Canonical Signal Recognition Particle Components Can Be Bypassed for Posttranslational Protein Targeting in Chloroplasts

Abstract

The chloroplast signal recognition particle (cpSRP) and its receptor (cpFtsY) target proteins both cotranslationally and posttranslationally to the thylakoids. This dual function enables cpSRP to utilize its posttranslational activities for targeting a family of nucleus-encoded light-harvesting chlorophyll binding proteins (LHCPs), the most abundant membrane proteins in plants. Previous in vitro experiments indicated an absolute requirement for all cpSRP pathway soluble components. In agreement, a cpFtsY mutant in Arabidopsis thaliana exhibits a severe chlorotic phenotype resulting from a massive loss of LHCPs. Surprisingly, a double mutant, cpftsy cpsrp54, recovers to a great extent from the chlorotic cpftsy phenotype. This establishes that in plants, a new alternative pathway exists that can bypass cpSRP posttranslational targeting activities. Using a mutant form of cpSRP43 that is unable to assemble with cpSRP54, we complemented the cpSRP43-deficient mutant and found that this subunit is required for the alternative pathway. Along with the ability of cpSRP43 alone to bind the ALBINO3 translocase required for LHCP integration, our results indicate that cpSRP43 has developed features to function independently of cpSRP54/cpFtsY in targeting LHCPs to the thylakoid membranes.

INTRODUCTION

Thylakoids are the internal chloroplast membranes and contain essential proteins for photoautotrophic growth. A vast majority of thylakoid proteins are encoded by the nuclear genome, synthesized in the cytoplasm, and translocated by a two-step process. In the first step, precursors directed by a cleavable N-terminal targeting domain are imported into the chloroplast across the envelope membranes by a unique translocon system (Kessler and Schnell, 2006). During or soon after import, the chloroplast-targeting domain is cleaved by a stroma-localized processing protease. In the second step, processed pathway intermediates in the stroma enter one of four distinct thylakoid localization pathways, which appear substrate-specific in studies that reconstitute localization in vitro (for review, see Keegstra and Kline, 1999). The polypeptides that reside in the thylakoid lumen are mainly targeted by the chloroplast Sec (cpSec) and chloroplast TAT (cpTAT) pathways. Integral membrane proteins rely mainly on the chloroplast signal recognition particle (cpSRP) pathway or use a mechanism that lacks any known protein components, termed the spontaneous pathway.

The only known nucleus-encoded proteins to utilize the cpSRP pathway are the light-harvesting chlorophyll binding proteins (LHCPs). They constitute up to 50% of the protein bulk in thylakoids (Green and Salter, 1996) and therefore are considered the most abundant membrane proteins on earth. General steps of the LHCP targeting mechanism appear conserved with targeting by SRPs that function in the cytosol of prokaryotes and eukaryotes.

Cytosolic SRPs bind hydrophobic signal sequences at the N terminus of newly synthesized polypeptides as they emerge from the ribosome (for review, see Pool, 2003; Luirink et al., 2005). SRP then directs the bound ribosome/nascent polypeptide chain complex to a SRP receptor that functions at the endoplasmic reticulum or the cytoplasmic membrane in bacteria. GTP binding by SRP and its receptor coordinate the release of the nascent polypeptide/ribosome complex to a Sec translocase, while GTP hydrolysis promotes the release of SRP from its receptor. A distinct functional difference in LHCP targeting by cpSRP stems from the absence of a ribosome. A cpSRP-LHCP complex termed the transit complex assembles in the stroma and represents the targeted form of LHCP (Li et al., 1995; Delille et al., 2000). Targeting culminates at the thylakoid owing to interactions between cpSRP, cpFtsY, and ALBINO3 (ALB3) (Moore et al., 2003), a thylakoid insertase required for LHCP integration into the lipid bilayer (Moore et al., 2000, 2003; Bellafiore et al., 2002).

The unique posttranslational substrate binding function of cpSRP stems from its peculiar subunit composition (for review, see Schuenemann, 2004). Cytosolic SRPs are minimally composed of an RNA moiety in both prokaryotes and eukaryotes bound to a 54-kD GTPase (SRP54; Ffh in Escherichia coli) that functions in binding hydrophobic signal sequences as well as the SRP receptor, SRα, or its homolog FtsY in E. coli. cpSRP contains a conserved 54-kD subunit (cpSRP54) but lacks an RNA moiety, reflected by the absence of an ancestral RNA binding site in cpSRP54 (Groves et al., 2001; Rosenblad and Samuelsson, 2004). The novel posttranslational binding activity of cpSRP is partially based on a plant-specific 43-kD subunit (cpSRP43; Schuenemann et al., 1998) that forms a heterodimer with cpSRP54 (Groves et al., 2001). Whereas cpSRP54 in the transit complex is thought to bind the C-terminal transmembrane span in LHCPs (High et al., 1997), cpSRP43 binds an internal 18–amino acid polar motif termed L18 that is conserved among LHCPs (Delille et al., 2000; Tu et al., 2000).

Despite the lack of an RNA moiety, cpSRP also exhibits a classical cotranslational function important to the biogenesis of chloroplast-encoded proteins. Biochemical experiments demonstrated cpSRP54 binding to ribosome-associated D1, a component of the photosystem II (PSII) core complex (Nilsson and van Wijk, 2002; Zhang and Aro, 2002), which is found at reduced levels along with other chloroplast-encoded reaction center proteins in cpSRP54-deficient Arabidopsis thaliana (Pilgrim et al., 1998). The duality in cpSRP function is consistent with the identification of two cpSRP54 pools in stroma, one bound to cpSRP43 and the other bound to ribosomes. cpSRP43 appears absent in the ribosome-bound pool of cpSRP54 (Franklin and Hoffman, 1993). In this context, deletion mutants of cpSRP54 (ffc) and cpSRP43 (chaos) also support a posttranslational and cotranslational function of cpSRP; the absence of cpSRP43 affects the accumulation of LHCPs, whereas the loss of cpSRP54 alters the accumulation of both LHCPs and chloroplast-synthesized proteins (Pilgrim et al., 1998; Amin et al., 1999; Klimyuk et al., 1999). Comparison of the strongly chlorotic chaos ffc double mutant with mutants lacking either cpSRP43 or cpSRP54 demonstrated that each subunit contributes to LHCP localization in an independent and additive manner (Hutin et al., 2002).

Discovery of the chloroplast SRP receptor homolog cpFtsY led to the demonstration that, along with GTP and cpSRP, LHCP integration into isolated thylakoids requires cpFtsY (Kogata et al., 1999; Tu et al., 1999). A cpFtsY-cpSRP54 complex, stabilized at the thylakoid membrane by nonhydrolyzable GTP, associates with ALB3, indicating that neither substrates nor cpSRP43 are required for cpSRP interaction with ALB3 (Moore et al., 2003).

In maize (Zea mays) mutants lacking cpFtsY, there is an accumulation of the cpSRP-LHCP transit complex in the stroma, although the extent to which different LHCPs rely on cpFtsY varies among LHCP family members (Asakura et al., 2004). Nevertheless, maize cpFtsY mutants also exhibited a seedling-lethal phenotype, as plants died 2 weeks after germination (Asakura et al., 2004). Such a drastic phenotype, compared with the less severe mutant phenotypes deficient for cpSRP43 or cpSRP54, supports a central role for cpFtsY in targeting LHCPs and in thylakoid biogenesis.

