Intertwined translational regulations set uneven stoichiometry of chloroplast ATP synthase subunits
2007; Springer Nature; Volume: 26; Issue: 15 Linguagem: Inglês
10.1038/sj.emboj.7601802
ISSN1460-2075
AutoresDominique Drapier, Blandine Rimbault, Olivier Vallon, Françis-André Wollman, Yves Choquet,
Tópico(s)Photosynthetic Processes and Mechanisms
ResumoArticle26 July 2007free access Intertwined translational regulations set uneven stoichiometry of chloroplast ATP synthase subunits Dominique Drapier Dominique Drapier UMR 7141 CNRS/UPMC, Institut de Biologie Physico-Chimique, Paris, France Search for more papers by this author Blandine Rimbault Blandine Rimbault UMR 7141 CNRS/UPMC, Institut de Biologie Physico-Chimique, Paris, France Search for more papers by this author Olivier Vallon Olivier Vallon UMR 7141 CNRS/UPMC, Institut de Biologie Physico-Chimique, Paris, France Search for more papers by this author Francis-André Wollman Francis-André Wollman UMR 7141 CNRS/UPMC, Institut de Biologie Physico-Chimique, Paris, France Search for more papers by this author Yves Choquet Corresponding Author Yves Choquet UMR 7141 CNRS/UPMC, Institut de Biologie Physico-Chimique, Paris, France Search for more papers by this author Dominique Drapier Dominique Drapier UMR 7141 CNRS/UPMC, Institut de Biologie Physico-Chimique, Paris, France Search for more papers by this author Blandine Rimbault Blandine Rimbault UMR 7141 CNRS/UPMC, Institut de Biologie Physico-Chimique, Paris, France Search for more papers by this author Olivier Vallon Olivier Vallon UMR 7141 CNRS/UPMC, Institut de Biologie Physico-Chimique, Paris, France Search for more papers by this author Francis-André Wollman Francis-André Wollman UMR 7141 CNRS/UPMC, Institut de Biologie Physico-Chimique, Paris, France Search for more papers by this author Yves Choquet Corresponding Author Yves Choquet UMR 7141 CNRS/UPMC, Institut de Biologie Physico-Chimique, Paris, France Search for more papers by this author Author Information Dominique Drapier1, Blandine Rimbault1, Olivier Vallon1, Francis-André Wollman1 and Yves Choquet 1 1UMR 7141 CNRS/UPMC, Institut de Biologie Physico-Chimique, Paris, France *Corresponding author. UMR 7141 CNRS/UPMC, Institut de Biologie Physico-Chimique, 13, rue Pierre et Marie Curie, Paris 75005, France. Tel.: +33 1 58 41 50 75; Fax: +33 1 58 41 50 22; E-mail: [email protected] The EMBO Journal (2007)26:3581-3591https://doi.org/10.1038/sj.emboj.7601802 PDFDownload PDF of article text and main figures. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info The (C)F1 sector from H+-ATP synthases comprises five subunits: α, β, γ, δ and ε, assembled in a 3:3:1:1:1 stoichiometry. Here, we describe the molecular mechanism ensuring this unique stoichiometry, required for the functional assembly of the chloroplast enzyme. It relies on a translational feedback loop operating in two steps along the assembly pathway of CF1. In Chlamydomonas, production of the nucleus-encoded subunit γ is required for sustained translation of the chloroplast-encoded subunit β, which in turn stimulates the expression of the chloroplast-encoded subunit α. Translational downregulation of subunits β or α, when not assembled, is born by the 5′UTRs of their own mRNAs, pointing to a regulation of translation initiation. We show that subunit γ, by assembling with α3β3 hexamers, releases a negative feedback exerted by α/β assembly intermediates on translation of subunit β. Moreover, translation of subunit α is transactivated by subunit β, an observation unprecedented in the biogenesis of organelle proteins. Introduction Energy transduction in mitochondria or chloroplasts is performed by membrane-associated oligomeric proteins that comprise subunits of dual genetic origin. Many subunits are expressed in the nucleo-cytosol before being imported into the organelle, but a subset of subunits is organelle-encoded and therefore translated next to the membranes, where functional assembly of the proteins takes place. Studies of chloroplast proteins in the unicellular green alga Chlamydomonas reinhardtii, or of mitochondrial proteins in the yeast Saccharomyces cerevisiae, have disclosed a unique trait in the biogenesis of these multimeric