A protein import receptor of chloroplasts is inserted into the outer envelope membrane by a novel pathway.

The outer envelope protein OEP86 functions as a receptor for precursor proteins in the chloroplastic import machinery. In contrast to most other organellar outer membrane proteins it is synthesized as a precursor polypeptide (preOEP86) in the cytosol and is post-translationally targeted to the organelles. PreOEP86 is targeted to and productively inserted into the chloroplastic outer envelope mediated by a bipartite signal consisting of the presequence and the COOH terminus of the precursor protein. The cleavable presequence alone does not seem to contain sufficient information to target preOEP86 without the COOH terminus or a hybrid protein consisting of the presequence of preOEP86 and the mature form of the small subunit of ribulose bisphosphate carboxylase to intact chloroplasts. The presequence seems to be required to maintain preOEP86 in an integration competent state, whereas interaction of preOEP86 with chloroplasts is accomplished by a short sequence of amino acids in the COOH-terminal portion of the mature protein. The COOH-terminal portion of preOEP86 contains enough information to also direct mature OEP86 into the outer envelope membrane of pea chloroplasts. However, mature OEP86 enters the productive folding pathway much less efficiently than preOEP86. The COOH terminus of preOEP86 not only serves as a membrane anchor but seems to be required for a productive translocation through an interaction with other outer envelope proteins. Although the binding was ATP-dependent, productive folding was not. PreOEP86 seems to follow a unique road into the chloroplastic outer envelope.

The outer envelope protein OEP86 functions as a receptor for precursor proteins in the chloroplastic import machinery. In contrast to most other organellar outer membrane proteins it is synthesized as a precursor polypeptide (preOEP86) in the cytosol and is posttranslationally targeted to the organelles. PreOEP86 is targeted to and productively inserted into the chloroplastic outer envelope mediated by a bipartite signal consisting of the presequence and the COOH terminus of the precursor protein. The cleavable presequence alone does not seem to contain sufficient information to target preOEP86 without the COOH terminus or a hybrid protein consisting of the presequence of preOEP-86 and the mature form of the small subunit of ribulose bisphosphate carboxylase to intact chloroplasts. The presequence seems to be required to maintain pre-OEP86 in an integration competent state, whereas interaction of preOEP86 with chloroplasts is accomplished by a short sequence of amino acids in the COOH-terminal portion of the mature protein. The COOH-terminal portion of preOEP86 contains enough information to also direct mature OEP86 into the outer envelope membrane of pea chloroplasts. However, mature OEP86 enters the productive folding pathway much less efficiently than preOEP86. The COOH terminus of pre-OEP86 not only serves as a membrane anchor but seems to be required for a productive translocation through an interaction with other outer envelope proteins. Although the binding was ATP-dependent, productive folding was not. PreOEP86 seems to follow a unique road into the chloroplastic outer envelope.
The vast majority of mitochondrial and chloroplastic protein constituents are nuclear encoded, synthesized in the cytosol, and post-translationally imported into the organelles (1,2). In general, proteins are made as precursors with NH 2 -terminal transit sequences, which contain the information necessary for targeting to the proper organelle and in many cases also for routing inside the organelle (1). In contrast, nuclear encoded polypeptide constituents of the chloroplastic outer envelope membranes are generally not synthesized as larger precursor proteins but contain internal targeting information (3)(4)(5)(6). This insertion process does not require the hydrolysis of ATP or protease-sensitive chloroplast surface components (3). In some cases ATP seems to stimulate insertion (5,7). In contrast, import of precursor proteins into chloroplasts requires the hy-drolysis of ATP at several steps during the process and involves protease-sensitive chloroplast surface components (1,2,8). Several constituents of the precursor protein import machinery of the chloroplastic outer envelope membrane have been identified recently, namely the outer envelope proteins of 86, 75, 44, and 34 kDa (OEP86, OEP75, OEP44, and OEP34) (9 -15). Biochemical data suggest that OEP86 functions as a receptor for precursors (12), whereas OEP75 could form the central protein conducting channel of the outer envelope translocase (11). OEP86 and OEP75 are both synthesized with NH 2 -terminal cleavable presequences (12,15). To date OEP86 and OEP75 are the only known proteins of an organellar outer membrane, which are made with a presequence. The presequence of preOEP75 contains typical features of stroma directing envelope transfer domains of "normal" chloroplastic precursors and biochemical evidence indicated that preOEP75 uses components of the general chloroplast protein import machinery (15). In contrast to stroma targeting signals, the presequence of preOEP86 is unusually long (15 kDa) and highly negatively charged. The productive translocation of preOEP86 into the chloroplastic outer envelope is dependent on ATP and requires protease-sensitive chloroplast surface components but does not use the general import machinery (12). These data indicate that a unique pathway has developed for the translocation of the protein import receptor into the chloroplastic outer envelope membrane.
