A nuclear-encoded sigma factor, Arabidopsis SIG6, recognizes sigma-70 type chloroplast promoters and regulates early chloroplast development in cotyledons

Authors


(fax +81 75 703 5448; e-mail shiina@kpu.ac.jp).

Summary

Eubacterial-type multi-subunit plastid RNA polymerase (PEP) is responsible for the principal transcription activity in chloroplasts. PEP is composed of plastid-encoded core subunits and one of multiple nuclear-encoded sigma factors that confer promoter specificity on PEP. Thus, the replacement of sigma factors associated with PEP has been assumed to be a major mechanism for the switching of transcription patterns during chloroplast development. The null mutant (sig6-1) of plastid sigma factor gene AtSIG6 exhibited a cotyledon-specific pale green phenotype. Light-dependent chloroplast development was significantly delayed in the sig6-1 mutant. Genetic complementation of the mutant phenotype by the AtSIG6 cDNA demonstrated that AtSIG6 plays a key role in light-dependent chloroplast development. Northern and array-based global analyses for plastid transcripts revealed that the transcript levels of most PEP-dependent genes were greatly reduced in the sig6-1 mutant, but that the accumulation of nuclear-encoded RNA polymerase (NEP)-dependent transcripts generally increased. As the PEP α subunit and PEP-dependent trnV accumulated at normal levels in the sig6-1 mutant, the AtSIG6 knockout mutant probably retained functional PEP, and the transcriptional defects are likely to have been directly caused by AtSIG6 deficiency. Most of the AtSIG6-dependent genes are preceded by σ70-type promoters comprised of conserved −35/−10 elements. Thus, AtSIG6 may act as a major general sigma factor in chloroplasts during early plant development. On the other hand, the mutant phenotype was restored in older seedlings. Arabidopsis probably contains another late general sigma factor, the promoter specificity of which widely overlaps with that of AtSIG6.

Introduction

Plastids evolved from a cyanobacterial ancestor and have retained several prokaryotic characteristics in their transcriptional and translational apparatuses. Higher plant plastids contain two types of RNA polymerases: a bacteria-type multi-subunit RNA polymerase (PEP) and a nuclear-encoded bacteriophage-type RNA polymerase (NEP) (Hedtke et al., 1997; Hess and Börner, 1999; Liere and Maliga, 2001). These RNA polymerases are responsible for the transcription of distinct types of plastid genes (Allison et al., 1996; Hajdukiewicz et al., 1997). Plastid-encoded photosynthesis genes are transcribed exclusively by PEP (class I genes). On the other hand, non-photosynthetic housekeeping genes are mostly transcribed by both PEP and NEP (class II genes), whereas a few genes, such as rpoB and accD, are transcribed exclusively by NEP (class III genes). The timed expression of photosynthetic genes during plant development is dependent on an increase in PEP transcription activity during chloroplast development (Baumgartner et al., 1993; Bisanz-Seyer et al., 1989). Thus, PEP likely plays an important role in early plant growth.

PEP originates from ancestral cyanobacteria and consists of a core catalytic complex (α2βββ′′), whose subunits are encoded by plastid genes, and a sigma factor that is presumed to confer promoter specificity to the core complex (Allison, 2000; Liu and Troxler, 1996; Tanaka et al., 1996). Most eubacterial genomes encode several sigma factors, each of them recognizing a set of promoters (Ishihama, 2000). Bacterial sigma factors have been grouped into two families, the σ70 and σ54 families (Wösten, 1998). The σ70 family is further divided into three subgroups, namely the primary sigma factors (group 1), which are essential for cell growth; non-essential primary-like sigma factors (group 2); and so-called alternative sigma factors (group 3), which direct RNA polymerase to initiate transcription at specific promoters in response to environmental signals (Lonetto et al., 1992). Similarly, the ancestor of chloroplasts, cyanobacteria, also contains multiple σ70-type sigma factors that exhibit promoter preference. The chromosome of Synechosystis sp. Strain PCC6803 contains nine sigma factors: one primary sigma factor, four primary-like sigma factors, and four alternative sigma factors (Kaneko et al., 1996).

Multiple sigma factors have also been identified in the plastids of higher plants. In Arabidopsis, sigma factor genes form a small gene family consisting of six genes (AtSIG1AtSIG6) (Fujiwara et al., 2000; Hakimi et al., 2000; Isono et al., 1997; Tanaka et al., 1997). All of the plastid sigma factors are related to bacterial groups 1 or 2 sigma factors of the σ70 family and contain conserved subdomains, including regions 2.4 and 4.2, that are responsible for promoter sequence recognition (Lonetto et al., 1992). No σ54-like sigma factor has been identified in plants (Allison, 2000). Phylogenetic analysis of plant sigma factors suggests that they may be grouped into several subgroups (Allison, 2000; Fujiwara et al., 2000; Tsunoyama et al., 2004).

