THE JOURNAL OF B I O ~ I C A L CHEMISTRY 0 1994 by The American Society for Biochemistry and Molecular Biology, Inc

Vol. 269, No. 9, Issue of March 4, pp. 6437-6443, 1994 Printed in V.S.A.

The Use of Chemical Cross-linking to Identify Proteins That Interact with a Mitochondrial Presequence*

(Received for publication, June 21, 1993, and in revised form, October 21, 1993)

Amos S. Gaikwad and Michael G. CumskyS From the Department of Molecular Biology and Biochemistry, University of California, Irvine, California 92717

Previous work has shown that when yeast mitochon- dria are incubated in the presence of the presequence peptide pL4(1-22), the peptide is imported and accumu- lates within the mitochondrial membranes, presumably at the import sites. If the extramitochondrial concentra- tion of peptide is sufficiently high, enough peptide ac- cumulates within the import sites to prevent the uptake of authentic precursor proteins. We have used chemical cross-linking to probe the interaction of this peptide with yeast mitochondrial proteins. We found that radio- labeled pL4(1-22) could be reproducibly cross-linked to a number of polypeptides. Interestingly, nearly all were membrane proteins. Several of the cross-linked proteins were located in the outer membrane, while others were located in the inner membrane. The interaction between the peptide and many of the cross-linked products was shown to be specific by two independent criteria. First, an excess of unlabeled peptide acted as a competitor in the cross-linking reaction, and, second, treatment of the peptide with the alkylating agent N-ethylmaleimide dra- matically reduced its ability to form cross-links. Two of the cross-linked species corresponded to the outer mem- brane proteins, Mas7Op and ISP42. Significantly, both of these proteins have previously been shown to play criti- cal roles in mitochondrial protein import. While the role of the other cross-linked proteins in the import process remains to be determined, the results of this study dem- onstrate that our experimental approach may be useful in identifying components of the import machinery as well as proteins that interact with mitochondrial prese- quences.

Most mitochondrial proteins are encoded on nuclear DNA and translated in the cytosol. They must then be targeted to mitochondria, cross one or both mitochondrial membranes, and be correctly sorted into one of the four mitochondrial compart- ments. Despite considerable experimental effort, the molecular mechanism(s) underlying the import of proteins into mitochon- dria are only beginning to be understood. A detailed knowledge of this process is fundamental to our overall understanding of both mitochondrial biogenesis and protein transport into and across all biological membranes.

In the vast majority of cases, nuclear encoded mitochondrial proteins are initially synthesized with an NH2-terminal exten- sion called a presequence or leader peptide. Over the last dec- ade, the mitochondrial presequence and its role in protein im-

and National Institutes of Health Grant GM 36675. The costs of pub- * This work was supported by American Cancer Society Grant BE 120

lication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertzsement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

$ An Established Investigator of the American Heart Association. To whom correspondence should be addressed. Tel.: 714-856-7766; Fax: 714-856-8551.

port have been studied extensively. While presequences share no sequence homology, several structural features appear to be conserved. They (almost always) contain a high percentage of basic and hydroxylated amino acids, are devoid of acidic resi- dues, and appear to form amphiphillic structures (Roise and Schatz, 1988). Mitochondrial presequences are also known to mediate several of the early events in protein import. After the initial recognition and binding of the precursor protein by pro- teinaceous components at the mitochondrial surface, the trans- location process begins by movement of the presequence into the matrix of the organelle (reviewed recently in Pfanner and Neupert (1990) and Glick and Schatz (1991)). This process, which requires the electrical component of the proton motive force (A+), is driven by the electrophoretic attraction of the positively charged presequence to the net negatively charged matrix (Martin et al . , 1991). Translocation occurs at sites of contact between the two mitochondrial membranes (Pfanner et al., 1988; Pon et al . , 19891, which appear to be composed of distinct translocation pores or channels within the individual bilayers (Glick et al., 1991; Pfanner et al . , 19921. At present, the outer membrane translocation site has been partially charac- terized; it consists of at least two receptor-like proteins and several other membrane proteins that presumably form a chan- nel within the bilayer (Kiebler et al., 1990; Sollner et al . , 1992). On the other hand, very little is known about the inner mem- brane translocation site. While the nature of this site is actively being investigated in several laboratories, only a few proteins that are essential for transport across the inner membrane have thus far been identified (Maarse et al . , 1992; Scherer et al . , 1992; Emtage and Jensen, 1993).

It has been known for many years that the presequence targets proteins to mitochondria; presequences have been shown to be both necessary for mitochondrial delivery and suf- ficient to direct heterologous, non-mitochondrial proteins to the organelle (Hurt et al., 1984, 1985; Honvich et al., 1985). Yet, despite considerable experimental effort, it has still not become clear precisely how, on a molecular level, the information in the presequence is decoded. Furthermore, there is currently little evidence suggesting a direct interaction between the prese- quence and any mitochondrial protein, excepting only the proc- essing peptidase located in the matrix (Hawlitschek et al . , 1988; Jensen and Yaffe, 1988).

Do mitochondrial presequences interact with proteins of the translocation machineries within the mitochondrial mem- branes? In the present study, we have addressed this question by taking advantage of the unique properties of a synthetic presequence peptide, pL4(1-22) (Glaser and Cumsky, 1990a, 1990b). Using chemical cross-linking methods, we have identi- fied several membrane proteins that specifically interact with this presequence. Two of the proteins, Mas70p and ISP42, are of particular note, because they correspond to previously iden- tified components of the outer membrane translocation site. Several others, which are located within either the inner or the outer mitochondrial membrane, may represent previously uni-

6437

6438 Proteins That Interact with Mitochondrial Presequence dentified proteins that comprise part of the import machinery a n d o r facilitate the import process.