In this study we have analyzed the role of the cleavable presequence of preOEP86 and that of internal sequences for targeting and insertion into the chloroplastic outer envelope membrane. Surprisingly the presequence of preOEP86 alone does not contain sufficient targeting information for chloroplasts but a productive translocation process requires a bipartite signal that is present in the presequence as well as in the COOH terminus of preOEP86. Our data suggest further that preOEP86 interacts initially with proteinaceous chloroplast surface components in an ATP-dependent step. The COOH terminus of preOEP86 seems to contain two domains, one that serves as a membrane anchor and one that is necessary for productive folding.

Construction of preOEP86 and mOEP86 for in Vitro Transcription-Translation
The cDNA coding for preOEP86 in the vector pET17b (12) was recloned into pET24c (Novagen, Madison, WI) in two steps after restriction with NdeI and XhoI resulting in preOEP86-pET24c. To obtain a cDNA coding for mOEP86, which could be used for in vitro transcription and translation, an additional NdeI site was introduced at the processing site of preOEP86 by in vitro mutagenesis. The primers used for the PCR 1 were 5Ј-CGTCCCAATCATATGGCTCC-3Ј (forward) and 5Ј-CGC-* This work was supported by the Deutsche Forschungsgemeinschaft and the Fonds der Chemischen Industrie. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Construction of Different Fusions from preOEP86, preSSU, and SSU
pre86SSU-A SphI site was introduced into preOEP86-pET17b by in vitro mutagenesis using the following primers for the PCR 5Ј-GCTC-TACGCATGCAAGGAGC-3Ј (reverse) and T7 universal primer (forward). The PCR product was subcloned into the vector pGEM5Zf(ϩ) (Stratagene, La Jolla, CA) after restriction with NdeI and SphI resulting in clone pre86-pGEM5Zf(ϩ). A cDNA fragment coding for SSU was obtained by restriction of preSSU-pSP64 (17) with SphI and SacI and was ligated into pET24c (Novagen) after restriction with NdeI and SacI together with the NdeI/SphI-fragment from pre86-pGEM5Zf(ϩ). The coding sequence of the resulting clone pre86SSU-pET24c was verified by sequencing. It contained amino acids 1-149 of preOEP86 and the entire coding sequence of mature SSU.

In Vitro Transcription-Translation and Expression of the Various Constructs
mRNAs were obtained using either SP6 or T7 RNA-polymerase depending on the promoter present in the vector construct. The mRNAs were translated in a reticulocyte lysate system in the presence of 35 S-labeled methionine and cysteine as described before (4). OEP86-C 699 -879 was overexpressed in E. coli BL21De3 cells and recovered in inclusion bodies (18).

Isolation of Chloroplasts and Protein Import
Chloroplasts were isolated from pea leaves as described before (13). A standard import assay contained chloroplasts equivalent to 30 g of chlorophyll (19) and was carried out for 15-25 min at 25 or 4°C in a final volume of 200 l and 2-5% (v/v) translation product. Chloroplasts were treated with the protease thermolysin either before (750 g of protease mg Ϫ1 chlorophyll, 30 min, 4°C) or after import (100 g of protease mg Ϫ1 chlorophyll, 15 min, 4°C). Intact chloroplasts were recovered prior to further experimentation or analysis. The detailed procedure is as outlined in Ref. 13. ATP was removed from the translation mixture by the addition of hexokinase and glucose (20).

