�ective inhibitors of the electroneutral Pi

uptake in mitochondria [27].It has been observed that PIC can be function-

ally converted to a uniport mode by HgCl2 modi-

®cation of a single cysteine residue (C28 in yeastPIC) [28]. E�ux of phosphate and various diva-lent anions (sulfate, aspartate), of glucose andlysine [23,25] as well as of Clÿ [29] has beenobserved in this mode. It is not yet understood ata molecular level why mercury bound to a singlecysteine converts the carrier into an unspeci®cuniporter. In particular, it is not known whetheranions are translocated via the ``regular'' phos-phate translocation pathway in this mode or ifthe uniport pathway is di�erent. Nothing isknown about the existence of the uniport modein vivo or about a putative physiological role.

We previously found that mitochondrial PIC isinhibited by FAs, most strongly by 12-(4-azido-2-nitrophenylamino)dodecanoic acid (AzDA) andthat PIC can be speci®cally photolabeled by the azi-donitrophenyl FA derivatives [27]. This led us tospeculate that PIC might mediate FA cycling aswell. Consequently, in this paper, we investigated ifPIC was able to mediate FA cycling and whetherthe FA uniport has any relation to the HgCl2-induced uniport mode observed for this carrier.As a result, we found (i) that PIC does mediateFA cycling and (ii) that common inhibitors doexist for these two non-classical functions of PIC.

2. Material and methods

Most of the chemicals were purchased fromSigma, St.Louis, MO, USA. Hydroxylapatite,BIO-GEL HTP was from Bio-Rad, Richmond,VA, USA. Octylpentaoxyethylene was fromBachem Feinchemikalien, Bubendorf, Switzer-land; all other chemicals were of a reagent grade.

2.1. Expression of yeast phosphate carrier inEscherichia coli

The expression of yeast phosphate carrier in E.coli was carried out as previously described [28].Brie¯y, E. coli expression strain BL21 [30] wastransformed with plasmid pNYHM31 containingcDNA coding for yeast PIC. One liter of 2� YT(1% yeast extract, 1.6% bactotryptone, 0.5%NaCl, pH 7) medium, supplied with carbenicillin(100 mg), was inoculated with an overnight col-

M. ZÏaÂcÏkovaÂ et al. / The International Journal of Biochemistry & Cell Biology 32 (2000) 499±508500

ony of transformed BL21. Cells were grown toan A600 of 0.6 under vigorous shaking at 378C.Isopropyl-1-thio-b-galactopyranoside (1 mM) inthe presence of carbenicillin, was used to initiatePIC expression while continuing cell growth for3 h. The pellet from 125 ml of culture was sus-pended in TE-bu�er (10 mM Tris±HCl, 0.1 mMEDTA, 1 mM DTT, pH 7) and passed twicethrough a French pressure cell, followed by10 min of centrifugation at 12,000 g. The reho-mogenized pellet was centrifuged at 1100 g andthe supernatant was again centrifuged at 12,000 g.The inclusion bodies obtained were stored atÿ708C.

2.2. Isolation of phosphate carrier

Isolation of yeast PIC from inclusion bodieswas done according to a method previouslydescribed [28]. The pelleted inclusion bodies(about 3 mg wet weight) were suspended andwashed two times by centrifugation in TE bu�er.The washed pellet was presolubilized by 1.5 mlof 5 mM tetraethylammonium N-tris[hydroxy-methyl]methyl-2-aminoethane sulfonate (TEA)±TES, 30 mM TEA2SO4, 0.1 mM Tris±EDTA,pH 7.2, containing 0.3% sodium lauroylsarcosi-nate (SLS). After centrifugation at 14,000 g for2 min, the resulting pellet was solubilized in0.75 ml of 5 mM TEA±TES, 30 mM TEA2SO4,0.1 mM Tris±EDTA, pH 7.2, containing 1.67%SLS and 1% octylpentaoxyethylene (OctylPOE)(w/w). This procedure yields a crude protein con-taining some contaminating proteins from the in-clusion bodies. However, the majority of theseare eliminated by the presolubilization step. Inmost cases, this product was used for reconstitu-tion.

