Expression of the Arabidopsis Sigma Factor SIG5 Is Photoreceptor and Photosynthesis Controlled

Two collections of Arabidopsis GAL4 enhancer trap lines were screened for light-intensity dependent reporter gene activation. Line N9313 was isolated for its strong light-intensity regulation. The T-DNA element trapped distant enhancers of the SIG5 promoter, which drives expression of a sigma factor involved in regulation of chloroplast genes for photosystem II core proteins. The T-DNA insertion 715 bp upstream of the transcription initiation site splits the promoter in a distal and proximal part. Both parts are sensitive to blue and red light and depend on photosynthetic electron transport activity between photosystem II and the plastoquinone pool. The mainblue-light sensitivity is localized within a 196-bp sequence (–887 to –691 bp) in the proximal promoter region It is preferentially CRY1 and PHYB controlled. Type-I and type-II phytochromes mediate red-light sensitivity via various promoter elements spread over the proximal and distal upstream region. This work characterizes SIG5 as an anterograde control factor of chloroplast gene expression, which is controlled by chloroplast signals in a retrograde manner.


Introduction
Light provides the main source of energy for plant life. Furthermore, light is a key regulator of plant development and metabolism. Many light-responses include gene expression regulation [1,2]. Cis-acting regulatory elements control the combination, spacing and relative orientation of transcription factors, which in turn make the promoters sensitive to light intensities and light qualities in a gene-specific manner [2,3]. Most light responsive elements (LREs) have been identified by characterization of the promoter regions of photosynthesis-associated genes [4]. Motifs, such as G-boxes (CACGTG; [5]), Z-boxes (ATACGTGT; [6]), I-boxes (GATAAGR; [7]) and GATA-motifs (GATA; [8]) have mainly been identified in close vicinity (within 300 bp upstream) to the transcription initiation site. In the core-promoters, specific TATA-boxes/Inr-elements increase the light sensitivity [9]. More distant elements were identified by foot-printing and gel-retardation assays in screens for binding motifs of light-responsive transcription factors [1,10]. Despite these intensive gene-specific studies, knowledge on light-regulation of gene expression is still fragmentary [11,12].
Light regulates organellar and nuclear gene expression [13,14]. It controls germination, photomorphogenesis and plant development [15,16]. Light signals are sensed by photoreceptors, such as phytochromes, cryptochromes and phototropins, and the photosystems [17,18]. Some light signals are directly targeted to promoter element-bound transcription factors [10], others are interwoven in developmental and stress-regulated signaling cascades [14,19]. Light, besides water and nutrients, is the most important environmental regulator in plants.
To screen for light-regulated enhancers, irrespective of the function of the regulated gene and the position relative to the transcription initiation site, we analyzed Arabidopsis thaliana GAL4-GFP enhancer trap lines [20,21] for light-responsiveness. If the enhancer trap T-DNA is inserted in the vicinity of an enhancer element, the GAL4-VP16 transcriptional activator is expressed. The GAL4-factor subsequently induces expression of an ER-targeted GFP (green fluorescent protein) (mGFP5ER) which can be monitored non-invasively [20,21].
Previously, enhancer trap lines have been used to identify regulatory sequences controlling developmental and organ-specific expression patterns, e.g., senescence [22], stomatal guard cell development [21] or lateral root development [23]. Here, we used GAL4 GFP enhancer trap populations (GAL4 GFP ET) to screen for light-responsive promoter elements irrespective of their location relative to the transcription start site. Lines with strong GFP expression in mesophyll cells were isolated and sub-selected for responsiveness to light intensity variation. A trap insertion in the distal part of the SIG5 upstream region gave novel insights into the light-regulation of a nuclear-encoded regulator of chloroplast gene expression. SIG5 (At5g24120) encodes the sigma factor, which activates expression of the D1 and D2 protein genes (psbA and psbD) inside chloroplasts [24]. It is required for light dependent regulation during the day [25] and, therefore, an essential factor in the nuclear control of chloroplast function. Here, we show that SIG5 is regulated by the combined action of multiple blueand red-light sensitive elements spread over the full promoter and by photosynthetic electron transport.

Selection of Enhancer-Trap Lines
To identify distal promoter elements involved in light-regulation of gene expression in mesophyll cells, two collections of light grown GAL4 GFP enhancer trap lines [20,21] were screened by fluorescence microscopy for mesophyll activity of the enhancer at an age of 7-13 days. At this age, plants are fully shifted from heterotrophic lipid consumption to photoautotrophy and the first true leaves develop [26].
For each selected line, twenty four 10 day old seedlings were screened for light-intensity regulation by comparison of the GFP-activity at 10, 100 or 200 µmol·photons·m −2 ·s −1 white light using a top reader fluorometer. The crude data were corrected by subtraction of the background fluorescence of wildtype C24 plants, which is caused e.g., by cell wall fluorescence. In several enhancer trap lines the GFP fluorescence positively correlated with the growth light intensity. The strongest gradual light-intensity regulation was observed for the lines N9266 and N9313 ( Figure 1A,B). Smaller leaves in low light and larger ones at higher light intensities can mock light intensity regulation, while longer hypocotyls would partly mask it. Corrections were performed by standardization of the GFP activity on the leaf area. In N9266 and N9313, light intensity regulation was pronounced confirming light-intensity dependent regulation. Light-intensity regulation in N9266 and N9313. Seedlings were grown for 10 days in 10, 100 and 200 µmol·photons·m −2 ·s −1 white light. One-way ANOVA was performed comparing all groups with Bonferroni's post-test. Statistical significance of difference is indicated as asterisks above bars (p < 0.05); (C and D): Transverse confocal section of spongy mesophyll of cotyledons. The fluorescence of ER-targeted GFP (mGFP5ER) is presented in green, Chlorophyll fluorescence in red.

