JXB Advance Access originally published online on August 8, 2005
Journal of Experimental Botany 2005 56(419):2287-2320; doi:10.1093/jxb/eri243
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
REVIEW ARTICLE |
Recognition and envelope translocation of chloroplast preproteins
Department of Biology, University of Leicester, University Road, Leicester LE1 7RH, UK
* To whom correspondence should be addressed. Fax: +44 (0)116 252 3330. E-mail: rpj3{at}le.ac.uk
Received 14 March 2005; Accepted 7 June 2005
 |
Abstract |
|---|
Top
Abstract
IntroductionPlastids are a diverse group of plant organelles that perform essential functions including important steps in many biosynthetic pathways. Chloroplasts are the best characterized type of plastid, and constitute the site of oxygenic photosynthesis in plants, a process essential to all higher life forms. It is well established that the majority (>90%) of chloroplast proteins are nucleus-encoded and must be post-translationally imported into these envelope-bound compartments. Most nucleus-encoded chloroplast proteins are translated in precursor form on cytosolic ribosomes, targeted to the chloroplast surface, and then imported across the double-membrane envelope by translocons in the outer and inner envelope membranes of the chloroplast, termed TOC and TIC, respectively. Recently, significant progress has been made in our understanding of how proteins are targeted to the chloroplast surface and translocated across the chloroplast envelope into the stroma. Evidence suggesting the existence of multiple import pathways at the outer envelope membrane for different classes of precursor proteins has been presented. These pathways appear to utilize similar TOC complexes equipped with different combinations of homologous GTPase receptors, providing preprotein recognition specificity.
Key words: Chloroplast, chloroplast protein import, membrane translocation, organelle biology, plastid, protein targeting, Tic, Toc, translocon
|
Introduction |
|---|
Plastids, their evolution and functions
Chloroplasts belong to a family of organelles known as plastids. They are ubiquitous in plants and algae, and constitute the site of photosynthesis in these photoautotrophic organisms. Oxygenic photosynthesis converts absorbed sunlight energy into a usable form of chemical energy, and is essential for plant growth and development. The redox reactions associated with photosynthesis lead to the generation of oxygen from water, and provide the energy necessary for the fixation of carbon dioxide. Photosynthesis is the most important producer of oxygen and organic matter on earth, and therefore provides the fuels essential for all higher forms of life (Nelson and Ben-Shem, 2004
).
Different types of non-photosynthetic plastids can be found in non-green tissues of plants (e.g. leucoplasts and chromoplasts), and these usually carry out specialized biosynthetic and storage functions (Whatley, 1978; Mullet, 1988). For example, leucoplasts normally accumulate lipids, starch, or proteins and can be found in petals and roots, whereas chromoplasts accumulate carotenoid pigments that provide colour in fruits, flowers, and senescing leaves. However, all plastids are separate cellular compartments containing their own genetic machinery and play essential roles in numerous biosynthetic pathways. These include steps in amino acid, lipid, nucleotide, hormone, vitamin, and secondary metabolite biosynthesis (Leister, 2003).
It is generally accepted that plastids originated from an endosymbiotic event in which a eukaryotic cell engulfed a photosynthetic prokaryote, an ancestor of today's cyanobacteria (Margulis, 1970; McFadden, 2001). These organisms thereafter developed a symbiotic relationship in which the photosynthetic prokaryote provided energy to the host, possibly in exchange for essential nutrients and protection. Through evolution, the endosymbiont lost its autonomy and came under the genetic control of the host cell. This occurred through the process of gene transfer, in which genes from the prokaryotic genome were transferred to the nuclear genome of the eukaryotic host cell (for a review, see Timmis et al., 2004). The mechanism by which gene transfer occurs is largely unknown, but transfer of plastid DNA to the nucleus has been experimentally demonstrated to occur at a relatively high frequency (Huang et al., 2003).
Gene transfer allowed the integration of plastid development into the developmental programme of plants. Chloroplasts, as well as other types of mature plastids, generally develop from undifferentiated plastids termed proplastids. Proplastids are found in undifferentiated meristematic cells and undergo division in parallel with cell division in such a way that a constant number of proplastids are found in all cells. Indeed, new plastids are not assembled de novo within plant cells but arise only from the division of pre-existing plastids (for a review, see Aldridge et al., 2005). As meristematic cells differentiate into different cell types, similarly, proplastids differentiate into different plastid types. For example, in leaves, as meristem cells develop into mesophyll cells, proplastids differentiate into chloroplasts, a process characterized by the rapid accumulation of photosynthetic proteins and the green chlorophyll pigments used to absorb light energy.
The processes of plastid division, differentiation and maturation are largely under the control of the plant cell nucleus. The cell exerts control over these events mainly by regulating the expression of plastid proteins. As a consequence of gene transfer, greater than 90% of chloroplast proteins are nucleus-encoded (Leister, 2003). Thus, plant and algal cells have had to engineer a way to ensure that plastid proteins, now translated in the cytosol, are faithfully and efficiently targeted to and imported into plastids.
General principles of chloroplast protein import
The study of plastid protein biogenesis has largely been limited to the study of chloroplast proteins. The predominance of chloroplasts over other types of plastids in plants, their ease of isolation, and the enormous importance of the photosynthetic functions that they perform are the reasons why they have received the most attention. Chloroplasts, like all plastid types, are delimited by a double membrane envelope and, in addition, contain a third membrane system, the thylakoid membranes, in which are integrated and assembled the various photosynthetic protein complexes. The chloroplast is consequently made up of three different membranes, namely the outer envelope membrane (OEM), the inner envelope membrane (IEM), and the thylakoid membrane, and three aqueous sub-compartments, namely the inter-membrane space (IMS), the stroma, and the thylakoid lumen. In order for nucleus-encoded chloroplast proteins to be correctly integrated into this complex organelle, elaborate protein sorting mechanisms are essential. Because the different membranes of the chloroplast are highly impermeable to large molecules such as proteins, selective protein translocation machineries are necessary to allow newly translated chloroplast proteins to reach internal sub-compartments.
