An intermolecular disulfide-based light switch for chloroplast psbD gene expression in Chlamydomonas reinhardtii

Authors

  • Christian Schwarz,

    1. Molekulare Pflanzenwissenschaften, Biozentrum Ludwig Maximilian University Munich, Grosshaderner Straße 2–4, 82152 Planegg-Martinsried, Germany
    Search for more papers by this author
  • Alexandra-Viola Bohne,

    1. Molekulare Pflanzenwissenschaften, Biozentrum Ludwig Maximilian University Munich, Grosshaderner Straße 2–4, 82152 Planegg-Martinsried, Germany
    Search for more papers by this author
  • Fei Wang,

    1. Molekulare Pflanzenwissenschaften, Biozentrum Ludwig Maximilian University Munich, Grosshaderner Straße 2–4, 82152 Planegg-Martinsried, Germany
    Search for more papers by this author
  • Francisco Javier Cejudo,

    1. Instituto de Bioquimica Vegetal y Fotosintesis, University of Seville and Consejo Superior de Investigaciones Cientificas, Avenida Americo Vespucio 49, 41902 Seville, Spain
    Search for more papers by this author
  • Jörg Nickelsen

    Corresponding author
    1. Molekulare Pflanzenwissenschaften, Biozentrum Ludwig Maximilian University Munich, Grosshaderner Straße 2–4, 82152 Planegg-Martinsried, Germany
    Search for more papers by this author

(e-mail joerg.nickelsen@lrz.uni-muenchen.de).

Summary

Expression of the chloroplast psbD gene encoding the D2 protein of the photosystem II reaction center is regulated by light. In the green alga Chlamydomonas reinhardtii, D2 synthesis requires a high-molecular-weight complex containing the RNA stabilization factor Nac2 and the translational activator RBP40. Based on size exclusion chromatography analyses, we provide evidence that light control of D2 synthesis depends on dynamic formation of the Nac2/RBP40 complex. Furthermore, 2D redox SDS–PAGE assays suggest an intermolecular disulfide bridge between Nac2 and Cys11 of RBP40 as the putative molecular basis for attachment of RBP40 to the complex in light-grown cells. This covalent link is reduced in the dark, most likely via NADPH-dependent thioredoxin reductase C, supporting the idea of a direct relationship between chloroplast gene expression and chloroplast carbon metabolism during dark adaption of algal cells.

Introduction

Due to the endosymbiotic origins of the chloroplast, its gene expression machinery is basically of prokaryotic origin. However, during the evolutionary development of chloroplasts, this machinery was extensively modified by recruitment of nucleus-encoded regulatory factors, which now constitute an intracellular network dedicated to coordination of gene expression in the nucleus and the organelle (Barkan, 2011). While recent years have seen the identification and characterization of a number of these trans-acting factors, much less is known about their precise molecular modes of action with regard to light-dependent regulation.

In this context, the idea of redox control of chloroplast gene expression has attracted much attention, as it provides an appealing basis for a direct link between photosynthetic activity and expression of photosynthesis-related chloroplast genes (Dietz and Pfannschmidt, 2011). Indeed, many elements of chloroplast gene expression, including RNA transcription, stabilization, processing and splicing, and translation have been shown to be affected directly or indirectly by the redox state of the organelle (Barnes and Mayfield, 2003). However, translation appears to represent the rate-limiting step for synthesis of chloroplast-encoded proteins (Eberhard et al., 2002; Zerges and Hauser, 2009).

In the green alga Chlamydomonas reinhardtii, synthesis of the large subunit of ribulose-1,5-bisphosphate carboxylase/oxygenase (Rubisco) encoded by the rbcL gene has been shown to be regulated via the redox state of the chloroplast glutathione pool, which in turn is modulated by light-induced oxidative stress (Irihimovitch and Shapira, 2000). Interestingly, the RbcL protein possesses an intrinsic non-specific RNA binding activity located within its N-terminal region (Yosef et al., 2004). It has therefore been postulated that binding of RbcL to its own mRNA blocks its translation if either its redox-controlled interaction with the chloroplast chaperone system or the Rubisco subunit assembly is disturbed (Cohen et al., 2005).

The most elaborate (but also the most controversial) model for redox-controlled translational regulation in chloroplasts has been described in C. reinhardtii for the psbA gene that encodes the D1 protein of the photosystem II (PSII) reaction center (Barnes and Mayfield, 2003; Zerges and Hauser, 2009). This model postulates that redox-controlled binding of a protein complex to the 5′ UTR of the psbA mRNA leads to recruitment of ribosomes. The heart of this complex is the RNA-binding protein RB47, whose activity is modulated by RB60, a disulfide isomerase homolog (Kim and Mayfield, 2002). RB60 was shown to form intermolecular disulfide bonds with RB47 in vitro, suggesting tight cooperation of these factors in vivo (Alergand et al., 2006). It was proposed that light-dependent reduction of the involved thiol groups in RB60 provides the molecular basis for light-dependent increases in D1 synthesis (Trebitsh et al., 2000).

We have previously shown that expression of the chloroplast psbD gene in C. reinhardtii is under the control of a high-molecular-weight (HMW) complex containing the RNA stabilization factor Nac2 and the translational activator RBP40 (Schwarz et al., 2007). Furthermore, translation of psbD mRNA depends on an U-rich element within its 5′ UTR that serves as a binding site for the translational activator RBP40 (Nickelsen et al., 1999; Ossenbühl and Nickelsen, 2000). Deletion of this element (ΔU) results in complete loss of D2 synthesis (Nickelsen et al., 1999; Figure 1a,b) but is partially restored in genetically selected second-site suppressor lines, namely suΔU+9 and suΔU–3 (position of the second-site suppressor mutations relative to the translation start), which harbor point mutations in a downstream RNA stem-loop structure encompassing the AUG start codon (Klinkert et al., 2006). In these lines, the psbD mRNA can be translated in the absence of RBP40 binding, leading to a model in which Nac2-assisted binding of RBP40 to the U-rich element affects the RNA conformation at the initiation codon, and thereby makes the initiation site accessible to the translational machinery (Schwarz et al., 2007). Thus, both RBP40 and the RNA stem-loop have the capacity to form a molecular switch that regulates psbD gene expression, and, as a direct consequence of the so-called CES (control by epistasis of synthesis) process, accumulation of the entire PSII in C. reinhardtii (Minai et al., 2006).

