A substrate-independent, 14:3:3 protein-mediated plastid import pathway of NADPH:protochlorophyllide oxidoreductase A

Abstract

Plastids are semiautonomous organelles that contain only limited coding information in their own DNA. Because most of their genome was transferred to the nucleus after their endosymbiotic origin, plastids must import the major part of their protein constituents from the cytosol. The exact role of cytosolic targeting factors in the regulation of plastid protein import has not been determined. Here, we report that the nucleus-encoded NADPH:protochlorophyllide (Pchlide) oxidoreductase A plastid precursor (pPORA) can use two different plastid import pathways that differ by the requirements for cytosolic 14:3:3 proteins and Hsp70. pPORA synthesized in a wheat germ lysate segregated into different precursor fractions. While import of free pPORA and only Hsp70-complexed pPORA was Pchlide-dependent and involved the previously identified Pchlide-dependent translocon, 14:3:3 protein- and Hsp70-complexed pPORA was transported into Pchlide-free chloroplasts through the Toc75-containing standard translocon at the outer chloroplast membrane/translocon at the inner chloroplast membrane machinery. A 14:3:3 protein binding site was identified in the mature region of the ³⁵S-pPORA, which governed 14:3:3 protein- and Hsp70-mediated, Pchlide-independent plastid import. Collectively, our results reveal that the import of pPORA into the plastids is tightly regulated and involves different cytosolic targeting factors and plastid envelope translocon complexes.

Import of most plastid precursor proteins is transit sequence-dependent, requires ATP, and is mediated by translocon at the outer chloroplast membrane (Toc) and the translocon at the inner chloroplast membrane (Tic) complexes (1–6). In Arabidopsis, different versions of the Toc complex have been identified that differ by an interchange of receptor and GTPase components as well as precursor specificities (7–11). Ivanova et al. (12) and Kubis et al. (13) discovered that the main presequence receptor at Toc159 interacts with precursors to photosynthetic proteins, but it displayed little activity in binding assays with nonphotosynthetic precursors, including that of NADPH:protochlorophyllide (Pchlide) oxidoreductase A (pPORA) (14). Ivanova et al. (12) and Kubis et al. (13) discovered that the main presequence receptor atToc159 interacts with precursors to photosynthetic proteins, but it displayed little activity in binding assays with nonphotosynthetic precursors including that of NADPH:protochlorophyllide (Pchlide) oxidoreductase A (pPORA) (14). Import of this nucleus-encoded enzyme was shown to depend on the presence of Pchlide (15–17) and involve a translocon not identical with the Toc159- and Toc75-containing translocon used by photosynthetic precursors (18–20). Biochemical studies identified the pPORA translocon to consist of several unique components, called Pchlide-dependent translocon (Ptc) proteins, of which Ptc52 most likely operated as a Pchlide a oxygenase. The detection of such activity would imply a tight coupling of tetrapyrrole precursor biosynthesis and protein import and suggest a photoprotective role of the Pchlide-dependent pPORA import pathway for plant de-etiolation (15–20).

Interestingly, Dahlin et al. (21) and Aronsson et al. (22) described a substrate-independent import mode of pPORA in vitro. They used wheat germ-translated pPORA not pretreated with urea rather than urea-denatured or rabbit reticulocyte-translated precursor for their import experiments. This experimental difference suggests that cytosolic targeting factors present in wheat germ lysates may target pPORA to an import site other than the Ptc complex. In the present study, we identified such factors and demonstrate that pPORA is capable of interacting with 14:3:3 proteins and cytosolic Hsp70, which bound to a RTPpTFT motif in the mature part of pPORA and sorted the precursor to the trimeric Toc159/Toc75/Toc34 import machinery.

Results

Precursor Conformation Affects pPORA Import in Vitro.

