A unique sequence motif in the 54 kD subunit of the chloroplast signal recognition particle mediates binding to the 43 kD subunit

Chloroplasts contain a novel type of signal recognition particle (cpSRP) that consists of two proteins, cpSRP54 and cpSRP43. cpSRP is involved in the post-translational targeting of the nuclear encoded light-harvesting chlorophyll-binding proteins (LHCPs) to the thylakoid membrane by forming a soluble cpSRP.LHCP transit complex in the stroma. Despite high sequence homology between chloroplast and cytosolic SRP54 proteins, the 54-kDa subunit of cpSRP is unique in its ability to bind cpSRP43. In this report, we identified a 10-amino acid long segment of cpSRP54 that forms the cpSRP43-binding site. This segment is located at position 530-539 close to the C terminus of cpSRP54. In addition, we demonstrate that arginine at position 537 is essential for binding cpSRP43 and that mutation of arginine 536 drastically reduced cpSRP43 binding. Mutations within the cpSRP43-binding site of cpSRP54 that reduced or completely abolished cpSRP complex formation also did inhibit transit complex formation and integration of LHCP into the thylakoid membrane, reflecting the importance of these residues for LHCP targeting. Alignment studies revealed that the cpSRP43-binding site is conserved in chloroplast SRP54 proteins and is not present in any SRP54 subunit of cytosolic SRPs.


INTRODUCTION
The cytosolic signal recognition particle (SRP) is part of a ubiquitous protein targeting machinery that mediates the cotranslational insertion of membrane proteins into the endoplasmic reticulum of eukaryotes and the cytoplasmic membrane of prokaryotes. All known cytosolic SRPs are ribonucleoproteins and their minimal functional core is formed by an RNA component and a conserved ~54 kD protein (SRP54). SRP54 consists of an Nterminal NG-domain encoding a GTPase function and a C-terminal located M-domain, that binds to the signal sequence of the elongating substrate protein (1)(2)(3). Recently, it was demonstrated that chloroplasts contain an SRP that is involved in the posttranslational targeting of members of the nuclear encoded light harvesting chlorophyll binding protein family (LHCPs) to the thylakoid membrane (4)(5)(6). LHCPs form the peripheral antenna of photosystem I and II and comprise approximately one third of the thylakoid membrane proteins. Like all known cytosolic SRPs chloroplast SRP contains a 54 kD subunit (cpSRP54). Interestingly, in contrast to cytosolic SRPs chloroplast SRP does not contain a 4 Several thylakoid membrane proteins (e. g. D1, D2, PSI-A, PSI-B) are encoded by the chloroplast genome and cotranslationally inserted into the thylakoid membrane. Recent reports describe that cpSRP54 is involved in the cotranslational targeting of D1 to the thylakoid membrane by binding to the first transmembrane domain of the elongating nascent chain of D1 (12, 13). Notably, no interaction of cpSRP43 with the D1 protein was detected (12). Further evidence for an involvement of cpSRP54 in cotranslational targeting of thylakoid membrane proteins came from the analysis of Arabidopsis mutants lacking functional cpSRP54 . The young leaves of these plants showed a reduced level of the plastid encoded photosystem I and II reaction center proteins (14,15). These results supported the earlier observation that the chloroplast stroma contains two different pools of cpSRP54. One pool is bound to cpSRP43 and active in transit complex formation with LHCP, whereas a second pool of cpSRP54 was found to be associated with 70 S ribosomes in absence of cpSRP43 (6,16).

Yeast two-hybrid assay
The yeast two-hybrid assays were done as described in Jonas-Straube et al. (17), except for the following modifications. pGBKT7 constructs (see above) were used instead of pAS2 constructs as prey plasmids. For cpSRP54M constructs that showed only weak or no interaction with cpSRP43 in the yeast two-hybrid experiments, expression levels comparable with full-length cpSRP54M were verified by Western blot analysis using antibodies against the c-Myc epitope (BD Biosciences). Growth of the yeast cells on medium lacking leucine, tryptophan and histidine (-leu,-trp,-his) was classified in (++), (+) and (-), whereby (++) means that most colonies have a diameter of > 1,5 mm and (-) indicates normal background growth (whitish colonies < 0,6 mm). The filter lifts to measure β-galactosidase activity were incubated for at least 1.5 h to develop a blue color ((++): strong blue color; (+): medium blue color; (-): no blue color development). All pGBKT7-constructs were cotransformed with pACT2 in yeast cells and showed no self-activation of the reporter genes.