Figure 1.

 Rates of D2 synthesis depend on light conditions.
(a) Thylakoid membrane proteins (equivalent to 9 μg chlorophyll) from indicated strains were pulse-labeled with 35S-sulfur, fractionated by SDS–PAGE in 16% gels containing 8 m urea, and visualized by autoradiography. Lanes D, cells grown in the dark for 38 h prior to analysis; lanes L, cells weregrown in continuous light at 30 μmol m−2 sec−1. The gel on the right (marked ‘ee’) shows extended exposure of signals from suΔU+9.
(b) Coomassie-stained version of the gel shown in (a) to show equal loading.
(c) Densitometric quantification of D2 protein synthesis rates shown in (a). The results are representative of five independent experiments. D2 signals were normalized relative to the AtpA signal indicated by the asterisk in (a). The relative increase of D2 synthesis upon light induction is indicated.
(d) Northern blot analysis of psbD and atpB transcripts from the indicated strains grown under the same conditions as in (a).

Here, we report on the molecular mechanisms that underlie light-controlled regulation of psbD gene expression via the RBP40/RNA stem-loop switch. We provide evidence showing that RBP40 is required for this control, and that light-dependent formation of the active Nac2/RBP40 complex is likely to be mediated by establishment of an intermolecular disulfide bridge between the two factors. The redox state of this connection appears to represent a key determinant for synthesis of D2, and therefore PSII.

Results

Light regulation of D2 synthesis depends on the RBP40/RNA stem-loop switch

To test whether the RBP40/psbD RNA stem-loop switch is involved in the well-known light-dependent regulation of D2 synthesis in C. reinhardtii, we analyzed light-dependent D2 synthesis rates in deletion strain ΔU (which lacks the U-rich element) and the suppressor lines suΔU–3 and suΔU+9 which have lost the RBP40 binding site but contain a less stable RNA stem-loop (Malnoe et al., 1988; Klinkert et al., 2006). In pulse-labeling experiments, wild-type cells grown in the light exhibited an 2.7-fold increase in rates of D2 synthesis relative to cells that had been adapted to dark conditions for 38 h (Figure 1a,c). As described previously, in the dark, the ΔU strain showed no D2 protein synthesis, while both suppressor lines exhibited reduced D2 synthesis compared to the wild-type (Nickelsen et al., 1999). Moreover, in the light, D2 expression increased only 1.8- and 1.9-fold in suΔu–3 and suΔU+9 strains, respectively (Figure 1a,c). These results suggest that, in the suppressor lines, both the overall rate of psbD mRNA translation and the degree of induction of D2 synthesis by light are affected. Reduced levels of D1 synthesis in strains with affected D2 synthesis are due to secondary CES effects (Minai et al., 2006). Northern analyses verified that the observed differences are due to translational effects, as no significant alterations in psbD mRNA levels were observed under the conditions tested (Figure 1d). In conclusion, these findings suggest that RBP40 is required for efficient regulation of D2 synthesis by light as bypass of RBP40 function in the suppressor lines results in reduced levels of light control. This supports the hypothesis of a light switch constituted by the negatively acting psbD mRNA stem-loop at the AUG start codon and RBP40, in which this protein activates translation by changing the conformation of this RNA structure.

RBP40 contains a single Cys residue.  How does light affect this molecular switch? Redox reactions have been postulated to play critical roles during light activation of chloroplast gene expression (Barnes and Mayfield, 2003; Dietz and Pfannschmidt, 2011), and, interestingly, RBP40 was identified as a target for glutathionylation under conditions of oxidative stress in a proteomic analysis in C. reinhardtii (Michelet et al., 2008). Inspection of the amino acid sequence of RBP40 revealed the presence of only a single cysteine residue at amino acid position 11 (Cys11), which may serve as a target for glutathione binding (Figure 2a). However, Cys11 is located within the predicted N-terminal transit sequence of RBP40 that is supposed to be cleaved off upon import by the chloroplast (Barnes et al., 2004). Instead, Cys11 is still present in the mature protein. The predicted size of the putative RBP40 precursor is 44 kDa, while the in silico calculated mature form lacks the first 18 amino acids, resulting in a polypeptide of 42 kDa. The apparent molecular weight of immunodetected stromal RBP40 is approximately 40 kDa, which matches neither of the predicted sizes (Figure 2b). However, to test for the presence of a Cys residue in mature RBP40, various amounts of stromal proteins from C. reinhardtii were treated with the thiol-alkylating reagent PEG5000-maleimide for 120 min, and subsequently analyzed by immunoblotting. As shown in Figure 2(b), approximately 40% of the applied RBP40 material showed an alkylation-dependent size shift, indicating that Cys11 is still present in the mature RBP40. It remains unclear why only a subfraction of RBP40 was PEGylated, but, as reviewed by Veronese (2001) and Mero et al. (2011), a huge variety of parameters affect PEGylation efficiency, in some cases leading to incomplete alkylation reactions. These parameters include the thiol-reacting group, pH, molar ratios between target protein and alkylation reagent, molecular sizes of alkylation reagents, protein structure, oxidation state of thiol groups and THE PEG–SDS interaction, some of which may also have affected RBP40 PEGylation (Bernazzani et al., 2004; Hu et al., 2005; Kubetzko et al., 2006; Xiong et al., 2006; Tsui et al., 2009; Onoue et al., 2011). To rule out the possibility that PEGylated RBP40 represents a non-imported precursor form of RBP40, we tested the purity of the stromal protein preparation. Localization of the cytoplasmic marker protein NAB1 was followed by immunological means, revealing only minor contaminations of chloroplasts with cytoplasmic material, whereas ∼40% of RBP40 was alkylated (Figure 2c; Mussgnug et al., 2005). These data imply that either RBP40 is imported into the chloroplast by an alternative pathway that does not involve N-terminal processing of proteins or it contains an unusually short transit sequence of <11 amino acid residues (Schwenkert et al., 2011). Both possibilities are compatible with previous in vitro import experiments, which detected no size change in RBP40 after transport into chloroplasts (Barnes et al., 2004).

