The ins and outs of mitochondrial protein import from

0 downloads 0 Views 1MB Size Report
Transworld Research Network. 37/661 (2), Fort P.O., ..... functional properties. For this reason conductance is considered the hallmark of any channel's identity.
Transworld Research Network 37/661 (2), Fort P.O., Trivandrum-695 023, Kerala, India

Recent Res. Devel. Biophys., 3(2004): 413-474 ISBN: 81-7895-130-4

22

The ins and outs of mitochondrial protein import from an electrophysiological point of view Pablo M.V. Peixoto1, Sonia Martínez-Caballero1,2, Sergey M. Grigoriev2 Kathleen W. Kinnally2 and María Luisa Campo1 1 Dept. de Bioquímica y Biología Molecular y Genética, Facultad de Veterinaria Universidad de Extremadura, Aptdo. 643, 10071 Cáceres, Spain; 2New York University, College of Dentistry, Dept. Basic Sciences, 345 East 24th Street New York NY 10010, USA

I. Abstract The application of electrophysiological techniques, in particular patch-clamping, to mitochondrial membranes represented a turning point on the biophysics of mitochondria. The electrophysiological approaches used in these studies as well as the information they provide are outlined. The advantages and drawbacks of each one of these techniques are also discussed. The existence of channels, most of which are very large, in particular in the inner membrane of mitochondria, was an astonishing Correspondence/Reprint request: Dr. María Luisa Campo, Dept. de Bioquímica y Biología Molecular y Genética Facultad de Veterinaria, Universidad de Extremadura, Aptdo. 643, 10071 Cáceres, Spain E-mail: [email protected]

414

Pablo M.V. Peixoto et al.

finding. A survey of the state of the art of investigations on each one of these mitochondrial channel is presented. A special emphasis is prone to the physiological role each of these channels may play; particularly the involvement in protein translocation processes of some of the larger channels found in both mitochondrial membranes. The versatility of patch-clamping combined with biochemical and genetic alterations of the protein import translocases of mitochondria has been integral to tackle structure-function relationship studies of the three protein import complexes yet described in mitochondria. Essential pore components have been identified, and specific roles could be assigned to others. An analysis of the character of the aqueous pores through which mitochondrial proteins are translocated has been obtained.

II. Introduction Multicellular organisms probably could not exist without mitochondria. These bacteria-sized organelles make efficient use of nutrient molecules requiring oxygen in the process. For this reason they are found in the cytoplasm of virtually all eukaryotic cells, and are especially abundant in cells and organelles that are associated with active processes. Apart from a central role in energy-generating processes, mitochondria are involved in other complex processes such as apoptosis and cardioprotection. In addition, mitochondrial dysfunctions play pivotal roles in neurodegenerative disorders from Parkinson’s to Huntington’s to Alzheimer’s diseases. In recent years there has been a renewed interest in the field of mitochondria and our view of these organelles changes with each new discovery of their involvement in yet another fundamental biological process. A complete description of the regulation of mitochondrial energy metabolism would not be possible without taking into account the spatial organization of mitochondrial compartments. The two membranes that surround mitochondria create distinct compartments within the organelle, and are themselves very different in structure and function. The outer membrane is a relatively simple phospholipid bilayer containing proteins. In the past few years it has become clear that much of the molecular traffic crossing the outer membrane does so via the Voltage Dependent Anion-selective Channel called VDAC or mitochondrial porin (1). This abundant protein represents more than 60% of the outer membrane proteins in some organisms, and makes this membrane permeable to molecules up to 5 kDa. Nutrient molecules, adenine nucleotides (ADP, ATP), and ions pass through the outer membrane with ease. The large size of many of the permeating species led many investigators to view the outer mitochondrial membrane as a leaky structure of little importance to the mitochondrion’s function. However, the channels isolated from the outer membrane are not only large enough to account for the

Channel activity of mitochondrial protein import

415

permeability of this membrane to non-protein molecules, but have also been shown to be controllable (2,3). Thus, VDAC can be regulated in a number of ways that implicate it as a site for the regulation of mitochondrial function, yet technical limitations have prevented the extension of these studies to a relevant cellular context (4). In this sense, it is becoming clear that VDAC probably acts as a convergence point for a variety of life-or-death signals (5-8). The possibility that, in vivo, these channels are regulated raises an important issue: i.e. the function of the inner mitochondrial membrane could depend on the permeability of the outer membrane. Therefore, the outer mitochondrial membrane is located at the crossroads of a great deal of traffic through mitochondria; nevertheless, its structure and function can not be regarded as trivial. In contrast, the inner membrane is freely permeable only to O2, CO2, NH3, and water. Other hydrophilic metabolites and all inorganic ions of biological importance can cross this membrane due to the presence of specific carrier proteins and channels. The molecular composition of this membrane is highly complex, including all of the electron transport system, the ATP synthetase complex, and many transport proteins. The total surface area is greatly increased by the wrinkles, or folds, highly organized into lamellae or cristae (9). The cristae are not lamelliform sheets (as depicted in textbooks) but irregularly shaped internal compartments connected by long, narrow (ca. 20 nm diameter) tubes to each other and to the periphery of the inner membrane (10). Although the outer and the inner mitochondrial membranes are well defined structures, each of them possessing different sets of enzymes and fulfilling different functions, intimate contacts between the two membranes have been identified on both morphological and functional grounds (3,11-13). In this review we will give a short account of our joint effort to solve the problem of the existence of large conductance channels, especially in the inner membrane of mitochondria, and then focus on further developments, with the intent of stimulating additional investigation on this topic.

III. Electrophysiologycal mitochondrial channels

approaches

to

study

Playing as gatekeepers to signal transduction and metabolism, channels are present in most cell membranes. It is undoubted that ion channel research has been stimulated by the progressive advancement of electrophysiological techniques, making available for innumerable studies a large quantity of cellular and sub-cellular channels that were not suitable for it before the development of these techniques. The existence of channels in mitochondria, especially in their inner membrane, although premised by several biochemical studies in the early 80’s (14), required the application of electrophysiological

416

Pablo M.V. Peixoto et al.

techniques. Before, the first direct approach to measure mitochondrial membrane potential was the impalement of single mitochondrion with microelectrodes, initially used in Henry Tedeschi’s laboratory (15-17). However, no direct evidence for the existence of channels could be inferred from this approach. The reason was the thermal noise of the signal source, which is the basic limitation for any current measurement, disregarding instrumentation noise. The majority of in-depth studies on mitochondrial membrane channels have been carried out in reconstituted systems, mainly done in planar bilayers or using tip-dip techniques. In the late 80’s a big step forward was taken when the technical difficulties were overcome and it became possible to apply patchclamp techniques on isolated mitochondria or mitoplasts (18,19).

A. Bilayer studies Black lipid membranes become electrically conductive in the presence of certain proteins added in trace amounts. The step-like changes in conductance observed account for the insertion into the membrane of single pore-like structures (20). The most common approach to obtain bilayers is the “paintedbilayer” technique in which lipids dissolved in organic solvents are spread with a small brush onto a micropore (50-200 µm diameter) that is punched through a Teflon® (DuPont, Wilmington, DE) or plastic septum (12-25 µm thickness) separating two different aqueous compartments (Figure 1A). Alternatively, a bilayer can be obtained if the solutions of the two compartments containing lipid monolayers, spontaneously formed at the air-water interface, are evenly and simultaneously raised above the micropore (21). In both cases, the presence of the planar bilayer is determined by measurement of the characteristic bilayer capacitance (22), and proteins can be incorporated by fusing proteoliposome vesicles made with charged lipids to the preformed bilayers. This system can be used to study permeation and gating properties of ion channels in a chemically isolated environment. An additional advantage is the easy exchange of solutions in either side of the membrane. The majority of the electrophysiological studies of VDAC have been carried out with this approach.

B. The “tip-dip” method Bilayer formation at the tip of a pipette (23) is fundamentally analogous to that of bilayer formation from two monolayers at the air-water interface across a small aperture. In principle, a microelectrode tip (0.5-1 µm diameter) is dipped once into a solution containing a lipid-protein monolayer at the airwater interface, while positive pressure is applied. The pressure is released, and the electrode is withdrawn through the interface to pick up a monolayer on

Channel activity of mitochondrial protein import

417

the tip. Then the pipette is reimmersed into the solution, leading to the apposition of the second monolayer, thereby forming a bilayer at the tip of the microelectrode (Figure 1B). A high salt concentration (i.e. 1 M) or a concentration gradient, or the inclusion of a small amount of detergents, are Measure electrode

Cis chamber Agar bridge

Reference electrode

Measure electrode

Reference electrode Trans chamber

Septum

A

BB

Micromanipulator Reference electrode

C Perfusion

Measure Headstage electrode

D

Figure 1. Experimental setups for bilayer, “tip-dip”, and patch-clamp techniques. (A) Photograph of the chambers for the formation of planar lipid bilayers across an aperture in a Teflon® septum. Two Ag/AgCl electrodes are coupled to the cis and trans reservoirs via salt bridges (KCl in 2% w/v agar). One of the electrodes is connected to ground, and the other to the amplifier. (B) Setup for the reconstitution of ions into a bilayer formed at the tip of a microelectrode in the “tip-dip” method. The photograph shows the chamber, a salt bridge or reference electrode and the measure electrode fixed by a Teflon® holder. Both experimental setups should be located inside a Faraday cage and mounted on a vibration-isolation platform. (C) A detail of the patch-clamp setup showing the recording chamber mounted on the stage of an inverted microscope. The chamber contains a perfusion system for the exchange of solutions, a salt bridge or reference electrode and a microelectrode. The microelectrode mounted in a holder is connected to the headstage of the amplifier (rectangular box at the right) via highquality chassis BNC connector. A three-axis micromanipulator is attached to the headstage. (D) Example of a complete patch-clamp setup including an inverted microscope and a micromanipulator system enclosed inside a Faraday cage and placed on an air table. A rack containing the amplifier, a digital frequency filter, oscilloscopes and a digital recorder is shown on the left.

418

Pablo M.V. Peixoto et al.

conditions that can be used to regulate channel insertion in tip dip as well as bilayers. The Tom channel, previously called PSC (Peptide Sensitive Channel), of the outer membrane of mitochondria was discovered by this technique.

C. Patch-clamp techniques Model membrane systems have proven to be useful tools for studying the biophysics of lipid bilayers and transmembrane proteins. At the time bilayer and tip-dip techniques were used to study channels in these artificial membranes, no direct evidence was available from biological preparations, since the methods for recording currents in living cells have background noise levels about a hundred times higher than the single current observed in bilayers. A better method for recording currents from biological preparations became available when Neher and Sakmann succeeded in isolating a small area of membrane surface (a “patch”) and obtained a “pipette-to membrane seal” with Giga-ohm electrical resistance (a “Giga-seal”). This and the improvement of the electronic amplifiers allowed resolution of picoamperesized currents (24), and therefore to record square-wave signals, the proof that channels in biological membranes open and close stochastically in an all-ornone manner. The main difficulty in obtaining a Giga-seal from mitochondria or mitoplasts membranes arises from the small size of these organelles (1-2 µm and 2-5 µm diameter, respectively). Thus, the relationship between the shape and the electrical resistance of the microelectrode becomes a crucial issue. Micropipettes are obtained with a pipette puller from thin wall borosilicate glass capillaries with an inner filament. The fine tip (0.2-0.5 µm), with a short taper in the micropipette does not require further fire polishing. The micropipette is then filled with a conducting solution (usually 150 mM KCl) prior to mounting on a micromanipulator with the aid of a micropipette holder (Figure 1C). A microelectrode is attained when a silver/silver chloride wire inserted into the micropipette is connected to the headstage of a patch-clamp amplifier. The circuit is completed when the tip of the microelectrode and that of an agar reference electrode are immersed into a conducting bath (Figures 1C and 1D). Under these conditions the electrical resistance of the microelectrode should not be higher than 30 MΩ. To obtain a Giga-seal the microelectrode is placed at about 1 µm distance from the mitochondrion, while positive pressure is continuously applied, in order to keep the tip clean from any debris. The pressure is released and the microelectrode is further advanced to gently touch the organelle membrane. Quite often this is sufficient to form a tight seal between the glass tip and the membrane. Sometimes however, a small negative pressure is applied through the microelectrode in order to form the seal. Failure to attain a Giga-seal implies changing the microelectrode every time an attempt is made, since the small tip gets “dirty” once it touches the membrane.

Channel activity of mitochondrial protein import

419

One of the several advantages of patch-clamping is the versatility it offers to study channels (and single channels) from both sides of the membrane and in different ambiances, due to the possibility to achieve different configurations. If the organelle remains intact after the seal formation, the patch is in the attached configuration. Although the physical basis of the Gigaseal formation is still not clear, it is recognized to provide mechanical stability, so that patches could be “excised” from the mitochondrion or mitoplast and studied under a different configuration just by drawing the microelectrode away. Alternatively, the attached patch can be broken (without the loss of the seal) if short (3-300 ms) voltage steps of 200-500 mV are applied (25). The rupture of the membrane at this spot provides electrical continuity between the patch microelectrode and the mitochondrial matrix leading to a wholemitoplast configuration, similar to conventional microelectrode impalement. The advantage raises if we consider that in small organelles like mitoplasts (25 µm) the resistance of the impalement microelectrode and the mitoplast impedance are of the same order of magnitude, thus requiring feedback circuitry for voltage clamping. On the contrary, the internal resistance of the patch microelectrode is orders of magnitude higher than the mitoplast impedance. A disadvantage is however, that any regulatory molecule of a channel could be lost when a patch is excised. In addition, even in wholemitoplast recording mode the internal medium (mitochondrial matrix) is quickly exchanged with that of the microelectrode. Nevertheless, the ability to observe, in real time, the dynamics of individual mitochondrial channels in a steady state has important implications for studies on the molecular biology of ion-permeable channels.

D. Conductance as the hallmark of channel properties These three general approaches to apply electrophysiological techniques to mitochondria rely upon current measurements under voltage clamp conditions. To voltage clamp a membrane an operational amplifier is used as a current-tovoltage converter. In this arrangement the membrane potential is held at a preselected value (holding potential) by injecting current. The current flowing through a single channel or a set of channels in a bilayer or at the tip of a microelectrode tends to cause a voltage drop. The current that is applied to maintain the membrane at the precise holding potential is then measured. It is remarkable that much of what we know about ion channels is thus deduced from current measurements and based on the very essential principle formulated in Ohm’s law (I = gV): a relation between current (I), voltage (V), and conductance (g). Conductance (i.e. the ease of flow of current between two points, like both ends of a membrane channel) is measured in siemens (S), and depends entirely on the channel’s physical and functional properties. For this reason conductance is considered the hallmark of any channel’s identity.

420

Pablo M.V. Peixoto et al.

Mitochondrial channels, like all other channels, open and close quantum and stochastically allowing the passive flux of ions (current) down their electrochemical gradient. This behavior is the reflection of abrupt conformational transitions taking place in a time scale briefer than a millisecond. The amplitude of the transitions or flickering between the open and closed states is dependent on the voltage applied across the membrane. However, the actual size of the channel (the conductance of the channel) is constant, and is calculated according to Ohm’s law, dividing the current measured by the voltage applied. The average distribution of the current for a certain period of time gives rise to the total amplitude histogram, in which the percentage of time that the channel spends at a particular current level is plotted (26). The mean current level is usually taken from such histograms. In single channel experiments, the conductance of the channel is calculated from the ratio between mean current level and the applied voltage. Total amplitude histograms also allow calculations of the open probability of the channel and that of any other defined substate. In addition, the number of channels present can be determined by dividing the peak conductance by the single-channel conductance. When the channel’s behavior is recorded long enough, and if only one channel is present, several parameters can be statistically determined, such us the rate of flickering, mean open and mean closed time, mean length of the activity periods (burst length), as well as total open and closed times. Kinetic models can be derived from these parameters (27,28). The selectivity of the channel can be determined from current-voltage relationships under asymmetrical conditions, that is having a different salt concentration in the microelectrode and in the bath. The polarity of the reversal potential (i.e. the voltage necessary to cancel any ion flow under asymmetrical conditions) indicates the cationic or anionic nature of the channel; while its magnitude determines the relative permeability to different ions (27,29). Another source of information is provided by looking at the probability of the channel for being open as a function of voltage (voltage dependence). This is a characteristic of each channel that could be affected by many factors, including drugs (30). In fact, a close look at the current transitions and their kinetics enables the complete fingerprint of any channel.

E. Native and reconstituted mitochondrial membranes To date it is clear that to study channels involved in mitochondrial processes, it is desirable to use the molecular resolution provided by patchclamp techniques. Until a few years ago this was a considerable problem, mainly because mitochondria are small, and difficult to access within living cells. Nevertheless, these techniques were first applied to mitoplasts (mitochondria stripped of their outer membranes) isolated from different

Channel activity of mitochondrial protein import

421

mammalian tissues (30,31), as well as the tiny mitochondria isolated from yeast cells (32), and those from cell lines (33). Most recently, recording channel activity from mitochondria within a cell has become possible, using a “patch within a patch” technique, in which the patch microelectrode is kept clean inside a double barrel micropipette while the outside barrel perforates the plasma membrane (34). In this study, mitochondria located within the presynaptic terminals of the squid giant axon were patch-clamped in a region that contained mitochondria, but no other obvious intracellular membranes as assessed by electron microscopy. While it is preferable to perform studies under physiological conditions, it is not always feasible to form the Gigaseal necessary for single channel analysis. For example, VDAC is a high conductance channel occurs at surface densities of 103 to 104 per µm2 on native outer membranes of fungal and animal mitochondria (35). We can estimate that in a patch pipette with a tip diameter of 0.5 µm there are hundreds of VDAC molecules, assuming they are randomly distributed. The high conductance of VDAC (650 pS in 150 mM KCl) together with its high density makes the formation of Giga-seals difficult on intact mitochondria. While the Jonas lab has recently overcome these technical difficulties by fashioning even smaller microelectrodes (Jonas et al, 2004), earlier studies relied upon reconstitution. Proteoliposomes, enlarged sufficiently to be amenable to the patch pipette, were used to check whether purified mitochondrial outer or inner membranes reconstituted into pure lipid liposomes contained ion channels. This method, which involves careful dehydration-rehydration of the protein-lipid mixture has proven to be successful for all mitochondrial channels tested (32,36-38). Whenever a comparison has been made with data obtained in situ (as for mitochondria or mitoplasts), the reconstituted systems seemed to reproduce native conductances with good fidelity (38-41). Nevertheless, validation of method requires a precise scrutiny of every single channel property (38). The reconstitution method described by Criado and Keller (42) has turned out to be a useful tool for the study of mitochondrial electrophysiology. Briefly, 0-50 µg of purified membranes are mixed with 1 mg azolectin small liposomes (downsized from large multilamellar lipid vesicles by sonication in water) and brought to a final volume of 50 µL with 5 mM Tris, pH 7.4. Small drops of the mixture are placed on a glass coverslip and desiccated for 2-3 hours at 4ºC, till completely dry. Rehydration is done by covering each drop with 5 mM Tris for an overnight incubation at 4ºC. The vesicles are collected by thoroughly washing the coverslip with 0.5 mL 5 mM Tris. While heterogeneous in size, many are unilamellar proteoliposomes of several microns in diameter. Alternatively, proteoliposomes containing outer and inner membranes can be formed by freezing and thawing. In this case, the same mixture of small liposomes and purified membranes are rapidly frozen in

422

Pablo M.V. Peixoto et al.

liquid nitrogen and thawed on ice three times. The resulting proteoliposomes are smaller than those prepared by dehydration-rehydration but channel activity is readily detected (40). In addition, the simpler model system of proteoliposomes overcomes some other obstacles presented by native membranes. The reconstituted system offers the possibility to vary the protein to lipid ratio and thus to find the optimum ratio at which most patches contain just a single channel. Nonetheless, the very time consuming process of isolating mitochondria and mitoplasts, especially from yeast, prior to patch-clamping makes these experiments a truly laborious approach. A great advantage comes from the use of proteoliposomes prepared in advance and stored at –80ºC. Even if freezing breaks about half the giant unilamellar proteoliposomes, these preparations can be used for several months with no detectable changes on the electrophysiological properties of the mitochondrial channels.