EXPERIMENTAL PROCEDURES Synthetic Peptides-The synthesis, deprotection, and purification of

the synthetic presequence peptide pU(1-22) has been described previ- ously (Glaser and Cumsky, 1990a). Each lot of peptide was tested for purity and biological activity by assaying its ability to block mitochon- drial protein import over the concentration range of 5-20 w.

Because covalent modification of any type has been shown to alter the biological activity of pU(1-22) or pU(1-23) (which contains tyrosine as the COOH-terminal residue) (Glaser and Cumsky, 1990a, 1990b), a radiolabeled derivative for use in cross-linking studies was synthesized by the following procedure. First, Fm~c-[~~SlMet was synthesized in a reaction that contained 0.49 M Fmoc-C1 ( i n acetone), 0.38 M [35S]Met, 0.38 M Na2C08. The reaction was carried out a t room temperature with constant stirring for 12 h. The acetone was then evaporated, the pH value was adjusted to 6.5, and the mixture was extracted several times with ethyl acetate to separate Fm~c-[~~SlMet from the unreacted C3%1Met. The ethyl acetate layers (which contained the Fmo~-[~~SlMet) were pooled, washed four to five times with HzO, and dried over anhy- drous MgSO,. A small aliquot was counted for radioactivity to deter- mine the degree of incorporation of [35SlMet into Fm~c-[~~SlMet. The purity of the product was also confirmed by thin layer chromatography.

The Fm~c-[~~SlMet was coupled to the partially synthesized peptide (residues 2-22) in a single cycle using standard synthesis conditions (Atherton et al., 1981; Dryland and Sheppard, 1986). The labeled pep- tide was then deprotected by incubation for 8 h at room temperature in a mixture of 90% trifluoroacetic acid, 5% thioanisol, 3% ethanediol, 2% anisole (reagent R). It was then cleaved from the resin for 2 h at room temperature in a 20% solution of methanolic ammonia. The free depro- tected peptide was dissolved in reagent R, precipitated by the addition of 10 volumes of anhydrous ethyl ether, pelleted, washed four additional times with ether, and lyophilized. The dried peptide was dissolved in 0.1% trifluoroacetic acid and purified by reverse phase chromatography on a Sep-Pak C18 column (Waters Associates). The column was devel- oped using a 20-35% acetonitrile gradient. The purified peptide, which eluted at 35% acetonitrile, was then concentrated in a speed vac. The purity of the peptide was determined by SDS-polyacrylamide gel elec- trophoresis (PAGE)' in the presence of 6 M urea (12.5% polyacrylamide gel). The biological activity of the peptide was confirmed by its ability to block mitochondrial protein import (see "Results").

In Vitro Import Reactions-Procedures for the in vitro transcription and translation of mitochondrial precursor proteins, preparation of yeast mitochondria, and in vitro import into yeast mitochondria have been described previously (Glaser et al., 1990; Miller and Cumsky, 1991, 1993). Peptide inhibition of mitochondrial import was carried out as described in our earlier study (Glaser and Cumsky, 1990a).

Cross-linking of pU(1-22) to Mitochondria-Cross-linking of pL4(1- 22) was accomplished using a modification of the standard import assay. Radiolabeled pU(1-22) (approximately 5-20 w) was incubated with 40-50 pg of mitochondria in 40 pl of SH buffer, which contained 250 m sucrose, 2 m HEPES-KOH (pH 7.2), 1 m MgC12, 1 ~ M A T P , and 10 m K+-malate/7.5 rm K+-succinate. m e r 15-30 min at room temperature, the membrane-permeable homobifunctional cross-linker ethylene gly- colbis(succinimidy1succinate) (EGS) was added to a final concentration of 1 m. The cross-linking reaction was allowed to proceed for 10 min at room temperature, terminated by the addition of glycine to 20 m, and the mitochondria were reisolated by centrifugation.

In certain cases, reactions were chased with cold pL4(1-22) prior to cross-linking. To accomplish the chase, cold pL4(1-22) (40-200 w, final concentration) was added to the reaction after the initial 15-30-min incubation with radiolabeled pL4(1-22) (see above). After an additional 30 min, cross-linking was performed as described in the preceding para-

Fractionation of Mitochondria-Fractionation of cross-linked mito- chondria into total soluble and membrane fractions was accomplished essentially as described in our earlier work (Glaser and Cumsky, 1990b). Mitochondria (0.25 mg) were cross-linked to radiolabeled pU(1- 22) as described above. After terminating the reaction with glycine, the mitochondria were reisolated by centrifugation and resuspended in 50 p1 of SEM buffer (250 mM sucrose, 1 rm EDTA, 10 m MOPS). They

graph.

The abbreviations used are: PAGE, polyacrylamide gel electropho- resis; Dm, dithiothreitol; EGS, ethylene glycolbis(succinimidy1succi- nate); MOPS, 3-(N-morpholino)propanesulfonic acid; NEM, N-ethylma- leimide.

were then diluted with 9 volumes of ME buffer (10 m MOPS, pH 7.2, 1 m EDTA) containing 1 m~ phenylmethylsulfonyl fluoride, 5 pg/ml leupeptin, and 10 pg/ml pepstatin. The suspension was swollen for 20 min on ice and then sonicated four times for 20 s using a microtip and an output setting of two-thirds the maximum (the sample was cooled on ice for 30-40 s between rounds of sonication). The mixture was centri- fuged at 15,000 rpm for 10 min (4 "C) to remove unbroken mitochondria, and the supernatant was collected. It was then centrifuged for an ad- ditional hour at 60,000 rpm in TLA 100.1 rotor (4 "C) to separate the total membranes from the soluble fraction.