Analogous with bacterial transcription regulation, the replacement of plastid sigma factors has been postulated to play a key role in PEP regulation (Allison, 2000). Interestingly, all of the plastid sigma factors are nuclear-encoded and are likely to mediate direct nuclear control on plastid gene expression (Liu and Troxler, 1996; Tanaka et al., 1996). Standard promoters recognized by bacterial primary sigma factors, such as Escherichia coliσ70, are characterized by consensus −10 (TATAAT) and −35 (TTGACA) core promoter elements (σ70-type promoter). Most of the PEP-dependent photosynthesis and housekeeping genes (class I and class II) are preceded by σ70-type promoters. Intensive promoter mutation analyses revealed that these promoter elements were essential for transcription initiation by PEP in vitro and in vivo (reviewed in Hess and Börner, 1999; Liere and Maliga, 2001; Link, 1996; Stern et al., 1997; Sugiura, 1992). In addition to the standard σ70-type PEP promoters, some PEP promoters exhibit unique promoter architectures, such as the psbD blue light-responsive promoter (psbD LRP) which lacks a functional −35 element (Kim et al., 1999; Nakahira et al., 1998; Thum et al., 2001; To et al., 1996). These facts suggest that chloroplasts may contain at least two types of sigma factors. One of these groups would contain general sigma factor(s) responsible for the transcription of many photosynthesis and housekeeping genes. On the other hand, some factor-dependent promoters, such as psbD LRP, may be transcribed by PEP holoenzyme-containing specialized sigma factors. We recently found that AtSIG5 acts as a specialized sigma factor in chloroplasts and is responsible for the light-dependent transcription at the psbD LRP (Tsunoyama et al., 2004). However, general sigma factors have not yet been identified in chloroplasts.

We identified a T-DNA-generated mutation in the AtSIG6 gene of Arabidopsis. Inactivation of AtSIG6 resulted in chlorophyll-deficient cotyledons in young seedlings and reduced the transcription of the majority of PEP-dependent genes preceded by σ70-type promoters. AtSIG6 likely acts as a general sigma factor in chloroplasts and is probably responsible for the recognition of σ70-type standard PEP promoters in young cotyledons. The present study also suggested the presence of another general sigma factor that functions mainly at a later stage of seedling growth.

Results

T-DNA insertion inactivated AtSIG6 in the sig6-1 mutant

To date, AtSIG6 homologs have been found in Arabidopsis (AB029916; At2g36990), maize (AF099112) and rice (AB096012) and can probably be classified into a structurally conserved SIG6 group (Allison, 2000; Fujiwara et al., 2000). A detailed comparison of the amino acid sequences of plastid sigma-like factors with E. coliσ70 suggested that AtSIG2 and AtSIG6 are the most similar to the E. coli primary sigma factor, σ70 (Privat et al., 2003). To analyze the function of AtSIG6 in plants, we obtained seeds with a T-DNA insertional mutant (Figure 1a) in the AtSIG6 gene (Garlik 893c09.b.1a.Lb3Fa) from a collection of Arabidopsis (Columbia ecotype) T-DNA insertion lines generated by Syngenta. The insertion point of the left and right ends of the T-DNA were mapped by sequencing the PCR products generated by the primer pairs RB3 and SIG6-P2 (left end), and LB3 and SIG6-P1 (right end). The mutant allele contains a T-DNA insertion in exon 4 of AtSIG6, 407 bp downstream of the translation start site, and generates a truncated protein missing the conserved regions 1.2 –4.2, which are essential for σ70-type sigma factors (Figure 1a). A DNA gel-blot analysis using a T-DNA-specific probe (phosphoinothricin acetyltransferase (bar)-specific probe) indicated that multiple T-DNAs had been inserted at this site, but the exact arrangement and number of the T-DNAs was not determined (data not shown). An RNA gel-blot analysis using an AtSIG6-specific probe demonstrated that AtSIG6 transcripts were barely detectable in the sig6-1 mutant (Figure 1b). Thus, we concluded that the T-DNA insertion in the AtSIG6 gene completely inactivated the gene (null mutant) and named the mutant sig6-1. T4 plants generated by self-pollination of homozygous T3 plants were used for further experiments.

Figure 1.

Molecular analysis of the T-DNA insertion in the AtSIG6 locus.
(a) Exons (boxes) and the position of the T-DNA insertion in exon 4 of the AtSIG6 gene are shown.
(b) The analysis of AtSIG6 transcript levels. Total RNA was isolated from 4- and 8-day-old seedlings and the RNA gel blot (5 μg) was probed with the 600 bp cDNA fragment for AtSIG6. Endogenous AtSIG6 transcript was barely detected in the sig6-1 mutant.
(c) Phenotype of the sig6-1 mutant and wild-type plants during early plant development.
(d) Phenotype of 4-day-old seedlings of wild-type, sig6-1 and sig6-1/35S::SIG6.

Light-dependent chloroplast development was delayed in the sig6-1 cotyledons

The homozygous sig6-1 mutant plants had pale green cotyledons with drastically reduced chlorophyll content (<20% of wild-type plants) in 3- or 4-day-old seedlings (Figure 1c). The chlorophyll deficiency was limited to the young cotyledons and was restored to the wild-type level in 8-day-old seedlings. Interestingly, a few newly emerging true leaves following the cotyledons also initially showed a chlorophyll-deficient phenotype but turned green within 1 day (data not shown). On the other hand, later true leaves were mostly green and healthy from the beginning of leaf development.

We next performed segregation analyses of the pale green phenotype in the sig6-1 mutant plants. All progenies of self-pollinated homozygous mutant plants exhibited a pale green phenotype. Furthermore, progenies obtained from self-pollinated heterozygous (SIG6/sig6) seedlings were segregated in a 3.5:1 ratio (i.e. 117 green:33 pale green). These results are the same as those for a characteristic inherited as a single recessive Mendelian trait, supporting the hypothesis that the inactivation of AtSIG6 caused a pale green phenotype.

We compared electron micrographs of chloroplasts from wild-type and sig6-1 mutant cotyledons in 4-day-old light-grown seedlings. The mutant cotyledons contained smaller and spongiform plastids with markedly reduced internal thylakoid membranes, compared with wild-type chloroplasts (Figure 2a,b). Confocal microscopy observations of developing chloroplasts revealed that 3-day-old mutant seedlings contained smaller plastids exhibiting weaker chlorophyll fluorescence, compared with wild-type chloroplasts, whereas the chloroplasts in 7-day-old mutant plants were comparable to wild-type chloroplasts in size and chlorophyll fluorescence (Figure 2c).