In Arabidopsis, a viable cpftsy mutant was recently identified (Durett et al., 2006), but the repercussions of this mutation on the cpSRP pathway were not studied. In this article, we identified a second mutant allele of Arabidopsis cpftsy. The phenotype of this cpftsy mutant was compared with the phenotype of Arabidopsis mutants affected in the other cpSRP components. Importantly, the viability of the cpftsy mutant made it possible to combine mutations, which led to the discovery that removal of both cpSRP54 and cpFtsY relieves the severity of the cpftsy mutant phenotype. These data demonstrate that in the absence of conserved SRP targeting components, cpSRP54 and cpFtsY, a second targeting mechanism accommodates LHCP localization to the thylakoids and most likely relies on cpSRP43 and its ability to bind both LHCP and ALB3.

RESULTS

cpFtsY Is Not Required for Arabidopsis Viability

An Arabidopsis mutant defective for cpFtsY production was identified by protein gel blot screening of an ethyl methanesulfonate (EMS)–mutagenized collection using a polyclonal antibody directed against cpFtsY (Figures 1A and 1B). The cpftsy gene present in the mutant was amplified by PCR and sequenced. This revealed a C-to-A transversion in exon 6, creating a stop codon that truncated half of the protein (Figure 1C). The absence of a cpFtsY fragment in the mutant suggests that the identified allele was most probably a null mutant. The mutant exhibited a strong chlorotic phenotype, which remained linked with the cpFtsY mutation after three backcrosses that were required to remove another albino mutation present in this line. In order to confirm that the observed phenotype of the backcrossed line resulted only from lack of cpFtsY, the mutant was successfully complemented with cpFtsY cDNA fused to a modified cauliflower mosaic virus 35S promoter from plasmid pKYLX71-35S² (Figures 1A and 1B). It is noteworthy that transformants overexpressing the cpFtsY cDNA accumulated normal levels of cpFtsY protein compared with the wild type (data not shown), which could stem from posttranscriptional control of the cpFtsY level. Despite its marked chlorosis, the Arabidopsis cpftsy mutant is fully viable and can easily be grown in vitro on agarose or on soil.

Figure 1.

The Arabidopsis cpftsy Mutant Exhibits a Severe Chlorotic Phenotype That Is Complemented by the Expression of a cpFtsY Transgene.

(A) Phenotypic comparison of the wild type, the cpftsy mutant, and the cpftsy mutant transformed with a cpFtsY expression construct (cpftsy/35ScpFtsY) is shown for plants grown to the rosette stage.

(B) Total protein extracts from each plant shown in (A) were separated by SDS-PAGE, blotted, and probed with antibody directed against cpFtsY.

(C) EMS mutation in the CS3171 line introduced a stop codon in the coding region of the cpFtsY gene.

The Arabidopsis cpftsy Mutant Exhibits a Very Similar Phenotype to the ffc chaos Double Mutant

The very strong chlorosis exhibited by the cpftsy mutant (Figure 2A) was due to the loss of ∼80% of the photosynthetic pigments (Figure 2B). More specifically, there was a 67% reduction of xanthophylls, an 80% reduction of β-carotene, and an 84% reduction of chlorophylls (Figure 2B). This pigment decrease did not alter the chlorophyll a/b ratio, which was ∼2.4 in the cpftsy mutant and the wild-type control. Electron micrographs of cpftsy mutant chloroplasts revealed important alterations in the ultrastructure of this organelle, with a drastic reduction in the amount of both appressed and unappressed thylakoid membranes (Figure 2C). Taken together, these traits are much like those of the ffc chaos (cpSRP54⁻/cpSRP43⁻) double mutation (Hutin et al., 2002), which exhibits a similar loss of photosynthetic pigments and alteration of chloroplast ultrastructure (Figure 2B). Also like the ffc chaos double mutant, the cpftsy mutation greatly reduced the accumulation of various LHCPs relative to the wild type. However, the extent of the reduction differs between the various LHCPs (Figure 3A). The two most abundant antennae of PSII (Lhcb1 and Lhcb2) have a similar size and could not be distinguished by the antibody used, which recognizes both proteins (LHCII in Figure 3A).

Figure 2.

The cpftsy Mutant and ffc chaos Double Mutant Exhibit Similar Severe Phenotypic and Ultrastructural Alterations.

(A) Wild-type, double mutant ffc chaos, and cpftsy mutant plants are presented at the rosette stage.

(B) Pigments were extracted from the plant lines shown in (A) and analyzed as described in Methods. Data are mean values of five measurements ± sd. The scale (ng/mm²) for minor carotenoids is shown at right, and the scale for chlorophylls is given at left. antera, antheraxanthin; βcar, β-carotene; chla, chlorophyll a; chlb, chlorophyll b; lut, lutein; neo, neoxanthin; vio, violaxanthin; zea, zeaxanthin.

(C) Chloroplast ultrastructure was examined by electron microscopy using leaf sections from the plant lines shown in (A). Representative electron micrographs are shown. Bar = 1 μm.

Figure 3.

Photosystem Protein Accumulation and Photosystem Activity of cpftsy and chaos ffc Mutants Indicate Similar but Distinct Functions of cpSRP and cpFtsY.

Total membrane proteins were extracted from leaves of wild-type and mutant (cpftsy and chaos ffc) plants as described in Methods and analyzed by protein gel blotting ([A] to [C]) using antibodies directed against the proteins indicated at right. Samples were loaded at three different concentrations to detect and avoid possible saturation of the immunodetection (quantification is indicated in Supplemental Table 1 online), and blotting was repeated at least three times to ensure reproducibility. A representative blot is shown for each protein detected.

(A) Antenna proteins of PSI and PSII (Lhca and Lhcb, respectively). The antibody dilution used (in parentheses) and the amount of membrane proteins loaded on each lane were as follows: Lhca1 (1:100), 3/5/7 μg; Lhca2 (1:200), 3/5/7 μg; Lhca3 (1:1000), 1.5/3/5 μg; Lhca4 (1:100), 5/7/10 μg; LHCII (1:1000), 1/2/4 μg; Lhcb3 (1:30), 1.5/3/5 μg; Lhcb4 (1:30), 1/3/5 μg; Lhcb5 (1:100), 0.75/1.5/3 μg; and Lhcb6 (1:750), 1/2/4 μg.

(B) PSI and PSII reaction center binding proteins. The antibody dilution used (in parentheses) and the amount of membrane proteins loaded on each lane were as follows: D1 (1:100), 2/4/6 μg; and PsaA/B (1:1000), 5/10/15 μg.

(C) Proteins targeted by thylakoid pathways distinct from cpSRP. Fibrilin C34 (Fibr.C34) is targeted by the ΔpH pathway; plastocyanin (PC) and the 33-kD subunit of the oxygen evolving complex (OE33) use the Sec pathway; and PsbS is targeted by the spontaneous insertion pathway. ClpC is a component of the chloroplast envelope translocase included for comparison. The antibody dilutions used (in parentheses) and the protein amounts loaded in each lane were as follow: ClpC (1:4000), 3/5/7 μg; fibrilin C34 (1:1000), 3/5/7 μg; PC (1:5000), 2/5/10 μg; OE33 (1:10,000), 0.5/1.25/3 μg; and PsbS (1:4000), 2/5/10 μg.

(D) and (E) Photochemistry. Chlorophyll fluorescence characteristics were measured from isolated leaves as described in Methods to compare maximum quantum yield of PSII (Fv/Fm) (D) and quantum yield of PSII-mediated electron transport (ΔF/Fm′) (E). Data are mean values of four separate experiments ± sd.