Fractionation of total membranes into inner and outer membrane vesicles was accomplished essentially as described by Hwang et al. (1989) except that the amount of starting material was generally 40-50 mg of mitochondria.

Fractions from the linear sucrose gradient (0.85-1.6 M sucrose) were collected and analyzed by immunoblotting to confirm the purity of each membrane preparation. The representative marker proteins used were either OMP45 or ISP42 for the outer membrane and subunit Va of cytochrome c oxidase for the inner membrane. Inner membrane vesicles prepared by this procedure were routinely found to be fully competant to import and process precursor proteins to a protease-protected form.

NEM neatment of Peptides-N-ethylmaleimide (NEM) treatment of pLl(1-22) was performed essentially as described by Murakami et al. (1988) with the following modifications. The peptide (either radiola- beled or cold) was first incubated with dithiothreitol ( D m ) (40-50 p t ) for 10 min at room temperature. Freshly prepared NEM was then added to a final concentration of 20 m, and the mixture was incubated for 30 min at 37 "C. In certain cases, a second aliquot of NEM was added and the incubation continued for an additional 30 min. The reaction was then stopped by adding DIT to a final concentration of 150 m. Cross- linking of mitochondria with NEM-treated peptide was performed as described for the native peptide.

Miscellaneous Methods-Procedures for protein determinations, SDS-PAGE, immune blotting, immunoprecipitation, and fluorography have been described previously (Glaser and Cumsky, 1990a; Glaser et al., 1990; Miller and Cumsky, 1991, 1993).

RESULTS

Cross-linking of the Presequence Peptide pU(1-22) to Iso- lated Yeast Mitochondria-pL4(1-22) is a synthetic peptide cor- responding to the NHz-terminal 22 residues of the presequence to subunit IV of yeast cytochrome c oxidase. Our earlier results demonstrated that pL4(l-22) was taken up by mitochondria and that it accumulated within the mitochondrial membranes (Glaser and Cumsky, 1990a, 1990b). In turn, the high intrami- tochondrial concentration of pL4(1-22) prevented the uptake of authentic precursor proteins; it acted as a reversible inhibitor of mitochondrial protein import. Thus, we reasoned that chemi- cal cross-linking of accumulated intramitochondrial peptide would be a viable means by which to examine the interaction between mitochondrial presequences and the import machin- ery. Because covalent modification of any type has been shown to destroy the biological activity of pL4(1-22) (Glaser and Cum- sky, 1990a, 1990b), we synthesized a radiolabeled form of the peptide that contained [35S]Met at the NH2 terminus for use in cross-linking experiments (see "Experimental Procedures"). This peptide behaved identically to the unlabeled derivative in that it accumulated within the mitochondrial membranes (not shown) and reversibly blocked protein import at a concentra- tion of 10-20 p 4 . As shown in Fig. 1, the labeled peptide was able to block import of the precursors to both subunit IV of cytochrome c oxidase (pIV) and the p subunit of the F1-ATPase (pF1p). It was, therefore, used in all subsequent cross-linking experiments. To identify proteins that interact with pL4(1-22), we used

EGS, a homobifunctional cross-linker that is membrane-per- meable but can cross-link soluble proteins as well. Mitochon- dria were allowed to accumulate the labeled peptide for 15-30 min at room temperature, were incubated with the cross-linker as described under "Experimental Procedures," and were then analyzed by SDS-PAGE. As shown in Fig. 2A, both major and minor cross-linked products were generated. Importantly,

Proteins That Interact with Mitochondrial Presequence 6439

V a l - - + + - - Peptide - - - - + +

Prot. K - + - f - +

plV - I V -

m 1

I

Peptide - - -I- + Pr0t .K. - + - +

FIG. 1. Radiolabeled pL4(1-22) blocks protein import into yeast mitochondria. Isolated mitochondria (50 pg) in 100 pl ofTRB (250 mM sucrose, 80 mM KCI, 5 mM MgC12, 20 mM MOPS, pH 7.2, 3% bovine serum albumin, 1 mM ATP, and 10 mM K+-malate/7.5 mM K+-succinate) were preincubated on ice in the presence of approximately 20 PM radio- labeled pU(1-22). After 5 min, 1.5 x IO4 cpm :35S-labeled precursor (either F,P or subunit IV of cytochrome c oxidase, pF,P and pW, respec- tively) was added to each tube. The reactions were allowed to proceed for 15 min a t 30 “C and were then stopped by the addition of 10 pg/ml valinomycin (Val). 50-pl aliquots from each reaction were removed, and the mitochondria were reisolated by centrifugation a t 4 “C and pro- cessed for SDS-PAGE. The remainder of each reaction (50 pl) was di- gested with proteinase K (Prot. K) (250 pg/ml) for 30 min on ice. The reactions were then stopped by the addition of phenylmethylsulfonyl fluoride to 1 mM, and the mitochondria were collected by centrifugation and processed for SDS-PAGE. Import reactions were analyzed by SDS- PAGE and fluorography; reactions involving F,P were resolved on 7.5% gels, those involving subunit IV on 15% gels. Where indicated for pF,P, valinomycin was included in the import reaction mixture a t 10 pg/ml. The positions of the precursor and mature forms of each protein are indicated (pFIP and F,P, p N and IV, respectively). Only one portion of the fluorographed gel is shown; the front of each lune contained a large radiolabeled band corresponding to peptide that had been taken up by the mitochondria.