enzymes: the expression of several organelle-encoded subunits is regulated at the translational step in an assembly-dependent manner. We have termed this process Control by Epistasy of Synthesis (CES), whereby the presence of one subunit is required for sustained synthesis of another organelle-encoded subunit from the same protein complex (for reviews see Wollman et al, 1999; Choquet and Vallon, 2000; Choquet and Wollman, 2002). In all instances studied so far, the CES process involves transmembrane subunits that assemble in a 1 to 1 stoichiometry in the final protein. In these cases, the molecular basis for this regulation has been consistently identified as a negative feedback of the unassembled subunit on the initiation of its translation. By contrast with the major subunits of photosystem I, photosystem II, cytochrome bc/b6f complexes and cytochrome oxidase, which are integral membrane proteins, the catalytic sector of the proton ATP synthase, F1, which is responsible for the reversible synthesis of ATP is made of extrinsic subunits that assemble in a hydrophilic environment (Abrahams et al, 1994). The other sector, F0, that behaves as a selective proton channel, is membrane embedded and assembles within the lipid bilayer (Deckers-Hebestreit and Altendorf, 1996). A unique feature of the biogenesis of the ATP synthase is the requirement for an uneven stoichiometry between subunits in both F1 and F0. In F1, 3 copies each of the α and β subunits assemble with only one copy of the γ, δ and ε subunits, whereas 10–14 copies of suIII (named subunit c in bacteria) assemble with 1 copy of subunits I, II and IV in F0. In bacteria and mitochondria, the two sectors may assemble independently. In Escherichia coli, a functional F0 is inserted into the membrane in the absence of F1 synthesis (Aris et al, 1985; Fillingame et al, 1986), while mutants lacking F0 accumulate a functional F1 sector, showing ATPase activity (Klionsky and Simoni, 1985). Mitochondrial F0 contains organelle-encoded subunits, whereas mitochondrial F1 is made exclusively of nuclear-encoded subunits that may assemble independently of the presence of F0 (for a review see Ackerman and Tzagoloff, 2005). In marked contrast, both CF0 and CF1 are made of subunits of mixed genetic origin in chloroplasts (Lemaire and Wollman, 1989a; for a review see Strotmann et al, 1998). SuI, III and IV from CF0 are chloroplast-encoded, but SuII is nucleus-encoded. CF1 subunits α, β and ε are expressed from the chloroplast genes atpA, atpB and atpE, respectively, while subunits γ and δ are expressed from the nuclear genes ATPC and ATPD. The biogenesis of the chloroplast ATP synthase thus requires control mechanisms to ensure not only these specific and uneven stoichiometries between its constitutive subunits, but also a cross-talk between two distinct genetic compartments. Studies of ATP synthase mutants of C. reinhardtii have shown that defects in the expression of any of its constitutive subunits lead to a pleiotropic loss in most polypeptides, from both CF0 or CF1 (Lemaire and Wollman, 1989b). Thus, assembly of the chloroplast ATP synthase is a concerted process. These studies also pointed to a tight coupling between the rates of synthesis of several subunits: mutants lacking subunit β display a markedly decreased synthesis of subunit α, whereas mutants defective in the expression of subunit α show some increase in synthesis of subunit β (Lemaire and Wollman, 1989b; Drapier et al, 1992). Whether interactions between rates of synthesis of the two major CF1 subunits α and β play a role in their three time accumulation over that of their assembly partners in CF1 has not been investigated. In the present work, we used reporter genes translated under the control of the 5′ untranslated regions (UTRs) of atpA and atpB to demonstrate that the final α3β3γ stoichiometry required for functional assembly of CF1 indeed results from an assembly-dependent control of translation initiation of the α and β subunits. In addition we report here the first instance where one subunit (subunit β) activates in trans the expression of its assembly partner, subunit α. Results Coordinated synthesis of CF1 subunits A convenient way to assess the