Identification of Enhancer Trapped Sequences and T-DNA Numbers
For line N9266, hybridization of Southern-Blots with probes against the GAL4 element gave two bands ( Figure 2A). Since DNA was digested with BglI, which does not cut within the GAL4 element, two bands indicate two T-DNA insertions. In contrast, the hybridization pattern of line N9313 showed that the GFP fluorescence results from a single T-DNA insertion (Figure 2A).
To identify the T-DNA insertion site of N9313, 3-step thermal asymmetric interlaced PCR (TAIL-PCR; [27]) was performed with genomic DNA. The tertiary TAIL-PCR product was cloned into pJET1.2/blunt cloning vector (Fermentas, St. Leon-Rot, Germany) and sequenced using a vector specific primer. The sequencing product contained at one end the 35S-minimal promoter demonstrating that it is specific for the enhancer trap insertion site. Comparison of the sequence upstream of the 35S-minimal promoter with Arabidopsis thaliana genome sequence data revealed that the T-DNA of line N9313 is inserted 1198 bp upstream of the coding sequence of SIG5 (At5g24120) ( Figure 2B). The SIG5 gene encodes a sigma factor which is involved in diurnal light regulation of plastid gene expression [24,25]. The T-DNA is also inserted in the vicinity (1302 bp upstream) of the coding sequence of a protein with unknown function (At5g24130) (Figure 2A). Publicly available microarray data (eFP browser; [28]; data not shown) suggest that it encodes an almost seed-specifically expressed gene. The direction of the inserted enhancer-trap construct matches with the direction of the SIG5 gene, which is, like the reporter gene of N9313, leaf expressed and light-responsive [22,29,30]. Based on EST comparison, a 483 bp intron-containing 5'-UTR has been predicted for SIG5. Since it is only substantiated by two ESTs (and alternative transcription initiation cannot be excluded), we refer to the translation start site as position +1 for description of the SIG5 promoter in this manuscript.
The T-DNA insertion site was confirmed by PCR with primers designed to anneal to the genomic DNA flanking the insertion site and to the right border of the T-DNA. The enhancer trap element is inserted 715 upstream of the predicted transcription initiation site of SIG5 in N9313 and separates the SIG5 promoter in a proximal part controlling SIG5 transcription and a distal part regulating GFP expression ( Figure 2B).  At5g24120 (

Light Quality Regulation of SIG5 by the Distal and Proximal Upstream Regions
SIG5 has previously been described to be blue-light regulated in a light-intensity-dependent manner [29][30][31][32]. To test whether and to which extent blue-light regulation is mediated by the trapped promoter elements, GFP fluorescence activity was tested in N9313 plants grown for 10 days in blue light (470 nm, 10 µmol photons m −2 s −1 ). The reporter gene activity was only 1.25-fold higher in blue light than in darkness ( Figure 3A) demonstrating that it is not sufficient to mediate the strong blue-light induction reported in the literature [30].
In parallel, seedlings were grown under red light (>600 nm). Originally, these plants were thought to function as controls, since it was reported that SIG5 is not red-light regulated [31]. However, the GFP pattern showed that the distal upstream region is sensitive to red light ( Figure 3A).
For further analysis of light quality sensing, GFP mRNA levels (resulting from activation of the enhancer trap by the distal upstream region) and the SIG5 mRNA levels, which are under control of the proximal 715 bp of the upstream region and the SIG5-5'-UTR, were compared in the enhancer trap line N9313 and in C24, which is the wildtype background of N9313 and expresses SIG5 under the full-length promoter. RNA-decay analysis ( Figure A1) revealed that during a 24 h treatment 99% of SIG5 mRNA and 97% of GFP mRNA can be expected to depend on de-novo synthesis. Here, plants were pre-cultivated for 12 days in 100-120 µmol·photons·m −2 ·s −1 white light to minimize germination and early seedling development effects. Afterwards, they were transferred for 24 h into darkness for relaxation and exposed for 24 h to 100-120·µmol·photons·m −2 ·s −1 monochromatic blue (471 nm), red (673 nm) or far-red light (745 nm), respectively. Controls were kept in the dark. transcript levels in N9313 and C24. Ten day old light grown seedlings were dark adapted for 24 h and then exposed to 100-120 µmol·photons·m −2 ·s −1 monochromatic blue, red or far-red light for 24 h. Relative transcript levels were determined by qRT-PCR relative to ACT2 transcript level and normalized on the transcript levels of dark adapted seedlings prior to the transfer to monochromatic light. Bars represent means (±SEM). * indicates significant differences from the dark adapted samples (Student's t-test, p < 0.05). Blue light illumination resulted in low GFP mRNA levels in N9313 confirming that the distal upstream region is not blue-light sensitive. Consistent with the previous indication for red-light sensitivity of the distal upstream region, stronger GFP mRNA accumulation was observed in red-light illuminated N9313 seedlings ( Figure 3B). Far-red light resulted in only very low GFP expression.
SIG5 transcript levels, which are under control of the proximal upstream region in N9313, were 260-fold higher in blue light in the enhancer trap line than in darkness showing that the previously reported blue-light sensitivity is mediated within the proximal upstream region. Red light resulted in almost 100-times the dark-level and far-red light gave around 15-times the dark mRNA level of SIG5 in N9313 ( Figure 3D).
In C24 ( Figure 3C), the relative red-light and, especially, the far-red-light response were stronger compared to the blue-light reaction than in N9313 ( Figure 3D) demonstrating that SIG5 transcript abundance regulation in C24 combines the regulatory effects of proximal and distal upstream regulation in N9313 ( Figure 3B,D).

