The plastid genome in photosynthetic higher plants encodes subunits of an Escherichia coli‐like RNA polymerase (PEP) which initiates transcription from E.coli σ70‐type promoters. We have previously established the existence of a second nuclear‐encoded plastid RNA polymerase (NEP) in photosynthetic higher plants. We report here that many plastid genes and operons have at least one promoter each for PEP and NEP (Class II transcription unit). However, a subset of plastid genes, including photosystem I and II genes, are transcribed from PEP promoters only (Class I genes), while in some instances (e.g. accD) genes are transcribed exclusively by NEP (Class III genes). Sequence alignment identified a 10 nucleotide NEP promoter consensus around the transcription initiation site. Distinct NEP and PEP promoters reported here provide a general mechanism for group‐specific gene expression through recognition by the two RNA polymerases.
The plastid genome of photosynthetic higher plants encodes proteins which are homologous to the Escherichia coli DNA‐dependent RNA polymerase α, β and β′ subunits. The subunit structure of the plastid enzyme is similar to that of the E.coli enzyme, except that the plastid β′ and β″ subunits are equivalent to the N‐ and C‐termini of the bacterial β′ subunit, respectively. The plastid genes were named rpoA, rpoB, rpoC1 and rpoC2 to indicate homologies with the E.coli genes. Promoter selection by this plastid‐encoded plastid RNA polymerase (PEP) is dependent on nuclear‐encoded σ‐like factors. The promoters utilized by PEP are similar to E.coli σ70 promoters, consisting of −35 and −10 consensus elements (reviewed in Igloi and Kössel, 1992; Gruissem and Tonkyn, 1993; Link, 1994, 1996). Transcription activity from some PEP promoters is modulated by nuclear‐encoded transcription factors interacting with elements upstream of the core promoter (Sun et al., 1989; Iratni et al., 1994; Allison and Maliga, 1995; Kim and Mullet, 1995).
Several reports have suggested the existence of an additional plastid‐localized, nuclear‐encoded RNA polymerase (reviewed in Igloi and Kössel, 1992; Gruissem and Tonkyn, 1993; Mullet, 1993; Link, 1994, 1996). By deleting the rpoB gene encoding the essential β subunit of the tobacco PEP, we established the existence of a second nuclear‐encoded plastid RNA polymerase (NEP) in photosynthetic higher plants (Allison et al., 1996). Deletion of rpoB yielded photosynthetically defective, pigment‐deficient plants. An examination of ΔrpoB plants revealed proplastid‐like structures containing low levels of mRNAs for the photosynthetic genes rbcL, psbA and psbD due to the lack of PEP promoter activity. In wild‐type tobacco leaves, the ribosomal RNA operon (rrn) is transcribed by PEP. Interestingly, in the ΔrpoB plants the rrn mRNA accumulated close to wild‐type levels due to transcription by NEP acting at a downstream non‐σ70‐type promoter. The rRNA operon is the first plastid transcription unit for which a promoter for both PEP and NEP was identified.
We report here that the rRNA operon is not unique, but represents a class of plastid transcription units which have at least one promoter each for PEP and NEP. These genes or operons have a potential for expression by either of the two plastid RNA polymerases. Furthermore, some genes are transcribed by only one of the two RNA polymerases. We propose that transcription by NEP and PEP, through recognition of distinct promoters, is a general mechanism of group‐specific gene regulation during chloroplast development. A tentative NEP promoter consensus is derived by the alignment of the transcription initiation sites.