A I 2 3 4 B 5

F“ “

,98K - 98K .u

- 68K

P -43K .68K

- 43K

EGS - + + + T M comp.peptide ( J J M ) 0 0 100 200

FIG. 2. Cross-linking of pL4(1-22) to mitochondrial proteins. A, cross-linking or cross-linking/competition reactions containing 10 p~ radiolabeled pL4(1-22), 40 pg of mitochondria, and 1 mM final concen- tration EGS (except lune 1 where the cross-linking reagent was omitted)

“Experimental Procedures.” Reactions were terminated by the addition were carried out for 10 min a t room temperature as described under

of glycine to 20 mM, and the mitochondria were reisolated by centrifu- gation. They were then analyzed by SDS-PAGE and fluorography on 10% gels. The position of representative marker proteins is indicated at the right; the arrows at the left indicate the positions of the major cross-linked products that were significantly (or completely) competable by excess nonlabeled peptide. B, Coomassie Blue staining pattern of 50 pg of the total mitochondrial membrane proteins analyzed by SDS- PAGE on a 10% gel. The position of representative marker proteins is indicated a t the right; TM, total membranes.

while the relative intensity of each cross-linked species varied slightly from experiment to experiment, the profile of cross- linked products was highly reproducible and distinct from the profile of mitochondrial membrane proteins stained with Coo- massie Blue (Fig. 2B). Several of the major cross-linked prod- ucts, including those with apparent electrophoretic mobilities

of approximately 19, 25, 30, 34, 45, 60, 65, and 75 kDa, were found to be competable by a 5- or 10-fold molar excess of unla- beled peptide (Fig. 2 A , lanes 3 and 4) . The same concentration of a nonspecific competitor, a basic 22-residue peptide referred to as pATIII (Glaser and Cumsky, 1990a), failed to compete in the cross-linking reaction (not shown). Thus, the interaction between pL4(1-22) and a number of yeast mitochondrial pro- teins appears to be specific.

Inactive pLA(1-22) Will Not Form Cross-links with Yeast Mi- tochondrial Proteins-If the interaction between pL4(1-22) and the major cross-linked products was indeed specific, we pre- dicted that a biologically inactive form of the peptide (i.e. one that did not block the import of mitochondrial precursor pro- teins) would be a poor substrate for cross-linking. To inactivate the peptide, it was pretreated with NEM, which alkylates the cysteine residue a t position 19 (Glaser and Cumsky, 1990a). As shown in Fig. 3, the peptide was clearly inactivated by this procedure; the NEM-treated derivatives (both the labeled and nonlabeled forms) lost the ability to block mitochondrial import of the precursor to subunit Va of yeast cytochrome c oxidase (pVa; Fig. 3, lanes 5-12). The loss of import inhibition was not due to the presence of NEM or DTT in the reaction mixture, since, when present at the same concentration as in the reac- tions of lanes 5-12, neither reagent had an apparent effect on the import efficiency of pVa (Fig. 3, lanes 13 and 14).

Fig. 4 shows that when radiolabeled pL4(1-22) was pre- treated with NEM prior to incubation with mitochondria, i t lost the ability to form cross-links (i.e. interact) with essentially all of the proteins identified in Fig. 2 (Fig. 4, lane 2 ) . Again, the inability to form cross-links was not due to excess NEM or DTT in the reaction mixture, since the pattern of cross-linked bands was nearly identical to that of Fig. 2 when 1 mM DT’I-inacti- vated NEM was included in the reaction (Fig. 4, lane 4 ) . Thus, the results presented in Figs. 3 and 4 are in line with those of Fig. 2 and strongly suggest that pL4(1-22) interacts specifically with a number of mitochondrial proteins.

Cross-linked Products Reside in Both the Inner and Outer Mitochondrial Membranes-When mitochondria that had been treated with peptide and subsequently cross-linked were sepa- rated into soluble and membrane fractions, we found that the majority of the cross-linked products were membrane proteins (Fig. 5A). This result was consistent with our earlier observa- tion that the vast majority of the imported peptide resided within the membranes (Glaser and Cumsky, 1990a, 1990b). To determine whether the products resided in the inner or outer membrane, total mitochondrial membranes were further sepa- rated by centrifugation on a sucrose density gradient. Fractions were then collected and analyzed by immunoblotting (to deter- mine the presence of representative marker proteins) and by cross-linking to pL4(1-22). The results of the immunoblot anal- ysis confirmed that this procedure resulted in a typical sepa- ration of the two membranes (Fig. 5C). The results of the cross- linking experiment are shown in Fig. 5B. Nearly all of the proteins that were strongly cross-linked to the peptide in intact mitochondria could also be cross-linked with the isolated mem- branes. We also found that these proteins did not reside exclu- sively in one of the two membranes; while a majority of the strongly cross-linked proteins that were competable with ex- cess cold peptide (Fig. 2) fractionated with the outer mem- brane, a t least one of this group clearly resided within the inner membrane (Fig. 5B 1. Moreover, recent results have shown that several of the less strongly cross-linked proteins, which are not clearly resolved in Figs. 2 and 5, are also inner membrane proteins (see “Discussion”).2

In addition to providing evidence that cross-linked products

A. S. Gaikwad, unpublished results.

6440 Proteins That Interact with Mitochondrial Presequence

1 2 3 4 5 6 7 8 9 10 11 12 13 14 PVOL

Var - -

Va I + + " "

Prot. K - + - + - + - + - + - f - + """"

Peptide - - - - + + + + + + + + - - ""