Both Mature and preOEP86
Interact with Intact Pea Chloroplasts-To investigate the insertion of mOEP86 and its precursor form preOEP86, both were synthesized in a reticulocyte lysate and incubated with intact pea chloroplasts (Fig. 1A). PreOEP86 binds to chloroplasts and is processed to the mature form, OEP86 (Fig. 1A, lane 1). Upon protease treatment, processed mature OEP86 is cleaved and a residual 52-kDa proteolytic fragment can be detected (Fig. 1A, lane 2). This 52-kDa fragment of OEP86 represents a COOH-terminal portion of the protein (12) and can be used as a parameter for the correct folding of the protein in situ and in vitro (12,23). These data indicate that preOEP86 added to chloroplasts had inserted and folded correctly in the chloroplastic outer envelope. In contrast when translation product of mOEP86 was offered to chloroplasts, binding was observed but much less 52 kDa breakdown product was seen (Fig. 1A, lanes 3 and 4), indicating that the presequence of OEP86 was necessary for a productive translocation process. To characterize further the nature of the interaction between chloroplasts and preOEP86 or mOEP86, respectively, organelles were separated into a soluble and a membrane fraction. The membranes were then extracted at high pH (pH 11, 0.1 M Na 2 CO 3 ). As shown (Fig. 1B) preOEP86, processed mature OEP86 or mOEP86 were recovered exclusively in the membrane fraction (Fig. 1B, lanes 1 and 2 and lanes 5 and 6) and were resistant to extraction at high pH ( Fig.  1B, lanes 3 and 4 and lanes 7 and 8), indicating that preOEP86 as well as mOEP86 interacted with chloroplasts by a proteinlipid interaction. The respective translation products were recovered in the soluble fraction (not shown and Ref. 12). These data indicate that an insertion or targeting signal exists in the mature form of OEP86, which might help to anchor the protein into the membrane, whereas the presequence of preOEP86 seems to stimulate the productive translocation process.
To analyze the role of the presequence of preOEP86 in more detail, a hybrid protein was constructed, which consisted of the presequence of preOEP86, i.e. amino acids 1-149, fused in frame with the mature SSU of Rubisco (pre86SSU) to compare its targeting properties with that of preSSU. We were unable to detect any binding or interaction of pre86SSU translation product to chloroplasts either at 50 M ATP or 2 mM ATP ( Fig. 2A,  lanes 1 and 2 and lanes 3 and 4) not even under very rapid recovery conditions by centrifugation through a silicon oil layer  (21) and fluorography (22). B, chloroplasts were lysed in 10 mM Hepes-KOH, pH 7.6. A membrane fraction (P) and soluble proteins (S) were obtained by centrifugation (250,000 ϫ g, 15 min). A second membrane fraction was extracted with 0.1 M Na 2 CO 3 and again separated into an insoluble fraction (P) and high pH soluble protein fraction (S). Samples were further analyzed as in A. In order to estimate the translocation efficiency 10% of the translation product (TL) added to a 200-l assay is shown separately.
into HClO 4 (24) (not shown). Although it cannot be excluded that targeting information present in the presequence of preOEP86 is hidden in the hybrid protein pre86SSU in vitro due to improper folding, we conclude that the presequence of preOEP86 does not contain enough information on its own to target a protein to chloroplasts, but additional internal OEP86 sequences are required for a productive targeting and insertion process.