IV. The channels of the mitochondrial membranes In the last few years, the field of mitochondrial transport has been transformed by the application of patch-clamping. For instance, an unexpected diversity of currents in both mitochondrial membranes has been revealed. Also striking is the size of these membrane conductances that varies from a few pS, like those of prototypical ion channels, to several nS. Classification of these channel activities is based mostly on functional properties like size, voltagedependence, pH and cation sensitivity, and pharmacology. Biochemical progress toward the structural and functional characterization of mitochondrial channels would not be possible without an extensive molecular pharmacology, including a variety of physiological effectors. In fact, most of our early knowledge of the functional architecture of mitochondrial channels came from pharmacological experiments. We performed a systematic study of the sensitivity, under patch-clamp conditions, of mitochondrial channels to chemicals known to have effects on biological functions [for a review see (43)]. The benzodiazepine receptor ligands (44), inhibitors of the respiratory chain like Antimycin A (28), and classical uncouplers of the oxidative phosphorylation (26), were examined. At least two distinct features emerged from these studies. One is a pleiotropic effect, since many drugs affect many mitochondrial channels. The evidence comes not only from indirect permeability studies using mitochondrial suspensions, but also from patch-clamp studies. To the mentioned substances we have to include amphiphilic drugs such as the cardiac antiarrhythmics propranolol and amiodarone (45) and also quinine, local anesthetics as dibucaine or dupivacaine as well as anxyolitic and muscle relaxing drugs such as the benzodiazepines RO-5-4865 or clonazepan. This

Channel activity of mitochondrial protein import

423

pleiotropic effect might have several explanations ranging from a nonspecific effect on the membrane, to different channels sharing the same receptor, or the existence of homologous domains in different channel proteins. Although most of these drugs are hydrophobic and they may cause a general effect on the membrane, like it has been described for tranquilizers and anesthetics (46,47), many data argue against this hypothesis. For example, while some of them are channel inhibitors, they have opposite effects on the membranes. In addition, closely related compounds exert very different effects on channels, suggesting the involvement of specific receptors. The alternative, that is a single receptor capable of interacting with many channels, cannot be completely ruled out. The mitochondrial benzodiazepine receptor binds a broad spectrum of drugs (48,49) including many affecting several mitochondrial channels. Finally, it is worth noting that only few of the channels are inhibited by specific compounds.

A. The channels of the outer mitochondrial membrane 1. The Voltage Dependent Anion Channel (VDAC) VDAC was the first mitochondrial channel discovered, and it stood out from other observed conductances because it showed voltage-gating properties (50). As a result of the intensive research carried out, many VDAC isoforms and orthologs have been purified and the amino acid sequence of several has been deduced by cDNA sequencing (51,52). Also, a comprehensive characterization of VDAC from different sources ranging from protist (50,53,54), fungi (55), or plants (56-59) to mammalian mitochondria (60,61) is available. From an electrophysiological point of view, VDAC’s characterization has been carried out essentially with partially purified protein incorporated into planar bilayers. Under these conditions VDAC exhibits two main conductance levels. The fully open state has a single channel conductance of 650 pS (in 150 mM KCl) and remains open most of the time at low voltages of either polarity, up to 10 mV. The open state shows a slight preference for anions over cations (PCl − /PK + = 2). A lower conductance state (substate) of 300 pS (in 150 mM KCl) is present at both positive and negative higher potentials (typically over ±30 mV). The substate is also permeable to simple salts, but the selectivity is now slightly cationic. Interestingly, the permeability of both states to mitochondrial metabolites (mostly organic anions) is dramatically different with just the fully open state being permeable to metabolites like ATP and ADP (62). Occupancy of the lower conductance substate is favored by certain polyanions (63) and endogenous effectors like NADH (64,65) as well as a soluble mitochondrial protein (66), suggesting that these states may be occupied by VDAC in its mitochondrial environment (65,67). The bell-shaped dependence of the channel’s open probability on the applied voltage is one of

424

Pablo M.V. Peixoto et al.

the main characteristics of VDAC gating, although the molecular mechanism of this process is still under discussion (68,69). Nevertheless, a detailed study of the gating properties of VDAC has revealed that under some conditions these channels can be completely or almost completely closed (70). It is established that a single polypeptide chain of 30-32 KDa forms one VDAC channel (71) with a large pore size (2.5-3.0 nm diameter in the open state and roughly 1.9 nm in the substate). Moreover, a three-dimensional structure model has been developed based on sequence analysis and electron microscopy of two-dimensional crystals (69) (72). The pore is a Beta barrel formed from a series of ß sheets in which hydrophobic amino acids face the membrane interior and adjacent hydrophilic amino acids line the aqueous ion conduction pathway. Besides a highly functionally-conserved archetype of VDAC, present in all cell types, several isoforms can be found in different organisms (73-75). They retain some of the basic properties and their differences probably rely upon specialized functions (76). VDAC is by far the best-studied mitochondrial channel. There is no question about its function as a passive permeability pathway mediating the exchange of metabolites across the outer membrane. In addition, the pore formed by VDAC can be modulated in a number a ways that implicate this channel as a site for the regulation of mitochondrial function in a cellular context. The challenge is now to gain direct evidence for the role of VDAC’s gating in intact cells, since direct correlations between in vitro observations and any in vivo processes have yet to be demonstrated. VDAC plays an important role in coordination of communication. A substantial aspect of this management is a transient formation of complexes with other proteins. For example, several ATP dependent enzymes, such as hexokinase, glucokinase, or creatine kinase, are associated to the outer membrane by binding to VDAC (8,77-79). Also, VDAC is generally assumed to be in contact sites (80) and to participate as a key component of the complex constituting the permeability transition pore (PTP) (81). Mitochondria contain an unidentified structure that forms a large unspecific pore under conditions of high matrix Ca2+, Pi and oxidative stress (82). Electrophysiological measurements revealed properties of the pore related to VDAC, suggesting that it may be a component of the whole structure (83,84). Finally, various and often conflicting mechanisms are proposed to account for the increased permeability of the mitochondrial outer membrane that is responsible for cytochrome c release from the intermembrane space during apoptosis. In any of them, the implication of VDAC in yet another fundamental biological process as apoptosis is gaining strong support (7,8,85,86). 2. Tom channel In 1987, Tedeschi and co-workers showed by patch clamping giant mitochondria that the macroscopic currents they were able to record from outer

Channel activity of mitochondrial protein import

425

membranes could not be accounted for by VDAC only (18). Therefore, they proposed that another class of channels still not described was present in these membranes. Using the tip-dip method Thieffry and Henry found a voltagedependent cation channel of large conductance in the outer membrane of mammalian (87,88) and yeast mitochondria of a VDAC deficient strain (89). They were able to characterize this channel in bilayers formed by the three methods described (i.e. planar bilayer, patch clamp of giant liposomes and tip dipping), concluding that most of its properties were not affected by the physico-chemical characteristics of the membrane (90). The peak conductance was estimated in about 850 pS in 150 mM NaCl and the presence of two conductance jumps of equal amplitude were described in yeast. However, a somehow more complicated pattern was described in mammalian outer membranes, with five conductance levels of 750, 530, 310, 210, and 110 pS respectively. Complete closures were never observed (91). According to these authors it was difficult to describe the voltage dependence due to the large variability among the records. In general they accounted for a higher open probability at cytoplasmic negative potentials. The Vo, the voltage at which the channel spends half its time open, varied from –20 to 0 mV depending on the species from which the channel was derived, while the gating charge (measure of the effective charge that moves across the membrane to fully open the channel) was estimated about 2 (89). The channel was slightly cationic (PK+ > PNa+ > PCl-) with PNa+/PCl- ≈ 3, which means 3 sodium ions pass through the channel for every chloride ion (91). The most striking characteristic of this channel was a rapid and reversible blockage induced by a positively charged 13-residue synthetic peptide, whose sequence corresponded to the amino terminus of the nuclear-coded subunit IV of yeast cytochrome c oxidase (yCoxIV(1-13)) (92,93). Very soon it was realized the possible functional implication of this finding and a new research venue was established to determine the involvement of this channel, then named PSC (Peptide Sensitive Channel), in the process of protein import into mitochondria. The details of these investigations will be outlined in section VI. 3. The Mitochondrial Apoptosis-induced Channel (MAC) Very recently, a new megachannel has been discovered in the outer membrane of mitochondria (94). It was founded present only in mitochondria undergoing apoptosis, and its activity correlated with the onset of this highly regulated process. Indeed, overexpression of the anti-apoptotic protein Bcl-2 prevented the appearance of this channel, named MAC (Mitochondrial Apoptosis-induced Channel). The channel has been described by patch clamp of mitochondria and proteoliposomes containing purified mitochondrial outer membranes from apoptotic FL5.12 cells. In addition, it has been detected in mitochondrial outer membranes of yeast expressing the pro-apoptotic protein

426

Pablo M.V. Peixoto et al.

Bax (95). Electrophysiological analysis indicated that MAC is a heterogeneous channel with variable conductance. Multiple conductance levels could be recorded, with peak single-channel transitions of 2.5 ± 0.6 nS. The large current transitions occurred at a lower frequency than the smallest transitions of 1-1.5 nS. According to Hille (27), assuming a pore length of 7 nm, this maximum conductance corresponds to a pore diameter of 4.0 ± 0.5 nm. The open probability was independent of voltage, and showed slight cationic selectivity with a permeability ratio PK+/PCl- = 3. There is considerable evidence that members of the Bcl-2 protein family exert their pro- and antiapoptotic effects by regulating the release of cytochrome c from the intermembrane space of mitochondria (96-98). While the permeability transition pore is implicated in this release in some systems (99,100), recent studies indicate cytochrome c is released early in apoptosis without loss of integrity of the mitochondrial outer membrane in some cell types (101-104). Altogether, these findings have implicated MAC as a candidate for the outer-membrane pore through which cytochrome c and possibly other factors exit mitochondria during apoptosis. Thus, the pore diameter allows the passage of 17 kDa Dextran, which should be sufficient to allow the passage of 12.5 kDa cytochrome c (105). Also, the failure of proteoliposomes containing MAC to retain cytochrome c is an indirect indication of the release of this protein through this particular channel (106). Recently, the modification of the electrophysiological activity of MAC by physiological levels of cytochrome c has provided a more direct evidence that this molecule may cross the mitochondrial outer membrane via MAC (105). That MAC and PTP are the expression of different entities has been corroborated by recent pharmacological studies in which concentrations up to 10 µM cyclosporin A, the most specific inhibitor of PTP in mitochondria (IC50 0.1-1 µM) (107-109) had no effect on MAC activity (110). Whereas the molecular identity of MAC is not yet known, it has been reported that oligomeric recombinant Bax (human) spontaneously inserts and forms high-conductance ion channels in artificial membranes (97,111,112). MAC may be composed of Bax and/or Bak.

B. The channels of the inner mitochondrial membrane Two decades ago the existence of channels in this membrane that tightly controls its permeability to ions, was challenged. Today, however, it is totally accepted and has proven a very active field of research. Electrophysiology has been decisive and provided a wealth of information. However, few of the proteins responsible for the electrophysiological activities have been identified and the challenge is now to bestow the meaning of this information by its functional significance.

Channel activity of mitochondrial protein import

427

1. The mitochondrial Centum pico-Siemens channel (mCS) The first application of patch-clamp to the native inner mitochondrial membrane resulted in the discovery of a 107 pS channel, termed mCS (mitochondrial Centum pico-Siemens) (113). This activity was recorded in several labs using mitoplasts from different mammalian tissues and species, such as mice liver, heart, brain, pancreas and adrenal, rat liver, heart and brown adipose tissue, ox heart or mitoplasts from human tissue culture cells (39,114-118). The channel also withstood the dehydration rehydration procedure in reconstituted proteoliposomes containing inner membrane fragments (36,119). However, it was never detected in intact yeast mitoplasts nor in thousands of patches of yeast inner membrane proteoliposomes (32). The channel was slightly anion-selective (PCl-/PK+ = 4.5) and showed a strong voltage dependence. The open probability had a marked sigmoidal profile. It remained closed at negative potentials and the occupation of the open states was favored by positive potentials (30) (the sign of the potential differences is reported respect to the mitochondrial matrix side). The activity of mCS responded to a typical bursting behavior, i.e. many openings are grouped together separated by brief (less than 5 ms) closed periods. This and a detailed kinetic analysis of single channels allowed more than one open and closed states to be ascribed to the mCS conductance (117,120,121). At least two closed and two open states of the same conductance were proposed, as two exponentials were generally needed to fit the open and closed dwell time distributions (26,28). A 50% sub-conductance level was detected (28), as well as infrequent channel openings 1/3 and 2/3 of the most frequent conductance level of 107 pS (117). In addition, it was possible to observe peak transitions as high as 140 pS in the presence of some effectors like amiodarone, (45). All these conductances had the same voltage dependence, and at least the 50 pS substate showed the same anionic selectivity. Whereas mCS was largely unaffected by variations of pH from 6 to 9, on the matrix or cytoplasmic side of the membrane (19), a possible modulation by Ca2+ has not yet clarified. While Sorgato et al. recorded the channel activity without any specific activation procedure, we founded the channel was normally quiescent but could be activated if calcium was chelated from the cytoplasmic side of mitoplasts during the isolation. However, once activated, the activity was not influenced by Ca2+ from 10-9 to 10-5 M (122). An extensive pharmacological study was carried out using a variety of drugs, some of which were interesting because of their interaction with the inner membrane of mitochondria (36,43). Amphiphilic cationic drugs such as amiodarone and propranolol, as well as inhibitors and uncouplers of oxidative phosphorylation, like antimycin A and carbonyl cyanide phenylhydrazones (FCCP) respectively, affected the channel activity. At least the inhibition of mCS by the later two type of effectors of the respiratory chain was attributed to

428

Pablo M.V. Peixoto et al.

binding at a site distinct from that involved with oxidative phosphorylation, as indicated by their reversible effects and differences in IC50 (26,28). Ligands of the mitochondrial benzodiazepine receptor such as Ro5-4864, PK 11195, and protoporphyrin IX, were high-affinity effectors. On other hand, clonazepam, a central benzodiazepine receptor ligand, showed no effect (123). Furthermore, quinine, cyclosporin A, oligomycin, N,N’-dicyclohexilcarbodiimide, and brongkrekic acid had no effect on mCS activity (124). Several substances affecting some plasma membrane channels were tested to establish a possible relationship between mCS and these better known channels, however, none of them modified the channel activity (36). Pharmacology has enabled the elimination of a variety of proteins that might underlie mCS activity. Andreyev et al. postulated that the mitochondrial adenine nucleotide translocator would, in the presence of low concentrations of fatty acids, become a heat generator in muscle as well as in liver, without losing its sensitivity to the classical inhibitors carboxyatractylate and brongkrekic acid (125). However, mCS was not affected by carboxyatractylate and brongkrekic acid indicating ANT was not involved in mCS. Similarly, mCS was not associated with the F0 region of the ATP-synthase as its activity was insensitive to oligomycin and DCCD (36). Klitsch and Siemen reported that mCS was inhibited by submillimolar levels of di- and trinucleoside phosphates, as well as GMP, when added to the outside of patched mitoplasts (117). These results indicated that mCS was not related to the uncoupling protein, thermogenin, which was insensitive to GMP. Furthermore, Inoue et al. (116) reported that mCS was not affected when millimolar Mg2+ or ATP and micromolar ADP was applied on the matrix side of excised patches. The mitochondrial inner membrane anion channel (IMAC) is a channel inferred from light scattering (matrix swelling) studies, which had broad anion selectivity and conducted mono-, di-, and trivalent anions (126). This channel is believed to be an important component of the volume homeostatic mechanism of mitochondria, and was maintained closed or inactive by matrix Mg2+ and H+. IMAC, like many other chloride/anion channels, was reversibly inhibited by DIDS (stilbene-2-2’-disulfonate) (127). The lack of effect on mCS activity of Mg2+ (107), DIDS, quinine or DCCD (dyciclohexylcabodiimide) (36), all of them well known blockers of IMAC (126,128,129) speaked against the identification of the mCS with this particular channel. Despite many observations that appeared to be in conflict with this identification, Borecký et al. [contrary to Sorgato et al. (113)] founded mCS of brown fat mitochondria was sensitive to changes in the matrix pH (130). According to these authors, this was a very important property that mCS and IMAC had in common, suggesting their possible identity. In addition they observed the anion pattern of mCS and IMAC, as that of blockers, were also very similar. It is clear that a