2.3. Reconstitution of phosphate carrier and¯uorescent monitoring of H+ and Clÿ ¯uxes

Reconstitution was performed by the detergentremoval method designed for transport monitor-ing by ¯uorescence dyes. The ¯uorescence ofdyes (see below) was monitored on a Shimadzu,RF5301 PC (Shimadzu, Japan), equipped withcross-oriented Polaroid polarization ®lters (exci-

tation vertically). Lipids (39 mg egg yolk lecithin,1.66 mg cardiolipin and 0.66 mg phosphatidicacid) were solubilized by 76 mg OctylPOE in ali-quots of stock solutions, so that the resultant1 ml mix contained 150 ml of solubilized inclusionbodies (approx. 150 mg of total protein) and itscomposition matched the desired internal med-ium of the vesicles to be formed. In addition,2 mM of the ¯uorescent anion probe SPQ(6-methoxy-N(3-sulfopropyl)-quinolinium) wasincluded. After this, the mixture was incubatedfor 2 h in the ®rst 5 ml column ®lled with Bio-Beads SM2 (Bio-Rad). The proteoliposomesformed were removed by centrifugation and thenincubated for 10 min in the second Bio-Beads-SM2 column and then removed again. The sec-ond step removed the residual detergent. ExternalSPQ was washed by the passage through Sepha-dex G25-300 spin columns [2,6,7]. All columnswere pre-washed by the given internal medium.

In order to assay H+ ¯uxes, monitored asdH+, i.e. changes in internal [H+] relative to theinitial state, a method based upon SPQ quench-ing by TES anion was employed [2,6±8]. The in-ternal medium contained 28.8 mM TEA±TES(9.2 mM TEA), pH 7.2, 84.4 mM TEA2SO4, and0.6 mM TEA±EGTA, whilst the external med-ium consisted of 84.4 mM K2SO4 instead ofTEA2SO4. Vesicles (25 ml; 1 mg lipids) wereadded to 2 ml of external medium. After 10 s, aFA, usually lauric acid, was added and its redis-tribution on both sides of the membrane lead tointerior acidi®cation which is indicated by theSPQ ¯uorescence increase [2,5,6]. After another10 or 20 s, an H+ e�ux was initiated by the ad-dition of 0.1±1 mM valinomycin. Measurementswere performed at 258C. Calibration of the ¯uor-escence related to [H+] was carried out usingproteoliposomes suspended in the internal med-ium by the addition of 2 M KOH aliquots in thepresence of 2 mM nigericin and the resultant ¯u-orescence, F, was ®tted according to the modi®edStern±Volmer equation, while considering the ex-perimental ¯uorescence to be composed of thenet ¯uorescence plus background (mostly lightscattering), L. In the resulting equation

�H�� 1=Kq�: �F0 ÿ F �=�Fÿ L�, �1�

M. ZÏaÂcÏkovaÂ et al. / The International Journal of Biochemistry & Cell Biology 32 (2000) 499±508 501

F0 is the unquenched ¯uorescence and Kq is thequenching constant. Parameters Kq and L wereobtained by linear regression of the plot of F vs(F0ÿF )/[KOH]. The measured ¯uorescence traces

were converted using Eq. (1) into ``H+ traces'',from which the rates in mM.sÿ1 were derived.When multiplied by the internal volume of pro-teoliposomes (V ), which was estimated from

Fig. 1. (a) Fatty acid-induced H+ ¯uxes in proteoliposomes with the reconstituted yeast phosphate carrier. Calculated ``H+ traces''

(see Section 2.3) are shown for H+ e�ux (monitored as dH+, i.e. changes in internal [H+] relative to the initial state) from the

proteoliposomes containing yeast PIC, induced by 0.1 mM valinomycin after pre-equilibration with 300 mM lauric acid, in the

absence or presence of inhibitors: +MDPh Ð 5 mM methylenediphosphonate±Tris salt; +IDPh Ð 3.75 mM iminodi(methylene-

phosphonate±Tris salt). The initial rate of H+ e�ux in the control corresponded to H+ ¯ux density of 51.10ÿ6 pmol H+ sÿ1 mmÿ2

(proteoliposome volume was 0.55 ml.mg lipidÿ1) that with a given amount of total protein gave the minimum turnover number of

1.5 sÿ1 for PIC. (b) Protein-independent fatty acid-induced basal H+ ¯uxes in liposomes. Similar ``H+ traces'' are illustrated for

protein-free liposomes of high internal volume (2.88 ml.mg lipidÿ1) when induced by 0.1 mM valinomycin and 300 mM lauric acid in

the absence or presence of the inhibitors (+MDPh,+IDPh). The initial rates re¯ected H+ ¯ux density of 35.10ÿ6, 38.10ÿ6, and36.10ÿ6 pmol H+ sÿ1 mmÿ2, respectively.