Arabidopsis SIG5 Transcription Is Red Light Sensitive
Some of the sigma factors, e.g., SIG2 and SIG5, are strongly induced in far-red light [33], while SIG5 is the fastest blue-light induced sigma factor. Based on previous analyses, it was more or less excluded that SIG5 is red-light sensitive [29][30][31][32]. Since most work was done on rosette leaves of 4 week old Arabidopsis plants of the accession Landsberg erecta (Ler) and Columbia-0 (Col-0), while we observed red-light responses in up to 2 week old C24 seedlings, we first adjusted our growth conditions, illumination period and the duration of dark-adaptation and red-light illumination according to [31]. SIG5 transcript levels were increased in 4 week N9313 plants after 3 h and after 24 h red-light illumination ( Figure 4A). In a test for accession specific variation, SIG5 mRNA levels were also increased in C24, Ler and Col-0 in response to red light ( Figure 4B).

In Silico Analysis of the SIG5 Promoter for Putative Light Responsive Elements
In an in silico scan of the 2 kb upstream of the CDS of SIG5 using the PlantCARE [34] and PLACE [35] databases for prediction of plant cis-acting regulatory DNA elements, 36 putative light-sensitive motifs were identified (Table 1)  10 day old light grown seedlings were dark adapted for 24 h and then exposed to 100-120 µmol·photons·m −2 ·s −1 monochromatic red light for 24 h. * indicates significant differences from the dark adapted samples (Student's t-test, p < 0.05). Table 1. Light responsive motifs identified in the 2 kb sequence upstream of the CDS of SIG5 as predicted by PlantCARE and PLACE databases. A graphical overview is provided in Figure A2.

Mapping Revealed Locally Restricted Blue-Light Sensitivity and Disperse Red-Light Sensitivity
To map the SIG5 light sensitive promoter elements, the SALK-collection of T-DNA insertion lines [36] was screened for lines which interrupt the SIG5 promoter at different positions. Five lines suitable for subsequent analysis were identified: SALK_015625 carries a T-DNA insertion at position −1618 relative to the SIG5 CDS, SALK_077048 at position −1032, SALK_072457 at −887, SALK_019261 at −691 and SALK_133729 at −515. The T-DNA insertion sites were confirmed by PCR with T-DNA border primers and primers flanking the insertion site. Homozygous lines were isolated from segregating populations by PCR-based genotyping. For the light response tests, 10 day old seedlings grown in white light were dark-adapted for 24 h and treated with monochromatic blue, red and far-red light for 24 h prior to RNA isolation.
In blue-light, the SIG5 transcript levels were similar in the three T-DNA lines with insertions at nucleotide positions −1618, −1032 and −887 in Col-0 ( Figure 5A). Seedlings with T-DNA insertions at position −691 and −515 were strongly impaired in their blue light induction of SIG5 transcription demonstrating that the 196-bp region between position −887 and position −691 mainly mediates blue-light regulation. Weaker blue-light sensitivity differences were observed between −515 and −691 in the proximal upstream region and upstream of −1618 in the distal region.
When the same lines were compared in red and far-red light, the strongest red-light difference was observed between the −887 and −1032 insertion lines ( Figure 5B). Weaker, but still significantly strong positive elements are located between −515 and −691 and between −691 and −887. In the distal upstream region negative red-light regulation was observed if a SALK-line T-DNA was inserted upstream of the position of the enhancer trap insertion in N9313 (−1089), as if the T-DNA insertion destroyed the otherwise inducing element. The overall low far-red-light sensitivity increased gradually with the length of the upstream region ( Figure 5C) indicating widely distributed regulation. sensitive regions of the Arabidopsis SIG5 promoter by qRT-PCR analysis of SIG5 transcript levels in Col-0 wild-type and SALK T-DNA insertion lines exposed for 24 h to monochromatic light after 10 days growth in darkness. The numbers give the distances of the T-DNA insertions relative to the SIG5 CDS. Values were normalized to the SIG5 transcript level of 24 h dark-adapted seedlings. The data are means of 2-4 biological replicates (±SEM). * indicates significant differences from the dark adapted samples (Student's t-test, * p < 0.05). ℓ indicates significant differences from the line with the next longer upstream region (Student's t-test, p < 0.05).

Blue-and Red-Light Regulation of the Proximal Upstream Region
As a first approach of signal transduction analysis, SIG5 transcript levels were analyzed in photoreceptor mutant lines ( Figure 6) grown next to their respective wild-type for 24 h in monochromatic light after 10 days pre-cultivation in white light and 24 h in darkness. In the phot1phot2 and the cry2 mutants, which are deficient in phototropins and cryptochrome 2, respectively [37,38], SIG5 transcript levels were not significantly different from the respective wildtype lines. In cry1 and cry1cry2 mutants [39], it was strongly decreased ( Figure 6A) demonstrating that CRY1 is involved in the blue light induction of SIG5. SIG5 transcript levels were also decreased in the homozygous offspring of N9313 × cry1cry2 ( Figure 6E) demonstrating that also the weak distal blue-light response is mediated by cryptochromes.
Blue light can alternatively be sensed via activation of the Soret absorption bands of phytochromes [40]. The comparison of phyA, phyB and the phyAphyB double mutants demonstrated that phyB mutants are impaired in the blue light induction of SIG5 transcription, whereas phyA mutants are not ( Figure 6B).
Comparison of phyA, phyB and phyAphyB mutants in red light showed that the red-light response of the SIG5-promoter is mediated by PHYB, but not by PHYA ( Figure 6C). Consistent with a general low FR-response, the transcript levels of SIG5 were low in phyA, phyB and phyAphyB mutants ( Figure 6D).