To determine under which conditions pPORA import is Pchlide-dependent or Pchlide-independent, wheat germ-translated, ³⁵S-labeled pPORA was pretreated with urea and diluted or left urea-untreated and diluted (henceforth referred to as mock-incubated). The precursor was then added to barley chloroplasts that had been isolated on a sucrose/Percoll gradient (15). Fig. 1 shows that urea-denatured and diluted precursor was taken up only by Pchlide-containing chloroplasts [compare lane a versus lane c; see also supporting information (SI) Fig. 8A, lanes 7 and 8 versus 9 and 10]. By contrast, a fraction of nonurea-pretreated, but diluted, precursor was imported also into Pchlide-free chloroplasts (Fig. 1, lane d versus lane c and SI Fig. 8A, lanes 4 and 5 versus 2 and 3). Because pPORA translated in rabbit reticulocyte lysates was poorly imported under both conditions (Fig. 1, lanes g and h versus e and f and SI Fig. 8A, lanes 11–20), we assumed that a factor may be present in wheat germ extract that triggered substrate-independent import of pPORA. Consistent with this view, import of reticulocyte-translated pPORA into Pchlide-free chloroplasts could be rescued by addition of wheat germ extract (Fig. 1, lane j and SI Fig. 8C, lanes 2 and 3). A similar effect was observed for leaf extract (Fig. 1, lane l), but not for cotyledonary extract (Fig. 1, lane m) isolated from Arabidopsis thaliana. The wheat germ factor is most likely a protein. Wheat germ extract pretreated with urea (Fig. 1, lane i) or denatured by heat (Fig. 1, lane k) was inactive. The wheat germ factor is responsible for substrate-independent import of pPORA into chloroplasts but had no effect on import of ³⁵S-Pchlide oxidoreductase B (pPORB) (SI Fig. 8B). Consistent with the documented location of Pchlide oxidoreductase in plastids (23) ³⁵S-pPORA and ³⁵S-pPORB were not imported into isolated mitochondria regardless of what conditions were used (SI Fig. 8C, lanes 6–10 and 16–20, respectively). Controls with wheat germ- or rabbit reticulocyte-translated mitochondrial precursors, such as cytochrome c oxidase subunit Vb (At3g15640 gene product) and succinate dehydrogenase subunit 3 (At5g09600 gene product) (23) (SI Fig. 9), confirmed the functionality of the prepared barley mitochondria in terms of protein import.

Fig. 1.

The effect of urea on the import competence of wheat germ- and rabbit reticulocyte-translated pPORA into chloroplasts. ³⁵S-pPORA was synthesized in wheat germ (lanes a–d) or rabbit reticulocyte (lanes e–m) lysates. Then the translation mixtures were split. One aliquot was pretreated with urea and diluted (lanes a, c, e, g, and i), whereas the other aliquots were mock-incubated and diluted (lanes b, d, f, and h), before being added to isolated barley chloroplasts containing (+) or lacking (−) Pchlide produced by 5-ALA pretreatment. Import was studied for 15 min in darkness. After import, the assays were centrifuged. Plastids obtained in the sediment fraction were treated with thermolysin and extracted with trichloroacetic acid, and protein was resolved by SDS/PAGE and autoradiography. Protein found in the supernatant fraction obtained after centrifugation was precipitated with trichloroacetic acid, processed, and detected by SDS/PAGE autoradiography. In five additional incubations, the effect of wheat germ extract (WGE) (lanes i–k), Arabidopsis leaf extract (ALE) (lane l) and Arabidopsis cotyledonary extract (ACE) (lane m) on reticulocyte-synthesized pPORA was examined. Aliquots of the wheat germ extract were denatured with urea (lane i) or heat (lane k) or left untreated (lane j) before addition to the import reaction. Light gray and dark gray columns define the amounts of precursors and mature proteins, respectively, relative to the total amount of radioactivity (sum of imported and unimported precursors and mature proteins) in the import mixtures, set as 100. n.d., not detectable.

Wheat Germ-Translated pPORA Interacts with 14:3:3 Proteins and Hsp70.