SPQ volume distribution [2], and related to theapproximate protein amount, H+ ¯ux ratescould be calculated in nmol H+.(s.mg protein)ÿ1.These rates represent minimum values of turn-over numbers. When the rates expressed in nmolH+.(s.mg lipid)ÿ1 were further divided by a sur-face of 1 mg liposomes (in mm2, calculated fromthe internal volume) H+ ¯ux density per mm2

was obtained.Clÿ uptake into proteoliposomes was measured

using an internal medium composed of 12.8 mMTEA±Pi, 143.5 mM TEA2SO4, 0.6 mM TEA±EDTA, 0.1 mM KCl, pH 7.2. Consequently, theClÿ uniport was monitored as Clÿ quenching ofSPQ after the addition of 25 ml vesicles (1 mglipids) to 2 ml of external medium containing12.8 mM TEA±Pi, 215 mM KCl, 0.6 mM TEA±EDTA, pH 7.2, initiated by the addition of0.1 mM valinomycin in the presence of 25 mMHgCl2. The calibration of ¯uorescence to Clÿ

¯ux and calculation of the ¯ux rates were carriedout in a similar way as that described above forH+ ¯ux. For calibration, aliquots of 2 M KClwere added to proteoliposomes in the presence of1 mM nigericin and 10 mM tributyltin chloride.For calculations, Eq. (1) was used substituting[Clÿ] for [H+]. For measuring Clÿ e�ux an in-ternal medium was used consisting of 150 mM[Clÿ] (composed of 100 mM KCl and 50 mMTris±HCl, pH 7.2), 25 mM TEA2SO4, 0.6 mMTris±EGTA. The internal medium of the vesiclelumen contained 150 mM TEA±TES, 25 mMTEA2SO4, 0.6 mM Tris±EGTA, pH 7.2. Clÿ

e�ux was initiated by 0.1 mM valinomycin in thepresence of 25 mM HgCl2 and quanti®ed asdescribed above.

3. Results

3.1. Fatty acid-induced H+ e�ux inproteoliposomes containing the reconstitutedphosphate carrier

We monitored the transmembrane H+ ¯uxescorrelating to the presence and activity of thereconstituted PIC in proteoliposomes (Fig. 1a).As previously noted, in control liposomes with-out the inserted PIC, acidi®cation of the interiorof the vesicles occurs upon the addition of lauricacid [5]. This phenomenon signals the redistribu-tion of FA molecules across both lea¯ets of thelipid bilayer leading to an H+ release inside [5].This has been termed ¯ip±¯op acidi®cation [3]and also proceeds in the PIC-containing proteoli-posomes (Fig. 1a). After the subsequent additionof valinomycin, an increased H+ e�ux isobserved, similar to that reported in comparableexperiments with reconstituted uncoupling pro-teins [2,6±8] and the ADP/ATP carrier [9]. In thecase of reconstituted uncoupling proteins, wehave provided evidence that this observationre¯ects FA cycling, i.e. FA anions are supposedto be translocated via the carrier and they returnin protonated form, thereby giving rise to trans-membrane H+ transport. The FA-induced H+

e�ux mediated by yeast PIC in the presence of200 or 300 mM lauric acid and 1 mM valinomycin

Fig. 2. Fatty acid dose±responses for H+ e�ux in proteolipo-

somes containing the yeast phosphate carrier and in protein-

free liposomes. Proteoliposomes: (Q), 1 mM valinomycin, (R),

0.1 mM valinomycin; liposomes:(q), 1 mM valinomycin, (r),

0.1 mM valinomycin. H+ ¯ux density re¯ecting the H+ e�ux,

induced by various concentrations of lauric acid, is plotted vs

membrane concentrations of lauric acid that was derived

according to Eq. 1 in Ref. [4]. H+ ¯ux density per square mmwas calculated from the obtained H+ e�ux rates (in mM.sÿ1)and the experimentally determined volume (0.8 ml mg lipidÿ1

in proteoliposomes; 1.3 ml mg lipidÿ1 in liposomes).