Channel activity of mitochondrial protein import

429

definitive answer to this question will require the protein(s) responsible(s) for both the biochemical IMAC phenomenon and mCS channel must be found. The possibility that mCS might be a stretch-activated channel was also considered. Indeed, they shared some similarities in their conductance value and voltage dependence (131,132). However, the extent of suction on patches affected neither the number nor open times of mCS. In addition, gandolinium, which inhibited mammalian and plant stretch-activated channels (133), had no effect on mCS (36). We are thus left with an open question regarding the role that could tentatively be assigned to this large conductance channel in the inner membrane of mitochondria. 2. The pH sensitive channels (AAA and ACA) Two other low conductance and pH-sensitive channel activities were founded in liver mitoplasts. Both of them displayed greater open probability upon alkalinization of the matrix side of the membrane, and both were activated by depletion of Mg2+ (134,135). One was a 15 pS cation channel while the other was a 45 pS anion channel. They were referred as ACA (Alkaline-induced Cation-selective Activity) and AAA (Alkaline-induced Anion-selective Activity), respectively. However, the selectivity among the ions of the same charge was very slight. AAA was initially thought to correspond to IMAC because of their similar induction by alkaline pH or depletion of Mg2+. An argument against considering IMAC and AAA reflections of the same channel activity came from discrepancies on selectivity. Whereas the ratio of permeabilities to Cl- over glucuronate was reported to be higher than 400 for IMAC (126) it was found to be only approximately 3 for AAA (135). The conductance and voltage dependence of ACA were similar to those of the ATP-sensitive K+ channel (see below). However, unlike the later channel, ACA was relatively non-selective for cations and was not affected by 4aminopyridine and glibenclamide plus ATP. The selectivity and inhibition by Mg2+ suggested that ACA channel activity may correspond to one of the cation uniporters implicated in volume homeostasis, whose existence was inferred from suspension studies (136,137). However, present information on this channel is not sufficient for a firm conclusion. In summary, a clear correlation between electrophysiological and biochemical data has not yet been established for these two alkaline induced channel activities. 3. The K+-selective channel (mitoKATP) While it has not yet been recorded in the native inner membrane, an exquisitely K+-selective channel, with no clear voltage dependence and about 10 pS conductance in 100 mM salt, was observed in large vesicles obtained by

430

Pablo M.V. Peixoto et al.

fusion of liver mitoplasts (116). Some differences in conductance were observed that were either species- or tissue-dependent (138,139). Localization of this channel in the inner membrane was supported by detection of its activity in patches also containing mCS activity. However, confirmation awaits recording this channel in mitoplasts. This small conductance acquired a particular relevance for its sensitivity to drugs and physiological effectors. Thus, the channel activity was inhibited by 4-aminopyridine and glybenclamide plus ATP, blockers of the plasma membrane K+ channel. Although, there was a longstanding suggestive evidence for the existence of an inner membrane protein designed to catalyze electrophoretic K+ uptake into mitochondria (140,141), in this particular case, patch-clamp techniques preceded permeability studies on mitochondrial suspensions in defining this channel activity. Reconstitution studies carried out by Paucek et al. confirmed that the inner membrane contained an ATP-dependent channel (142). These findings were bolstered by evidence that a variety of K+ channel openers and inhibitors influenced mitochondrial function (142,143). An important outcome of these studies was the finding that mitoKATP was the receptor for KATP openers that protect the heart against ischemia-reperfusion injury (143). Today there is some evidence linking the opening of this channel, and not the sarcolemmal variant, to cellular protection against metabolic stress in a variety of tissues, including liver, gut, brain, and kidney, and is thought to be an essential component of the mechanism of ischemic preconditioning in the heart (144-147). Investigation of the mechanism of protection by mitoKATP channel opening has spawned several hypothesis, which in general, are not mutually exclusive and probably all of them contribute to preservation of mitochondrial and contractile function. A consensus is building that mitoKATP opening evokes a response involving several protective mechanisms, including matrix swelling, ROS modulation, and effects on mitochondrial Ca2+ homeostasis (144, 148). Again, structural information on this particular channel is scarce. Only a few studies have tried to identify the putative mitoKATP subunits. A 53 kDa protein was tentatively identified by Diwan as a component of this channel (141), and a 55 kDa protein has been reported by Mironova et al. (140,149). Although not conclusive, these data, together with the similar pharmacology of known plasma membrane KATP, suggest that mitoKATP might be composed of an inward rectifier potassium channel subunit (Kir) (150,151) in association with a sulfonylurea receptor (SUR) (152,153). The absence of a clear molecular definition of the mitoKATP components has hampered structure-function studies of this mitochondrial channel. For this reason, determination of the structure of the pore-forming protein(s) underlying mitoKATP conductance will undoubtedly resolve many of the yet unanswered questions.

Channel activity of mitochondrial protein import

431

4. The mitochondrial Ca2+ uniporter Considerable evidence indicates that mitochondria play a key role in physiological intracellular Ca2+ homeostasis (137,154-158). Intramitochondrial Ca2+ homeostasis is controlled by separate Ca2+-uptake and release pathways. Early studies established that Ca2+ uptake into mitochondria was mediated by a transport system that may have the properties of a channel. This system was defined as the Ca2+ uniporter (159,160). Thus, the rapid uptake of cytoplasmic Ca2+ by mitochondria during intracellular Ca2+ signaling is mediated by this uniporter (161,162), located in the inner membrane (163), which permits transport of the ion down its electrochemical gradient. A vast amount of information concerning the functional properties of this mitochondrial transport system is available (137,161). However, there is very little information in relation to its structural features, despite an enormous amount of literature and the importance of the problem. Recently, Kirichok et al. patch clamped mitoplasts isolated from COS-7 cells to measure whole-mitoplast and single channel Ca2+ currents (25). They identified a previously unknown Ca2+-selective ion channel that is sensitive to inhibitors of mitochondrial Ca2+ uptake (i.e. nanomolar concentrations of rhuthenium 360 blocked the channel). The single channel conductance is small and has multiple subconductance states between 2.6 pS and 5.2 pS (in 105 mM CaCl2). The current measured as a function of holding voltage is inwardly rectifying. The open probability is close to 100% at –200 mV and declines to 10% at –80 mV, making this channel suitable for Ca2+ uptake into energized mitochondria. It is a Ca2+ selective channel, even at low cytoplasmic [Ca2+] (dissociation constant ≤ 2 nm), that preferentially conducts Ca2+ into mitochondria (selectivity Ca2+ ≈ Sr2+ >> Mn2+ ≈ Ba2+). Also, the channel can provide sufficient current densities to explain mitochondrial Ca2+ uptake by the uniporter. After a detailed electrophysiological characterization, these authors have concluded that the properties of the current mediated by this novel channel are those of the mitochondrial Ca2+ uniporter. In agreement with this conclusion, we have to consider that the main difference between carriers and ion channels resides in just a lower turnover number of the former. It is known, for instance, that many integral membrane proteins, biochemically defined as carriers may display “channel-like” behavior under some conditions. Such behavior has been reported for the (Na+,K+)-ATPase (164), the asparte/glutamate carrier (165), the phosphate carrier (166,167), the uncoupling protein UCP1 (168) and the ADP/ATP translocator (169). The channel behavior described in these cases would appear to be an observation of little or no physiological relevance. However, there is mounting evidence suggesting that altered carrier function could be the underlying mechanism of pathophysiological situations such as the opening of the permeability transition (see below) or the carrier-mediated fatty acid uncoupling (170). In addition, it

432

Pablo M.V. Peixoto et al.

should be mentioned that the particular complexity of Ca2+ movements in and out of mitochondria has been responsible for underestimation of the kinetics of the uniporter. Indeed, in respiring mitochondria the kinetics of Ca2+ uptake becomes rapidly limited by the rate of H+ pumping as [Ca2+]o is raised above ~10 µM, with the consequent underestimation of both the Vmax and the apparent Km for Ca2+ (171). Even more, if a high turnover is prima facie evidence of a channel, the opposite is not true, i.e. a channel may approach the kinetic behavior of a carrier if it goes through several conformational states. It is surprising however, that under their experimental conditions, only small channel activity was reported by these authors, without records of any of the large conductance channels that are abundant in the inner membrane of mitochondria, e.g., mCS and Tim channel activities (see below). While the electrophysiological characteristics of this calcium channel match the pharmacology and kinetic aspects for identification as the calcium uniporter, molecular and immunological studies are needed to validate this assignment. 5. The ADP/ATP translocator (ANT-linked channel) Like the Ca2+ uniporter, there is evidence indicating the mitochondrial ADP/ATP carrier or Adenine Nucleotide translocator (ANT) can also display channel activity. Patch-clamp experiments of purified ANT from bovine heart behaved as a large Ca2+ dependent channel when reconstituted into liposomes (169). According to these authors, the mean conductance of the open state was 500-700 pS (in symmetrical 100 mM KCl) with several substates. The channel was slightly cationic selective (PK+/PCl- = 4.3) and reversibly activated by Ca2+. It was inhibited by bongkrekate and ADP, but remained insensitive to carboxyatractylate, in accordance with the known effects of these specific ligands on the ANT. Also, they reported a distinctive kinetic response to holding potentials. When the voltage was raised up to 80-100 mV of either polarity, the channel was in the open state without current fluctuations to the close or sublevels states. At higher positive potentials rapid gating transitions to the lower conductance substates could be observed, whereas at high negative holding potential the conductance remained unchanged. The same results were obtained if the reconstituted system contained recombinant ANT from Neurospora crassa expressed in exclusion bodies of E. coli (172). In this case the preparation was free from residual mitochondrial components and presumably free from other possible contaminating channels. The study was extended to investigate whether cyclophilin (a protein binding cyclosporine A) influenced the ANT-linked channel making it sensitive to cyclosporin A. Since the immunosuppressive cyclosporin A was established as the most specific inhibitor of the mitochondrial permeability transition pore (PTP) (see below), the clear intention of these experiments was to find a possible correlation between both of these entities. Whereas cyclophilin

Channel activity of mitochondrial protein import

433

increased the probability of channel detection and at the same time decreased the probability of voltage gating under high positive holding potentials, cyclosporin A added alone had no effect on the channel activity. However, when both of these agents were added together the suppression of voltage gating by cyclophilin was prevented. The authors reasoned that ANT could be a conducting component of PTP. However, several issues need to be addressed before a firm conclusion can be reached. For instance, the effect of cyclophilin was transient, detected only in the first 20-25 min. after the addition. Also, in agreement with the effect on PTP, an inhibition of ANT (in 150 mM KCl) by cyclosporine A (with or without cyclophilin) was expected. Instead, this immunosuppressive drug just reverted the effect of cyclophilin, which in turn increased the open probability of the channel. Even more, recent results by Kokoszka et al. with knock out mice confirm that the ANT is not an essential component of PTP, although it does have an essential role in its regulation (see below) (163). 6. The permeability transition pore The permeability transition (PT) was originally described in the 1970s (173,174) as a reversible phenomenon in which mitochondria become freely permeable to solutes of less than ~1500 Da. Ca2+, Pi, and many oxidants induce this increase in permeability, while Mg2+, ADP, and H+ inhibited the onset of the PT. The immunosuppressive cyclosporin A was an effective and specific inhibitor of the PT at submicromolar concentrations (175). The saturable nature of the inhibition implied that a protein channel or pore in the membrane mediated this phenomenon. The pore had a big diameter (~3 nm) and therefore the conductance was high. Opening of a single pore could be sufficient to uncouple oxidative phosphorylation, release some intramitochondrial solutes, and induce large-amplitude mitochondrial swelling. The PT has been considered a pathophysiological mechanism required for cell killing. There is renewed interest in PT as it has been implicated not only in determining whether cells live or die, but also whether death occurs by necrosis or apoptosis (176,177). Patch-clamping mitoplasts from diverse sources, including several mammalian tissues (31,45,115) and human cell lines (114) revealed a large conductance channel, up to ~1200 pS (in 150 mM KCl), composed of several subconductance states from 10 to 1000 pS. In retrospect, the channel activity previously referred to as the Multiple Conductance Channel (MCC) (115) [also called MMC for mitochondrial megachannel by Petronilli et al. (118)] in early studies actually arose from two different channels, i.e., the Permeability Transition Pore (PTP) and the Translocase of the Inner Membrane (Tim) channels (see below). While Tim channel activity crosses all species examined thus far, the PTP does not exist in yeast (32). Below is a description of the

434

Pablo M.V. Peixoto et al.

retrospective analysis of MCC/MMC behavior, which is presumably the electrophysiological representation of the PTP. The PTP was a high conductance channel of ~1.2 nS with multiple subconductance levels. The channel was slightly cation selective (PK+/PCl- is 3-6 depending on the species (32) and could be induced by elevated Ca2+ levels (122,178) or applied potentials of greater than 60 mV of either sign (31,124). Once activated, the PTP activity persisted at all voltages. The size of current transitions from low to high-conductance states was progressively achieved upon application of voltage (31). The high-conductance states were often occupied for long periods of time (seconds to minutes), especially at negative voltages of >50 mV) (124). The channel occupied lower conductance levels at low positive potentials. This voltage profile was typically observed in mammalian mitoplasts (37). PTP is thus a very high conductance channel whose uncontrolled opening should decrease coupling of oxidative phosphorylation. Therefore, it is not surprising that PTP activity was influenced by a variety of physiological effectors. These included divalent cations, pH, ADP and voltage (107). Several pharmacological agents affected PTP, in electrophysiological experiments including antimycin A (28), cyclosporin A (178), specific ligands of the mitochondrial benzodiazepine receptor like Ro5-4864, and protoporphyrin IX, but not ligands of the central benzodiazepine receptor like clonazepam (123), and the uncouplers of the oxidative phosphorylation CCCP and FCCP (26). In addition, the list of compounds affecting PTP activity included the amphiphilic cations amiodarone, propranolol and quinine, as well as dibucaine [see Campo et al. 1998 for summary (43)]. The lack of information about the molecular basis of the PTP has made difficult to extend these studies. The dominant hypothesis is that the PTP spans the two mitochondrial membranes and is composed, in part, of proteins from both of these membranes (137,179). Among the proteins proposed to be part of PTP are VDAC of the outer membrane and the ANT of the inner membrane (172). Binding to the matrix protein cyclophilin D, is required for the Ca2+-dependent onset of the permeability transition (PT). Cyclophilin D also confers sensitivity of the PTP to cyclosporin A, an effective and specific inhibitor at submicromolar concentrations. Other proteins forming the PTP may include hexokinase from the cytoplasm and creatine kinase at the surface of the inner membrane. However, the recent results by Kokoszka et al. (163) have shown that ANT is not an essential component of PTP, although it does have an essential role in its regulation. Our earlier electrophysiological results with carboxyatractyloside are consistent with these findings (180). Mitochondria from mice liver lacking the two isoforms of ANT could still undergo a PT that is inhibited by cyclosporin A and induced by addition of Ca2+, even though more Ca2+ was required. In

Channel activity of mitochondrial protein import

435

addition, hepatocytes remained competent in responding to apoptotic stimuli (163). Future studies are needed to clarify the electro-physiological data with that of the PTP. In particular, molecular and immunological identification of components would advance the field. Nevertheless, we then focused our attention to the possible involvement of the high conductance channel activities in mitochondrial protein import.

V. Protein translocation membranes

across

mitochondrial

About half of the proteins synthesized in the cytoplasm must be translocated across at least one membrane before reaching their final destinations. Thus, sorting of proteins is an essential process for many vital cellular functions such as signaling, secretion, compartmentalization, differentiation, or apoptosis. Mitochondria are of particular interest to this process, due to their elevated protein content (15-20% of all cellular proteins) and the fact that the small mitochondrial genome encodes only a few inner membrane proteins [i.e. 13 in human cells and only 8 in yeast (181,182)]. These proteins, in fact, add to the complexity of protein trafficking pathways in mitochondria (see below). Also, two membranes confine four different mitochondrial compartments where proteins can be directed. In addition, the need to maintain a low permeability for coupling oxidative phosphorylation must be balanced with the passage of large molecules, presumably through water-filled channels. How does this transport process actually take place; how are mitochondrial proteins targeted specifically to mitochondria as opposed to other membranes present in the cytosol; or how does a cell control its overall protein traffic? These questions represent a major challenge to modern cell and molecular biology. The yeast Saccharomyces cerevisiae is a useful experimental model for the study of mitochondrial protein import. This is primarily due to the well developed genetics of this organism, the ease by which it can be transformed, and the possibility to isolate large quantities of cells for biochemical examination. It is likely, however, that similar mechanisms function in human mitochondria, since their precursor proteins can be efficiently imported into yeast mitochondria and human homologues of yeast mitochondrial protein translocases have been identified (183-186). The importance of this process is demonstrated by the fact that most mitochondrial proteins essential for yeast viability are involved in the import of proteins into the mitochondria. A more complete picture of the mitochondrial protein import pathways has emerged over the years [for a review see (187-191)]. Figure 2 represents an overall view of these pathways.

436

Pablo M.V. Peixoto et al.

Figure 2. The protein import machinery of mitochondria. The illustration shows the three multisubunit complexes that are responsible for mitochondrial protein import including the TOM complex in the outer mitochondrial membrane (MOM) and the TIM23 and TIM22 complexes in the inner membrane (MIM). The protein constituents of these complexes are referred to as either Tom or Tim followed by their molecular weight. Precursor proteins with mitochondrial targeting peptides bind to specific receptors, e.g. Tom70, Tom22, Tom20, which direct import via Tom40, the pore forming protein of the translocase of the outer membrane. Precursor proteins are translocated as unfolded, linear polypeptides. Proteins with N-terminal presequences are directed to the TIM23 complex for import across the inner membrane. The proteins Tim23, Tim17, Tim44, Tim50, and Tim14 are part of the TIM23 complex. Translocation of precursor proteins across the inner membrane requires a membrane potential and the ATP-driven action of Hsp70. Precursor proteins with internal targeting signals (blue) are guided to the TIM22 complex by chaperones in the intermembrane space (IMS), e.g., Tim9, Tim10. Tim22, Tim18 and Tim54 are components of this complex.

A. Delivery and targeting of proteins to mitochondria The transfer of mitochondrial proteins from their site of synthesis in the cytosol to the surface of the mitochondrial outer membrane, where translocation begins, preferentially although not exclusively, occurs in a posttranslational manner (192-194,194). Since mitochondria can only import unfolded or loosely folded proteins, the cell has to ensure that the precursor proteins remain in a transport-competent state, typically a loosely folded

Channel activity of mitochondrial protein import

437

conformation, on the route from the cytosol to their final destination inside the organelle. In addition, the extensive hydrophobic regions of some proteins, that will eventually form transmembrane spans, have to be shielded in the aqueous environments of cytoplasm or intermembrane space, in order to prevent protein aggregation. Cytosolic factors like chaperones assist the import of the polytopic membrane proteins preventing aggregation, misfolding or proteolysis, and helping in recognition and transfer of cargo proteins across membranes (195,196). Information for targeting to mitochondria is contained in the mitochondrial proteins themselves (197). In principle, there are two types of signals that direct proteins to mitochondria: amino-terminal (or presequence signals) and internal signals. The common presequence signal consists of the first 10-40 amino acids of the protein. It is a cleavable sequence containing many positively charged, hydrophobic and hydroxilated residues (198). A characteristic feature of presequences is their probability to form amphipathic α-helices with one positive and one hydrophobic face (199). Most mitochondrial matrix proteins utilize this type of amino-terminal signal. Some proteins that are destined for the other mitochondrial compartments also use variations of presequences, such as a hydrophobic sorting sequence, adjacent to the presequence, acting as a stop signal across the inner membrane (200-202). Approximately 30% of mitochondrial targeted proteins do not contain recognized presequences. Instead, they carry internal signals, which are less well characterized. In this case, the targeting information is spread throughout the length of the protein (190,203205). Many multiple membrane spanning metabolite transporters of the inner membrane, a number of intermembrane directed proteins, as well as most mitochondrial outer membrane proteins belong to this group. Three protein complexes or translocases, one located in the outer membrane and two in the inner membrane, are the main machineries for specific recognition, sorting and translocation of mitochondrial precursor proteins. The translocase of the outer membrane or TOM complex transports all precursor proteins destined to the IMS, matrix or MIM. The presequence translocase of the inner membrane (TIM23 complex) translocates preproteins with cleavable presequences. The protein insertion complex of the inner membrane (TIM22 complex) mediates the insertion of hydrophobic membrane proteins that carry internal targeting signals. While the import machineries of the outer and inner membranes can operate independently, current models favor their transient linkage at contact sites (206-208).