Photoreceptor Control of the Distal SIG5 Upstream Region
For the analysis of the regulatory function of photoreceptors on the distal SIG5 upstream region, the enhancer trap line N9313 was crossed with phot1phot2 [37] and with cry1 and cry2 [39] mutants. Lines homozygous for the T-DNA and the respective mutant were selected and analyzed for GFP Due to the severe differences in growth and greening between the mutants and wildtype plants, the reporter gene activities were standardized on chlorophyll-a levels after blue-light treatment ( Figure 6F) and on fresh weight after red-light treatment ( Figure 6G). The GFP fluorescence of N9313 x phot1phot2 mutant plants was similar to that of N9313. The GFP values of N9313 × cry1 were only slightly decreased, while the GFP fluorescence of cry1cry2 double mutants were hardly detectable ( Figure 7A), demonstrating that regulation of the distal upstream region depends on the availability of CRY2.  Relative SIG5 transcript levels of old HY5 deficient hy5-1 and SALK_056405C and the respective wild-type in response to blue light or red light. Ten day old seedlings were dark adapted for 24 h and then exposed to 100-120 µmol photons m −2 s −1 monochromatic blue light or red light for 24 h. Relative transcript level of dark adapted seedlings prior to the transfer to monochromatic light was set to 1.0. (C) GFP fluorescence in N9313 and N9313 crossed with the HY5 deficient SALK_056405C (mean values of 4-8 samples with 10 seedlings each (±SEM). (D) Relative SIG5 transcript abundances in Col-0 wild-type and in cop1 mutants as determined by qRT-PCR relative to ACT2 values. Light grown seedlings were dark-adapted for 24 h prior to RNA isolation. SIG5 transcript levels of Col-0 were set to 1.0. * indicates significant differences from the N9313 or Col-0 value (Student's t-test, p < 0.01).
The reporter gene activity was also decreased in crosses of N9313 to phyA and phyB ( Figure 6G). In the phyAphyB double mutant it was less than in each single mutant indicating a combined control of the promoter. Besides PHYB, Arabidopsis expresses three more type-II phytochromes. To investigate their overall impact, N9313 was crossed with the hy1 mutant, which is deficient in phytochromobiline biosynthesis [41]. In response to the hy1 mutation GFP activity was slightly more decreased than by phyAphyB ( Figure 6G), demonstrating that the distal upstream region of SIG5 is regulated by the combined action of various phytochromes, with strongest impact of PHYA and PHYB.

Regulation of the Distal SIG5 Upstream Region by HY5 and COP1
The bZIP transcription factor HY5 translates cryptochrome and phytochrome signals into gene expression regulation [42]. Its stability is under control of COP1 [43]. In the two HY5-deficient lines, hy5-1 and SALK_056405C, SIG5 transcript abundance was strongly decreased in red and in blue light ( Figure 7A,B). qRT-PCR analysis showed a very strong accumulation of SIG5 transcripts in dark-adapted cop1-6 seedlings ( Figure 8D), confirming that SIG5 transcription is under control of HY5.
In the homozygous offspring of the cross of N9313 with the hy5-1 mutant, GFP fluorescence was reduced to almost undetectable levels ( Figure 7C) demonstrating that HY5 has strong impact on the distal upstream region.  (Table 2). To test whether the predicted motifs are involved in light quality regulation, they were amplified by PCR and fused to a truncated 35S promoter upstream of an open reading frame for GFP-GUS. In a PCR-based approach [44] the putative HY5 sites were mutagenized. Young tobacco leaves were transfected with mutagenized and non-mutagenized constructs. Expression activity was quantified from the GUS activity in leaf disks excised from the transfection site. As negative control, the minimal 35S-promoter (without regulatory upstream elements) was fused to GFP-GUS. The comparison of mutagenized and non-mutagenized constructs showed white-and red-light sensitivity for the G-box, but not for the other two predicted elements ( Figure 8A).

Yeast-1-Hybrid Analysis of HY5 Binding to the SIG5 Promoter
To test for direct binding of HY5 to the SIG5 promoter, yeast-one-hybrid assays were performed with SIG5 promoter fragments as baits and HY5 as prey. The promoter of RUBISCO small subunit 1A (pRBCS1A) was used as positive control [43]. HY5 decreased auto-activation of the proximal SIG5 upstream region in yeast demonstrating binding of the plant transcription factor ( Figure 8B). Analysis of the distal SIG5 upstream region gave no indication for direct HY5 interaction with the distal SIG5 promoter ( Figure 8B mid).

Impact of Photosynthetic Electron Transport on Light Regulation
Red light can also drive photosynthesis. Monochromatic red-light is predominantly absorbed by photosystem II [47]. During acclimation, expression of psaAB, which encodes the reaction center proteins of photosystem I, decreases and expression of the photosystem-II reaction center protein D1 (encoded by psbA) increases in order to optimize photosynthetic efficiency [48]. SIG5 regulates transcription of the photoreaction center proteins D1 and D2 (encoded by psbA and psbD) [24,32]. Here, SIG5 expression was shown to be light-intensity dependent ( Figure 1) and red-light regulated ( Figure 9).
To test the importance of photosynthetic electron transfer on SIG5 expression, half of the 8 day old seedlings grown in white light were sprayed with 10 µM DCMU prior to shifting them for 24 h to red light. DCMU blocks the Q B -site [49]. Electrons cannot be transferred from photosystem II to the plastoquinone pool. Therefore, DCMU should simulate over-excitation of photosystem II and function antagonistically to non-saturating doses of red light.
In C24, DCMU application decreased red-light dependent induction of the SIG5 promoter ( Figure 9A). In N9313 the SIG5 transcript levels were hardly detectable after DCMU treatment ( Figure 9B). The results demonstrated that regulation of the proximal upstream region depends strongly on photosynthetic electron transport. Red-light regulation of the distal upstream region is only partly dependent on electron transport activity as indicated from a weaker decrease in response to DCMU ( Figure 9C).