amounted typically to 0.4 and 0.6 nmolH+.(s.mg lipid)ÿ1, respectively, which corre-sponds to 106 and 160 nmol H+.(s.mg totalprotein)ÿ1. From this result a minimum turnovernumber for a PIC monomer of 3.6 or 5.4 sÿ1, re-spectively, can be calculated. When 300 mM lau-ric acid and 0.1 mM valinomycin was used, theH+ e�ux rates amounted to 0.16 to 0.3 nmolH+.(s.mg lipid)ÿ1 (Fig. 1a). Palmitic acid inducedsimilar H+ ¯uxes. In order to exactly quantifythe contribution of PIC to this ¯ux, we havecompared the lauric-acid-induced H+ ¯uxes inproteoliposomes containing reconstituted yeastPIC to those observed in liposomes preparedidentically (Fig. 2). Since liposomes withoutreconstituted membrane proteins inevitably havea higher internal volume as compared to proteo-liposomes, (in Fig. 2, 1.3 vs 0.8 ml mg lipidÿ1 inproteoliposomes), the H+ ¯ux density per mm2,JH+, was used for comparison. Furthermore, the

actual lauric acid concentration in the membranewas calculated from the total lauric acid concen-tration, as described in detail in Ref. [4]. At atotal lauric acid concentration exceeding 100 mM,which corresponds to a concentration in themembrane of 143 mM, the protein-mediated H+

e�ux could clearly be discriminated from thebasal H+ e�ux in the protein-free liposomes.This applied to the two di�erent valinomycinconcentrations used, namely 0.1 and 1 mM.Although a relatively high basal protein-indepen-dent H+ ¯ux in liposomes was observed (Fig. 1b,2) which could be caused either by disturbancesin the lipid bilayer or by the presence of ion pair-ing resulting from FA interaction with the vali-nomycin/K+ complex, a signi®cant contributionof PIC to the total observed H+ e�ux in proteo-liposomes is easily recognized.

3.2. Diphosphonate inhibition of the FA-inducedH+ e�ux mediated by the phosphate carrier

The observed FA cycling, mediated by yeastPIC was not inhibited by ATP or other purinenucleotides, nor by mersalyl or N-ethylmaleimide(results not shown). The nucleotides belong to re-

Fig. 4. Clÿ uniport mediated by HgCl2-modi®ed phosphate

carrier Ð Clÿ uptake. Calculated ``Clÿ traces'' (see Section

2.3) are shown for a typical experiment testing the Clÿ uptake

in the proteoliposomes containing yeast PIC modi®ed with

25 mM HgCl2, when induced by 0.1 mM valinomycin (val) in

the absence or presence of inhibitors: +MDPh Ð 7.5 mM

methylenediphosphonate±Tris salt; +IDPh Ð 1.25 mM imi-

nodi(methylenephosphonate)±Tris salt. The trace ``no HgCl2''

illustrates the basal Clÿ permeability. Initial rates of Clÿ

uptake amounted to 3.6 nmol Clÿ(s.mg lipid)ÿ1 in control,

corresponding to the minimum turnover number of 32 sÿ1.

Fig. 3. Dose±response for diphosphonate inhibition of fatty

acid-induced H+ e�ux mediated by the yeast phosphate car-

rier. Relative rates of H+ e�ux, induced by 200 mM lauric

acid and 0.1 mM valinomycin from the proteoliposomes

containing yeast PIC are plotted vs log concentration of in-

hibitors: (*), methylenediphosphonate; (r) iminodi(methyl-

enephosphonate). Solid lines represent the theoretical

dose±responses calculated on the basis of Hill equation, with

parameters of nH 0.8 and 1; and Kis of 4.9 and 8.8 mM,

respectively, derived from the linear regressions of the Hill

plots (relative errors of estimations of regression parameters

were 11%).


agents which would e�ect reconstituted uncou-pling protein UCP1 [2,8]. We found, however,that diphosphonates which were previously foundto inhibit the electroneutral phosphate uptake inrat liver mitochondria [27], were able to decreasethe FA cycling mediated by the reconstitutedyeast PIC (Fig. 1a). Tris-salts of methylenedipho-sphonate and iminodi(methylphosphonate) inhib-ited the PIC-mediated H+ e�ux with anapparent Ki of 4.9 and 8.8 mM, respectively, in atypical experiment (Fig. 3). Unbu�ered dipho-sphonic acids of the same compounds (withoutTris) inhibited with Kis of 0.4 and 0.6 mM, re-spectively. Here the two inhibitory e�ects maysum up, the sole diphosphonate inhibition andthe superimposed inhibitory e�ect caused by theexternal acidi®cation. The latter was found toprevent ¯ip±¯op of neutral FAs from the innerlipid lea¯et to the outer one (ZÏ aÂ cÏ kovaÂ and JezÏ ek,unpublished). The basal protein-independentH+e�ux in liposomes was not inhibited by thesetwo diphosphonates (Fig. 1b). In conclusion,since the two diphosphonates were previouslyfound to speci®cally inhibit the electroneutral Pi

uptake [27], we interpret the observed inhibitionof FA cycling as speci®cally related to theinserted PIC.