To identify the wheat germ factor interacting with pPORA and leading to its Pchlide-independent import, size exclusion chromatography was performed. Wheat germ-translated barley ³⁵S-pPORA and ³⁵S-pPORB, as well as ³⁵S-presmall subunit of ribulose-1,5-bisphosphate carboxylase/oxygenase (pSSU) of soybean and ³⁵S-precursor ferredoxin (pFd) of Silene pratensis were cotranslated and applied to a Superose 6 column. Fig. 2 A and B shows that both ³⁵S-pPORA (44 kDa) and ³⁵S-pPORB (46 kDa) segregated into at least three different populations, displaying molecular masses of ≈250 kDa (fraction A), 125 kDa (fraction B), and 45 kDa (fraction C), respectively. Based on the differential elution behavior, we concluded that the ³⁵S-pPORA and ³⁵S-pPORB were present in at least two different, protein-bound forms, eluting in fractions A and B, respectively. The third fraction is most likely to be caused by free, nonprotein-bound precursor (fraction C). A fourth fraction, constituting 5–10% of total radioactivity, migrated with the void volume. This fraction was disregarded in subsequent experiments because it comprised import-incompetent precursor aggregates. For ³⁵S-pFd and ³⁵S-pSSU only two different precursor populations were found of which the faster eluting population had an elution time between fractions A and B containing the ³⁵S-pPORA and ³⁵S-pPORB (Fig. 2B, fractions 20–24). Western blotting using a heterologous anti-Hsp70 antiserum of tomato identified a 70-kDa band, which was most abundant in fractions 4–20 (Fig. 2C Upper). Parallel gel blot analyses with a 14:3:3 protein antiserum of tobacco detected a strong 35-kDa band and a weaker 33-kDa band, which both fractionated in highest amounts in fractions 10–14 and 26–28, respectively (Fig. 2C Lower). Indistinguishable 70-, 35-, and 33-kDa bands were obtained when wheat germ extract was subjected to immunoprecipitations with either antiserum (Fig. 2D). Respective controls with preimmune sera were negative (Fig. 2D, PIS1 and PIS2). Arabidopsis leaf extract provided similar results, whereas extracts of Arabidopsis cotyledons produced only tiny amounts of the 35- and 33-kDa bands that, in most cases, were below the limit of detection (Fig. 2E).

Fig. 2.

Size fractionation of wheat germ-translated ³⁵S-pPORA and ³⁵S-pPORB of barley, ³⁵S-pSSU of soybean, and ³⁵S-pFd of S. pratensis. ³⁵S-pPORA, ³⁵S-pPORB, ³⁵S-pSSU, and ³⁵S-pFd were coproduced from corresponding cDNA clones by coupled transcription/translation and cofractionated on a Superose 6 column. Molecular mass standards were run in parallel. (A) Individual fractions were harvested and an aliquot of each fraction was used for A₂₈₀ (▾) and radioactivity (■) measurements. (B and C) Another aliquot was resolved by SDS/PAGE and detected by either autoradiography (B) or Western blotting (C), by using antisera against tomato Hsp70 (Upper) and tobacco 14:3:3 protein (Lower). (D) In a control experiment, wheat germ protein was precipitated with the same anti-Hsp70 and anti-14:3:3 antisera as well as respective preimmune sera (PIS) and detected by SDS/PAGE and Western blotting. (E) For comparison, wheat germ extract (WGE) as well as Arabidopsis leaf (ALE) and cotyledonary extracts (ACE) were precipitated and processed identically. Stars mark nonspecific background bands.

pPORA Complexed with 14:3:3 Proteins and Hsp70 Is Imported into the Plastids in a Pchlide-Independent Manner.

To assess whether the size-fractionated precursors interacted with the detected 14:3:3 proteins and Hsp70, coimmunoprecipitations were performed. Fig. 3 shows that the ³⁵S-pPORA and ³⁵S-pPORB resolved in fraction A (see Fig. 2B) were precipitated by both the Hsp70 (Fig. 3A) and 14:3:3 protein (Fig. 3B) antisera. In fraction B, only Hsp70-complexed ³⁵S-pPORA and ³⁵S-pPORB were found, as deduced from their coprecipitation with the anti-Hsp70, but not anti-14:3:3 protein, antiserum (Fig. 3). In fraction C, low amounts of free ³⁵S-pPORA and ³⁵S-pPORB cofractionated with Hsp70 during gel filtration and gave rise to nonspecific background signals (Fig. 3A). For ³⁵S-pFd and ³⁵S-pSSU different results were obtained. Whereas ³⁵S-pFd was present in a 14:3:3-plus Hsp70-bound form, ³⁵S-pSSU could only be precipitated with the anti-Hsp70 antiserum (Fig. 3 A versus B). This result indicated that ³⁵S-pSSU, unlike ³⁵S-pFd, did not form stable complexes with 14:3:3 proteins.