possible effect of the state of assembly of CF1 on the rates of synthesis of its constitutive subunits is to examine how mutations specifically preventing the expression of one of the main CF1 subunits in C. reinhardtii affect the expression of other subunits. To this end, we generated a ΔatpA deletion strain (Table I; Figure 1). We also used the Fud50 chloroplast mutant strain, hereafter referred to as ΔatpB, lacking expression of subunit β because of a partial deletion of the atpB gene (Woessner et al, 1984; Figure 1), and the nuclear mutant strain atpC1 defective for the expression of subunit γ as it lacks ATPC mRNA (Smart and Selman, 1991; Figure 1B and D). Figure 1.Impaired CF1 assembly feeds back on the expression of subunits α and β. (A) Synthesis of subunits α and β, as determined in a representative 5 min pulse-labelling experiment in the wild type and in mutant strains defective for CF1 assembly. Quantification of the levels of synthesis—normalised to that of the four major PSII subunits (apoCP47, apo-CP43, D1 and D2) to correct from variations in 14C uptake and incorporation—are shown below. Note that, at variance with previous reports (Lemaire and Wollman, 1989b; Drapier et al, 1992), cycloheximide and 14C-acetate were added simultaneously. Positions of subunits α, β and cytochrome f are shown, as well as those of two subunits from PSII: apo-CP47, -CP43, and that of EFTu. (B) Accumulation of subunits α, β and γ in the same strains, detected with specific antibodies. In the bottom panel, immunoblots reacted with antibody against subunit γ were overexposed in order to detect residual γ in strains ΔatpA and ΔatpB; these conditions allow detection of a cross-reaction with mitochondrial subunit γ. Cytochrome f, subsequently used as a reporter, displays similar expression levels in all strains. OEE2, loading control. (C) Accumulation of atpA and atpB transcripts in the same strains. petA transcript, loading control. (D) Accumulation of transcripts for ATPC and petA (as a loading control) in the wild-type and atpC1 strains. Download figure Download PowerPoint Table 1. Transformation experiments Transformantsa Recipient strainsa Transforming plasmid ΔatpAb WT pΔatpA∷Kr βTrb WT patpB335StKr dBf WT pKdBf {βTr, dAf}b dAf patpB335StKr atpC1 {dBf} atpC1 pKdBf atpC1 {ΔatpA} atpC1 pΔatpA∷Kr atpC1 {βTr}b atpC1 patpB335StKr {βTr, dBf} βTrc pKdBf atpC1 {βTr, dBf} atpC1 {βTr}c pKdBf {ΔatpA, dAf} dAf pΔatpA∷Kr {ΔatpA, βTr, dAf} {βTr, dAf}c pΔatpA∷Kr αTr WT pKratpA300St {αTr, dAf}d dAf pKratpA300St {αTr, βTr, dAf}d {βTr, dAf}c pKratpA300St aAαe WT pKaAα {βTr, aAα}e βTrc pKaAα All recipient strains were sensitive to spectinomycin. Transformants were selected for resistance to spectinomycin (100 μg ml−1). a Strains are named by their genotype. By convention, the chloroplast genotype is indicated between accolades for strains containing more than one mutation and follows, when required, the nuclear genotype. b These strains were initially selected for spectinomycin resistance due to the presence of the recycling spectinomycin resistance cassette (Kr). Once homoplasmic with respect to the ATP synthase mutation, they were grown on TAP medium for several generations to allow the spontaneous loss of the recycling cassette, according to Fischer et al (1996), but not that of the mutated ATP synthase allele. c They, therefore, became spectinomycin sensitive again and could be used as a recipient strain in a new round of transformation experiments based on selection for spectinomycin resistance. d These strains are shown in the right panel in Figure 7C and D, but are described in Supplementary Figure S1 and related text. e These strains were used to obtain the results presented in Table II, but they are described in Supplementary Figure S2 and related text. Translation of subunits α and β was assessed in wild-type and mutant strains by 5 min pulse-labelling experiments by adding 14C-acetate simultaneously with cycloheximide, an inhibitor of cytosolic translation (Figure 1A). In the wild-type strain, the rates of synthesis of subunits α and β were similar. In strain ΔatpA, synthesis of subunit β showed some upregulation. In strain ΔatpB, synthesis of subunit α was considerably