Sec-dependent Pathway and ΔpH-dependent Pathway Do Not Share a Common Translocation Pore in Thylakoidal Protein Transport*

Thylakoidal proteins of plant chloroplasts are transported to thylakoids via several different pathways, including the ΔpH-dependent and the Sec-dependent pathways. In this study, we asked if these two pathways utilize a common translocation pore. A fusion protein consisting of a 23-kDa subunit of the oxygen evolving complex and Escherichia coli biotin carboxyl carrier protein was biotinylated in E. coli cells and purified. When incubated with isolated pea thylakoids in the absence of avidin, the purified fusion protein was imported into the thylakoids via the ΔpH-dependent pathway. However in the presence of avidin, the fusion protein became lodged in the thylakoid membranes, with its N terminus reaching the thylakoidal lumen, while its C-terminal segment complexed with avidin exposed on the thylakoidal surface. The translocation intermediate of the fusion protein inhibited the import of authentic 23-kDa subunit, suggesting that it occupies a putative translocation pore for the ΔpH-dependent pathway. However the intermediate did not block import of the 33-kDa subunit of the oxygen evolving complex, which is a substrate for the Sec-dependent pathway. These results provide evidence against the possibility of a common translocation pore shared by the Sec-dependent pathway and the ΔpH-dependent pathway.

Nuclear-encoded thylakoidal proteins are synthesized in the cytosol, are imported into the chloroplast stroma, and are subsequently inserted into or translocated across the thylakoid membranes. Recent evidence has shown that thylakoidal proteins are transported to the thylakoids via several different pathways, including the Sec-dependent and the ⌬pH-depend-ent pathways (1,2). In bacterial cells, secretory proteins are translocated across the cytoplasmic membrane through a channel consisting of SecY/E/G proteins with the aid of a translocation ATPase, SecA. In chloroplasts, homologues of SecA (cpSecA), SecY (cpSecY) and SecE (cpSecE) have been identified and shown to mediate transport of a class of thylakoid lumenal proteins in vitro and in vivo (3)(4)(5)(6)(7)(8)(9)(10)(11)(12). Another class of thylakoid lumenal proteins is translocated across the thylakoid membranes via a transport pathway that requires ⌬pH across the thylakoid membranes as a sole energy source and Hcf106 protein in the thylakoid membranes (13)(14)(15). This ⌬pH pathway appears to have ability to translocate folded proteins that are not accepted by the Sec-dependent pathway (16 -18). Recent studies have revealed that a protein transport pathway similar to the thylakoidal ⌬pH-dependent pathway is present in bacterial cells (19 -21).
Although the Sec-dependent and the ⌬pH-dependent pathways appear to utilize some pathway-specific components, they could share common translocation pore in the thylakoid membranes. In vitro competition experiments have shown that the import of a certain substrate protein into isolated thylakoids is effectively competed with saturating amounts of another competitor protein which utilizes the same thylakoidal transport pathway as the substrate protein but not with competitor proteins using other pathways (21). However the competition could occur only in the rate-limiting step of the transport, and the passage through a possible common translocation pore might not correspond to this step. A maize mutant lacking cpSecY shows more severe defects in thylakoid biogenesis than mutants lacking cpSecA, suggesting a possibility that cpSecY is involved in the thylakoidal protein transport along the pathways other than the Sec-dependent one (12).
One of the crucial tests for the possibility of the common translocation pore shared by different thylakoidal transport pathways would be to generate a saturating amount of a translocation intermediate that occupies most of the translocation pores for one pathway and to see its competition effects on the protein transport along other pathways. Such a translocation intermediate will span the thylakoid membranes with topology in which its N terminus reaches the thylakoid lumen while its C terminus remains outside the thylakoids. Since precursor proteins bearing avidin, which forms a stable tetrameric structure and tightly binds biotin, have been successfully used to generate a translocation intermediate spanning the mitochondrial membranes (23) and that spanning the chloroplast envelope (24), we supposed that the bulky avidin would not be accommodated by the translocation pore for the ⌬pH pathway. In this study, a fusion protein (23K-BCCP) 1 consisting of a wheat 23-kDa protein of the oxygen evolving complex (23K) and Escherichia coli biotin carboxyl carrier protein (BCCP), which is efficiently biotinylated in vivo (25), was expressed in E. coli cells. In in vitro import experiments with isolated pea thylakoids in the presence of avidin, the purified fusion protein became lodged in the thylakoid membranes, with its N terminus reaching the lumen and its C-terminal segment complexed with avidin exposed on the thylakoid surface. The translocation * This work was supported by a grant for the "Biodesign Research Program" from the Institute of Physical and Chemical Research (RIKEN) and by a grant-in-aid for Scientific Research from the Ministry of Education, Science and Culture of Japan. 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  intermediate of the fusion protein specifically inhibited the translocation via the ⌬pH pathway but not that via the Sec pathway. These results provide strong evidence against the possibility of a common translocation pore shared by the two pathways.