Transit complex formation
Transit complex formation was measured essentially as described (6)

Protein expression and FTIR spectroscopy
Recombinant cpSRP54M and cpSRP54M(∆536-540) were expressed from their corresponding pET-29b(+) constructs (see above) in the E. coli strain BL21(DE3). Cells were grown in LB medium containing 50 µg/ml ampicillin at 37 o C to an OD 600 of 0.6 and expression was induced by the addition of 1 mM isopropyl-ß-D-thiogalactopyranoside for 3 h.
Overexpressed protein was purified on Ni 2+ -NTA agarose (Qiagen) under native conditions as changed to 50 mM Tris-DCl, pH 8.0 in D 2 O and the protein samples were concentrated to 10-15 µg/µl using spin concentrators (Vivascience).
FTIR spectroscopy was carried out at 22°C on a Bruker IFS 88 spectrometer. For each spectrum, a 256-scan interferogram was collected at single beam mode with a 2 cm -1 resolution and a 1 cm -1 interval from the 2000 to 1000 cm -1 region. Reference spectra were recorded under identical conditions with solute buffer only. Each measurement was repeated

RESULTS
Previously, it was demonstrated that the interaction of cpSRP43 and cpSRP54 is mediated via the C-terminal located M-domain of cpSRP54 (cpSRP54M; residues 371-564) (17). In order to define more clearly the region of cpSRP54M that mediates binding to cpSRP43, serial deletions from either the amino-or the carboxyl-terminal end of cpSRP54M were cloned into the bait plasmid pGBKT7. The interaction of cpSRP54M and the various deletion constructs with mature cpSRP43 was tested in the yeast two-hybrid system using pACT2-43 as prey plasmid (Table 3). No obvious change in binding intensity was observed between cpSRP43 and the N-terminal deletion constructs cpSRP54M(485-564), cpSRP54M(521-564) and cpSRP54M(530-564) in comparison with full-length cpSRP54M (Table 3) To confirm this result we tested whether a synthetic peptide corresponding to the putative cpSRP43-binding site competes with cpSRP54M for the complex formation with cpSRP43 in in vitro pull-down experiments. Fig. 2 A shows that the binding of radiolabeled cpSRP54M to GST-cpSRP43 in the presence of increasing amounts of the synthetic peptide is progressively inhibited. In pull-down assays conducted in the presence of 7.5 µM synthetic peptide binding of cpSRP54M to cpSRP43 was blocked completely, whereas the addition of an unrelated peptide did not lead to a reduction of complex formation ( Fig. 2 A/B). Furthermore, we were able to corroborate the finding that residues 530-539 of cpSRP54 are involved in the formation of the cpSRP43-binding site by demonstrating that a deletion construct of cpSRP54M lacking this region is not able to bind cpSRP43 (Table 3).
A noticeable feature of the amino acid region of cpSRP54 comprising the cpSRP43-binding site is the presence of the positively charged pentapeptide (RRKRK) located at amino acid position 536-540. To determine whether this region is involved in binding of the highly negatively charged cpSRP43, we constructed cpSRP54(∆536-540) and tested its ability to bind cpSRP43 by yeast two-hybrid experiments and in vitro pulldown assays. As shown in Table 3 and Fig. 4, no interaction was observed between cpSRP54(∆536-540) and cpSRP43.
We proved that the removal of the positively charged pentapeptide did not lead to an overall structural rearragement by performing FTIR spectroscopy of highly purified recombinant cpSRP54M and cpSRP54(∆536-540). Both proteins generated very similar spectra indicating that their secondary structure is almost identical (Fig. 3). This result, together with the above mentioned observation that a synthetic peptide containing the pentapetide RRKRK inhibits binding of cpSRP54 to cpSRP43, demonstrates clearly that residues within this motif are essential for the formation of the cpSRP43-binding site.
We next aimed to define the role of the individual amino acids within the pentapeptide RRKRK (536-540) in binding cpSRP43. Therefore, site-directed mutagenesis was used to exchange one or more of these positively charged residues into uncharged amino acids. The ability of the generated cpSRP54M mutants to bind cpSRP43 was initially tested in yeast twohybrid experiments (Table 4). No significant change in binding intensity to cpSRP43 was observed when using the constructs cpSRP54M(K538M), cpSRP54M(R539G) or cpSRP54M(K540M). Even the simultaneous change of the positively charged residues K358, R539 and K540 (cpSRP54M(K538M, R539G, K540M)) or the deletion of these amino acids (cpSRP54M(∆538-540)) did not lead to a measurable loss of interaction with cpSRP43 in this system. However, the single mutation R537G caused a complete loss of binding of the corresponding construct cpSRP54M(R537G) to cpSRP43, since no β-galactosidase activity was detectable in the yeast two-hybrid system. Consistently, all other tested constructs containing the R537G mutation (cpSRP54M(R537G, K538M), cpSRP54M(R537G, K538M, R539G, K540M)) were also unable to interact with cpSRP43 ( To further support these results and to quantify them we measured the differences in binding of radiolabeled cpSRP54M or various constructs containing mutations within the RRKRK (536-540) motif to GST-cpSRP43 by in vitro pulldown experiments (Fig. 4 A). In accordance with the yeast two-hybrid experiments, results of the binding reactions show that cpSRP54M constructs containing the mutation R537G did not interact with cpSRP43 and that binding of cpSRP54M(R536G) to cpSRP43 was reduced on average by ~90 % compared to cpSRP54M.