Figure 2.

Transmission electron microscopic images of chloroplasts in 4-day-old wild type (a) and sig6-1 seedlings (b). Bars indicate 1 μm in each panel.
(c) Laser-scanning confocal microscopic analysis of chlorophyll fluorescence in mesophyll cells of cotyledons in light-grown seedlings.
(d) Phenotype of greening seedlings that were grown in the dark for 4 days and subsequently illuminated with white light for 24 h.

Dark-grown sig6-1 seedlings showed a normal etiolated phenotype characterized by long hypocotyls, closed apical hooks and small rudimentary cotyledons. The size and number of etioplasts were indistinguishable between the sig6-1 and wild-type seedlings (data not shown). After 6 h of exposure to light, the wild-type etiolated seedlings quickly reverted to a normal growth pattern with open apical hooks, and expanded and green cotyledons. However, the light-dependent greening of etiolated seedlings and chloroplast development were markedly retarded in the sig6-1 mutant (Figure 2d). Of note, the opening of the apical hooks and expansion of the cotyledons proceeded normally in the mutant. It is likely that AtSIG6 is indispensable for light-dependent transformation of etioplasts to chloroplasts, but not for other phototomorphogenic programs. Taken together, we concluded that AtSIG6 plays a key role in the light-dependent chloroplast development in cotyledons at an early stage of growth in Arabidopsis.

The transcript levels of most PEP-dependent genes were greatly reduced in the sig6-1 cotyledons

Retardation of chloroplast development might be caused by a defect in plastid gene expression in the AtSIG6-deficient mutant. To investigate this possibility, we first measured the global transcript pattern in the mutant using a plastid DNA microarray. The microarray contained duplicates of 257 spots of 500-bp PCR products covering the entire genome of the Arabidopsis chloroplast without overlaps (NAIST microarray consortium). In the control experiment with independently isolated RNAs from 9-day-old wild-type seedlings, almost all of the reliable hybridization signals were within +/− twofold changes (data not shown). In contrast, the hybridization signals varied extensively when 4-day-old wild-type seedlings were compared with 4-day-old sig6-1 seedlings exhibiting a pale green phenotype. The histogram shown in Figure 3 indicated that decreased signals were seen mostly for class I (green bars) and class II (yellow bars) PEP-dependent genes. The decreased class I genes included rbcL, psbA, psbB, psbC, psbD, psbH, psbN and psbT, while the decreased class II genes included rRNA genes such as rrn16, rrn23, rrn5 and rrn4.5 (see Table S1). In Arabidopsis, PEP has been reported to be responsible for the majority of transcription activity in rRNA transcription (Sriraman et al., 1998). It should also be noted that no reductions in class III NEP-dependent gene transcripts were seen in the sig6-1 mutant seedlings. In contrast, increased signals were mostly derived from NEP-dependent class II (yellow bars) and class III (orange bars) genes, including clpP, rps15, ndhB, ycf1 genes and rpoB, rpoC1, rpoC2 genes, respectively, but scarcely from class I genes in the sig6-1 mutant.

Figure 3.

Histogram of average ratios of chloroplast transcripts from the sig6-1 mutant and wild-type seedlings. Total RNA from 4-day-old mutant and wild-type seedlings was labeled reciprocally with Cy3 and Cy5. The ratio for each spot is the average of two replicates. The green bars represent class I genes, yellow bars represent class II genes, orange bars represent class III genes, and gray bars represent unidentified genes (see Table S1).

To confirm the array-based expression data, we further examined the effects of AtSIG6 disruption on plastid transcript pattern during early plant development using Northern analyses (Figure 4). Transcripts of all the PEP-dependent photosynthesis genes examined here had already accumulated in 4-day-old light-grown wild-type seedlings (Figure 4a). The transcript levels increased markedly in 8-day-old seedlings, indicating that the expression of PEP-dependent genes is regulated developmentally during seedling growth. In 4-day-old sig6-1 seedlings, the steady state transcript levels of PEP-dependent genes, including psbA, rbcL, rrn16, psaA, atpB and psbD decreased drastically, compared with wild-type levels. The reduced accumulation of these transcripts was virtually restored in 8-day-old sig6-1 seedlings. In contrast, the sig6-1 seedlings accumulated higher levels of transcripts for NEP-dependent genes, such as rpoA, rpoB, clpP, accD, ycf1 and ycf2, in 4(5)-day-old seedlings (Figures 4b and 5a). These results suggested that AtSIG6 may be responsible for the transcription of the majority of PEP-dependent genes at an early stage of light-dependent chloroplast development in cotyledons.

Figure 4.

Accumulation of plastid gene transcripts in the sig6-1 mutant and wild-type seedlings during early plant development. Total RNA was isolated from seedlings grown on MS-agar plate containing sucrose under a 16 h light and 8 h dark cycle. Plants were illuminated with white light for 4 h before sampling. Total RNAs (2–4 μg) were subjected to RNA gel-blot analysis.
(a) Expression of PEP-dependent genes. Asterisks show unprocessed precursor transcripts.
(b) Expression of NEP-dependent genes.
(c) Expression of tRNAs and rbcL.

Figure 5.

