Structural Basis for Specific Substrate Recognition by the Chloroplast Signal Recognition Particle Protein cpSRP43

See allHide authors and affiliations

Science  11 Jul 2008:
Vol. 321, Issue 5886, pp. 253-256
DOI: 10.1126/science.1158640


Secretory and membrane proteins carry amino-terminal signal sequences that, in cotranslational targeting, are recognized by the signal recognition particle protein SRP54 without sequence specificity. The most abundant membrane proteins on Earth are the light-harvesting chlorophyll a/b binding proteins (LHCPs). They are synthesized in the cytoplasm, imported into the chloroplast, and posttranslationally targeted to the thylakoid membrane by cpSRP, a heterodimer formed by cpSRP54 and cpSRP43. We present the 1.5 angstrom crystal structure of cpSRP43 characterized by a unique arrangement of chromodomains and ankyrin repeats. The overall shape and charge distribution of cpSRP43 resembles the SRP RNA, which is absent in chloroplasts. The complex with the internal signal sequence of LHCPs reveals that cpSRP43 specifically recognizes a DPLG peptide motif. We describe how cpSPR43 adapts the universally conserved SRP system to posttranslational targeting and insertion of the LHCP family of membrane proteins.

Most membrane proteins are cotranslationally inserted into membranes by the signal recognition particle (SRP) acting as an adaptor between the ribosome and the translocation channel. Although the composition of SRP varies, the ribonucleoprotein core of SRP and its guanosine triphosphate (GTP)–dependent interaction with a membrane-bound SRP receptor is universally conserved (1, 2). Helix 8 of the SRP RNA, one of the most conserved RNA structures known, was shown to play an essential, catalytic role in canonical SRPs (3, 4). It accelerates the rate of SRP-receptor complex formation, it significantly stimulates the guanosine triphosphatase (GTPase) activity of the SRP-receptor complex, and the activation can be ascribed to the so-called “tetraloop” (3, 4). SRP of chloroplasts (cpSRP) differs strikingly from these canonical SRP systems (5). cpSRP lacks the SRP RNA (6) and targets both chloroplast-encoded and nucleus-encoded substrates to the thylakoid membrane. Chloroplast-encoded substrates are cotranslationally targeted to the thylakoid membrane by the interaction of chloroplast ribosomes and cpSRP54 (7), whereas the nucleus-encoded LHCP substrates are imported into the chloroplast stroma and are posttranslationally targeted to the thylakoid membrane. LHCPs serve as antenna systems in photosynthesis (8, 9). They contain three trans-membrane helices (TM1 to 3) to which chlorophyll and carotenoid cofactors bind during insertion into the thylakoid membrane (10). In the stroma, LHCPs and cpSRP form the transit complex, which allows the transport of the hydrophobic LHCPs to the thylakoids in an insertion-competent conformation (11). In this posttranslational targeting system, cpSRP forms a heterodimer consisting of cpSPR54, the chloroplast homolog of the SRP core protein, and cpSRP43, a protein specifically dedicated to this targeting complex (12, 13).

cpSRP43 is a multidomain protein of 43 kD with a unique arrangement of chromodomains (CD1 to 3) and ankyrin repeats (Ank1 to 4) (Fig. 1A) that provide a robust scaffold for protein interactions (1417). Previously, only structures determined by nuclear magnetic resonance (NMR) spectroscopy of isolated chromodomains of cpSRP43 were available (18). The crystal structure of cpSRP43 from Arabidopsis thaliana (residues 85 to 267) was determined at 1.5 Å resolution (19) (table S1). The structure contains the N-terminal chromodomain CD1 (residues 85 to 129) and four ankyrin repeats (Ank1 to 4, residues 130 to 267), which together form an elongated horseshoe common to ankyrin-repeat proteins (Fig. 1, B and C) (16, 20). Although Ank2 and Ank3 have the typical fold of ankyrin repeats with a helix-turn-helix motif, the flanking Ank1 and Ank4 (Cap-repeats) contain unusual elongated helices. CD1 displays the classical chromodomain fold known from Polycomb (21) and other CD-containing proteins (15), which is built from a central three-stranded antiparallel β sheet supported by a perpendicular α helix on the concave surface of the β sheet. CD1 in the crystal structure differs significantly from the NMR structure (18) and forms a tight interface with the ankyrin repeats of about 240 Å2. CD1 and Ank1 are connected by the fusion of the CD helix with the N-terminal ankyrin helix (α1a/b), making it difficult to determine the domain boundaries. The most striking deviation from the canonical ankyrin-repeat structure is, however, observed in Ank4. Here, both α helices (α7 and α8) are elongated by an extension of 16 residues that protrude from the convex side of the horseshoe. The high conservation of this extension suggests a role as a protein-protein interaction site.