EXPERIMENTAL PROCEDURES
Materials-Most of the reagents were purchased from Nacalai Tesque, Sigma, Takara, and Toyobo. Chloroplasts, chloroplast lysate prepared by osmotic lysis, thylakoids, and stromal fraction were prepared from pea seedlings (Pisum sativum, var. Alaska) as described previously (3).
Plasmids-A DNA fragment for 69 -156 amino acid residues of BCCP (25) was amplified from E. coli chromosome and 3Ј-terminally fused with that for a c-Myc epitope by PCR. The fragment was inserted into pET-21d (Novagen) at the HindIII/XhoI site to yield pET-BCCP/MHd. A DNA fragment encoding the processing intermediate form of wheat 23K (i23K) (3) was amplified by PCR and inserted into the NcoI/EcoRI site of pET-BCCP/MHd. The resulting plasmid, pET-i23K/BCCP, expresses an i23K-BCCP fusion protein shown in Fig. 1A. To construct pBirA, a biotin ligase overproduction plasmid, the birA gene with its own promoter was amplified by PCR from E. coli chromosome and was subcloned into pSTV28 (Takara).
Production of Biotinylated i23K-BCCP-A BL21(DE3)/pBirA strain harboring pET-i23K/BCCP was cultured in terrific broth with 0.1 mM d-biotin, 34 g/ml chloramphenicol, and 50 g/ml ampicillin at 37°C and exposed to 1 mM isopropyl-␤-D-thiogalactopyranoside for 2 h. The cells were harvested and disrupted by sonication in 20 mM Tris-HCl, pH 7.9, 0.5 M NaCl, and 5 mM imidazole. A supernatant fraction from the centrifugation at 100,000 ϫ g for 30 min was applied onto a His Bind Resin column (Novagen). i23K-BCCP was eluted from the column with the buffer containing 0.2 M imidazole, dialyzed against 20 mM HEPES-KOH, pH 8.0, and stored at Ϫ80°C.
In Vitro Import into Pea Thylakoids-For import assays with radiolabeled substrates, 35 S-labeled processing intermediate forms of 23K (i23K) and 33K (i33K) were translated in a wheat germ cell-free system in the presence of Tran 35 S-label™ (ICN, Inc.). In vitro import of i23K, i33K, and i23K-BCCP into isolated pea thylakoids were performed as described previously (3). The radiolabeled 23K and 33K were detected by radioimaging with Storm 860 image analyzer (Molecular Dynamics). 23K-BCCP was detected by immunoblotting with anti-c-Myc monoclonal antibody 9E10.

RESULTS AND DISCUSSION
Preparation of a Biotinylated Substrate, i23K-BCCP-In order to halt the translocation of substrate proteins for the ⌬pHdependent pathway across thylakoid membranes, we utilized a biotinylated 23K protein complexed with a stably folded and bulky avidin tetramer. We introduced a biotin moiety into i23K, a wheat 23K protein with a truncated transit peptide lacking a chloroplast targeting signal, by expressing it as a fusion protein between i23K and residues 69 -156 of BCCP in E. coli cells (Fig. 1A). It is known that Lys-122 in the above segment of BCCP is biotinylated in E. coli cells (25). A c-Myc/ (His) 6 -tagged version of the fusion protein was expressed together with a BirA protein (biotin ligase) in E. coli BL21(DE3) cells, which were cultivated in the medium containing 0.1 mM biotin. The expressed 42-kDa fusion protein was soluble in the cytosol and was purified by metal affinity chromatography for the (His) 6 tag (Fig. 1B, lane 1). About 70% of the purified i23K-BCCP was found to be biotinylated (Fig. 1B, lane 2).