Figure 9.
(A-C) Relative SIG5 and GFP transcript levels in C24 and N9313 in response to red and far-red light and red light 24 h after treatment with 10 µM DCMU and light exposure. The seedlings were pre-cultivated 10 days in 120 µmol photons m −2 s −1 white light followed by 24 h dark incubation. The transcript levels were standardized on ACT2 transcript levels and normalized to the transcript levels prior to the transfer to monocromatic light. Statistical significance of difference (Student T-Test) is indicated as asterisks above bars (p < 0.05); (D-F) GFP expression mediated by the 2 kb full-length SIG5 promoter, the SIG5 proximal and SIG5 distal upstream regions fused to the Δ35S promoter and the Δ35S promoter after 4 days cultivation of the transfected tobacco leaves in monocromatic red and far-red light. In 1/3 of the plants the transfected leaves were sprayed with 1 mM DCMU prior to the exposure to the monochromatic light. The scale of the false color pictures was standardized on the same color scale.  To test how strong the regulation by photosynthetic electron transport is manifested in the promoter, red-light-and DCMU-responses were analyzed in young tobacco leaves which were transfected with constructs expressing GFP under the control of the SIG5 full length promoter ( Figure 9D,E), the distal or proximal upstream regions (fused to the Δ35S minimal promoter), or the Δ35S minimal promoter. Similar as in seedlings, DCMU antagonized the red-light dependent induction of the full-length, the distal and the proximal SIG5 upstream regions.
In non-transfected plant material, the redox state of the plastoquinone pool (1-qP) and the quantum yield of photosystem II (F V /F M ) were determined ( Figure 10). The redox state of the plastoquinone pool was highly oxidized in red light in the presence and absence of DCMU. In red light the quantum yield of photosystem II was slightly higher than in white light in tobacco and in the range of white light treated plants in Arabidopsis in absences of DCMU and decreased after DCMU treatment reflecting PSII damage.

Sugar Effect on SIG5 Transcription
Intensity dependent light responsiveness can be mediated by photosynthate signatures. Vice versa, high levels of carbohydrates can indicate high photosynthetic activity. As part of the feed-back mechanisms, they inhibit the activity of the Calvin-Benson-Cycle and, therefore, the main sink for electrons transported in the photosynthetic light reaction. The plastoquinone pool is more reduced and the quantum yield on photosystem II decreased [50].
The impact of carbohydrates on the light-regulated SIG5 promoter was tested by application of sucrose. The osmotic side effects of sucrose were simulated by sorbitol. Forty eight hours after application of sucrose or sorbitol to 10 day old light-grown seedlings, the SIG5 mRNA and GFP mRNA levels were quantified in C24 and N9313. Sucrose resulted in decreased GFP-levels in N9313. SIG5 transcript amounts were decreased in C24, but not in N9313 ( Figure 11) demonstrating that only the distal upstream region is sucrose-responsive. Sorbitol did not alter SIG5 expression significantly. Figure 11. Relative SIG5 and GFP transcript levels in C24 and N9313 in response to sucrose and sorbitol. The transcript levels were standardized on ACT2 and normalized on the transcript levels determined on 0% sucrose. Statistical significance of difference (Student's t-test) is indicated as asterisks above bars (p < 0.05).
For sterile growth, Arabidopsis seeds were surface sterilized for 1 min in 80% (v/v) ethanol followed by 8 min in 20% (v/v) household bleach. After five washing steps in sterile water, they were transferred

Extraction of Genomic DNA, TAIL-PCR and Sequence Analysis
For small scale isolation of genomic DNA, plant material was homogenized in 200 µL 50 mM Tris/HCl pH 8.0, 25 mM EDTA, 250 mM NaCl and 0.5% (w/v) SDS. After extraction with 200 µL phenol-chloroform-isoamyl alcohol (25:24:1), the samples were centrifuged at 20,000 ×g for 5 min at room temperature. The DNA was precipitated for 1 h at −20 °C after transfer of the upper phase to 200 µL isopropanol. Following 15 min centrifugation at 20,000 ×g, the precipitated DNA was resuspended in 100 µL water.
TAIL-PCR was performed with 2 µL DNA solution per 20 µL total volume as described in [27] with the primers listed in Table 3. PCR products were cloned into pJET1.2/blunt (Fermentas/Thermo-Fischer, St. Leon-Rot, Germany). Sanger sequencing was performed by GATC (Konstanz, Germany). Database searches were performed using the BLAST program on the NCBI server [51]. The insertion sites were confirmed by PCR, designed to anneal the genomic DNA flanking the mapped insertion sites and T-DNA primer TAIL-TR3.2.

Southern Blot Analysis
DNA was extracted from plant material using the DNeasy Plant Mini Kit (Qiagen, Hilden, Germany). Three µg of genomic DNA was digested with BglII (Fermentas) and purified by sodium acetate precipitation prior to separation on 1% (w/v) TAE agarose gels and capillary transfer onto Hybond TM -N membranes (GE-Healthcare, Freiburg, Germany). A 504 bp digoxigenin (DIG) labeled GAL4 probe was amplified from genomic DNA of N9313 with the DIG Probe Synthesis Kit (Roche, Germany) by PCR using primers mPPR1-5 (5'-CGGCAAGCTTGGATCCAACAATG-3') and mPPR1-3 (5'-CCCGG AGCTCGTCCCCCAGGCTG-3'). After crosslinking of the DNA to the membrane by UV-light, the blot was hybridized under high stringency conditions using DIG Easy Hyb Granules (Roche, Mannheim, Germany). Following two washing steps at 68 °C in 75 mM NaCl, 7.5 mM sodium citrate (pH 7.0) with 0.1% (w/v) SDS and two in 15 mM NaCl, 1.5 mM sodium citrate (pH 7.0) with 0.1% (w/v) SDS, the bound probes were detected in a Luminescent Image Analyzer LAS-4000 (GE-Healthcare) using the CDP-Star, ready to use Kit (Roche).