B. Translocation across the outer membrane: The TOM complex The TOM complex contains at least seven different integral membrane proteins that form a large, dynamic complex. The proteins Tom20, Tom22, and

438

Pablo M.V. Peixoto et al.

Tom70 function as receptors. Presequences of cleavable precursor proteins are first recognized by Tom20 (190,199,209), while precursor proteins with internal targeting signals are typically recognized by the receptor Tom70, which forms a dimer (190) and might possess chaperone like properties (203). Tom22 functions as a central receptor that accepts precursor proteins from both of these receptors and mediates their transfer to the channel (210). The protein-conducting channel across the outer membrane is formed by Tom40 (211,212). The import pathways of all mitochondrial proteins converge during passage across this channel. Three small Tom proteins –Tom5, Tom6 and Tom7- complete the stable core of this complex. Tom5 is involved in the transfer of all types of precursors from Tom22 to the Tom40 channel (213). It might also have receptor properties for some intermembrane proteins (214). Tom6 functions as an assembly factor of this translocase, whereas Tom7 favors its dissociation (215). Preproteins with a presequence are initially recognized by Tom20, transferred to Tom22, Tom5, and Tom40, and translocated across the outer membrane. When the presequence emerges in the intermembrane space side of the TOM complex, it can bind to the carboxy-terminal domain of Tom22 before transfer to the inner membrane. Precursor proteins with internal signals are initially bound by multiple molecules of the receptor Tom70. They are transferred in an ATP-dependent step to Tom22 and Tom5 before they are inserted into the pore (216,217). The precursors of the outer membrane proteins, such as VDAC and Tom40, with multiple β-strands are first imported through the Tom complex to the intermembrane space side (216), where they follow a bacteria-like pathway of export and integration into the outer membrane. In this pathway, the small Tim proteins of the intermembrane space (see below) pass the precursors to the so-called SAM complex (Sorting and Assembly Machinery). Mas37 (218) and Sam50 (219-221) are identified components of this complex and are highly conserved throughout evolution.

C. Translocation across the inner membrane: The TIM23 complex After guidance across the outer membrane by TOM complex, preproteins with presequences interact with the TIM23 complex of the inner membrane. The core of the TIM23 complex consists of the integral membrane components Tim23, Tim17, and Tim50. Tim23 and Tim17 are encoded by essential genes. Although their membrane domains are homologous, they cannot substitute for each other. Tim23 consists of four predicted transmembrane helices and forms a channel across the inner membrane. Channel formation by Tim23 was demonstrated directly by reconstitution of the purified protein into liposomes (222). The intermembrane space domain of this protein extends into the outer

Channel activity of mitochondrial protein import

439

membrane (223), providing a mechanism for the two translocases (TOM and TIM23) to get into close contact. At the same time, this domain is important for recognizing the presequences as they emerge form the TOM complex (224,225), as well as for the high cation selectivity of the channel. Tim23 is tightly associated with Tim17. The exact role of Tim17 remains unknown, but it has been proposed to influence the channel activity (190). Tim50 exposing its large C-terminal domain to the intermembrane space is believed to have a role in guiding the proteins through the intermembrane space from the TOM complex to the channel of TIM23, (226-228) and to be crucial for generation of interactions between the TOM and TIM23 complexes (12). The TIM23 translocase consists of the inner membrane integrated complex and a peripherally attached import motor at the matrix side. This motor is composed of the peripheral membrane protein Tim44 that recruits the mitochondrial mtHsp70 and its cofactor Mge1 (229,230). All three are essential for viability. mtHsp70 binds extended segments of preproteins as they emerge on the matrix side, and is powered by the binding and hydrolysis of ATP. Tim44 functions as a membrane anchor for the ATPase domain of mtHsp70 directly at the exit site of the import channel (231), whereas the peptide-binding domain of mtHsp70 interacts with the preprotein in transit. Mge1 is a nucleotide-exchange factor that promotes the reaction cycle of mtHsp70. Recently another protein, Tim14, anchored to the inner membrane by a single transmembrane segment and exposing a J-domain to the matrix side, has been identified as a new component of the mitochondrial import motor. Tim14 interacts with Tim44 and mtHsp70 in an ATP-dependent manner thereby activating mtHsp70 to allow rapid and efficient trapping of the precursor proteins (232,233). Additionally, a novel essential co-chaperone, Tim16, forms a stable complex with Tim14 (234). The function of mtHsp70 in holding preproteins upon their exit from the import channel is thus coordinated by the membrane bound co-chaperones Tim44, Tim14, and Tim16 (235). The molecular mechanism of the import motor probably involves both passive trapping and active pulling of the preproteins (194,236). Once precursors emerge in the mitochondrial matrix, their amino-terminal presequences are generally cleaved off by the dimeric mitochondrial processing peptidase (MPP), a zinc-dependent metalloendopeptidase (237,238).

D. Insertion of proteins in the inner membrane: The TIM22 complex The TIM22 import machinery mediates the biogenesis of polytopic inner membrane proteins, such as those of the mitochondrial carrier family (239), and the import components Tim17, Tim22, and Tim23 (187). In yeast, the TIM22 import machine is formed by an inner membrane complex of 300 kDa,

440

Pablo M.V. Peixoto et al.

consisting of the membrane components Tim22, Tim18, and Tim54 (240-243) with the peripheral membrane protein Tim12p and small amounts of Tim9 and Tim10 (243-246). Tim22 is the core component of this complex. It is structurally related to Tim23 and Tim17, and like them is an essential protein for yeast viability. When isolated and reconstituted into liposomes, it forms a voltage-gated channel (247). Tim54 exposes a large domain to the intermembrane space and is required for the stability of Tim22 but may not interact directly with the translocating proteins (241). Tim18 is structurally related to subunit IV of the succinate dehydrogenase and it may be involved in the assembly and stability of the TIM22 complex (240,242). Small Tim proteins including Tim8, Tim9, Tim10, and Tim13, in the intermembrane space, are part of the TIM22 import pathway. They are thought to function as chaperones that guide hydrophobic membrane proteins across the aqueous environment of the intermembrane space and maintain them in an import-competent conformation. These small proteins assemble in a particular manner. Tim9 partners with Tim10 and Tim8 partners with Tim13 to form 70 kDa complexes of equimolar ratios (240,244,248,249). The Tim9-Tim10 complex interacts with the membrane spanning domains of carrier protein precursors as they emerge from the TOM complex and delivers them to the TIM22 translocase for insertion into the membrane, probably via interaction with Tim12 (203,239,243-246,250,251). The homologous yet non-essential Tim8-Tim13 complex functions during the import of different classes of proteins entering both mitochondrial membranes, such as the inner membrane Tim23 (248,252,253) and the outer membrane Tom40 (254). The Tim8-Tim13 complex traps the incoming precursors in the intermembrane space thereby maintaining them in a translocation competent conformation. Many mitochondrial proteins inherited from the prokaryotic progenitor cell are inserted into the inner membrane in an export step following translocation into the matrix. Also, most mitochondrially encoded proteins are hydrophobic membrane proteins, which are integrated into the lipid bilayer during their synthesis on mitochondrial ribosomes. The mitochondrial inner membrane possesses a further protein insertion machinery, for proteins that are exported from the matrix into the inner membrane (255,256). This pathway bears similarities to Sec-independent protein export in bacteria and requires the Oxa1 protein. Oxa1 is a component of a general protein insertion site in yeast mitochondrial inner membrane used by both nuclear and mitochondrial DNA encoded proteins. Oxa1 is a member of the conserved Oxa1/YidC/Alb3 protein family found throughout prokaryotes and eukaryotes (where it is found in mitochondria and chloroplasts) (257,258). The mitochondrial machinery for protein import and assembly has proven to be complex. Variations or different combinations of the two general pathways outlined above are likely to occur, and one can envision a network of protein import pathways. The tools of molecular biology, genetic engineering

Channel activity of mitochondrial protein import

441

and gene technology have been traditionally used to study protein sorting into mitochondria at the molecular level. The innovation of our approach to this problem consisted in the introduction of electrophysiological techniques.

VI. In search for the pore components of the mitochondrial protein translocases In 1975, Blobel and Dobberstein proposed the formation of a transient tunnel in the membranes of the endoplasmic reticulum, through which the nascent protein chain would be transferred in light of theoretical considerations of the “signal hypothesis” (259). Experimental evidence compatible with a protein-conducting channel came from biochemical studies of signal sequence insertion and translocation (260). The proof of the existence of large aqueous channels in membranes derived from rough endoplasmic reticulum was obtained years later with electrophysiological techniques (261,262). Around the same time a general protein import pathway in mitochondria was outlined and followed the same steps (263). The existence of a specific site or component that performed the membrane insertion of precursor proteins was then proposed and termed “general insertion protein” (GIP) (264).

A. The protein import pores of the TIM23 and TOM complexes If channels are integral to the process of protein translocation, the evidence that signal peptides were sufficient for opening the protein-conducting channels in E. coli (265) represented a turning point for investigations on channels or pores involved in protein import. Thus, the most significant characteristic of the cationic channel described by Henry et al. in the outer membrane of mitochondria was the blockade induced by a 13-residue synthetic peptide with the sequence of the amino terminal end of a mitochondrial directed protein (subunit IV of yeast cytochrome oxidase) (92). They named this channel PSC (Peptide Sensitive Channel) and suggested it could be involved in protein translocation (266). As discussed below, PSC is now referred to as the Tom channel. Similarly, the first link between the high conductance channel activity of mitoplasts (previously referred to as MCC in yeast and now as the Tim23 channel) and protein import across the inner membrane was made as synthetic peptides whose sequence mimic signal peptides, transiently blocked this channel, either in mouse or yeast mitochondria. These peptides caused a momentary closure of Tim23 channel; that is, a reversible, voltage-, and dose-dependent flicker blockade. The same peptides had no effect on two other mitochondrial channels: mCS in the inner and VDAC in the outer membrane. In contrast, similar peptides whose sequences do not support protein import had no effect on the Tim23 channel

442

Pablo M.V. Peixoto et al.

(267). The specific interactions of signal peptides with the Tim23 channel suggest that this channel could be involved in protein import across the inner mitochondrial membrane (268). In addition, as expected from the low permeability of the mitochondrial inner membrane, Tim23 channel is normally closed in mitoplasts under metabolizing conditions, but it is opened by µM concentrations of signal peptides (269). Besides modulation of the opening of the protein import pores by signal peptides, other predictions can be made for the characteristics of these import pores based on the general protein translocation scheme. For example, it is expected they shall have diameters consistent with their ability to allow the selective passage of unfolded polypeptides (see below). Also, structural modifications of the translocases like mutations or controlled proteolytic digestion, as well as antibodies against some of their components that modify protein import, may alter the electrophysiological properties of the channels involved. Experimental approaches based on these criteria have been used to evaluate if these high conductance activities, the two channels previously identified by electrophysiological techniques, could be assigned to the corresponding pores of the TIM23 and TOM complexes respectively. Several lines of evidence support this notion. We founded antibodies against Tim23 that inhibit protein import, at similar concentrations blocked the conductance through MCC but not PSC in yeast (267). Pre-immune serum, antibodies against VDAC, and antibodies against an iron sulfur protein of the inner membrane had no effects. Furthermore, a point mutation on Tim23 was associated with inhibition of protein import (270) and at the same time with the loss of regulation of the mitoplast channel by signal peptides (267). More recently, expression of Tim23 was down-regulated if a yeast strain carrying a Gal promoter before the Tim23 gene is grown in the absence of galactose (271). No channel activity was observed if proteoliposomes containing Tim23 depleted membranes were patch-clamped (unpublished results of Martinez, Kinnally and Campo). The identification of the outer membrane activity (PSC) as the protein import pore of the TOM complex is also supported by many experimental observations. Immunoprecipitation of extracts with antibodies against Tom40 correlated with loss of this channel activity in bilayer experiments (272) and antibodies against Tom40 modified the conductance (272,273). Significantly, channel activity similar to that of outer membranes was detected in bilayers upon incorporation of purified TOM complex and of bacterially expressed Tom40 (211,212,273). Hence, it has been proposed that the channel activities of previously referred to as MCC and PSC of yeast are identical to the pores of the TIM23 and TOM complexes. Therefore, MCC and PSC are now designated the Tim23 and Tom channel activities, respectively (274).

Channel activity of mitochondrial protein import

443

1. Architecture of the Tim23 and Tom channel pores Channels putatively involved in protein import are expected to have large pore diameters, which would enable the translocation of precursors across membranes. The pore diameter of a channel can be estimated from its conductance, which is proportional to the resistance to ion flow. According to Hille, R = ρ(L / πa2), were R is the resistance of the channel and ρ the resistivity of the solution, L the pore length and a the pore radius (27). Electrophysiology can be used to probe the dimensions of the Tim23 and Tom channels. Simplistically, a single pore diameter of 2.4 to 2.7 nm can be estimated from the peak conductance of the Tim23 and Tom channels, assuming a pore length of 5.5 to 7 nm, the average thickness of the outer membrane (275). However, a closer look at the single channel behavior for both pores reveals a predominant substate state of 500 pS consistent with two semi-pores of 1.7 to 1.9 nm gating cooperatively as a double barrel pore. Analysis of the open probabilities is consistent with the prediction of a doublebarrel pore structure for the Tim23 and Tom channels. Histograms compiling the percentage of the total time spent in each current level reflect the probability of any of the conductance states of a given channel; each of which can be accounted as defined peaks in these histograms. For a single channel, the probability of being in the closed state is 1-Po, if Po is the probability of being in the open state. The total current histograms typically fit a binomial distribution, providing transitions between the open and closed states are independent. In fact, a binomial distribution converges into Gaussian when the probability of a given state is high. In the same way, two or more independent channels present within the same patch can be revealed if the current recordings follow a binomial equation {f(x) = [n!px(1-p)n-x]/x!(n-x)!)} where n is the number of channels, x is the number of channels in the open state, and p is the open probability (27). Both Tim23 and Tom channels have a substate with a conductance (500 pS) half of the open state. Occupation of the open, half-open and closed states does not follow a Gaussian distribution for either Tim23 or Tom channels except at low potentials, consistent with double barrel pore that gates cooperatively. Figure 3 shows typical total current histograms obtained at two voltages. Clearly, the distribution of any defined channel state only followed a Gaussian distribution when the open probability was high (i.e. at +10 mV). At higher voltages, as the open probability decreased and more conductance states became apparent (i.e. at +30 mV), deviations from binomial distributions support the hypothesis that Tim23 and Tom channels are double barrel pores that cooperatively gate. Consistent with the predictions of a double-barrel pore structure based on single channel behavior, single-particle electron micrographs of purified TOM complex showed two, and possibly three, apparent ~2 nm holes (212).

444

Pablo M.V. Peixoto et al.

A second approach to estimate pore sizes comes from the transport of rigid molecules of increasing diameters like polymers, i.e. polyethylene glycols and dextrans, or gold particles (276). In essence, if the rigid molecule is smaller than the pore it permeates throughout the pore displacing the normal conducting ions, thus reducing the channel conductance. On the other hand, if the molecule is not permeable, it does not significantly decrease the channel conductance. With this method the predicted diameters of the pores of the Tim23 and Tom channels were 1.8 and 2.0 nm, respectively (manuscript in preparation). These results indicate the aqueous pathways through both channels are similar, although the pore of the Tim23 channel may be slightly smaller than that of the Tom channel. Importantly, the inferred pore diameter of both channels is sufficient to allow the passage of unfolded polypeptides, regardless

Figure 3. Total amplitude histograms for the TIM and TOM channels can deviate from predicted binomial distributions. Total amplitude histograms are shown for TIM (A at +20 mV; B, +40 mV) and TOM (C, +20 mV; D, +30 mV) and TOM (C, +20 mV; D, +30 mV) channels for experimental data (black) and data simulations (gray) fit to the probability of occupying the 1 nS level, assuming two independent channels and a binomial distribution. Simulations were generated by Electrophysiology Data recorder V-2.2.3. software (J. Demspter, University of Strathclyde) after providing open probability, transition amplitude, mean open and mean closed time, with five openings per burst for each data set. Interval durations were binned at a resolution of 200-500 bins. Histograms are not leak subtracted. Taken form Muro et al. with permission (38).

Channel activity of mitochondrial protein import

445

whether they are single- or double-barrel pores. Several mitochondrial channels, including the mitochondrial centum-picosiemen channel (mCS), the alkaline pH-activated anion and cation channels, and the ATP-sensitive K+ channel, were eliminated as potential candidates for protein import pores since their estimated pore diameters would limit the passage of polypeptides. Finally, it is worth noting other channel activities implicated in protein import in the endoplasmic reticulum and E. coli, have the same conductance, and therefore similar size as the Tim23 and Tom channels (262,265,277,278).

B. The Tim23 and Tom channels are the expression of distinct molecular entities As reviewed above, electrophysiological techniques applied to mammalian and yeast mitochondria have revealed the presence of large conductance channels in both membranes of this organelle. Under normal conditions, the outer membrane contains the Tom channel and Tim23 channel is present in the inner membrane. These two mega-channels share many general characteristics. Aside from their apparent distinct distribution in mitochondrial membranes, the similarities between the electrophysiological properties and behavior of these two channels made it difficult to clearly discriminate between them. In particular, Tim23 and Tom channel activities are sensitive to signal peptides and have the same voltage dependence and conductance. Therefore, we undertook a direct comparison of the Tim23 and Tom channel activities, under identical conditions and eliminated the possible contribution of VDAC to these activities. 1. Purity of the isolated mitochondrial membranes One of the main concerns when working with isolated outer or inner membranes reconstituted into proteoliposomes is the purity of these preparations. The issue becomes not trivial when trying to assign a defined location for a given channel, and the problem augments by the existence of “contaminating” contact sites between the two membranes. As stated above, contact sites where the two mitochondrial membranes are closely and tightly apposed, have long been known to exist. They are involved in protein precursor uptake and energy transfer. Hexokinase and creatine kinase, as well as the outer membrane VDAC, and the inner membrane ANT are thought to be part of these dynamic structures (279). As apparent from morphological electron microscopy studies, the swelling and shrinking treatment to obtain mitoplasts, though severe and prolonged, is unable to completely remove the outer membrane (119). Outer membrane fragments of different extensions remain firmly bound to the inner membrane. In our hands, the French press treatment of mitochondria gives cleaner mitoplasts. Figure 4 shows the cross contamination of mitochondrial inner membrane (MIM) with outer membrane

Pablo M.V. Peixoto et al.