Fig. 1.

Structure of cpSRP43. (A) Scheme of the domain structure of cpSRP43 with chromodomains (CD1 to 3) and ankyrin repeats (Ank1 to 4). Domains present in the crystal structure are given by residue numbers and are indicated in rainbow colors. (B) Side view of cpSRP43 in ribbon representation. The domains are labeled. (C) Side view of cpSRP43 [90° rotation with respect to (B)]. Secondary structure elements are numbered. The N- and C-termini are labeled. (D) Front view of cpSRP43. The surface representation shows two hydrophobic grooves separated by a positive ridge. The molecular surface is colored blue and red according to positive and negative electrostatic potential, respectively. The asterisk highlights Tyr204 in Ank3. (E) Back view of cpSRP43 [same view as in (C)] showing the highly negatively charged surface with a spacing of negative charges reminiscent of RNA (see fig. S2).

Because the SRP RNA is absent in chloroplasts, cpSRP43 (isoelectric point 4.4) might be a structural and functional substitute for SRP RNA. Molecular mimicry between RNA and protein has been observed before with the complex formed by transfer RNA and the Thermus aquaticus elongation factor Tu (tRNA/EF-Tu) and EF-G of the translation apparatus representing the most prominent example (22). The analysis of the cpSRP43 surface shows high sequence conservation (fig. S1). One side exposes two predominantly hydrophobic grooves (grooves 1 and 2), which are separated by a positive ridge (Fig. 1D). The grooves may be involved in the interaction with LHCPs and/or cpSRP54. The opposite surface of cpSRP43 is highly negatively charged (Fig. 1E), with an intriguing regular spacing of negative charges of about 5.5 Å and 18 Å, reminiscent of the backbone phosphate groups of an RNA double helix. This pattern extends over the convex side of the ankyrin horseshoe. Overall, the shape and surface charge of cpSRP43 resembles helix 8 of the SRP RNA (fig. S2) (23), which binds to SRP54 in canonical SRPs and plays a catalytic role (3, 4). cpSRP43 also interacts with cpSRP54 (11, 12), but the molecular details of this interaction are not known. It induces only a moderate stimulation of the GTPase activity of the cpSRP54-receptor complex (24). Compared with cytosolic SRPs, a higher intrinsic rate of complex formation between cpSRP54 and cpFtsY (the chloropast receptor) was observed, which allows the requirement for the SRP RNA to be bypassed (25). Preliminary data indicate that cpSRP43 does not further accelerate complex formation (25). Therefore, cpSRP43 may only partially substitute for SRP RNA function. In cotranslational targeting with a translating ribosome present, cpSPR43 and the SRP RNA seem dispensable, whereas in posttranslational targeting, cpSRP43 is crucial for the formation of the transit complex.

