•In this study, we examined the biochemical and physiological functions of Nicotiana benthamiana S1 domain-containing Transcription-Stimulating Factor (STF) using virus-induced gene silencing (VIGS), cosuppression, and overexpression strategies.
•STF : green fluorescent protein (GFP) fusion protein colocalized with sulfite reductase (SiR), a chloroplast nucleoid-associated protein also present in the stroma. Full-length STF and its S1 domain preferentially bound to RNA, probably in a sequence-nonspecific manner.
•STF silencing by VIGS or cosuppression resulted in severe leaf yellowing caused by disrupted chloroplast development. STF deficiency significantly perturbed plastid-encoded multimeric RNA polymerase (PEP)-dependent transcript accumulation. Chloroplast transcription run-on assays revealed that the transcription rate of PEP-dependent plastid genes was reduced in the STF-silenced leaves. Conversely, the exogenously added recombinant STF protein increased the transcription rate, suggesting a direct role of STF in plastid transcription. Etiolated seedlings of STF cosuppression lines showed defects in the light-triggered transition from etioplasts to chloroplasts, accompanied by reduced light-induced expression of plastid-encoded genes.
•These results suggest that STF plays a critical role as an auxiliary factor of the PEP transcription complex in the regulation of plastid transcription and chloroplast biogenesis in higher plants.
Transcription of plastid-encoded genes involves at least two distinct types of RNA polymerase: plastid-encoded multimeric RNA polymerase (PEP) and nucleus-encoded phage-type RNA polymerase (NEP) (Link, 1996; Stern et al., 1997; Maliga, 1998; Shiina et al., 2005). PEP is similar to the eubacterial multisubunit RNA polymerase and consists of a core catalytic complex (α2ββ′β′′) whose subunits are encoded by plastid genes and a nucleus-encoded sigma factor thought to confer promoter specificity to the core complex (Shiina et al., 2005; Schweer et al., 2010). In addition, Arabidopsis possesses three NEPs, which show different subcellular localization: RpoTp and RpoTm are imported into chloroplasts and mitochondria, respectively, and RpoTmp is targeted to both chloroplasts and mitochondria (Hedtke et al., 2000). Promoters of different groups of plastid-encoded genes are recognized by PEP, NEP, or both NEP and PEP (Hajdukiewicz et al., 1997; Maliga, 1998; Baba et al., 2004).
Plastid-encoded multimeric RNA polymerase core subunits are present in two chloroplast protein preparations, the membrane-associated plastid transcriptionally active chromosome (pTAC) and the soluble RNA polymerase (sRNAP) (Pfannschmidt & Link, 1994). pTAC contains 40–60 polypeptides, including two core subunits of PEP (Suck et al., 1996). Both protein preparations can carry out transcription of plastid protein-encoding genes, rRNA, and tRNA (Rajashekar et al., 1991; Krupinska & Falk, 1994). Furthermore, the protein composition of pTAC and sRNAP fractions is different in proplastids, chloroplasts, and etioplasts, indicating dynamic regulation of the plastid transcription complexes depending on developmental stage (Reiss & Link, 1985; Pfannschmidt & Link, 1994; Suck et al., 1996). To identify nuclear-encoded proteins involved in plastid transcription, the protein composition of pTAC was analyzed by mass spectrophotometry, and the functions of several components were investigated in Arabidopsis (Pfalz et al., 2006). The mutants of the PTAC factors showed significantly reduced expression of plastid genes with PEP promoters, while expression of NEP-dependent genes was not affected or was elevated (Pfalz et al., 2006). Interestingly, this expression pattern is similar to that of Δrpo tobacco mutants that lack a functional PEP (Legen et al., 2002) and of a knockout Arabidopsis mutant of thioredoxin z, a novel subunit of PEP (Arsova et al., 2010; Schröter et al., 2010).
The Arabidopsis pTAC preparation contains a protein with a single S1 domain, designated PTAC10 (Pfalz et al., 2006). The S1 domain is an RNA-binding domain of c. 70 amino acids that is present in many RNA-associated proteins, including bacterial ribonucleases, and transcription and translation initiation factors (Bycroft et al., 1997). A tobacco homolog of PTAC10 was also identified as a major protein associated with the tobacco PEP complex that was affinity-purified using the α subunit with a C-terminal His tag (Suzuki et al., 2004). However, cellular function of this S1 domain-containing protein has not been characterized. In this study, we investigated protein characteristics and in vivo functions of the protein, S1 domain-containing Transcription-Stimulating Factor (STF), in N. benthamiana. We showed that STF binds to RNA via its S1 domain, and is colocalized with the chloroplast nucleoids. Function analyses of STF using virus-induced gene silencing (VIGS), cosuppression, and overexpression suggest that STF stimulates plastid gene expression and chloroplast biogenesis in N. benthamiana.
