Plant Physiol. (1998) 116: 1179-1190

Transit Peptide Mutations That Impair in Vitro and in Vivo Chloroplast Protein Import Do Not Affect Accumulation of the -Subunit of Chloroplast ATPase¹

Karen L. Kindle^* and Susan D. Lawrence²

Plant Science Center, Biotechnology Program, 151 Biotechnology Building, Cornell University, Ithaca, New York 14853

	ABSTRACT

Top Abstract Introduction Methods Results Discussion References

We have begun to take a genetic approach to study chloroplast protein import in Chlamydomonas reinhardtii by creating deletions in the transit peptide of the -subunit of chloroplast ATPase-coupling factor 1 (CF₁-, encoded by AtpC) and testing their effects in vivo by transforming the altered genes into an atpC mutant, and in vitro by importing mutant precursors into isolated C. reinhardtii chloroplasts. Deletions that removed 20 or 23 amino acid residues from the center of the transit peptide reduced in vitro import to an undetectable level but did not affect CF₁- accumulation in vivo. The CF₁- transit peptide does have an in vivo stroma-targeting function, since chimeric genes in which the stroma-targeting domain of the plastocyanin transit peptide was replaced by the AtpC transit peptide-coding region allowed plastocyanin to accumulate in vivo. To determine whether the transit peptide deletions were impaired in in vivo stroma targeting, mutant and wild-type AtpC transit peptide-coding regions were fused to the bacterial ble gene, which confers bleomycin resistance. Although 25% of the wild-type fusion protein was associated with chloroplasts, proteins with transit peptide deletions remained almost entirely cytosolic. These results suggest that even severely impaired in vivo chloroplast protein import probably does not limit the accumulation of CF₁-.

	INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

Most chloroplast proteins are encoded in the nucleus, synthesized in the cytosol as precursors, and imported into the chloroplast posttranslationally. The targeting information for translocation across the chloroplast envelope membranes resides in an N-terminal TP (for reviews, see Keegstra et al., 1989; de Boer and Weisbeek, 1991; Theg and Scott, 1993; Gray and Row, 1995; Schnell, 1995; Cline and Henry, 1996). Proteins destined for chloroplast compartments other than the stroma require additional targeting information. For inner envelope and integral thylakoid membrane proteins, the targeting signal appears to be in the mature protein. Precursors of thylakoid lumen proteins have a bipartite TP. The N-terminal part contains the stroma-targeting domain, which is removed by stromal processing to yield a translocation intermediate. The C-terminal part contains a hydrophobic region resembling bacterial signal sequences that directs translocation across the thylakoid membrane into the lumen using one of two targeting pathways (for reviews, see Keegstra, 1989; Robinson and Klösgen, 1994).

Although there is no primary amino acid sequence consensus among the stroma-targeting domains of TPs, they do share certain features. For example, they are enriched in basic and hydroxylated residues and lack acidic residues and Tyr. The stroma-targeting domains of vascular plant TPs have three domains: (a) an N-terminal part that is deficient in charged residues, Gly, and Pro; (b) a middle domain that is enriched in Ser, Thr, Lys, and Arg; and (c) a C-terminal portion that is predicted to form an amphiphilic -strand (von Heijne et al., 1989). Chlamydomonas reinhardtii TPs are significantly shorter than those of vascular plants. Although their stroma-targeting domains are also predicted to have a tripartite domain structure, the uncharged N-terminal region is short, and the central region has a predicted -helical nature more typical of a mitochondrial targeting signal. As in vascular plants, the C-terminal part is predicted to form an amphiphilic -strand (Franzén et al., 1990).

Numerous in vitro studies have established that the TP is both necessary and sufficient for targeting chloroplast or foreign proteins to chloroplasts (for reviews, see Keegstra et al., 1989; de Boer and Weisbeek, 1991). In most cases, deletions throughout the TP have dramatic effects on in vitro import (Reiss et al., 1987; Archer and Keegstra, 1993; Bassham et al., 1994; Pilon et al., 1995; Lawrence and Kindle, 1997), suggesting that most parts of the TP contribute to the import process. Mutations in various parts of the stroma-targeting domain affect binding, translocation, and stromal processing to different degrees. However, the results have differed somewhat among various TPs; therefore, a detailed model for functional regions of the stroma-targeting region has not emerged. The N terminus appears to be involved in binding to and translocation across the envelope, whereas the C terminus plays a stronger role in defining the site and efficiency of stromal processing (Pilon et al., 1995; Cline and Henry, 1996).