Next, we tested whether the purified fusion protein can be imported into isolated thylakoids via the ⌬pH-dependent pathway like authentic i23K. Incubation of purified i23K-BCCP with isolated thylakoids converted it to a 39-kDa mature form (m23K-BCCP), which was protected against exogenous thermolysin (Fig. 1C, lanes 1 and 2), indicating that the fusion protein was imported into the thylakoids. The amount of proteaseprotected m23K-BCCP was not dependent on the extent of biotinylation of the BCCP segment (not shown). The transport into thylakoids was inhibited by nigericin (lanes 3 and 4), a protonophore dissipating ⌬pH across the thylakoid mem-branes, but was not affected by NaN 3 (lanes 5 and 6), an inhibitor for the Sec pathway. These results indicate that addition of the C-terminal BCCP moiety and its biotinylation do not impair the transport of the 23K part into the thylakoids via the ⌬pH-dependent pathway.

Translocation of Biotinylated i23K-BCCP Complexed with Avidin across the Thylakoid Membranes Is Arrested after the Signal Cleavage
Step-We tried to interfere with thylakoidal transport of i23K-BCCP via the ⌬pH-dependent pathway by adding avidin to the reaction. Biotinylated i23K-BCCP was first incubated with or without avidin, and isolated thylakoids were subsequently added to the reaction. In the absence of avidin, the fusion protein was imported into the thylakoids ( Fig. 2A, lanes 2 and 3). However, when preincubated in the presence of avidin, the fusion protein complexed with avidin was converted to the mature form (lane 4) but remained sensitive to externally added protease (lane 5). A large amount of the unprocessed form of the fusion protein bound to the surface of the thylakoids (lane 4) probably because basic protein avidin itself interacts with acidic phospholipids of the membranes. In control experiments, the import of authentic i23K into thylakoids was not inhibited by this amount of avidin alone (data not shown).
After the import reaction for the fusion protein in the presence of avidin, the thylakoids recovered by centrifugation were treated with various reagents and fractionated into the membranes and the soluble fractions. When the thylakoids were washed with hypotonic buffer alone, both the unprocessed and the mature forms were mainly recovered in the membrane fraction (Fig. 2B, lane 3). A large fraction of the unprocessed and the mature forms were extracted from the membranes with as low as 0.5 M NaCl (lanes 5 and 6). The mature form was completely extracted with 0.2 M Na 2 CO 3 , pH 11.0 (lanes 9 and 10) and also with 4 M urea (data not shown). These results suggest that the mature form is associated with the thylakoid membranes through protein-protein interactions. This would be consistent with a possibility that the mature form spans the thylakoid membranes through a protein translocation pore.
23K-BCCP-Avidin Complex Blocks the ⌬pH-dependent Transport Pathway but Not the Sec-dependent Pathway-The above results suggest that the i23K-BCCP complexed with avidin initiates translocation, but fails to complete it because the bulky avidin cannot be accommodated by the translocation pore for the ⌬pH-dependent pathway. If the protease-sensitive mature form of 23K-BCCP represents a genuine translocation intermediate that is trapped in the translocation pore, it should block import of authentic substrates for the ⌬pH pathway into the thylakoids. Thylakoids were thus incubated with large amounts of biotinylated i23K-BCCP in the presence or absence of avidin, recovered by centrifugation to remove excess amounts of the fusion protein and avidin, and subjected to the second incubation with radiolabeled i23K, an authentic substrate for the ⌬pH pathway. When i23K-BCCP was omitted during the first incubation, the thylakoids could import i23K efficiently whether or not avidin was present during the first incubation (Fig. 3A, lanes 2, 5, and 8). In contrast, when thylakoids were incubated with increasing amounts of i23K-BCCP in the presence of avidin, the reisolated thylakoids retained less ability to import i23K (Fig. 3A, lanes 3, 4, 6, and 7). The presence of avidin during the first incubation was essential to block the import of i23K during the second incubation (Fig. 3A,  lanes 8 -11). The first incubation with 10 g of i23K-BCCP and avidin blocked the import of i23K in the second reaction to about 20% of the control reaction (Fig. 3A, lane 7). Therefore, we conclude that at least a part of protease-sensitive 23K-BCCP-avidin molecules represents a true translocation intermediate. The 23K-BCCP-avidin complex likely occupies most of the translocation pores for the ⌬pH-dependent pathway.