446

(MOM). Western blots of 80 µg MIM and varying amounts of MOM were decorated with antibodies against Tom40, a protein of the translocase complex located in the outer membrane of mitochondria. Quantitative analysis of these blots with Scion Imaging (Frederick, MD) software allows one to establish the purity of our membranes at approximately 90-95%. MIM 80

MOM 10

15

25

30

µg protein

O.D.

2

Tom40

1

0 0

10 20 Total protein (µg)

30

Figure 4. Purity of the mitochondrial inner membrane preparation is high. Western blot using antisera against Tom40, a protein of the outer membrane, show minimal cross-contamination of the inner membrane (MIM). Increasing amounts of outer membranes (MOM) were placed on the SDS-gels. Densitometric analysis of the westerns with Scion Imaging (Frederick, MD) software quantified the membranes were over 92% pure.

2. Electrophysiological comparison of the Tim23 and Tom channel activities We conducted a comparative study of the electrophysiological characteristics of the Tim23 and Tom channel activities and their responses to signal and non-signal peptides (38). The outer and inner membranes separately fused with liposomes by the dehydration-rehydration procedure described before, and the resulting proteoliposomes were patch-clamped in 150 mM KCl, under symmetrical conditions. Importantly, a yeast strain lacking VDAC isoforms 1 and 2 was used to probe the distinct identity of VDAC from the Tim23 and Tom channels while reducing the background channel activity of the preparations in these studies. The single channel behavior of the Tim23 and Tom channels are remarkably similar as shown in Table 1 and Figure 5. Typical current traces of the Tim23 and Tom channels at various voltages are shown in Figure 5. Both channels tended to open at negative potentials, while they occupy lower conductance levels at positive potentials e.g. 30 mV, as shown in the open probability plots of Figures 5C and 5D. The most significant feature was however, the similarities between the two channel activities.

Channel activity of mitochondrial protein import

447

Table 1. Electrophysiological comparison of Tim23 and Tom channel activities

A

B

Tim23

30 mV

Tom

o s c

10 mV

o

-10 mV -20 mV

o s o

-30 mV

s o 50 ms

D

Po

C

Po

10pA

20 mV

o s

1

1

0.5

0.5

-80

-60 -40 -20 0 20 Voltage (mV)

40

60

80

-60

E

-20 0 20 Voltage (mV)

40

60

F

10

10 Ln [Po/(1-Po)]

Ln [Po/(1-Po)]

-40

0

-10

0

-10 -20

0

20 40 Voltage (mV)

60

-20

0 20 40 Voltage (mV)

60

Figure 5. Single channel properties of Tim23 and Tom are identical. (A, B) Current traces of Tim23 and Tom channels are shown at various voltages with 2 kHz filtration in symmetrical 150 mM KCl, 5 mM HEPES, pH 7.4. (C, D) The Tim23 and Tom channels have the same voltage dependence. The open probabilities (Po) were determined at various voltages for recordings of 30 seconds duration of Tim23 and Tom channels respectively. (E, F) Gating charge is proportional to the slope of Ln[Po/(1-Po)] versus voltage plots and show that the gating charge of Tim23 (E) and Tom (F) are similar. Po was determined from the occupation of the 1000 pS level of total amplitude histograms. Points were the mean ± SD of at least eight independent patches.

Pablo M.V. Peixoto et al.

448

A closer look to the properties of these channels confirmed this observation. As shown in Tables 1 and 2, from conductance to single channel analysis, the outstanding coincidence made impossible to distinguish between them. The parameters associated with voltage dependence were indistinguishable for both channels. In particular, the gating charge, related to the number of charges moving across the membrane as the channel changes from open to closed, and V0, the voltage where the probability of being open or closed is the same (Figures 5E and 5F). The selectivity of both channels can be described as low and slightly cationic, since the ability to distinguish anions from cations is 5 to 1 for K+ over Cl-. This could be related to the big pore size necessary to accommodate a protein chain during translocation and the recognition of the positively charged presequences. Even a more detailed electrophysiological characterization of the single channel kinetics was unable to reveal any significant differences between the Tim23 and Tom channel activities. As shown in Table 2, the mean open and close times as well as the dwell constants, a reflection of the several conductance and kinetic states, were virtually identical. Considering both of these channels have the same function, it might not be surprising they share common properties. However, as they are composed of different proteins, some variations are expected. Table 2. Kinetic analysis of Tim23 and Tom single channel parameters

Analyzed at 2 kHz filtration and 5 kHz sampling rate

Channel activity of mitochondrial protein import

449

3. Regulation of the Tim23 and Tom channel activities by pH The pH dependence of the Tim23 and Tom channel activities was determined in order to probe their relationship with protein import (280). The open probabilities for Tim23 and Tom channels had similar pH dependences, as shown in the current traces and plots of Figure 6. Increasing proton concentration

Figure 6. The pH dependence of Tim23 and Tom channel activities. Current traces of Tim23 (A) and Tom (B) channels were recorded from excised patches from proteoliposomes after perfusion of the bath with media at the indicated pH with the microelectrodes filled with 150 mM KCl, 5 mM HEPES, pH 7.4. O, S, and C correspond to the open, sub-, and closed states, respectively. (C) The open probability (Po) of the Tim23 (z,„) and Tom ({,…) channels was calculated from total amplitude histograms of 30 seconds of current traces at +20 mV after perfusion of the bath with media at the indicated pH (z,{). Alternatively, pH indicates that one inside the pipette, with medium pH 7.4 in the bath (z,…). The proton EC50 is 6.5 and 5.7 and the Hill coefficients are 1.1 ± 0.1 and 1.8 ± 0.3 for Tim23 and Tom respectively. (D) The probability of occupying the substate (Ps) of the Tim23 and Tom channels was calculated as above. The proton IC50s for increasing Ps are pH 6.4 and 5.9 and the Hill coefficients were -1.4 ± 0.2 and -3.7 ± 0.3 for Tim23 and Tom respectively. (E) The probability of the Tim23 and Tom channels occupying the closed state (Pc) was calculated as above. The proton IC50s are pH 5.25 and 3.2 and the Hill coefficients were 1.4 ± 0.1 and -0.3 ± 0.1 for Tim23 and Tom, respectively. Taken from Grigoriev et al. with permission (280).

450

Pablo M.V. Peixoto et al.

on the intermembrane or matrix faces decreased the open probabilities for both channels. Accordingly, the probabilities of the substate and closed states increased with acidification of the medium. A small shift on the IC50 of proton-induced half-closure could be observed. While the IC50 was pH 6.4 for the Tim23 channel, a value of pH 5.7 was recorded for the Tom channel. The differences in IC50s indicate the Tim23 channel is more easily inhibited by protons than the Tom channel, as indicated by the different conformation states occupied by the two channels at low pH (see current traces of Figure 6 at pH 5.0 and 5.5). The inhibition of the Tim23 and Tom channels by protons may be related to H+ binding sites on the channels that recognized the positively charged precursors of preproteins. Alternatively, the destabilization of the open states of the channels may be due to effects of pH on the lipid bilayer, proteinlipid interactions, or a direct effect on the channel proteins themselves. The pH dependence of the open probability for the two protein import channels was correlated with that of the protein import itself. Importantly, The IC50 for protein import into mitochondria (pH 6.5) was indistinguishable from that of the fully open state of the Tim23 channel (Figure 7). While the Tom channel did not show a linear relationship with protein import, a correlation could be established between the open probability of Tim23 as a function of pH and protein import. Therefore, the open probability of Tim23 may limit import of preproteins into mitochondria at acidic pH. Furthermore, the parallel decreases in import and open probability of the Tim23 suggest that both pores of this channel may need to be open in order for import to occur. 4. Regulation of the Tim23 and Tom channel activities by signal peptides The targeting regions or presequences at the amino-termini of mitochondrial precursors are not conserved, but the domains fold as cationic amphiphilic α-helices (197,281). As stated before, earlier studies showed that several synthetic peptides, whose sequences mimic the mitochondrial targeting regions, e.g., cytochrome oxidase subunits IV and VI, regulated the conductance of ions through Tim23 and Tom channels in patch-clamp experiments (267,268,272). The effects of these signal peptides were dependent on dose and voltage (92,268). Although the peptide SynB2 is similar in structure, i.e., cationic and α-helical, it does not support protein import into mitochondria (282). Notably, SynB2 also did not modify Tim23 channel activity (267). Furthermore, we found that nVDAC, an α-helical peptide with one negative charge (283), cVDAC, an uncharged peptide that is not α-helical (283), as well as iVDAC a cationic peptide predicted to form a βsheet (284), did not cause any flicker increase. These investigations were extended to directly compare the effects of synthetic peptides on Tim23 and Tom channels in a strain of yeast deficient in both isoforms of VDAC.

Channel activity of mitochondrial protein import

451

C

Figure 7. Mitochondrial protein import correlates with the open probability of the Tim23 channel. (A) Autoradiographs show inhibition of protein import into mitochondria at low pH. P and M indicate preprotein and mature protein. Mature protein is processed after import and hence has a lower molecular weight. (B) Protein import, as indicated by mature protein band density, was normalized to that at pH 7.4. The IC50 was pH 6.5. (C) Protein import capability is linked to Tim23 open probability. The relationship between the relative protein import function and the open probabilities of the Tim23 (●) and Tom (○) channels is shown. While the best fit line for the Po of Tim23 and protein import had a correlation coefficient of 0.975, the data for the Po of Tom did not have a linear relationship with protein import. Taken from Grigoriev et al. with permission (280).

452

Pablo M.V. Peixoto et al.

The native orientation of the inner and outer membranes is maintained in the reconstituted system as indicated by similar accessibility of specific inner or outer membrane proteins to controlled proteolytic digestion both in the native system (mitoplasts and mitochondria) and the reconstituted system (inner and outer membrane proteoliposomes) (not shown). Additionally, the same asymmetry of voltage dependence was recorded from >95% of all Tim23 and Tom channels. Therefore, the illustrations of Figures 8A and 8B represent the general configuration of the proteoliposome patch and microelectrode with respect to the TIM23 and TOM complexes. Addition of peptides to the bath solution represented their application to the matrix-face of the TIM23 complex and to the intermembrane space face of the TOM complex in the inner and outer membrane proteoliposomes, respectively. Conversely, inclusion of peptides in the micropipettes represented their exposure to the intermembrane space side on the inner membrane and the cytoplasmic side of the outer membrane in the two types of proteoliposomes. The Tim23 and Tom channels respond similarly to the presence of signal peptides. They remained predominately open at their peak conductance of ~1000 pS at low positive potential, e.g. +20 mV. As shown in Figures 8A and 8B, inclusion of the signal peptide yCOX-IV(1-13) in the medium modified the flow of ions through the pores of both channels. Transitions to the half-open level of 500 pS could be visualized as downward deflections in the current traces. Whereas transitions to this substate were relatively infrequent in the absence of signal peptide or in the presence of the control peptide, SynB2, the current traces revealed large amplitude, rapid flickering between the open (~1000 pS), half-open (500 pS) and closed states in the presence of yCOX-IV(113). This effect was specific for targeting peptides. Typically, there was an increase of four- to eightfold in the frequency of half and full closures (or transition events) in the presence of yCOX-IV(1-13) compared to the absence of peptide (control) and the presence of SynB2, as represented in the histograms of Figure 8E. Interestingly, the dose dependence for the effect, as well as the maximal flicker rate induced, were the same for the Tim23 and Tom channels (Figure 8F). In agreement with this effect, presequence peptides modified the kinetics of Tim23 and Tom channel activities. While the mean closed time was basically unaffected, a significant reduction in mean open time was apparent (Table 3), consistent with a destabilization of the open state. Once again, the effects of various signal peptides on both channel activities were indistinguishable. Notably, the µM concentration of the signal peptides necessary to induce these effects on the Tim23 and Tom channels (267,268) were similar to those known to competitively inhibit protein import (285). The effect of presequences was reversible by perfusion with media in the absence of these peptides, as shown in the sample current traces and total amplitude histograms of Figure 9.

Channel activity of mitochondrial protein import

453

Tim23

Tom

A

B

C

D

Control SynB2

yCOX-IV

F

500 250 0

Tim23

control

SynB2

Tom yCOX

Flicker (events/s)

Flicker (events/s)

E 750

Tom Tim23

500 250 0

1

10

100

Log[yCOX-IV1-13 ] µM

Figure 8. The channel activities of Tim23 and Tom are modulated by synthetic targeting peptides. (A) and (B) Illustrations show the general orientation of the proteoliposome patch and the electrode with respect to the TIM23 and TOM complexes, as well as the side of the membrane perfused with the peptide. (C) and (D) Typical current traces of Tim23 and Tom were recorded at 2 kHz from patches excised from proteoliposomes containing mitochondrial inner and outer membranes, respectively. Traces were obtained in the absence (control) and presence of 50 µM targeting peptide (yCOX-IV from cytochrome oxidase subunit IV) or control peptide (SynB2) in symmetrical 150 mM KCl, 5 mM HEPES, pH 7.4, under voltage-clamp conditions at +20 mV. O, S, and C indicate the open, substate and closed current levels. (E) The histograms of flicker rates (number of transition events/second) in the absence (control) and presence of SynB2 or yCOX-IV1-13 are similar for the Tim23 and Tom channels. (F) The plot shows the dependence of the flicker rate of the Tim23 (z) and Tom ({) channels on the concentration of the targeting peptide yCOX-IV1-13. Modified from Kinnally et al. with permission (274).

454

Pablo M.V. Peixoto et al.

Table 3. Presequence peptides modify the kinetics of Tim23 and Tom channel activities

*50 µM yCOX-IV(1-13) in the bath at +20 mV & microelectrode at -20 mV

Figure 9. The effect of targeting peptides is reversible. Current traces (A) and current amplitude histograms (B) (from current traces 20-30 seconds in duration of yeast Tim23 channel activity at +30 mV are shown in the presence (middle) and absence (left) of 50 µM yCOX-IV peptide. Effects were reversed by replacing the bath with media containing no peptide (right). Zero current level was measured at 0 mV under symmetrical conditions. Taken from Lohret et al. with permission (268).

Peptide sensitivity was a voltage-dependent phenomenon in that rapid flickering of Tim23 and Tom channels was observed when the electrical gradient favored movement of the signal peptide across the membrane. Hence, the flicker rates increased with positive potentials if the peptides were located in the bath and had little effect at negative potentials. Conversely, they increased with negative potentials if the peptides were located in the microelectrode. Intriguingly, the flicker rates induced by 50 µM peptide in the bath at +20 mV were consistently greater than the rates if the peptides were in the microelectrode at –20 mV for both Tim23 and Tom channels. These observations may be related to the voltage dependence of the channels as suggested by Kasianowicz et al. (286) in blockade of hemolysin by singlestranded DNA molecules. Alternatively, the disparity of flicker rates may be due to variations in the effective peptide concentration (possibly by chelation

Channel activity of mitochondrial protein import

455

of the peptide by the micropipette glass). These observations may also be a reflection of different binding affinities at sites located on the two membrane faces of the channels. The nature of the flicker events is not yet totally understood. The duration of the events was typically 400-700 µs suggesting an interaction of the signal peptides with the channel, with the consequent destabilization of the open state. On the other hand, it was proposed that these events reflected transient occlusions of the pore of the channels during translocation of the peptides. However, such translocation events were expected to be significantly faster, tens of nsec rather than hundreds of µsec, for polymers of this length (286). Akeson et al. found correlations between the duration of blockade and polymer length in studies of RNA and DNA translocation through the hemolysin channel (287), and the type of blockade of VDAC by nucleotides (288, 289). It is also possible that both effects are concomitant, and the fast translocation events are masked by the long duration blockage associated with signal peptides interaction with the channels, e.g., receptors. While the regulation of Tim23 and Tom channel activities by targeting peptides does not discriminate between them, these studies generated a functional assay in which we can now evaluate the effects of structural changes made to the TIM23 and TOM complexes by biochemical and genetic manipulations. 5. Differential response of the Tim23 and Tom channels to proteolytic digestion Limited proteolytic digestion was undertaken to further characterize the Tim23 and Tom channels. Trypsin treatment of the matrix side and the intermembrane side (see Figure 10) of excised proteoliposome patches containing inner and outer mitochondrial membranes, respectively, did not modify the peptide sensitivity of the Tim23 and Tom channel activities. As expected, perfusion of the patches with signal peptides increased the flicker rate (Figure 10C). A subsequent perfusion with media containing trypsin (200 µg/ml) reversed the effect and relieved most of the flicker blockade. After 10 min in the presence of trypsin, perfusion with trypsin inhibitor followed by signal peptide showed the targeting peptide effect was insensitive to this proteolytic treatment, and the predictable increase in flickering was again recorded. These findings suggest the function of the components of the TIM23 and TOM complexes important in signal peptide regulation and/or closely associated with the Tim23 and Tom channels are not modified by this regime of proteolysis. In contrast, the peptide effects on Tim23, but not Tom channel activity are disrupted if the outer face of proteoliposomes is treated with 200 µg

Pablo M.V. Peixoto et al.

456

trypsin before patch-clamping. Figure 11 shows the typical current traces of Tim23 and Tom channels recorded at -20 mV under control conditions, i.e. no trypsin treatment and no signal peptide. The expected increase in flickering induced by signal peptides was observed in the middle current traces for both channels.

200 100 0

Tim23

yCOX-IV1-13

KCl+Trypsin

300 Control

Events/sec

400

Control

500

yCOX-IV1-13

C

KCl+Trypsin

yCOX-IV1-13

B

Tom

Figure 10. The matrix side of the Tim23 channel and the intermembrane space side of the Tom channel are trypsin insensitive. (A, B) The illustrations show the membrane face treated with trypsin ( ) and peptides with respect to the orientation of the TIM23 (A) and TOM (B) complexes. (C) Histograms of flicker rates show normal Tim23 and Tom channel activities at +20 mV in the absence (control) and presence of 50 µM of yCOX-IV1-13. The bath was then perfused with media without peptide (not shown), media containing trypsin at 200 µg/ml for 10 min (KCl + Trypsin), and finally media again containing 50 µM of yCOX-IV1-13. The targeting peptide effect was not altered by trypsin-treatment of matrix or intermembrane space side of Tim23 and Tom channels, respectively.