Furthermore, we found that the observed FAcycling in proteoliposomes containing the recon-stituted yeast PIC was inhibited by 0.5 mMundecanesulfonate and 2.5 mM hexa¯uoropho-sphate. These two agents inhibit FA-induced H+

¯uxes in proteoliposomes containing reconsti-tuted UCP1, PUMP, and the ADP/ATP carrier(ZÏ aÂ cÏ kovaÂ and JezÏ ek, unpublished). In particular,undecanesulfonate has been considered to inhibitFA cycling speci®cally [2,7]. Other hydrophobicagents such as butylphosphonate, decanepho-sphonate or agaric acid in concentrations of upto 1 mM did not interfere with the FA-inducedH+ ¯uxes mediated by yeast PIC.

Fig. 6. Dose±responses for diphosphonate inhibition of Clÿ

uniport mediated by HgCl2-modi®ed phosphate carrier. Rela-

tive rates of Clÿ uptake (T,*) and Clÿ e�ux (t), induced

by 25 mM HgCl2 in the proteoliposomes containing yeast PIC

are plotted vs log concentration of inhibitors: (*) methylene-

diphosphonate; (T,t) iminodi(methylenephosphonate). Solid

lines represent theoretical dose±responses calculated on the

basis of Hill equation, with parameters of nH 1.3, 0.8 and 0.8;

and Kis 4.2, 0.25, and 0.58 mM, respectively, derived from the

linear regressions of the respective Hill plots (relative errors of

estimations of regression parameters were 12, 4 and 20%, re-

spectively).

Fig. 5. Clÿ uniport mediated by HgCl2-modi®ed phosphate

carrier Ð Clÿ e�ux. Calculated ``Clÿ traces'' (see Section 2.3)

are shown for the Clÿ e�ux from the proteoliposomes con-

taining yeast PIC modi®ed with 25 mM HgCl2, when induced

by 0.1 mM valinomycin (val) in the absence or presence of

+IDPh Ð 2.5 mM iminodi(methylenephosphonate)±Tris salt.

Trace ``no HgCl2'' illustrates the basal Clÿ permeability. In-

itial rate of Clÿ uptake in control amounted to 8.8 nmol

Clÿ(sÿ1. mg total protein)ÿ1, corresponding to the minimum

turnover number of 80 sÿ1.


3.3. Clÿ uniport mediated by the phosphate carrierafter functional conversion

Previously it has been found that several mito-chondrial anion carriers can be converted to anunspeci®c uniport mode by chemical modi®cationof cysteine residues [23,25]. Since the FA cyclingmechanism in principle requires uniport of FAsacross the membrane, we investigated a possiblerelationship between a HgCl2-induced unspeci®cClÿ uniport via PIC and the FA interaction withthis carrier. When we assayed proteoliposomescontaining reconstituted yeast PIC for Clÿ

uptake, a slow Clÿ uniport in the range of thebasal liposomal Clÿ permeability was observedÐ 0.002 nmol Clÿ.(s.mg lipid)ÿ1 in the presenceof 0.1 mM valinomycin and 0.16 nmol Clÿ.(s.mglipid)ÿ1 with 1 mM valinomycin (Fig. 4). Ad-dition of 25 mM HgCl2 to the PIC-containingproteoliposomes, however, induced fast Clÿ uni-port Ð 4 nmol Clÿ.(s.mg lipid)ÿ1 at 0.1 mM vali-nomycin and 5.2 nmol Clÿ.(s.mg lipid)ÿ1 at 1 mMvalinomycin. With a given protein amountit accounted for a minimum turnover of 36and 46 sÿ1, respectively. This activity could beinhibited by the addition of mersalyl and bymethylenediphosphonate or iminodi(methylene-phosphonate) (Fig. 4). As a control, we appliedreverse K+ and Clÿ gradients. In fact, the sameHgCl2-induced Clÿ uniport via PIC could beobserved, in this case, however, as a Clÿ e�uxfrom proteoliposomes (Fig. 5). These Clÿ e�uxrates typically reached 7±8 nmol Clÿ.(s.mglipid)ÿ1 at 0.1 mM valinomycin and 25 mMHgCl2. Iminodi(methylenephosphonate) wasactive as an inhibitor of the Clÿ¯ux under theseconditions, whereas methylenediphosphonate wasnot e�ective. Phosphonoformic acid (3 mM) onlyslightly inhibited the Clÿ e�ux. The apparent Ki

