Protein translocation at the envelope and thylakoid membranes of chloroplasts.

Endosymbiotic evolution has resulted in the transfer of genes encoding the vast majority of the protein components of plastids to the nuclear genome. In response to this displacement of genetic material, plastids have evolved a system to post-translationally import nuclear encoded preproteins from their site of synthesis on cytoplasmic ribosomes. The protein import process can be viewed as a cascade of protein targeting events that are governed by a hierarchy of topogenic sequences. The targeting sequences are sequentially decoded resulting in the localization of the polypeptide to the appropriate organellar subcompartment (for review see Refs. 1 and 2). Recent studies have begun to uncover the components that underlie the mechanism of targeting and translocation at the envelope and thylakoid membranes of chloroplasts. These studies suggest a single, common mechanism for recognition and translocation of cytoplasmic preproteins across the double membrane of the envelope. With the exception of two members of the hsp70 family of molecular chaperones, the translocation components of the envelope that have been identified and sequenced are unique and, surprisingly, show no similarity in primary structure to the known components of the mitochondrial import apparatus. In contrast, the thylakoid membrane appears to have evolved a variety of targeting pathways with certain pathways retaining elements that are closely related to bacterial and endoplasmic reticulum translocation systems.

Endosymbiotic evolution has resulted in the transfer of genes encoding the vast majority of the protein components of plastids to the nuclear genome. In response to this displacement of genetic material, plastids have evolved a system to post-translationally import nuclear encoded preproteins from their site of synthesis on cytoplasmic ribosomes. The protein import process can be viewed as a cascade of protein targeting events that are governed by a hierarchy of topogenic sequences. The targeting sequences are sequentially decoded resulting in the localization of the polypeptide to the appropriate organellar subcompartment (for review see Refs. 1 and 2).
Recent studies have begun to uncover the components that underlie the mechanism of targeting and translocation at the envelope and thylakoid membranes of chloroplasts. These studies suggest a single, common mechanism for recognition and translocation of cytoplasmic preproteins across the double membrane of the envelope. With the exception of two members of the hsp70 family of molecular chaperones, the translocation components of the envelope that have been identified and sequenced are unique and, surprisingly, show no similarity in primary structure to the known components of the mitochondrial import apparatus. In contrast, the thylakoid membrane appears to have evolved a variety of targeting pathways with certain pathways retaining elements that are closely related to bacterial and endoplasmic reticulum translocation systems.

The General Pathway of Envelope Translocation
Protein import across the chloroplast envelope consists of the specific recognition of preproteins at the outer envelope and subsequent translocation of the protein simultaneously across both the outer and inner membranes (for review see Refs. 3 and 4). Envelope translocation is facilitated by the coordinate interaction of protein-conducting machineries in the outer and inner membranes at contact zones where the two membranes are held in close apposition (5). The targeting signal for envelope translocation resides within an amino-terminal extension of the preprotein designated the transit sequence. Although there is no apparent similarity in primary sequence among the transit sequences of different precursor proteins, import competition studies and transit sequence swapping experiments support the proposal that all preproteins use the same mechanism for envelope translocation (3).
Two general stages in envelope translocation can be distin-guished by their distinct energy requirements (Fig. 1). The first stage represents the high affinity association of the precursor protein with the outer envelope (6,7) and requires the hydrolysis of low concentrations (Ͻ100 M) of both ATP (8,9) and GTP (10) in the cytoplasm or intermembrane space. At this stage, the precursor is irreversibly bound to the envelope and has been designated an early translocation or import intermediate (5,11). The second stage represents the complete translocation of the precursor into the stroma and requires the hydrolysis of higher concentrations of ATP (Ͼ1 mM) within the stromal compartment (12,13). Unlike protein import into mitochondria, a membrane potential is not involved in envelope translocation. Recent studies suggest that each energy-dependent stage of envelope translocation corresponds to sequential insertion of the preprotein across the outer and inner membranes, respectively (14,15). Upon entering the stromal compartment, the transit sequence is removed by a specific metalloendopeptidase called the general stromal processing peptidase (SPP) 1 (16,17). The SPP is a soluble stromal protein of 140 kDa that contains a zinc-binding domain that is conserved in several metalloendopeptidases including the ␤ subunit of the processing peptidase involved in the maturation of proteins imported into mitochondria (18).
The simultaneous translocation of preproteins across the outer and inner envelope membranes is facilitated by contact zones (5). These zones correspond to specialized membrane domains where the outer and inner membranes come into close contact. A recent study suggests that energy-dependent insertion of a precursor across the outer membrane is accompanied by the engagement or formation of contact sites (14). However, it is apparent that the outer and inner import machineries can function independently if the two membranes are physically separated by subjecting chloroplasts to a hypertonic shock (15). Therefore, it is unlikely that the import machineries are permanently linked but that their interactions at contact zones are dynamic. The biochemical nature of these zones remains to be established, but it is clear that these structures hold clues to the coordinate interactions of the import machineries in the outer and inner membrane.