Materials and Methods
Chloroplast run-on transcription assay
Intact chloroplasts were purified from mature leaves of wild-type Nicotiana benthamiana L. plants and pale-green leaves of the CS1 lines as described (Kwon & Cho, 2008) with slight modification. Leaves were ground in ice-cold CIB buffer (0.45 M sorbitol, 50 mM Hepes-KOH pH 7.8, 2 mM EDTA, 0.1% BSA, 2.5 mM MgCl2, protease inhibitor cocktail). After filtering through two layers of Miracloth, the leaf extract was centrifuged for 5 s at 1000 g to remove debris. After centrifugation for 10 min at 2000 g, the pellet was resuspended in 2 ml of CIB buffer and then layered on a discontinuous 40% (3 ml)/80% (5 ml) Percoll gradient. After centrifugation for 15 min at 1500 g using a centrifuge (Hanil Science Industrial, Incheon, South Korea, model UNION 32R), intact chloroplasts were isolated from the interface between the two layers, which were colored green and yellow in the wild-type and the CS1 lines, respectively. The chloroplasts were washed twice in HMS buffer (0.33 M Sorbitol, 50 mM Hepes-KOH pH 7.8, 2.5 mM MgCl2 and protease inhibitor cocktail) by centrifugation for 5 min at 1500 g. The purified intact chloroplasts were resuspended in a small volume of HMS buffer, and counted under a light microscope using a hemacytometer (Marienfeld, Lauda-konigshofen, Germany).
The chloroplast run-on transcription assay was performed using lysed chloroplasts as described (Mullet & Klein, 1987; Rapp et al., 1992). Duplicate samples containing c. 2.5 × 106 chloroplasts were vortexed and incubated for 5 min in 50 μl run-on transcription buffer containing 50 mM HEPES [KOH], pH 8.0, 10 mM MgCl2, 25 mM potassium acetate, 10 mM dithiothreitol, 125 μM ATP, CTP, GTP and 50 μCi α-[32P]-UTP. The radiolabeled RNA was isolated using QIAzol (Qiagen), and hybridized to DNA blots prepared by denaturing and spotting 100 pmol of PCR products representing selected coding regions of plastid genes psaB, psbA, rps14, rbcL, rpoB, and ycf2, and the nuclear gene N. benthamiana Rae1 (Lee et al., 2009). Signal intensities were visualized on a BAS-2500 bioimage analyzer (Fuji film, Tokyo, Japan) and quantified using the Analysis Life Science Research imaging system (Olympus, Tokyo, Japan).
To observe the effect of added STF on plastid transcription, the chloroplast run-on assay was carried out with control maltose binding protein (MBP) or the recombinant MBP : STF, MBP : STFΔ5 and MBP : STFΔ2 fusion proteins. Lysed chloroplasts (c. 2.5 × 106) were incubated with MBP (2 pmol μl−1), MBP : STFΔ5 (1 pmol μl−1), MBP : STFΔ2 (1 pmol μl−1) or MBP : STF (0.7 pmol μl−1) in run-on transcription buffer for 10 min at 25°C in the presence or absence of tagetitoxin (10 μM Tagetin™; Epicentre Biotechnologies, Madison, WI, USA) in the final volume of 290 μl. Ten microlitres of α-[32P]-UTP (10 μCi μl−1) were added to the reaction mixture, followed by further incubation at 25°C for indicated times. Aliquots of the samples were taken periodically and mixed with QIAzol (Qiagen) to stop the reaction before RNA extraction. Incorporation of α-[32P]-UMP in the newly synthesized RNAs was measured using a scintillation counter (Beckman (Model LS6500), Indianapolis, IN, USA).
Nucleic acid binding assay
33P-labeled RNA substrate (c. 160 nucleotides long) was prepared by transcribing the BamH1-digested pET-22b(+) plasmid using T7 RNA polymerase as previously described (Kim et al., 2007). For RNA–protein interaction analysis, the labeled RNAs (0.5 pM) were incubated with purified recombinant proteins in binding buffer (10 mM Tris-HCl (pH 8.0), 50 mM NaCl, 1 mM EDTA, and 7% glycerol) for 30 min on ice. The mixture was separated on 6% nondenaturing polyacrylamide gel, and RNA bands were detected by Phosphor Imager (Fuji, Japan). For electrophoretic mobility shift assay (EMSA) competition assays, 25 pmol of the recombinant proteins was incubated with variable ratios of the radiolabeled (0.5 pM) and unlabeled RNAs, ranging from 1 : 0 to 1 : 20. For DNA–protein interaction analysis, 150 ng single-stranded DNA (M13mp8 ssDNA; 7229 nucleotides; Sigma Wako Nippon Gene) and 200 ng double-stranded DNA (M13mp8 RF1 (replicative form) dsDNA; 7249 bp; Sigma Wako Nippon Gene) were incubated with purified recombinant proteins in binding buffer (20 mM Tris-HCl, pH 7.5, 2 mM EDTA, 4 mM KCl, 5% glycerol, and 50 μg ml−1 BSA) for 40 min on ice as previously described (Karlson et al., 2002). The mixture was separated on 1% agarose gels, and the DNA bands were visualized under UV light.