for methylenediphosphonate inhibition of the Clÿ

uptake was 4.2 mM (Fig. 6). Iminodi(methylene-phosphonate) exhibited a lower Ki in the case ofthe Clÿ uptake (cis-inhibition) than for Clÿ e�ux(trans-inhibition), amounting to 250 and 580 mM,respectively, in a typical experiment (Fig. 6).Interestingly, mersalyl prevented initiation ofboth Clÿ uptake and Clÿ e�ux mediated byHgCl2-modi®ed PIC. This re¯ects the partici-

pation of SH groups in the induction of the Clÿ

uniport mode, as mersalyl competes with HgCl2on these SH groups.

4. Discussion

Several carriers belonging to the mitochondrialgene family have been found, at least in vitro, tomediate FA cycling, monitored as FA-inducedH+ ¯uxes in proteoliposomes [3]. FA cycling cri-tically depends on the presence of appropriatecarrier proteins, as proved by the slower basicH+ ¯uxes observed in liposomes alone in thepresence of FAs up to 0.5 mM concentrations,which correlates to nearly molar in the mem-brane (see Eq. 1 in [4]). In this work we addedone more carrier, PIC, to the list of transportproteins which are able to catalyze FA cycling inproteoliposomes. Further studies, however, arerequired to decide whether FA cycling mechan-isms actually play a role in vivo. The presentresults indicate that PIC seems to be less sensitiveto FAs than UCP1, based on the preliminarykinetic data for PIC (Fig. 2) and the reported Km

for lauric acid of 8 mM for UCP1 [2].It should be mentioned here that it is more

appropriate to express these constants as mem-brane FA concentrations. Indeed, FAs in themitochondrial inner membrane can reach milli-molar concentrations in the presence of a fewmM of free FAs in the surrounding water phase[4].

The fact that PIC mediates FA cycling is basedon its ability to mediate uniport of FA anions.We have demonstrated that the uniport activityof PIC with respect to other anions, e.g. Clÿ, isnearly zero, unless anion uniport is induced byHgCl2 modi®cation of PIC. Interestingly, thesame inhibitors, methylenediphosphonate andiminodi(methylenephosphonate) a�ect both FAanion uniport and HgCl2-induced anion uniport.These diphosphonates did not inhibit the HgCl2-induced anion uniport activity of other mito-chondrial carrier proteins (ZÏ aÂ cÏ kovaÂ and JezÏ ek,unpublished). Consequently, we suggest thatdiphosphonates act near the Pi binding site,whereas FAs interact with a domain which is


possibly common to all carriers that mediate FAcycling. This suggestion is in line with our in-terpretation that the phosphate translocationpathway (Pi binding site) is not identical to theFA binding site. On the contrary, we considerthe inhibition by hexa¯uorophosphate and unde-canesulfonate to result from some competition ina hydrophobic domain of the membrane, as suchit must proceed in all transport proteins testedup to date.

It is not clear whether FA cycling evolvedduring phylogenesis or whether it is an accidentalside e�ect of these transporters. It should be keptin mind that there is obviously another mechan-ism of H+ translocation inherent to the functionof this carrier, i.e. H+ cotransport (or OHÿ anti-port), concomitant with the Pi translocation,which is actually the physiological activity ofPIC. This catalytic pathway mediating phosphatetranslocation must be distinct from the oneenabling FA cycling, otherwise, PIC would act asa phosphate/FA antiporter. In this case, phos-phate transport had to be activated by FAs, how-ever, the opposite result was found to be true,i.e. inhibition of phosphate transport by FAs[27]. Nevertheless, FA cycling could represent aminor function of PIC, independent of the phys-iological phosphate transport. A slight uncou-pling which might occur in vivo due to FAcycling via PIC accelerates respiration, which isbene®cial for metabolism. Since a series of car-riers, regulated either by ligands or transcription-ally, is involved in FA cycling, namelyuncoupling proteins [4,8], the ADP/ATP carrier[1,9,10], glutamate/aspartate carrier [15], dicar-boxylate carrier [14], and now the phosphate car-rier, the phenomenon of FA-induced uncouplingin mitochondria needs further investigation toreveal all possible physiological functions.