Figure 2.

 Thiol labelling and RNA binding activity of RBP40.
(a) Schematic representation of the RBP40 polypeptide showing the location of the four repeats involved in RNA binding relative to the single Cys residue (Cys11). The C-terminal end of the putative transit peptide (18 amino acids, Barnes et al., 2004) is indicated by the scissors symbol.
(b) Cys11 alkylation. Indicated amounts of chloroplast proteins (cp) were treated with 10 mm methoxy polyethylene glycol 5000-maleimide (PEG-Mal) for 120 min, and then subjected to immunoblot analysis using an anti-RBP40 antibody and an alkaline phosphatase-coupled secondary antibody.
(c) Immunodetection of RBP40 and the cytoplasmic marker NAB1 in 2 μg of soluble proteins from either whole cells (wc) or chloroplasts (cp).
(d) RNA-binding activity of full-length recombinant RBP40 protein (rRBP40). Aliquots (3 μg) of rRBP40 were pre-treated with 25 mm GSSG, 50 mm GSH or a 50-fold molar excess of NEM, UV-crosslinked to a radiolabeled psbD 5′ UTR probe, and fractionated by SDS–PAGE.

To test whether the redox state of Cys11 directly affects the RNA-binding activity of RBP40, UV crosslinking experiments with full-length recombinant RBP40 (rRBP40) and a psbD 5′ UTR RNA probe were performed under various redox conditions (Figures 2c and S1). Addition of oxidized (GSSG) or reduced (GSH) glutathione to the reaction mixture had no effect on RNA recognition (Figure 2c). Moreover, alkylation with N-ethylmaleimide (NEM) had no significant effect on RNA binding, indicating that binding of RBP40 to RNA is not dependent upon the redox state of Cys11.

Light- and redox-dependent formation of the Nac2/RBP40 complex  We have previously shown that RBP40 forms a complex with the RNA stabilization factor Nac2, and that this interaction specifies recognition of the psbD 5′ UTR by RBP40, as on its own, RBP40 binds to any RNA, at least in vitro (Ossenbühl and Nickelsen, 2000; Barnes et al., 2004; Schwarz et al., 2007). We therefore wished to know whether the interaction with Nac2 is affected by the redox state of Cys11. To this end, we analyzed the distribution of stromal RNA/protein (RNP) complexes by size-exclusion chromatography (SEC; Johnson et al., 2010; Schwarz and Nickelsen, 2010). When wild-type cells were grown in the light, the previously described Nac2/RBP40 complex was identified by both Nac2 and RBP40 antibodies in the range of ∼450–650 kDa (Figure 3a, fractions 6–9; Schwarz et al., 2007). In addition, even larger complexes with a size of approximately 1000 kDa were detected with the RBP40 antibody (Figure 3a, fractions 4 and 5). These latter complexes have not been observed in previous experiments using time-consuming glycerol gradient centrifugation for RNP complex separation, probably because they are relatively labile. As RNA co-immunoprecipitation experiments have previously shown that ribosomal RNA can be precipitated by an RBP40 antibody, these RBP40-specific HMW complexes may represent associations with ribosomes/ribosomal subunits during the initiation phase of translation when Nac2 has already dissociated from the psbD mRNA (Schwarz et al., 2007). This idea is further supported by data revealing that the ribosomal protein S1, and thus at least the small ribosomal subunit, partially co-elutes with these larger RBP40-containing complexes (Figure 3a).

Figure 3.

 Light-dependent formation of Nac2/RBP40 complexes.
Wild-type (a) and mutant (b) cells were grown under the conditions indicated on the left (see Figure 1 for details) and subjected to SEC. Fractioned proteins were subjected to Western blotting and labeled with the antibodies indicated on the right. The sample marked LR was treated with 5 mm glutathione prior to SEC. Fraction numbers and molecular weights are indicated at the top.

Intriguingly, when dark-adapted cells were analyzed, RBP40 accumulated only in the low molecular weight range, peaking at approximately 160 kDa (Figure 3a, fractions 9–15). Concomitantly, the Nac2 signal shifted towards the fractions 8–10, corresponding to a complex of smaller size in the range of ∼400–450 kDa (Figure 3a). This suggests that, in the dark, most of RBP40 is detached from the Nac2 complex, and, as a consequence, psbD mRNA translation is down-regulated. Hence, dynamic formation of the Nac2/RBP40 complex may provide the molecular basis for the observed light-dependent regulation of D2 synthesis (Figure 1). To determine whether formation of this complex is redox-dependent, we performed SEC analysis on RNP complexes from light-grown cells in the presence of reduced glutathione. As shown in Figure 3(a), these reducing conditions resulted in detachment of RBP40 from the Nac2 complex, although the effect was less pronounced than that observed in dark-grown cells (Figure 3a, fractions 8–15). Nevertheless, the data strongly suggest that the redox state does play a critical role in Nac2/RBP40 complex formation.

Moreover, SEC analysis of RNP complexes from the suppressor line suΔU+9 revealed the presence of HMW Nac2/RBP40 complexes, similar to the situation in the wild-type (Figure 3b). This indicates that interaction of RBP40 with its cognate binding site on the psbD 5′ UTR is not a prerequisite for Nac2/RBP40 complex formation, which is consistent with the earlier finding that this complex can form even in a mutant strain lacking the psbD mRNA (Schwarz et al., 2007).