Detailed descriptions of the following techniques are included in the legends of the Supporting Information, Methods S1: Virus-induced gene silencing, northern blot analysis and semiquantitative reverse transcription polymerase chain reaction (RT-PCR), real-time quantitative RT-PCR, 4’,6-diamidino-2-phenylindole (DAPI) staining, confocal microscopy for subcellular localization of STF, chloroplast fractionation and STF localization, determination of chlorophyll fluorescence parameters, transmission electron microscopy, protein extraction and immunoblotting, purification of recombinant STF proteins, and Generation of STF-overexpression transgenic lines in N. benthamiana.
Identification of STF
We carried out functional genomics using Tobacco rattle virus (TRV)-based VIGS in N. benthamiana (Cho et al., 2004; Ahn & Pai, 2008; Kang et al., 2010). This screening revealed that gene silencing of N. benthamiana STF (S1 domain-containing Transcription-Stimulating Factor) caused severe leaf yellowing. N. benthamiana STF (GenBank accession number HM012811) encodes a 680-amino-acid polypeptide with a molecular mass of 80 019 Da (Fig. S1). The ChloroP algorithm (http://www.cbs.dtu.dk/services/ChloroP/; Emanuelsson et al., 1999) identified a chloroplast transit peptide of 34 amino acids at the N-terminus of the predicted STF protein. In addition, STF has a single S1 domain in the middle of the protein (Fig. S1). Interestingly, the S1 domain of STF is homologous to the S1 domain of the conserved transcription accessory protein Tex in prokaryotes. N. benthamiana STF is highly homologous to related proteins in Arabidopsis (STF-At), rice (STF-Os), and tomato (STF-Le) (Fig. S1). STF is a single copy gene in the Arabidopsis and rice genomes. According to the Genvestigator program (https://www.genevestigator.com/), the Arabidopsis STF gene is highly expressed in the embryo, endosperm, imbibed and germinating seeds, shoot apex, and leaf primordia, but expressed at a low level in the stem, roots (including the root tip), and calli, indicating that STF is preferentially expressed in the seeds and in young aerial tissues with active cell division (Fig. S2).
VIGS phenotypes and silencing of endogenous STF transcripts
For VIGS of STF, we cloned three different STF cDNA fragments into the TRV-based VIGS vector pTV00 (Cho et al., 2004), and infiltrated N. benthamiana plants with Agrobacterium containing each plasmid (Fig. 1a). TRV : STF(N) and TRV : STF(C) contain a 500 bp N-terminal and a 380 bp C-terminal region of the cDNA, respectively, whereas TRV : STF(F) contains cDNA corresponding to the full-length coding region. VIGS with each construct resulted in the same phenotype of leaf yellowing without greatly affecting the overall growth and development (Fig. 1b,d). The colors of the affected leaf sectors in TRV : STF plants varied from pale green to yellow or white. The amount of endogenous STF mRNA in the VIGS lines was measured by semiquantitative RT-PCR to determine the effect of gene silencing (Fig. 1f). RT-PCR using STF-A primers showed that the TRV : STF(C) line had much lower amounts of PCR product in the yellow sectors of the leaves compared with the TRV control, indicating reduced abundance of STF transcript. The same primers detected high abundance of viral genomic transcripts containing the N-terminal region of STF in TRV : STF(N) lines (Fig. 1f). Similarly, RT-PCR with the STF-B primers produced lower amounts of PCR product in the TRV : STF(N) line than in the TRV control, but detected viral genomic transcripts containing the C-terminal region of STF in TRV : STF(C) lines. As a control, the transcript abundance of actin remained constant (data not shown). These results confirmed silencing of the endogenous STF gene in the TRV : STF plants.
Generation of STF overexpression and cosuppression lines
To generate STF overexpression (OE) transgenic lines, we transformed N. benthamiana plants with STF : His constructs in which the STF protein-coding region was fused to a His-tag under control of the CaMV35S promoter (p35S::STF:His). The OE lines showed no visible differences in pigment biosynthesis or plant growth compared with wild-type N. benthamiana plants (Fig. 1c,e). Semiquantitative RT-PCR using the STF-specific primer revealed high abundances of transgene transcripts in three representative OE lines (Fig. 1g). Interestingly, of > 25 independent OE lines, two transgenic lines displayed the leaf yellowing phenotype. Semiquantitative RT-PCR using STF-specific primers generated greatly reduced amounts of PCR products in these plants compared with wild-type plants, suggesting cosuppression of the endogenous STF gene (Fig. 1g). When the T1 seeds from the cosuppression (CS) lines were germinated on the medium, some seeds produced a callus-like structure and soon died (results not shown), while others developed into seedlings with chlorotic cotyledons (Fig. 1c). The surviving T1 seedlings of the CS lines grew slowly and exhibited the leaf yellowing phenotype to a varying degree (Fig. 1c–e). This result suggests that STF deficiency in early developmental stages inhibited plant growth as well as pigment biosynthesis.