PL4 N-pL4 *pL4 N-*pL4 N t DTT FIG. 3. NEM-inactivated pCd(1-22) does not block protein import into yeast mitochondria Import reactions using 50 pg of mitochondria

and 1.5 x lo4 radiolabeled precursor (subunit Va of yeast cytochrome c oxidase, pVa) were performed as described in the legend to Fig. 1. Where indicated, the mitochondria were preincubated with various forms of the peptide pU(1-22); pL4 and N-pL4 denote the native and NEM-treated forms of the peptide, respectively, while *pL4 and N - * p U denote the radiolabeled native and radiolabeled NEM-treated forms of the peptide, respectively. The positions of the precursor (pVa) and mature forms (Vu) of subunit Va are indicated. Shown in lanes 13 and 14 are the results of

All import reactions were analyzed by SDS-PAGE and fluorography on 15% gels. Addition of valinomycin or proteinase K treatment of the reaction a control import reaction performed in the presence of 2 m~ NEM (inactivated with D m , which was also present a t a final concentration of 40 mM).

mixture is also indicated.

1 2 3 4 "

- 6 8 K " - -43K - -

- 29K

EGS + + + -I-

pL4 + + + + - N

cornp.peptide - - +: N-DTT

FIG. 4. NEM-inactivated pIA(1-22) is not efficiently cross- linked to mitochondrial proteins. Mitochondria were pretreated with either radiolabeled pL4(1-22) (reactions of lanes 1, 3, and 4 ) or NEM-inactivated pL4(1-22) (lane 2) as described in the legend to Fig. 2. The reaction shown in lune 3 was chased with 100 PM nonlabeled pep- tide for 30 min prior to adding the cross-linker (see "Experimental Procedures"). The reaction in lane 4 shows the results of cross-linking pL4(1-22) in the presence of 1 mM Dl"-inactivated NEM (DTT was present a t a final concentration of 20 mM). The samples were analyzed by SDS-PAGE and fluorography on 10% gels. The position of several molecular mass marker proteins is indicated at the right.

reside in both mitochondrial membranes, the results of the experiment presented in Fig. 5 are important for another rea- son. It has recently become clear that distinct translocation machineries reside within both the outer and inner mitochon- drial membranes (Glicket al., 1991; Pfanner et al., 1992). It has also been established that specific import of precursor proteins can be achieved across isolated outer or inner membrane vesicles (Hwang et al., 1989; Mayer et al., 1993). The latter observation suggests that the respective translocation machin- eries of each membrane are largely intact in appropriate prepa- rations of membrane vesicles. Our observation that essentially the same cross-linking profile is seen with either intact mito- chondria or membrane vesicles (Figs. 2 and 5) suggests that our vesicle preparations are clean. More significantly, the results shown in Fig. 5 underscore the fact that our cross-linking re- actions are highly specific.

l h o Cross-linked Products Are Components of the Mitochon- drial Import Machinery-As a first step toward characterizing the proteins that interacted with pL4(1-22), we asked whether any of the cross-linked products corresponded to known mito- chondrial proteins. To accomplish this, mitochondria were first incubated with the radiolabeled peptide and cross-linked using EGS as before. Aliquots of the cross-linked organelles were then solubilized and analyzed by immunoprecipitation using antisera that had been prepared against a number of different

yeast mitochondrial proteins, including OMP45p, a 45-kDa outer membrane protein whose function is presently unclear (Yaffe et al., 1989); Mas7Op, an outer membrane surface recep- tor (Hines and Schatz, 1993); ISP42p, a component of the outer membrane translocation site (Vestweber et al., 1989; Baker et al., 1990); p32, the 32-kDa product of the Mirl gene, proposed to be an outer membrane import receptor (Murakami et al., 1990; Pain et al., 1990; Murakami et al., 1993); MasGp, the product of the MAS6 gene, which is a 23-kDa integral inner membrane protein that may be involved in transport across the inner membrane (Emtage and Jensen, 1993); porin, a 29-kDa outer membrane protein that forms an anionic channel in the outer membrane (Dihanich, 1990; Pfaller et al., 1990), and AAC (the ADP/ATP carrier), an abundant 30-kDa inner membrane protein (Pfanner and Neupert, 1987) whose mammalian hom- olog was previously shown to be cross-linked to a presequence peptide (Meyer, 1990).3 Importantly, each of the antisera used has been shown to be able to immunoprecipitate its correspond- ing antigen in control experiments performed in both this and other laboratories (data not shownh4

Most of the antisera failed to yield radiolabeled products when the immunoprecipitates were resolved by SDS-PAGE and fluo- rography. Included in this group were antibodies directed against OMP45 (Fig. 6 A , lane 21, porin (Fig. 6B, lane 101, p32 (Fig. 6B, lane 11 ), Mas6p (Fig. 6B, lane 121, and AAC (Fig. 6B, lane 13). The results suggest that these proteins were not among those strongly cross-linked to pU(1-22) under the conditions of our experiments. However, antisera against two of the proteins, ISP42 and Mas7Op, immunoprecipitated distinct radiolabeled products of the appropriate size (Fig. 6, A, lanes 3 and 8, and B, lane 91, thereby suggesting that these proteins did interact with the peptide. For ISP42, we showed that immunoprecipitation could be significantly diminished if the radiolabeled peptide was chased with an excess ofthe cold derivative prior to cross-linking (Fig. 6 A , lane 4) . In contrast, if we used a nonspecific competitor for the chase (pATIII, see above), ISP42 could still be immuno- precipitated (Fig. 6 A , lane 5 ) . ISP42 could also be immunopre- cipitated, albeit weakly, with antisera prepared against total yeast mitochondrial outer membranes (Fig. 6, lane 1). For Mas70p, the antisera precipitated two radiolabeled proteins, which, based upon electrophoretic mobility, appeared to be au- thentic Mas70p and ISP42 (Fig. 6B, lane 9). This result is in agreement with previous reports that the Neurospora crassa homologs of ISP42 and Mas70p (MOM38p and MOM72p, re- spectively) are part of a receptor complex in the mitochondrial outer membrane (Kiebler et al., 1990; Sollner et al., 1992) and

G. Shore, personal communication. M. Douglas, R. Jensen, D. Pain, G. Schatz, and M. Yaffe, personal

communications.