The mutations K540M and K538M did not influence binding significantly, whereas the conversion of R539 into G539 resulted in a considerable reduction of binding by ~45 %. This reduction value was not detected in the semiquantitative yeast two-hybrid system.
We next sought to analyse whether the positive charge or the specific structure of the arginine side chain at position 536, 537 or 539 is required for interaction of cpSRP54 with cpSRP43.
Therefore, arginines at these positions were replaced individually with lysine and the interaction of the resulting constructs with cpSRP43 was tested by pull-down experiments. As shown in Fig. 4 B, cpSRP54M(R536K) and cpSRP54M(R537K) were not able to bind cpSRP43. A reduced binding efficiency was observed for cpSRP54(R539K). Hence, constructs containing the R/K mutations behaved in the same way as the R/G mutants, demonstrating that a positive charge at position 536, 537 or 539 is not sufficient to mediate interaction with cpSRP43. We then analysed whether the polar charged side chain of R537 can be functionally replaced by the polar side chains of glutamine or asparagine. As shown in  (18). This report describes that the Cterminal 26 residues of cpSRP54 are essential for complex formation with cpSRP43. In the present study, we performed a detailed analysis of the cpSRP43-binding site of cpSRP54. We demonstrate, that the essential binding site is located within residues 530-539 of cpSRP54.
Furthermore, we show that two residues R536 and R537 are crucial for binding cpSRP43.
R539 is also involved in binding cpSRP43, but is not essential for this process. These data demonstrate that the essential amino acids mediating binding to cpSRP43 are located in the Cterminal region of cpSRP54. However, they are not located within the last 26 amino acids (539-564) of cpSRP54, explaining our initial finding that removal of these amino acids did not completely abolish binding to cpSRP43.
Based on the above results one would expect that the cpSRP43 binding region (residues 530-539), including the essential amino acids R536 and R537 of Arabidopsis cpSRP54, is conserved in all chloroplast SRP54 proteins. In Fig. 7 an alignment of the C-termini of all known chloroplast SRP54 proteins is shown. As expected, the double arginine motif is conserved in all sequences. In addition, in all species this motif is followed by two positively charged amino acids (KR or KK). Two other residues that are conserved throughout all sequences (P532, G533) were identified. They are also located within the cpSRP43 binding region and current work is in progress to test whether these amino acids may also play an important role in binding cpSRP43. Alignment studies revealed that a region homologous to the cpSRP43-binding site is not present in any cytosolic SRP54. This is particularly noticeable in the case of the SRP54 homologue of E. coli, since chloroplast SRP54 differs from the cytosolic homologue of E. coli by a C-terminal extension containing the cpSRP43 binding site (Fig. 7). This finding explains our previous observation that bacterial SRP54 (Ffh) cannot form a complex with cpSRP43 (7).
In E. coli the Ffh protein is involved in the cotranslational transport of membrane proteins to the plasma membrane. During this process Ffh, that is bound to the bacterial SRP-RNA, interacts with the signal sequence of the nascent protein and also contacts the ribosome at the ribosomal subunit L23 that is located close to the nascent chain exit site (21-23). In plastids, cpSRP54 is a component of two fundamentally different mechanisms, since it is involved in post-and cotranslational targeting pathways. From these observation the question concerning the molecular details of switching between the post-and cotranslational mode of action of cpSRP54 arises. Provided that the cotranslational targeting mechanism in chloroplasts is similar to that in bacteria, the results from the present work show that the cpSRP43 binding site is located at a position of cpSRP54 that is apparently not required for the cotranslational pathway. However, since no RNA was identified yet as a component of the cotranslationally acting cpSRP, it might be possible that the cotranslational targeting mechanisms in chloroplasts and bacteria exhibit substantial differences. Therefore, more work is necessary to analyse the molecular details underlying the recruitment of cpSRP54 for functioning in the posttranslational targeting of LHCP and the cotranslational targeting of chloroplast encoded proteins to the thylakoid membrane.      The translation vectors encoding the indicated cpSRP54 constructs were generated by sitedirected mutagenesis using the cpSRP54 translation vector pAF1 (10) as template.   Nucleotides coding for the changed amino acids are underlined.