Effects of STF silencing on chloroplasts
Transverse sections of TRV : STF leaves showed the typical leaf structure of dicotyledonous plants, except for diminished chloroplast numbers, when compared with TRV control leaves (Fig. 2a). Leaf protoplasts were generated from TRV, TRV : STF, wild-type, OE5, and CS1 lines, and examined by light microscopy and confocal laser scanning microscopy (Fig. 2b). Although chloroplast size and number were comparable in TRV, wild-type, and OE5 protoplasts, TRV : STF and CS1 protoplasts exhibited reduced chloroplast numbers and drastically decreased chloroplast size and autofluorescence (Fig. 2b,d,e). TRV : STF chloroplasts contained chloroplast nucleoids (cp-nucleoids) with significantly increased DAPI fluorescence despite their small size, while TRV control chloroplasts exhibited cp-nucleoids with faint DAPI staining (Fig. 2c).
Reduced accumulation of chloroplast proteins and decreased photosynthetic capacity
The severe leaf yellowing phenotypes indicate impaired photosynthetic activity of the STF-silenced plants. Indeed, western blot analyses revealed significant reduction in the amounts of plastid-encoded photosynthetic proteins rbcL (Rubisco large subunit), cyt f (Cyt b6f complex subunit), D1 (PSII reaction centre core subunit), atpB (ATP synthase subunit), and psaA (PSI core subunit) in leaves of the TRV : STF VIGS lines and the CS1 lines (Fig. 3a,b). The STF OE5 line exhibited a similar pattern of accumulation of the plastid-encoded proteins to that of TRV control and wild-type plants (Fig. 3a,b). It has been reported that accumulation of photosynthetic complex subunits is highly coordinated, and defective synthesis of any single subunit in the photosynthetic enzyme complex leads to defective accumulation of the entire complex (Barkan, 1998). Thus, reduced accumulation of individual core subunits in the STF VIGS lines and CS1 lines indicates a global defect in the assembly and function of the photosynthetic enzyme complexes.
To understand the physiological impact of defective protein accumulation, we examined the photosynthetic capacity of wild-type, OE5, and CS1 leaves (Fig. 3c,d). The ratio of variable fluorescence to maximum fluorescence (Fv/Fm) reflects the maximal photochemical efficiency of photosystem II (Fig. 3c). The Fv/Fm value in the CS1 leaves was significantly reduced under light growth conditions when compared with those of wild-type or OE5 leaves, suggesting a severe reduction in functional photosystem II centers (Fig. 3c). STF deficiency also resulted in a decrease in the electron transport rate (ETR), the noncyclic electron transport flux through PSII, in the CS1 line (Fig. 3d). These results suggest that the photosynthetic capacity of the CS1 line was severely compromised.
Chloroplast targeting of STF
To investigate the subcellular localization of STF, we generated fusion proteins in which full-length STF protein (Met-1 to Asp-680) or its N-terminal region (Met-1 to Gln-199) was fused to green fluorescent protein (GFP) under the control of the CaMV35S promoter to generate STF : GFP and STF-N : GFP, respectively. DNA constructs encoding the different GFP fusion proteins were introduced into protoplasts isolated from N. benthamiana seedlings, and gene expression was analyzed by confocal laser scanning microscopy (Fig. 4a). Green fluorescent signals of STF : GFP were detected as distinct particles within the chloroplasts, and the STF : GFP signal colocalized with DAPI-stained cp-nucleoids (Fig. 4b). By contrast, the STF-N : GFP signal was evenly distributed in the chloroplasts (Fig. 4a). Next, the protoplasts were cotransformed with a DNA construct encoding SiR : red fluorescent protein (RFP), in which sulfite reductase (Cannon et al., 1999; Sato et al., 2001; Kang et al., 2010) was fused to RFP (Fig. 4c). SiR is associated with cp-nucleoids, but also present in the stroma (Sekine et al., 2007), and it was identified as a component of pTAC (Pfalz et al., 2006). The STF : GFP signal overlapped with the SiR : RFP signal as distinct particles within the chloroplasts (Fig. 4c). These results are consistent with previous findings on the association of STF with the PEP complex (Suzuki et al., 2004; Pfalz et al., 2006). N. benthamiana STF (STF : GFP) was colocalized with Arabidopsis STF (STF-At : GFP), suggesting that the two proteins are functional homologs (Fig. S3).
STF is localized in both the stromal and thylakoid membrane fractions
S1 domain-containing Transcription-Stimulating Factor : GFP was expressed in wild-type N. benthamiana leaves by agroinfiltration, and chloroplast subfractions were purified from the leaves at 2 d after infiltration to examine the protein localization within chloroplasts. Arabidopsis STF : GFP (STF-At : GFP) was also examined in a similar way to test whether there are any differences in the subplastidic localization between the STF homologs. Western blot analysis with anti-GFP antibody detected STF : GFP and STF-At : GFP proteins in both the stromal and thylakoid membrane fractions (Fig. 5a). As a control, specific antibodies detected rbcL only in the stromal fraction and the D1 subunit of photosystem II only in the thylakoid membrane fraction (Fig. 5a). Since STF is copurified with a transcriptionally active chromosome preparation and cp-nucleoids are mainly membrane-associated, we examined whether STF is associated with the thylakoid membrane via a DNA or RNA tether (Fig. 5b). Treatment of the thylakoid membrane fractions with either DNase or RNase did not release a portion of the thylakoid membrane-associated STF into the soluble fractions, suggesting that STF association with the membrane is not mediated by chloroplast DNA or RNA.