In addition to the LHCPs, the cpftsy mutation affected chloroplast-encoded proteins, which require the cpSRP54 subunit to reach their thylakoid location (Amin et al., 1999; Nilsson et al., 1999; Nilsson and van Wijk, 2002). These include the reaction center proteins of PSI (PsaA and PsaB) and PSII (D1). As observed in Figure 3B, both complexes are reduced in the cpftsy mutant compared with the wild-type control. This decrease in reaction center proteins also resembles closely the phenotype of the ffc chaos double mutant (Figure 3B).

As described previously for all of the cpSRP mutants, equal loading of total membrane proteins for immunoblot analyses resulted systematically in an overrepresentation of chloroplast envelope membrane proteins from mutant plants relative to the wild type, due to the lack of LHCP (main membrane proteins). Therefore, in order to quantify the loading differences, we used an antibody raised against the stromal protein ClpC, which associates with the chloroplast inner membrane translocon (Akita et al., 1997). As observed previously (Hutin et al., 2002) for ffc chaos, an excess of ClpC protein was also found for the cpftsy mutant (Figure 3C). Therefore, the differences in specific protein levels observed between mutants and the wild-type control in Figures 3A and 3B appear underrepresented, owing to the fact that a greater number of chloroplast membranes are present in the mutant-derived samples to achieve an equal amount of total membrane protein in mutant and wild type samples. The function of the various other thylakoid targeting pathways was investigated by immunodetection of proteins localized specifically by the spontaneous (PsbS), Sec (OE33 and plastocyanin [PC]), and cpTAT (fibrilin CDSP34) pathways. In all cases, with the exception of OE33, the proteins targeted by these various pathways were found in greater amounts in both mutants (Figure 3C). This confirmed that the mutation studied here does not inhibit these pathways. The higher levels observed are consistent with the observation that both mutant-derived samples contain more chloroplast proteins than the wild-type control (except of course for cpSRP-targeted proteins). The slight decrease in the level of OE33 protein, which is a subunit of the PSII oxygen-evolving complex, could be attributed to the decreased PSII levels observed in these plants.

Chlorophyll Fluorescence Characteristics Distinguish cpftsy from the ffc chaos Double Mutant

Previous analyses indicated that the ffc chaos double mutation abolishes the cpSRP pathway required for thylakoid targeting of LHCP in chloroplasts (Hutin et al., 2002). Similarity between the cpftsy mutant phenotype and the ffc chaos double mutant suggests that cpFtsY is also absolutely required for cpSRP function. Yet, chlorophyll fluorescence analyses point to distinct features between both mutants. The photochemical activity of PSII was analyzed in vivo using chlorophyll fluorescence measurements (Fv/Fm) in dark-adapted leaves. The Fv/Fm ratio is a measure of the maximal photochemical efficiency of PSII when all of the primary quinonic PSII electron acceptors are oxidized (Maxwell and Johnson, 2000). As found previously (Hutin et al., 2002), the ffc chaos double mutants did not differ significantly from the wild-type control, whereas this value was noticeably affected (28% decrease) in the cpftsy mutant (Figure 3D). This suggests that, in contrast with ffc chaos, the proper assembly and the photochemical competence of PSII reaction centers were affected in the cpftsy mutant. The actual quantum yield of PSII photochemistry (ΔF/Fm′) was measured at different photon flux densities in cpftsy and ffc chaos. ΔF/Fm′ is a measure of the efficiency of linear electron transport in the chloroplasts (Genty et al., 1989). At illumination levels of <250 μmol·m⁻²·s⁻¹, the quantum yield was more reduced in the cpftsy mutant relative to the ffc chaos double mutant, indicating that photosynthetic electron transport was more affected and was more rapidly saturated by light in the cpftsy mutant. There are many factors that could contribute to the reduction of the efficiency of photosynthetic electron transport in cpftsy, including the inhibition of PSII (indicated by the reduced Fv/Fm value) or a block of the electron transport chain after PSII (e.g., an inhibition of PSI).

Lack of cpSRP54 Abolishes the cpFtsY Effect on LHCP Targeting to Thylakoids

The Arabidopsis ffc mutant lacking cpSRP54 accumulates more LHCP (Amin et al., 1999) than the cpftsy mutant. This may indicate that cpSRP54 functions only to improve the efficiency of protein targeting by a mechanism that relies on cpFtsY even in the absence of cpSRP54. Alternatively, the requirement for cpFtsY may be bypassed in the absence of cpSRP54, owing to a need for cpFtsY only when targeting substrates are bound by cpSRP54. These hypotheses may be genetically tested by creating a ffc cpftsy double mutant lacking both cpSRP54 and cpFtsY. The resulting double mutant would exhibit a severe phenotype similar to that of the cpftsy mutant if cpSRP54 simply improves the efficiency of a cpFtsY-dependent routing mechanism. By contrast, the ffc cpftsy double mutant would exhibit a less severe ffc-like phenotype if cpFtsY function were dedicated to substrate targeting by cpSRP54. This scenario implicates the function of a second LHCP localization pathway that is independent of both cpSRP54 and cpFtsY and that operates only when LCHP targeting substrates fail to interact with cpSRP54.

The cpftsy and ffc mutants were crossed to produce F1 plants. As expected (both mutations are recessive), the resulting hybrids exhibited a wild-type phenotype. In the progeny of these F1 plants, we analyzed various individuals using antibodies raised against cpFtsY and cpSRP54. This allowed the identification of a ffc cpftsy double mutant that exhibited a ffc mutant phenotype (Figures 4A and 4B). The double mutant was also verified by backcrosses with ffc or cpftsy mutants. The resulting F1 progeny exhibited ffc or cpftsy phenotypes, respectively, as expected.

Figure 4.

The Double Mutant ffc cpftsy Exhibits an ffc Phenotype.

(A) ffc cpftsy and ffc mutants are shown at the rosette stage.

(B) Total leaf protein extracts from wild-type and mutant (ffc, cpftsy, and ffc cpftsy) plants were examined by protein gel blotting using antibody against cpSRP54 or cpFtsY as indicated.

(C) Pigments extracted from leaves of ffc cpftsy and ffc mutants were characterized as described in Methods. Data are mean values of four measurements ± sd. The concentrations (ng/mm) of carotenoids are indicated at left and those of chlorophylls are indicated at right. antera, antheraxanthin; βcar, β-carotene; chla, chlorophyll a; chlb, chlorophyll b; lut, lutein; neo, neoxanthin; vio, violaxanthin; zea, zeaxanthin.

(D) to (F) Relative levels of PSI and PSII antenna proteins (Lhca and Lhcb, respectively) (D), reaction center binding proteins (D1 and PsaA/B) (E), and proteins targeted by the ΔpH (fibrilin [Fibr.C34]), Sec (PC and OE33), and spontaneous insertion (PsbS) pathways (F) were examined by protein gel blotting as described in the legend to Figure 3.