Identification and Isolation of Homozygous T-DNA Insertion Lines by PCR
T-DNA insertion lines were tested for the T-DNA insertion site and homozygosity by PCR with T-DNA border primer LBb1.3 (5'-ATTTTGCCGATTTCGGAAC-3') and specific primers (Table 3) binding upstream and downstream of the predicted T-DNA insertion site. For confirmation, the PCRs were repeated with 10 seedlings of the progeny.

Determination of Reporter Gene Activity
GFP activity screens were performed in 96-well microtiter plates on 100 µL MS medium per well supplemented with 0.5% (w/v) sucrose in a Fluoroskan Ascent FL fluorometer (Thermo Fisher Scientific, St. Leon-Rot, Germany) in the top-reader modus with 500 ms integration time, 485 nm excitation at 527 nm. Each well was covered with nine measurement points to scan all plants equally.
Background fluorescence was subtracted from the means by analyzing the autofluorescence at 527 nm of parallel grown C24 seedlings. Normalization was performed based on area determination using the ImageJ software package [52].
In crosses of N9313 with the photoreceptor mutants, the GFP activity was quantified in triplicates from 100 µL extracts because of the high variability in seedling morphology. Plant material (10-20 mg) was homogenized in 500 µL sodium phosphate, pH 7.0. After 2 min centrifugation at 20,000 ×g GFP fluorescence was quantified in a Mithras LB 940 top reader fluorometer (Berthold Technologies GmbH & Co. KG, Bad Wildbad, Germany) with 100 ms counting time (excitation filter 460/10 and emission filter F510). Chlorophyll contents were determined according to [53].
GUS expression was quantified in 15-30 mg leaf material and standardized on protein levels as described in [54].

Transcript Abundance Analysis
RNA was isolated from 10-15 frozen seedlings in Precellys ® 24 (Peqlab, Erlangen, Germany) using the RNeasy Plant Mini Kit. The purity of the RNA was determined spectrophotometrically from the A 260 /A 280 ratio. cDNA was synthesized with the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Darmstadt, Germany). Primers for qRT-PCR were designed by the QuantPrime software [55,56] (Table 3). The primer specificity was checked by assaying that the melting curves display only a single peak for each PCR product of interest, and by gel electrophoresis. Real-time amplification was performed according to the MIQE standards [57] using the Brilliant II SYBR ® Green Master Mix (Agilent Technologies, Böblingen, Germany) in a Stratagene MX3005P Cycler (Agilent Technologies, Böblingen, Germany). No-template-controls were performed for each gene. Each biological replicate represents an independent RNA isolation. Levels of each transcript relative to the constitutively expressed ACT2 control gene [58] were quantified as described by Pfaffl [59].

mRNA Decay Analysis
Transcription was inhibited by transferring 10 day old N9313 seedlings to liquid MS medium containing 200 µM Actinomycin D (Act D) (Sigma-Aldrich, Taufkirchen, Germany) at 120 µmol photons m −2 s −1 . Control plants were incubated in MS medium. The mRNA half-life was quantified for SIG5 and GFP transcripts from the C t values determined by ACT2-standardized qRT-PCR in samples treated for 1, 2, 4 and 8 h. Since mRNA decay generally obeys first-order kinetics [60,61], an exponential regression model (A = 1e −kt ) was applied for determination of the decay coefficient (k decay ). The mRNA half-life was then calculated using the following equation: t½ = ln(2)/k decay .

Reporter Gene Construct Design, Site-Directed Mutagenesis and Tobacco Transfection
SIG5 promoter fragments were amplified from genomic DNA of the accession C24 by PCR with specific primers containing PstI and a BcuI restriction sites (Table 3) and cloned into the TA cloning site of the pCR8/GW/TOPO vector (Invitrogen, Darmstadt, Germany). The truncated (-48) 35S minimal promoter from cauliflower mosaic virus (CaMV) which was amplified by PCR from T-DNA of the enhancer trap line and cloned into the restriction sites. The promoter fragments were transferred into the binary vector pHGWFS7.0 [62] with LR Clonase II enzyme mix (Invitrogen, Darmstadt, Germany). Agrobacterium tumefaciens strain GV3101 [63] was transformed using the freeze-thaw method [64].
For transfection of tobacco leaves [65], transformed Agrobacteria were cultivated in 10 mL YEB medium at 28 °C and 180 rpm to OD 600 0.5. For co-infiltrations, GV3101(pMP90) and the respective SIG5-promoter strain were mixed in a 2:3-ratio. Cells were collected by 8 min centrifugation at 3000 ×g at room temperature and resuspended in 100 mM MES pH 5.6 plus 10 mM CaCl 2 supplemented with 150 µM acetosyringone. After 2 h incubation, the agrobacteria suspensions were infiltrated into tobacco leaves according to [66]. Site-directed mutagenesis was performed according to [44].

Yeast One-Hybrid Analysis
SIG5 and RBCS1A promoter fragments and HY5 cDNA were generated by PCR using genomic DNA of the ecotype C24 as template. The HY5 cDNA was cloned into the BamHI/XhoI site of pACT2 (Clontech, Montain View, CA, USA). SIG5 promoter fragments and a 196-bp fragment of the Arabidopsis RBCS1A promoter were cloned into the SacI/EcoRI site of pHIS2 (Clontech, Mountain View, CA, USA) using the LigaFast Rapid DNA Ligation System (Promega, Fitchburg, Madison, WI, USA). The Saccharomyces cerevisiae strain Y187 (MATα,112, gal4Δ, met-, gal80Δ, URA3:GAL1 UAS GAL1 TATA -lacZ, MEL1, with reporter genes HIS3 and lacZ) was transformed with lithium acetate (LiAc) [69]. Transformants were selected on minimal synthetic dropout (SD) medium (Clontech, Montain View, CA, USA) lacking the amino acids corresponding to the autotrophy markers of the plasmids. For yeast colony-PCR, a single colony was resuspended in 50 µL water supplemented with 60 U mL −1 lyticase and incubated at 30 °C for 30 min. After cell lysis at 95 °C (10 min) and 2 min centrifugation at 20,000 ×g, 2 µL of the supernatant was used for PCR in a total volume of 20 µL.