Proteins That Interact with Mitochondrial Presequence 6441

A T M S

- 29K

B

- 98K

- 68K

- 43K

OM IDF I M

- 68K

- 43K , . .-.

1

L

M 2 3 3 4 6 7 8 9 9 0

C

- Isp42

- Va

Fxn: M 2 3 4 5 6 7 8 9 10 FIG. 5. pU(1-22) is primarily cross-linked to membrane proteins. A, mitochondria (0.25 mg) were cross-linked to radiolabeled pL4(1-22)

and then separated into total membrane and soluble fractions as described under “Experimental Procedures.” Samples representing 50 pg of intact mitochondria (T) (pretreated with radiolabeled peptide and cross-linked as described in the legend to Fig. 2) or the total membrane ( M ) and soluble fractions ( S ) derived from approximately 50 pg of mitochondria were analyzed by SDS-PAGE and fluorography on 10% gels. The positions of representative molecular mass standards are indicated at the right. B, 50 mg of mitochondria were separated into outer and inner membrane vesicles using the procedure of Hwang et al. (1989). The pelleted membrane vesicles were resuspended in 75 p1 of SH buffer and cross-linked to radiolabeled pL4(1-22) using the procedure described for mitochondria under “Experimental Procedures.“ Cross-linking reactions contained approximately 100 pg of membrane vesicles in 10 pl. After stopping the reactions with glycine, the samples were then processed and analyzed by SDS-PAGE and fluorography on a 10% gel. The lane designated M represents cross-linked products from 50 pg of intact mitochondria. The number under each lane represents the sucrose gradient fraction from which the vesicles were derived, numbering from the top to the bottom of the gradient. Fractions containing outer membrane (OM) or inner membrane (ZM) vesicles were determined by immunoblotting samples from each fraction against known marker proteins (below). IDF, intermediate density fraction. The arrowheads denote the positions of the major cross-linked products that were shown to be competable in Fig. 2. An arrowhead marking a protein in fraction 2 indicates that the protein was enriched in the outer membrane. The protein marked with the arrowhead in fraction 7 indicates that the product appeared to be enriched in the inner membrane. The arrowhead marking the protein in fraction 5 indicates that this product appeared essentially uniform throughout the gradient. The position of representative molecular mass standards is indicated at the right. C, 50 pg of total mitochondrial protein (MI or 50 pg of membrane vesicles derived from the indicated gradient fractions were fractionated by SDS-PAGE on a 15% gel. Proteins from the gel were then transferred to nitrocellulose and analyzed by immunoblotting using antibodies prepared against the outer membrane protein ISP42 (Isp42) and the inner membrane protein subunit Va of cytochrome oxidase (Vu).

Ab: MOM 45 42 42 42 Ab: 42 70 P p32 M 6 AAC “

Comp: - - - S NS FIG. 6. pU(1-22) forms cross-links with ISP42 and Mas?@. A, cross-linking or cross-linking/competition using 40 pg of mitochondria and

radiolabeled pL4(1-22) was performed as described in the legend to Fig. 2. After stopping the reactions, the samples were subjected to immuno- precipitation as described (Glaser and Cumsky, 1990a). Antibodies used were raised against yeast mitochondrial outer membranes (MOM), OMP45 (45) , or ISP42 (42 ) (see text). Competitor peptides were nonlabeled pL4(1-22) (S, specific) or pATIII (NS, nonspecific). Lane 6 contains 40 pg of cross-linked mitochondria. B, experiment is the same as in A except antibodies used were raised against ISP42 (42) , Mas7Op (70), porin (P), p32 (32) , Mas6p (M6), or AAC (AAC) (see text). Lane 7 contains 40 pg of cross-linked mitochondria. ISP42 was included as a control, since the experiments shown in panels A and B were performed with different preparations of peptide and mitochondria.

that, under certain conditions, the two proteins can be immu- noprecipitated together (Kiebler et al., 1990). When taken to- gether, the results of Fig. 6, A and B , strongly support the con- clusion that ISP42 and Mas70p were cross-linked to pL4(1-22) and, therefore, must interact with the peptide under the con- ditions of our experiments.

DISCUSSION

The results presented here extend those of our earlier studies on the interaction of the presequence peptide pL4(1-22) with

isolated yeast mitochondria. Consistent with our earlier work, we found that pL4(1-22) was efficiently taken up by mitochon- dria and, using chemical cross-linking as an assay, that it in- teracted with a number of mitochondrial membrane proteins. The interaction between pL4(1-22) and several of those pro- teins was found to be specific by two independent criteria. First, cross-linking to radiolabeled pU(1-22) could be competed by an excess of the nonlabeled derivative. Second, pL4(1-22) that had been inactivated by NEM could not be cross-linked to any significant degree.

6442 Proteins That Interact with Mitochondrial Presequence

The apparent molecular mass of several major cross-linked products fell into distinct size ranges: 30, 40-45, and 70 kDa. Therefore, we asked if those products corresponded to previ- ously identified yeast mitochondrial membrane proteins. Using immunological techniques, we determined that the outer mem- brane proteins, porin, OMP45, and p32, and the inner mem- brane proteins, Mas6p and AAC, had not been cross-linked and, therefore, did not appear to strongly interact with pL4(1-22) under the conditions of our experiments. On the other hand, the outer membrane proteins, Mas7Op and ISP42, had become cross-linked and, therefore, must somehow interact with this presequence peptide.