The observed redox control of RBP40 association with the Nac2 complex raises the question of whether photosynthetic electron transport (PET) is directly involved in controlling the synthesis of D2. To assess this, two photosynthetic mutants with defects in either PSII or PSI were examined with regard to formation of Nac2/RBP40 complexes in the light. In the mbb1 mutant, a nuclear factor is mutated that is required for stabilization of the chloroplast psbB mRNA encoding the CP47 subunit of PSII (Vaistij et al., 2000), as Nac2 is required for psbD stability. As shown in Figure 3(b), the distribution of both Nac2 and RBP40 following SEC analysis resembled that found for light-grown wild-type cells, that the absence of PSII does not affect Nac2/RBP40 complex formation per se. In the psaA trans-splicing mutant raa1, PSI is absent, causing severe oxidative stress when cells are grown in the light (Merendino et al., 2006). Under these conditions, partial disassembly of the Nac2/RBP40 complex was observed, as indicated by the shift of both the Nac2 and the RBP40 signal towards lower molecular weights during SEC (Figure 3b). This suggests that although PET is not required for formation of the Nac2/RBP40 complex, it may be involved in its disassembly, which would lead to down-regulation of synthesis of D2 and consequently PSII.

Light- and redox-dependent disulfide bridge formation of Nac2 and RBP40.  The data obtained so far support the idea of a light-dependent control of Nac2/RBP40 complex formation that may involve Cys11 of RBP40. To test this more directly, two-dimensional SDS–PAGE analyses were performed in which stromal proteins were first fractionated by SDS–PAGE in the absence of reducing agents, that is preserving pre-formed disulfide bridges (Ströher and Dietz, 2008), and then orthogonally electrophoresed under reducing conditions. Consequently, polypeptides that contain no disulfide bonds in their native state lie on a diagonal across the second-dimension gel, while intermolecular or intramolecular disulfide bridges cause deviations from the diagonal to the left or right, respectively (Ströher and Dietz, 2008).

When stromal proteins from light-grown wild-type cells were analyzed by following this procedure, some RBP40 was detected on the diagonal, but substantial amounts were also found in a smear in the HMW range up to approximately 170 kDa (Figure 4a). When stromal proteins were pre-treated with reduced glutathione, no such HMW signals were detectable, indicating that RBP40 forms an intermolecular disulfide bridge via its single cysteine Cys11 (Figure 4b). On the other hand, most Nac2 was found on the diagonal at 140 kDa, but minor amounts co-migrated in the range of the RBP40 signal at approximately 170 kDa (Figure 4a). Reduction prior to electrophoresis in the first dimension eliminated this 170 kDa HMW form, confirming that its formation is redox-dependent (Figure 4b). These findings are consistent with the existence of a direct disulfide bridge between Cys11 in RBP40 (40 kDa) and one of the several Cys residues present in Nac2 (140 kDa), causing the signal at approximately 170 kDa (Figure S2). Analysis of the nac2-26 mutant, which fails to accumulate any Nac2 protein, further substantiated this idea: no HMW RBP40 signals were obtained in this strain (Figure 4b). The smear-like RBP40 signal below 170 kDa probably reflects partial breakdown products of the Nac2/RBP40 dimer and/or disulfide bonds of RBP40 to other proteins. If the latter is the case, these interactions are Nac2-dependent, because they are not detected in the nac2-26 mutant background (Figure 4b). Strikingly, HMW RBP40 signals were also lacking when stromal proteins from dark-adapted wild-type cells were assayed, suggesting that the putative disulfide bridge linking Nac2 and RBP40 is reduced in the dark (Figure 4b). Furthermore, the Nac2 signal appeared to be shifted towards lower molecular weight on the right side of the diagonal, suggesting enhanced formation of intramolecular disulfide bridges in Nac2 in the dark (Figure 4b). The transient formation of one or more of such Nac2 internal disulfide bridges probably contributes to the complex Nac2 signal pattern seen in light-grown cells (Figure 4a).

Figure 4.

 Light-dependent formation of a disulfide bridge between Nac2 and RBP40.
(a) Aliquots (100 μg) of stromal proteins from light-grown wild-type cells were fractionated by 2D redox SDS–PAGE, and Nac2 and RBP40 were localized by immunoblot analysis. The diagonal along which polypeptides that form no S–S bonds are expected to lie is indicated.
(b) Sections of 2D redox gels showing immunodetected Nac2 and RBP40 signals after 2D electrophoresis of samples from strains grown under the indicated conditions. The samples marked LR were reduced with 5 mm glutathione prior to electrophoresis in the first dimension. +DBMIB indicates treatment of cells with 10 μm DBMIB for 2 h prior to analysis. The arrows indicate Nac2 material at 170 kDa. For further details, see also Figures 1 and 3.

We also analyzed Nac2/RBP40 disulfide bond formation in the genetic backgrounds used to study complex formation by SEC. In the suppressor line suΔU+9, Nac2 and RBP40 signals at 170 kDa were actually enhanced, indicating efficient binding of RBP40 to Nac2 despite the absence of its cognate binding site on the psbD 5′ UTR (Figure 4b). Furthermore, several additional intermediate RBP40 signals appeared, reflecting the situation in the wild-type. In mbb1, even more enhanced Nac2/RBP40 binding was detected, confirming that the effect seen in the nac2-26 mutant is not due to a deficiency in PSII but is Nac2-specific (Figure 4b). In addition, and in contrast to the wild-type, almost all Nac2 and RBP40 migrated at approximately 170 kDa, suggesting extremely efficient disulfide bridge formation under PSII-deficient conditions and/or fewer internal disulfide bridges in Nac2 (Figure 4b). This situation allowed us to better resolve the complex Nac2 signal pattern seen in the wild-type (Figure 4a). Stromal proteins from mbb1 were reduced by adding 5 mm glutathione, and subsequent electrophoresis showed that not only RBP40 but also Nac2 clearly deviated from the diagonal towards the left, indicating intermolecular disulfide bridge formation in Nac2 (Figure S3).