Fig. 3.

Coimmunoprecipitation of size-fractionated ³⁵S-pPORA, ³⁵S-pPORB, ³⁵S-pSSU, and ³⁵S-pFd with antisera against tobacco 14:3:3 proteins and tomato Hsp70. (A and B) ³⁵S-pPORA and ³⁵S-pPORB were either cotranslated (A) or translated separately (B) and size-fractionated, and protein was found in the indicated fractions subjected to coimmunoprecipitation using antisera against tomato Hsp70 (A) or tobacco 14:3:3 proteins (B). In vitro-translated and size-fractionated ³⁵S-pSSU and ³⁵S-pFd were used as controls.

The import competence of the different size-fractionated precursors was studied in subsequent experiments. Fig. 4A shows that whereas free ³⁵S-pPORA or only Hsp70-complexed ³⁵S-pPORA were imported into 5-aminolevulinic acid (5-ALA)-pretreated, Pchlide-containing chloroplasts (lanes 2 and 3 plus 7 and 8 versus 4 and 5 plus 9 and 10), 14:3:3 protein-plus Hsp70-bound ³⁵S-pPORA was also taken up by Pchlide-free chloroplasts (lanes 12–15). For ³⁵S-pPORB, ³⁵S-pFd, and ³⁵S-pSSU, no such differences were observed (Fig. 4 B–D). The precursors were transported with similar efficiencies into Pchlide-free and Pchlide-containing chloroplasts regardless of whether free, 14:3:3 protein-bound and/or Hsp70-bound precursors were used for studying import (Fig. 4 B–D).

Fig. 4.

Import competence of ³⁵S-pPORA, ³⁵S-pPORB, ³⁵S-pSSU, and ³⁵S-pFd after their fractionation into different 14:3:3 protein- and/or Hsp70-complexed states. Wheat germ-translated ³⁵S-pPORA (A), ³⁵S-pPORB (B), ³⁵S-pSSU (C), and ³⁵S-pFd (D) were separated into free, Hsp70-complexed and 14:3:3 protein-plus Hsp70-complexed forms as described in Fig. 2. The different precursors then were added to isolated barley chloroplasts containing (+) or lacking (−) Pchlide produced by 5-ALA pretreatment. Import was tested in darkness for 15 min. Thl, thermolysin.

Identification of a 14:3:3 Protein Binding Site in the Mature PORA.

A common 14:3:3 protein binding site has been identified in pSSU of tobacco, the precursor of the major light-harvesting protein of photosystem II (pLHCII) of pea, and the pea 23- and 33-kDa precursor proteins of the oxygen-evolving complex (pOE23 and pOE33, respectively) (24–26) (Fig. 5A). A related RTPpTFT motif is present in the mature region of PORA (Fig. 5A, pPORA-mat; pT identifies the presumed phosphorylation site), whereas pPORB's transit sequence contains a RVNpTSS motif (Fig. 5A, pPORB-TP). We hypothesized that either motif might be involved in forming larger complexes with 14:3:3 proteins and Hsp70.

Fig. 5.

In vitro mutagenesis of the pPORA and pPORB and import of the resulting mutant precursors into chloroplasts. (A) Site-directed mutagenesis was used to replace Thr with Ala residues at positions 171 and 52 of the mature pPORA (pPORA-mat) and transit peptide (TP) of pPORB (pPORB-TP), respectively, in the defined 14:3:3 protein recognition motifs. Note that related motifs are present in the transit peptides of pSSU, OE23, and OE33 of pea. (B) The resulting pPORA* and pPORB* mutant precursors were translated in the wheat germ lysate in the presence of [³⁵S]methionine and tested for their import competence with barley chloroplasts containing (+) or lacking (−) Pchlide produced upon 5-ALA pretreatment. After import, aliquots were treated with (+) or without (−) thermolysin (Thl). (D) An aliquot of the generated mutant precursors was subjected to size fractionation and subsequent coimmunoprecipitation using the anti-Hsp70 antiserum. (C) As references, the authentic nonmutated pPORA and pPORB were used. To directly compare the elution behavior of the different precursors under identical conditions, mixtures of pPORA plus pPORB as well as pPORA* plus pPORB* were cofractionated (Top), whereas immunoprecipitations were carried out with precursors that had been separated individually (Middle and Bottom). Numbers below the autoradiograms indicate precursor levels (dpm × 10³) precipitated with the anti-Hsp70 antiserum relative to the total amount of radioactivity subjected to size fractionation (200 and 150 × 10³ dpm for pPORA* and pPORB*, respectively).