Pigment analyses of the ffc cpftsy double mutant revealed that the levels of chlorophyll a and b, β-carotene, lutein, and neoxanthin were similar in the ffc and ffc cpftsy mutants. However, in the ffc cpftsy double mutant, violaxanthin was almost completely converted to zeaxanthin and antheraxanthin (Figure 4C), suggesting the presence of a light stress in this mutant (Demmig-Adams and Adams, 2000). Analysis of nucleus-encoded (LHCP; Figure 4D) and chloroplast-encoded (reaction centers PSI and PSII; Figure 4E) proteins targeted by cpSRP revealed no striking difference between ffc and ffc cpftsy mutants. This suggests that, for the proteins targeted by the cpSRP pathway, the ffc mutation is epistatic to the cpftsy mutation. Examination of proteins targeted by other thylakoid-targeting pathways did not reveal striking differences between ffc and ffc cpftsy, with the exception of plastocyanin, which accumulated to a higher level in the double mutant (Figure 4F). Taken together, these data are in agreement with a model in which cpFtsY activity requires the presence of cpSRP54.

In assays that reconstitute LHCP integration into isolated thylakoids, integration requires all cpSRP pathway soluble targeting components (cpSRP43, cpSRP54, and cpFtsY). Yet, much like the cpSRP54 mutant alone (ffc), eliminating both cpSRP54 and cpFtsY in vivo (ffc cpftsy) affected LHCP accumulation only moderately. This indicates the presence of an alternative pathway for LHCP targeting that bypasses these two proteins in plants when cpSRP54 is absent and unable to interact with LHCP to form a cpSRP-LHCP transit complex. Since a previously reported double mutant lacking both cpSRP54 and cpSRP43 (ffc chaos) accumulates a much lower level of LHCP than either the ffc mutant or the ffc cpftsy double mutant, it can be assumed that cpSRP43 is required for this alternative pathway. The cpSRP54-independent activity of cpSRP43 can be examined using a cpSRP43 mutant that lacks the ability to bind cpSRP54, since cpSRP54-dependent LHCP integration activity is lost in vitro when cpSRP54-cpSRP43 heterodimer formation is disrupted (Goforth et al., 2004).

cpSRP43-Deficient Mutants Can Be Fully Complemented by Expression of a Modified cpSRP43 Protein That Does Not Bind cpSRP54

Combined functional and interaction studies demonstrated that CD2, the second of three chromodomains present in cpSRP43 (amino acids 270 to 320), is required for cpSRP43 binding to cpSRP54 (Goforth et al., 2004). Arabidopsis cpSRP43 cDNA deleted of the CD2 domain (ΔCD2cpSRP43) driven by the cauliflower mosaic virus 35S promoter was introduced into the cpSRP43-null mutant (chaos). Several independent transformants were produced, and seven of them (named ΔC2cpSRP43-A to -G) containing only one copy of the construct were selected by DNA gel blot for further analysis (data not shown). As reported previously, the chaos mutant exhibited a pale green phenotype (Klimyuk et al., 1999), resulting from the loss of ∼45% of the photosynthetic pigments (Figure 5A). Interestingly, all of the chaos transformants complemented with ΔC2cpSRP43 exhibited increased pigmentation, reaching in most cases wild-type levels (Figure 5A). The ability of ΔC2cpSRP43 expression to complement the chaos mutant was independent of cpSRP heterodimer formation, as indicated by the results of coprecipitation assays conducted with anti-cpSRP43 antibody, which binds both native cpSRP43 and ΔC2cpSRP43 (Figure 5B). Although cpSRP54 was found to coprecipitate with cpSRP43 using soluble leaf extract from wild-type plants, cpSRP54 was absent when ΔC2cpSRP43 was immunoprecipitated from soluble leaf extract derived from chaos transformants expressing ΔC2cpSRP43 (Figure 5B). These results are consistent with sizing chromatography assays, which showed that cpSRP54 coelutes with cpSRP43 using wild-type plant extracts but elutes separately from ΔC2cpSRP43 using leaf extracts derived from the chaos transformants (data not shown). Together, these findings verify the results of in vitro studies showing that ΔC2cpSRP43 is unable to bind cpSRP54 (Goforth et al., 2004). Importantly, they also demonstrate that normal LHCP targeting can take place in the absence of cpSRP43/cpSRP54 dimer formation, adding further support for a cpSRP54/cpFtsY-independent function of cpSRP43.

Figure 5.

cpSRP43-cpSRP54 Heterodimer Formation Is Not Required for cpSRP43 to Complement chaos Mutants.

(A) chaos mutants lacking cpSRP43 were transformed with DNA coding for cpSRP43 that lacks the cpSRP54 binding domain. Pigment accumulation (chlorophylls a and b and carotenoids) of the chaos mutant and different T1 plants resulting from transformation of chaos (chaos ΔC2cpSRP43-A to -G) was analyzed as described in Methods and expressed as a percentage of wild-type values (ecotype Landsberg erecta).

(B) Soluble protein extract was prepared from leaves of the wild type corresponding to ffc (Col), ffc mutant, the wild type corresponding to chaos (Landsberg erecta), chaos, and two of the chaos transformants described for (A) (chaos ΔC2cpSRP43-F and -G) as indicated. The top and middle panels (input) show the level of cpSRP43 (closed arrowheads), ΔC2cpSRP43 (open arrowhead), or cpSRP54 in each extract detected by protein gel blotting with the antibodies (α) indicated at right. The bottom panel shows cpSRP43 (closed arrowhead) and ΔC2cpSRP43 (open arrowhead) copurified with cpSRP54 following immunoprecipitation with cpSRP54 antibody (IP α-cpSRP54).

cpSRP43 Interacts with ALB3 in the Absence of cpSRP54 and cpFtsY

Genetic data supporting a cpSRP54/cpFtsY-independent targeting pathway that relies on the function of cpSRP43 raises the possibility that cpSRP43 is able to interact directly with ALB3 in targeting LHCPs. To investigate this possibility, double immunolabeling experiments were performed in various mutant backgrounds to examine the colocalization of cpSRP43 and ALB3 (Figure 6A). A cryofixation protocol was employed to freeze the localization of proteins occurring in the cell, thereby avoiding vagaries more common to classical fixation protocols. The association between cpSRP54 and cpSRP43 in the stroma of intact chloroplasts was previously reported using a similar approach (Hutin et al., 2002). In micrographs from wild-type and mutant backgrounds examined, immunolabeling of cpSRP43 was observed about twice as often as ALB3. Therefore, colocalization events were evaluated in relation to the less frequently detected protein ALB3 and were calculated for independent micrographs by counting the number of colocalization events (Figure 6A). χ² analysis was used to confirm the statistical validity of the results (data not shown).

Figure 6.

cpSRP43 Interaction with ALB3 Takes Place in the Absence of cpSRP54 and cpFtsY.

(A) cpSRP43 colocalizes with ALB3 in the absence of cpSRP54 and cpFtsY. Immunolabeling of ALB3 (white arrowheads; 10-nm gold particle) and cpSRP43 (black arrowheads; 5-nm gold particle) in wild-type, cpftsy, ffc, and ffc cpftsy backgrounds was conducted as described in Methods. Representative micrographs are shown for each sample. Bars = 100 nm. Analysis of independent micrographs was performed to evaluate the frequency of cpSRP43 and ALB3 colocalization events. For each genotype, the percentage of ALB3 colocalized with cpSRP43, the total amount of cpSRP43 immunolabeled observed, and the total amount of ALB3 immunolabeled observed in all micrographs, respectively, are indicated as follows: wild type, 0%/147/78; ffc, 18%/92/40; cpftsy, 1%/181/87; and ffc cpftsy, 16%/185/75.