During the last decade, a considerable amount of progress has been made in our understanding of how proteins are targeted to the chloroplast, and how these reach their final destination. Most nucleus-encoded proteins destined to the internal compartments (i.e. the IMS, IEM, stroma, thylakoid membrane, and thylakoid lumen) of the chloroplast are expressed as precursor proteins (preproteins) carrying an N-terminal extension called a transit peptide (TP). The TP plays an important role in the targeting of the preprotein to the chloroplast, as well as in directing its translocation across the envelope membranes (Bruce, 2001). It is generally believed that preproteins are post-translationally imported into the chloroplast, and that their targeting to the organelle is assisted by soluble cytosolic factors that maintain them in a loosely-folded, import-competent conformation. Once a preprotein arrives at the chloroplast surface, translocons at the outer and inner envelope membranes of the chloroplast, termed TOC and TIC, respectively, mediate its translocation (Fig. 1). These translocons are hetero-oligomeric protein complexes that actively drive the movement of the preprotein across the envelope membranes with the help of various molecular chaperones. The individual proteins making up these complexes are named using the Toc or Tic prefix, followed by a number corresponding to the calculated molecular weight of the mature protein in kilodaltons (kDa) (Schnell et al., 1997).
|
Preproteins arriving at the chloroplast surface are recognized by receptor components of the TOC. The receptors specifically and selectively interact with the TP, before they mediate its insertion into the channel of the TOC. The receptor components may also act to gate the protein-conducting channel in order to maintain a barrier at the outer membrane. It is widely believed that the TOC and TIC are able to interact physically in order to allow the direct translocation of proteins simultaneously through the outer and inner envelope membranes. Components of both the TOC and the TIC protruding into the IMS, and possibly soluble IMS proteins, are likely to promote the formation of this TOC–TIC supercomplex. The TIC complex mediates the translocation of the preprotein through the IEM, and is believed to do so with the help of chaperones recruited at the stromal face of the TIC. As the preprotein emerges into the stroma, the TP is cleaved off by a stromal processing peptidase (SPP). Molecular chaperones may then assist the processed protein to reach its functional conformation, or, alternatively, the protein may be further directed to other sub-compartments of the chloroplast by additional targeting signals.
The translocation of preproteins across the chloroplast envelope is an energy-dependent process. By characterizing the import of preproteins into isolated chloroplasts in vitro, three basic steps of the translocation process have been defined based on their energy requirements. First, the initial interaction of preproteins with the TOC receptors is an energy-independent and reversible step called energy-independent binding (Perry and Keegstra, 1994). Second, under energy-limiting conditions, initial translocation leads to the formation of early import intermediates. These intermediates are formed most efficiently in the presence of GTP and low levels of ATP (
100 µM) (Olsen and Keegstra, 1992; Young et al., 1999), and correspond to precursors in which the TP has crossed the OEM and is in contact with components of the TIC complex (Wu et al., 1994; Ma et al., 1996). Similar early import intermediates can be generated by exposing preproteins to energized chloroplasts at low temperatures (
4 °C) (Leheny and Theg, 1994). In the literature, the formation of early import intermediates has also been called binding and docking; in this review, in order to be consistent and to avoid confusion with energy-independent binding, these alternative terms will be avoided. Lastly, in order for complete preprotein translocation to occur, higher levels of ATP (>100 µM) are required in the stroma (Theg et al., 1989). This higher energy requirement is attributed to stromal molecular chaperones (Pain and Blobel, 1987).
The aim of this review is to provide a detailed synthesis of the research that has improved our understanding of how preproteins bearing transit peptides are recognized at the chloroplast surface, and then imported into the stroma by the TOC and TIC complexes. Other, less well understood pathways of protein import into the chloroplast are not discussed in detail. These include protein insertion into the OEM and IEM, protein import into the IMS, and other putative, alternative pathways for protein import into the chloroplast (Jarvis, 2004; Jarvis and Robinson, 2004). Proteins destined to the chloroplast OEM generally lack a transit peptide, and have been proposed to insert directly into the outer envelope membrane, independently of the TOC apparatus (for a review, see Schleiff and Klösgen, 2001). However, recent studies on outer envelope protein 14 kDa (OEP14) from pea (Pisum sativum), a model membrane protein carrying a single transmembrane domain at its N-terminus, have revealed that insertion is assisted by Toc75, perhaps disassociated from the TOC complex (Tu and Li, 2000; Tu et al., 2004); Toc75 is the channel-forming component of the TOC complex (see section entitled, ‘The TOC channel’).
Other recent data suggest that alternative import pathways, distinct from the general TOC–TIC import pathway, do exist for certain proteins (Jarvis, 2004). For example, two nucleus-encoded chloroplast IEM proteins, the chloroplast envelope quinone oxidoreductase homologue (ceQORH) and Tic32, have each been found to lack a cleavable TP (Miras et al., 2002; Hörmann et al., 2004; Nada and Soll, 2004). Both proteins are targeted to chloroplasts by signals contained in their mature sequences, and, in addition, Tic32 was shown to be imported independently of the general TOC–TIC pathway. It is likely that these two proteins represent only the first of many, however, since a proteomics study recently identified other nucleus-encoded chloroplast proteins that appear to lack canonical TP sequences (Kleffmann et al., 2004). Interestingly, Kleffman et al. (2004) also identified nucleus-encoded chloroplast proteins that carry N-terminal sequences similar to the signal peptides that direct proteins to the endoplasmic reticulum. These proteins may therefore utilize a chloroplast targeting pathway involving the endomembrane system. Alternatively, since outer envelope proteins such as OEP14 were also found to carry signal-peptide-like N-termini (Lee et al., 2001, 2004), it is possible that these proteins follow a similar targeting pathway involving Toc75. Finally, the chlorophyll biosynthetic enzyme, NADPH:protochlorophyllide oxidoreductase (POR) A, has been proposed to utilize a substrate-dependent import pathway that is independent of the general TOC–TIC pathway (Reinbothe et al., 2000). In vivo data in support of this novel import pathway have been presented (Kim and Apel, 2004; Kim et al., 2005), and components of a putative import apparatus have been identified (Reinbothe et al., 2004a, b, 2005), but the existence of this pathway has been strongly disputed by others (Aronsson et al., 2000, 2001, 2003b; Dahlin et al., 2000; Jarvis and Soll, 2002).