We constructed a number of preOEP86 deletions to define further regions in the polypeptide that were responsible for a targeting and productive insertion (Fig. 2B). The mode of interaction between chloroplasts and preOEP86 deletions, for which a binding to chloroplasts was observed, was further analyzed by extraction at pH 11 (0.1 M Na 2 CO 3 ) (25). The detection of a protease-resistant fragment of a preOEP86 construct was taken as a measure for a certain extent of folding and insertion into the membrane. Thus, by using alkaline extraction and proteolytic fragmentation, we could distinguish between two stages in the interaction process of preOEP86 with chloroplasts. Our results (Fig. 2B) indicate that neither the presequence of preOEP86, namely preOEP86 137 itself, nor preOEP86 347 bound to chloroplasts, whereas preOEP86 542 and preOEP86 697 bound with low efficiency (10 -20% of full-length preOEP86) (Fig. 2B). Only preOEP86 and preOEP86 775 bound to chloroplasts with similar yield, i.e. 5-10% of the total translation product added to a binding experiment. Both preproteins were completely recovered in the alkaline insoluble fraction, whereas preOEP86 542 and preOEP86 697 were recovered in the alkaline-soluble fraction (Fig. 2B). Thus, a short COOH-terminal stretch between amino acids 697-775 of preOEP86 seems to be responsible for the alkaline-resistant interaction with the chloroplastic outer envelope. To obtain further evidence for this notion a NH 2 -terminal deletion of OEP86 was constructed, namely OEP86-C 699 -879 , which contained only the COOH-terminal 180 amino acids. OEP86-C 699 -879 inserted into the outer envelope membrane in a way that was resistant to alkaline extraction (Fig. 2B). When chloroplasts were treated with the protease thermolysin after an insertion experiment, no protease protected product could be detected (Fig. 2B) from any of the preOEP86 deletion constructs or from OEP86-C 699 -879 except for the full-length preOEP86 (Fig. 2B). From these data we conclude that a COOH-terminal portion of preOEP86 anchors the protein into the lipid bilayer. In addition the COOH-terminal portion of preOEP86 and its presequence might have to cooperate to insure not only targeting but a productive integration process of preOEP86 as evidenced by the appearance of the 52-kDa breakdown product from the full-length precursor only. Indeed this conclusion is consistent with studies presented below.
The COOH terminus of OEP86 might act as a sorting sequence and membrane anchor for the chloroplastic outer envelope membrane. To study this point we constructed hybrid proteins consisting of either the mature or the precursor form of SSU fused in frame to amino acids 653-879 of preOEP86 (Figs. 3 and 4), resulting in the constructs SSUc86 and pre SSUc86. When intact chloroplasts were incubated with SSUc86 translation product in the presence of different ATP concentrations, about 5-10% of the added protein bound to chloroplasts (Fig. 3A, upper panel). The yield of binding was not signifi-FIG. 2. The presequence of preOEP86 cannot target a passenger protein to chloroplasts but is necessary together with a COOH-terminal portion of OEP86 for a productive translocation process. A, pre86SSU does neither bind nor import into chloroplasts independent of the ATP concentration. Experimental conditions were as outlined in the legend to Fig. 1. TL, translation product, 10% of which was added to an experiment. Th, thermolysin. B, truncated preOEP86 polypeptides were synthesized by in vitro transcriptiontranslation after treatment of preOEP86-pET17b with different restriction enzymes (see "Materials and Methods"). OEP86-C 699 -879 was constructed and synthesized as described under "Materials and Methods." Binding and interaction with chloroplasts was done and analyzed as above. The yield of binding was quantified by laser densitometry of the exposed x-ray films. ϩϩϩ, Ͼ5% binding; ϩ, Ͻ2% binding of added translation product. Integration of preOEP86 deletion proteins was also assessed by the appearance of a protease-resistant fragment (thermolysin (Th) fragment).

FIG. 3. The COOH terminus of OEP86 targets the mature form of SSU to chloroplasts and anchors it into the outer envelope.