RNA-binding activities of STF and its S1 domain
To investigate RNA-binding activity, full-length STF and the S1 domain were expressed in E. coli as fusion proteins to glutathione-S-transferase (GST) and purified for gel-mobility shift assays (Fig. 6a). An approx. 160-nucleotide, 33P-labeled, single-strand RNA was synthesized by transcription of pET-22b(+) plasmid using T7 RNA polymerase, incubated with increasing concentrations of GST : STF and GST : S1, and run-on native polyacrylamide gels to visualize the RNA–protein complexes. Both GST : STF and GST : S1 readily formed stable RNA–protein complexes, while GST alone did not bind the RNA probe (Fig. 6a). To examine the specificity of the STF–RNA interaction, EMSA competition assays were performed using two His-tag fusion proteins, STF : 6 × His and S1 : 6 × His, with variable ratios of the radiolabeled and unlabeled RNAs (Fig. S4a,b). Excessive amounts of the cold RNA competitor competed with the labeled RNA, displacing STF : 6 × His and S1 : 6 × His from the labeled probe (Fig. S4a,b). These results suggest specificity of the interaction between STF and RNA. GRP7 : 6 × His, a fusion protein between GRP7 RNA-binding protein (Kim et al., 2008) and the 6 × His tag, was used as a positive control for RNA binding (Fig. 6a). These results demonstrate that STF is able to bind to RNA, and suggest that the S1 domain is the primary RNA-binding domain of STF, although we cannot rule out the possibility that other protein regions make minor contributions to the binding. Since an artificial RNA derived from vector sequence was used as a probe, these assays could not detect any sequence-specific interaction between STF and native plant RNA. However, the observation that STF binds to the artificial RNA indicates that cellular substrates for STF may be sequence-nonspecific RNA transcripts.
To test DNA-binding activities, ssDNA (M13mp8 ssDNA; 7229 nucleotides) and dsDNA probes (M13mp8 RF1 dsDNA; 7249 bp) were incubated with increasing concentrations of GST : STF and GST : S1 and analyzed on agarose gels. GST : STF and GST : S1 exhibited no affinity to DNA in single- or double-stranded form (Fig. 6b). As positive controls, GST : CspA, a GST fusion protein of the single-stranded DNA binding protein CspA (E. coli cold shock protein A: Genbank accession number AP_004238.1), and GST : CSDP1, a GST fusion protein of CSDP1 (Arabidopsis cold shock domain protein 1; Park et al., 2010) that binds to both single- and double-stranded DNA, shifted the position of the DNA bands by formation of a DNA–protein complex (Fig. 6b). GST alone did not bind to DNA. Taken together, these results suggest that STF preferentially binds to RNA.
Differential effects of STF silencing on transcript accumulation of plastid-encoded genes
Next we investigated whether STF deficiency affects the mRNA expression profiles of the plastid-encoded genes (Fig. 7). Total RNA was prepared from leaves of TRV, TRV : STF, wild-type, OE1, OE5, and CS1 lines, and the rRNA amounts were determined as a control for RNA loading (Fig. 7d). Since white and yellow sectors of TRV : STF and CS1 leaves contained severely defective chloroplasts (Fig. 2), we used only the pale green sectors for RNA preparation. Northern blot analyses were carried out using diverse plastid-encoded genes as probes; these included genes transcribed by PEP (Fig. 7a), NEP (Fig. 7c), or both PEP and NEP (Fig. 7b). The relative transcript abundances were also quantified (Fig. 7e–g). Steady-state transcript abundances of PEP-dependent photosynthesis genes psbD, petB, psaB, psbA, and rbcL (Hajdukiewicz et al., 1997; Baba et al., 2004) decreased significantly in the TRV : STF and CS1 lines, in contrast to TRV, wild-type, and OE lines (Fig. 7a,e). The TRV : STF and the CS1 lines also exhibited greatly reduced transcript abundances of atpA, atpB, clpP, ndhB, and ndhF, which have both NEP and PEP promoters (Silhavy & Maliga, 1998; Magee & Kavanagh, 2002) (Fig. 7b,f). However, steady-state transcript abundances of NEP-dependent plastid genes accD, rpoB, rpoA, and ycf2 (Hajdukiewicz et al., 1997; Baba et al., 2004) remained constant in all six lines (Fig. 7c,g). Thus, depletion of STF caused pronounced down-regulation of PEP-dependent gene expression without greatly affecting NEP-dependent transcript accumulation, indicating that STF is required for proper function of the PEP transcription machinery.
Chloroplast run-on transcription assay
Reduced accumulation of PEP-dependent gene transcripts in the STF-silenced plants could result from a decreased rate of PEP-mediated transcription or decreased mRNA stability. To distinguish these possibilities, we carried out chloroplast run-on transcription assays (Fig. 8a,b). Chloroplasts prepared from mature leaves of the wild-type and the pale green sectors of the CS1 leaves were pulse-labeled with α-[32P]-UTP, and the radiolabeled RNAs were purified and hybridized to a blot spotted with PCR products of the indicated plastid genes and a nucleus-encoded gene, N. benthamiana Rae1 (Lee et al., 2009), as a negative control. Each probe was presented in duplicate. The signal intensities of PEP-dependent genes psaB, psbA, rps14, and rbcL were significantly lower in the CS1 line than in the wild-type, whereas there were no obvious differences between the lines in the signal intensities of the NEP-dependent plastid genes rpoB and ycf2 (Fig. 8a,b). Quantification of the signal intensities indicated that the decrease in transcription rate in the CS1 line correlated with the decrease in mRNA accumulation (Fig. 8b). These results suggest that STF positively regulates PEP to stimulate the transcription rate.