Following import, or envelope translocation, four different pathways are known to be responsible for the further targeting of proteins from the stroma to the thylakoids. These pathways, which are not discussed in detail here, include the signal recognition particle (SRP)-dependent and ‘spontaneous’ insertion pathways for the insertion of proteins into the thylakoid membrane, and the Sec and Tat pathways for the translocation of proteins into the thylakoid lumen. Unlike the general protein translocation pathway of the chloroplast envelope described earlier, which appears to have evolved as a response to gene transfer and shows no clear homology to other known protein translocation systems, most of these thylakoidal pathways are highly homologous to protein translocation systems of the bacterial cell membrane. Like the protein import pathway of the envelope, these pathways are important for the proper biogenesis of the photosynthetic machinery associated with the thylakoids, and have been the subject of intensive research. For more general reviews of chloroplast protein targeting, including thylakoid protein import, readers are referred to Keegstra and Cline (1999) and Jarvis and Robinson (2004).
History of chloroplast protein import research
Chloroplast protein import research stretches back over several decades. Blair and Ellis (1973) proposed an ‘envelope carrier hypothesis’ to account for the fact that isolated pea chloroplasts are able to synthesize only a small number of their constituent proteins (including the large subunit of Rubisco, RbcL); the rest (including the small subunit of Rubisco, RbcS), it was correctly presumed, are synthesized outside of the chloroplast on cytosolic ribosomes. The hypothesis proposed the existence of specific mechanisms for the transport into chloroplasts of all those chloroplast proteins which are synthesized in the cytosol: ‘one possibility is that a membrane protein exists in the outer envelope which recognizes a site common to those proteins destined for the plastid’ (Blair and Ellis, 1973). The existence of such a ‘recognition site’ received considerable support when Dobberstein et al. (1977) demonstrated that Chlamydomonas reinhardtii RbcS is synthesized as a high molecular weight 20 kDa precursor that can be processed to a mature form of 15 kDa by a specific endoproteolytic activity present in Chlamydomonas cells. The final proof of the ‘envelope carrier hypothesis’ was provided in 1978 when two groups independently reconstituted the import of RbcS precursor (pRbcS) into isolated chloroplasts (Chua and Schmidt, 1978; Highfield and Ellis, 1978); in each case, precursor processing was observed to occur concomitantly with chloroplast import. Import occurred after precursor synthesis was complete, or post-translationally, which is consistent with the observation that the pRbcS is synthesized on free cytosolic ribosomes (Dobberstein et al., 1977), and with the fact that the chloroplast surface, unlike that of the endoplasmic reticulum (ER), is not coated with ribosomes. A year later, the amino acid sequence and N-terminal position of the Chlamydomonas RbcS TP was determined (Schmidt et al., 1979).
The nature of the envelope protein(s) mediating chloroplast protein import took longer to establish. Treatment of isolated chloroplasts with proteases such as thermolysin, which degrades only surface-exposed protein domains, provided the initial evidence for the existence of a proteinaceous import apparatus (Cline et al., 1984, 1985; Friedman and Keegstra, 1989). The demonstration that precursor binding to chloroplasts was saturable provided further support for the existence of specific protein import sites at the chloroplast surface (Friedman and Keegstra, 1989). The application of protein cross-linking techniques and the isolation of translocation complexes, led to the tentative identification of import apparatus components (Waegemann and Soll, 1991; Perry and Keegstra, 1994), and in 1994 and 1995 the identity of the first TOC and TIC components was firmly established (Hirsch et al., 1994; Kessler et al., 1994; Schnell et al., 1994; Wu et al., 1994; Seedorf et al., 1995; Tranel et al., 1995). Since then, several more components of the import apparatus have been identified, and the exact sequence of events that occurs during import has been established (Keegstra and Cline, 1999; Chen et al., 2000b; Hiltbrunner et al., 2001a; Jarvis and Soll, 2002). More recently, attention has turned to the determination of the specific functions performed by individual TOC and TIC components, as discussed in detail in this review. In addition, the model organism, Arabidopsis thaliana, has been utilized so that genomics, genetics, and other in vivo approaches can be used to complement the biochemical approaches that have been used traditionally, with pea chloroplasts, to study protein import (Jarvis et al., 1998; Bauer et al., 2000; Chen et al., 2002; Chou et al., 2003).
Overview of protein import into mitochondria
Mitochondria are organelles found in almost all eukaryotic cells. They are mainly responsible for the generation of usable energy by oxidative respiration, but also perform other essential metabolic functions. In many ways, mitochondria are strikingly similar to chloroplasts. They also arose from an endosymbiotic event, and their evolution similarly involved gene-transfer from the endosymbiont to the nucleus of the eukaryotic progenitor cell. Like chloroplasts, mitochondria are semi-autonomous organelles: they retain a functional genetic system and a few protein-coding genes, but most mitochondrial proteins are nucleus-encoded and translated on cytosolic ribosomes. Thus, in order to carry out their metabolic functions, mitochondria must import most of their proteome post-translationally from the cytosol. Mitochondria are composed of two membranes, namely the outer membrane and the inner membrane, and two aqueous compartments, namely the inter-membrane space and the matrix. Since the majority of mitochondrial proteins are found either in the matrix or the inner membrane, each membrane is equipped with hetero-oligomeric protein complexes which mediate protein translocation across the hydrophobic membrane barriers. These are the translocases of the outer and inner membranes of mitochondria, or the TOM and TIM complexes, respectively.
Most components of the chloroplast envelope and mitochondrial translocases are not homologous, but the processes of chloroplast protein import into the stroma and mitochondrial protein import into the matrix appear to be generally analogous. Due to the importance of mitochondrial biology for human health, and the ease of working with yeast as a model organism, the processes of mitochondrial protein targeting and import are understood in much greater detail than the analogous processes in chloroplasts. Because of this, mitochondrial import is often used as a model to predict how chloroplast protein targeting and import may function. A short summary of mitochondrial preprotein targeting and import is therefore included below, in order to provide background for the comparisons that will be drawn throughout this review. For a more detailed description of protein import into mitochondria, readers are referred to the following reviews: Neupert and Brunner (2002), Endo et al. (2003), Truscott et al. (2003), Rehling et al. (2004), and Wiedemann et al. (2004).