Acknowledgements

Excellent technical assistance of Jana Bruck-nerovaÂ and Jana KosÏ arÏ ovaÂ is gratefully acknowl-edged. The project was supported by grants ofthe program Kontakt for Czech±German mutualcooperation from the Czech Ministry of Edu-

cation (ME 085) and by the Grant Agency of theCzech Republic, Grant No. 301/98/0568. Fluo-rometer was purchased from the funds of theCzech±U.S. Science and Technology Program,grant No. 86043.

References

[1] V.P. Skulachev, Fatty acid circuit as a physiological

mechanism of uncoupling of oxidative phosphorylation,

FEBS Lett. 294 (1991) 158±162.

[2] K.D. Garlid, D.E. Orosz, M. ModrianskyÂ , S. Vassanelli,

P. JezÏ ek, On the mechanism of fatty acid-induced pro-

ton transport by mitochondrial uncoupling protein, J.

Biol. Chem. 271 (1996) 2615±2620.

[3] P. JezÏ ek, H. EngstovaÂ , M. ZÏ aÂ cÏ kovaÂ , A.E. Vercesi,

A.D.T. Costa, P. Arruda, K.D. Garlid, Fatty acid

cycling mechanism and mitochondrial uncoupling pro-

teins, Biochim. Biophys. Acta 1365 (1998) 319±327.

[4] K.D. Garlid, M. JabuÊ rek, P. JezÏ ek, The mechanism of

proton transport mediated by mitochondrial uncoupling

proteins, FEBS Lett. 438 (1998) 10±14.

[5] P. JezÏ ek, M. ModrianskyÂ , K.D. Garlid, Inactive fatty

acids are unable to ¯ip±¯op across the lipid bilayer,

FEBS Lett. 408 (1997) 161±165.

[6] P. JezÏ ek, M. ModrianskyÂ , K.D. Garlid, A structure±ac-

tivity study of fatty acid interaction with mitochondrial

uncoupling protein, FEBS Lett. 408 (1997) 166±170.

[7] P. JezÏ ek, A.D.T. Costa, A.E. Vercesi, Reconstituted

plant uncoupling mitochondrial protein allows for pro-

ton translocation via fatty acid cycling mechanism, J.

Biol. Chem. 272 (1997) 24272±24278.

[8] M. JabuÊ rek, M. VarÏ echa, R.E. Gimeno, M. Dembski,

P. JezÏ ek, M. Zhang, P. Burn, L.A. Tartaglia, K.D.

Garlid, Transport function and regulation of mitochon-

drial uncoupling proteins 2 and 3, J. Biol. Chem. 274

(1999) 26003±26007.

[9] N. Brustovetski, M. Klingenberg, The reconstituted

ADP/ATP carrier can mediate H+ transport by free

fatty acids, which is further stimulated by mersalyl, J.

Biol. Chem. 269 (1994) 27329±27336.

[10] P. SchoÈ nfeld, P. JezÏ ek, J. BoreckyÂ , E. Belyaeva, V.S.

Slyshenkov, M.R. Wiec° kowski, L. Wojtczak,

Photomodi®cation of mitochondrial proteins by azido

fatty acids and its e�ect on mitochondrial energetics.

Further evidence for the role of the ADP/ATP carrier in

fatty-acid-mediated uncoupling, Eur. J. Biochem. 240

(1996) 387±393.

[11] P. PolcÏ ic, L. SÏ abovaÂ , J. Kolarov, Fatty acids induced

uncoupling of Saccharomyces cerevisiae mitochondria

requires an intact ADP/ATP carrier, FEBS Lett. 412

(1997) 207±210.

[12] J.E. Walker, M.J. Runswick, The mitochondrial trans-

port protein superfamily, J. Bioenerg. Biomembr. 25

(1993) 435±445.


[13] B. El Moualij, C. Duyckaerts, J. Lamotte-Brasseur, F.E.

Sluse, Phylogenetic classi®cation of the mitochondrial

carrier family of Saccharomyces cerevisiae, Yeast 13

(1997) 573±581.