An in vitro-mutagenesis approach was taken to explore the involvement of the RTPpTFT and RVNpTSS motifs in pPORA and pPORB import. In either case, the presumed phosphorylation site, pT, was exchanged by an Ala residue. Subsequently, the derivatized pPORA* (T → A) and pPORB* (T → A) were translated in wheat germ lysate. Half of the incubation mixtures were size-fractionated as described in Fig. 2, whereas the other halves were immediately incubated with Pchlide-free or Pchlide-containing chloroplasts in a standard import reaction. Fig. 5B shows that ³⁵S-pPORA* (T → A) was taken up only by Pchlide-containing chloroplasts (lanes 2 and 3 versus lanes 4 and 5). By contrast, ³⁵S-pPORB* (T → A) was imported with similar efficiencies into Pchlide-free and Pchlide-containing chloroplasts (Fig. 5B, compare lanes 7 and 8 versus 9 and 10).

We next examined binding of 14:3:3 proteins and Hsp70 to the ³⁵S-pPORA* (T → A) and ³⁵S-pPORB* (T → A) mutant precursors by size exclusion chromatography. As controls, the nonmutagenized wild-type proteins were used. To trace potential differences in their elution behaviors, pPORA* (T → A) and pPORB* (T → A) as well as pPORA and pPORB were coseparated on the same Superose 6 column. Fig. 5 C Top and D Top illustrates that introduction of the mutation in either case drastically affected the elution behavior of the precursor. ³⁵S-pPORA* (T → A) and ³⁵S-pPORB* (T → A) no longer cofractionated in terms of larger, 14:3:3 protein- and Hsp70-containing complexes but were present in free and only Hsp70-bound forms (compare Fig. 5 C and D).

Photocross-Linking of Plastid Envelope Proteins that Interact with pPORA During Pchlide-Dependent and Pchlide-Independent Import.

pPORA was expressed as carboxyl-terminally tagged, hexa-His (His₆) precursor in Escherichia coli, purified, and derivatized with ¹²⁵I-N-[4[(p-azidosalicylamido)butyl]-3′(2-pyridyldithio) propionamid (APDP) (19). As controls, ¹²⁵I-APDP-pPORB and a ¹²⁵I-APDP-pSSU-protein A fusion protein, consisting of pea pSSU and protein A (6, 7), were produced identically. Then an aliquot of each precursor was reconstituted into the 14:3:3 protein- and Hsp70-complexed state and purified by size exclusion chromatography. Another aliquot was mock-incubated with wheat germ reaction buffer and gel-filtered.

The different precursors were incubated with isolated, energy-depleted barley chloroplasts in either the absence of added nucleoside triphosphates (−NTPs), the presence of 0.1 mM Mg-ATP, or the presence of 0.1 mM Mg-ATP plus 0.1 mM Mg-GTP. Taking into account previous findings of Kouranov and Schnell (6), these different nucleoside triphosphate combinations and concentrations were used to distinguish energy-independent binding (−NTPs) and energy-dependent binding (0.1 mM Mg-ATP) of the precursors to the plastids from their insertion (0.1 mM Mg-ATP plus 0.1 mM Mg-GTP) across the respective import machineries. After 15 min in darkness, the cross-linker was activated by UV light exposure. ¹²⁵I-labeled envelope proteins cross-linked by the different precursors were solubilized from isolated envelope membranes and detected by SDS/PAGE and autoradiography.