Three general mitochondrial protein import pathways have been defined in yeast and animal cells. (i) The presequence import pathway utilized by preproteins carrying an N-terminal targeting sequence (presequence), which includes most proteins targeted to the mitochondrial matrix and some proteins of the inner membrane and inter-membrane space. (ii) The carrier protein import pathway utilized by inner membrane proteins that carry multiple internal targeting signals. (iii) The outer membrane protein insertion pathway utilized by ß-barrel proteins which also carry internal targeting signals. All three pathways share a common path through the outer membrane at the TOM complex. The TOM complex is equipped with different receptors that can recognize preproteins carrying presequences and internal targeting signals. Once translocation across the outer membrane has been achieved, the different pathways diverge. Outer membrane ß-barrel proteins are targeted to the sorting and assembly machinery (SAM), which assists their insertion into the membrane, whereas preproteins following the other two pathways are targeted to distinct TIM complexes. Presequence preproteins are targeted to the presequence translocase, or TIM23 complex, and carrier proteins are targeted to the protein insertion machinery of the inner membrane, or TIM22 complex.
The import pathway followed by chloroplast stromal proteins carrying an N-terminal TP is generally believed to be most analogous to the presequence protein import pathway of mitochondria. After initial binding at the TOM receptors, the positively-charged presequence of mitochondrial precursors is thought to be translocated through the outer membrane in step-wise fashion, by progression along a chain of binding sites on the TOM translocase. Forward movement of the preprotein is driven by the increasing affinity of the binding sites for the presequence along the translocase. Like the TOC and TIC of the chloroplast envelope, the TOM and TIM are known to form supercomplexes, allowing for simultaneous translocation through both membranes. Translocation of the presequence through the TIM translocase is driven by a pulling force exerted by the electric potential across the inner membrane. Once the presequence emerges at the inner face of the TIM, translocation of the preprotein is completed by the action of a molecular chaperone in the matrix, called matrix Hsp70 (mtHsp70). Together with soluble and membrane-bound co-chaperones, mtHsp70 forms a presequence translocase-associated protein import motor (PAM) that drives protein import in an ATP-dependent fashion. Translocation is completed by the removal of the presequence by the mitochondrial processing peptidase (MPP).
|
Targeting of preproteins from the cytosol to the chloroplast |
|---|
N-terminal targeting signals are commonly used in eukaryotic cells to direct nucleus-encoded proteins to various cellular sub-compartments, including the ER, mitochondria, and chloroplasts (Blobel and Sabatini, 1971
; Schnell and Hebert, 2003). In order to prevent mis-targeting of proteins, these signals must be specifically recognized by the cellular protein sorting machinery. Like the presequence of mitochondrial preproteins, the TP of chloroplast precursors is an essential targeting signal and plays several key roles to ensure that the protein reaches its destination (Bruce, 2001). First, as preproteins are synthesized on cytosolic ribosomes, TPs are probably recognized by soluble cytosolic factors that assist the targeting of the preprotein to the chloroplast. Later, as they reach the chloroplast, they are known to interact directly and specifically with receptors of the TOC complex, leading to translocation initiation. TPs also interact with other downstream components of the translocation machinery, and are therefore likely to play important roles to ensure the progression of the protein through the TOC and TIC complexes. Finally, they must be correctly recognized by the stromal processing machinery so that preproteins can attain their mature and functional form, or so that downstream N-terminal targeting signals are recognized to allow further targeting (Keegstra and Cline, 1999; Jarvis and Robinson, 2004).
Surprisingly, although TPs play such important roles, they are highly variable in length, amino acid composition, and sequence. Typically, higher plant TPs are between 20 and 100 residues in length, are rich in hydrophobic and hydroxylated residues, and almost completely lack acidic residues such that they show an overall positive charge (Zhang and Glaser, 2002). Since TPs lack detectable blocks of conserved amino acid sequence, they are thought to share similar secondary or tertiary structural properties that would allow their specific recognition and targeting. The detailed analysis of a few TPs from a small number of algal and higher plant proteins suggests that they are largely unstructured in aqueous environments and form a random coil (von Heijne and Nishikawa, 1991; Bruce, 2001). However, they have been found to form amphipathic helix domains when exposed to membrane-like environments. These properties are very similar to those of mitochondrial presequences, and it is therefore unclear how specificity of targeting is achieved in plants, which possess both types of organelle (Macasev et al., 2000; Chew and Whelan, 2004). The targeting sequences of plant mitochondrial and chloroplast proteins are proposed to be made of several domains, which may mediate their different functions and may also display different properties that could account for the specificity of targeting (Bruce, 2000).
The proposed, largely unstructured nature of presequences and TPs in the cytosol has led to the suggestion that they may serve as ideal substrates for the binding of Hsp70 molecular chaperones (von Heijne and Nishikawa, 1991). Indeed, the majority of known presequences and TPs have been found to display at least one putative Hsp70 binding site, and several have been shown to interact with Hsp70s experimentally (Ivey and Bruce, 2000; Ivey et al., 2000; Rial et al., 2000; Zhang and Glaser, 2002). Cytosolic Hsp70s, therefore, probably interact with the TP in the cytosol, as it emerges from the ribosome, in order to prevent preprotein aggregation and to maintain a translocation-competent conformation. In addition, Hsp70s have been suggested to participate in protein translocation across the chloroplast envelope membranes at several stages, at both the outer and inner faces of the outer envelope, as well as in the stroma (Marshall et al., 1990; Tsugeki and Nishimura, 1993; Schnell et al., 1994; Kourtz and Ko, 1997). It is possible that the TP is important in promoting binding by these Hsp70s during translocation.
Interestingly, TPs of chloroplast preproteins are proposed to be phosphorylated at a serine or threonine residue within a loosely conserved motif by a soluble, cytosolic, ATP-dependent kinase (Waegemann and Soll, 1996). Furthermore, the phosphopeptide motif of TPs appears to be specifically recognized by 14-3-3 proteins (May and Soll, 2000). Because plant mitochondrial preproteins are neither phosphorylated nor recognized by 14-3-3 proteins, it was proposed these TP-specific events may confer specificity for targeting to the chloroplast. It was further shown that chloroplast preproteins translated and phosphorylated in a wheat-germ lysate associate with 14-3-3 and Hsp70 proteins in a complex termed the guidance complex. Preproteins assembled into such hetero-oligomeric complexes were imported into isolated chloroplasts at a rate three- to four-times higher than free preproteins, suggesting that these complexes constitute true targeting intermediates. Recently, however, the physiological relevance of TP phosphorylation and 14-3-3 protein interaction was disputed, since disruption of the TP phosphorylation site did not affect targeting or import in vivo (Nakrieko et al., 2004).