In the transit complex, cpSRP43 binds to the L18 region, a highly conserved region in the stromal loop connecting TM2 and TM3 [residues 152 to 169 of Lhcb1 (Fig. 2A)] (12, 26), whereas cpSRP54 interacts with the TM regions of LHCP with a possible preference for TM3 (12, 27). The TM regions of LHCPs were shown to contribute to, but not to be sufficient for, promoting transit complex formation. However, the L18 region is strictly required (26, 28). In order to characterize this complex in more detail, we performed binding studies of cpSRP43 variants with a synthetic L18 peptide (L18p) using isothermal titration calorimetry (ITC). A high-affinity interaction with a 1:1 stoichiometry was observed (Fig. 2B and fig. S3). Because a truncated cpSRP43 construct comprising only CD1/Ank1 to 4 is sufficient for binding L18p without a significant drop in affinity (fig. S3), it was used for cocrystallization. The crystal structure of cpSRP43 in complex with L18p (Fig. 2C) is very similar to the free cpSRP43 (table S1). L18p binds to groove 1 across the inner (concave) surface of Ank2 to 4 (Fig. 2, B and C). It interacts with the ankyrin repeats at a similar location, as recently described for an ankyrin-repeat protein involved in transcriptional control (fig. S3) (29). L18p adopts an extended conformation except for a DPLG signature motif (Asp-Pro-Leu-Gly) typical for type I turns (residues 162 to 165) (Fig. 3A and fig. S4). The structure of the DPLG motif bound to cpSPR43 is basically identical to the one observed in the LHCP crystal structures (8, 9) and in other proteins containing a DPLG motif (30, 31), which emphasizes the stability of this structural element (fig. S5). We characterized the importance of the DPLG motif for the cpSRP43/L18p interaction by ITC with truncated and mutated L18 peptides (Fig. 3B and fig. S3). The motif is essential for binding to cpSRP43, as no interaction was observed in its absence. Replacement of Leu164 by Lys also results in a loss of binding to cpSRP43 (Fig. 3B). Although the replacement of the leucine in the DPLG motif does not alter the conformation of the type I turn (fig. S5), the long and charged side chain of a lysine cannot be accommodated in the hydrophobic binding groove. To test the functional relevance of this observation, membrane insertion of LHCP was analyzed by in vitro thylakoid import assays (Fig. 3C) (32). Wild-type LHCP is efficiently inserted as judged by the presence of a characteristic protease-resistant fragment [asterisk (Fig. 3C)], but the Leu164Lys mutant of LHCP does not insert into thylakoids at all. Therefore, the recognition of the DPLG motif by cpSPR43 is essential for the formation of a productive transit complex.

Fig. 2.

Structure of the cpSRP43/L18p complex. (A) LHCP topology with three transmembrane helices (TM1 to 3). The sequence of the L18 region is given for the major LHCP, Lhcb1 from Pisum sativum, which was used in this study (red, the DPLG motif is underlined). TM3 starts immediately after the L18 region. (B) Typical isothermal titration calorimetry (ITC) experiment of the cpSRP43 interaction with L18p. (C) Ribbon representation of cpSRP43 (blue) with bound L18p (as a ball-and-stick model, gray). The N- and C-termini are indicated. (D) Surface representation of the cpSRP43/L18p complex. The peptide (labeled by residue numbers) binds in the hydrophobic groove 1. Four residues at the N terminus and two residues at the C terminus are not resolved.

Fig. 3.

Specific recognition of the DPLG motif by cpSRP43. (A) Close-up view of the DPLG motif wrapped around the Tyr204 hook. The peptide is shown together with a 2mFobs – DFcalc electron density contoured at 2 σ. Residues essential for binding as tested by mutagenesis [see (B)] are marked with an asterisk. (B) Summary of ITC data. The DPLG motif within L18p is essential for cpSRP43 binding. The Tyr204Ala mutation in cpSRP43, as well as the Leu164Lys mutation in L18p, totally abolish the interaction (see fig. S3). (C) Thylakoid import assay. Wild-type pre-Lhcb1 (pLhcb1) and the Leu164Lys mutant were synthesized by in vitro translation (lanes 1), incubated with pea thylakoids (lanes 2) and treated with trypsin (lanes 3). The asterisks (*) mark the protease-protected fragment of membrane-inserted p-Lhcb1, which is absent in the Leu164Lys mutant. Note: The Leu164Lys mutant was in vitro translated in the presence of a suppressor Lys-tRNAamber, which always yields a mix of full-length protein (pLhcb1) and truncated protein (marked by ◯) (19).