It is generally assumed that in vitro import assays accurately reflect processes that occur in vivo, although this assumption has not been rigorously tested. In fact, different results were obtained in vivo and in vitro with two constructs in which the SSU TP-coding region was fused to that of neomycin phosphotransferase (NptII) with or without 23 amino acid residues from the N terminus of a mature SSU (Kuntz et al., 1986; Wasmann et al., 1986). Similarly, a fusion of the PC TP to -lactamase resulted in a chimeric precursor that was correctly localized to the lumen in vivo but not in vitro (de Boer et al., 1991). Fusion proteins that included the TP and 5 or 23 N-terminal amino acid residues of a chloroplast inner envelope protein were imported into both chloroplasts and mitochondria in vitro. However, in vivo the shorter fusion protein remained in the cytosol, and the longer one was localized in chloroplasts (Silva-Filho et al., 1997). The interpretations in these studies have been somewhat limited, because they involved chimeric proteins rather than native chloroplast proteins and were tested using different species for in vitro and in vivo analyses.

To investigate the mechanism of chloroplast protein import using the powerful in vivo and genetic approaches available for C. reinhardtii, we have made deletions throughout the TPs of two different chloroplast proteins: PC, a lumen-localized protein, and the -subunit of chloroplast ATPase, an extrinsic thylakoid membrane protein. The effects of these mutations have been examined in vivo, by transforming the mutant genes into C. reinhardtii, and in vitro, by importing mutant precursors into isolated C. reinhardtii chloroplasts. For PC, the in vivo and in vitro results agreed in most cases. However, deletions generally affected in vitro import more significantly than in vivo protein accumulation. An exception was a deletion that removed residues near the N terminus of the TP, which severely impaired in vivo protein accumulation, but affected in vitro import only moderately (Lawrence and Kindle, 1997).

We report here the disparate effects of TP mutations on in vivo accumulation and in vitro import of the -subunit of chloroplast ATPase. The chloroplast thylakoid ATPase uses the proton gradient generated by photosynthetic electron transport to synthesize ATP and therefore is required for photoautotrophic growth. Chloroplast ATPase has two discrete parts: CF₀, which is an integral thylakoid membrane protein complex composed of four polypeptides that carry out proton translocation, and CF₁, which has five polypeptide substituents with catalytic and regulatory activity. Three of the CF₁ subunits (, , and ) are encoded by the chloroplast genome and two ( and ) are encoded in the nucleus (for reviews, see McCarty and Carmeli, 1982; Futai et al., 1989; Junge, 1989). CF₁- is an extrinsic, stroma-facing thylakoid membrane protein that is encoded by the nuclear AtpC gene. C. reinhardtii cDNA (Yu et al., 1988) and genomic (Smart and Selman, 1991b) clones have been isolated and sequenced. C. reinhardtii CF₁- is synthesized as a precursor and can be imported into pea (Pisum sativum L.) chloroplasts (Yu et al., 1988). A synthetic peptide identical to the 35-amino acid residue CF₁- TP competes with pea SSU for import into pea chloroplasts and is presumed to have a stroma-targeting function (Theg and Geske, 1992). Additional information for assembly of CF₁- into the ATPase complex presumably resides within the mature protein.

We have made a series of deletions in the TP-coding region of AtpC and tested their ability to complement the ATPase deficiency of an atpC1 mutant (Smart and Selman, 1991a). CF₁- precursors containing the same deletions have been synthesized and tested in an in vitro C. reinhardtii chloroplast protein-import assay. As described below, the consequences of these mutations were very different in these two assays. By expressing chimeric genes in C. reinhardtii transformants, we demonstrate that the CF₁- TP has an in vivo chloroplast-targeting function, which is compromised by deletions identical to ones that have no discernible effect on CF₁- accumulation. We discuss the implications of these results and speculate that chloroplast protein import may not become the rate-limiting step for ATPase accumulation until it is impaired beyond even our most severe mutations.

	MATERIALS AND METHODS

Top Abstract Introduction Methods Results Discussion References

Chlamydomonas reinhardtii Strains

7

Plasmid DNAs and Nuclear Transformation

View larger version (31K): [in this window] [in a new window]   Figure 1. AtpC gene structure and locations of introduced restriction sites. A schematic representation of the SalI-NcoI AtpC genomic DNA fragment used for transformation is shown at the top, with the sequence between an EcoRI site introduced 55 bp upstream of the initiation codon and the end of the TP-coding region shown below. Open boxes represent the five exons (I-V). The site of transcription initiation is indicated by a horizontal arrow (Yu et al., 1988; Quinn and Merchant, 1995). Stromal processing occurs at the end of the TP, as indicated by the vertical arrow (Yu and Selman, 1988). The initiation codon (ini) is shown in bold. The sites of introduced restriction sites are underlined in the wild-type sequence, with the nucleotide changes shown above. The sequence of the PstI site used for the fusion to the ble-coding region is shown in italics.