Fig. 6 depicts cross-link products of 130, 52, 37, 33, 30, 26, 22, 20, and 16 kDa that were formed with pPORA lacking 14:3:3 protein and Hsp70. The labeling intensity of these bands varied with the incubation conditions, most likely reflecting different stages of the import process (Fig. 6A). Cross-linking of ¹²⁵I-APDP-pPORA containing 14:3:3 protein and Hsp70 yielded a different polypeptide pattern, comprising main 159-, 86-, 75-, 68-, 64-, 60-, 52-, 46-, 40-, and 22-kDa bands (Fig. 6B). Immunoprecipitations (Fig. 6A) proved that the 16-, 52-, and 130-kDa proteins were caused by Ptc/Oep16, Ptc52, and Ptc130, respectively. By contrast, the 75-kDa product cross-linked to the 14:3:3 protein- and Hsp70-complexed precursor was identified as barley Toc75 (Fig. 6B). The 159- and 86-kDa bands, which were likewise labeled, most likely represent orthologs of Arabidopsis and pea Toc159 and its major proteolytic fragment, Toc86 (Fig. 6B). Barley Toc159 appears to be hypersensitive to proteolysis as are its pea (9, 27) and Arabidopsis (10) counterparts, and it was degraded into several lower molecular mass products, including Toc86, during sample handling (Fig. 6B).

Fig. 6.

¹²⁵I-APDP cross-linking of 14:3:3 protein and/or Hsp70-complexed as well as free precursors to plastid envelope proteins. (A) Cross-linking (lanes a–c) and immunoprecipitations (lanes a′–c′ and a″-c″) of plastid envelope proteins interacting with the free, non-14:3:3 protein and Hsp70-complexed ¹²⁵I-APDP-pPORA. For the immunoprecipitations, anti-Ptc130, anti-Ptc16, and anti-Ptc52 antisera (lanes a′–c′) as well as respective preimmune sera (PIS) (lanes a″–c″) were used. (B) Cross-linking (lanes d–f) and immunoprecipitations (lanes d′, f′, d″, and f″) obtained for the 14:3:3 protein and Hsp70-complexed ¹²⁵I-APDP-pPORA with the indicated anti-Toc159 and anti-Toc75 antisera (lanes d′ and f′) and respective preimmune sera (lanes d″ and f″). (C–F) Cross-linking of plastid envelope proteins interacting with free (lanes a–c) and 14:3:3 protein and Hsp70-complexed (lanes d–f) ¹²⁵I-APDP-pPORB (C and D) and free (lanes a–c) and Hsp70-complexed (lanes d–f) ¹²⁵I-APDP-pSSU-protein A (E and F).

Cross-linking of ¹²⁵I-APDP-pPORB (Fig. 6 C and D) and ¹²⁵I-APDP-pSSU-protein A (Fig. 6 E and F) yielded basically the same cross-link patterns as that found for 14:3:3 protein and Hsp70-containing ¹²⁵I-APDP-pPORA. Interestingly, in neither case could difference be observed for the 14:3:3 protein plus Hsp70-complexed versus free precursors. However, the signal intensities of the detected cross-link products were in several cases lower than that observed for the ¹²⁵I-APDP-pPORA, which may be a reflection of the proximity and/or strength of interaction between the precursors studied and respective components of the protein import machinery. Similar conclusions had been drawn by Kouranov and Schnell (6) and Ma et al. (7).

Chlorina Mutants Lacking Chlorophyllide a-Oxygenase Import pPORA in a Protochlorophyllide b-Dependent Manner.

Ptc52 was previously proposed to play a key role in the Pchlide-dependent import pathway of pPORA (19). Time courses over import suggested that most likely Ptc52 operated as Pchlide a-oxygenase to provide Pchlide b as the import trigger (19). To test this hypothesis, activity tests were performed with isolated Ptc complex. Import intermediates containing pPORA-(His)₆ trapped in junction complexes between the outer and inner envelope membranes were prepared and purified from lysed Arabidopsis wild-type and chlorina (ch)1–3 plastids, solubilized (28), and incubated with Pchlide a, O₂, Fd, and a Fd-reducing system comprising glucose-6-phosphate, NADPH, glucose-6-phosphate dehydrogenase, and Fd:NADPH oxidoreductase (28). Both plastid types were used to exclude a potential redundancy between the Ptc52 and chlorophyllide a oxygenase (CAO) reactions.