(a) Genetic complementation of the reduced accumulation of PEP-dependent transcripts and enhanced accumulation of NEP-dependent transcripts by the AtSIG6 cDNA. Total RNA was isolated from 4-day-old seedlings of the sig6-1 mutant, complemented sig6-1 mutant and wild-type. Plants were grown on MS-agar plate containing sucrose under a 16 h light and 8 h dark cycle, and illuminated with white light for 4 h before sampling. Total RNAs (2–10 μg) were subjected to RNA gel-blot analysis.
(b) Immunoblot analysis of α subunit of PEP. Total soluble proteins (20 and 30 μg) extracted from 4-day-old wild-type and sig6-1 seedlings were hybridized with an antibody against the rice α subunit. CBB-stained gel image is also shown below. Arrowhead indicates the predicted size of PEP α subunit.

The blue light-responsive promoter of psbD (psbD LRP) is a unique PEP promoter that is distinct from standard σ70-type PEP promoters. Primary transcripts (4.5 and 3.7 kb) from the psbD LRP were not detectable in 4-day-old wild-type seedlings (Figure 4a), suggesting that a special sigma factor responsible for psbD LRP activity, AtSIG5 was absent in young chloroplasts. On the other hand, psbD LRP transcripts were detected in 8-day-old seedlings of both wild-type and sig6-1 plants. AtSIG6 is unlikely to be responsible for psbD LRP activity.

Recently, AtSIG2 deficiency was shown to reduce the accumulation of several PEP-dependent tRNAs, such as trnE(UUC), trnD(GUC), trnV(UAC) and trnM(CAU), without affecting the transcript levels of most of the PEP-dependent photosynthesis genes (Hanaoka et al., 2003; Kanamaru et al., 2001; Privat et al., 2003). We examined the accumulation of several tRNAs in the sig6-1 mutant to compare the role of AtSIG6 with that of AtSIG2 in tRNA transcription (Figure 4c). Similarly to the AtSIG2 knockout and anti-sense plants, the expression of the trnE operon was diminished in young cotyledons of the sig6-1 mutant. The accumulation of trnQ(UUG) also decreased in sig6-1 young seedlings. On the other hand, trnV(UAC) level did not decrease in the sig6-1 mutants in contrast to the AtSIG2 knockout mutants, suggesting that AtSIG6 and AtSIG2 recognize different sets of tRNA promoters. Furthermore, this observation strongly supports our assumption that PEP is functional in the sig6-1 mutant.

Of final note, unprocessed precursor mRNAs for atpB and rrn16 accumulated to a significant degree in the sig6-1 mutant (Figure 4a). This finding suggests that AtSIG6 inactivation not only affected transcription of PEP-dependent genes, but may also have had an indirect effect on mRNA processing.

The mutant phenotype of the sig6-1 was complemented by the AtSIG6 cDNA

We performed genetic complementation studies to determine that AtSIG6 was responsible for the mutant phenotype of sig6-1. We transformed a homozygous sig6-1 mutant with full-length AtSIG6 cDNA fused to the cauliflower mosaic virus 35S promoter (sig6-1/35S::SIG6) and analyzed the T2 seedlings of seven independent lines. All of the complementation lines exhibited green cotyledons in 4-day-old seedlings (Figure 1d). These results indicate that AtSIG6 disruption is the direct cause of the pale-green cotyledons that were observed in the mutant phenotypes.

We also analyzed the expression of several plastid-encoded genes in a sig6-1 mutant complemented with AtSIG6 cDNA. As shown in Figure 1(c), AtSIG6 gene expression was not detected in the sig6-1 mutant. However, Northern blot analysis revealed a higher level of AtSIG6 transcripts in the complemented sig6-1 mutant than in wild-type plants (Figure 5a). Similarly, the reduced accumulation of all examined PEP-dependent transcripts (psbA, psbB, rbcL and rrn16) was almost restored in the complemented sig6-1 mutant. In contrast, the enhanced accumulation of NEP-dependent transcripts, such as rpoB, ycf1 and ycf2, was significantly reduced in the complemented sig6-1 mutant. These results strongly suggest that the AtSIG6 deficiency was the direct cause of the reduced transcription of most PEP-dependent genes in the sig6-1 mutant. The accumulation of trnV(UAC) was not affected in either the sig6-1 mutant or the complemented sig6-1 line that overexpressed AtSIG6, suggesting that AtSIG6 is not involved in trnV(UAC) transcription.

PEP is not deficient in 4-day-old sig6-1 mutant seedlings

The reduced accumulation of PEP-dependent transcripts and the enhanced accumulation of NEP-dependent transcripts are characteristics of mutant plants lacking one of the PEP genes. Thus, PEP might be deficient in the sig6-1 mutant as a result of the reduced translation activity in the chloroplasts caused by the insufficient accumulation of rRNA and tRNA molecules. To exclude this possibility, soluble protein extracts from 4-day-old wild-type and sig6-1 mutant seedlings were immunoblotted with anti-PEP α subunit antibodies raised against the rice α subunit. As shown in Figure 5(b), the same amount of α subunit protein was detected in both wild-type and mutant plants. Furthermore, the AtSIG6 deficiency did not reduce the accumulation of trnV(UAC), the expression of which depends on a PEP holoenzyme (Kanamaru et al., 2001; Legen et al., 2002). Taken together, these findings suggest that functional PEP is present in the sig6-1 mutant seedlings.