PetE (PC) genes in which the stroma-targeting domain of the bipartite TP was replaced by the CF₁- TP were constructed as follows. Similarly to the strategy described above, an EcoRI site was introduced 28 bp upstream of the initiation codon, and NheI sites were engineered into Arg-Ser-coding sites in the PC TP (K.L. Kindle, unpublished data). NheI sites e and f lie on either side of the putative stroma-processing site between the predicted stroma-targeting and lumen-translocation domains. A series of chimeric PetE genes that were expressed from the PetE promoter was created by replacing EcoRI-NheI fragments from pPC-Nhe-e and pPC-Nhe-f with the corresponding fragments from p Nhe-c and p Nhe-d. A series of chimeric PetE genes driven by the AtpC promoter was created by replacing the KpnI-NheI fragments in the pPC constructs with the corresponding fragments from the p plasmids. The names for these constructs include the fusion sites in the CF₁- and PC TPs. For example, 26-PC20 contains 26 N-terminal amino acid residues from the CF₁- TP fused to the PC TP at amino acid residue 20.

To create chimeric genes in which the CF₁- TP-coding region was fused to that of the BRP (encoded by the ble gene), a PstI site was introduced immediately upstream of the BRP initiation codon in pSP109 (V. Lumbreras and S. Purton, personal communication) by site-directed mutagenesis using the oligonucleotide 5-CACTCAACATCTGCAGATGGCCAAGCTG. The BamHI-PstI fragment of pEco was cloned upstream of this PstI site to create p-ble, which encodes a chimeric precursor protein in which 32 amino acid residues from the CF₁- TP are fused to the BRP immediately adjacent to the initiation codon. Deletion derivatives were created by replacing the KpnI-PstI fragment of p-ble with the corresponding fragments from p-7-26, p-7-29, and p-27-29.

Mutant and wild-type AtpC genes were cotransformed into the nuclear genome of strain CC-3022 using the glass bead method (Kindle, 1990), with pMN24 DNA (Fernández et al., 1989), which contains the Nit1 gene, for selection. Chimeric PetE genes were cotransformed into strain AP6, using BamHI-digested pArg7.8 DNA for selection (Debuchy et al., 1989; K.L. Kindle, unpublished data). KpnI-digested pSP109, p-ble, or p-ble TP-deletion derivatives were transformed into nit1-305cw15 or CC-3396, and phleomycin-resistant transformants were selected on SGII-NH₄ or high-salt acetate plates with 2 to 5 µg/mL phleomycin (Stevens et al., 1996).

To produce CF₁- precursors for in vitro import assays, the AtpC cDNA insert of pTZ18R-CF₁-1A (Yu et al., 1988) was cloned into the SP6 transcription vector pGEM7f⁺ (Promega) to produce pcD1. The wild-type and deleted TP-coding regions were cloned from the genomic context into pcD1 in three steps. First, a 310-EcoRI-EagI fragment from pcD1, which includes the PstI site in exon I, was subcloned into BluescriptII KS to produce pcD1-EE. Second, the EcoRI-PstI fragment from pcD1-EE was replaced by EcoRI-PstI fragments from p-Eco, p-27-29, p-7-26, and p-7-29. Third, the EcoRI-EagI fragments from these plasmids, containing the wild-type and deleted genomic TP-coding regions, were then cloned back into EcoRI+Eag-digested pcD1.