decreased, as previously reported (Lemaire and Wollman, 1989b; Drapier et al, 1992), and its accumulation was barely detectable. This behaviour qualifies subunit α as a CES subunit (Wollman et al, 1999; Choquet and Wollman, 2002) requiring the presence of subunit β to be expressed at a significant rate. In strain atpC1, translation and accumulation of subunit β were drastically decreased. Subunit β, therefore, is also a CES protein whose rate of synthesis depends on the presence of subunit γ. We note that subunit α, despite its inability assemble with subunit γ, resisted proteolytic degradation, as did subunit β in the absence of subunit α: both accumulated to significant amounts in exponentially growing cells from strains atpC1 and ΔatpA, respectively (Figure 1B). The decreased synthesis of the CES subunits, β in the absence of γ, or α in the absence of β, could not be accounted for by the limited changes in the amount of atpA or atpB mRNAs (Figure 1C), suggesting a translational or early post-translational regulation. The atpB-5′UTR confers a γ-dependent translation to a reporter protein We investigated the mechanism by which the nucleus-encoded γ subunit regulates synthesis of the chloroplast-encoded β subunit, by testing whether it resulted from a regulation of translation initiation. To this end, we constructed a chimeric gene, made of the atpB promoter and 5′UTR regions fused to the petA coding sequence for cytochrome f, previously shown to be a convenient reporter gene (Wostrikoff et al, 2004; Minai et al, 2006). This chimeric gene (5′atpB-petA) was associated with an aadA spectinomycin resistance cassette (Goldschmidt-Clermont, 1991) to allow the selection of transformed cells (Figure 2A and Table I). In order to compare the expression of this chimera in strains expressing, or not, the γ subunit, this construct was introduced by biolistic transformation at the petA locus on the chloroplast genome of wild-type and atpC1 strains. Spectinomycin resistant transformants, hereafter referred to as dBf or atpC1 {dBf}—because they express a complex d (ATP synthase) 5′atpB driven cytochrome f -, were recovered from both strains. Transformants dBf were capable of phototrophic growth and expressed the cytochrome f reporter at a similar level as endogenous cytochrome f in the wild type (Figure 2B and C). By contrast, transformants atpC1 {dBf}, lacking subunit γ, displayed a strong decrease in synthesis of the cytochrome f reporter (Figure 2B) leading to the accumulation of only 5% of cytochrome f in strain atpC1 {dBf} as compared to strain dBf (Figure 2C). Figure 2.The CES behaviour of subunit β corresponds to a translational regulation mediated by the atpB 5′UTR. (A) Map of the chloroplast petA gene in wild-type and dBf strains. Arrows indicate transcription start sites. K stands for the aadA cassette, in opposite orientation with respect to petA. (B) Chloroplast translates in strains dBf and atpC1 {dBf} (2 and 3 independent transformed strains are shown, respectively) that express the reporter cytochrome f. WT and atpC1 strains expressing the regular petA gene are shown as control. The positions of cytochrome f, subunits α and β, CP43 and CP47 are indicated. Levels of synthesis of the cytochrome f reporter, normalised to those of apo-CP43, are indicated. (C) Accumulation of α, β, γ subunits, cytochrome f and OEE2 (loading control). (D) Accumulation of transcripts for petA, either endogenous or chimeric (slightly shorter) and atpB (loading control). (E) Distribution of psbD and atpB mRNAs on sucrose gradients in WT and atpC1 strains. Polysomes are recovered in the pellet and fractions 1–5, while free mRNAs and dissociated 50S and 30S ribosome subunits are recovered in fractions 6–10, as described in (Minai et al, 2006). psbD mRNA serves as a control for a chloroplast transcript unrelated to CF1 biogenesis. Download figure Download PowerPoint The reduced expression of the cytochrome f reporter did not result from a reduced accumulation of its chimeric mRNA in strain atpC1{dBf} (Figure 2D). Thus, the decreased synthesis of subunit β in the absence of its assembly partner—subunit γ- is governed by the atpB 5′UTR and is most likely controlled at the level of