Expression of AtSIG6 transcripts is dependent on tissue, light and seedling development

To elucidate the role of AtSIG6 in chloroplast development during seedling growth, we examined the expression of all Arabidopsis sigma factor genes during early seedling growth. The AtSIG2, AtSIG3, AtSIG4 and AtSIG6 transcripts were detected in 4- or 5-day-old wild-type seedlings (Figure 6a). The accumulation of AtSIG6 transcripts increased in 8-day-old seedlings, but AtSIG2, AtSIG3 and AtSIG4 transcript levels were almost constant during seedling growth. In contrast, AtSIG1 and AtSIG5 transcripts were barely detected in 4-day-old seedlings and accumulated to a significant degree in 8-day-old seedlings. These results suggest that AtSIG2, AtSIG3, AtSIG4 and AtSIG6 likely function in cotyledons from an early stage of seedling growth, whereas AtSIG1 and AtSIG5 may function mainly at a later stage of seedling growth. Next, we examined the tissue specificity of AtSIG6 expression. AtSIG6 transcripts were abundantly accumulated in leaves, but not in roots (Figure 6b). Furthermore, the expression of AtSIG6 transcripts was obviously under light control, but was not induced to a significant degree by salt or cold stresses.

Figure 6.

Transcript accumulation for plastid sigma factor genes. Total RNA was isolated from seedlings grown on MS-agar plate containing sucrose under a 16 h light and 8 h dark cycle.
(a) Accumulation of sigma factor genes at an early stage of seedling growth. Plants were illuminated with white light for 4 h before sampling. Total RNAs (5 μg) were subjected to RNA gel-blot analysis.
(b) Leaf-specific and light-induced accumulation of AtSIG6 transcripts. Three-week-old plants were treated with 300 mm NaCl for 5 h (lane 3) or treated with cold temperature (4°C) for 48 h (lane 4) or dark-adapted for 24 h (lane 6) or illuminated with white light for 3 h after 24 h dark-adaptation (lane 5). Total RNA was isolated from leaves (lane 1, 3, 4, 5 and 6) and roots (lane 2). Total RNAs (5 μg) were subjected to RNA gel-blot analysis.

Discussion

AtSIG6 is responsible for the global transcription of PEP-dependent genes in cotyledons during early plant development

Arabidopsis contains six nuclear-encoded sigma-like factors that are expected to confer promoter specificity to the bacteria-type plastid RNA polymerase, PEP. In E. coli, most of the standard σ70-type promoters are recognized by a primary sigma factor, σ70, in the exponential stage, whereas alternative sigma factors direct RNA polymerase to specific promoters in response to environmental signals. If a chloroplast contains a single primary sigma factor that is essential for chloroplast development, as in the case of most eubacteria, deficient mutants are expected to produce an albino phenotype. In this study, we have identified an AtSIG6 null mutant, sig6-1 and found that the sig6-1 mutant exhibited pale green cotyledons containing small and undifferentiated plastids in 4-day-old seedlings. Light-dependent chloroplast development in etiolated seedlings was also significantly delayed in the sig6-1 mutant. Genetic complementation of the pale green phenotype of the sig6-1 mutant using AtSIG6 cDNA demonstrated that AtSIG6 plays a key role in early light-dependent chloroplast development in cotyledons.

To elucidate the role of AtSIG6 in plastid gene expression, we compared the global transcript patterns of plastid genes between wild-type and sig6-1 mutant seedlings using plastid DNA microarray and Northern blot analyses. The expression of many PEP-dependent class-I photosynthesis genes and class-II rDNA genes decreased in the sig6-1 mutant, whereas increased signals were derived mostly from class-III and some class-II genes that are largely dependent on NEP activity. This transcript pattern is reminiscent of that of PEP-deficient mutants (Allison et al., 1996; Hajdukiewicz et al., 1997). Furthermore, AtSIG6 cDNA was able to complement the reduced accumulation of PEP-dependent genes. These evidences strongly suggest that AtSIG6 is responsible for the recognition of standard PEP promoters and is indispensable for PEP activity in young seedlings. Our results best support the role of AtSIG6 as a general transcription initiator during early cotyledon development.

On the other hand, AtSIG6 might be a specialized sigma factor that recognizes a set of unique PEP promoters for essential genes involved in plastid translation, such as rRNAs, tRNAs and ribosomal proteins, and an AtSIG6 deficiency would result in a reduced translational activity in plastids. In this case, a deficiency of the PEP core would be expected in AtSIG6 knockout plants because of the reduced translation of plastid-encoded PEP core genes. To exclude this possibility, we immunologically measured the accumulation of PEP α subunit and found that the α subunit did not decrease in the AtSIG6 knockout plants, compared to that in wild-type plants. In addition, trnV(UAC) was actively transcribed in the sig6-1 mutant. trnV(UAC) has been reported to be exclusively transcribed by PEP (Hanaoka et al., 2003; Kanamaru et al., 2001). These data clearly suggest that active PEP is present in the sig6-1 mutant and that a PEP deficiency is not the primary reason for the reduced transcription of most PEP-dependent genes.

Another possibility is that a reduction in the accumulation of trnE, a precursor of ALA biosynthesis, may cause chlorophyll deficiency and indirectly suppress PEP activity or the mRNA stability of PEP-dependent genes in the AtSIG6-deficient mutant. However, a reduced accumulation of trnE and chlorophyll has also been reported in AtSIG2 knockout and anti-sense plants, with no affect on the overall abundance of PEP-dependent photosynthesis transcripts (Hanaoka et al., 2003; Kanamaru et al., 2001; Privat et al., 2003). Thus, trnE deficiency is unlikely to be responsible for the reduced accumulation of PEP-dependent transcripts. Taken together, we concluded that AtSIG6 likely acts as a general and major sigma factor during early plant development.