Fig. 7 highlights Pchlide a to Pchlide b conversion both with Ptc complex isolated from Arabidopsis wild-type and ch1–3 plastids. The identity of Pchlide b was confirmed by absorbance measurements and mass spectrometry, using synthetic standards (28). Because CAO-deficient ch1–3 plastids imported wheat germ-translated, urea-denatured pPORA in a Pchlide-dependent manner, we conclude that not CAO, but Ptc52, was the enzyme responsible for Pchlide b synthesis in these plants.

Fig. 7.

Ptc52 activity in relation to pPORA import in wild-type and chl1–3 plastids. Ptc complex was isolated from wild-type and ch1–3 plastids of Arabidopsis. Enzyme activity measurements were carried out with Pchlide a, O₂, and an electron-producing system. Pigments were extracted with 100% acetone containing 0.1% (vol/vol) diethyl pyrocarbonate and subjected to HPLC on a reversed-phase column. Pigment detection was made by absorbance measurements at 455 nm, the Soret band of Pchlide b. (A) The curves show pigments extracted from isolated Ptc complex of WT (dashed line) and ch1–3 (red dotted line) plastids after 15 min in darkness, as compared with time 0 (solid line). Peak 1 eluting at 12.5 min corresponds to Pchlide b, as determined by absorbance measurements, whereas peak 2 eluting at 15 min corresponds to Pchlide a (data not shown). (B) (Upper) The identity of Pchlide b was established by matrix-assisted laser desorption/ionization mass spectroscopy. (Lower) The theoretical isotopic distribution of C₃₅H₃₀N₄O₆Mg corresponding to Pchlide b is indicated. (C) Time course of Pchlide b synthesis obtained with ch1–3 and WT Ptc complexes. (D) Import of wheat-germ-translated, urea-denatured pPORA into ch1–3 and WT chloroplasts, which had been pretreated with (+) or without (−) 5-ALA, to give rise to intraplastidic Pchlide synthesis and repurified. Note the lack of imported and processed PORA in mock-incubated ch1–3 and wild-type plastids and the restoration of import by 5-ALA pretreatment in either case. Thl, thermolysin.

Discussion

In the present study, evidence is provided for a default in vitro-import pathway of pPORA. This pathway, which leads to pPORA's uptake into Pchlide-free chloroplasts, requires wheat germ 14:3:3 proteins and Hsp70, Interestingly, rabbit reticulocyte lysate also contains 14:3:3 proteins and Hsp70, but the established larger complexes were not functional in terms of pPORA's chloroplast default import. SI Figs. 8–10 prove the specificity of plant and animal 14:3:3 proteins and Hsp70 in chloroplast and mitochondrial protein import, respectively.

Binding of wheat germ 14:3:3 proteins occurs to a RTPpTFT motif in the mature region of PORA. This result disproves the previously postulated specificity of the guidance complex comprising 14:3:3 proteins and Hsp70 for phosphorylated transit sequences (26). Jackson-Constan et al. (29) pointed out that many transit sequences, including that of pSSU of soybean used in this study, do not contain the 14:3:3 protein consensus RSX_1,2pSX(P) motif. We observed that ³⁵S-pPORB and ³⁵S-pFd containing the 14:3:3 recognition motif were imported with similar efficiencies into isolated barley chloroplasts before and after denaturation with urea. Nakrieko et al. (30) showed that removal of the presumed phosphorylation site in a plastid presequence had no impact on subsequent chloroplast protein import in vivo.

The Pchlide dependency of pPORA import, which is mediated by Ptc52, but not CAO, is developmentally regulated and was observed for cotyledonary cells, but not for leaf mesophyll cells of Arabidopsis (31). Interestingly, only trace amounts of 14:3:3 protein interacting with our antiserum were detectable in extracts of Arabidopsis cotyledons and were obviously insufficient to trigger substrate-independent import in vitro (this study) and in planta (31). The protein kinase implicated in the phosphorylation of the target Thr residue in the 14:3:3 recognition motif may likewise be absent or inactive in cotyledonary cells. The relative prevalence and implicit precursor affinities (K_D values) of the trimeric Toc159/Toc75/Toc34 and Ptc import machineries in plastids may distinguish cotyledonary and mesophyll cells.