Identification of genes with promoters for the NEP polymerase
To facilitate mapping of additional NEP promoters, we examined mRNA steady‐state concentrations in ΔrpoB plants for several plastid genes (Figure 1). The plastid genes were divided into three groups based on mRNA steady‐state concentrations in fully‐extended leaves of wild‐type and ΔrpoB plants. Group I includes genes for which the mRNAs accumulate to high levels in wild‐type leaves, and to very low levels in the leaves of ΔrpoB plants (Figure 1A). Genes in this class are psaA (photosystem I gene); psbB and psbE (photosystem II genes); petB, a cytochrome b6/f complex gene (for references see Shinozaki et al., 1986b); ndhA, a respiratory chain NADH dehydrogenase homologue (Matsubayashi et al., 1987); and the rps14 ribosomal protein gene (Meng et al., 1988). Group II includes plastid genes encoding mRNAs that accumulate to significant levels in both wild‐type and ΔrpoB leaves (Figure 1B). Group II includes atpB (ATP synthase gene; Orozco et al., 1990); clpP encoding the proteolytic subunit of the Clp ATP‐dependent protease (Gray et al., 1990; Maurizi et al., 1990); ndhB and ndhF, two respiratory chain NADH dehydrogenase homologues (Matsubayashi et al., 1987); the rps16 ribosomal protein gene (Shinozaki et al., 1986a); and ycf1, a gene with unknown function (ORF1901; Wolfe et al., 1992; Hallick and Bairoch, 1994; data not shown). Group III includes genes for which there is significantly more mRNA in the ΔrpoB leaves than in the leaves of wild‐type plants (Figure 1C). Among these are: accD, encoding a subunit of the acetyl‐CoA carboxylase (Sasaki et al., 1993); ribosomal protein genes rpl33 and rps18 (Shinozaki et al., 1986b); and ycf2, a putative ATPase with unknown function (ORF2280; Hallick and Bairoch, 1994; Wolfe, 1994).
Apparent are the more complex RNA patterns in ΔrpoB plants as compared with wild‐type plants (Figure 1B and C). The reason for the more complex patterns may be activation of additional promoters upstream of the tested genes, differences in mRNA processing and stability, and differences in the transcription termination signals for the two polymerases. The origin of the complex RNA patterns in the ΔrpoB plants is outside the scope of the present study.
The atpB and atpI ATP synthase genes have both NEP and PEP promoters
The RNA gel blot analysis identified a number of genes and operons for which transcript levels are maintained or elevated in ΔrpoB leaves (Figure 1B). To identify NEP promoters, 5′ ends of transcripts were mapped by primer extension analysis. To distinguish between 5′ ends that represent transcripts from a NEP promoter from those generated by RNA processing, the 5′ ends were capped using guanylyltransferase.
Transcript 5′ ends for the tobacco atpB gene have been identified by Orozco et al. (1990) at nucleotide positions −255, −289, −488, −502 and −611 relative to the translation initiation codon (ATG; the nucleotide directly upstream of the A‐occupying position −1). Primer extension analysis identified each of these 5′ ends in wild‐type plants (Figure 2A). These RNA species are not resolved distinctly in Figure 1B (see also Kapoor et al., 1994). In the ΔrpoB plants only the −289 RNA species was present. The 5′ end of this transcript was capped using guanylyltransferase (Figure 2B). Therefore, we propose that the −289 RNA is transcribed from a NEP promoter, termed PatpB–289. Interestingly, the −289 transcript is also present in the wild‐type leaves, although it is significantly less abundant than in the ΔrpoB plants. The −255, −488, −502 and −611 transcripts are absent in the ΔrpoB plants (Figure 2A), suggesting that they are transcribed by PEP in plastids.
The atpI operon includes the atpI–atpH–atpF–atpA genes (Figure 3C). In wild‐type tobacco leaves, we mapped mRNA 5′ ends to the −207 region (−212, −209 and −207) and at nucleotides −130 and −85. Interestingly, in ΔrpoB leaves only the −207 transcript is detectable (Figure 3A; data not shown). This transcript could be capped in the ΔrpoB RNA sample (Figure 3B), demonstrating that it originates directly from a NEP promoter. A signal at the same position was obtained in the in vitro capping reaction of wild‐type RNA samples, corresponding to the −207, −209 and −212 transcripts which were not resolved in the assay. We could also cap the −130 transcript which is present only in wild‐type leaf (Figure 3A and B). Thus, it is likely that PatpI–130 is recognized by PEP.