AtSIG6 generally recognizes σ70-type PEP promoters

In Figure 7, we aligned the pre-sequences of all PEP-dependent genes whose transcripts decreased in the sig6-1 mutant. All of the AtSIG6-dependent genes were preceded by typical σ70-type promoters harboring conserved −35/−10 core promoter elements. AtSIG6-dependent promoters include some unique σ70-type promoters, such as rbcL (Kim et al., 2002; Lam et al., 1988; Shiina et al., 1998), psbA (Eisermann et al., 1990; Satoh et al., 1999), psaA (Cheng et al., 1997) and rrnP1 (Suzuki et al., 2003), whose activities are regulated by extra promoter cis elements other than −35/−10 elements. This study clearly revealed that AtSIG6 directs PEP to initiate transcription at most of the σ70-type PEP promoters, including these unique promoters. On the other hand, the AtSIG6-independent accumulation of trnV(UAC) in the sig6-1 mutant suggests that other sigma factors are involved in the initiation of transcription at the trnV(UAC) promoter, although the trnV(UAC) promoter also exhibits typical σ70-type promoter elements. Transcription of trnV(UAC) may be preferentially mediated by AtSIG2 (Hanaoka et al., 2003; Kanamaru et al., 2001; Privat et al., 2003).

Figure 7.

Comparison of AtSIG6-dependent promoters in Arabidopsis. −10 and −35 promoter sequences are boxed. AtSIG6-independent trnV(UAC) promoter and psbD LRP are also shown.

In contrast, AtSIG6 did not recognize another unique PEP promoter, psbD LRP, whose activity is dependent on an upstream AAG box and transcription activator AGFs. AtSIG5 was recently shown to be responsible for the recognition of psbD LRP (Tsunoyama et al., 2004). In vitro reconstitution and transcription analyses would provide further insights on the promoter preferences of AtSIG6.

Chloroplasts likely contain another general sigma factor besides AtSIG6

The inactivation of AtSIG6 produced no obvious phenotype in older cotyledons and true leaves. These observations suggest that Arabidopsis contains another general sigma factor that functions mainly at a later stage of seedling growth but is absent in young cotyledons. Multiple general sigma factors may support active transcription from highly polyploid plastid DNAs during late plant development. One of the candidates for the late general sigma factors would be AtSIG1. AtSIG1 transcripts were not detected in 4-day-old seedlings and started to accumulate in 8-day-old seedlings, suggesting that AtSIG1 is mainly active at a later stage of seedling growth. Promoter-GUS assay experiments have shown that AtSIG1 promoter is activated 1 day after AtSIG2 promoter activation (Kanamaru et al., 1999). Previous in vitro experiments demonstrated that recombinant SIG1 proteins could function with the E. coli core enzyme and recognized σ70-type promoters including psbA and/or rbcL promoters, in mustard (Homann and Link, 2003; Kestermann et al., 1998), Arabidopsis (Hakimi et al., 2000; Privat et al., 2003) and maize (Beardslee et al., 2002). Furthermore, all higher plants examined thus far likely contain at least one SIG1 gene. These facts suggest that the SIG1 protein may play a crucial role as a general sigma factor in mature chloroplasts.

AtSIG2-knockout plants also exhibited a pale green phenotype (Shirano et al., 2000). As chlorophyll deficiency was observed in both young and mature mutant plants, AtSIG2 is likely indispensable for chloroplast transcription in all stages of plant development. Array-based global analysis of plastid transcript revealed that AtSIG2 and ATSIG6-deficient mutants showed different transcript profiles of entire chloroplast genome (Nagashima et al., 2004a). AtSIG2 inactivation resulted in the reduced accumulation of several plastid-encoded tRNAs, including trnV(UAC), and increased accumulation of NEP-dependent transcripts, but did not affect the transcript levels of most of the PEP-dependent photosynthesis genes (Hanaoka et al., 2003; Kanamaru et al., 2001; Nagashima et al., 2004a). Moreover, the accumulation of psbA mRNA was also reduced in developing cotyledons in the AtSIG2 anti-sense mutants (Privat et al., 2003). These data suggested that AtSIG2 preferentially initiates transcription at several tRNA promoters including trnE(UUC) promoter, and possibly psbA promoter, but less efficiently recognizes other standard PEP promoters. The AtSIG2-dependent promoters are similar to σ70-type promoters and are characterized by a conserved AT-rich pentanucleotide element in the spacer region (Hanaoka et al., 2003). However, we cannot exclude the possibility that AtSIG2 is able to initiate transcription at broad σ70-type promoters, as hybrid proteins of E. coliσ70 with C-terminal fragments of AtSIG2 containing regions 1.2–4.2 could complement E. coli rpoD mutants (Hakimi et al., 2000). This fact suggested that AtSIG2 is able to initiate transcription at broad σ70-type promoters in E. coli. However, AtSIG2 deficiency did not reduce the accumulation of most of the PEP-dependent photosynthesis genes. This discrepancy might be explained by complementary expression of other sigma factors in the AtSIG2-deficient mutant seedlings. Enhanced accumulation of plastid sigma factor transcripts including AtSIG6 was observed in the AtSIG2 knockout mutant (Nagashima et al., 2004a). Furthermore, the expression of AtSIG3 protein was enhanced in the AtSIG2 anti-sense mutant (Privat et al., 2003). These data suggested that AtSIG2 deficiency might be complemented by an increased accumulation of AtSIG6 and/or other plastid general sigma factors in the AtSIG2-deficient mutants. AtSIG2 knockout plants exhibited a pale green phenotype in all stages of plant development, whereas chlorophyll deficiency was recovered in mature leaves in AtSIG2 anti-sense plants (Privat et al., 2003). It has been proposed that AtSIG2-dependent expression of trnE(UUC) may play important roles in biosynthesis of chlorophyll and photosynthesis proteins (Hanaoka et al., 2003; Kanamaru et al., 2001). In AtSIG2 anti-sense plants, AtSIG6 and/or other general sigma factors may backup the expression of trnE(UUC) in mature leaves. To characterize the redundant functions of AtSIG6 and AtSIG2 in detail, a crossing between sig6-1 and sig2 mutant lines is now under investigation.