Crossing Arabidopsis thaliana Plants and Mutant Selection
Arabidopsis lines were cross-pollinated. From the progenies, homozygous cry1, phyA or phyB mutants were selected according to their hypocotyl length in blue, red or white light, respectively [18]. Plants carrying the homozygous phot1 allele were identified by measuring the phototropic response [70]. The presence of the GFP-enhancer element was tested by PCR.

Fluorescence Microscopy
Tissue-specific GFP expression was analyzed by confocal laser scanning microscopy using a Zeiss LSM 510 Meta with a multiline argon ion laser (Carl Zeiss, Jena, Germany). Cells were examined with a 40X Zeiss oil-immersion objective (1.3 numerical aperture). GFP was excited at 488 nm and the emission was recorded through the meta-channel at 497-550 nm. Fluorescence images were analyzed with LSM Image Browser software [71].

Chlorophyll-a Fluorescence Analysis
The quantum yield of photosystem II and the redox state of the plastoquinone pool were determined with a Dual-PAM-100 (Heinz Walz, Effeltrich, Germany) at room temperature after 20 min dark incubation with saturating light flashes of 6000 µmol photons m −2 s −1 as F V /F M [72]. The redox state of the plastoquinone pool was calculated from the steady state photochemical quench (qP) in 100 µmol photons m −2 s −1 red light (600 nm) as (1-qP) with light flashes of 6000 µmol photons m −2 s −1 .

Transcriptional Regulation of SIG5 Promoter
SIG5 is one of six sigma factors of Arabidopsis [73,74] controlling plastid gene expression [75]. It activates psbD and psbA transcription upon dark-light shifts and controls diurnally chloroplast gene expression in a light-dependent manner [24,25,32]. Whereas the structure of the psbD promoter was a subject of intensive investigation [76,77], the structure of the SIG5 promoter has been barely investigated. From transcript abundance analysis [24,29,30] and GUS-reporter gene assays [32], SIG5 transcription was shown to be strongly blue-light responsive and stress-sensitive. Here, we showed that it is also red-light responsive in a photoreceptor-and photosynthesis-dependent manner (Figures 6, 7 and 9).
In blue light, SIG5 is the fastest transcriptionally activated sigma factor [30]. SIG5 binds to the −948 upstream region of the psbD promoter, which drives the psbD-psbC-psbZ operon encoding core-proteins of photosystem II, in response to blue and white light and various stress conditions [32,76]. SIG5 is hardly expressed in far-red light ( Figure 3C and [33]) and, therefore, can be assumed to be less important than other sigma factors, especially SIG2 and SIG6, in shade avoidance reactions, such as hypocotyl elongation, and activation of greening upon germination [78,79]. As shown by promoter-GUS analysis [32], SIG5 transcription activity quickly decreases three and four days after onset of germination and becomes activated again in rosette leaves. SIG5 transcript abundance increases in response to high light, NaCl, mannitol and cold in rosette leaves [32]. Analysis of SIG5-deficient T-DNA insertion lines demonstrated that SIG5 is essential for maintaining photosynthetic activity upon stress [32].
In contrast to the blue/far-red type of sigma factors, e.g., SIG2 and SIG6, SIG5 is induced by red and blue light (Figures 3 and 5). The regulatory elements are spread widely over the proximal and distal upstream region. Comparison of distal and proximal regulatory activities ( Figure 3) and stepwise deleted promoters ( Figure 5) show complex combinational regulation. Most of the red-light induction, part of the photosynthetic control and the sugar regulation take place in the distal upstream region (Figures 3, 9 and 11), while the proximal upstream region mediates the main blue-light sensitivity ( Figure 5).

Blue-light Regulation
The 196 bp blue-light responsive region, located 208-404 bp upstream of the predicted transcription initiation site ( Figure 5A), is under control by CRY1 and PHYB ( Figure 6A,B). CRY2 and the phototropins PHOT1 and PHOT2 are of negligible importance ( Figure 6A).
The CRY2-mediated blue-light responsive enhancers of the distal upstream region ( Figures 3A,B  and 5A) may support CRY1 regulation of the distal upstream region ( Figure 7A) under low light, such as upon shading or during germination. In the rosette stage, Arabidopsis leaves are frequently exposed to high light intensities. Under these conditions, CRY2 can be expected to be light inactivated [80,81].