A fusion protein, SSUc86, was synthesized that contained the entire coding sequence for the small subunit of Rubisco (SSU) and as a carboxyl-terminal extension amino acids 653-879 of preOEP86. A, chloroplasts were incubated with SSUc86 or SSU translation product in the presence of different concentrations of ATP. All other manipulations were as outlined on top of the figure and as described in the legend to Fig. 1. B, SSUc86 was recovered together with chloroplast membranes (P) and remained in the Na 2 CO 3 insoluble fraction (P) and not in the soluble protein fractions (S) (methods as outlined in the legend to Fig.  1). TL, translation product; Th, thermolysin. cantly influenced by the various ATP levels (Fig. 3A, upper  panel, lanes 1 and 3). No interaction was detected between chloroplasts and SSU translation product (Fig. 3A, lower panel,  lanes 1-4), demonstrating that the COOH terminus of OEP86 was responsible for the interaction of SSUc86 with chloroplasts. Chloroplasts bound SSUc86 was susceptible to protease treatment and was completely degraded (Fig. 3A, upper panel,  lanes 2 and 4), indicating that import to the inside of the organelle had not occurred. Furthermore the data demonstrate that the SSU portion of the hybrid protein was exposed on the organellar surface and not to the intermembrane space. SSUc86 was recovered in the membrane fraction (Fig. 3B, lanes  1 and 2) and was resistant to extraction by 0.1 M Na 2 CO 3 (Fig.  3B, lanes 3 and 4). Together these data indicate that the COOH terminus of OEP86 can anchor an otherwise soluble protein into the outer envelope membrane in a way that it is exposed to the cytosol.
The presequences of chloroplast proteins localized in the inner envelope, the stroma, or the thylakoids function as an envelope transfer stroma targeting domain, even allowing membrane proteins, which contain multiple ␣-helical transmembrane helices in vivo, to pass across both the outer and inner chloroplastic envelope membranes (1). We thus wanted to know if preSSU could function as an envelope transfer carrier for the COOH terminus of preOEP86 and constructed a hybrid protein in which preSSU was fused to amino acids 653-879 of preOEP86 (see above) resulting in preSSUc86. The fusion protein preSSUc86 bound to chloroplasts in the presence of 50 M ATP (Fig. 4A, lane 1), conditions that also favor binding of preSSU but that do not allow maximal translocation (Fig. 4A, lane 5) (26,27). When chloroplasts were treated with the protease thermolysin after such a binding experiment, translocation intermediates of preSSU were detected, namely Tim 3 and Tim 4 (Fig. 4A, lane 6), which have been described before (13,28). Chloroplast-bound preSSUc86 gave rise to identical translocation intermediates as preSSU (Fig. 4A, lanes 2 and 6) although with low yield, indicating that some preSSUc86 had entered the general import route across the envelope membranes. Under conditions that allow complete translocation of a precursor protein, i.e. 2 mM ATP, processed mature SSU was detected protease protected inside chloroplasts from preSSU, whereas very little precursor remained at the organellar surface (Fig. 4A, lanes 7 and 8). In contrast most of preSSUc86 remained bound to the chloroplast surface in a protease-accessible way in the presence of 2 mM ATP (Fig. 4A, lanes 3 and 4). Between 10 -20% of preSSUc86 was found in the processed form, SSUc86, protease protected inside chloroplasts (Fig. 4A,  lanes 3 and 4) corroborating our earlier notion that preSSU can indeed function, although with low efficiency, as a carrier for c86. Differences in the interaction of preSSU and preSSUc86 with intact chloroplasts are indicated by differences in localization and mode of interaction. At 50 M ATP preSSU is recovered together with the membranes, but the majority (between 70 and 80%) is extractable by pH 11 (Fig. 4B, lower  panel, lanes 1-4). This is in agreement with data reported before (13) under these conditions. Only very little of pre SSUc86 is extractable by 0.1 M Na 2 CO 3 (Fig. 4B, upper panel,  lanes 3 and 4). The alkaline soluble form of preSSUc86 is most likely the fraction that is bound to the general protein import machinery of chloroplasts via the preSSU portion of the hybrid protein, whereas the insoluble form of preSSUc86 represents the fraction that interacts with the chloroplastic outer envelope via the COOH terminus of OEP86. Processed mature SSU and SSUc86 are recovered in the soluble stroma, and the remaining protein in the membrane fraction is completely alkaline extractable (Fig. 4B, lanes 5-8).