The change in structure of parts of the TP from a random coil to an amphipathic helix in membrane-mimetic conditions led to the proposal that chloroplast preproteins may interact specifically with the unique lipids of the OEM prior to association with TOC receptors (Bruce, 2000). Although this remains a possibility, the TP has now been shown to interact directly and specifically with recombinant TOC receptors in aqueous solution (Sveshnikova et al., 2000b; Becker et al., 2004b; Ivanova et al., 2004; Smith et al., 2004). It is possible that a similar change in structure may be induced directly via an interaction with the receptor proteins at the TOC, since mitochondrial presequences have been shown to assume an amphipathic helical structure upon interaction with the Tom20 receptor (Abe et al., 2000; Muto et al., 2001). The amphipathic nature of the helices formed by chloroplast preproteins appears to be provided by the clustering of hydroxylated residues on one side of the helix, rather than by the clustering of basic residues as observed in mitochondrial presequence helices (Bruce, 2000). The different nature of these secondary structures may therefore also contribute to the specificity of targeting. However, a larger number of TP structures and the nature of the interactions between TPs and the TOC receptors must be established before this can be confirmed.
Finally, the TOC component, Toc159, has recently been proposed to play a role in preprotein targeting from the cytosol to the chloroplast. Toc159 has traditionally been thought of as a receptor for chloroplast precursor proteins acting exclusively at the chloroplast OEM (see below). However, re-investigation of its localization profile suggested that a large proportion (50%) of the protein behaves as a soluble cytosolic protein (Hiltbrunner et al., 2001b). This prompted the suggestion that Toc159 may shuttle between the cytosol and the chloroplast OEM to mediate the transport of nucleus-encoded preproteins between their site of synthesis and the TOC complex. This particular role of Toc159 needs to be investigated further, however, since the existence of a soluble form of the protein has been disputed (Becker et al., 2004b).
|
Preprotein binding and recognition at the TOC complex |
|---|
Receptor components of the TOC complex
Preprotein recognition at the OEM is mediated by at least two receptor components of the TOC complex, Toc34 and Toc159 (Fig. 1). These were originally identified as proteins in close proximity to preproteins bound to the chloroplast OEM translocon under energy-limiting conditions at the initial stage of translocation (Perry and Keegstra, 1994
; Schnell et al., 1994). Both proteins display homology to GTPases (Bourne et al., 1991), and show substantial levels of similarity to each other across their GTPase (G) domains. In Toc34, the G-domain spans almost the entire length of the protein except for short amino acid extensions at the N- and C-termini that lack homology with Toc159. The C-terminal extension contains a single, short stretch of hydrophobic amino acids, predicted to form a transmembrane 
-helix, which anchors Toc34 in the OEM (Kessler et al., 1994; Seedorf et al., 1995). By contrast, in Toc159, the central G-domain is flanked by a large acidic (A) domain at the N-terminus, and an unusual membrane-anchoring (M) domain at the C-terminus (Bölter et al., 1998a; Chen et al., 2000a). Initially, Toc159 was identified as an 86 kDa proteolytic fragment (Toc86) lacking the large A-domain. The full-length pea protein was only recently discovered (Bölter et al., 1998a) and characterized (Chen et al., 2000a) as a result of the Arabidopsis genome project (The Arabidopsis Genome Initiative, 2000), which led to the identification of a homologous gene encoding a protein much larger than 86 kDa. The A-domain of Toc159 was found to be rich in acidic residues, and to be essentially made up of two different tandemly-repeated motifs (Chen et al., 2000a).
Isolated chloroplasts in which the Toc159 protein remained intact displayed a significantly greater import rate than those in which Toc159 had been proteolysed (Bölter et al., 1998a), demonstrating that this non-essential domain does perform some role in import. Due to its acidic nature, the A-domain was proposed to interact directly with the basic TPs of precursor proteins and to help mediate the preprotein receptor function of Toc159. The M-domain of Toc159 lacks a clear stretch of hydrophobic amino acids that could span the membrane as an -helix, and it predominantly consists of hydrophilic residues (Hirsch et al., 1994; Kessler et al., 1994). However, the M-domain was shown to be necessary for the correct insertion of Toc159 into the OEM (Hirsch et al., 1994; Kessler et al., 1994). In addition, it was shown to become highly resistant to proteases as well as to salt and alkaline extraction upon membrane insertion, indicating that it behaves as a true integral membrane anchor (Hirsch et al., 1994; Kessler et al., 1994). The bulk of Toc34 and a large proportion of Toc159, consisting of its A- and G-domains, protrude into the cytosol where they can interact specifically with the TPs of incoming preproteins (Hirsch et al., 1994; Kessler et al., 1994; Ma et al., 1996; Chen et al., 2000a; Sveshnikova et al., 2000b).
A third putative preprotein receptor, Toc64, has been described (Sohrt and Soll, 2000). Toc64 was initially identified as a pea OEM protein co-fractionating with the solubilized TOC complex in sucrose density gradients. It was further shown, by cross-linking in isolated chloroplasts, to be in close physical proximity with known components of the TOC and TIC complexes, as well as with an arrested translocation intermediate. Interestingly, Toc64 carries three tetratricopeptide repeat (TPR) motifs that are exposed in the cytosol. Such TPR domains have previously been shown to mediate protein–protein interactions (Lamb et al., 1995), and commonly mediate the functional interaction of proteins with molecular chaperones (e.g. Hsp70, Hsp90) (Frydman and Höhfeld, 1997). Thus, Toc64 has been proposed to participate in import as a docking receptor for phosphorylated preproteins delivered to the chloroplast by cytosolic Hsp70s (Sohrt and Soll, 2000).
Receptor proteins equipped with TPR motifs are also known to play roles in the recognition of preproteins targeted to other protein translocation systems. For example, Pex5 is a soluble protein involved in the translocation of proteins to the peroxisome, and Tom70 is a receptor for TIM22 substrates (e.g. carrier proteins) at the outer membrane of mitochondria. The Pex5 TPR motifs are known to interact directly with preproteins (Terlecky et al., 1995; Gatto et al., 2000), whereas the TPR motifs of Tom70 serve as a docking site for the cytosolic chaperones, Hsp70 and Hsp90, and also interact directly with internal targeting domains of carrier proteins (Young et al., 2003). It is therefore possible that Toc64 could bind directly to chloroplast preproteins or, as suggested previously, interact with a soluble targeting factor associated with chloroplast preproteins.