(B) cpSRP43 specifically coprecipitates ALB3. Salt-washed thylakoids (200 μg) were incubated with different amounts of cpSRP43-his protein (0 to 30 μg) as indicated in the presence of 0.5 mM GMP-PNP. Treated thylakoids were washed, solubilized in maltoside, and mixed with Talon resin to repurify His-tagged cpSRP43 and all proteins bound to cpSRP43 as described previously (Moore et al., 2003). cpSRP43 and any copurified proteins were eluted from the Talon resin as described in Methods. The eluted proteins were then examined on protein gel blots probed with antibodies directed against the proteins indicated at left. A lane of salt-washed thylakoids was included to verify the specificity of each antibody.

In the wild-type background and in the cpftsy mutant, where cpSRP54/cpSRP43 heterodimers persist, cpSRP43 did not or only occasionally colocalized with ALB3 (0 and 1% of colocalization events observed, respectively). These data are consistent with previous biochemical studies showing that cpSRP interacts poorly with ALB3 in the absence of cpFtsY and only takes place efficiently in the presence of cpFtsY when nonhydrolyzable GTP is added to stabilize cpSRP–cpFtsY interaction at the membrane in a complex with ALB3 (Moore et al., 2003). Interestingly, in the absence of cpSRP54 (ffc and ffc cpftsy backgrounds; Figure 6A), colocalization of cpSRP43 with ALB3 can clearly be observed (18 and 16%, respectively). Taken together, these results support the possibility that in both the ffc mutant and the ffc cpftsy double mutant, cpSRP43 may act to target LHCPs to ALB3, owing to an affinity of cpSRP43 for ALB3.

To understand whether cpSRP43 alone possesses the ability to target LHCP substrates through interactions with ALB3, thylakoids salt-washed to remove detectable cpSRP54, cpSRP43, and cpFtsY were incubated with increasing amounts of His-tagged cpSRP43. Following removal of the unbound cpSRP43, thylakoid-bound cpSRP43 was repurified from maltoside-solubilized membranes and examined by protein gel blotting for the presence of copurified proteins (Figure 6B). Both OE23 and cpSecY were undetectable in these coprecipitation assays. LHCP was detectable at the same low level whether cpSRP43 was absent or present in amounts as high as 4 μg, suggesting that the small amount of copurified LHCP observed in these assays resulted from nonspecific binding to the metal affinity beads.

When 10 μg of cpSRP43 was included in the assay, a slight increase in the amount of copurified LHCP was observed relative to levels seen in the absence of cpSRP43. However, this most likely stemmed from nonspecific binding of cpSRP43 to thylakoids, since no additional LHCP signal resulted from incubation of thylakoids with 30 μg of cpSRP43. This also indicates that LHCP, once properly integrated and assembled into light-harvesting complexes, is no longer available to interact specifically with cpSRP43, in contrast with the specific cpSRP43–LHCP interaction that takes place in solution as LHCP targeting substrate crosses the stroma (Tu et al., 2000). By contrast, ALB3 copurified with cpSRP43 in a concentration-dependent manner, revealing a clear ability of cpSRP43 to interact with ALB3. The demonstration of an interaction between ALB3 and cpSRP43 without a requirement of cpFtsY or cpSRP54 supports the in situ colocalization data and the putative bypass of these two components by cpSRP43 in posttranslational targeting of LHCPs to the thylakoid membrane.

DISCUSSION

cpFtsY Is Not Essential for the Survival of Arabidopsis Plants

To date, the impact of cpFtsY deficiency on the cpSRP protein targeting pathway in vivo has been investigated only in maize (Asakura et al., 2004), in which the absence (or reduction) of cpFtsY produced seedling-lethal plants that died 2 weeks after germination. The situation is clearly distinct in Arabidopsis, in which the null mutant identified in this study exhibited a severe chlorotic mutant phenotype similar to that in maize but is viable. This observation is in agreement with the recent map-based cloning of fro4, a mutation residing in cpFtsY (Durett et al., 2006). The fro4 mutants were identified for their inability to increase root Fe(III) chelatase activity when exposed to low iron availability. No studies were conducted to examine the effect of cpFtsY deficiency on the cpSRP pathway, despite fro4 mutants exhibiting a similar severe chlorotic phenotype. It is noteworthy that in both Arabidopsis and maize, chloroplast ultrastructure is similarly perturbed during the first weeks after germination. However, growing cpFtsY-deficient Arabidopsis under low light allowed the recovery of normal chloroplast ultrastructure after 3 to 4 months of culture (data not shown), suggesting that the cpFtsY mutant may be sensitive to photooxidative stress.

The Function of cpFtsY Extends beyond Its Role as a cpSRP Receptor in the Biogenesis of Chloroplast-Synthesized Thylakoid Proteins

It has been demonstrated that cpSRP54 binds nascent D1 polypeptides (Nilsson et al., 1999) and that ffc mutants lacking cpSRP54 exhibit a reduction of D1 insertion into the thylakoid membranes (Amin et al., 1999). This was accompanied by reduced accumulation of other PSI and PSII reaction center components in the ffc mutants, indicating that cpSRP54 contributes to the targeting of chloroplast-encoded proteins through a cotranslational process. Nevertheless, a substantial level of these chloroplast-encoded proteins accumulated in the absence of cpSRP54, suggesting that a less efficient cpSRP54-independent routing mechanism functions in the localization of chloroplast-synthesized thylakoid proteins (Figure 4) (Amin et al., 1999). The comparison of cpftsy and chaos ffc phenotypes indicates that the chloroplast-encoded proteins (D1, PsaA/B) are affected similarly in both mutants, consistent with the role of cpFtsY as a receptor for cpSRP in cotranslational targeting to the thylakoid. However, two observations suggest that cpFtsY also functions in photosystem assembly. First, PSII photochemical activity (Fv/Fm) was affected in the cpftsy mutant but not in the chaos ffc double mutant. Second, photosynthetic electron transport (ΔF/Fm′) was diminished in the cpftsy mutant relative to the chaos ffc double mutant, mainly at light levels of <250 μmol·m⁻²·s⁻¹. Inhibition of Fv/Fm and ΔF/Fm′ occurred despite the fact that these two mutants accumulate similar pigment and antenna levels. Similarly, photosynthetic electron transport was inhibited in ffc cpftsy double mutants relative to ffc mutants (see Supplemental Figure 1 online).

Although we have not examined the accumulation of all photosystem core polypeptides, these differences do not correlate with the level of photosystem core polypeptides examined. In this context, it is possible that cpFtsY plays a role in photosystem assembly beyond its receptor action in cpSRP targeting. This is supported by recent results showing that cpftsy is linked to the posttranslational reduction of Fe(III) chelate reductase activity (Durett et al., 2006). The absence of cpFtsY may reduce iron content in the very early stage of development. PSI and the cytochrome b₆/f complex contain iron–sulfur clusters; therefore, iron is crucial for the function of those photosynthetic complexes. Iron deficiency would lead to an inhibition of electron transfer through PSI and/or cytochrome b₆/f, thus explaining the fluorescence results described above, which showed a marked reduction of photosynthetic electron transport efficiency when cpFtsY was lacking. Inhibition of PSII photochemistry, indicated by a decreased Fv/Fm, could be a secondary effect of the impairment of the PSI and cytochrome b₆/f electron transfer activities.