Finally, reduction of the Nac2–RBP40 disulfide bridge was found to occur in light-grown raa1 cells, suggesting that oxidative stress and/or a reduced plastoquinone (PQ) pool lead to down-regulation of D2 synthesis via the Nac2/RBP40 redox switch (Figure 4b). To differentiate between these two possibilities, wild-type cells were incubated for 2 h with DBMIB (2,5-dibromo-6-methyl-3-isopropyl-1,4-benzoquinone) prior to analysis. This treatment inhibits PET to the cytochrome b6f complex, and thus leads to over-reduction of the PQ pool. Under these conditions, no disulfide bond formation between Nac2 and RBP40 was observed, supporting the idea that the redox status of the PQ pool may be involved in down-regulation of psbD mRNA translation. In conclusion, our data reveal a clear correlation between Nac2/RBP40 complex formation as visualized by SEC analysis and formation of a disulfide bridge between Nac2 and RBP40.

Blue or red light do not affect Nac2/RBP40 complex formation.  In higher plants, both chloroplast transcription and translation have been shown to be regulated by exposure to low levels of blue light (Gamble and Mullet, 1989; Barneche et al., 2006). In particular, psbD gene transcription depends on a blue light-responsive promoter element that is recognized by a specific sigma factor, namely sig5 (Lerbs-Mache, 2011). However, in the chloroplast of C. reinhardtii, only a single sigma factor has been shown to operate, and thus no obvious changes in psbD mRNA levels were observed under the light conditions applied here (Figure 1d; Carter et al., 2004; Bohne et al., 2006). Nevertheless, we tested whether exposure of dark-adapted cells to low-level blue or red light induces Nac2/RBP40 disulfide bridge formation. As shown in Figure 5, irradiation with red or blue for 3 h does not induce formation of the Nac2/RBP40 complex. This suggests that disulfide bridge formation is not dependent on signal relays activated by blue or red light.

Figure 5.

 Blue and red light do not affect Nac2/RBP40 disulfide bridge formation.
Stromal proteins were isolated from dark-adapted wild-type cells that had been exposed to low levels (5 μmol m−2 sec−1) of either blue (BL) or red light (RL) for 3 h, and subjected to 2D redox PAGE. Nac2 and RBP40 proteins were localized by immunoblot analysis.

Chloroplast NTRC may be involved in reduction of the Nac2/RBP40 disulfide bridge in the dark.  The emerging picture of light-dependent regulation of D2 synthesis suggests a central role for the redox state of the Nac2/RBP40 complex. In the light, assembly of this complex requires formation of a disulfide bond between the two proteins. In the dark or under oxidative stress/PQ pool over-reduction, this bond is reduced, and consequently RBP40 is detached from Nac2. While oxidative stress is likely to lead to the previously observed glutathionylation of RBP40 (Michelet et al., 2008), how reduction in the dark can be achieved remains obscure.

One candidate for this role is the recently identified NTRC (chloroplast NADPH-dependent thioredoxin reductase C) enzyme (Serrato et al., 2004). NTRC reduces disulfides in the dark using electrons from NADPH generated by the oxidative pentose phosphate pathway (Kirchsteiger et al., 2009). To test the possibility that NTRC may be involved in redox regulation of the Nac2/RBP40 complex in the dark, 2D redox PAGE was performed using stromal proteins from light-grown cells that had been pre-incubated with recombinant NTRC (rNTRC) from the cyanobacterium Anabena sp. PCC 7120 in the presence of 250 μm NADPH (Figure 6a). Whereas NADPH alone had no effect on the covalent link between Nac2 and RBP40, the disulfide bridge was reduced when the cyanobacterial enzyme was added. In contrast, rNTRC from rice had no effect on the Nac2/RBP40 complex (data not shown), suggesting that specific recognition of the disulfide target has diverged during evolution. Nevertheless, in C. reinhardtii, it does appear that the Nac2/RBP40 complex represents a target for chloroplast NTRC, which may therefore be the enzyme that mediates down-regulation of D2 synthesis in the dark in this species.

Figure 6.

 The Nac2/RBP40 disulfide bridge can be reduced by rNTRC.
(a) Stromal proteins isolated from light-grown wild-type cells were incubated with 250 μm NADPH in the presence or absence of 2 μm rNTRC enzyme from Anabena sp. PCC 7120, and analyzed by 2D redox PAGE. The rNTRC enzyme was prepared as described by Pascual et al. (2011).
(b) A 1:1 mixture of 50 μg each of stromal proteins from dark-adapted and light-grown cells was incubated in the presence or absence of 250 μm NADPH prior to 2D redox PAGE. Nac2 and RBP40 proteins were detected by immunoblot analysis.

To confirm the possibility that dark-grown cells contain an activity that reduces the Nac2/RBP40 disulfide bridge, we mixed stromal protein extracts from light- and dark-grown wild-type cells in a 1:1 ratio. When this mixture was assayed, a drastic decrease in the level of the Nac2/RBP40 complex was observed that cannot be explained by a dilution effect (Figure 6b). In the presence of 250 μm NADPH, this effect was even more pronounced, strongly suggesting that dark-adapted chloroplasts from C. reinhardtii contain an activity, probably the algal NTRC, that severs the link between Cys11 of RBP40 and Nac2, and thereby down-regulates psbD gene expression. In addition, these data support the idea that chloroplast NTRC is less active when cells are grown in the light.