Channel activity of mitochondrial protein import

457

200

Tim23 yCOX-IV1-22

yCOX-IV1-13

yCOX-IV1-13

Tom

SynB-2

300 Control

Events/sec

400

yCOX-IV1-22

yCOX-IV1-13

yCOX-IV1-13

100

0 Untreated

Trypsin treated

Figure 11. Tim23 and Tom channels respond differently to trypsin treatment of the intermembrane space side and the cytosolic side of the membranes. The illustrations represent the face of the inner membrane (MIM) or the outer membrane (MOM) of proteoliposomes treated with trypsin ( ) and the administration of peptides. Current traces and histograms of flicker rates show the characteristic behavior of Tim23 and Tom channels at -20 mV in the absence and presence of 50 µM yCOX-IV1-13 in the microelectrode. In proteoliposomes containing mitochondrial inner membranes the typical increase in flickering induced by the signal peptides yCOX-IV1-13 or yCOX-IV122 was disrupted after trypsin treatment (200 µg/mg protein for 10 min). The same treatment did not affect the Tom channel activity.

458

Pablo M.V. Peixoto et al.

However, the lower traces show that after trypsin treatment with 200 µg/mg for 10 min, followed by trypsin inhibitor, Tim23 channel activity lost its peptide sensitivity, whereas Tom channel activity was not affected. The histograms of Figure 11 quantify these two distinct effects, i.e., only the Tim23 channel activity was modified by mild trypsin treatment of the intermembrane space. Furthermore, the same behavior was observed if the native membranes, i.e. intact mitoplasts and mitochondria, were treated with trypsin prior to reconstitution (not shown), confirming the reconstituted system maintains the orientation of the native membranes. Most importantly, these data provided a means to distinguish these two channel activities. Furthermore, these results corroborate previous findings indicating the independence of Tom and Tim23 channels. e.g., the channel activity of the inner membrane was blocked by antibodies against Tim23, while the channel activity of the outer membrane was not affected (268). Similarly, a point mutation in Tim23 resulted in loss of peptide sensitivity for patches containing inner membranes, but had no detectable effect on outer membrane patches (267) (see below). 6. Functional relationship of VDAC with the Tim23 and Tom channels These and other studies (see below) were carried out in a strain of yeast in which the two isoforms of VDAC were deleted. No differences on the regulation by signal peptides of the Tim23 and Tom channel activities were detected when compared to those of the wild-type strain. Furthermore, the channel activity recorded from the native inner membranes of mitochondria isolated from a yeast strain carrying a single VDAC deletion was identical to that recorded from proteoliposomes prepared with the inner membranes of mitochondria isolated from the wild-type (VDAC-containing) strain (37) or the double deletion mutant (38). These findings indicate that VDAC is not tightly linked to the normal channel activities of the two import complexes under reconstituted conditions.

C. Structure-function studies of the mitochondrial protein import complexes Despite the similarities in the electrophysiological properties of the Tim23 and Tom channel activities, their distinctiveness was revealed by differences in their behavior after structural modification by proteolysis. Hence, patch clamping can serve as an effective functional assay for structural modifications of the mitochondrial protein import complexes. This approach was extended so that the protein import complexes of mitochondria were altered by biochemical procedures or genetic manipulations, and the functional effects on the channel activities were analyzed and interpreted in the light of these results.

Channel activity of mitochondrial protein import

459

1. The Tim23 channel In previous studies, trypsin treatment of mitoplasts resulted in loss of the Tim23 protein, which was essential for protein import (270). In our studies, proteoliposomes containing inner membranes were prepared from mitoplasts similarly treated with trypsin, and the cleavage of Tim23 was verified by western blots using antibodies against Tim23 that recognize epitopes located on the external face of the inner membrane (290, 291) (Figure 12). Alternatively, trypsin treatment was performed after reconstitution of the inner membranes into liposomes. The results were identical in both cases. Importantly, the peptide sensitivity of the Tim23 channel activity was abolished by trypsin. Presequence peptides did not induce rapid flickering of the Tim23 channel in inner membrane proteoliposomes treated with trypsin (see Figure 12). Similarly, presequence peptides in the bath or microelectrode had no effect on Tim23 channel activity reconstituted from inner membranes of trypsin-treated mitoplasts (Figure 12). The western blots show other main components of the TIM23 complex, i.e., Tim17, Tim44, and Hsp-70 were not affected by the mild trypsin treatment, indicating the cleavage of Tim23p was sufficient to eliminate peptide regulation of the TIM23 complex. Thus, trypsin treatment of mitoplasts cleaved Tim23 and abolished presequence peptide regulation of Tim23 channel activity.

Mitoplasts

Control yCoxIV1-13 microelectrode yCoxIV1-13 bath

500 400 300 200 100 0

Untreated

Trypsin treated

Mitoplasts

Figure 12. Trypsin treatment of the intermembrane space side of mitoplasts affects the protein Tim23 of the inner membrane translocase. Western blots using various antibodies against components of the TIM23 translocase show trypsin treatment (200 µg/mg protein for 10 min) of the intermembrane space side of mitoplasts resulted in loss of Tim23 but not Tim17, Tim44 or Hsp70. The histograms of flicker rates show, the Tim23 channel activity of patches excised from trypsin treated mitoplasts is no longer sensitive to targeting peptides placed on either side of the membrane. The data indicate loss of Tim23 is associated with loss of recognition of targeting peptides.

460

Pablo M.V. Peixoto et al.

In a previous study, a point mutation on Tim23 (tim23-1) resulting in a substitution of aspartate for glycine at position 186, had effects similar to this trypsin treatment on the peptide sensitivity of the Tim23 channel activity (267). That is, presequence peptides did not induce flickering of the Tim23 channel from either side of the membrane in the tim23-1 mutant. Other than that, the electrical properties of the Tim23 channels recorded from wild-type and tim23-1 strains were virtually identical. Importantly, mitochondria isolated from yeast carrying this mutation are defective, in a temperature-dependent manner, in the import of several precursor proteins, including subunit IV of yeast cytochrome oxidase (270). Antibodies against Tim23 were checked for their effect on Tim23 channel activity (267). Preincubation of proteoliposomes with these antibodies resulted in the complete blockade of the channel activity (not shown). In contrast, equivalent amounts of pre-immune serum or antibodies against VDAC or and iron sulfur protein of the inner membrane, did not affect the electrophysiological behavior of the Tim23 channel. Taken all together, these findings indicate Tim23 protein is closely associated with the pore of the TIM23 complex, needed to maintain native-like channel activity, and support a role for Tim23 protein as a receptor for cationic α-helical presequences. 2. The Tom channel While Tom70 and Tom20 are putative receptors in the Tom complex, they were not found to be essential proteins and import of precursors with presequences was not inhibited in deletion mutants (292). Our observations that Tom channel activity, including sensitivity to presequence peptides, was not modified in the double deletion mutant, D20D70 (not shown) are consistent with these findings. In addition, protease treatment was used to modify the structure of the TOM complex to evaluate the contribution of various components to the regulation of the Tom channel activity by presequence peptides. As expected, digestion of Tom70 and Tom20 did not alter the peptide sensitivity of the channel (not shown). However, trypsintreatment resulting in proteolysis of the essential protein Tom22 caused the loss of peptide sensitivity. Figure 13 shows digestion of mitochondria by 400 µg/mg trypsin for 10 minutes eliminated the cytosolic epitopes of Tom22, while the pore forming Tom40 remained unaltered. Patch-clamp of proteoliposomes containing these digested membranes were no longer sensitive to presequence peptides, as reflected in the histograms. These results support a role for Tom22p as a receptor in the TOM complex and indicate it may be an integral part of the TOM complex as suggested by others e.g., (292, 293).

Channel activity of mitochondrial protein import

461

Mitochondria

Untreated Trypsin

Tom22

Flicker (events/sec)

Tom40

400

Control yCoxIV1-13 microelectrode yCoxIV1-13 bath

300 200 100 0

Untreated

Trypsin treated

Mitochondria

Figure 13. Cleavage of Tom22p causes loss of peptide sensitivity of the Tom channel. The illustration shows the membrane face of a mitochondria treated with trypsin ( ) prior to patch-clamping with respect to the orientation of the TOM complex. The western blots were done on aliquots of mitochondria untreated or treated with 400 µg trypsin/mg protein for 10 minutes. Antibodies against Tom40 showed trypsin digestion did not modify the signal of this protein. Antibodies against Tom22 detected the loss of this protein from the digested membranes. Proteolytic modification of Tom22 was accompanied by loss of peptide sensitivity, as represented in the histograms of flicker rates obtained at -20 mV. A milder trypsin treatment (200 µg trypsin/mg protein for 10 min) affected Tom70 and Tom20 but not the channel activity of the TOM complex (not shown).

3. The Tim22 channel While import of proteins into the mitochondrial matrix requires the function of the TIM23 complex, insertion of multitopic proteins into the inner membrane is mediated by the TIM22 translocase. To date, all electrophysiological studies carried out on native and isolated inner membranes reconstituted into proteoliposomes, have revealed only one channel activity related to protein import. As discussed before several studies associate this channel activity with the TIM23 complex. Yet, under the same conditions, no channel activity could be assigned to the TIM22 complex. It has been reported that recombinant Tim22 forms a hydrophilic voltageactivated channel after reconstitution into bilayers (247), and a similar approach has been used to characterize the channel activity of the purified TIM22 complex (294). The reconstituted Tim22 is a single pore with multiple subconductance levels and responds to an internal targeting signal, but not to presequences. However, all recordings with the TIM22 complex showed the

462

Pablo M.V. Peixoto et al.

characteristics of two coupled pores. Also, single particle analysis of this complex, appears to be a double/triple barrel pore. Importantly, electrophysiological differences have been reported between the channel activities of the TIM22 complex and those of Tim22 alone i.e., difference in sensitivity and threshold voltage needed to induce the increased flickering caused by internal targeting signals. It is thus possible that the nature of the preparation (i.e. recombinant Tim22p vs detergent solubilized TIM22 complex) and/or the reconstitution system (i.e. bilayers vs patch clamping) affect the electrophysiological properties of this channel. Efforts are currently underway to detect the channel activity of the TIM22 complex in the native system.

VII. Perspectives Tom, Tim23, and Tim22, the mitochondrial protein import channels are high conductance, voltage dependent and slightly cation selective, in keeping with their functions. They have a half open state, and are likely double pores. Importantly, they are sensitive to presequence and/or internal signal peptides. Even if a complete picture of the protein translocation processes has been outlined, many queries still remain unanswered. For example, the exact role of Tim17 and its implication on the channel structure of the TIM23 complex needs to be elucidated. In addition, the mechanism(s) by which the Tim44/mtHsp70 facilitates translocation across the inner membrane is presently controversial. Two main hypotheses are in lead. One puts forward diffusion across the membrane by Brownian motion; in which case binding of mtHsp70 would prevent backsliding. The other proposes that mtHsp70 unfolds preproteins and generates a pull upon ATP hydrolysis, resulting from ratcheting of preprotein-bound mtHsp70 on its membrane-anchoring site on Tim44. In addition, for many years the TIM22 complex has remained elusive, especially from the electrophysiological point of view. Conditions for detection of its channel activity in the native system have been hard to establish, and more effort is needed to clarify its properties and functioning. The use of specific mutants has proven a very useful tool and could serve to unravel these types of questions. The application of genetic strategies in a number of experimental systems, in particular the study of recombinant proteins in planar bilayers has been a very successful approach to understanding the protein import channels of mitochondria. The activities recorded this way come close to the behavior observed in the native/proteoliposome model systems with regard to pore size and peptide sensitivity. However, some significant differences have been detected in conductances with that reported for the native membrane or reconstituted in proteoliposomes. Such differences also are observed when the isolated translocases are reconstituted. It is possible that the loss of regulatory

Channel activity of mitochondrial protein import

463

components could be the cause of such differences. Something similar may occur with Tim23 channel detection. Direct patch-clamp of mitoplasts show that unless activated, e.g., by signal peptides, the channel is normally closed under metabolizing conditions. However, Tim23 activity does not require previous activation and is typically detected when the membranes reconstituted into proteoliposomes. Identification of such factors should give a closer physiological prospective of the functioning of these translocases. Of particular interest is the insertion of proteins into the inner membrane of mitochondria. Also intriguing is the existence of a channel in the TIM22 complex responsible for such function. The exact physiological role played by the Tim22 channel in this process needs to be addressed. In short, correlations between the electrophysiological activity and the translocase function of this channel are fundamental issues. Furthermore, studies on the functional effects of loss of the smaller Tim components will provide additional insight into the architecture of this translocase. Finally, despite the identification of the core components of the three translocases, new proteins are continuously being described (e.g., Tim54, Tim16, and Tim14). A more complete picture of mitochondrial protein import will emerge when their particular roles within the complex are established. While Tom, Tim23, and Tim22 channels are responsible for the import of unfolded proteins into mitochondria, the apoptosis-induced channel (MAC) is a putative export channel, at least for some proteins of the intermembrane space like cytochrome c. At this early point of time since the discovery of this channel, no evidence exists that a motor-like structure extrudes cytochrome c. However, the relationship between MAC and the TOM complex becomes an attractive possibility requiring some attention. In summary, the approach of using patch-clamp techniques to establish structure-function relationships of the mitochondrial protein import complexes not only allows the identification of their essential components but also enables an analysis of the character of the aqueous pores through which proteins are translocated. The scope of the new venues opened by these investigations extent from identification of pharmacological agents targeting the channels of the mitochondrial translocases to the discovery of the molecular mechanisms underling some inherited diseases based on genetically deficient protein import.

VIII. Acknowledgements This research was supported by the grants DGI (Spanish Ministry of Science and Technology) PB98-0988 and Consejería de Educación, Ciencia y Tecnología (Junta de Extremadura) 2PR02B007. NSF grants MCB-0235834 and INT003797, and NIH grant GM57249 to KWK. PMVP was a recipient of

Pablo M.V. Peixoto et al.

464

a CAPES 1047091 fellowship from the Brazilian Ministry of Science and Education. We thank Carla Koheler and Jeff Schatz (U. Basel, Switzerland), Mike Forte (Oregon Health State University) and Robert Jensen (Johns Hopkins University School of Medicine) for the yeast mutant strains.

IX. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27.

Mannella, C.A. and Colombini, M. 1984. Biochim Biophys Acta 774:206-214. Colombini, M. 1979. Nature 279:643-645. Dolder, M., Wendt, S., and Wallimann, T. 2001. Biol. Signals Recept. 10:93-111. Blachly-Dyson, E. and Forte, M. 2001. IUBMB Life 52:113-118. Tsujimoto, Y. and Shimizu, S. 2002. Biochimie 84:187-193. Chen, M., Zhai, L., Christensen, S. B., Theander, T. G., and Kharazmi, A. 2001. Antimicrob. Agents Chemother. 45:2023-2029. Madesh, M. and Hajnóczky, G. 2001. J. Cell Biol. 155:1003-1015. Azoulay-Zohar, H., Israelson, A., Abu-Hamad, S., and Shosham-Barmatz, V. 2004. Biochem. J. 377:347-355. Mannella, C. A., Pfeiffer, D. R., Bradshaw, P. C., Moraru, I. I., Slepchenko, B., Loew, L. M., Hsieh, C. E., Buttle, K., and Marko, M. 2001. IUMBM Life 52:93100. Mannella, C.A., Marko, M., Penczek, P., Barnard, D., and Frank, J. 1994 . Microsc. Res. Tech. 27:278-283. Brdiczka, D. 1991. Biochim Biophys Acta 1071:291-312. Chacinska, A., Rehling, P., Guiard, B., Frazier, A. E., Schulze-Specking, A., Pfanner, N., Voos, W., and Meisinger, C. 2003. The EMBO J. 22:5370-5381. Crompton, M. 2003. Curr. Med. Chem. 10:1473-1484. Mironova, G. D., Fedotcheva, N. I., Makarov, P. R., Pronevich, L. A., and Mironov, G. P. 1981. Biofizika 26:731-734. Maloff, B. L., Scordilis, S. P., and Tedeschi, H. 1977. Science 195:898-900. Campo, M. L., Bowman, C. L., and Tedeschi, H. 1984. Eur. J. Biochem. 141:1-4. Campo, M. L. and Tedeschi, H. 1984. Eur. J. Biochem. 141: 5-7. Tedeschi, H., Mannella, C. A., and Bowman, C. L. 1987. J. Membrane Biol. 97:21-29. Sorgato, C. M., Keller, B. U., and Stühmer, W. 1987. Nature 330:498-500. Hladky, S. B. and Haydon, D. A. 1970. Nature 225:451-453. Montal, M. and Mueller, P. 1972. Proc. Natl. Acad. Sci. USA 69:3561-3566. Wonderlin, W. F. and Finkel, A. F. R. J. 1990. Biophys. J. 58:289-297. Wilmsen, U., Methfessel, C., Hanke, W., and Boheim, G. 1983. Channel current fluctuation studies with solvent-free lipid bilayers using Neher-Sakmann pipettes. Elsevier, Amsterdam.479-485. Hamill, O. P., Marty, A., Neher, E., Sakmann, B., and Sigworth, F. J. 1981. Pflugers Arch 391:85-100. Kirichok, Y., Krapivinsky, G., and Clampham, D. E. 2004. Nature 427:360-364. Campo, M. L., Tedeschi, H., Muro, C., and Kinnally, K. W. 1997. J. Bioenerg. Biomembr. 29:223-231. Hille, B. 2002. Ionic channels of excitable membranes. Sinuaer Associates Inc., Sunderland, Massachusetts.