Next, we asked if the translocation intermediate for the ⌬pH-dependent pathway described above can block import of substrates for the Sec-dependent pathway. Thylakoids were treated as in Fig. 3A and subjected to the second incubation with radiolabeled i33K, an authentic substrate for the Sec-dependent pathway. The stromal fraction was added to the reaction to supply cpSecA, which is essential for the Sec-dependent transport (3,4). When thylakoids were incubated with 10 g of i23K-BCCP and avidin, which were sufficient to block 80% import of i23K during the second incubation, the reisolated thylakoids did not lose the ability to import i33K (Fig. 3B). Therefore, inhibition of thylakoidal protein transport by the arrested translocation intermediate of i23K-BCCP-avidin complex is specific for the ⌬pH-dependent pathway. This means that the translocation pore for the ⌬pH-dependent pathway is not shared by substrates for the Sec-dependent pathway. The ⌬pH-dependent pathway and the Sec-dependent pathway most likely have their own translocation pores for protein passage across the thylakoid membranes.
Successful generation of the membrane-spanning translocation intermediate that occupies ⌬pH-dependent translocation pores enables us to estimate the number of the import sites. Titration of the import sites for the ⌬pH-dependent pathway in the thylakoid membranes with i23K-BCCP-avidin complex showed that the import sites were saturated by 73 Ϯ 7 ng of the fusion protein/75 g of chlorophyll of thylakoids (Fig. 4). This means that, since an average chlorophyll content per chloroplast is Ϸ1 pg, 2 a single chloroplast is able to accumulate 15,000 molecules of the i23K-BCCP-avidin complex occupying the import machinery for the ⌬pH-dependent pathway. However, maximal import rates of only 700 -900 molecules of proteins/one chloroplast/min have been reported for the ⌬pH pathway (22,26). This may mean that passage through the translocation pore is not a rate-limiting step in the ⌬pH-dependent pathway across the thylakoid membranes, although we cannot rule out the possibility that a part of the 23K-BCCP molecules is not associated with the translocation pore.
The present study has revealed that the ⌬pH-dependent pathway and the Sec-dependent pathway for thylakoidal protein transport do not converge at the translocation pores. This is consistent with the observation that the translocation across the cytoplasmic membrane via the Tat system in E. coli, a 2 T. Nohara, unpublished data.  lanes 3 and 5). Lane 1 contains 0.1 g of i23K-BCCP as a standard. B, after the import reaction with avidin, the thylakoids were recovered by centrifugation at 4,000 ϫ g for 10 min, treated with 10 mM HEPES-KOH, pH 8.0 ( lanes  2 and 3), 0.5 M NaCl (lanes 5 and 6), 1.0 M NaCl (lanes 7 and 8), or 0.2 M Na 2 CO 3 (lanes 9 and 10), and separated into supernatants (lanes marked by "S") and pellets (lanes marked by "P") by centrifugation at 200,000 ϫ g for 30 min. Lanes 1 and 4 contain 0.1 g of i23K-BCCP. FIG. 3. i23K-BCCP blocks protein transport via the ⌬pH pathway but not that via the Sec pathway. A, indicated amounts (0 -10 g) of biotinylated i23K-BCCP were incubated with 150 g of chlorophyll of chloroplast lysate in the absence (lanes 8 -11) or presence of avidin (lanes 2-7) at 25°C for 30 min in the light. The thylakoids were then recovered as described in the legend to Fig. 2B, suspended in the import buffer with the stromal fraction, and subjected to the second import reaction for radiolabeled i23K. After 15-min incubation, the thylakoids were recovered, treated with thermolysin, and subjected to protein analysis by SDS-PAGE. Amounts of the mature form of 23K were quantified; the amount for m23K without i23K-BCCP and avidin during the first incubation (lane 8) was set to 100%. B, two-step import was performed as in A except that radiolabeled i33K was used as a substrate in the second import. Amounts of the mature form of 33K were quantified and expressed as in A. Lane 1 of each panel contains 2.5% of i23K or i33K added to the reaction. i and m represent intermediate and mature forms, respectively. Bands marked with asterisks are unrelated to the import substrates.
translocation pathway related to the thylakoidal ⌬pH pathway, is independent of SecY and SecE (20). In analogy to the bacterial Sec system, the translocation pore for the Sec pathway likely consists of cpSecY/cpSecE and unidentified other chloroplast homologues of Sec proteins. The subunits constituting the translocation pore for the ⌬pH-dependent pathway have not been identified yet, but Hcf106 is at least functionally linked to the pore. The successful generation of the translocation intermediate lodged in the thylakoid membranes will provide a valuable tool to identify components of the translocation pore for the ⌬pH-dependent pathway in the future study.