Discussion

We have previously postulated that the translational activator RBP40, together with an RNA stem-loop structure encompassing the AUG start codon, form a molecular switch with the capacity to regulate chloroplast psbD gene expression (Klinkert et al., 2006). Here, we demonstrate that this molecular switch is indeed involved in controlling D2 synthesis in a light-dependent manner (Figure 7). Suppressor lines in which the requirement for RBP40 is bypassed exhibited reduced levels of light-induced D2 synthesis (Figure 1b). This suggests light-dependent change of the RNA structure by RBP40. We have previously shown that recruitment of RBP40 by the Nac2 complex specifies its interaction with the psbD 5′ UTR (Schwarz et al., 2007). Intriguingly, this recruitment process and the subsequent formation of a Nac2/RBP40 complex is light-dependent, and thus most likely forms the critical step during dark/light transitions in the pattern of psbD gene expression (Figure 3b).

Figure 7.

 Working model for redox regulation of psbD gene expression.
In the light, RBP40 binds to Nac2 via an intermolecular disulfide bridge, and, as a consequence, the RNA conformation at the AUG start codon is altered. This allows ribosome access to the initiation site and enables efficient translation of psbD mRNA. In the dark, the disulfide bridge between Nac2 and RBP40 is reduced via NTRC, leading to detachment of RBP40 from Nac2 and down-regulation of D2 synthesis.

Moreover, the data strongly suggest a direct interaction of RBP40 with Nac2 via a light-dependent disulfide bridge involving the single Cys residue in RBP40 at position 11. This conclusion, is based on the findings that parts of both Nac2 and RBP40 co-migrate at 170 kDa in a redox-dependent manner, and the 170 kDa signal is absent in the dark and in the nac2-26 mutant but present in an unrelated PSII-deficient mutant, that is mbb1. Furthermore, in mbb1, a clear increase of Nac2 and RBP40 signal intensity at 170 kDa is observed when comparing the mbb1 mutant to the wild-type. However, it remains to be shown what causes the RBP40 signals between 40 kDa and 170 kDa in the wild-type and the suppressor suΔU+9 (Figure 4b). It should be noted that Nac2 is a rather unstable protein that rapidly degrades upon storage, and consequently all experiments were performed with freshly prepared material. As the αNac2 antibody was generated against the C-terminal 60 amino acids of Nac2, only a limited set of transiently formed breakdown products are detected. Hence, it is possible that the intermediate RBP40 signals represent dimers with partially degraded Nac2. Alternatively, under certain conditions, RBP40 may form disulfide connections in a Nac2-dependent manner to other, as yet to be identified, subunits of the Nac2 complex that are smaller than Nac2. Future identification of such subunits is required to unambigously answer if these RBP40 signals between 40 kDa and 170 kDa are caused by degradation products or additional interaction partners of RBP40.

Nevertheless, the data support the possibility that at least parts of Nac2 and RBP40 form a disulfide bridge and that the redox state of Cys11 of RBP40 is the main target for the light control mechanism. To date, no RBP40 knockout mutant lines are available, which hampers site-directed genetic approaches to confirm the role of Cys11 in vivo. However, the redox state of Cys11 apparently has no direct influence on the RNA-binding activity of RBP40, which is consistent with the localization of Cys11 in the N-terminal segment of RBP40, relatively remote from its predicted RNA-binding domain, which starts at position 39 (Figure 2a; Barnes et al., 2004). This RNA-binding domain is made up of four conserved repeats, each spanning 70 amino acids, and is structurally related to other RNA-binding domains of the RBD or KH type (Figure 2a; Barnes et al., 2004). In agreement with this, a truncated version of RBP40 lacking the first 18 N-terminal amino acids (including Cys11) has been shown to retain general RNA-binding activity (Barnes et al., 2004).

The question arises as to which Cys residue in the Nac2 protein may interact with Cys11 of RBP40. Nac2 encodes a total of 11 Cys residues at various positions, some of which are likely to form intramolecular disulfide bonds, as suggested by the 2D redox PAGE analyses (Figure 4). Interestingly, the most probable disulfide bridge is bioinformatically predicted to be formed between Cys981 and Cys1008, which are both located within the tetratricopeptide repeat (TPR) domain of Nac2 (Figure S2). This domain has previously been shown to play a critical role for Nac2 function, probably by mediating the interaction with other subunits of the Nac2 complex (Boudreau et al., 2000). In addition, a putative dinucleotide recognition domain is predicted at position 402–413 of Nac2, which may be involved in modulation of the redox state of one or more of its Cys residues. However, only a systematic evaluation of these sites will uncover the residue that forms the link with RBP40.

The most interesting question concerns the light-mediated redox switch at Cys11. The current models for redox control of chloroplast gene expression usually involve light-catalyzed reduction processes that are linked to PET via either PSI and thioredoxin or the redox state of the PQ pool (Barnes and Mayfield, 2003). In case of Nac2/RBP40 complex formation, PET is apparently not required as Cys11 oxidation is also observed in the PSII mutant mbb1. This is consistent with previously measured wild-type levels of D2 synthesis rates in this mutant (Vaistij et al., 2000). However, both PSI deficiency and a reduced PQ pool obviously cause Nac2/RBP40 complex disintegration via RBP40 Cys11 reduction, and result in shutdown of de novo PSII synthesis, thereby avoiding harmful photosynthetic electron overflow. The reductive detachment of RBP40 from Nac2 under oxidative stress conditions is likely to be mediated via glutathionylation of RBP40 as has reported previously (Michelet et al., 2008; Figure 4b). How the redox state of the PQ pool is communicated to Cys11 of RBP40 remains to be determined.

However, in the dark, a different regulatory redox system appears to operate on the Nac2/RBP40 disulfide bridge, namely the NTRC system. This system uses NADPH, produced during darkness by the oxidative pentose phosphate pathway, as source of reducing power (Neuhaus and Emes, 2000). Thus it was proposed that NTRC enables redox regulation in the chloroplast during the night. This possibility is supported by the hypersensitivity of the Arabidopsis NTRC knockout mutant to prolonged darkness (Pérez-Ruiz et al., 2006). Recently, regulation of the ADP-Glc pyrophosphorylase involved in starch synthesis in Arabidopsis thaliana was shown to be mediated by NTRC (Michalska et al., 2009). Our data suggest that this enzyme, which is active in the dark, is also involved regulation of chloroplast psbD gene expression, at least in C. reinhardtii. Thus, NTRC directly links the regulation of chloroplast gene expression to carbon metabolism in the chloroplast, that is the oxidative pentose phosphate pathway.