A clpP NEP promoter is highly active in chloroplasts
Primer extension analysis with wild‐type plants identified clpP RNA 5′ ends at nucleotide positions −53, −95 and −173. In contrast, in ΔrpoB leaves, the 5′ ends mapped to nucleotides −53, −173 and −511 (Figure 4A). Since in vitro capping reaction verified that each of these are primary transcripts (Figure 4B), it would seem likely that these transcripts derive from NEP promoters. The PclpP–53 promoter is highly expressed in both wild‐type‐and ΔrpoB plants; thus, it is a NEP promoter with a potential for high‐level expression in different tissue types. The PclpP–53 promoter is well conserved in spinach, in which it is the only promoter transcribing the clpP gene (Westhoff, 1985). Additional NEP promoters are PclpP–173 and PclpP–511. The PclpP–511 transcript accumulates only in ΔrpoB plants (Figure 4A). Note also, that PclpP–511 is located within the psbB coding region (Figure 4C); therefore, its activity in wild‐type plastids may be affected by the convergent psbB PEP promoter (not marked).
Transcripts from PclpP–95 accumulate only in wild‐type leaves. Therefore, it is likely that this promoter is recognized by PEP.
The accD gene is transcribed exclusively from a NEP promoter
RNA for the lipid biosynthetic gene accD accumulates to high levels only in ΔrpoB plants. A major transcript initiating at nucleotide position −129 (Figure 5A) could be capped in vitro (Figure 5B). Therefore, this RNA is transcribed from a NEP promoter. The mRNA from PaccD–129 accumulates to very low levels in the photosynthetically active leaf mesophyll cells, whereas it is abundant in the non‐photosynthetic plastids of ΔrpoB plants.
NEP promoters share a loose consensus adjacent to the transcription initiation site
Sequences flanking 10 transcription initiation sites were aligned to identify conserved NEP promoter elements (Figure 6). Included in the sequence alignment are nine promoters identified in this study and Prrn–62, the NEP promoter described in Allison et al. (1996) and Vera and Sugiura (1995). We have included sequences for Pycf2–1577 and Pycf1–41 for which the 5′ ends were shown to be primary transcripts by capping in vitro (data not shown). Both of these promoters are active in ΔrpoB leaves, but not in the leaves of wild‐type plants. Included in the sequence alignment are tentative NEP promoters for rps2 and rps16, for which there is mRNA in ΔrpoB leaves. Transcripts for these tentative NEP promoters were mapped by primer extension analysis. However, the in vitro capping assays were inconclusive due to low abundance of the mRNAs (data not shown).
Multiple sequence alignment of the regions immediately flanking the NEP 5′ ends identified a loose 10 nucleotide consensus around the transcription initiation site (Figure 6). Clustering of conserved sequences around the transcription initiation site is reminiscent of the promoters of the single‐subunit mitochondrial and phage T3/T7 RNA polymerases (Raskin et al., 1993; Tracy and Stern, 1995). Conservation of additional nucleotides upstream and downstream of the NEP transcription initiation site is also apparent. Clustering of conserved sequences around the NEP transcription initiation site contrasts PEP promoters for which −35/−10 elements are localized upstream of the transcribed region (Igloi and Kössel, 1992; Gruissem and Tonkyn, 1993; Link, 1994, 1996). All these NEP promoters are inactive in chloroplasts, and are utilized only in the proplastid‐like organelles of the ΔrpoB plants. Interestingly, the region around the transcription initiation site is less well conserved for PclpP–53, the only NEP promoter highly active in chloroplasts (bottom of Figure 6). Identification of functionally relevant sequences for PclpP–53 and the other NEP promoters will require dissection of the regions containing the transcription initiation sites.