Presumably, any additional function of cpFtsY, beyond its role as a membrane receptor for cpSRP, relies on its ability to localize to the thylakoid membrane, since immunolocalization failed to detect cpFtsY in the stroma (data not shown). Despite the fact that cpFtsY is distributed almost equally between stroma and thylakoids prepared from pea (Pisum sativum) chloroplasts (Tu et al., 1999), our immunolocalization data are in agreement with biochemical results obtained with maize (Asakura et al., 2004) showing that cpFtsY is almost entirely associated with the thylakoid membrane.

cpSRP54 Activity Is Detrimental to Thylakoid Biogenesis in the Absence of cpFtsY

Similar to maize mutants lacking cpFtsY (Asakura et al., 2004), the Arabidopsis cpftsy mutant exhibited a stronger phenotype than mutants lacking either cpSRP43 or cpSRP54. This observation supports the proposal formulated by Asakura and coworkers (2004) that cpFtsY plays an essential role for LHCP targeting as well as the assembly of other membrane components of the photosynthetic apparatus. Nevertheless, analysis of the ffc cpftsy double mutant clearly modifies our interpretation of the cpftsy phenotype described above. Epistasy of the ffc mutation on the cpftsy mutation (at least concerning proteins targeted by cpSRP) indicates that both genes have a complementary action and that cpSRP54 must be present for cpFtsY to function in the chloroplast SRP targeting pathway; the severe phenotype of plants lacking cpFtsY (cpftsy mutant) is largely reversed by additionally removing cpSRP54 (ffc cpftsy double mutant). The resulting phenotype of the ffc cpftsy double mutant closely resembles that of plants lacking cpSRP54 alone (ffc). Taken together, our data indicate that an alternative sorting pathway functions in the chloroplast independent of both cpSRP54 and cpFtsY. Our data also suggest that targeting substrates, normally targeted by cpSRP/cpFtsY, gain access to the alternative pathway only in the absence of cpSRP54. In the presence of cpSRP54, cpFtsY is strictly required.

In cotranslational protein targeting by mammalian and bacterial SRPs, the targeting reaction is initiated in the cytosol by the interaction of a cpSRP54 homolog (SRP54 and Ffh, respectively) with N-terminal hydrophobic signal sequences in the targeting substrates. SRP then directs its ribosome-associated targeting substrate to an SRP receptor homolog of cpFtsY at the membrane, owing to the affinity of SRP54 for the receptor, an interaction that stimulates GTP binding by both SRP54 and its receptor. In the presence of an available translocase, the targeted substrate is released from SRP54 followed by GTP hydrolysis to release SRP from its receptor (Pool, 2005). Biochemical studies indicate that these interactions are conserved in chloroplasts. A stable complex is formed between cpSRP54 and cpFtsY at the thylakoid in the presence of nonhydrolyzable GTP to prevent their dissociation (Moore et al., 2003). The cpSRP54-cpFtsY complex can form independently of cpSRP43 and associates with the ALB3 translocon. In this context, it would be expected that targeting substrates would remain bound to cpSRP54 in the absence of cpFtsY and thereby be unavailable for membrane insertion during thylakoid biogenesis. In maize that lacks cpFtsY, LHCP bound to cpSRP was found to accumulate in the stroma and on the surface of the thylakoids, supporting the need for cpFtsY to release targeting substrates from cpSRP (Asakura et al., 2004). While the severe nature of the observed cpftsy phenotype is consistent with substrates entering a dead-end targeting pathway upon interaction with cpSRP54, the less severe ffc-like phenotype exhibited by the cpftsy ffc double mutant indicates that in the absence of substrate binding by cpSRP54, substrates can be successfully routed to the thylakoid by an alternative pathway that leads to their proper integration and assembly. Although cpFtsY appears to play an additional role in photosystem assembly, as discussed above, the similar accumulation of LHCPs in ffc and cpftsy ffc mutants indicates that once LHCP localization is independent of cpSRP54, it also becomes independent of cpFtsY, even if cpFtsY is present. Since the proper localization of LHCPs relies on the function of ALB3 for proper integration into the thylakoid membrane, it would be expected that a cpSRP54-independent (or cpSRP54/cpFtsY-independent) routing mechanism that is active in both the ffc and cpftsy ffc mutants would also possess the ability to direct LHCP-targeting substrates to ALB3.

cpSRP43 Is a Component of a cpSRP54/cpFtsY-Independent LHCP-Targeting Pathway

Insight into the nature of the cpSRP54-independent targeting of LHCP comes from consideration of the ffc chaos double mutant, which lacks both cpSRP54 and cpSRP43 (Hutin et al., 2002). Unlike the ffc or cpftsy ffc mutants, the absence of both cpSRP subunits results in a marked decrease in LHCP accumulation, indicating that cpSRP43 is required for LHCP localization even in the absence of cpSRP54. This theory has been tested by expression in the chaos background of a cpSRP43 variant that is unable to interact with cpSRP54 and to support the formation of a cpSRP-LHCP transit complex (Goforth et al., 2004). Reversion of transgenic plants with the cpSRP43 variant, which contains its LHCP binding site, confirmed the possibility that cpSRP43 can function independent of cpSRP54 in the delivery of LHCP to thylakoids. Moreover, we found that cpSRP43 colocalizes with ALB3 in the absence of cpSRP54 (Figure 6). While colocalization data alone may not be interpreted as a direct indication of protein–protein interaction, they are consistent with the finding that recombinant cpSRP43 alone exhibits affinity for ALB3 (Figure 6). Taken together with the fact that ALB3 is an essential component for thylakoid biogenesis and for posttranslational integration of LHCPs by the cpSRP-targeting pathway (Sundberg et al., 1997; Moore et al., 2000, 2003), it seems likely that cpSRP43 functions in targeting LHCPs to ALB3 when cpSRP54 is absent (ffc and cpftsy ffc mutants). We are currently examining details of the cpSRP43–ALB3 interaction to understand whether this interaction plays a role in the cpSRP54/cpFtsY-dependent LHCP targeting mechanism.

In vitro (Schuenemann et al., 1998; Tu et al., 1999; Yuan et al., 2002), all soluble components of the cpSRP pathway (cpSRP54, cpSRP43, and cpFtsY) are required to achieve the integration of LHCP into the thylakoids. The absence of only one component reduces LHCP integration to levels that are barely detectable. Prior to the studies reported here, this result was difficult to reconcile with the partial LHCP targeting observed in both ffc and chaos mutants and with the additive phenotype of the ffc chaos double mutant (Hutin et al., 2002). The existence of an alternative pathway based on cpSRP43 provides information necessary to better understand the results of previous in vivo studies. The fact that cpSRP43 alone appears unable to support LHCP integration into isolated thylakoids suggests that cpSRP43 may function along with other stroma proteins to mediate LHCP integration in ffc and cpftsy ffc mutants. The use of chloroplast subfractions from these mutants should aid in understanding details of this new LHCP routing mechanism. While other components may be required for such a pathway, sizing chromatography of cpSRP43 from soluble extracts of plants lacking cpSRP54 suggests that if other components do function with cpSRP43, they are not bound to cpSRP43 prior to interaction with LHCP-targeting substrates (data not shown).