Stimulation of chloroplast transcription by exogenously added STF proteins
To demonstrate that STF plays a direct role in the regulation of plastid transcription, we examined whether exogenous addition of STF increases the transcription rate in chloroplast run-on transcription assays (Fig. 8c–h). We first prepared recombinant proteins of full-length STF (lacking the N-terminal 40 amino acids containing the chloroplast transit peptide), STFΔ5 and STFΔ2 as MBP fusion proteins (Fig. 8c). The corresponding cDNA fragments were cloned into the pMAL expression vector and expressed in E. coli. The recombinant MBP : STF, MBP : STFΔ5, and MBP : STFΔ2 fusion proteins were affinity-purified using the N-terminal MBP tag yielding polypeptides of c. 118, 90, and 82 kDa in size, respectively (Fig. 8d). Chloroplasts were prepared from mature leaves of wild-type N. benthamiana plants. Chloroplast run-on assays were carried out with α-[32P]-UTP for the indicated times after incubation with the purified MBP : STFΔ5 or MBP as a control (Fig. 8e). The radiolabeled RNAs were purified and incorporation of α-[32P]-UMP into plastid RNA was quantified. Quantitative analysis revealed that plastid transcription rate in the presence of MBP : STFΔ5 was significantly higher than that in the control reaction with MBP (Fig. 8e). The results were highly reproducible with independent preparations of chloroplasts and MBP : STFΔ5/MBP proteins. MBP : STFΔ2 failed to elevate the plastid transcription rate, indicating that the S1 domain is required for the transcription stimulation by STF (Fig. 8f). Incorporation of α-[32P]-UMP was effectively suppressed by tagetitoxin, which specifically inhibits PEP (Mathews & Durbin, 1990), in the reaction with MBP as well as with MBP : STFΔ5, further supporting the notion that STF mainly operates in the PEP-mediated transcription system (Fig. 8g). It was previously shown that tagetitoxin abolishes transcription in mature chloroplasts (Sakai et al., 1998; Sekine et al., 2002). Finally, exogenous addition of MBP : STF increased the plastid transcription rate to a level comparable to MBP : STFΔ5, although only two-thirds of the protein amounts were used for MBP : STF (Fig. 8h). Taken together, these data demonstrate that STF can directly stimulate PEP-catalyzed transcription processes in chloroplasts.
Delayed transition from etioplasts to chloroplasts
Since a high level of plastid gene expression is required for the transition from etioplasts to chloroplasts, we examined the phenotypic effects of STF deficiency during this transition in seedlings from wild-type, OE5, and CS1 lines (Fig. 9a). The seedlings were dark-grown for 1 wk (Dark), and then transferred to light for 3 h (D→L 3 h) or 16 h (D→L 16 h). Compared with the wild-type and OE5 seedlings, which exhibited normal cotyledon greening after transfer, greening in the CS1 seedlings was greatly delayed (Fig. 9a). To investigate the delayed greening at the cellular level, thin cotyledon sections were prepared from wild-type and CS1 seedlings at the indicated time and observed by transmission electron microscopy (TEM) (Fig. 9b–o). The cotyledon chloroplasts of the dark-grown wild-type seedlings displayed extensive prolamellar bodies (PLBs) from which unstacked prothylakoids protruded into the stroma (Fig. 9b,c). Upon light exposure, primarily unstacked thylakoids developed without PLBs or with PLBs of reduced sizes at 3 h after transfer, followed by formation of numerous grana stacks at 16 h after transfer (Fig. 9d–h). In the CS1 line, some of the etioplasts were already degenerating before light exposure (Fig. 9k). Furthermore, the CS1 line did not show normal progression of thylakoid development during the transition (Fig. 9i–o). At 3 and 16 h after transfer, the plastids contained a poorly developed thylakoid system and some of the plastids were degenerating or filled with vesicles. In addition, while protein bodies that had accumulated in mature seeds had almost disappeared in the cotyledons of etiolated seedlings of the wild-type, the cotyledons of the CS1 seedlings still contained a large number of protein bodies, even at 16 h after transfer to light (Fig. 9i,m; cf. control in 9b). These results suggest that nutrient mobilization in the cotyledons was significantly delayed in the CS1 line, consistent with its disrupted thylakoid biogenesis.
The chloroplasts in mature leaves from the TRV, TRV : STF VIGS, and CS1 lines were also examined by TEM. While TRV chloroplasts exhibited well-developed thylakoid membranes and large starch granules (Fig. 9p), leaf chloroplasts from the TRV : STF and CS1 lines did not show stacked grana but instead contained numerous vesicles or premature unstacked thylakoids, frequently present with multiple plastoglobuli (Fig. 9q–s). These results indicate that STF is required for the light-triggered biogenesis of chloroplasts from etioplasts.