What is the nature of the interaction between pL4( 1-22) and Mas70p/ISP42? Both of these proteins are often grouped within the general category of import receptors. Mas7Op appears to function in the initial recognition and binding of precursor pro- teins to mitochondria, while ISP42 is thought to play a critical role in the recognition or uptake of bound precursors at the outer membrane translocation site (Pfanner and Neupert, 1990; Glick and Schatz, 1991). One possible interpretation of our data is that the function of ISP42 and Mas7Op includes the ability to recognize, and therefore interact directly with, mito- chondrial presequences. Alternatively, it is also possible that the interaction between pL4(1-22) and ISP42/Mas70p results from the fact that pL4(1-22) is present in high concentrations within the outer membrane translocation site, which is thought to include both of these proteins (Vestweber et al., 1989; Kiebler et al., 1990; Sollner et al., 1992).

At present, the available data do not permit us to distinguish between the two interpretations. However, several observa- tions suggest that the association between pL4(1-22) and the proteins to which it was strongly cross-linked, including ISP42 and Mas7Op, results from something more specific than the accumulation of peptide within the translocation sites. First, pL4(1-22) was reproducibly cross-linked to a discrete set of proteins, even when a different cross-linking reagent was used.2 Second, several of the proteins cross-linked to pL4(1-22) were distinct from those observed in cross-linking studies per- formed in other laboratories (Gillespie, 1987; Vestweber et al., 1989; Sollner et al., 1992; Scherer et al., 1992), although we note that both ISP42 and Mas70p (or their N. crassa homologs, MOM38 and MOM72, respectively) have been cross-linked to intermediates spanning the outer and inner membrane trans- location sites (Vestweber et al., 1989; Sollner et al., 1992). Fi- nally, as shown in Fig. 5B, we were able to cross-link pL4(1-22) to the same set of proteins even after the mitochondria had been fractionated into preparations of outer and inner mem- brane vesicles. Considering the heterogeneous nature of these membrane preparations, we find it striking that such a high degree of cross-linking specificity was retained in the absence of mitochondrial integrity. We suggest that when taken to- gether, these results provide additional support for the notion that the interaction between pL4(1-22) and yeast mitochondria is a highly precise one.

Could both ISP42 and Mas70p have the ability to specifically recognize and decode the information contained within mito- chondrial presequences? Clearly, the answer to this question remains to be conclusively determined. It has long been thought that the so-called surface receptors (Mas7Op and Mas20p in yeast and MOM72 and MOM19 in N. crassa) carried out this function, based primarily on the observations that the receptors mediate the initial interaction between precursors and mitochondria (Pfaller et al., 1988; Pfanner et al., 19881, and that precursors lacking a functional presequence fail t o bind to mitochondria efficiently, if at all (Hurt et al., 1984, 1985). On the other hand, it now seems apparent that mitochondrial rec- ognition and binding of precursor proteins is a complex process

that may require several different mitochondrial components (Kiebler et al., 1990; Hines and Schatz, 19931, cytosolic factors (Murakami and Mori, 1990; Murakami et al., 1992), and por- tions of the precursor other than the presequence (Pfanner et al., 1987; Hines and Schatz, 1993). The precise role of ISP42 in the import process is even less well defined, although ISP42 function is essential for both protein import and viability of yeast (Baker et al., 1990). Since surface receptors appear to enhance import efficiency and are not obligatory for uptake (Pfaller et al., 1989; Miller and Cumsky, 1991; Hines and Schatz, 19931, it is clear that mitochondria must contain an alternative receptor (or receptors) that first, accept precursors that are already bound to the outer membrane and second, can mediate recognition and binding of precursor proteins in the absence of surface receptor (Mas7Op, MasBOp, MOM 19, MOM 72) function. I t seems likely that such a receptor resides at the outer membrane translocation site; whether it is ISP42 re- mains to be demonstrated.

Currently, we are in the process of performing experiments that we hope will allow us to determine whether ISP42 and Mas70p recognize presequences directly or whether cross-link- ing occurred simply because of the proximity of pL4(1-22) to these proteins while it was within the translocation sites. We argue, however, that in either case, the experimental approach described here clearly provides a means to identify membrane proteins that may be fundamentally involved in the process of mitochondrial protein import. It is possible that one or more of the polypeptides that become strongly cross-linked to pL4( 1-22) under the conditions of our experiments may represent previ- ously uncharacterized proteins that comprise part of the trans- location machinery or play some other role in the uptake of precursor proteins. Recent work in our laboratory has resulted in the purification of one of these proteins (the 75-kDa outer membrane protein, which we do not think is Mas70p) and the partial purification of several others (60, 40, and 30 m a ) , which we are currently in the process of characterizing in more detail.2 We have also found that several less strongly cross- linked proteins can be identified on the basis of their immuno- logical recognition by a complex antisera generated against total mitochondrial inner membranes (a gift of G. Schatz). We are hopeful that the purification and characterization of these proteins will lead us to their corresponding structural genes and ultimately to the elucidation of what role, if any, they play in mitochondrial import.

M. Klingenberg, D. Pain, G. Schatz, and M. Yaffe for the generous gifts Acknowledgments-We are very grateful to Drs. G. Blobel, R. Jensen,

of antisera. We also thank Joseph Kosmoski and Dr. C. Glabe for con- tinued advice and assistance on peptide synthesis and purification. We acknowledge several members of our laboratory, including Dr. B. Miller, Dr. G. Blickenstaff, Dr. V. Bilanchone, and L. Jung for advice, helpful discussions, and comments on the manuscript.