The results obtained so far indicated that firstly the presequence of preOEP86 was necessary for an efficient and productive translocation pathway, although it did not contain chloroplast targeting information. Secondly, the COOH terminus of OEP86 is necessary to anchor the protein into the chloroplastic outer envelope. Thirdly, the COOH terminus could also play a role as a sorting signal.
As demonstrated in Fig. 2, OEP86-C 699 -879 was able to insert in an alkaline-resistant way into the chloroplastic outer membrane. Therefore, we wanted to know if overexpressed OEP86-C 699 -879ex was able to influence binding and insertion of preOEP86 into chloroplasts. In the presence of overexpressed OEP86-C 699 -879ex proper insertion and folding of preOEP86 translation product was reduced by about 70%, as demonstrated by the reduced yield of mature OEP86 and the 52-kDa fragment (Fig. 5, lanes 2 and 4). To detect competition at the level of preOEP86 binding to chloroplasts is much more difficult because a significant amount of binding in vitro seems to be nonproductive (see above). Furthermore we do not expect that competition occurs at the level of nonspecific and nonproductive interaction with the bulk lipid phase. However, the results show that OEP86-C 699 -879ex associates with and partially inhibits most likely proteinaceous components involved in the productive targeting and insertion pathway (Fig. 5). In an attempt to differentiate further between the role of lipids FIG. 4. PreSSU cannot function as an efficient envelope transfer stroma targeting domain for the carboxyl terminus of OEP86. A, the radiolabeled precursor proteins preSSUc86 or preSSU were incubated with chloroplasts under binding, i.e. 50 M ATP, or import, i.e. 2 mM ATP, conditions. The experiment was stopped and chloroplasts either not treated or treated with the protease thermolysin (Th) (see Fig. 1). The preSSU translocation intermediates are labeled Tim 3 and Tim 4 (13,28). B, the suborganellar localization of pre-SSUc86, preSSU, SSUc86, and SSU was tested under binding and import conditions by separating the organelles into a soluble protein (S) and membrane fraction (P). The membrane fraction was further subjected to extraction by pH 11 (0.1 M Na 2 CO 3 ) (P, pellet; S, soluble protein; for details see Fig. 1). TL, translation product. and proteinaceous components in the preOEP86 insertion pathway, chloroplasts were treated with the protease thermolysin prior to an insertion experiment. PreOEP86 bound to chloroplasts either not treated or treated with protease in the presence of ATP. This binding was largely resistant to alkaline extraction (not shown). Only untreated chloroplasts, however, were able to properly insert and fold the protein as deduced from the 52-kDa fragment (not shown). Binding of mOEP86 or OEP86-C 699 -879 to protease-treated chloroplasts occurred with similar yields as to nontreated chloroplasts. These data indicate that proteinaceous components are involved in the productive binding and insertion process of preOEP86 (12). When these components are removed a nonproductive interaction occurs in vitro between the outer envelope lipids and preOEP86 most likely via the COOH terminus.
To analyze the role of ATP in the translocation pathway, the translation products of preOEP86, preOEP86 775 , mOEP86, OEP86-C 699 -879 , and preSSU were depleted of ATP by the addition of hexokinase and glucose (20). This treatment should not influence a potential protein-lipid interaction but only an energy requiring binding process. In the absence of ATP preOEP86, preOEP86 775 , and preSSU bound only with low yield to chloroplasts in contrast to parallel experiments in the presence of 20 M ATP, which resulted in a 2-3-fold stimulation of binding depending on the protein used (see Fig. 6). The absence or the presence of ATP, however, did not influence significantly the extent of interaction of those OEP86 forms, which did not contain a presequence, namely mOEP86 and OEP86-C 699 -879 (Fig. 6). The presequence of OEP86 seems to hinder an unproductive and premature interaction of the COOH terminus with the lipid-phase of the outer envelope membrane. Whether the ATP-requiring step is due to a cytosolic factor present in the reticulocyte lysate or a component in the envelope membranes is not known at the moment.