Based on their promoters, plastid genes and operons can be assigned to three classes: those with PEP (Class I), NEP and PEP (Class II), and NEP promoters only (Class III). Class I mainly includes photosynthetic genes which are transcribed from σ70‐type PEP promoters, including genes shown in Figure 1A and rbcL, psbA and psbD (Allison et al., 1996). RNA steady‐state concentrations for these genes are high in wild‐type leaves, and very low—if at all detectable—in the leaves of ΔrpoB plants (Group I). Class II transcription units rrn, atpB, atpI and clpP have both NEP and PEP promoters (Allison et al., 1996; also see Results). The mRNAs for these operons accumulate to significant levels in both wild‐type and ΔrpoB plants (Figure 1B; see also Figure 4B in Allison et al., 1996). Although no attempt was made to fully characterize rps16, rpl33–rps18, ndhB, ndhF and ycf1 transcription, it is likely that high transcript levels for these genes in both leaf types is also due to transcription by both the plastid‐encoded and nucleus‐encoded RNA polymerases. Class II genes encode diverse functions, although none is a photosystem I or II polypeptide. Class III is small, and includes accD and ycf2 transcribed from the PaccD–129 and Pycf2–1577 NEP promoters. rpoB is also likely to be a Class III gene (Hess et al., 1993). With respect to mRNA steady‐state concentrations, these genes belong to Group III: there is significantly more mRNA in the leaves of ΔrpoB plants than in those of wild‐type plants (Figure 1C).
It appears that genes with similar functions are transcribed by PEP, or by both PEP and NEP as a group. A good example is transcription of all tested photosystem I (psa) and photosystem II (psb) genes by PEP (see above). Also, the mRNAs for most ribosomal protein genes accumulate to relatively high levels in both leaf types, including those of rps16 and rpl33–rps18 (Figure 1). An exception is rps14, for which there is very little mRNA in the ΔrpoB leaves (Figure 1A). The rps14 gene is transcribed as part of the psaA operon which contains two photosystem I genes (psaA, psaB; Meng et al., 1988) and apparently has no dedicated NEP promoter. Puzzling is the differential accumulation of the mRNAs for the ndh genes, since they are assumed to be components of the same complex (Matsubayashi et al., 1987): ndhB and ndhF accumulate to high levels in both wild‐type and ΔrpoB plants, whereas ndhA mRNA is present only in the wild‐type leaves. Observed differential accumulation may indicate complementation of some, but not all plastid ndh genes by mitochondrial copies of the same gene (Matsubayashi et al., 1987).
Given the general need for lipid biosynthesis, accD mRNA should be present in all cell types. However, in wild‐type leaves, the level of accD mRNA is low (Figure 1C). The plastid gene accD encodes a subunit of the prokaryotic‐type acetyl‐CoA carboxylase. It is feasible that, in chloroplasts, lipid biosynthesis may depend on the nucleus‐encoded eukaryotic enzyme (Sasaki et al., 1995). It remains to be seen in which cell types the plastid accD gene is highly expressed from the newly identified NEP promoter (Figure 5).
Interestingly, clpP is transcribed from the PclpP–53 NEP promoter in wild‐type leaves while most other NEP promoters are inactive in the same tissue. Differential accumulation of mRNAs from NEP promoters is possibly due to regulation by nuclear‐encoded factors via gene‐specific promoter elements, as reported for σ70‐type PEP promoters (Sun et al., 1989; Iratni et al., 1994; Allison and Maliga, 1995; Kim and Mullet, 1995).
The general rule emerging from the data is that photosystem I and II genes are transcribed by the PEP polymerase, whereas most other genes have both PEP and NEP promoters. This is in agreement with the observed major role for PEP in the transcription of all plastid genes in chloroplasts, including photosynthetic and housekeeping genes (reviewed in Igloi and Kössel, 1992; Gruissem and Tonkyn, 1993; Link, 1994, 1996). As to the role of NEP, we assume that it plays an important role in non‐green plastids, in tissues in which PEP is absent or is present only in limited amounts. A similar role for the two plastid RNA polymerases was proposed in the expression of plastid housekeeping genes from alternative promoters in photosynthetic and non‐photosynthetic tissues based on the study of tobacco tissue culture cells (Vera and Sugiura, 1995; Vera et al., 1996). Tissue‐ and cell type‐specific transcription from NEP promoters will have to be determined individually, by following accumulation of RNA and proteins for reporter genes in transgenic plastids.