Components of the Envelope Translocation Machinery
Five proteins of the outer envelope import machinery have been identified (Fig. 1). Three of these components, IAP34 (OEP34), IAP75 (OEP75), and IAP86 (OEP86), form a complex in the outer membrane that stably associates with early translocation intermediates (19,20). All three IAPs are integral membrane proteins (10,(21)(22)(23). IAP34 and IAP86 are related in primary sequence, and both contain cytoplasmically exposed GTP-binding domains (10,21,22). The latter characteristic presumably accounts for the GTP requirement in import. Protease sensitivity and membrane extraction experiments have shown that IAP75 is deeply embedded in the outer membrane (19,23); however, its primary sequence is remarkably hydro-* This minireview will be reprinted in the 1996 Minireview Compendium, which will be available in December, 1996. This is the third article of five in the "Protein Translocation Minireview Series." ‡ To whom correspondence should be addressed: Dept. of Biological Sciences, Rutgers, The State University of New Jersey, 101 Warren St., Newark, NJ 07102. Tel.: 201-648-1082; Fax: 201-648-1007. E-mail: schnell@andromeda.rutgers.edu. 1 The abbreviations used are: SPP, stromal processing peptidase; IAP, import intermediate associated proteins; STD, stromal targeting domain; LTD, lumen targeting domain; PC, plastocyanin; OE33, 33-kDa subunit of the oxygen-evolving complex; LHCP, light-harvesting complex protein; SRP, signal recognition particle; ER, endoplasmic reticulum. philic and exhibits no typical membrane-spanning helices. Secondary structure predictions suggest that this component may contain extensive ␤-sheet structure similar to the bacterial porins (19,23).
IAP75 and IAP86 were shown to directly cross-link to an early import intermediate indicating a central role for these components in the import process (11). Subsequent studies have shown that IAP75 and IAP86 cross-link directly to the transit sequence in precursor binding assays even in the absence of nucleoside triphosphates suggesting that these two components form an initial receptor site in the outer membrane import complex (14). A role for IAP86 as a receptor component is supported by the observation that antibodies to IAP86 inhibit binding of precursor proteins to the envelope (22).
In the presence of concentrations of ATP and/or GTP that promote partial insertion of the preprotein across the outer membrane, cross-linking of the transit sequence occurs predominantly to IAP75 (14). This observation suggests that, in addition to forming part of the receptor site, IAP75 may form all or part of the protein-conducting channel in the outer membrane. This assignment is consistent with the observation that anti-IAP75 antibodies block protein import (23) and is especially attractive considering the predicted porin-like structure of this component.
The function of IAP34, the third component of the outer membrane complex, remains to be determined. A direct association of IAP34 with precursor proteins has not been detected (14), and antibodies to IAP34 do not block import. The GTP binding activities of IAP34 and IAP86 have prompted investigators to speculate that they may work in concert to regulate the presentation of the transit sequence to the membrane translocation machinery through cycles of GTP binding and hydrolysis (19). This function is reminiscent of the roles played by the two GTP binding subunits of the signal recognition particle receptor in targeting the nascent chain-ribosome complex to the ER membrane (24).
The two additional import components of the outer envelope are homologs of the hsp70 family of molecular chaperones. One of these proteins, Com70, is peripherally associated with the cytoplasmic surface of the outer membrane and is closely related in primary structure to the major cytoplasmic eukaryotic hsp70s (25,26). The additional hsp70 homolog, the hsp70-IAP, was identified based on its presence in detergent-soluble complexes containing IAP34, IAP75, IAP86, and an early import intermediate (19,20). The hsp70-IAP is tightly bound to the outer membrane with the bulk of the protein apparently located in the intermembrane space of the envelope (19). These features distinguish it from the soluble cytoplasmic and matrix hsp70s that are essential for protein import into mitochondria.
Several groups have proposed that the two proteins function as molecular chaperones on either side of the outer membrane (19,26). The presence of the chaperones would serve to maintain the import competence of the precursor as it enters the import machinery on the cis side of the membrane (Com70) and as it emerges into the intermembrane space on the trans side of the membrane (hsp70-IAP) prior to engaging the inner membrane import machinery. This proposal is attractive as it is consistent with models in which hsp70 chaperones provide the driving force for membrane translocation (27,28) and also because the binding of the chaperones would account for the ATP requirement for the formation of early import intermediates.
Recent studies have suggested that the unique lipid composition of the envelope may play an important role in the specific targeting of precursor proteins to the outer membrane. Model membrane studies have shown that chloroplast preproteins and transit peptides interact specifically with outer membrane lipids and that this binding is specific for the transit sequence (for a summary see Ref. 29). Of particular interest is the observation that the lipid-protein interactions induce regular secondary structure into the normally disordered structure of transit sequences observed in aqueous solutions (30). These observations suggest that an initial, reversible interaction of transit sequences with the lipid component of the bilayer may induce regular structure, thereby facilitating the binding of the transit sequence to receptor components of the import machinery.
Late stage import intermediates have been identified that span the inner envelope membranes (5,15,26), and several candidates for components of the inner membrane import machinery have been identified (Fig. 1). Two envelope components, IAP100 (Cim97) (5, 26) and IAP36 (5), copurify with late stage import intermediates. IAP36 has not been characterized, but IAP100 is an integral inner membrane protein with a single apparent transmembrane domain (31,32). The IAP100 associates with the the plastid hsp60 homolog in an ATP-dependent manner and has been proposed to function in recruiting the hsp60 chaperone to the site of import, thereby facilitating folding of newly imported proteins (31). Two additional proteins, IAP21 and IAP25, recently have been shown to crosslink directly to the transit sequence of an early import intermediate (14). The localization of IAP25 has not been confirmed, but IAP21 appears to be an inner membrane protein. The function of neither protein has been investigated in detail, but their interactions with the transit sequence make them candidates for components that mediate the presentation of the precursor to the import machinery of the inner membrane.
A set of at least two immunologically related envelope proteins of 44 kDa, designated Com44/Cim44, has been detected in a covalently cross-linked aggregate from envelope membranes that contain a partially translocated import intermediate (26). Variants of the Com44/Cim44 have been localized to both the outer and inner envelope membrane (33), but their role in translocation remains to be investigated in detail.