In conclusion, the following scenario is postulated to describe the molecular events that underlie light-dependent regulation of D2 synthesis (Figure 7). In the dark, reduction of the disulfide bond between Nac2 and RBP40 via NTRC leads to detachment of RBP40 from the Nac2 complex, resulting in down-regulation of D2 synthesis. In the light, psbD mRNA translation is activated by RBP40, which is tightly bound via its Cys11 residue to Nac2, and is thereby targeted to its cognate binding site within the psbD 5′ UTR. RBP40 binding then alters the RNA conformation at the initiation codon, making it accessible to the translation machinery.

Although the role of NTRC in dark regulation of psbD mRNA translation is supported by the data presented here, the enzymatic components as well as the final electron acceptor during light-dependent formation of the disulfide bridge between Nac2 and RBP40 remain elusive. As discussed by Wittenberg and Danon (2008), indirect evidence strongly suggests that an efficient system for disulfide bond formation must exist in the stroma of chloroplasts. This system is likely to be related to the well-characterized systems in the periplasm of prokaryotes, the endoplasmic reticulum or the mitochondrial intermembrane space of eukaryotes, but its constituents as well as its working mode remain to be determined (Mesecke et al., 2005; Sevier and Kaiser, 2006). It is assumed that thiol oxidases and protein disulfide isomerases play a primary role in catalyzing specific disulfide bond formations, and that reactive oxygen species, GSSG and O2 serve as final electron acceptors. In a recent study, a thiol oxidase from chloroplasts of A. thaliana, named LTO1, was identified that is localized to the thylakoid lumen, where it introduces disulfide bonds into the PsbO subunit of PSII (Karamoko et al., 2011). Thus, LTO1 cannot be involved in regulating stromal chloroplast gene expression. Obviously, more knowledge and tools, including genetic resources such as mutants in redox regulation, are required to clarify this issue. Nevertheless, the data presented here suggest that the redox state of the Nac2/RBP40 disulfide bridge represents the key control point for regulation of D2 synthesis, which, given the CES principle of PSII assembly, is the key player in determining PSII levels in the green alga C. reinhardtii (Minai et al., 2006).

Experimental Procedures

Strains and culture conditions

Chlamydomonas reinhardtii strains were grown under continuous light (30 μE m−2 sec−1) at 23°C in Tris-acetate-phosphate medium containing 1% sorbitol (TAPS; Gorman and Levine, 1965). For dark adaption, cells were transferred to complete darkness for 38 h prior to analysis.

Analysis of nucleic acids

Whole-cell RNA was prepared using TriReagent (Sigma-Aldrich, http://www.sigmalaldrich.com) according to the manufacturer’s instructions, and 2 μg aliquots were electrophoretically fractionated on gels, blotted onto positively charged nylon membranes, and hybridized to atpB and psbD probes. Probes were generated by PCR using DIG-11-dUTP (digoxigenin-11-desoxyuridine-5′-triphosphate, Roche Diagnostics, http://www.roche-applied-science.com) and primers specific to the respective target genes (psbD, 5′-GTAATACGACTCACTATAGGGCCACAATGATTAAAATTAAA-3′ and 5′-GTTGGTGTCAACTTGGTGG-3′; atpB, 5′-ATGTTGTCCAGCGTGCGC-3′ and 5′-TTACTTCTTGGGCAGGAG-3′). Visualization of hybridization signals was performed by enhanced chemiluminescence using alkaline phosphatase-conjugated anti-DIG-antibody and CDP* substrate (disodium 4-chloro-3-(methoxyspiro {1,2-dioxetane-3,2′-(5′-chloro)tricyclo [3.3.1.13,7]decan}-1-4-yl)phenyl phosphate, Roche Diagnostics).

Pulse labeling of proteins

Chlamydomonas liquid cultures were grown in TAPS medium to a density of approximately 2 × 106 cells ml−1, pelleted (1000 g, 5 min, 23°C), resuspended in TAPS medium in which all sulfur-containing ingredients were replaced by the respective chloride salts (TAPS-S), and incubated for 16 h at 23°C in the dark or light. Cells were pelleted (1000 g, 5 min, 23°C), washed with TAPS-S/-T (lacking both sulfur salts and trace elements), and resuspended in TAPS-S/-T and grown in the dark or light for 2 h. Cells were then washed again, and resuspended in TAPS-S/-T to a concentration of 80 μg chlorophyll ml−1. Aliquots (225 μl) of the cell suspension were incubated with cycloheximide (10 μg ml−1) for 10 min. Subsequently, 100 μCi H235SO4 (Hartmann Analytic, http://www.hartmann-analytic.de) was added to each aliquot, followed by incubation for 15 min under the same light conditions as before. After centrifugation (1000 g, 5 min, 23°C), sedimented cells were frozen in liquid nitrogen. Cells were resuspended in 10 mm HEPES/KOH pH 7.5, 10 mm EDTA in the presence of CompleteMini protease inhibitor cocktail tablets (Roche), and disrupted by repeated pipetting. The homogenate was then centrifuged at 20 000 g for 30 min. The pellet was resuspended in 10 mm HEPES/KOH pH 7.5, 10 mm EDTA. Samples were fractionated by by SDS–PAGE in 16% gels containing 8 m urea.