The structural requirements for DPLG recognition are reflected by the specific properties of groove 1 with a pronounced bend, marked by Tyr204 of Ank3. The DPLG motif wraps around Tyr204 (Fig. 3A), and the replacement of Tyr204 by alanine abolishes the interaction with L18p (Fig. 3B). Adjacent positively charged residues seem to stabilize the interaction with L18p. Replacement of the arginines Arg161, Arg192, and also Arg226 by alanines (fig. S4) results in increased dissociation constants (fig. S3). Arg161 is in direct contact with L18p and, therefore, shows the strongest effect. Sequence analysis of FCPs (fucoxanthine chlorophyll a/c–binding proteins) of diatoms shows that, in this family of antenna complexes, the DPLG motif is present at the N terminus (fig. S6) (33). FCPs are also targeted by cpSRP, which further supports the role of the DPLG motif as a specificity determinant for transit complex formation. In LHCPs, the DPLG motif is part of a conserved interaction site for carotenoids (9), which are crucial for LHCP folding (10). When the cpSRP43/L18p and LHCP structures are superimposed (fig. 5A), Tyr204 superimposes with the carotenoid head group. Therefore, the interactions in the transit complex might be a prerequisite for the attachment of cofactors during membrane insertion.

cpSRP43 forms a stable heterodimer with cpSRP54 also in the absence of LHCP (12). Whereas the third chromodomain (CD3) is dispensable for the interaction with cpSRP54, CD2 is strictly required (34). Although CD2 is not present in the crystal structure, its location with respect to CD1 and the ankyrin repeats can be inferred from crystal symmetry (Fig. 4A and fig. S7). In different crystals obtained for cpSRP43 (19) CD1 always packs against the extension of Ank4 (helix α8) protruding from a symmetry-related molecule. A contact of about 270 Å2 is formed by this interaction. Two conserved, surface-exposed hydrophobic residues (Phe249 and Ile253) within helix α8 form the core of this interaction. Phe249 is accommodated in an aromatic cage (fig. S7), which is reminiscent of the interaction between chromodomains and methylated lysines in chromatin remodeling (21). We therefore propose that the conserved extension of Ank4 serves as binding site for CD2 of cpSRP43 (Fig. 4B). CD2 was recently shown to interact with an RRKR peptide from the C terminus of cpSRP54 (34). Together, these data prompt us to propose a model for the interaction between cpSRP54 and cpSRP43 in which the signal sequence–binding site located in the M domain of cpSRP54 would be positioned in the immediate vicinity of the L18 binding site in cpSRP43. Thereby, TM3 and the L18 region of LHCPs could be recognized simultaneously as suggested before (12). The location of the DPLG motif within LHCPs might be critical to position TM3 in a way that it can interact with cpSPR54. Although it cannot be predicted how TM1 and TM2 of LHCP are bound in the transit complex or where the GTPase domain of cpSRP54 would be located, the model presented here provides the basis for further experiments to characterize these interactions.

Fig. 4.

Model of the transit complex. (A) Analysis of cpSRP43 crystal packing. Ank4 of one molecule (light blue) (with L18p as a ball-and-stick model) forms a contact with CD1 from a neighboring molecule (gray, CD1*) (see fig. S7). (B) Structure-based model of the transit complex. CD2 is suggested to interact with Ank4. The M domain of cpSRP54 interacts with CD2 (34). The signal sequence–binding site of the M domain might preferentially bind TM3 of LHCPs (12, 27) and should therefore be in close proximity to L18 in groove 1 of cpSRP43.

LHCPs are multispanning membrane proteins and present in high amounts in the thylakoids. The specific requirements for efficient, posttranslational LHCP targeting and insertion have been met by recruiting cpSRP54 to the transit complex. The subsequent interaction between cpSRP54 and cpFtsY allows the use of the conserved downstream components of SRP-mediated membrane targeting. Although signal sequences are usually recognized without sequence specificity in both co- and posttranslational targeting (35), cpSPR43 adds specificity and provides additional interaction sites. The formation of the transit complex may compensate for the absence of the ribosome with cpSRP acting as a chaperone for the hydrophobic regions of the LHCPs before membrane insertion. Taken together, we show how posttranslational membrane insertion is accomplished by the adaptation of a universally conserved machinery.

Supporting Online Material

Materials and Methods

Figs. S1 to S7

Table S1


References and Notes

View Abstract

Navigate This Article