In Arabidopsis, three proteins similar to pea Toc64 (psToc64) were identified by homology searches (Jackson-Constan and Keegstra, 2001). Recently, the localization of these Toc64-like proteins has been investigated (Chew et al., 2004; Lee et al., 2004). The most similar protein to psToc64 (67% identity), Arabidopsis Toc64-III (atToc64-III), was shown to be associated with the chloroplast OEM, indicating that it is likely the functional orthologue of psToc64. By contrast, atToc64-I, a truncated Toc64-like protein, was not found to localize to organelles. Interestingly, mtOM64 (atToc64-V), the third Toc64-like protein, which displays 52% identity to psToc64, was found to associate with the mitochondrial outer membrane. This suggests that plant mitochondria, which appear to have different preprotein recognition properties from the mitochondria of other organisms (Macasev et al., 2000), may also utilize a Toc64-like receptor. This led to the hypothesis that Toc64 may be specifically involved in the recognition of so-called dual-targeted proteins (Chew et al., 2004): proteins that are normally targeted to both organelle types (Peeters and Small, 2001).
Core components and structure of the TOC complex
The receptor proteins, Toc34 and Toc159, have been shown by co-immunoprecipitation studies to form a stable trimeric complex with the Toc75 channel protein (Ma et al., 1996; Hiltbrunner et al., 2001b) (see section entitled, ‘The TOC channel’). Recently, the core TOC complex, found to be composed solely of these three proteins, was successfully purified from pea chloroplast OEM vesicles (OEVs) by sucrose gradient centrifugation (Schleiff et al., 2003c). The absence of Toc64 from this complex suggests that it may only transiently associate with the complex. Analysis of the stoichiometry of the three core components within the complex revealed that Toc86 (the proteolytic fragment of Toc159), Toc75, and Toc34 are assembled in a ratio of 1:4:4. The predicted size of 522 kDa for such a complex was found to be consistent with the apparent molecular mass of 500 kDa observed by size exclusion chromatography.
The isolated translocon was shown to interact specifically with the TP of precursor proteins in a GTP-dependent manner, and was therefore proposed to represent an intact and functional TOC complex. Electron microscopy analysis of core translocon particles incubated with gold-labelled TPs indicated that each particle could interact with up to four peptides, suggesting that each translocon may contain four independent channels. This idea was further supported by two-dimensional structural analysis of the translocon, again using electron microscopy. The translocon was shown to form a spherical complex, made up of four curved channels separated by a finger-like central structure. It is possible, as suggested by Schleiff et al. (2003c), that individual Toc75 proteins, each associated with a Toc34 receptor, form the four observed channels, and that these are assembled around a single Toc159 unit that forms part of the central finger-like domain.
GTP-regulated receptors of the TOC
GTP is necessary during the early steps of protein translocation (Olsen and Keegstra, 1992; Kessler et al., 1994; Young et al., 1999). GTP has been shown to promote the formation of early import intermediates (Ma et al., 1996; Kouranov and Schnell, 1997; Young et al., 1999), whereas, by contrast, non-hydrolysable GTP analogues inhibit the formation of such intermediates (Kessler et al., 1994; Young et al., 1999). These observations established a role for GTP-hydrolysis in mediating translocation initiation. Because of their homology to GTP-binding proteins and their capacity to bind GTP in vitro, Toc34 and Toc159 were originally proposed to regulate the initial steps of protein import through a GTP binding and hydrolysis cycle (Kessler et al., 1994). Shortly after this, Toc34 was shown to display intrinsic GTPase activity, further supporting a function in GTP-dependent regulation of protein import (Seedorf et al., 1995). More recently, it was also revealed that Toc34 interacts directly and specifically with the TP of RbcS in vitro, and that this interaction is stimulated by GTP (Sveshnikova et al., 2000a). Since their discovery, a considerable amount of effort has been invested in experiments to elucidate how GTP influences the TOC GTPases in mediating preprotein recognition and translocation initiation.
Based on sequence analysis of the Toc34 and Toc159 G-domains, the receptors were proposed to form a novel class of GTPase proteins (Kessler et al., 1994; Seedorf et al., 1995). This was recently confirmed by the resolution of the crystal structure of the pea Toc34 (psToc34) G-domain (Sun et al., 2002), which, due to its similarity to the Toc159 G-domain, serves as general model for the TOC GTPases. Analysis of the crystal structure revealed that the motifs and residues most likely involved in GTP binding and hydrolysis (Bourne et al., 1991) are uniquely arranged in Toc34, and differ significantly in sequence compared with those of the model GTPase, Ras. This was taken to suggest that the TOC GTPases probably utilize a novel GTP binding and hydrolysis mechanism, an idea that has taken further support from the finding that atToc33, the Arabidopsis orthologue of psToc34, appears to display unusual nucleotide binding properties (Aronsson et al., 2003a).
Interestingly, GDP-bound Toc34 G-domains were found to crystallize as dimers in which residues of each Toc34 monomer contributed to the GDP/GTP-binding pocket of its partner. Moreover, it was suggested that each monomer provides a catalytic ‘arginine finger’ (Arg133) necessary to promote GTP hydrolysis by the other monomer. Because of this, dimerization was proposed to lead to reciprocal stimulation of GTPase activity in each monomer. In most common GTPase systems, such as those of the small GTPases family including Ras, a non-homologous GTPase-activating protein (GAP) associates with the GTP-bound GTPase to provide such an arginine finger (Scheffzek et al., 1998). In the case of Toc34, the interaction properties of the two monomers within a dimer led to the suggestion that Toc34 might serve as its own GAP. However, the discovery that preprotein binding to the TOC GTPases strongly stimulates their GTPase activity (Jelic et al., 2002; Becker et al., 2004b) argues against this possibility, and suggests that TPs may actually serve as GAPs. Furthermore, the idea that dimerization stimulates the GTPase activity of Toc34 is inconsistent with the finding that substitution of Arg130 for an alanine residue (R130A) in atToc33 does not affect its GTPase activity, but instead greatly inhibits dimerization (Weibel et al., 2003); Arg130 in atToc33 corresponds to Arg133 in psToc34, the arginine residue proposed to act as an arginine finger (Sun et al., 2002).