Plastid gene expression during the transition from etioplasts to chloroplasts
To examine expression of plastid-encoded genes during the transition, total RNA was prepared from seedlings of the wild-type and the CS1 line at 0, 3, and 16 h after transfer from dark conditions to light. Then real-time quantitative RT-PCR was carried out using gene-specific primers for the plastid-encoded genes (Fig. 10). In the wild-type, transcript abundances of all of the genes tested significantly increased at 3 h after transfer, and then decreased a little from the value at 16 h after transfer (Fig. 10). The CS line also exhibited light-dependent induction of plastid gene expression during the transition, but accumulated much reduced amounts of the gene transcripts at all time points, regardless of their promoter types (Fig. 10). Furthermore, the transcript abundances of most of the genes were highest at 16 h after transfer in the CS line, suggesting delayed induction of light-induced plastid gene expression. Thus during the transition from etioplasts to chloroplasts, STF deficiency affected both NEP- and PEP-dependent transcription, in contrast to the effects in mature leaves.
S1 domain-containing Transcription-Stimulating Factor was previously identified as an associated component in the tobacco PEP complex (Suzuki et al., 2004) and in the Arabidopsis pTAC complex ((plastid transcriptionally active chromosomes); Pfalz et al., 2006), but function of STF has not been investigated. In this study, we present evidence that STF is a nucleus-encoded factor directly stimulating plastid transcription. STF colocalized with the chloroplast nucleoids and exhibited the RNA-binding activity in vitro. STF deficiency caused significant reduction in mRNA accumulation of PEP-dependent plastid genes in mature leaves, without much affecting NEP-dependent transcript accumulation. The reduced mRNA accumulation of PEP-dependent genes in STF-silenced lines was mainly caused by reduced transcription rate of the genes. By contrast, exogenous addition of the recombinant STF protein significantly increased the plastid transcription rate, while the PEP-specific inhibitor tagetitoxin abolished the transcription stimulation by STF. These results, combined with its consistent presence in plastid transcription complex, indicate that STF plays an important role in chloroplast transcription in mature leaves as an auxiliary factor for PEP transcription machinery.
In both prokaryotes and eukaryotes, the enzymatic activity of RNA polymerase (RNAP) is tightly regulated by a group of transcription factors that act on RNAP, RNA, or DNA during all stages of the transcription cycle (Borukhov et al., 2005). Many of these are DNA-binding proteins that act during transcription initiation, but an increasing number of proteins have been shown to act during elongation and termination stages by directly modifying the properties of RNAP. The majority of studies investigating the regulation of plastid transcription have focused on transcription factors or DNA-binding factors that confer promoter specificity. Chloroplast RNAP sigma factors play a critical role in the control of plastid transcription (Ishizaki et al., 2005; Loschelder et al., 2006; Tozawa et al., 2007; Onda et al., 2008). CDF2 acts as a transcription initiation factor for plastid rDNA transcription by NEP-2 in spinach by binding to the rDNA PC promoter region (Bligny et al., 2000). SibI specifically interacts with the region four of the plastid sigma factor Sig1 that regulates promoter recognition and site-specific transcription by PEP (Morikawa et al., 2002). The spinach DNA-binding protein of c. 31 kDa interacts specifically with the plastid psaA-psaB-rps14 promoter region (Cheng et al., 1997). Interestingly, the 163-amino-acid Etched 1 (ET1) in maize contains a zinc ribbon motif and is homologous to the C-terminal part of eukaryotic transcription elongation factor TFIIS (da Costa e Silva et al., 2004). ET1 is chloroplast-targeted, and a lack of ET1 function causes defective plastid development (da Costa e Silva et al., 2004). However, since ET1 is significantly smaller than TFIIS, it is unclear whether ET1 plays an analogous role with TFIIS in plastids. Recently, a larger TFIIS homolog working in the nucleus has been shown to be involved in seed dormancy in Arabidopsis (Grasser et al., 2009).