Channel activity of mitochondrial protein import

465

28. Campo, M. L., Kinnally, K. W., and Tedeschi, H. 1992. J. Biol. Chem. 267:81238127. 29. Sorgato, M.C. and Moran, O. 1993. Crit Rev Biochem Mol Biol 28:127-71. 30. Campo, M.L., Muro, C., Tedeschi, H., and Kinnally, K.W. 1996. Electrophysiological studies of mitochondrial channels. Plenum Press, 11-21. 31. Zorov, D. B., Kinnally, K. W., Perini, S., and Tedeschi, H. 1992. Biochim. Biophys. Acta 1105:263-270. 32. Lohret, T.A. and Kinnally, K.W. 1995. Biophys J 68:2299-309. 33. Murphy, R. C., Diwan, J. J., King, M., and Kinnally, K. W. 1998. FEBS Letters 425:259-262. 34. Jonas, E. A., Buchanan, J., and Kaczmarek, L. K. 1999. Science 286:1347-1350. 35. De Pinto, V., Ludwig, O., Krause, J., Benz, R., and Palmieri, F. 1987. Biochim Biophys Acta 894:109-19. 36. Sorgato, M.C., Moran, O., De Pinto, V., Keller, B.U., and Stühmer, W. 1989. J. Bioenerg. Biomembr. 21:485-496. 37. Kinnally, K.W., Lohret, T.A., Campo, M.L., and Mannella, C.A. 1996. J Bioenerg Biomembr 28:115-23. 38. Muro, C., Grigoriev, S. M., Pietkiewicz, D., Kinnally, K. W., and Campo, M. L. 2003. Biophys. J. 84:2981-2989. 39. Moran, O., Sandri, G., Panfili, E., Stuhmer, W., and Sorgato, M. 1990. J. Biol. Chem. 265:908. 40. Moran, O., Sciancalepore, M., Sandri, G., Panfili, E., Bassi, R., Ballarin, C., and Sorgato, M.C. 1992. Eur Biophys J 20:311-9. 41. Mirzabekov, T., Ballarin, C., Nicolini, M., Zatta, P., and Sorgato, M. C. 1993. J. Membr. Biol. 133:129-143 . 42. Criado, M. and Keller, B. U. 1987. FEBS Lett. 224:172-176. 43. Campo, M.L., Tedeschi, H., Muro, C., and Kinnally, K.W. 1998. Mitochondrial membrane channels and their pharmacology. Oxforf University Press, New York.321-342. 44. Kinnally, K., Zorov, D., Antonenko, Y., Snyder, S., McEnery, M., and Tedeschi, H. 1993. PNAS 90:1374-1378. 45. Antonenko, Y., Kinnally, K. W., Perini, S., and Tedeschi, H. 1991. FEBS Lett. 285:89-93. 46. Seeman, P.M. 1966. Int. Rev. Neurobiol. 9:145-221. 47. Roth, S.H. 1986. Drugs Exp. Clin. Res. 12:719-722. 48. Gavish, M., Bachman, I., Shoukrun, R., Katz, Y., Veenman, L., Weisinger, G., and Weizman, A. 1999. Pharmacol Rev 51:629-650. 49. Casellas, P., Galiegue, S., and Basile, A. S. 2002. Neurochemistry International 40:475-486. 50. Schein, S. J., Colombini, M., and Finkelstein, A. 1976. J. Membr. Biol. 30:99-120. 51. Kleene, R., Pfanner, N., Pfaller, R., Link, T. A., Sebald, W., and Neupert, W. 1987. EMBO J. 6:2627-2633. 52. Mihara, K. and Sato, R. 1985. EMBO J. 4:769-774. 53. Ludwig, O., Benz, R., and Schultz, J. E. 1989. Biochim. Biophys. Acta 978:319-327. 54. Koppel, D. A., Kinnally, K. W., Masters, P., Forte, M., Blachy-Dyson, E., and Mannella, C. A. 1998. J. Biol. Chem. 273:3794-13800.

466

Pablo M.V. Peixoto et al.

55. Colombini, M., Blachly-Dyson, E., and Forte, M. 1996. Ion Channels 4:169-202. 56. Mannella, C.A. and Bonner, W.D. Jr. 1975. Biochim Biophys Acta 413:213-25. 57. Blumenthal, A., Kahn, K., Beja, O., Galum, E., Colombini, M., and Breiman, A. 1993. Plant Physiol. 101:579-587 . 58. Elkeles, A., Breiman, A., and Zizi, M. 1997. J. Biol. Chem. 272:6252-6260. 59. Lee. A., Xu, X., Blachly-Dyson, E., Forte, M., and Colombini, M. 1998. J. Membr. Biol. 161:173-181. 60. Linden, M. and Gellerfors, P. 1983. Biochim. Biophys. Acta 736:125-29. 61. Dolder, M., Zeth, K., Tittmann, P., Gross, H., Welte, W., and Wallimann, T. 1999. J. Struct. Biol. 127:64-71. 62. Rostovtseva, T. and Colombini, M. 1997. Biophys. J. 72: 1954-1962. 63. Colombini, M., Yeung, C. L., Tung, J., and Konig, T. 1987. Biochim. Biophys. Acta 905:279-286. 64. Zizi, M., Forte, M., Blachly-Dyson, E., and Colombini, M. 1994. J. Biol. Chem. 269:1614-1616. 65. Rostovtseva, T. K., Komarov, A., Bezrukov, S. M., and Colombini, M. 2002 . J. Membr. Biol. 187:147-156. 66. Holden, M. J. and Colombini, M. 1993. Biochim. Biophys. Acta 1144:396-402. 67. Hodge, T. and Colombini, M. 1997. J. Membr. Biol. 157:271-279. 68. Levadny, V., Colombini, M., Li, X. X., and Aguilella, V. M. 2002. Biophys. J. 82:1773. 69. Mannella, C. A. 1997. J. Bioenerg. Biomembr. 29:525-531. 70. Gincel, D., Zaid, H., and Shshan-Barmatz, V. 2001. Biochem. J. 358:147-155. 71. Peng, S., Blachly-Dyson, E., Colombini, M., and Forte, M. 1992. J Bioenerg Biomembr 24:27-31. 72. Casadio, R., Jacoboni, I., Messina, A., and De Pinto, V. 2002. FEBS Letters 520:1-7. 73. Sampson, M. J., Lovell, R. S., and Craigen, W. J. 1997. J. Biol. Chem. 272:1896618973. 74. Sampson, M. J., Ross, L., Decker, W. K., and Craigen, W. J. 1998. J. Biol. Chem. 273:30482-30486. 75. Blachly-Dyson, E., Song, J., Wolfgang, W. J., Colombini, M., and Forte, M. 1997. Moll. Cell. Biol. 17:5727-5738. 76. Wu, S., Sampson, M. J., Decker, W. K., and Craigen, W. J. 1999. Biochim. Biophys. Acta 1452:68-78. 77. Pastorino, J. G. and Hoek, J. B. 2003. Cuur. Med. Chem. 10:1535-1551. 78. Nakashima, R. A. and Hoek, J. B. 1989. J. Bioenerg. Biomembr. 21:461-470. 79. Östlund, A. K., Göhring, U., Krause, J., and Brdiczka, D. 1983. Biochem. Med. 30:231-245. 80. Vyssokikh, M. Y. and Brdiczka, D. 2003. Acta Biochim. Pol. 50:389-404. 81. Verrier, F., Mignotte, B., Jan, G., and Brenner, C. 2003. Ann NY Acad Sci 1010:126-142. 82. Haworth, R. A. and Hunter, P. R. 1980. J. Membr. Biol. 57:231-236. 83. Szabó, I., De Pinto, V., and Zoratti, M. 1993. FEBS Lett. 330:206-210. 84. Zoratti, M. and Szabó, I. 1995. Biochim. Biophys. Acta 1241:139-176. 85. Rostovtseva, T. K., Antonsson, B., Suzuki, M., Youle, R. J., Colombini, M., and Bezrukov, S. M. 2004. J. Biol. Chem. 279:13575-13583.

Channel activity of mitochondrial protein import

467

86. Cheng, E. H., Sheiki, T. V., Fisher, J. K., Craigen, W. J., and Korsmeyer, S. J. 2003. Science 301:513-517. 87. Thieffry, M., Chich, J.F., Goldschmidt, D., and Henry, J.P. 1988. EMBO J 7:1449-54. 88. Chich, J.F., Goldschmidt, D., Thieffry, M., and Henry, J.P. 1991. Eur J Biochem 196:29-35. 89. Févre, F., Chich, J. F., Lauquin, G. J., Henry, J.-P., and Thieffry, M. 1990. FEBS Lett. 262:201-204. 90. Thieffry, M., Neyton, J., Pelleschi, M., Févre, F., and Henry, J.-P. 1992. Biophys J 63:333-339. 91. Henry, J.-P., Juin, P., Vallette, F., and Thieffry, M. 1996. J. Biochem. Biophys. 28:101-108. 92. Henry, J.-P., Chich, J. F., Goldschmidt, D., and Thieffry, M. 1989. J. Membr. Biol. 112:139-147. 93. Fevre, F., Henry, J.P., and Thieffry, M. 1994. Biophys J 66:1887-94. 94. Pavlov, E. V., Priault, M., Pietkiewicz, D., Cheng, E. H. Y., Antonsson, B., Manon, S., Korsmeyer, S. J., Mannella, C. A., and Kinnally, K. W. 2001. J. Cell Biol. 155:725-732. 95. Priault, M., Camougrand, N., Kinnally, K. W., Vallette, F. M., and Manon, S. 2003. FEMS Yeast Research 4:15-27 . 96. Wei, M.C., Zong, W.X., Cheng, E.H., Lindsten, T., Panoutsakopoulou, V., Ross, A.J., Roth, K.A., MacGregor, G.R., Thompson, C.B., and Korsmeyer, S.J. 2001. Science 292:727-30. 97. Antonsson, B., Conti, F., Ciavatta, A., Montessuit, S., Lewis, S., Martinou, I., Bernasconi, L., Bernard, A., Mermod, J. J., Mazzei, G., Maundrell, K., Gambale, F., Sadoul, R., and Martinou, J. C. 1997. Science 277:370-372. 98. Martinou, J. C. and Green, D. R. 2001. Nat. Rev. Mol. Cell Biology 2:63-66. 99. Brenner, C. and Kroemer, G. 2000. Science 289:1150-1151. 100.Kroemer, G. and Reed, J.C. 2000. Nat. Med. 6:513-519. 101.Eskes, R., Antonsson, B., Osen-Sand, A., Montessuit, S., Richter, C., Sadoul, R., Mazzei, G., Nichols, A., and Martinou, J.-C. 1998. J. Cell Biol. 143 :217-224. 102.Jurgensmeier, J. M., Xie, Z., Deveraux, Q., Ellerby, L., Bredesen, D., and Reed, J. C. 1998. PNAS 95:4997-5002. 103.Finucane, D. M., Bossy-Wetzel, E., Waterhouse, N. J., Cotter, T. G., and Green, D. R. 1999. J. Biol. Chem. 274:2225-2233. 104.De Giorgi, F., Lartigue, L., Bauer, M. K. A., Schubert, A., Grimm, S., Hanson, G. T., Remington, S. J., Youle, R. J., and Ichas, F. 2002. FASEB J. 16 :607-609. 105.Guo, L., Pietkiewicz, D., Pavlov, E. V., Grigoriev, S. M., Kasianowicz, J. J., Dejean, L. M., Korsmeyer, S. J., Antonsson, B., and Kinnally, K. W. 2004 . Am J Physiol Cell Physiol 286:C1109-1117. 106.Pavlov, E. V., Priault, M., Pietkiewicz, D., Cheng, E. H. Y., Antonsson, B., Manon, S., Korsmeyer, S. J., Mannella, C. A., and Kinnally, K. W. 2001. J. Cell Biol. 155:725-732. 107.Szabo, I., Bernardi, P., and Zoratti, M. 1992. J. Biol. Chem. 267:2940-2946. 108.Lenartowicz, E., Bernardi, P., and Azzone, G.F. 1991. J. Bioenerg. Biomembr. 23:679-688.

468

Pablo M.V. Peixoto et al.

109.Broekmeier, K.M., Carpenter-Deyo, L., Reed, D.J., and Pfeiffer, D.R. 1992. FEBS Lett. 304:192-194. 110.Martínez-Caballero, S., Dejean, L.M., and Kinnally, K.W. 2004. FEBS Lett. 111.Antonsson, B., Montessuit, S., Lauper, S., Eskes, R., and Martinou, J. C. 2000. Biochem. J. 345:271-278. 112.Saito, M., Korsmeyer, S. J., and Schlesinger, P. H. 2000. Nat. Cell Biol. 2:553555. 113.Sorgato, M.C., Keller, B.U., and Stühmer, W. 1987. Nature 330:498-500. 114.Murphy, R. C., Diwan, J. J., King, M., and Kinnally, K. W. 1998. FEBS Letters 425:259-262. 115.Kinnally, K.W., Campo, M.L., and Tedeschi, H. 1989. J Bioenerg Biomembr 21:497-506. 116.Inoue, I., Nagase, H., Kishi, K., and Higuti, T. 1991. Nature 352:244-7. 117.Klitsch, T. and Siemen, D. 1991. J Membr Biol 122:69-75. 118.Petronilli, V., Szabó, I., and Zoratti, M. 1989. FEBS Lett. 259:137-143. 119.Moran, O. and Sorgato, M. C. 1992. J. Bioenerg. Biomembr. 24:91-98. 120.Kinnally, K.W. and Tedeschi, H. 1994. Mitochondrial channels an integrated view. Springer-Verlag, Berlin Heidelberg.169-198. 121.Ballarin, C., Sorgato, M.C., and Moran, O. 1994. A minimal kinetic model of the 107 pS channel of the inner membrane of mitochondria. Springer-Verlag, Berlin, New York.131-136. 122.Kinnally, K.W., Zorov, D., Antonenko, Y., and Perini, S. 1991. Biochem Biophys Res Commun 176:1183-8. 123.Kinnally, K.W., Zorov, D.B., Antonenko, Y.N., Snyder, S.H., McEnery, M.W., and Tedeschi, H. 1993. Proc Natl Acad Sci U S A 90:1374-8. 124.Kinnally, K. W., Antonenko, Y., and Zorov, D. 1992. J. Bioenerg. Biomembr. 24:99-110. 125.Andreyev, A.Y., Bondareva, T.O., Dedukhova, U.I., Mokhova, E.N., Skulachev, V.P., and Volkov, N.I. 1988. FEBS Lett. 226:265-269. 126.Beavis, A.D. 1992. J Bioenerg Biomembr 24:77-90. 127.Beavis, A.D. and Davatol-Hag, H. 1996. J. Bioenerg. Biomembr. 28:207-24. 128.Beavis, A. and Garlid, K. 1988. J. Biol. Chem. 263:7574-7580. 129.Beavis, A. 1989. J. Biol. Chem. 264:1508-1515. 130.Siemen, D. 1997. J. Biol. Chem. 272:19282-19289. 131.Bowman, C.L., Ding, J.P., Sachs, F., and Sokabe, M. 1992. Brain Res. 584:272286. 132.Sachs, F. 1991. Mol. Cell Biochem. 104:57-60. 133.Yang, X.C. and Sachs, F. 1989. Science 243:1068-1071. 134.Antonenko, Y., Kinnally, K. W., and Tedeschi, H. 1991. J. Membr. Biol. 124:151158. 135.Antonenko, Y., Smith, D., Kinnally, K., and Tedeschi, H. 1994. Biochimica et Biophysica Acta (BBA) 1194:247-254. 136.Jung, D. W. and Brierley, G. P. 1999. Am J Physiol Cell Physiol 277:C1194-1201. 137.Kim, J.S., He, L., and Lemasters, J. 2003. Biochem. Biophys. Res. Com. 304:463470. 138.Dahlem, Y. A., Horn, T. F. W., Buntinas, L., Gonoi, T., Wolf, G., and Siemen, D. 2004. Biochimica et Biophysica Acta (BBA) - Bioenergetics 1656:46-56.

Channel activity of mitochondrial protein import

469

139.Zhang, D. X., Chen, Y.-F., Campbell, W. B., Zou, A.-P., Gross, G. J., and Li, P.-L. 2001. Circ Res 89:1177-1183. 140.Mironova, G.D., Fedotcheva, N.I., Makarov, P.R., Pronevich, L.A., and Mironov, G.P. 1981. Biofizika 26:451-457. 141.Diwan, J.J., Haley, T., and Sanadi, D.R. 1988. Biochem. Biophys, Res. Commun. 153:224-230. 142.Paucek, P., Mironova, G., Mahdi, F., Beavis, A., Woldegiorgis, G., and Garlid, K. 1992. J. Biol. Chem. 267:26062-26069. 143.Garlid, K. D., Paucek, P., Yarov-Yarovoy, V., Murray, H. N., Darbenzio, R. B., D'Alonzo, A. J., Lodge, N. J., Smith, M. A., and Grover, G. J. 1997. Circ Res 81:1072-1082. 144.O'Rourke, B. 2004. Circ Res 94:420-432. 145.Liu, Y., Sato, T., O'Rourke, B., and Marban, E. 1998. Circulation 97:2463-2469. 146.Gross, G.J. and Fryer, R.M. 1999. Cir. Res. 84:973-979. 147.Oldenburg, O., Cohen, M.V., Yellon, D.M., and Downey J.M. 2002. Cardiovasc. Res. 55:429-437. 148.Garlid, K. D., Dos Santos, P., Xie, Z.-J., Costa, A. D. T., and Paucek, P. 2003. Biochimica et Biophysica Acta (BBA) - Bioenergetics 1606:1-21. 149.Mironova, G. D., Negoda, A. E., Marinov, B. S., Paucek, P., Costa, A. D. T., Grigoriev, S. M., Skarga, Y. Yu., and Garlid, K. D. 2004. J. Biol. Chem. 279:32562-32568. 150.Suzuki, M., Kotake, K., Fujikura, K., Inagaki, N., Suzuki, T., Gonoi, T., Seino, S., and Takata, K. 1997. Biochemical and Biophysical Research Communications 241:693-697. 151.Bajgar, R., Seetharaman, S., Kowaltowski, A. J., Garlid, K. D., and Paucek, P. 2001. J. Biol. Chem. 276:33369-33374. 152.Szewczyk, A., Wojcik, G., Lobanov, N. A., and Nalecz, M. J. 1997. Biochemical and Biophysical Research Communications 230:611-615. 153.Grover, G. J. and Garlid, K. D. 2000. Journal of Molecular and Cellular Cardiology 32:677-695. 154.Csordas, G. and Hajnoczky, G. 2003. J. Biol. Chem. 278:42273-42282. 155.Poburko, D., Lee, C.-H., and van Breemen, C. 2004. Cell Calcium 35:509-521. 156.Landgraf, G., Gellerich, F.N., and Wusshings, M.H. 2004. Mol. Cell Biochem. 256-257:379-386. 157.Ichas, F., Jouaville, L.S., and Mazat, J.P. 1997. Cell 89:1145-1153. 158.Babcock, D.F., Herrington, J., Goodwin, P.C., Park, Y.B., and Hilee, B. 1997. J. Cell Biol. 136:833-844. 159.Selwyn, M. J., Dawson, A. P., and Dunnett, S. J. 1970. FEBS Letters 10:1-5. 160.Puskin, J.S., Gunter, T.E., Gunter, K.K., and Pussell, P.R. 1976. Biochemistry 15:3834-3842. 161.Rizzuto, R., Bernardi, P., and Pozzan, T. 2000. J Physiol (Lond) 529:37-47. 162.Montero, M., Alonso, M.T., Carnicero, E., Cuchillo-Ibañez, I., Albillos, A., García, A.G., García-Sancho, J., and Alvarez, J. 2000. Nat. Cell Biol. 2:57-61. 163.Kokoszka, J.E., Waymire, K.G., Levy, S.E., Sligh, J.E., Cal, J., Jones, D.P., MacGregor, G.R., and Wallace, D.C. 2004. Nature 427:461-465. 164.Last, T., Gantzer, M., and Tyler, C. 1983. J. Biol. Chem. 258:2399-2404.

470

Pablo M.V. Peixoto et al.