In Vitro Import Assay

Chloroplast Isolation

DNA and Protein Blots

For immunoblot analysis, cells from 1.5-mL cultures were harvested by centrifugation and saved as pellets at 80°C until use. Cells were resuspended in SDS sample buffer and fractionated in 12% SDS-polyacrylamide gels (Stern et al., 1991). After equilibration in transfer buffer, proteins were transferred to nitrocellulose membranes using a Bio-Rad semidry electroblotter. Blots were blocked, incubated with antisera, and washed as previously described (Stern et al., 1991). Crude antisera raised against the indicated proteins were used at the following dilutions: C. reinhardtii CF₁- (from B. Selman), 1:1,000; spinach CF₁- (from R. McCarty, Johns Hopkins University, Baltimore, MD), 1:100,000; C. reinhardtii PC (from S. Merchant, University of California, Los Angeles), 1:1,000; Streptoalloteichus hindustanus bleomycin-binding protein (from Cayla Sarl, Toulouse, France), 1:500; and C. reinhardtii Oee2, the 23-kD protein of the oxygen-evolving complex (from F.-A. Wollman, Institut de Biologie Physico-Chimique, Paris), 1:10,000. The identity of the CF₁- band was confirmed on blots not shown by using two antisera raised against spinach CF₁- (from A. Jagendorf and B. Baird, Cornell University, Ithaca, NY, and from R. McCarty and K.-H. Suess, Gatersleben, Germany). For detection via enhanced chemiluminescence, blots were incubated with Promega anti-rabbit IgG-conjugated horseradish peroxidase (1:2,000), washed, and treated with luminol and H₂O₂ (Durrant, 1990). X-ray film exposures were for 5 s to 2 min. To detect PC species, acetone-precipitated proteins were prepared and assayed by immunoblots as described by K.L. Kindle (unpublished data).

	RESULTS

Top Abstract Introduction Methods Results Discussion References

AtpC Genes That Have Large Deletions in the TP-Coding Region Complement the Leaky, Nonphotosynthetic Phenotype of atpC1

As a transformation recipient, we used CC-3022 (nit1atpC1), which was constructed by Smart and Selman (1991a) by performing nuclear transformation in the presence of herring-sperm DNA. This apparently caused an insertion event that resulted in the arsenate-resistant, ATPase-deficient phenotype that exhibited leaky, nonphotosynthetic growth. For this reason and because we anticipated that TP mutations might impair import and accumulation of CF₁-, mutant and wild-type AtpC genes were introduced into the nuclear genome by cotransformation with the Nit1 gene, which allows transformants to be selected on acetate-containing nitrate plates (SGII-NO₃).

In all transformations that included the wild-type AtpC gene or controls in which restriction sites were introduced upstream of or within the CF₁--coding region, both large and small Nit⁺ transformant colonies were recovered on selective plates. In the minus DNA control, no Nit⁺ colonies were recovered, whereas with the Nit1 gene alone, all of the Nit⁺ colonies were small. It seemed likely that the large Nit⁺ colonies were transformants in which the atpC mutation had been complemented. All of the large colonies accumulated wild-type levels of chloroplast ATPase, whereas small colonies accumulated the same low level of ATPase seen in the parental strain (data not shown). This suggested an easy way to identify cotransformants that expressed a functional AtpC gene.

Surprisingly, CC-3022 cells that were cotransformed with AtpC genes containing TP-coding region deletions (p-27-29, p-7-26, and p-7-29) also yielded small and large Nit⁺ colonies. Whole-cell proteins were prepared from large Nit⁺ colonies and analyzed for accumulation of chloroplast ATPase subunits. Figure 2 shows the immunoblot results for a representative pair of transformants from each mutant construct. Duplicate blots were reacted with antisera raised against spinach CF₁- and C. reinhardtii CF₁-. The CF₁- antiserum reacted with both mitochondrial F₁-, an indication of the amount of whole-cell protein loaded into each well, and chloroplast CF₁-, which accumulates in proportion to the total amount of chloroplast F₁. Figure 2 shows that all transformants accumulated significantly more CF₁- and CF₁- than the atpC1 recipient (CC-3022). Furthermore, the levels of CF₁- and CF₁- were similar in all transformants. The CF₁- protein that accumulated in 7-29 transformants migrated slightly behind wild-type CF₁-, suggesting that the precursors in these transformants were processed aberrantly or not at all.

View larger version (54K): [in this window] [in a new window]   Figure 2. Nit+ colonies cotransformed with AtpC genes carrying TP-deletion mutations accumulate wild-type levels of ATPase subunits. A, Immunoblot. Whole-cell proteins were isolated from the atpC1 transformation recipient transformed only with Nit1 (atpC1) or robust Nit+ transformants that were cotransformed with the wild-type AtpC gene (WT) or AtpC genes with TP deletions (). Proteins were fractionated in a 12% SDS-polyacrylamide gel and transferred to nitrocellulose, and immunoblot analysis was performed as described in ``Materials and Methods''. The upper blot was incubated with antiserum raised against spinach CF1-; both mitochondrial (mt-) and chloroplast (cp-) species were detected. The lower blot shows only the CF1- region of the blot that was incubated with antiserum raised against C. reinhardtii CF1- (cp-). B, Sequences of wild-type and deleted TPs; the extent and location of the deleted amino acid residues are indicated with a line.