[14] M.R. Wieckowski, L. Wojtczak, Involvement of the

dicarboxylate carrier in the protonophoric action of

long-chain fatty acids in mitochondria, Biochem.

Biophys. Res. Comm. 232 (1997) 414±417.

[15] V.N. Samartsev, E.N. Mokhova, V.P. Skulachev, The

pH-dependent reciprocal changes in contribution of

ADP/ATP antiporter and aspartate/glutamate antiporter

to the fatty acid-induced uncoupling, FEBS Lett. 412

(1997) 179±182.

[16] H. Wohlrab, Molecular aspects of inorganic phosphate

transport in mitochondria, Biochim. Biophys. Acta 853

(1986) 115±134.

[17] R. KraÈ mer, F. Palmieri, Molecular aspects of isolated

and reconstituted carrier proteins from animal mito-

chondria, Biochim. Biophys. Acta 974 (1989) 1±23.

[18] G.C. Ferreira, P.L. Pedersen, Phosphate transport in

mitochondria: past accomplishments, present problems,

and future challenges, J. Bioenerg. Biomembr. 25 (1993)

483±491.

[19] F. Palmieri, F. Bisaccia, L. Capobianco, V. Dolce, G.

Fiermonte, V. Iacobazi, V. Zara, Transmembrane top-

ology, genes, and biogenesis of the mitochondrial phos-

phate and oxoglutarate carriers, J. Bioenerg. Biomembr.

25 (1993) 493±501.

[20] R. KraÈ mer, Functional principles of solute transport

systems: concept and perspectives, Biochim. Biophys.

Acta 1185 (1994) 1±34.

[21] R. KraÈ mer, Structural and functional aspects of the

phosphate carrier from mitochondria, Kidney Int. 49

(1996) 947±952.

[22] A. Phelps, C. Briggs, L. Mincone, H. Wohlrab,

Mitochondrial phosphate transport protein.

Replacements of glutamic, aspartic, and histidine resi-

dues a�ect transport and protein conformation and

point to a coupled proton transport path, Biochemistry

35 (1996) 19757±19762.

[23] R. Stappen, R. KraÈ mer, Functional properties of the

reconstituted phosphate carrier from bovine heart mito-

chondria: evidence for asymetric orientation and charac-

terization of three di�erent transport modes, Biochim.

Biophys. Acta 1149 (1993) 40±48.

[24] R. Stappen, R. KraÈ mer, Kinetic mechanism of phos-

phate/phosphate and phosphate/OHÿ antiports cata-

lyzed by reconstituted phosphate carrier from beef heart

mitochondria, J. Biol. Chem. 269 (1994) 11240±11246.

[25] K. Herick, R. Stappen, R. KraÈ mer, Comparison of

functional and structural aspects of the reconstituted

phosphate and aspartate/glutamate carrier from mito-

chondria, in: F. Palmieri (Ed.), Thirty Years of Progress

in Mitochondrial Bioenergetics, Elsevier, Amsterdam

1995, pp. 83±87.

[26] H. Freitag, B. Kadenbach, Transport of phosphate ana-

logues in rat liver mitochondria, Eur. J. Biochem. 83

(1978) 53±57.

[27] H. EngstovaÂ , M. RuÊ zÏ icÏ ka, P. JezÏ ek, Interaction of mito-

chondrial phosphate carrier with fatty acids, EBEC

Rep. 10 (1998) 28.

[28] A. Schroers, R. KraÈ mer, H. Wohlrab, The reversible

antiport-uniport conversion of the phosphate carrier

from yeast mitochondria depends on the presence of a

single cysteine, J. Biol. Chem. 272 (1997) 10558±10564.

[29] T. Dierks, R. Stappen, R. KraÈ mer, Channel properties

of mitochondrial carriers, in: M. Forte, M. Colombini

(Eds.), Molecular Biology of Mitochondrial Transport

Systems, Vol. II 83, Springer, Berlin, Heidelberg, 1994,

pp. 117±129 NATO ASI Series.

[30] H. Wohlrab, C. Brigss, Yeast mitochondrial phosphate

transport protein expressed in Escherichia coli. Site-di-

rected mutations at Threonine-43 and at a similar lo-

cation in the second tandem repeat (Isoleucine-141),

Biochemistry 33 (1994) 9371±9375.


Documents

Interaction of mitochondrial phosphate carrier with fatty acids and hydrophobic phosphate analogs