translation initiation. This conclusion was further substantiated by the markedly decreased binding of atpB mRNA to polysomes (fractions P-5) in the atpC1 mutant: less than 10% of total atpB mRNA compared to 40% in the wild type (Figure 2E). Subunit γ does not act as a transactivator in translation of subunit β Down-regulation of atpB translation in the absence of subunit γ could be explained in either of two ways: the presence of subunit γ could contribute to activate translation of subunit β (transactivation hypothesis) or the failure to assemble subunit β within CF1 in absence of subunit γ could lead to a translational repression of subunit β (autoregulation hypothesis). Since the atpB-5′UTR conferred the γ-sensitivity to the translation of chimeric genes, these two hypotheses could be discriminated by looking at the expression of a reporter, here the 5′atpB-driven cytochrome f, in the absence of both the β and γ subunits. In the trans-activation hypothesis, the absence of the positive regulator should result in a poor expression of the chimeric construct, irrespective of the presence of subunit β. By contrast, in the autoregulation hypothesis, a strain lacking subunit β should show a high expression of the chimera, even in absence of subunit γ. Still, strains lacking subunit β because of the absence of the endogenous atpB mRNA—such as ΔatpB—may also show higher expression of 5′atpB-driven reporter genes because of an increased availability in specific translational activators normally bound to the regular atpB transcript in the wild type (Barkan and Goldschmidt-Clermont, 2000). To avoid such an ambiguity, we substituted the atpB gene from wild-type and atpC1 strains by a mutated version encoding a truncated β subunit, that allows accumulation and translation of the atpB mRNA but prevents accumulation of the β subunit. We introduced a stop codon in the atpB coding sequence, at position+335 with respect to the initiation codon (Figure 3A and M&M section), associated with the recycling spectinomycin resistance cassette (Fischer et al, 1996) to enable the selection of transformants on TAP-spectinomycin plates. In pulse-labelling experiments, the resulting βTr and atpC1 {βTr} strains (Table I) show no neosynthesised subunit β, but the presence of a very short-lived (t1/2: <10 min) truncation product with an apparent molecular weight of ∼25 kDa (see Figure 3B for strain βTr). Consequently, its accumulation remained below detection (Figure 3C), although a translatable atpB mRNA accumulates normally (Figure 3D). Figure 3.The {βTr} strain expresses a truncated and short-lived subunit β. (A) Strategy used to introduce a mutation (St) in the atpB gene, associated with the recycling spectinomycin resistance cassette, schematically depicted by (not to scale). Arrow indicates transcription start site. Relevant restriction sites are shown (K: KpnI, C: ClaI, N: NruI), as well as the position of the chloroplast inverted repeat (hatched arrow). Coding sequences are indicated in grey. Due to the St mutation, about one third of atpB mRNA is not translated in the mutant (white rectangle). (B) Expression of truncated subunit β: atpB translation products in strain βTr were analysed by pulse labelling followed by a chase for the indicated times in the presence of an excess of unlabelled acetate. Pulse-labelled wild type is shown for comparison. (C) Accumulation of subunit β in wild-type, βTr and ΔatpB (as a control) strains. OEE3, loading control. (D) Accumulation of transcripts for atpB, atpA and petA (loading control) in wild-type and βTr strains. Download figure Download PowerPoint Using the two steps procedure described in the legend of Table I we then substituted the resident petA gene of strains βTr and atpC1 {βTr} with the 5′atpB-petA chimera. In the resulting transformants {βTr, dBf} and atpC1 {βTr, dBf}, the 5′atpB-driven-cytochrome f reporter was more translated than in the control strain dBf (Figure 4A). This contrasts with the decreased expression observed in strain atpC1 {dBf} (Figure 2B) that accumulates full length subunit β unable to assembled in CF1. In agreement with pulse-labelling data, the accumulation of the cytochrome f reporter was increased by 50% in strains {βTr, dBf} and atpC1 {βTr, dBf} relative to dBf (Figure 4B), as opposed to its 95% decrease in strain atpC1 {dBf} (Figure 2C). Thus, subunit γ is not required to stimulate 5′atpB-driven translation, ruling out the trans-activation hypothesis. Our data point to a negative feedback mechanism: the production of unassembled β subunits represses translation initiation of atpB mRNA. Figure 4.The expression of 5′atpB-driven genes is no longer dependent on the presence of subunit γ when full-length subunit β is lacking. (A) Translation of the main chloroplast-encoded CF1 subunits and reporter cytochrome f in strains dBf, {βTr, dBf} and atpC1 {βTr, dBf}. *Points to truncated subunit β. Rates of synthesis of the cytochrome f reporter, relative to the reference strain dBf and normalised to that of the apo-CP43 subunit from PSII are shown. (B) Accumulation of subunits α, β and γ, of cytochrome f and OEE2 (loading control). In this figure and others, the vertical black line indicates lanes from the same gel that have been cropped because they were not originally adjacent on the gel. (C) Accumulation of transcripts for petA, atpB and psaA (loading control). Download figure Download PowerPoint Oligomeric forms of subunits α and β, but not 'free' subunit β, repress translation of atpB mRNA That the unassembled β subunit represses its own translation in the absence of the γ subunit but not in the absence of the α subunit (see Figure 1A) suggests that it does not adopt the same unassembled conformation in the two situations. A major difference between γ- or α-deficient mutant strains is that α/β oligomers can assemble in the former strain only. We thus deleted the atpA gene from the chloroplast genome of the atpC1 strain in order to prevent formation of α/β oligomers in absence of the γ subunit. Expression of subunit β in the resulting strains, atpC1 {ΔatpA}, is shown on Figure 5A and B. Whereas synthesis of subunit β was drastically reduced in strain atpC1, the double mutant, atpC1 {ΔatpA}, showed high rate of translation of subunit β, similar to that in the ΔatpA deletion strain. Accordingly, the accumulation of subunit β in strain atpC1 {ΔatpA} was similar to that in ΔatpA, much higher than in the atpC1 mutant. We conclude that subunit β per se, is unable to down-regulate its own synthesis: the negative feedback due to defective CF1 assembly requires the combined expression of α and β subunits, i.e. is rather born by α/β hetero-oligomers. Figure 5.In absence of subunit γ, the presence of subunit α is required for subunit β to repress translation of atpB mRNA. Levels of translation (A) protein (B) and transcript (C) accumulation for the main CF1 subunits in strains ΔatpA, atpC1 and atpC1 {ΔatpA} (two independent transformed clones). Levels of subunit β synthesis, relative to that of apoCP43, appear similar in the various strains, but strain atpC1. OEE2 (B) and petA transcript (C) provide loading controls. Download figure Download PowerPoint Oligomerisation of subunits α and β in a γ-deficient strain This prompted us to examine CF1 assembly intermediates in strains ΔatpA, ΔatpB, atpC1 and wild type. Soluble fractions, containing subunits not yet assembled into a membrane-bound ATP synthase, were analysed by CN–PAGE followed by denaturating Urea–SDS–PAGE in the second dimension and revealed after immunoblotting using a mixture of α and β specific antibodies (Figure 6). Figure 6.Oligomeric forms of α and β accumulate in the absence of subunit γ. CF1 assembly intermediates from soluble extracts of wild-type, ΔatpA, ΔatpB and atpC1 cells were separated on colourless native gels in the first dimension and denaturating 12–18% Urea-gels in the second dimension. Apparent molecular mass in the first dimension of molecular standards is shown on the top of the figure. After electrotransfer, the presence of subunits α and β was revealed using a mixture of antibodies specific for each subunit. The strength of the signal is therefore not an indication on the relative accumulation of the two subunits. The bottom black bar stresses the appropriate region for α/β heteromers in wild-type and atpC1 strains. Download figure Download PowerPoint In the wild type, subunits α and β were mainly