In conclusion, this study and that of Kim and Apel (31) confirm the existence of the substrate-dependent import pathway for pPORA. This pathway is most active in etiolated plants. At this stage of development, the newborn seedlings prepare for light harvesting and the switch from heterotrophic to photoautotrohic growth. In older plants, the pathway remains active as well, but its operation is superimposed by a second pathway that requires binding of 14:3:3 proteins and Hsp70 and can explain the previously observed substrate-independent import mode of pPORA (21, 22, 32). The role of this pathway could be to provide a back-up system for the sequestration of free pigment in a nonhazardous form.

Materials and Methods

DNA Constructs.

cDNAs for the different plastid precursors used in this study have been described (7, 18). Mitochondrial cytochrome c oxidase subunit Vb (At3g15640 gene product) and succinate dehydrogenase subunit 3 (At5g09600 gene product) cDNAs were isolated by a PCR-based approach (33).

Production and Purification of Precursors.

Precursors were synthesized by coupled in vitro-transcription/translation or produced in Escherichia coli. pPORA-(His)₆- and pPORB-(His)₆-tagged precursors were expressed in strain SG13009 (Qiagen, Valenica, CA) and purified as specified (18, 19).

Cell-free protein synthesis was carried out in self-made wheat germ and rabbit reticulocyte lysates. Size exclusion chromatography was performed on prepacked Superose 6 (model HR10/10; Amersham Biosciences, Piscataway, NJ) columns that had been equilibrated with buffer containing 50 mM Tris·HCl, pH 7.4, and 50 mM NaCl (26).

Protein Import.

Protein import was studied as described (15), by using in vitro-translated or bacterially expressed precursors and isolated plastids or mitochondria that had been prepared from 5-day-old, light- and dark-grown barley plants, respectively, in media containing sucrose. Treatment of isolated organelles with 5-ALA dissolved in 10 mM phosphate buffer, pH 8.0, or phosphate buffer alone was performed as described (15). Energy depletion of reisolated organelles was achieved according to Theg et al. (34). Details on the actual conditions used to perform the import reactions and postimport treatments have been described (35) and are provided in SI Text.

Cross-Linking.

Derivatization of pPORA-(His)₆ and pPORB-(His)₆ with APDP and subsequent cross-linking were carried out as described by Kouranov and Schnell (6) and Ma et al. (7). Organelle proteins were extracted with trichloroacetic acid, washed with acetone, ethanol, and ether, and resuspended in SDS-sample buffer (36). Heat-denatured proteins were analyzed electrophoretically and detected by autoradiography.

Activity Measurements of PTC52.

Ptc complex was isolated from junction complexes containing pPORA-(His)₆ trapped between the outer and inner envelope membranes as described (19). As plastid source, 5-day-old, light-grown Arabidopis wild-type or ch1–3 seedlings were used. Activity measurements were carried out according to Reinbothe et al. (28). Final 50-μl assays contained 100 μg of isolated Ptc complex and the following supplements: 2 mM Pchlide a, 10 μg of Fd (Sigma, St. Louis, MO), and a Fd-reducing system [2 mM glucose-6-phosphate; 1 mM NADPH; 50 milliunits of glucose-6-phosphate dehydrogenase; 10 milliunits of Fd-NADPH-oxidoreductase (Sigma)] (28). Separation and identification of pigments was done by reversed-phase HPLC, combined with absorbance measurements and matrix-assisted laser desorption/ionization TOF MS (Voyager DE STR Biospectrometry Work Station; Applied Biosystems, Foster City, CA) (28).

Other Procedures.

Immunoprecipitation was performed with heterologous antisera against 14:3:3 protein preparation of tobacco and Hsp70 of tomato (19).

Abbreviations

5-ALA

5-aminolevulinic acid

APDP

N-[4[(p-azidosalicylamido)butyl]-3′(2-pyridyldithio) propion amid

CAO

chlorophyllide a oxygenase

pFd

precursor ferredoxin

Pchlide

protochlorophyllide

pPORA

Pchlide oxidoreductase A

pPORB

Pchlide oxidoreductase B

pSSU

presmall subunit of ribulose-1,5-bisphosphate carboxylase/oxygenase

Ptc

Pchlide-dependent translocon

Toc

translocon at the outer chloroplast membrane.