To summarize, we propose a model (Figure 7) whereby LHCP binding to cpSRP43-cpSRP54 leads to the formation of a transit complex in the stroma to target LHCP from the chloroplast envelope to the thylakoid, owing to the ability of cpSRP54 to bind cpFtsY at the membrane. Upon transfer of LHCP to ALB3, GTP hydrolysis by both cpSRP54 and cpFtsY would release cpSRP54 from its receptor at the membrane. Thylakoid proteins encoded by the chloroplast and routed using the cotranslational cpSRP pathway (such as D1, PsaA, or PsaB) would use a similar mechanism that differs only by the absence of cpSRP43 and the presence of the ribosome. Nevertheless, an alternative pathway targeting nucleus-encoded antenna would coexist. Such a pathway would use cpSRP43 to stabilize substrate in the stroma and route proteins to ALB3, owing to the affinity of ALB3 for cpSRP43. This alternative mechanism appears to be independent of cpSRP54 and cpFtsY for protein integration. It should be noted that ELIP proteins, LHCP-related proteins induced by various stresses, have been described to require cpSRP43 but not cpSRP54 for their targeting (Hutin et al., 2003), suggesting that ELIP proteins are possible substrates for this alternative pathway. Increased accumulation of ELIPs observed in the ffc cpftsy double mutant compared with the ffc or cpftsy single mutant (data not shown) adds support to this view. Additional pathways also have to be considered, as the ffc chaos mutant still exhibited ∼20% of LHCP. We may assume that the alternative pathway described here may work without cpSRP43 less efficiently. Therefore, it would rely only on ALB3 proteins. This would be in agreement with the alb3 mutant phenotype (Sundberg et al., 1997), which has no pigments (suggesting that LHCPs do not integrate in such mutants). Unfortunately, such a hypothesis cannot be genetically tested, as the alb3 mutation is lethal.

Figure 7.

Model Proposed for LHCP Targeting to Thylakoids, Which Integrates the cpSRP54/cpFtsY-Dependent Pathway and the Independent Alternative Pathway Identified Here.

Refer to Discussion for a detailed explanation. TIC, translocon of the inner membrane of chloroplasts; TOC, translocon of the outer membrane of chloroplasts.

Eight years ago, Gunter Blobel received the Nobel Prize for his work on SRP protein targeting. One of his major achievements was the identification of signal peptides and the analysis of the ribonucleotide protein complex needed to mediate protein targeting to the endoplasmic reticulum. In chloroplasts, the cpSRP pathway was named after the discovery that thylakoid targeting of LHCP required cpSRP54 (Li et al., 1995), the canonical soluble SRP protein conserved in bacteria and mammals. Nevertheless, plants seem to have developed unique features for LHCP targeting: posttranslational SRP activity, loss of an SRP RNA component, the presence of a plant-specific SRP component (cpSRP43), and a cpSRP43 binding motif in LHCPs. Now it appears that the robust nature of LHCP targeting is maintained through the function of an alternative targeting system that can bypass the evolutionarily conserved SRP soluble components, cpSRP54 and cpFtsY, using cpSRP43, the only plant-specific component. The increased number of recently published genomes has made it possible to identify putative homologs of cpSRP43 in monocots (Os 03g0131900 in Oryza sativa), moss (clone ppls42j07 from Physcomitrella), and even the green alga Ostreoccocus tauri (Ot 02g03770). On the other hand, no obvious homolog of cpSRP43 is present in cyanobacteria. These data suggest that cpSRP43 is not restricted to higher plants and could have appeared during the evolution of LHCPs. In this context, it is noteworthy that the red algae possess an SRP RNA moiety that is absent in green algae, moss, and higher plants (Rosenblad and Samuelsson, 2004) and that to date no cpSRP43 has been found in the red algae. This could indicate that cpSRP43 is specific to green plants.

METHODS

Plant Materials and Growth Conditions

The Arabidopsis thaliana null mutant lacking cpSRP54 (ffc) and the double mutant lacking both cpSRP54 and cpSRP43 (ffc chaos) have been described previously (Klimyuk et al., 1999; Amin et al., 1999; Hutin et al., 2002).

The Arabidopsis mutant lacking cpFtsY (cpftsy; CS3171) was isolated from the George Redei chlorotic mutant collection deposited at the ABRC and identified by protein gel blot screening using antibody directed against cpFtsY (Tu et al., 1999). The mutation was identified by sequencing the PCR product obtained with primers (480) 5′-CTCTAGCACAACTGCCATGGCAACTTCT-3′ and (489) 5′-TAGTGAGACGAGACACAAGCAGTTCCTAT-3′, which hybridize upstream and downstream, respectively, of the cpFtsY coding sequence. The identified EMS cpftsy mutant was backcrossed three times with wild-type Columbia plants before conducting physiological and biochemical analyses. Plants were grown in soil at 23/17°C (day/night) with an 8-h illumination photoperiod (300 μmol·m⁻²·s⁻¹) and a RH of 60 to 85%. All studies were performed using leaves harvested during the vegetative phase, before flowering.

Complementation of the EMS cpftsy Mutant

DNA Constructs

The plasmid pKYLX71-35S² (Maiti et al., 1993) is a binary vector allowing expression of subcloned cDNA under the control of a modified cauliflower mosaic virus 35S promoter. cpFtsY cDNA digested from pTU1 (Tu et al., 1999) at the NcoI and HindIII sites was blunt-ended with Klenow and cloned into pKYLX71-35S² at the XhoI site to form p35S²-cpFtsY. The resulting DNA was amplified in Escherichia coli BL21 supplied with 10 μg/mL tetracycline and subsequently introduced into Agrobacterium tumefaciens strain GV3101 (Koncz et al., 1984) using tetracycline (2 μg/mL) for selection.

Plant Transformation

Due to the severe cpftsy phenotype, transformation of plants containing the cpftsy mutation was performed with heterozygous (cpFtsY/cpftsy) plants. The binary plasmid described above containing the 35S²-cpFtsY cassette was introduced into heterozygous plants by Agrobacterium vacuum infiltration (Clough and Bent, 1998). Transformed seeds were selected in the T2 generation from primary transformants on a medium supplied with kanamycin (50 μg/mL). The presence of at least one copy of 35S²-cpFtsY in plants homozygous for the cpftsy genomic mutation was confirmed by sequencing PCR products amplified with forward and reverse primers that will hybridize to the intron sequence flanking the EMS mutation, 5′-CAAATAGGTGAGAAATTTTGCC-3′ (FtMUT5′) and 5′-CACCAAACACCGATACTGGCAAGCC-3′ (FtMUT3′).

Electron Microscopy and Immunocytochemistry

Leaf discs (3 mm in diameter) were taken from the lamina of detached leaves and covered with hexadecene. The discs were rapidly degassed in a syringe filled with hexadecene, transferred to a specimen carrier filled with hexadecene, and flash-frozen with a high-pressure freezing system (EMPACT; Leica Microsystems) prior to transfer in liquid nitrogen to cryovials filled with dry acetone containing 0.2% uranyl acetate. Following incubation for 3 d at −90°C in an automatic freeze substitution system (AFS; Leica Microsystems), the temperature of the substitution medium was raised to −60°C (by increasing the temperature at 4°C/h), maintained at −60°C for 8 h, raised to −30°C (4°C/h over a period of 7.5 h), and maintained at −30°C for 12 h. Substitution medium was changed (2 h at −30°C) and then rinsed twice with pure acetone for 2 h at −20°C and 1 h at −15°C prior to progressive infiltration of the specimens with London Gold resin at 15°C for 24 h. Finally, the resin was polymerized in UV light, first at −15°C and then at 20°C.