The initial presequence-dependent interaction of preOEP86 with chloroplasts requires ATP (Fig. 6). We wanted to know if further steps, e.g. proper folding were also ATP-dependent. To analyze this, we first studied the temperature dependence of preOEP86 interaction with and insertion into in chloroplasts. Productive translocation of preOEP86 and processing and hence the yield of the 52-kDa proteolytic fragment were clearly temperature-dependent events (Fig. 7A, lanes 1-6). PreOEP86, which had bound to chloroplasts at 4°C, could be chased on the productive translocation pathway after reisolation and supplementation of the chloroplasts with fresh import buffer in the presence of ATP at 25°C as evidenced by the appearance of the 52-kDa fragment (Fig. 7A, lanes 7-10). This experimental set up was subsequently used to study the ATP requirement downstream of binding. Chloroplasts were incubated with preOEP86 in the presence of ATP at 4°C. Organelles were then isolated from the binding assay, washed, and supplemented with new import buffer in the absence or the presence of ATP or a nonhydrolyzable ATP analog (AMP-PNP), and the reaction was allowed to continue at 25°C for 20 min. The results show that ATP is not required for the final translocation steps of preOEP86. The yield of the 52-kDa fragment was similar in the absence or the presence of ATP or the nonhydrolyzable analog AMP-PNP at 25°C (Fig. 7B). DISCUSSION To date only two proteins have been identified, namely OEP86 and OEP75, that require an NH 2 -terminal cleavable presequence for productive targeting to an organellar outer membrane. The common denominator of these proteins is that both are localized in the chloroplastic outer envelopes and both are constituents of the protein import machinery (2,8). Although preOEP75 follows the general import pathway (15) it seemed evident from preliminary studies (12) that preOEP86 uses a unique translocation pathway. In this study we demonstrate that the presequence has to cooperate with a COOHterminal portion of preOEP86 to guarantee the efficient and productive translocation into the outer envelope. The function of the presequence of preOEP86 seems at least in part different from that of other presequences, because it has no ability to target preOEP86 137 and preOEP86 347 or a passenger protein, namely pre86SSU to chloroplasts. Our results demonstrate, however, that the presequence of preOEP86 is required for a productive translocation pathway. Firstly, the yield of OEP86 insertion is stimulated 3-4-fold when preOEP86 instead of mOEP86 is offered to chloroplasts. Secondly, only preOEP86 and preOEP86 775 show a significant dependence on ATP for binding to chloroplasts, whereas mOEP86 and OEP86-C 699 -879 do not. The presequence of preOEP86 is negatively charged in contrast to normal chloroplast directing presequences, which carry an overall positive charge. The negative charges might therefore inhibit the interaction with receptor components of the general protein import machinery of chloroplasts and lead to a specific and novel translocation pathway for preOEP86.  6. The presequence of preOEP86 causes ATP dependence of chloroplast binding. In vitro translation products were depleted of ATP by the addition of hexokinase and glucose (20). Binding to chloroplasts was then assayed either in the absence of exogenously added ATP or in the presence of 20 M ATP. Binding was quantified by laser densitometry of the exposed x-ray films. The bars represent the x-fold stimulation of binding in the presence of ATP in comparison with the absence of ATP. A typical result out of three repeats is shown.
Although the presequence of OEP86 is necessary for productive insertion of the preprotein, a specialized receptor for it seems not to be present in the outer envelope. PreOEP86 137 , preOEP86 347 , and pre86SSU had no capacity to bind to intact chloroplasts. Longer constructs, namely preOEP86 542 and preOEP86 697 , bound with low efficiency; only preOEP86 775 , OEP86-C 699 -879 , and the full-length precursor exhibited tight binding. These data indicated that preOEP86 contains also internal targeting information. A carboxyl-terminal region between amino acid positions 697-775 was required for an initial binding maybe via an interaction with the chloroplast envelope lipids as indicated by its resistance to extraction at pH 11. However, even preOEP86 775 did not yield a protease protected insertion product, suggesting that further information is present in the final carboxyl-terminal stretch of 100 amino acids of preOEP86, which is necessary to ensure a productive folding and insertion cycle of preOEP86. This hypothesis is corroborated by our finding (Fig. 5) that overexpressed OEP86-C 699 -879 ex could compete the productive translocation process of preOEP86. Thus, OEP86 requires bipartite information for targeting and translocation. The NH 2 -terminal presequence and the extreme COOH terminus seem to be involved in protein-mediated reactions of the translocation pathway, whereas an intermediate sequence (amino acid 697-775) could function as a membrane anchor. Further work is required to prove this idea.