REFERENCES

Atherton, E., Logan, C. J., and Sheppard, R. C. (1981) J. Chem. Soc. Perkin Z-ans.

Baker, K. F!, Schaniel, A,, Vestweber, D., and Schatz, G. (1990) Nature 548,605-

Dihanich, M. (1990) Experientia (Basel) 46, 14C152 Dryland, A., and Sheppard, R. C. (1986) J. Chem. SOC. Perkin Z-ans. I, 125-137 Emtage, J., and Jensen, R. E. (1993) J. Cell Biol. 122, 1003-1012 Gillespie, L. L. (1987) J. Biol. Chem. 262, 7939-7942 Glaser, S. M., and Cumsky, M. G. (1990a) J. Bwl. Chem. 285,880%8816 Glaser, S. M., and Cumsky, M. G. (1990b) J. Biol. Chem. 285,8817-8822 Glaser, S. M., Miller, B. R., and Cumsky, M. G. (1990) Mol. Cell. Biol. 10,1873-

Glick, B., and Schatz, G. (1991) Annu. Reu. Genet. 25, 2 1 4 4 Glick, B. S., Wachter, C., and Schatz, G. (1991) Dends Cell Biol. 1,99-103 Hawlitschek, G., Schneider, H., Schmidt, B., Tmpschug, M., Hartl, F.-U., and

Hines, V., and Schatz, G. (1993) J. Biol. Chem. 268,449-454 Honvich, A. L., Kalousek, F., Mellman, I . , and Rosenberg, L. E. (1985) EMBO J. 4,

Hurt, E. C., Pesold-Hurt, B., and Schatz, G. (1984) EMBO J. 3 , 3 1 4 9 4 1 5 6

I, 538-546

609

1881

Neupert, W. (1988) Cell 53,795-806

1129-1135

Proteins That Interact with Mitochondrial Presequence 6443 Hurt, E. C . , Pesold-Hurt, B., Suda, K., Oppliger, W., and Schatz, G. (1985) EMBO

Hwang, S., Jascur, T., Pon, L., and Schatz, G. (1989) J. Cell Biol. 109, 487-493 Jensen, R. E., and Yaffe, M. P. (1988) EMBO J. 7,3863-3871 Kiebler, M., Pfaller, R., Sl lner , T., Griffiths, G., Horstmann, H., Pfanner, N., and

Maarse, A. C. , Blom, J., Grivell, L. A., and Meijer, M. (1992) EMBO J. 11, 3619-

Martin, J., Mahlke, K., and Pfanner, N. (1991) J. Biol. Chem. 266, 18051-18057 Mayer, A,, Lill, R., and Neupert, W. (1993) J. Cell Biol. 121, 1233-1243 Meyer, D. I. (1990) Nature 347,424-425 Miller, B. R., and Cumsky, M. G. (1991) J. Cell Biol. 112, 833-841 Miller, B. R., and Cumsky, M. G. (1993)J. Cell Biol. 121, 1021-1029 Murakami, K., and Mori, M. (1990) EMBO J. 9, 3201-3208 Murakami, H., Pain, D., and Blobel, G. (1988) J. Cell Biol. 107, 2051-2057 Murakami, H., Blobel, G., and Pain, D. (1990) Nature 347,488491 Murakami, K., Tanase, S., Morino, Y., and Mori, M. (1992) J. Biol. Chem. 267,

J. 4,2061-2068

Neupert, W. (1990) Nature 348,610-616

3628

13119-13122

Pfaller, R., Steger, H. F., Rassow, J., Pfanner, N., and Neupert, W. (1988) J. Cell

Pfaller, R., Pfanner, N., and Neupert, W. (1989) J. Biol. Chem. 264,3439 Waller, R., Kleene, R., and Neupert, W. (1990) Experientia (Basel) 46, 153-161 Pfanner, N., and Neupert, W. (1987) J. Biol. Chem. 262,752%7536 Wanner, N., and Neupert, W. (1990)Annu. Reu. Biochern. 59,331353 Pfanner, N., Muller, H. K., Harmay, M. A,, and Neupert, W. (1987) EMBO J. 11,

Pfanner, N., Hartl, F.-U., and Neupert, W. (1988) Eur. J . Biochem. 175,20&212 Pfanner, N., Rassow, J., van der Klei, I. J., and Neupert, W. (1992) Cell 68,999-

Pon, L., Moll, T., Vestweber, D., Marshallsay, B., and Schatz, G. (1989) J. Cell Biol.

Roise, D., and Schatz, G. (1988) J. Biol. Chem. 263,4509-4511 Scherer, P. E., Manning-Krieg, U. C., Jeno, P., Schatz, G., and Horst, M. (1992)

Sollner, T., Rassow, J., Weidmann, M., Schlosmann. J.. Keil. P., Neupert, W., and

Biol. 107,2483-2490

34494454

1002

109,2603-2616

Proc. Natl. Acad. Sci. U. S. A. 89, 1193C11934

Murakami, H., Blobel, G., and Pain, D. (1993) Proc. Natl. Acad. Sci. U. S . A. 90, Pfanner, N. (1992) Nature 355,8447 335W362 Vestweber, D., Brunner, J., Baker, A,, and Schatz, G. (1989) Nature 341,205-209

Pain, D., Murakami, H., and Blobel, G. (1990) Nature 347,444-449 Yaffe, M. P., Jensen, R. E., and Guido, E. C. (1989)J. Biol. Chern. 264,21091-21026