Transformants Contain an Ectopic Copy of the Introduced AtpC Gene

View larger version (82K): [in this window] [in a new window]   Figure 3. Transformants that accumulate CF1- contain an ectopic copy of the introduced AtpC gene. Whole-cell DNA was prepared from the CC-3022 strain, phenotypic revertants (rev.), the wild-type P17 strain (WT), or robust Nit+ transformants that had been cotransformed with AtpC deletion constructs (7-26 and 7-29). C. reinhardtii DNA or a plasmid containing an introduced EcoRI site in the SalI-NcoI fragment (p-Eco) was digested with BglII plus EcoRI (lanes 1) or BglII only (lanes 2). The DNAs were separated in an agarose gel, transferred to a nylon filter, and hybridized with the BglII fragment, which is indicated as a heavy line between the restriction maps. The EcoRI site in the introduced DNA is indicated on the upper map; symbols are as defined for Figure 1. The asterisk indicates that the TP deletions are in exon I. The horizontal bracket below the lower map indicates that a DNA rearrangement has occurred in the PstI-BclI fragment of CC-3022.

It was previously reported that atpC1 is a stable, null mutation (Smart and Selman, 1991a). However, in our hands strain CC-3022 grew slowly on medium lacking acetate, and colonies with a more robust photoautotrophic phenotype appeared at a low frequency (10⁷). To further characterize the recipient strain and to determine the amount of CF₁- that is required for a photoautotrophic phenotype, whole-cell proteins from the mutant and revertant strains were analyzed for accumulation of CF₁. Figure 4 shows that two photosynthetic revertants of CC-3022 accumulated about 10% of the wild-type level of CF₁- and CF₁-; this is about twice the amount of CF₁- observed in CC-3022. CF₁- abundance was below the level of detection in CC-3022. We wondered whether there had been an excision of part or all of the introduced DNA that accounted for the enhanced level of protein accumulation in the phenotypic revertants. However, as shown in Figure 3 (and data not shown), the BglII fragment characteristic of CC-3022 is still present in the revertant strains. It is interesting that all of the tested revertants contained a higher-molecular-weight fragment that hybridized weakly with the AtpC probe.

View larger version (38K): [in this window] [in a new window]   Figure 4. Phenotypic revertants accumulate low levels of CF1-. Whole-cell proteins from CC-3022, two photosynthetic revertants of CC-3022 (rev. 1 and 2), and the P17 wild-type strain were subjected to immunoblot analysis, as described in the legend to Figure 2. In the wild-type lanes, proteins loaded were equivalent to 1, 2, 5, 10, 25, and 100% of the amount loaded into the first three lanes.

In Vitro Import Is Undetectable for Precursors with TP Deletions 7-26 or 7-29

We recently established a homologous C. reinhardtii chloroplast protein-import assay (Goldschmidt-Clermont et al., 1989) and demonstrated that import is time and energy dependent (Lawrence and Kindle, 1997). As shown in Figure 5, when wild-type or 27-29 CF₁- precursor was added to the assay, a lower-molecular-weight, protease-protected species, which presumably represents imported and processed CF₁-, was detected after 10 min. It is interesting that a higher-molecular-weight, protease-resistant species was also detected in the assay with the 27-29 precursor, suggesting that import or processing may be partially impaired by this deletion. In contrast, no processed or protease-protected CF₁- precursor was detected for the 7-26 or 7-29 precursor after 15 min, a time when import of wild-type precursor had nearly reached a maximum. This suggests strongly that these deleted TPs do not allow significant precursor import in vitro. However, it should be noted that this in vitro import assay was capable of detecting only about 5% of the wild-type level of PC import (Lawrence and Kindle, 1997); therefore, inefficient import of CF₁- might not have been detected.

View larger version (20K): [in this window] [in a new window]   Figure 5. Deletions in the CF1- TP impair in vitro import. 35S-labeled precursors containing the wild-type CF1- TP (wt AtpC), or TP deletions between the designated amino acid residues were synthesized and incubated for the indicated times (in minutes) with isolated C. reinhardtii chloroplasts in the presence of 10 mm ATP, as described in ``Materials and Methods''. Following import, chloroplasts were treated with thermolysin and repurified, and proteins were separated in a 12.5% SDS-polyacrylamide gel. P, five percent of the in vitro translated precursor that was used for the import assays.