On the other hand, we recently demonstrated that AtSIG5 is a unique plastid sigma factor that specifically recognizes a non-σ70-type psbD LRP (Tsunoyama et al., 2004). In conclusion, Arabidopsis chloroplasts likely contain several general sigma factors, including AtSIG6, that are responsible for the recognition of σ70-type PEP promoters and at least one specialized sigma factor, AtSIG5, that mediates blue light (Tsunoyama et al., 2002, 2004) and other stress signaling (Nagashima et al., 2004b) and activates transcription at psbD LRP. To specify the promoter preference of each sigma factor in vivo, complementation experiments using the sig6-1 mutant with each plastid σ factor cDNA are now underway.

Multiple sigma factors may be involved in the developmental regulation of PEP

This study suggested that AtSIG6 might be a major general sigma factor during early cotyledon development, whereas other late general sigma factors likely play a role in older cotyledons and mature leaves. In mustard, rifampicin-sensitive B-type PEP is dominant in etioplasts and in immature chloroplasts during greening, whereas rifampicin-resistant A-type enzyme is associated with several accessory proteins and is present in mature chloroplasts (Pfannschmidt and Link, 1994, 1997). Similarly, in wheat seedlings, the light-independent base-type PEP is converted to the tip-type PEP, whose activity is dependent on light during chloroplast development (Satoh et al., 1999). The tip-type PEP in mature chloroplasts recognizes an extended −10 promoter of the psbA gene, but not the base-type PEP. Taking into account the possible roles of AtSIG6 during early plant development, the AtSIG6 homolog may be a major sigma factor in mustard B-type PEP and/or wheat base-type PEP. The maize AtSIG6 homolog protein ZmSIG3 was expressed preferentially in the leaf base and roots, suggesting a possible role in immature chloroplasts (Lahiri and Allison, 2000). The conserved region 2.5 was involved in the recognition of the extended −10 promoter in σ70 (Kumar et al., 1993). Interestingly, AtSIG1 and AtSIG2 contain two highly conserved amino acid residues in region 2.5, Glu-458 and His-455, but AtSIG6 does not. Furthermore, mustard sigma factors SaSIG1 (AtSIG1 homolog) and SaSIG2 (AtSIG2 homolog) were recently shown to have an affinity for the psbA extended −10 promoter in vitro (Homann and Link, 2003). Whether AtSIG6 is capable of initiating transcription from extended −10 promoters remains to be analyzed.

An unidentified regulatory network may coordinate chloroplast transcription and mRNA processing

In contrast to most PEP-dependent transcripts, NEP-dependent transcripts increased in the sig6-1 mutant. Furthermore, AtSIG6 inactivation resulted in the accumulation of precursor RNAs for atpB and rrn16, suggesting an unusual RNA processing activity in this mutant. Similar unusual NEP-dependent mRNA accumulations and RNA processing have been reported in various mutant plants lacking one of the PEP genes (Allison et al., 1996; De Santis-Maclossek et al., 1999; Hajdukiewicz et al., 1997; Krause et al., 2000; Legen et al., 2002) and in a virescent mutant in rice (Kusumi et al., 1997). Chloroplasts probably have a complex regulatory mechanism that coordinates PEP and NEP transcription activities and RNA processing. Comparison of the chloroplast transcription patterns between the AtSIG2 and AtSIG6-deficient mutants suggested that PEP inactivation is not primarily responsible for the enhanced NEP activity. Further analyses of plastid transcripts and NEP expression in the sig6-1 seedlings may shed light on the signaling network regulating NEP and PEP activities.

Experimental procedures

Plant materials and growth conditions

Arabidopsis thaliana ecotype Columbia was used as a control in all experiments described here. The sig6-1 mutant Arabidopsis ecotype Columbia was identified in a collection of T-DNA insertion lines generated by Syngenta. Surface sterilized seeds were sown on 0.5x Murashige and Skoog (MS) medium containing 1% sucrose and placed at 4°C for 2 or 5 days followed by germination under long day conditions (16 h light, 8 h dark) at 22°C. Light intensity was 40–70 μmol m−2 sec−1 unless otherwise specified. Seedlings were then transferred to Jiffy-7 (Sakata Seed Co., Yokohama, Japan) and grown under the same conditions. For greening experiments, dark-grown 4-day-old seedlings were transferred to 20 μmol m−2 sec−1 white light. For RNA analysis, seedlings were grown synchronously on 0.5x MS plates under the standard condition. To test the effects of various stresses on AtSIG6 mRNA expression, 3-week-old plants were treated with the following stress conditions. NaCl: plants were transferred to 0.5x MS plates containing 300 mm NaCl and kept on the plates for 5 h; cold temperature: plants were kept at 4°C for 2 days. Plants were also dark adapted for 24 h (dark) and re-illuminated with 50 μmol m−2 sec−1 white light for 3 h (light). Photon fluence rates were measured using a quantum photometer (LI-250; Li-Cor Inc., Lincoln, NE, USA).

Identification of the T-DNA insertion allele of the AtSIG6 gene

To identify plants containing a T-DNA insertion, PCR was carried out using AtSIG6-specific primers SIG6-P1 (5′-AGCGCGATCTTCATTATCAGCCCC-3′) and LB3 (5′-TAGCATCTGAATTTCATAACCAATCTCGATACAC-3′). The insertion site was confirmed by sequencing the amplified fragment. Identified lines were homozygous for the T-DNA insertion. The mutant allele was named sig6-1. All experiments described here were performed with T4 generation plants. The copy number of the insertion allele was determined by restriction digestion with XhoI, HindIII or EcoRV followed by DNA gel-blot analysis.