During chloroplast development, early transcription of plastid genes encoding the plastid's transcription and translation apparatus relative to genes encoding proteins involved in photosynthesis was documented in barley (Baumgartner et al., 1993) and pea (DuBell and Mullet, 1995). Selective transcription of the gene groups by the nucleus‐encoded and plastid‐encoded RNA polymerases was proposed as one possible mechanism of differential gene expression (Hess et al., 1993; Lerbs‐Mache, 1993; Mullet, 1993). Distinct NEP and PEP promoters reported here for a large number of transcription units provide a general mechanism for developmentally‐timed expression of groups of plastid genes by the two plastid RNA polymerases (Figure 7).
Materials and methods
RNA gel blots
Total leaf RNA was prepared from fully‐expanded leaves using TRIzol (Gibco‐BRL), following the manufacturer's protocol. The RNA was electrophoresed on 1% agarose–formaldehyde gels, then transferred to Hybond N (Amersham) using the Posiblot Transfer apparatus (Stratagene). Hybridization to random‐primer labeled fragment was carried out in Rapid Hybridization Buffer (Amersham) overnight at 65°C.
Double‐stranded DNA probes were prepared by random‐primed 32P‐labeling of PCR‐generated DNA fragments. The sequence of the primers used for PCR, along with their positions within the tobacco ptDNA (Shinozaki et al., 1986b) are as follows:
The rps16 mRNA was probed with an EcoRI fragment isolated from plasmid pJS40, containing sequences between nucleotides 4938–5363 and 6149–6656 of the tobacco ptDNA (Shinozaki et al., 1986b). The rps14 mRNA was probed with the end‐labeled oligonucleotide 5′‐CACGAAGTATGTGTCCGGATAGTCC‐3′ (5′ end at nt 38621 in the plastid genome; Shinozaki et al., 1986b).
The probe for tobacco 25S rRNA was from plasmid pKDR1 (Dempsey et al., 1993) containing a 3.75 kb EcoRI fragment from a tobacco 25S/18S locus cloned in plasmid pBR325. When hybridizing gel‐blots for 25S rRNA, 32P‐labeled double‐stranded DNA probe was mixed with unlabeled plasmid pKDR1 corresponding to a 2‐fold excess over the amount of RNA present on the filter.
Primer extension analysis
Primer extension reactions were carried out on 10 mg (wild‐type) or 10 mg (ΔrpoB) of total leaf RNA as described (Allison and Maliga, 1995). The primers are listed below. Underlined oligonucleotides were also used to generate the capping constructs.
Sequence ladders were generated with the same primers using the Sequenase II kit (USB).
Identification of primary transcripts by in vitro capping
Total leaf RNA (20 μg) from wild‐type and ΔrpoB plants was capped in the presence of [α‐32P]GTP (Kennell and Pring, 1989). Labeled RNAs were detected by ribonuclease protection (Vera and Sugiura, 1992) using the RPAII kit (Ambion). To prepare the protecting complementary RNA, an appropriate segment of the plastid genome was PCR‐amplified using the primers listed below. The 5′ primers were designed to add a SacI restriction site (underlined) upstream of the amplified fragment. The 3′ primers were designed to add a KpnI site (underlined) downstream of the amplified sequence. The amplified product was cloned as a SacI–KpnI fragment into SacI‐ and KpnI‐restricted pBSKS+ vector (Stratagene). To generate unlabeled RNA complementary to the 5′ end of RNAs, the resulting plasmid was linearized with Acc65I (KpnI isoschizomer), and transcribed in a Megascript (Ambion) reaction with T7 RNA polymerase. The only exception was the atpI gene, for which an EcoRI site was used for cloning at the 3′ end, and for the linearization of the plasmid. Markers (100, 200, 300, 400 and 500 nt) were prepared with the RNA Century Markers Template Set (Ambion), following the manufacturer's protocol.
DNA sequence analysis
DNA sequence analysis was carried out utilizing the Wisconsin Sequence Analysis Package (Genetics Computer Group, Inc.).
These studies were supported by the National Science Foundation Grants MCB 93‐05037 and MCB 96‐30763 to P.M. L.A.A. was a recipient of a Charles and Johanna Busch Memorial Fund Postdoctoral Fellowship.
- Copyright © 1997 European Molecular Biology Organization