The CF₁- TP Is Sufficient for in Vivo Stroma Targeting

View larger version (41K): [in this window] [in a new window]   Figure 6. A, Diagram of the genomic AtpC and PetE genes; coding regions are shown as open and closed boxes, respectively. B, Maps of constructs that use either the PetE promoter (top) or the AtpC promoter (bottom) to drive expression of genes that contain a chimeric TP-coding region linked to the PC-coding region from PetE. C, The sequence of the wild-type PC TP. The two sites used for fusion to the CF1- TP are shown in bold (labeled below as Nhe-e and Nhe-f), and the postulated location of the stromal protease-processing site is indicated by an arrowhead. D, Names and amino acid sequences of the four chimeric TPs; the fusion sites are shown in bold. E, Accumulation of PC in Arg+ transformants. Whole-cell protein extracts were isolated from the transformation recipient (AP6), from an Arg+ transformant, and from Arg+ transformants that expressed the indicated PetE genes (the promoter is indicated in parentheses). Proteins were separated in a 12 to 18% gradient SDS-polyacrylamide gel and subjected to immunoblot analysis with antiserum raised against C. reinhardtii PC. 1:25, 1:5, and undiluted proteins from a transformant containing the wild-type (WT) PetE gene are shown in the three right lanes.

Figure 6, B and D, shows the maps and sequences of chimeric TPs fused to the PC-coding region, with gene expression driven by either the AtpC or PetE promoter. These constructs were cotransformed into strain AP6 (arg7pcy1-1 mt⁺), which carries two mutations in the PC-coding region that cause a frameshift resulting in premature translation termination (Quinn et al., 1993); the Arg7 gene was used for selection. Twenty Arg⁺ transformants from each construct were screened for PC accumulation by immunoblot analysis. The immunoblot in Figure 6E illustrates the highest level of PC accumulation observed among transformants generated with each of the eight constructs. In transformants carrying constructs in which the CF₁- TP-coding region was fused to the PC TP N-terminal of the putative stroma-processing site, PC accumulated to nearly the same level as in wild-type cells.

Mature PC also accumulated in transformants expressing the 26-30PC construct, in which the precursor contained 26 amino acid residues from the CF₁- TP-fused C-terminal of the processing site, at position 30 in the PC TP. There was no significant difference in PC accumulation, whether the constructs were expressed from the PetE or AtpC promoter. Since PC did not accumulate to a significant level in PC TP mutants lacking N-terminal amino acids (2-8 and 2-18 in Fig. 6), the stroma-targeting function in the chimeric constructs was presumably provided by the CF₁- TP. In contrast, transformants that expressed the fusion of the longer CF₁- TP region to the C-terminal site in the PC TP (29-30PC) accumulated barely detectable levels of PC. Moreover, this species migrated slightly more slowly than wild-type, mature PC, suggesting that it was aberrantly processed. These results indicate that the AtpC TP can have an in vivo stroma-targeting function, but that the functionality of the chimeric TP depends on the details of the fusion.

Deletions in the CF₁- TP Impair Its in Vivo Stroma-Targeting Function

View this table: [in this window] [in a new window]   Table I. Recovery of phleomycin-resistant transformants Glass bead transformation was carried out with 2 µg of the indicated DNA and 4 × 107 cells (Kindle, 1990). Cells grew in acetate-containing liquid medium for 18 h following transformation, and cultures were split in half and plated on high-salt acetate (CC-3396) or SGII-NH4 (nit1-305cw15) plates containing 2 or 5 µg/mL phleomycin (Phl).

To determine whether the wild-type or mutant TPs caused the BRP to be associated with chloroplasts, plastids were isolated from transformants that expressed the protein at a high level in preliminary experiments, and immunoreactive proteins were compared with those from whole cells by performing immunoblots. All transformants accumulated an immunoreactive protein with an electrophoretic mobility consistent with the predicted molecular mass of the unprocessed BRP precursor (between 14 and 17.5 kD, depending on the TP, indicated by asterisks in Fig. 7), although the amount was somewhat less for the 7-26 and 7-29 transformants than for the wild-type and 27-29 constructs. In the case of the wild-type CF1- TP, approximately 25% of the total fusion protein was associated with the chloroplast fraction, which was enriched in a more rapidly migrating polypeptide that may represent a processed species (indicated by a bullet in Figure 7).

View larger version (72K): [in this window] [in a new window]   Figure 7. Deletions in the CF1- TP prevent the accumulation of chloroplast-localized BRPs. The sequences of fusions between wild-type (wt) or deleted CF1- TPs and the BRP are shown at the top, with the TP sequence in plain text and the N-terminal sequence of the BRP in italics. Transformants expressing the various fusion proteins were analyzed in immunoblots. Proteins from whole cells (WC) or chloroplasts (Cp) isolated as described in ``Materials and Methods'' were denatured in sample buffer and fractionated in a 15% SDS-polyacrylamide gel, transferred to nitrocellulose, and reacted with antiserum raised against BRP or Oee2, the 23-kD polypeptide of the oxygen-evolving complex, which is localized in the chloroplast lumen. In the top and bottom panels, all lanes were exposed equally. The middle panel shows a shorter exposure of the wild-type lanes, and a longer exposure of the remaining lanes, to more clearly demonstrate the level of chloroplast-associated BRP. Putative chimeric precursors are indicated by asterisks, and processed species are indicated by bullets.