Complementation of the sig6-1 mutant

The AtSIG6 full-length cDNA was amplified from Arabidopsis cDNA using the primer pair SF-SalI (5′-AATGTCGACATGGAAGCTACGAGGAACTTGGTTTCT-3′) and SF-NotI (5′-AAAGCGGCCGCCTAGACAAGCAAATCAGCATAAGCA-3′). The resulting fragment was ligated into the transient expression vector pS65T (Isono et al., 1997) to obtain pS65TSIG6, which is under the transcriptional control of the cauliflower mosaic virus 35S promoter and has omega sequences for efficient translation. The SIG6 cassette P35SΩ::SIG6::Tnos was inserted into the EcoRI–HindIII site of the plant expression vector pBI121, then introduced into Agrobacterium tumefaciens strain C-58. The floral-buds of sig6-1 mutant plants were transformed according to Clough and Bent (1998). The transformants were selected on 0.5x MS 1% sucrose plates containing 50 μg ml−1 kanamycin. Successful complementation was confirmed by chlorophyll accumulation in cotyledons.

Northern blot analysis

Total RNAs were extracted from seedlings of Arabidopsis by RNeasy Plant Mini Kit (Qiagen Inc., Valencia, CA, USA). Total RNA samples (10 or 4 μg) were separated by denaturing agarose gel electrophoresis. After capillary blotting onto Hybond-N nylon membrane, RNA gel blots were hybridized to the randomly primed DNA probes encoding psbA, psbD, psaA, rbcL, rrn16, atpB, rpoA, rpoB, clpP, accD, trnQ(UUG), trnV(UAC), and trnEYD operons. cDNA fragments [1030–1509 of AtSIG1 (AB019942), 1048–1719 of AtSIG2 (AB019943), 1049–1716 of AtSIG3 (AB019944), 739–1260 of AtSIG4 (AB021119), 1096–1554 of AtSIG5 (AB021120), and 1158–1644 of AtSIG6 (AB029916)] were also used as probes for Northern analyses. Final wash conditions were 0.2x SSC, 0.1% SDS at 65°C for 30 min. Radioactive signals were detected by autoradiography or using BAS 1800 (Fuji Photo Film Co., Ltd., Tokyo, Japan).

Microarray analysis

DNA microarrays were constructed by NAIST array consortium. A total of 257 spots of 500 bp PCR fragments were subcloned into plasmid vectors and DNA microarrays were generated using these plasmid DNAs. Standard PCR was carried out for all clones using universal vector primers (M13 forward/reverse). Purified PCR products were spotted in duplicate onto glass slides. Twenty micrograms of the total cellular RNA were labeled with Cy3 and Cy5 using Cyscribe First-Strand cDNA Labeling Kit (Amersham Biosciences, Piscataway, NJ, USA). Hybridization was carried out in Microassay hybridization buffer (Amersham Biosciences) at 65°C for 12 h. After washing microarrays were scanned and analyzed using accompanied software.

Immunoblotting

Total soluble proteins were prepared by homogenization of tissue in a standard extraction buffer. Following a brief centrifugation protein concentration was determined using the Bradford assay. Proteins were separated by SDS-PAGE and transferred to PVDF membrane (ATTO, Tokyo, Japan). Membranes were blocked with non-fat dried milk and incubated with polyclonal anti-rice SIG1 antibody (diluted 1:10 000). Signal detection was carried out by ECL-Plus using manufacturer-supplied protocols (Amersham-Pharmacia, Piscataway, NJ, USA).

Quantification of chlorophyll content

Whole seedlings (10–20) grown on agar medium were collected in Eppen tubes. Extraction with 1 ml 80% (v/v) acetone–water was carried out after measuring fresh weight. Quantification of chlorophyll content was performed with triplicates.

Transmission electron microscopy

Cotyledons were fixed with 5% glutaraldehyde and post-fixed with 1% osmium tetroxide. The samples were stained with uranyl acetate, dehydrated, and then embedded in Spurr's resin. Ultrathin sections were observed by transmission electron microscopy.

Confocal microscopy

Chlorophyll imaging was carried out using a confocal laser-scanning microscope (μRadiance; Bio-Rad Laboratories Inc., Hercules, CA, USA). Chlorophyll was excited with the 568 nm line. Fluorescence images were collected through red channel.

Acknowledgements

We thank the Torrey Mesa Research Institute (San Diego, CA, USA) and Syngenta for providing the sequence-indexed Arabidopsis T-DNA insertion mutants. We also thank Dr Y. Tozawa for the generous gift of an antibody. We thank Dr Y. Isozumi for providing us the facilities of the Radioisotope Research Center. This work was performed as one of the technology development projects of the ‘Green Biotechnology Program’ supported by NEDO (New Energy and Industrial Technology Development Organization). This work was also supported by Grants-in-Aids for Scientific Research to T.S. (14540598) and for young scientists to Y.N. (200203019) and the Foundation for Bio-venture Research Center from the Japanese Ministry of Education, Culture, Sports, Science and Technology. Y.N. is a Research Fellow of the Japan Society for the Promotion of Science.

Supplementary Material

The following material is available from http://www.blackwellpublishing.com/products/journals/suppmat/TPJ/TPJ2362/TPJ2362sm.htm

Table S1. Data set of microarray analysis.

Ancillary