165.Dierks, T., Salentin, A., and Kramer, R. 1990. Biochim. Biophys. Acta 1028:281288. 166.Schroers, A., Kramer, R., and Wohlrab, H. 1997. J. Biol. Chem. 272:10558-10564. 167.Herick, K., Krämer, R., and Lühring H. 1997. Biochim. Biophys. Acta 1321:207220. 168.Huang, S. G. and Klingenberg, M. 1996. Biochemistry 35:16806-16814. 169.Brustovetsky, N. and Klingenberg, M. 1996. Biochemistry 35:8483-8488. 170.Ledesma, A. and Rial, E. 2004. Toxicology Mechanisms and Methods 14:41-46. 171.Bernardi, P. 1999. Physiol. Rev. 79:1127-1155. 172.Brustovetsky, N., Tropschug, M., Heimpel, S., Heidkämper, D., and Klingenberg, M. 2002. Biochemistry 41:11804-11811. 173.Hunter, D.R., Haworth, R.A., and Southard, J.H. 1976. J Biol Chem 251:5069-77. 174.Hunter, D. R. and Haworth, R. A. 1979. Arch Biochem Biophys 195:453-59. 175.Crompton, M., Ellinger, H., and Costi, A. 1988. Biochem. J. 255:357-360. 176.Lemasters, J. J., Nieminen, A. L., Qian, T., Trost, L. C., Elmore, S. P., Nishimura, Y., Crowe, R. A., Cascio, W. E., Bradham, C. A., Brenner, D. A., and Herman, B. 1998. Biochimica et Biophysica Acta (BBA) 1366:177-196. 177.Lemasters, J. 1998. Gastroenterology 115:783-786. 178.Szabó, I. and Zoratti, M. 1991. J. Biol. Chem. 266:3376-3379. 179.Verrier, F., Mignotte, B., Jan, G., and Brenner, C. 2003. Ann. N.Y. Acad. Sci. 1010:126-142. 180.Lohret, T.A., Murphy, R.C., Drgon, T., and Kinnally, K.W. 1996. J Biol Chem 271:4846-9. 181.Kogelnik, A., Lott, M., Brown, M., Navathe, S., and Wallace, D. 1996. Nucl. Acids. Res. 24:177-179. 182.Foury, F., Roganti, T., Lecrenier, N., and Purnelle, B. 1998. FEBS Letters 440:325-331. 183.Hoogenraad, N. J., Ward, L. A., and Ryan, M. T. 2002. Biochimica et Biophysica Acta (BBA) - Molecular Cell Research 1592:97-105. 184.Johnston, A. J., Hoogenraad, J., Dougan, D. A., Truscott, K. N., Yano, M., Mori, M., Hoogenraad, N. J., and Ryan, M. T. 2002. J. Biol. Chem. 277:42197-42204. 185.Muhlenbein, N., Hofmann, S., Rothbauer, U., and Bauer, M. F. 2004. J. Biol. Chem. 279:13540-13546. 186.Guo, Y., Cheong, N., Zhang, Z., De Rose, R., Deng, Y., Farber, S. A., FernandesAlnemri, T., and Alnemri, E. S. 2004. J. Biol. Chem. 279:24813-24825. 187.Rehling, P., Pfanner, N., and Meisinger, C. 2003. J. Mol. Biol. 326 :639-657. 188.Truscott, K. N., Brandner, K., and Pfanner, N. 2003. Current Biology 13:R326R337. 189.Rassow, J. and Pfanner, N. 2000. Traffic 1:457-464. 190.Pfanner, N. and Geissler, A. 2001. Nat Rev Mol Cell Biol 2:339-349. 191.Endo, T., Yamamoto, H., and Esaki, M. 2003. J Cell Sci 116:3259-3267. 192.Beddoe, T. and Lithgow, T. 2002. Biochimica et Biophysica Acta (BBA) Molecular Cell Research 1592:35-39. 193.Wienhues, U., Becker, K., Schleyer, M., Guiard, B., Tropschug, M., Horwich, A., Pfanner, N., and Neupert, W. 1991. J. Cell Biol. 115:1601-1609. 194.Matouschek, A., Pfanner, N., and Voos, W. 2000. EMBO J. 1:404-410.

Channel activity of mitochondrial protein import

471

195.Voos, W. and Rottgers, K. 2002. Biochimica et Biophysica Acta (BBA) Molecular Cell Research 1592:51-62. 196.Mihara, K. and Omura, T. 1996. Experientia 52:1063-1068. 197.Schatz, G. and Dobberstein, B. 1996. Science 271:1519-1526. 198.Roise, D. and Schatz, G. 1988. J. Biol. Chem. 263:4509-4511. 199.Abe, Y., Shodai, T., Mihara, K., Torii, H., Nishikawa, S., Endo, T., and Kohda, D. 2000. Cell 100:551-560. 200.Hahne, K., Haucke, V., Ramage, L., and Schatz, G. 1994. Cell 79:829-839. 201.Gartner, F., Bomer, U., Guiard, B., and Pfanner, N. 1995. EMBO J. 14:6043-6057. 202.Glick, B.S., Brandt, A., Cunningham, K., Müller, S., Hallberg, R.L., and Schatz, G. 1992. Cell 69:809-822. 203.Wiedemann, N., Pfanner, N., and Ryan, M. T. 2001. EMBO J. 20:951-960. 204.Brix, J., Rudiger, S., Bukau, B., Schneider-Mergener, J., and Pfanner, N. 1999. J. Biol. Chem. 274:16522-16530. 205.Neupert, W. 1997. Annu Rev Biochem. 66:863-917. 206.Glick, B., Wacher, C., and Schatz, G. 1991. Trends in Cell Bio. 1:99-103. 207.Schulke, N., Sepuri, N. B. V., Gordon, D. M., Saxena, S., Dancis, A., and Pain, D. 1999. J. Biol. Chem. 274:22847-22854. 208.Schulke, N., Sepuri, N. B. V., and Pain, D. 1997. PNAS 94:7314-7319. 209.Brix, J., Dietmeier, K., and Pfanner, N. 1997. J. Biol. Chem. 272:20730. 210.van Wilpe, S., Ryan, M. T., Hill, K., Maarse, A. C., Meisinger, C., Brix, J., Dekker, P. J. T., Moczko, M., Wagner, R., Meijer, M., Guiard, B., Honlinger, A., and Pfanner, N. 1999. Nature 401:485-489. 211.Hill, K., Model, K., Ryan, M. T., Dietmeier, K., Martin, F., Wagner, R., and Pfanner, N. 1998. Nature 395:516-521 . 212.Kunkele, K.-P., Heins, S., Dembowski, M., Nargang, F. E., Benz, R., Thieffry, M., Walz, J., Lill, R., Nussberger, S., and Neupert, W. 1998. Cell 93:1009-1019. 213.Dietmeier, K., Honlinger, A., Bomer, U., Dekker, P. J. T., Eckerskorn, C., Lottspeich, F., Kubrich, M., and Pfanner, N. 1997. Nature 388:195-200. 214.Kurz, M., Martin, H., Rassow, J., Pfanner, N., and Ryan, M. T. 1999. Mol. Biol. Cell 10:2461-2474. 215.Honlinger, A., Bomer, U., Alconada, A., Eckerskorn, C., Lottspeich, F., Dietmeier, K., and Pfanner, N. 1996. EMBO J. 15:2125-2137. 216.Model, K., Meisinger, C., Prinz, T., Wiedemann, N., Truscott, K.N., Pfanner, N., and Ryan, M.T. 2001. Nat Struct Biol 8:361-70. 217.Pfanner, N. and Wiedemann, N. 2002. Current Opinion in Cell Biology 14:400411. 218.Wiedemann, N., Kozjak, V., Chacinska, A., Schonfisch, B., Rospert, S., Ryan, M.T., Pfanner, N., and Meisinger, C. 2003. Nature 424:565-571. 219.Kozjak, V., Wiedemann, N., Milenkovic, D., Lohaus, C., Meyer, H. E., Guiard, B., Meisinger, C., and Pfanner, N. 2003. J. Biol. Chem. 278:48520-48523. 220.Paschen, S. A., Waizenegger, T., Stan, T., Preuss, M., Cyrklaff, M., Hell, K., Rapaport, D., and Neupert, W. 2003. Nature 426:862-866. 221.Gentle, I., Gabriel, K., Beech, P., Waller, R., and Lithgow, T. 2004. J. Cell Biol. 164:19-24. 222.Truscott, K. N., Kovermann, P., Geissler, A., Merlin, A., Meijer, M., Driessen, A. J., Pfanner, N., and Wagner, R. 2001. Nat. Struct. Biol. 8:1074-1082.

472

Pablo M.V. Peixoto et al.

223.Donzeau, M., Kaldi, K., Adam, A., Paschen, S. A., Wanner, G., Guiard, B., Bauer, M. F., Neupert, W., and Brunner, M. 2000. Cell 101:401-412. 224.Komiya, T., Rospert, S., Koehler, C., Looser, R., Schatz, G., and Mihara, K. 1998. EMBO J. 17:3886-3898. 225.Bauer, M. F., Sirrenberg, C., Neupert, W., and Brunner, M. 1996. Cell 87 :33-41. 226.Mokranjac, D., Paschen, S. A., Kozany, C., Prokisch, H., Hoppins, S. C., Nargang, F. E., Neupert, W., and Hell, K. 2003. EMBO J. 22:816-825. 227.Geissler, A., Chacinska, A., Truscott, K. N., Wiedemann, N., Brandner, K., Meyer, H. E., Meisinger, C., Pfanner, N., and Rehling, P. 2002. Cell 111:507-518. 228.Yamamoto, H., Esaki, M., Kanamori, T., Tamura, Y., Nishikawa, S., and Endo, T. 2002. Cell 111:519-28. 229.Okamoto, K., Brinker, A., Paschen, S. A., Moarefi, I., Hayer-Hartl, M., Neupert, W., and Brunner, M. 2002. EMBO J. 21:3659-3671. 230.Liu, Q., D'Silva, P., Walter, W., Marszalek, J., and Craig, E. 2003. Science 300:139-141. 231.Krimmer, T., Rassow, J., Kunau, W.-H., Voos, W., and Pfanner, N. 2000. Mol. Cell. Biol. 20:5879-5887. 232.Mokranjac, D., Sichting, M., Neupert, W., and Hell, K. 2003. EMBO J. 22:49454956. 233.Truscott, K. N., Voos, W., Frazier, A. E., Lind, M., Li, Y., Geissler, A., Dudek, J., Muller, H., Sickmann, A., Meyer, H. E., Meisinger, C., Guiard, B., Rehling, P., and Pfanner, N. 2003. J. Cell Biol. 163:707-713. 234.Kozany, C., Mokranjac, D., Sichting, M., Neupert, W., and Hell, K. 2004. Nature Structural and Molecular Biology 11:234-241. 235.Wiedemann, N., Frazier, A. E., and Pfanner, N. 2004. J. Biol. Chem. 279:1447314476. 236.Neupert, W. and Brunner, M. 2002. Nat Rev Mol Cell Biol 3:555-565. 237.Gakh, O., Cavadini, P., and Isaya, G. 2002. Biochim Biophys Acta 1592 :63-77. 238.Taylor, A. B., Smith, B. S., Kitada, S., Kojima, K., Miyaura, H., Otwinowski, Z., Ito, A., and Deisenhofer, J. 2001. Structure 9:615-625. 239.Sirrenberg, C., Bauer, M. F., Guiard, B., Neupert, W., and Brunner, M. 1996. Nature 384:582-585. 240.Koehler, C. M., Murphy, M. P., Bally, N. A., Leuenberger, D., Oppliger, W., Dolfini, L., Junne, T., Schatz, G., and Or, E. 2000. Mol. Cell. Biol. 20:1187-1193. 241.Kerscher, O., Holder, J., Srinivasan, M., Leung, R. S., and Jensen, R. E. 1997. J. Cell Biol. 139:1663-1675. 242.Kerscher, O., Sepuri, N. B., and Jensen, R. E. 2000. Mol. Biol. Cell 11:103-116. 243.Sirrenberg, C., Endres, M., Folsch, H., Stuart, R. A., Neupert, W., and Brunner, M. 1998. Cell 391:912-915. 244.Koehler, C. M., Merchant, S., Oppliger, W., Schmid, K., Jarosch, E., Dolfini, L., Junne, T., Schatz, G., and Tokatlidis, K. 1998. EMBO J. 17:6477-6486. 245.Adam, A., Endres, M., Sirrenberg, C., Lottspeich, F., Neupert, W., and Brunner, M. 1999. EMBO J. 18:313-319. 246.Koehler, C.M., Jarosch, E., Tokatlidis, K., Schmid, K., Schweyen, R.J., and Schatz, G. 1998. Science 279:369-73. 247.Kovermann, P., Truscott, K. N., Guiard, B., Rehling, P., Sepuri, N. B., Muller, H., Jensen, R. E., Wagner, R., and Pfanner, N. 2002. Mol. Cell 9:363-373.

Channel activity of mitochondrial protein import

473

248.Paschen, S. A., Rothbauer, U., Kaldi, K., Bauer, M. F., Neupert, W., and Brunner, M. 2000. EMBO J. 19:6392-6400. 249.Koehler, C. M. 2004. Trends in Biochemical Sciences 29: 1-4. 250.Curran, S. P., Leuenberger, D., Oppliger, W., and Koehler, C. M. 2002. EMBO J. 21:942-953. 251.Endres, M., Neupert, W., and Brunner, M. 1999. EMBO J. 18:3214-3221. 252.Davis, A. J., Sepuri, N. B., Holder, J., Johnson, A. E., and Jensen, R. E. 2000. J. Cell Biol. 150:1271-1282. 253.Rothbauer, U., Hofmann, S., Muhlenbein, N., Paschen, S. A., Gerbitz, K.-D., Neupert, W., Brunner, M., and Bauer, M. F. 2001. J. Biol. Chem. 276:3732737334. 254.Hoppins, S. C. and Nargang, F. E. 2004. J. Biol. Chem. 279:12396-12405. 255.Herrmann, J. M. 2003. Trends in Microbiology 11:74-79. 256.Stuart, R. A. 2002. Biochimica et Biophysica Acta (BBA) - Molecular Cell Research 1592:79-87. 257.Herrmann, J. M. and Bonnefoy, N. 2004. J. Biol. Chem. 279:2507-2512. 258.Kuhn, A., Stuart, R., Henry, R., and Dalbey, R. E. 2003. Trends in Cell Biology 13:510-516. 259.Blobel, G. and Dobberstein, B. 1975. J. Cell Biol. 67:835-851. 260.Gilmore R. and Blobel, G. 1985. Cell 42:497-505. 261.Simon, S., Blobel, G., and Zimmerberg, J. 1989. Proc. Natl. Acad. Sci. USA 86:6176-6180. 262.Simon, S.M. and Blobel, G. 1991. Cell 65:371-380. 263.Pfanner, N. and Neupert, W. 1990. Ann. Rev. Biochem. 59:331-353. 264.Pfanner, N. and Neupert, W. 1987. J. Biol. Chem. 262:7528-7536. 265.Simon, S.M. and Blobel, G. 1992. Cell 69:677-684. 266.Henry, J.P., Chich, J.F., Goldschmidt, D., and Thieffry, M. 1989. Biochimie 71:963-968. 267.Lohret, T. A., Jensen, R. E., and Kinnally, K. W. 1997. J. Cell Biol. 137 :377-386. 268.Lohret, T.A. and Kinnally, K.W. 1995. J Biol Chem 270:15950-3. 269.Kushnareva, Y. E., Campo, M. L., Kinnally, K. W., and Sokolove, P. M. 1999. Arch. Biochem. Biophys. 366:107-115 . 270.Emtage, J.L. and Jensen, R.E. 1993. J Cell Biol 122:1003-12. 271.Milisav, I., Moro, F., Neupert, W., and Brunner, M. 2001. J. Biol. Chem. 276:25856-25861. 272.Juin, P., Thieffry, M., Henry, J.-P., and Vallette, F. M. 1997. J. Biol. Chem. 272:6044-6050. 273.Kunkele, K.-P., Juin, P., Pompa, C., Nargang, F. E., Henry, J.-P., Neupert, W., Lill, R., and Thieffry, M. 1998. J. Biol. Chem. 273:31032-31039. 274.Kinnally, K. W., Muro, C., and Campo, M. L. 2000. J. Bioenerg. Biomembr. 32:47-54. 275.Mannella, C.A. 1981. Biochim. Biophys. Acta 33-40. 276.Schwartz, M. P. and Matouschek, A. 1999. PNAS 96:13086-13090. 277.Beckmann, R. 1997. Science 278:2123-2126. 278.Menetret, J.F. 2000. Mol. Cell 6:1219-1232. 279.Brdiczka, D., Beutner, G., RÜck, A., Dolder, M., and Wallimann, T. 1998. Biofactors 8:235-242.

474

Pablo M.V. Peixoto et al.

280.Grigoriev, S. M., Jensen, R. E., and Kinnally, K. W. 2003. FEBS Letters 553:163166. 281.Schatz, G. 1993. Protein Sci 2:141-146. 282.Allison, D. S. and Schatz, G. 1986. Proc Natl Acad Sci U S A 83:9011-9015. 283.Guo, X.W., Smith, P.R., Cognon, B., D'Arcangelis, D., Dolginova, E., and Mannella, C.A. 1995. J Struct Biol 114:41-59. 284.Mannella, C.A., Neuwald, A.F., and Lawrence, C.E. 1996. J Bioenerg Biomembr 28:163-9. 285.Mannella, C. A. 1982. J. Cell Biol. 94:680-687. 286.Kasianowicz, J., Brandin, E., Branton, D., and Deamer, D. 1996. PNAS 93 :13770-13773. 287.Akeson, M., Branton, D., Kasianowicz, J.J., Brandin, B., and Deamer, D.W. 1999. Biophys. J. 77:3227-3233. 288.Rostovtseva, T. K. and Bezrukov, S. M. 1998. Biophys. J. 74:2365-2373. 289.Xian, X. and Colombini, M. 2002. Biophys. J. 83:2550-2559. 290.Ryan, K. R., Leung, R. S., and Jensen, R. E. 1998. Mol. Cell. Biol. 18:178-187. 291.Komiya, T., Rospert, S., Schatz, G., and Mihara, K. 1997. EMBO J. 16:42674275. 292.Lithgow, T., Junne, T., Wachter, C., and Schatz, G. 1994. J. Biol. Chem. 269:15325-15330. 293.Dekker, P. J. T., Ryan, M. T., Brix, J., Muller, H., Honlinger, A., and Pfanner, N. 1998. Mol. Cell. Biol. 18:6515-6524. 294.Rehling, P., Model, K., Brandner, K., Kovermann, P., Meyer, H. E., Kuhlbrandt, W., Wagner, R., Truscott, K. N., and Pfanner, N. 2003. Science 299:1747-1751.