A small amount of the total 27-29 CF1--BRP was associated with the chloroplast fraction, where two distinct species were visible. The overexposed panel shows a trace amount of immunoreactive material in the chloroplast fraction of transformants expressing the 7-26 CF1--BRP, whereas none of the BRP lacking a TP or 7-29 CF1--BRP appeared to be chloroplast localized. These results are consistent with the suggestion that 32 amino acid residues from the CF1- TP are sufficient to direct a foreign protein into C. reinhardtii chloroplasts in vivo. Since cytosolic BRP fusion proteins accumulated in transformants containing chimeric genes with TP-coding region deletions, but little if any protein was chloroplast associated, the simplest explanation is that these TP deletions significantly impaired import of the fusion protein. However, we cannot eliminate the possibility that these fusion proteins were imported into the chloroplast but rapidly degraded.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

We have analyzed the effects of TP deletions on import of the extrinsic thylakoid membrane protein CF₁- in C. reinhardtii. Import was assessed in two ways: (a) in a homologous in vitro import assay and (b) by accumulation of the protein in transformants that express the mutant genes. Very different results were obtained with the two approaches. Precursors lacking two-thirds of the TP failed to be imported to a detectable level in vitro, but transformants that expressed the corresponding mutant AtpC genes in vivo accumulated wild-type levels of ATPase subunits.

We were initially concerned that photosynthetic colonies might have arisen from reversion of the original mutation or that wild-type AtpC genes might have been generated by homologous recombination (Sodeinde and Kindle, 1993). However, large (photosynthetic) Nit⁺ colonies were recovered only when AtpC constructs were included in the transformation experiments. Furthermore, analysis of genomic DNAs suggested that only ectopic integration events had occurred, since the endogenous AtpC locus was unaltered. Finally, accumulation of a higher-molecular-weight CF₁- species in the 7-29 transformants suggested that this species was the product of the introduced gene.

The CF₁- TP

When TP deletions that had no effect on in vivo CF₁- accumulation were introduced into the context of the BRP, they appeared to eliminate in vivo chloroplast protein import partially or nearly completely, since little or none of the fusion protein that accumulated was chloroplast associated. Although we cannot eliminate the possibility that fusion proteins with deletions in the TP are imported into the chloroplast and then rapidly degraded, this seems unlikely. Fusion proteins with TP deletions do accumulate in the cytosol; therefore, their lability would have to be chloroplast specific. Furthermore, the 27-29 species, which was detectable in the chloroplast fraction only at a low level, differed by only three amino acids from the wild-type fusion protein, which accumulated to a relatively high level in the chloroplast. The simplest interpretation is that the in vivo chloroplast-targeting function of the CF₁- TP is impaired by these deletions, in some cases severely. It is possible that redundant stroma-targeting information resides within the mature CF₁- protein and accounts for import and accumulation of CF₁- protein in vivo, but if so, this information is recognized only in vivo. We favor the interpretation that import of CF₁- is impaired even in vivo by the long TP deletions, but that relatively little import is sufficient to support wild-type levels of ATPase complex assembly.

The 7-29 TP deletion either prevents in vivo processing of the CF₁- precursor or causes aberrant processing, since a higher-molecular-weight CF₁- species accumulated in transformants carrying this mutation. Because of the large size of the deletion in 7-29, the precursor would be only 12 amino acid residues longer than the correctly processed mature protein. Apparently, the aberrant CF₁- species is assembled and functional in the ATPase complex, since the strain had a robust photosynthetic growth phenotype. Transformants expressing either the 7-26 or 27-29 precursors accumulated processed CF₁-. However, when the 27-29 CF₁- precursor was imported into isolated chloroplasts in vitro, a protease-protected precursor species was observed in addition to mature, processed CF₁-. Since the ratio of the two forms was constant during the import reaction, we speculate that the protease-protected precursor may represent a population that has entered a nonproductive import pathway and cannot be translocated across the inner envelope, or has adopted a conformation that is resistant to stromal processing. The aberrant processing of the 7-29 and 27-29 precursors in vivo and in vitro, respectively, is consistent with previously published studies in which mutations near the cleavage site affected the extent and/or fidelity of processing (Wasmann et al., 1988; Ostrem et al., 1989; Archer and Keegstra, 1993; Bassham et al., 1994; Pilon et al., 1995). A positively charged amino acid residue at position 4 relative to the cleavage site was required for processing of the soybean and pea light-harvesting chlorophyll a/b-binding protein precursors (Clark and Lamppa, 1991). In this respect, it is interesting to note that an Arg residue is present at 8 in the wild-type and 7-26 CF₁- precursors, whereas a Lys residue is present at position 10 in the 27-29 precursor. There are no positively charged residues in the 7-29 TP, which is probably unprocessed in vivo.