UV crosslinking of RNA to recombinant RBP40

For expression of recombinant RBP40 protein, the DNA sequence encoding amino acids 1–382 was PCR-amplified from a cDNA clone using the primer pair BamHI-RBP40 (5′- aaggatccATGCTGACCTTGAGACGTGC-3′) and RB38-DN44revSalI (5′- ttgtcgacCTAGTAGCGGGCGCCC-3′), and inserted into the plasmid pQE30 (Qiagen, http://www.qiagen.com) via BamHI/SalI restriction sites. Lower case letters indicate addition of active sites, i.e. restriction sites for cloning, through primer design. Protein expression in Escherichia coli M15 blue cells (Stratagene, http://www.stratagene.com) was induced by addition of isopropyl thio-β-d-galactoside to a final concentration of 1 mm, followed by growth at 37°C for 3 h. The recombinant protein was purified according to the manufacturer’s protocol for purification of histidine-tagged recombinant proteins under native conditions using Ni-Sepharose 6 Fast Flow (GE Healthcare, http://www.gelifesciences.com). In preparation for the binding reactions described below, the purified protein was incubated for 1 h at room temperature with 25 mm GSSG, 50 mm GSH, or a 50-fold molar excess of NEM, respectively, in a buffer containing 100 mm HEPES/KOH pH 7.8, 25 mm MgCl2 and 300 mm KCl, followed by desalting using Amicon Ultra centrifugal filtration devices (Millipore, http://www.millipore.com) with a 10 kDa molecular weight cut-off, according to the manufacturer’s instructions.

The DNA template for in vitro synthesis of the psbD RNA probe was generated by PCR using T7psbD5 (5′-gtaatacgactcactatagggCCACAATGATTAAAATTAAA-3′; T7 RNA polymerase promoter in lower-case letters) and psbDUTR3 (5′-ACCGATCGCAATTGTCAT-3′) as primers. RNA synthesis was catalyzed by T7 RNA polymerase (Fermentas, http://www.fermentas.de) in the presence of [α-32P]UTP (3000 Ci mmol−1; Hartmann Analytic), according to the manufacturer’s protocol (Fermentas). After removal of the template by treatment with DNase I (Promega, http://www.promega.com), the RNA was extracted using phenol/chloroform and precipitated using ethanol in the presence of ammonium. Binding reactions (20 μl) were performed at room temperature for 5 min, and contained 500–1000 kcpm of 32P-labeled RNA probe, 20 mm HEPES/KOH pH 7.8, 5 mm MgCl2, 60 mm KCl and 3 μg pre-treated protein. After irradiation, the free RNA probes were digested by treatment with 10 units RNase One (Promega) for 30 min at 37°C, and the samples were fractionated by SDS–PAGE and analyzed by phosphorimaging.

Gel filtration analysis of native proteins

For analysis of native protein complexes, chloroplasts were isolated from cw15 strains as described by Zerges and Rochaix (1998), and lysed in non-reducing breaking buffer (10 mm EDTA, 10 mm Tricine/KOH pH 7.5, and Roche CompleteMini protease inhibitor cocktail tablets). Membrane material was removed by centrifugation on a 1 m sucrose cushion (100 000 g, 30 min, 4°C). Reducing conditions were achieved by adding 5 mm GSH to the stroma-containing supernatant prior to concentration using Amicon Ultra filtration devices (Millipore). Samples (approximately 2 mg protein) were loaded through an SW guard column onto a 2.15 × 30 cm G4000SW column (Tosoh, http://www.separations.eu.tosohbioscience.com), and elution was performed using gel filtration buffer (50 mm KCl, 5 mm MgCl2, 5 mmε-aminocaproic acid, 20 mm Tricine/KOH pH 7.5) at a flow rate of 2 ml min−1 (Johnson et al., 2010). All steps were performed at 4°C.

Diagonal 2D redox SDS–PAGE

Stromal proteins (100 μg) from cw15 strains were isolated as described by Zerges and Rochaix (1998) in the absence of reducing agents. To prevent thiol reoxidation, proteins were alkylated with 0.1 m iodoacetamide in the dark (15 min at 4°C). An appropriate volume of non-reducing Laemmli buffer was added, and the samples were separated by SDS–PAGE in the first-dimension resolving gel (10% polyacrylamide). After electrophoresis, gel lanes were excised and incubated in SDS running buffer containing 0.1 m dithiothreitol (10 min at room temperature), before incubation with 0.1 m iodoacetamide in the same buffer (10 min, room temperature). The gel strips were then horizontally applied to a polyacrylamide gel containing 10% SDS, and electrophoresis was performed in the second dimension (Ströher and Dietz, 2008; Stengel et al., 2009). Gels were blotted onto supported nitrocellulose membranes, and diagonals of non-disulfide bond-forming proteins were marked on the blots after Ponceau red staining. Immunodetection by enhanced chemiluminescence was performed using antibodies raised against Nac2 and RBP40 (Schwarz et al., 2007).

PEGylation

Indicated amounts of proteins (0.3–3.0 μg) were treated with 10 mm methoxypolyethyleneglycol-maleimide (5 kDa; Laysan, http://www.laysanbio.com) in alkylation buffer (120 mm Tris, pH 6.8, 3.8% w/v SDS, 10 mm methoxypolyethyleneglycol-maleimide and 10 mm EDTA) for 120 min at room temperature in the dark (Wobbe et al., 2009). The reaction was stopped by addition of Laemmli buffer in the presence of 0.1 m dithiothreitol (Balsera et al., 2009). Bis(2-hydroxy-ethyl)amino-tris(hydroxymethyl)methane-SDS–PAGE (10%) and MOPS running buffer were used to separate the proteins. RBP40 was detected by immunoblot analysis using an alkaline phosphatase-coupled secondary antibody.

Acknowledgements

We thank K. Findeisen for skilled technical assistance and M. Goldschmidt-Clermont (Laboratory of Chloroplast Molecular Genetics, Department of Botany and Plant Biology, University of Geneva) for providing the raa1 mutant. This work was supported by a grant from the Deutsche Forschungsgemeinschaft to J.N. (grant number Ni390/4-2). Work in F.J.C.’s laboratory was supported by European Regional Development Fund co-financed grants from the Spanish Ministry of Science and Innovation (BIO2010-15430) and the Junta de Andalucía (BIO-182 and CVI-5919).

Ancillary