In this study, the S1 domain is critical for STF function, since its removal abrogated the transcription-stimulating activities of STF (Fig. 8). Previous findings suggested that the S1 domain provides a RNA-binding surface with either broad specificity or no sequence specificity, and in RNase E and polynucleotide phosphorylase, the S1 domain plays an additional role in protein–protein interactions (Bycroft et al., 1997; Schubert et al., 2004; Amblar et al., 2007; Johnson et al., 2008). Involvement of the S1 domain in transcription control has been suggested in several bacterial transcription factors. NusA is a highly conserved essential factor that is involved in transcription elongation, anti-termination, pausing, and termination (Mah et al., 1999; Gusarov & Nudler, 2001; Borukhov et al., 2005). Based on structural studies of NusA in complex with a bacterial RNAP, the NusA C-terminal region containing the S1 and KH domains interacts with the RNAP β- and β′-subunits that form the RNA exit channel, while simultaneously binding to the emerging transcript (Borukhov et al., 2005; Yang et al., 2009). These results suggest that the S1 domain constitutes part of a functional module that binds to both RNAP and the emerging RNA. The S1 domain of STF shows modest homology (amino acid similarity of c. 50%) with the S1 domain of the bacterial Tex (Toxin expression) proteins, while other regions of STF share no sequence homology with Tex. Tex was originally identified as an essential transcription accessory protein involved in transcription of critical toxin genes in Bordetella pertussis and was later found to be ubiquitous and extremely well conserved among prokaryotes, although its molecular function in diverse bacteria has remained enigmatic (Fuchs et al., 1996; He et al., 2006). A recent study on crystal structure and RNA binding of Pseudomonas aeruginosa Tex revealed that Tex preferentially binds to single-stranded RNA via its C-terminal S1 domain in a sequence-nonspecific manner (Johnson et al., 2008). Furthermore, the structure of Tex resembles the core structure of the eukaryotic transcription elongation factor Spt6, indicating that the Tex structure may represent a conserved scaffold for RNA binding and transcriptional regulation in both eukaryotes and prokaryotes (Johnson et al., 2008). The sequence homology suggests that the S1 domain in STF and Tex may play a similar role in the regulation of transcription processes.
Recent evidence suggests that many prokaryotic transcription factors bind directly to RNA polymerase without binding to DNA, affecting specific steps in the transcription pathway (Haugen et al., 2008). STF lacks DNA-binding activity, and therefore it does not likely confer promoter specificity, at least not by itself, upon the PEP transcription system. However, exogenously added STF was able to stimulate PEP-catalyzed transcription significantly in the run-on assays, while STF depletion disturbed PEP-mediated plastid transcription in vivo. We speculate that STF may function more broadly during PEP-mediated plastid transcription. Possibly by binding to PEP subunits and nascent RNAs as a component of the PEP transcription complex, STF may increase transcription activity and processivity of PEP, inducing allosteric changes. Further studies are required to understand detailed molecular mechanisms of STF action during chloroplast transcription cycles.
It is well known that during dark-to-light transition there is a dramatic increase in plastid gene expression, which supports the development of photosynthetically active chloroplasts. Mature seeds already possess both PEP and NEP, and assembly of these two transcription systems appears to proceed in parallel during seed germination in Arabidopsis (Demarsy et al., 2006). STF deficiency significantly delayed the light-triggered plastid gene expression and chloroplast biogenesis in N. benthamiana cotyledons, accompanied by delayed nutrient mobilization even after transfer to light. These results indicate that STF plays a critical role in chloroplast biogenesis and plastid gene expression during the transition from etioplasts to chloroplasts, in addition to its role in the differentiation from proplastids to chloroplasts in leaves. The PLBs in the etioplasts contain protochlorophyllide (Pchlide), which is rapidly reduced and esterified to chlorophyll upon illumination (Ryberg & Sundqvist, 1991), and thus PLBs play a significant role in the rapid onset of chloroplast development and photosynthetic activity. Consistent with the results from the etioplast proteomes (von Zychlinski et al., 2005; Kleffmann et al., 2006; Kanervo et al., 2008), recent proteomic analysis of purified prolamellar bodies from wheat etioplasts revealed that PLBs contain many proteins involved in photosynthetic light reactions, pigment biosynthesis, and the Calvin cycle, in addition to the predominant NADPH : protochlorophyllide oxidoreductase A ((PORA); Blomqvist et al., 2008). The abundance of photosynthetic proteins in PLBs supports the idea that PLB membranes are precursors to the thylakoids that facilitate the rapid formation of photosynthetic membranes during greening, and suggests that PEP activity is likely required for formation of functional PLBs in the etioplasts.
One interesting finding is that STF deficiency affected both NEP- and PEP-dependent gene expression during the light-triggered chloroplast development, in contrast to its specific effects on PEP-dependent expression in mature leaves of light-grown plants. The results indicate that, during the transition from etioplasts to chloroplasts, STF may play a role in both NEP- and PEP-dependent transcription systems. Alternatively, the etioplasts in the STF-deficient CS1 seedlings might be already abnormal so that the whole plastid gene expression system was malfunctioning during the transition, regardless of the promoter type. The latter possibility is supported by several findings in the CS1 line, including the degenerating etioplasts, much lower expression of the plastid genes in the etiolated seedlings, and the delayed or defective nutrient mobilization in the etiolated cotyledon even after transfer to light. A high abundance of STF transcripts was found in dry, imbibing, and germinating seeds in Arabidopsis based on the Genvestigator program, suggesting STF function in embryonic and postembryonic plastids (Fig. S2). Our results suggest that STF may play a critical role in gene expression during the light-induced transition of etioplasts to photosynthetically functional chloroplasts.
The authors wish to thank Dr Jon Soderholm (Yonsei University, Korea) for manuscript editing. This research was supported by grants to H-S.P. from the Mid-career Researcher Program (no. 20110000066 and 20100026168; NRF/MEST) and the Next-Generation BioGreen 21 Program (SSAC, no. PJ008214; RDA) of the Republic of Korea. H.K. was supported by the National Research Foundation of Korea grant funded by the Ministry of Education, Science and Technology (2011-0017357).