The productive translocation of preOEP86 into the chloroplastic outer envelope is not a spontaneous process but requires yet to be identified proteinaceous components. This conclusion is drawn from the evidence that (i) protease-treated organelles are unable to integrate preOEP86; (ii) hydrolysis of ATP is required for productive binding of preOEP86 but not for the final folding reactions; (iii) preOEP86 translocation can be competed with the overexpressed OEP86-C 699 -879 ex ; and (iv) the translocation is strictly temperature-dependent. In general the translocation of other proteins into the chloroplastic outer envelope does not require ATP or protease-sensitive chloroplast surface components (3,4,6) The carboxyl terminus of OEP86 is necessary to anchor the protein into the envelope membrane. However, it can also function as a membrane anchor for a soluble protein as demonstrated for SSUc86. SSUc86 is oriented in the outer envelope in a way that exposes SSU to the cytosol. This orientation is like that in the authentic OEP86, the NH 2 terminus of which protrudes also into the cytosol (12). We were unable to detect a protease-protected fragment from inserted SSUc86 indicating that only a short transmembrane region of SSUc86 was protease-resistant, too short to warrant detection. The other possibility would be that SSUc86 represents a monotopic protein (29), a possibility that has been considered before for OEP86 (12).
Envelope transfer stroma targeting domains can be used as carriers to import passenger proteins that contain several hydrophobic transmembrane ␣-helices into the organelle. The COOH terminus of OEP86, i.e. amino acids 653-879, however, represents a strong detour signal, which directs preSSUc86 to the chloroplastic outer envelope. About 10 -20% of the hybrid protein enters the general import pathway as demonstrated by the appearance of processed SSUc86 inside the organelle at 2 mM ATP. At 50 M ATP translocation intermediates were detected that were identical for preSSU and preSSUc86 as expected if translocation proceeds from the NH 2 terminus to the COOH terminus and comes to a stop at ATP concentrations that do not allow complete translocation, i.e. 50 -100 M ATP (13,28). FIG. 7. Translocation of preOEP86 into chloroplasts is temperature-dependent, but the final folding process does not require ATP. A, a standard translocation assay of preOEP86 was carried out at the different temperatures indicated on top of the figure. Productive translocation was estimated by the appearance of the 52-kDa proteolytic fragment upon thermolysin treatment. In a parallel experiment (lanes 7-10) preOEP86 was bound to chloroplasts at 4°C in the presence of ATP, and organelles were reisolated, washed once, and supplemented with fresh import medium in the presence of ATP. The reactions were allowed to continue either at 4°C (lanes 7 and 8) or at 25°C (lanes 9 and 10) for 20 min. Th, thermolysin. B, five binding experiments of preOEP86 to chloroplasts were carried out at 4°C for 20 min in the presence of 20 M ATP. Binding assays were pooled, and organelles were reisolated, washed once, resuspended in fresh import buffer, and divided into five aliquots. One aliquot was analyzed prior to the chase period (prebinding, left column). Four aliquots were used for the different chase experiments either in the absence or the presence of ATP (2 mM) or in the presence of the nonhydrolyzable ATP analog AMP-PNP (2 mM). Translocation reactions were allowed to continue for 20 min either at 4 or 25°C. The amount of the 52-kDa proteolytic fragment was taken as a measure for productive translocation. A typical result out of three repeats is shown.