observed in a complex of 500–600 kDa, absent in either strains ΔatpA or ΔatpB, that probably corresponds to the CF1 moiety of the ATP synthase. However, free β subunits—but not free α subunits—were detected in significant amount in the region around 60 kDa. In addition, subunit α, together with subunit β, was found in the region around 250 kDa (indicated by a black bar), and to a lesser extent in the 120–140 kDa region. However, 120–140 kDa was the major position of subunit α when expressed in absence of subunit β, as in strain ΔatpB, even if trace amounts of larger α-containing oligomers could also be detected. In strain ΔatpA, most of the β subunits were found as free polypeptides, below 100 kDa, with traces in larger complexes (up to 250 kDa) that may correspond to β homo-oligomers. In strain atpC1, subunit β was mostly fully unassembled, below 100 kDa, but a significant fraction was also found in the region around 250 kDa suggesting the presence of oligomers that were dramatically enriched in subunit α. Thus, in absence of subunit γ, an increased fraction of subunit α relocates in oligomeric complexes of same app. MM as those containing subunit β. As a further confirmation, subunit β was co-immunoprecipitated with an antibody raised against subunit α from a soluble cellular extract of strain atpC1 (data not shown), indicating that both subunits were indeed associated as hetero-oligomers. This mixed population of α- and β-containing oligomers reflects the accumulation of α/β assembly intermediates and provides a molecular basis for the α-sensitive CES behaviour of the β subunit in a γ-deficient context. Unassembled subunit β trans-activates translation of the α subunit To study the role of the atpA 5′UTR in the CES behaviour of the α subunit, we used the dAf strain that expresses a protein complex d (ATP synthase) atpA-driven cytochrome f, since it bears instead of the endogenous petA gene a 5′atpA-petA chimera. This reporter gene is translated at a level similar to—or slightly higher than—regular cytochrome f (Choquet et al, 1998). To compare the expression of the atpA 5′UTR-driven cytochrome f in the presence or absence of subunit β, we introduced by transformation the truncated atpB allele into the dAf strain. Indeed, the βTr strain shows the same decrease in translation of subunit α (Figure 3B) as strain ΔatpB (Figure 1A): truncated subunit β did not sustain translation of subunit α. Transformants {βTr, dAf}, recovered on TAP-spectinomycin plates, were analysed for cytochrome f expression. The levels of synthesis and accumulation of the cytochrome f reporter were about three times lower when subunit β could not accumulate (compare lanes dAf and {βTr, dAf} on Figure 7A and B). Since the chimeric petA mRNA accumulated to high level in both strains (Figure 7D), we conclude that 5′atpA-driven translation is decreased in cells lacking subunit β, the assembly partner of α. The atpA 5′UTR is therefore sufficient to confer a β-dependent CES behaviour to a reporter gene, pointing to a specific down-regulation of synthesis operating at the level of translation initiation. Figure 7.Initiation of translation of subunit α is not autoregulated but transactivated by subunit β. (A) Chloroplast translation products in dAf, {ΔatpA, dAf}, {βTr, dAf} and {βTr, ΔatpA, dAf} strains. The rate of translation of cytochrome f in the various strains, normalised to that of apoCP47 and relative to the reference strain dAf is indicated below for this representative experiment. (B, C) Accumulation of cytochrome f and OEE2 as a loading control in strains dAf, {ΔatpA, dAf}, {βTr, dAf}, {βTr, ΔatpA, dAf} (B) or in wild type and strains expressing the 5′atpA-petA chimeric gene, either alone (dAf) or together with a truncated subunit α {αTr, dAf}, a truncated subunit β {βTr, dAf} or both {βTr, αTr, dAf}. (C) On panel B, two independent transformants are shown for strains {βTr, dAf} and {βTr, ΔatpA, dAf}. (D) Accumulation of mRNAs for atpA, petA and psbA (loading control) in the wild type and in strains dAf, {ΔatpA, dAf}, {βTr, dAf} {βTr, ΔatpA, dAf} and {βTr, αTr, dAf}. Note the reduced accumulation of atpA mRNA in presence of the 5′atpA-petA chimeric gene. Download figu
Referência(s)