Sections (60 nm thick) were collected on 200-mesh nickel grids (Gilder) covered with a parlodion film and incubated with primary antibodies against cpSRP43, cpSRP54, cpFtsY, and ALB3 proteins raised in rabbit, mouse, or chicken. Colocalization was revealed by incubating sections with mixtures of primary antibodies, then with mixtures of 1:45 diluted secondary antibodies (Biocell International) conjugated to 10- or 5-nm gold particles. The sections were contrasted with uranyl and lead prior to observation with an FEI CM10 electron microscope operated at 80 kV. Digital images were recorded as 6-megapixel files using an AMT XR 60 camera.

Pigment Analyses

Leaf discs (0.8 cm in diameter) were used for pigment quantification after extraction in pure methanol. Photosynthetic pigments were identified using HPLC (Lagarde et al., 2000).

Protein Extraction and Immunoblot Analyses

Immunodetection of cpSRP54, cpSRP43, and cpFtsY was performed using total soluble protein extract from fresh leaves isolated as described (Amin et al., 1999). Total membrane proteins were extracted as described (Gillet et al., 1998). Protein concentration was determined using the Lowry protein assay (Sigma-Aldrich). Membrane proteins (D1, PsaA/B, ClpC, fibrilin C34, PC, OE33, and PsbS) were resolved by SDS-PAGE, transferred to nitrocellulose membranes, and detected with alkaline phosphatase–conjugated secondary antibody. Antiserum against LHCPs, including Lhca1, Lhca2, Lhcb1, Lhcb2, Lhcb3, Lhcb5, Lhcb6, Lhca3, Lhcb4, and Lhca4, have been described (Hoffman et al., 1987; Knoetzel et al., 1992; Sigrist and Staehelin, 1992, 1994; Krol et al., 1995). Antisera against D1, PsaA/B, OE33, PC, cpSRP54, cpSRP43, and PsbS were used as described (Hutin et al., 2002). Antisera against fibrilin C34 was used at 1:1000 dilution with 3, 5, and 7 μg of total membrane proteins loaded in each lane. Protein gel blot analysis conducted with antiserum directed against cpFtsY (Tu et al., 1999) was performed using 10 μg of total soluble protein extract. The revelation of alkaline phosphate activity was performed using nitroblue tetrazolium (Roche 1,087,479)/5-bromo-4-chloro-3-indolyl phosphate (Interchim UP38047A) as described (Sambrook et al., 1989). Densitometry analysis of the immunoblots was performed using YabGelImageX1.0 (http://homepage.mac.com/yabyab/rb/gelimage.html).

Chlorophyll Fluorescence Measurements

In vivo chlorophyll fluorescence from attached or detached leaves was measured at room temperature using a PAM-2000 fluorometer (Walz) as described (Hutin et al., 2003). In brief, the maximum quantum yield of PSII photochemistry was measured in dark-adapted leaves by subtracting the initial level of chlorophyll fluorescence (Fo; induced by a dim red light modulated at 600 kHz) from the maximum fluorescence level (Fm; induced by an 800-ms pulse of intense white light): (Fm − Fo)/Fm = Fv/Fm. The actual quantum yield of PSII photochemistry was measured in light-adapted leaves with the fluorescence parameter ΔF/Fm′ = (Fm′ − Fs)/Fm′, where Fm′ and Fs are the maximal and steady state fluorescence levels, respectively, in the light.

Coimmunoprecipitation

Soluble protein extract (300 μL) prepared as described above was incubated for 2 h at 4°C with mouse anti-cpSRP54 antiserum (7.5 μL). After 1 h of incubation with 20 μL of protein G–Sepharose (Amersham), IgG beads were washed four times with 20 mM HEPES-KOH, pH 8.0, 5 mM MgCl₂, and 180 mM NaCl. Bound proteins were eluted in 40 μL of 1× Laemmli buffer for SDS-PAGE and analyzed by protein gel blotting.

Binding of cpSRP43 to ALB3

Construction, Expression, and Purification of His-Tagged cpSRP43

The coding sequence for mature cpSRP43 beginning with AAVQRNY was amplified from pGEX-4T-2 (Moore et al., 2003) using a forward primer that coded for a 5′ BamHI and a six-His tag in conjunction with a reverse primer corresponding to the C terminus of cpSRP43 and incorporating an EcoRI restriction site. The resulting PCR product was inserted into pGEX-6P-2 (GE Healthcare) using BamHI and EcoRI restriction sites to create his-cpSRP43-pGEX-6P-2. Sequence analysis confirmed the addition of the N-terminal amino acids GSHHHHHH to cpSRP43.

BL21 Star (Invitrogen) cells containing his-cpSRP43-pGEX-6P-2 were grown to mid log phase and induced with 0.5 mM isopropyl-β-d-thiogalactopyranoside for 2 h at 37°C to produce GST-cpSRP43-his. Initial purification utilized Glutathione Sepharose Fast Flow (GE Healthcare). GST-cpSRP43-his was desalted into 50 mM Tris-HCl, pH 7.0 (at 25°C), 150 mM NaCl, 1 mM EDTA, and 1 mM DTT utilizing a HiPrep 26/10 desalting column (GE Healthcare). The GST affinity tag was cleaved by incubation with PreScission Protease (GE Healthcare) overnight at 4°C prior to desalting into PBS. Cleaved GST and PreScission Protease were removed using Glutathione Sepharose Fast Flow resin. The resulting cpSRP43-his was desalted into 10 mM HEPES and 10 mM MgCl₂ as described above, divided into aliquots, and stored at −80°C for further use.

Formation of the cpSRP43-ALB3 Complex

Salt-washed thylakoids were prepared as described (Yuan et al., 2002) and analyzed by protein gel blot to confirm the absence of cpSRP43, cpSRP54, and cpFtsY. The association between cpSRP43 and ALB3 was examined by incubating salt-washed thylakoids (200 mg of chlorophyll) with cpSRP43-his in the presence of 0.5 mM GMP-PNP for 30 min at 25°C. Treated thylakoids were isolated by centrifugation, buffer-washed with IBM (50 mM HEPES-KOH, pH 8, 330 mM sorbitol, and 10 mM MgCl₂), and reisolated by centrifugation. The washed membranes were resuspended in 100 μL of 3% BSA. Subsequently, 100 μL of 2% maltoside in IBM was added and the samples were incubated at room temperature with gentle shaking for 10 min prior to centrifugation at 70,000g for 12 min. The soluble fraction was diluted to 0.1% maltoside with IBM mixed with 50 μL of Talon resin (Clontech) for 30 min at 25°C to precipitate His-tagged cpSRP43 and all coprecipitating proteins. After incubation, samples were washed three times with 0.1% maltoside in IBM and once with IBM prior to elution with 100 μL of SDS solubilization buffer containing β-mercaptoethanol and heating to 70°C for 15 min as described (Moore et al., 2003). Protein gel blots of the precipitates were probed to identify the presence of the proteins indicated.

Accession Numbers

Arabidopsis Genome Initiative locus identifiers for the genes mentioned in this article are At2g28800 (ALB3), At2g45770 (cpFtsY), At2g47450 (cpSRP43), and At5g03940 (cpSRP54).

Supplemental Data

The following materials are available in the online version of this article.

• Supplemental Table 1. Densitometry Analysis of the Immunoblots Presented in Figures 3A, 3B, and 3C Made Using YabGelImageX1.0.

• Supplemental Figure 1. Quantum Yield of PSII-Mediated Electron Transport (ΔF/Fm′) of ffc, cpftsy, and ffc cpftsy Mutants and the Wild-Type Control.