The Multiple Pathways for Thylakoid Targeting
Nuclear encoded thylakoid proteins contain dual targeting signals (for review see Ref. 1) that direct a two-step targeting process. The primary signal is a cleavable stromal targeting domain (STD) that directs the proteins across the envelope into the stroma and is removed by the SPP. These sequences are structurally and functionally equivalent to the transit sequences of stromal proteins. SPP processing generates a stromal intermediate that is targeted to the thylakoid membrane or lumen by a secondary targeting signal. Lumenal proteins contain a bipartide transit sequence (34) composed of an STD followed in tandem by a cleavable thylakoid lumen targeting domain (LTD). The targeting information for thylakoid membrane proteins is contained within the mature sequence of the polypeptide and most likely resides within one or more of the transmembrane domains (35)(36)(37). Envelope translocation and thylakoid targeting can be separated in in vitro assays indicating that they are not obligatorily coupled processes.
The elegant and systematic analyses of in vitro targeting assays using intact chloroplasts or isolated thylakoids have resulted in the definition of at least three, and perhaps four, distinct pathways for thylakoid targeting (Fig. 2). These pathways are distinguished by their energetic requirements and specificities for protein substrates (for review see Ref. 1). The first pathway is represented by targeting of the 33-kDa subunit of the oxygen-evolving complex (OE33) and plastocyanin (PC) to the thylakoid lumen. Both proteins contain bipartide, cleavable transit peptides. Translocation into the lumen requires ATP hydrolysis and is stimulated by a pH gradient (⌬pH) across the thylakoid membrane (38 -41). The energetic requirements for translocation and the similarity of the LTDs of these precursor proteins to bacterial signal sequences suggested that this class of proteins may use a mechanism related to the general pathway for bacterial protein export. In fact, the LTDs have been shown to function as signal sequences in Escherichia coli and are correctly cleaved by the bacterial signal peptidase (42). The hypothesis for a conserved mechanism of thylakoid targeting has now been confirmed. A chloroplast homolog of the E. coli SecA protein (CPSecA) has been identified in the stroma of pea chloroplasts and has been shown to be essential for the targeting of the stromal intermediates of OE33 and PC to thylakoids in in vitro assays (43,44). This function is analogous to the function of SecA in the targeting of preproteins to the SecY/E/G protein translocon of the E. coli inner membrane (see minireview by Wickner and Leonard in this series (53)). Recently, a homolog of the E. coli SecY protein (CPSecY) has been identified in Arabidopsis thaliana and has been localized to the thylakoid membrane (45). Although the function of CPSecY still is under investigation, it is predicted to function as part of the protein translocon in the thylakoid membrane. Its discovery provides additional support for the hypothesis that the OE33/PC pathway for thylakoid translocation is analogous to the general pathway of bacterial export.
The second targeting pathway also encompasses thylakoid lumen proteins with bipartide transit sequences and is exemplified by the targeting of the 17-kDa (OE17) and 23-kDa (OE23) subunits of the oxygen-evolving complex of photosystem II. Although the presence of a bipartide transit sequence suggests that these precursor proteins may use a CPSecA pathway, their translocation into the lumen requires only a ⌬pH at the thylakoid membrane and is not competed by the PC or OE33 stromal intermediates (38,40,46). In addition, CPSecA has been shown not to stimulate translocation of the OE17 or OE23 intermediates into isolated thylakoids (43). The requirement for only a ⌬pH is unique among known protein translocation systems, but the electrochemical potential is an essential component of bacterial protein export suggesting that the ⌬pH pathway also may retain elements in common with this system.
The third pathway for thylakoid targeting is represented by the insertion of the integral, major light-harvesting complex protein (LHCP) into the thylakoid membrane. LHCP integration is directed by one or more regions within the transmembrane domains of the protein, is GTP-dependent, and is stimulated by the pH gradient at the thylakoid membrane (47). Partial reconstitution of the LHCP integration using stromal extracts and isolated thylakoids has identified one of the essential components of the transit complex (48). Remarkably, this component is a chloroplast homologue (54CP) of the 54-kDa subunit of the signal recognition particle (SRP) and the E. coli SRP (49). The discovery of CP54 suggests that it may mediate the delivery of the LHCP transit complex to the thylakoid membrane through a GTP binding and hydrolysis cycle similar to that used by SRP in targeting the nascent chainribosome complex to the ER (24). Thus, LHCP integration represents a second conservative targeting pathway that is homologous to protein targeting pathways that function at the endoplasmic reticulum and the bacterial inner membrane.
The existence of a fourth pathway for thylakoid targeting has been suggested by the recent discovery that the integral membrane protein, CF 0 II, integrates into the membrane independent of a ⌬pH, nucleoside triphosphates, or stromal factors (50). The mechanism of CF 0 II integration remains to be examined in detail, but the observations have prompted the investigators to propose that this protein inserts into the thylakoid membrane independent of proteinaceous factors.