Red-light Sensitivity
SIG5 transcription and transcript abundance were regulated by various phytochromes via widely dispersed red-light sensitive promoter elements ( Figures 5 and 6). The region −887 to −1032 in the proximal upstream region showed preference toward PHYB-mediated regulation ( Figure 6B), which is also the most abundant phytochrome in light-grown Arabidopsis [82].
In blue-light around 78% and in red-light 59% of the SIG5 transcripts depended on PHYB after 24 h ( Figure 6B,C). In a similar experiment, in which light-grown Arabidopsis plants were transferred back to light after 30 h dark treatment, SIG5 transcript levels were wild-type-like in phyA and phyB mutants after 1.5 h of illumination [32]. Differences can be observed after 6 h. At this time, the overall SIG5 expression was low in the analysis [32]. The comfortable access to a wide range of transcript abundance studies from various labs via tools like the eFP browser [28] showed strong diurnal regulation of SIG5 expression. Therefore, we compared transcript levels exactly after 24 h (except Figure 4A; 3 h data) and harvested plant material at the time of highest transcript abundance 2-4 h after onset of light. This approach demonstrated that SIG5 expression is also strongly red-light sensitive. Light signals are transmitted mainly via CRY1, PHYA and PHYB in a HY5 and COP1 dependent manner. Consistent with ChIP-based prediction of HY5-binding sites in the SIG5 promoter, [83] and previously shown HY5-dependent transcript abundance regulation [32], HY5 binds to the proximal upstream region ( Figure 8B) and a HY5-binding motif, a G-box, was shown to transmit white-and red-light signals ( Figure 8A).

Antero-and Retrograde Signaling Connecting Chloroplast and Nuclear Gene Expression
SIG5 expression is promoted in red-and blue-light ( Figure 3) and in a light intensity dependent manner (Figure 1). Depending on SIG5 regulation, chloroplast gene expression and photoprotection are regulated [24,32] indicating that its main function of SIG5 is anterograde light acclimation, such as compensation of photosynthetic imbalances due to suboptimal photoreaction center innervation.
In addition to the anterograde function [32], SIG5 is regulated by retrograde signals (Figure 9): Red light preferentially activates photosystem II [84]. Blocking the Q B -site of photosystem II with DCMU [49], decreased activation of SIG5 expression in red light (Figure 9). In red light and after DCMU treatment, the plastoquinone pool was oxidized as determined by chlorophyll-a fluorescence (1-qP; Figure 10). This result excludes the possibility of regulation by the redox state of the plastoquinone pool and showed that full SIG5 expression depends on photosynthetic electron transfer from photosystem II to the plastoquinone-pool.
psbA and psaAB expression respond to light-quality changes in order to adjust photoreaction center I and II availability to the relative innervation of the photosystems [84]. Transcription of psaAB is under control of the redox state of the plastoquinone pool, but not psbA [85]. SIG5 controls expression of psbD and, less specifically, psbA [32]. It is the anterograde control factor [32] which can adjust plastid gene expression to light signals due to its complex light quality sensitivity (Figures 1, 3 and 5).
Excess carbohydrates inhibit photosynthetic electron transport [50,86]. Like DCMU, sucrose suppresses SIG5 expression ( Figure 11). If SIG5 would be regulated depending on the redox state of the plastoquinone pool, sucrose application should have resulted in a higher reduction status. The sucrose data give additional evidence for independence from plastoquinone redox regulation, which controls state transition [87], expression of photosystem I core proteins [84] and STN7-dependent regulation of nuclear gene expression [88]. Regulation of SIG5 expression by red light and DCMU (Figures 9 and 11) demonstrates that transcription of the nuclear encoded SIG5 gene responds to photosynthetic signals.
The antagonism of the sucrose effect to the red-light effect excludes carbohydrates as direct signaling molecules in chloroplast-to-nucleus transmission of the photosynthetic electron transport signal. Insensitivity to similar concentrations of sorbitol ( Figure 11B,C) and induction by higher concentrations of sorbitol ( Figure 11C) and mannitol [32] excludes osmotic signaling. The comparison of the red-light and DCMU response (Figures 3 and 9) and the general stress-induction reported before [32], which decrease photosystem II activity [89], let assume that the signal emerges at PSII in response to high excitation pressure. A link between generation of signaling by D1 damage or the D1 repair cycle, demonstrates that SIG5 controls a perfect regulatory circuitry: In chloroplasts, D1 damage reflects over-excitation of photosystem II [90]. In response to retrograde signals, SIG5 drives expression of D1 and its closest protein partners D2, PsbC and PsbZ [24,32] and stabilizes photosystem II activity ( Figure 12).

Conclusions
SIG5 expression is regulated by multiple light-responsive elements spread over the 2 kb upstream region ( Figure 5). Signal transduction is controlled by cryptochromes, especially CRY1, and phytochromes ( Figure 6) in a HY5 and COP1-dependent manner (Figure 7) and by photosynthetic signals (Figure 9). We conclude that the chloroplast sigma factor SIG5 is a retrograde and light controlled regulator of chloroplast function. It combines intrinsic and extrinisic information important in adjusting nuclear and plastid gene expression upon light acclimation processes and links the signaling and regulation potential of the eukaryotic-type extra-plastidic gene expression system with the chloroplast regulatory system. Figure 12. SIG5 expression control in the light-signaling pathway regulating the expression of the chloroplast genes psbD and psbA for the photoreaction center II core proteins D1 and D2 in response to light intensity and quality. Light signals regulate SIG5 expression by photoreceptors in a HY5-and COP1-controlled manner via various promoter elements. SIG5 is post-translationally targeted to chloroplasts, where the sigma factor is involved in tuning expression of psbD and psbA. Photoreaction center expression is involved in the PSII repair cycle. Retrograde signals transmit information correlating with the photosystem II excitation status and damage, respectively, to the nucleus to adjust the intensity of SIG5 transcription.

Acknowledgements
This work was supported by the German Research Foundation (Grant Ba2011/5-1), the Heinrich-Heine-Universität Düsseldorf and the Freie Universität Berlin. We thank Alfred Batschauer for providing seed stocks of cry2, phot1phot2, phyA, phyB phyAphyB and cop1. Special thanks to Rüdiger Simon for hosting M.M. for the final phase of her project, to our former lab neighbor, Peter Jahns, for providing his equipment for chlorophyll-a fluorescence analysis after we had moved ours to Berlin, and to Rainer Bode for critical reading.

Conflicts of Interest
The authors declare no conflict of interest.