Accumulation of the Chloroplast ATPase

In this study the atpC1 recipient strain (CC-3022) had a leaky nonphotosynthetic growth phenotype and accumulated CF₁- to about 5% of the wild-type level. Although CF₁- was not detected in CC-3022 using three different antisera, these antisera also failed to detect 5% of the wild-type level of CF₁- (Fig. 4; data not shown). Although unassembled subunits of ATPase complexes do not generally accumulate, a C. reinhardtii atpA translation mutant (F54) accumulates some CF₁- in the absence of CF₁-, partially due to 3-fold increased synthesis of CF₁- in this context (Drapier et al., 1992). Since CF₁- and CF₁- accumulated in parallel in the phenotypic revertants described here, the simplest interpretation is that there is enough residual expression of AtpC in CC-3022 to allow the ATPase complex to accumulate to about 5% of the wild-type level. DNA-blot analysis was consistent with an alteration in the PstI-BclI fragment that contains exons II and III, introns I and II, and part of intron III. We presume that an insertion of herring-sperm carrier DNA occurred in one of the introns during the course of nuclear transformation and that splicing is therefore impaired. However, a more complex rearrangement is also possible, since insertional mutagenesis sometimes results in deletions at the insertion site (Tam and Lefebvre, 1993.) The nature of the phenotypic reversion is unknown, since the original DNA rearrangement in atpC1 was still present in both revertants tested. Perhaps a slight increase in splicing efficiency due to an extragenic suppressor mutation could account for the small (approximately 2-fold) increase in gene expression.

Strategies for Isolating Envelope Translocation Mutants

Preliminary results with AtpC-ble fusion genes and similar constructs with the PC TP (data not shown) suggest that BRP function is not impaired by N-terminally fused oligopeptides, since phleomycin-resistant transformants could be selected in most cases. Because the transformants characterized in this work were selected for phleomycin resistance, it is not surprising that some of the fusion protein remains cytosolic. The amount of cytosolic BRP may determine the level of phleomycin resistance, and this is likely to be a consequence of the level of gene expression as well as the efficiency of chloroplast import. The recovery of fewer phleomycin-resistant transformants with the wild-type TP fusions than with most TP-deletion constructs may indicate that a higher level of gene expression is required to recover phleomycin-resistant transformants carrying the wild-type construct, since the fusion protein can be imported into the chloroplast. We are screening for phleomycin-sensitive transformants that express the ble gene and accumulate processed BRP in the chloroplast, which will be used as the starting strains for selecting envelope translocation mutants.

Concluding Remarks

	FOOTNOTES

   Received August 25, 1997; accepted November 21, 1997.

	ABBREVIATIONS

Abbreviations: BRP, bleomycin-resistance protein. CF₁- and CF₁-, - and -subunits of chloroplast ATPase-coupling factor 1. PC, plastocyanin. SSU, small subunit of Rubisco. TP, transit peptide.

	ACKNOWLEDGMENTS

We are thankful to Bruce Selman for providing the atpC1 mutant, the genomic and cDNA clones, and the C. reinhardtii CF₁- antiserum used in this work. Antisera were also generously contributed by Sabeeha Merchant, Francis-André Wollman, Dick McCarty, and André Jagendorf. We thank Ken Cline for advice concerning the in vitro chloroplast protein-import assay and David Stern and Clare Simpson for comments about the manuscript.

	LITERATURE  CITED

Top Abstract Introduction Methods Results Discussion References

Archer EK, Keegstra K (1993) Analysis of chloroplast transit peptide function using mutations in the carboxyl-terminal region. Plant Mol Biol 23: 1105-1115 [CrossRef][Medline]

Bassham DC, Creighton AM, Karnauchov I, Herrmann RG, Klösgen RB, Robinson C (1994) Mutations at the stromal processing peptidase cleavage site of a thylakoid lumen protein precursor affect the rate of processing, but not the fidelity. J Biol Chem 269: 16062-16066