In Vivo Models for Thylakoid Protein Targeting
The assignment of distinct thylakoid targeting mechanisms and studies of their biochemical nature are supported by the recent development of in vivo genetic models in higher plants and green algae. Two maize mutants have been identified that specifically affect distinct targeting pathways (51). The tha1 mutant is specifically affected in the OE33 and PC targeting pathway and not in the other pathways. This mutant also affects the targeting of a plastid-encoded protein, cytochrome f, indicating that plastid and nuclear encoded proteins can share a targeting mechanism. Cytochrome f is synthesized on plastid ribosomes as a precursor with a typical amino-terminal LTD. The tha1 gene has recently been cloned and has been shown to be the maize CPSecA (1). 2 A second mutant, hcf106, is defective only in OE17 and OE23 targeting. Neither mutant affects the integration of LHCP.
The analysis of a class of mutations in the LTD of cytochrome f of Chlamydomonas suggests that although certain elements of the targeting pathways of each class of thylakoid proteins may be unique, they may also share common components (e.g. the components of the membrane translocation machinery) (52). One of these mutants showed a dominant-negative effect on the the integration of LHCP and a second plastid-encoded protein, D1. These results suggest that cytochrome f, D1, and LHCP may share common elements at one point in their translocation process. Suppressors of the cytochrome f mutants have 2 A. Barkan, personal communication. been identified, and their analysis should provide important in vivo evidence that complements the biochemical studies of thylakoid targeting.

Perspective
Research over the last 2 years has provided a bountiful harvest of information on the pathways and components of protein targeting at the chloroplast envelope and the thylakoid membrane. The components that have been identified provide the needed markers with which to identify additional components. Within the not-so-distant future, it should be possible to reconstitute part or all of these targeting reactions and, thereby, address more fundamental questions such as the exact roles of nucleoside triphosphates or the ⌬pH in regulating and/or driving membrane transport. The identification of mutants affected in thylakoid targeting provides encouraging signs that in vivo models to address such fundamental questions can now be applied to protein targeting in chloroplasts. Molecular genetic systems aimed at manipulating components of the import apparatus are now being developed to provide in vivo models for envelope translocation. The knowledge obtained from the combination of these approaches will be essential for understanding plastid biogenesis but should also have important implications for understanding the general principles that govern protein translocation in other systems. For example, the relationship among the CPSecA and 54CP pathways and their bacterial and ER counterparts should provide knowledge of the essential, conserved elements of membrane translocation in these systems. In contrast, the apparent lack of similarity between chloroplast and mitochondrial import systems should provide insight into the distinct evolution of translocation systems as a consequence of dual endosymbiotic events.