Because the G-domains of Toc34 and Toc159 are very similar, and particularly because the residues determined to be involved in Toc34 dimerization are also conserved in Toc159, it was proposed that Toc34 and Toc159 may heterodimerize in vivo (Kessler and Schnell, 2002). This idea was supported by the finding that the G-domains of Toc34 and Toc159 interact in vitro (Hiltbrunner et al., 2001b). Furthermore, it can be proposed that such heterodimerization might be favoured in vivo, since, when Sun et al. (2002) assessed whether Toc34 G-domains could homodimerize in solution, only 20% of polypeptides were found as dimers. Another interesting feature of the dimer is that there appears to be no exit or entrance for nucleotide exchange at the GTP-binding pocket, suggesting that either a large conformational change or dimer dissociation must occur to allow GDP/GTP exchange. Many GTPases also associate with a second partner protein, a guanine nucleotide exchange factor (GEF), that promotes the exchange of GDP for GTP at the binding pocket of the GTPase (Bourne et al., 1991). As yet, however, GEFs for Toc34 and Toc159 have not been identified.
Overall, the study by Sun et al. (2002) provided significant insight into how the TOC GTPases may function together as GTP-regulated receptors (Kessler and Schnell, 2002). In addition, the identification of the actual motifs that form the GDP-binding pocket and the residues that are likely to mediate GTP binding and hydrolysis provided the information necessary to design useful mutants to analyse the function of the TOC GTPases further. Interestingly, biochemical assessment of the relationship between GTP- and preprotein- binding by Toc34 led to the suggestion that the GTPase activity of the TOC GTPases may be regulated by phosphorylation (Sveshnikova et al., 2000a). It was found that recombinant Toc34 could be phosphorylated in vitro, and that phosphorylation inhibits GTP binding and, as a consequence, preprotein binding. The possibility that protein import is regulated in this way is further supported by the fact that Toc34 and Toc159 can each be phosphorylated by a different protein kinase of the chloroplast OEM (Fulgosi and Soll, 2002). However, it has yet to be demonstrated that phosphorylation of the TOC GTPases truly occurs in vivo.
Recognition of preproteins by the TOC GTPases
Recent efforts to define the roles of Toc34 and Toc159 receptors in preprotein recognition have led to the emergence of two distinct models (Fig. 2): the targeting model, and the motor model (Jarvis and Robinson, 2004; Kessler and Schnell, 2004). The targeting model proposes that Toc159 is the primary receptor, and that it also plays a role in preprotein targeting from the cytosol to the chloroplast surface. By contrast, the motor model suggests that Toc34 carries out the role of primary receptor, whereas Toc159 acts as a molecular motor to drive preprotein translocation across the outer envelope.
|
The targeting model:
In the targeting model, after preprotein recognition, Toc159 associates with the secondary receptor, Toc34, to promote the transfer of the precursor to the Toc75 channel. The idea that Toc159 acts as the primary receptor is based on the early finding that Toc159 is the predominant protein cross-linking with preproteins arrested at the initial stages of translocation (Perry and Keegstra, 1994
; Ma et al., 1996; Kouranov and Schnell, 1997). This notion is further supported by the fact that Toc159 antibodies were found to block formation of early import intermediates in vitro (Hirsch et al., 1994). However, the model also incorporates the more recent finding that an abundant cytosolic form of Toc159 exists in addition to its integral OEM form (Hiltbrunner et al., 2001b). In this study, analysis of Arabidopsis Toc159 (atToc159) in Arabidopsis protoplasts by immunofluorescence microscopy suggested that a certain proportion of the protein is cytosolic. This observation was supported by subcellular fractionation experiments in Arabidopsis and pea, which revealed that Toc159, but neither Toc34 nor Toc75, could be found in a soluble, cytosolic form. The discovery of a soluble form of Toc159 led to the suggestion that the protein may shuttle between the cytosol and the chloroplast OEM in order to target newly translated, nucleus-encoded chloroplast proteins. Such a model implies that it is necessary for Toc159 to be able to both insert and retract from the membrane, and therefore provides an attractive explanation for its unusual, hydrophilic membrane-anchoring domain.
The discovery of a cytosolic soluble form of Toc159 prompted new initiatives to gain a better understanding of the factors that control and direct its targeting and insertion into the chloroplast OEM. In vitro targeting studies revealed that Toc159 binding and insertion are regulated in part by the protein's intrinsic GTPase activity. GTP binding by Toc159 promotes binding of the receptor to the chloroplast OEM, and GTP hydrolysis is required for efficient insertion of the M-domain into the membrane (Bauer et al., 2002; Smith et al., 2002). A role for the G-domain in Toc159 targeting has been supported by two independent in vivo studies (Bauer et al., 2002; Lee et al., 2003). These both showed that a functional GTPase domain is required for Toc159 transgenes to fully complement an Arabidopsis atToc159 null mutant called ppi2 (see section entitled, ‘Arabidopsis Toc159 homologues’). In addition, both studies revealed that Toc159 GTPase mutants with reduced GTP-binding and/or -hydrolysis activity failed to localize at the OEM in Arabidopsis chloroplasts and, instead, appeared to accumulate in the cytosol.
Surprisingly, however, the G-domain of Toc159 is not essential for targeting in vivo, since a truncated form of Toc159 composed solely of the M-domain was found to accumulate at the OEM, and to mediate partial complementation of ppi2 (Lee et al., 2003). These findings led Lee et al. (2003) to suggest that a non-functional Toc159 GTPase domain has a dominant negative effect on the targeting of the M-domain. In agreement with this idea, GTP hydrolysis by Toc159 does not appear to directly drive insertion of the M-domain into the membrane in vitro, but rather generates the GDP-bound form of the receptor required for efficient insertion (Smith et al., 2002). This was suggested by the fact that Toc159 was found to insert equally well in the presence of either GTP or GDP. It is therefore likely that GTPase mutants that cannot bind and/or hydrolyse GTP efficiently maintain an insertion-incompetent conformation, perhaps because the M-domain is masked by the G-domain. By contrast, the Toc159 M-domain (lacking the G-domain) appears to be in an insertion-competent form, although both targeting and insertion are most likely rather inefficient.
Studies by both Hiltbrunner et al. (2001a) and Smith et al. (2002) suggested that Toc159 targeting to the chloroplast is mediated by an interaction with Toc34. Smith et al. (2002) further showed that this specific interaction corresponds to a homotypic interaction between the G-domains of both proteins. While association did occur in the presence of GTP (Hiltbrunner et al., 2001a; Smith et al., 2002), this interaction was stimulated in the presence of GDP (Smith et al., 2002). In a recent study taking advantage of a newly-developed protocol allowing reconstitution of the TOC components into liposomes (Hinnah et al., 1997; Schleiff et al., 2003b), it was revealed that binding and insertion of Toc159 into the OEM is also dependent upon Toc34 GTPase activity (Wallas et al., 2003), confirming the importance of a GTP-regulated interaction between the homologous G-domains of Toc34 and Toc159. It is possible that initial binding occurs between Toc159-GTP and Toc34-GTP, and that this induces GTP hydrolysis by both GTPases. Hydrolysis would induce stable dimerization and insertion of the M-domain of Toc159 into the membrane. In addition, Wallas et al. (2003) revealed that the presence of Toc75 is essential for the membrane insertion of the M-domain of Toc159. It is possible that, due to its overall hydrophilic nature, the insertion of the Toc159 M-domain involves the formation of ß-strands (Schleiff et al., 2003a), and that these are stabilized through interactions with Toc75; alternatively, Toc75 may actively participate in the insertion of the Toc159 M-domain.
As described earlier, the regulated targeting and insertion of Toc159 into the OEM is proposed to be coupled with the targeting of precursor proteins to the TOC complex, leading to translocation initiation. Such a scenario is reminiscent of co-translational protein targeting to the ER and bacterial inner membrane by the SRP and the SRP-receptor (SR/FtsY) GTPases. In these systems, the nascent signal peptide of a preprotein associates with the SRP, and then makes its way as a complex along with the ribosome to the target membrane. At the membrane, a GTP-regulated interaction between the SRP GTPase and the homologous, membrane-associated SR/FtsY GTPase leads to the transfer of the precursor protein to the Sec translocase and the initiation of co-translational translocation (Keenan et al., 2001; Schnell and Hebert, 2003). The targeting hypothesis proposes similar roles for the Toc159 and Toc34 GTPases, as shown in Fig. 2A. Soluble Toc159 interacts with chloroplast preproteins in the cytosol and mediates their targeting to the chloroplast surface. Upon arrival at the chloroplast membrane, a similar homotypic interaction between Toc159 and membrane-bound Toc34, coupled with GTP hydrolysis, leads to transfer of the preprotein to the TOC channel. Further support for this hypothesis was recently provided by the finding that soluble Toc159, either from cell extracts or in vitro translation mixtures, can interact specifically with the TP of chloroplast preproteins (Smith et al., 2004).
The motor model:
In the recently proposed motor model, Toc159 functions as a GTP-dependent motor that pushes preproteins through the TOC channel via multiple rounds of GTP hydrolysis. This hypothesis stems from the results of recent studies in which the structure of the purified TOC complex, and the functions of individual TOC components of the reconstituted complexes (see section entitled, ‘Translocation through the TOC complex’), were examined. The proposed central function of Toc159 as a motor and the near 1:1 stoichiometric association of Toc34 with the Toc75 channel in the core complex (Schleiff et al., 2003c), led to the suggestion that Toc34 is more likely to function as the primary receptor (Becker et al., 2004b). This view can be supported by the fact that Toc34 was found to cross-link to precursors only at the initial energy-independent binding stage of import; by contrast, Toc159 appears to remain in close proximity with precursors throughout their translocation through the OEM (Kouranov and Schnell, 1997). In addition, the direct and specific interaction of Toc34 with the TP of various precursor proteins has been clearly demonstrated (Sveshnikova et al., 2000b; Schleiff et al., 2002; Becker et al., 2004b), further supporting its role as a receptor.
Recently, Becker et al. (2004b) presented strong evidence in support of the motor model. First, detailed subcellular fractionation studies suggested that the previously-observed soluble form of Toc159 (Hiltbrunner et al., 2001b) is an artefact produced by the fractionation method used. They argued that the Toc159 detected in the cytosolic fraction was associated with low-density membrane shreds released from the chloroplast during tissue homogenization. In light of these results, they concluded that Toc159 is only present as an integral protein of the chloroplast OEM. Having concluded that productive interactions between preproteins and the TOC GTPases occur only at the OEM, Becker et al. (2004b) demonstrated that, like Toc34, Toc159 interacts specifically with the TP of RbcS, and that GTP stimulates this preprotein binding. However, in contrast with Toc34, which was previously shown to bind phosphorylated TPs with high affinity (Sveshnikova et al., 2000a), Toc159 only recognizes the non-phosphorylated form of the RbcS TP. Furthermore, by using peptides spanning either the C- or N-proximal regions of the RbcS TP, they revealed that a Toc159 fragment spanning the G- and M-domains (Toc86) displays high affinity for the N-terminal part of the TP. This, again, is in contrast with Toc34, which binds the phosphorylated C-terminal peptide with high affinity (Sveshnikova et al., 2000b; Becker et al., 2004b).
The different peptide-binding properties of Toc34 and Toc159 were used to show that Toc34 functions before Toc159 in the binding and translocation of precursor proteins. This was demonstrated by using the different peptides to inhibit binding and import into proteoliposomes containing the reconstituted core TOC complex (TOC proteoliposomes). It was found that only the phosphorylated C-proximal peptide was able to significantly inhibit binding of pRbcS to TOC proteoliposomes, whereas both the phosphorylated C-proximal peptide and the N-proximal peptide inhibited import of the precursor. This indicates that in order to bind to the chloroplasts, precursors must first interact with Toc34. Subsequently, these precursors need to interact with Toc86 to be imported. These experiments were designed under the assumption that the TP is phosphorylated in the cytosol and requires dephosphorylation in order to become import-competent. However, as described previously, the importance and even the occurrence of TP phosphorylation in vivo is still a case open to debate. Regardless of this caveat, the results of these experiments may nevertheless be valid, since, even though a phosphorylated peptide was used as a specific competitor for Toc34 binding, the pRbcS used in the import assays was not phosphorylated. Inhibition of precursor binding to TOC proteoliposomes was only caused by the high-affinity binding of a peptide to Toc34, implying that Toc34 plays the initial receptor role.
In addition to assessing the hierarchy of the TOC receptors, Becker et al. (2004b)