Transcription of plastid genes is modulated by two nuclear-encoded α subunits of plastid RNA polymerase in the moss Physcomitrella patens


  • The nucleotide sequences reported in this paper have been submitted to the DDBJ/EMBL/GenBank database under accession numbers AB110071 (PpRpoA2 gene) and AB110072 (PpRpoA2 cDNA).

*(fax +81 52 789 3080; e-mail


In general, in higher plants, the core subunits of a bacterial-type plastid-encoded RNA polymerase (PEP) are encoded by the plastid rpoA, rpoB, rpoC1 and rpoC2 genes. However, an rpoA gene is absent from the moss Physcomitrella patens plastid genome, although the PpRpoA gene (renamed PpRpoA1) nuclear counterpart is present in the nuclear genome. In this study, we identified and characterized a second gene encoding the plastid-targeting α subunit (PpRpoA2). PpRpoA2 comprised 525 amino acids and showed 59% amino acid identity with PpRpoA1. Two PpRpoA proteins were present in the PEP active fractions separated from the moss chloroplast lysate, confirming that both proteins are α subunits of PEP. Northern blot analysis showed that PpRpoA2 was highly expressed in the light, but not in the dark, whereas PpRpoA1 was constitutively expressed. Disruption of the PpRpoA1 gene resulted in an increase in the PpRpoA2 transcript level, but most plastid gene transcript levels were not significantly altered. This indicates that transcription of most plastid genes depends on PpRpoA2-PEP rather than on PpRpoA1-PEP. In contrast, the transcript levels of petN, psbZ and ycf3 were altered in the PpRpoA1 gene disruptant, suggesting that these are PpRpoA1-PEP-dependent genes. These observations suggest that plastid genes are differentially transcribed by distinct PEP enzymes with either PpRpoA1 or PpRpoA2.


Plastids are semi-autonomous organelles that possess their own genetic information. The components of photosynthesis complexes and the translational and transcriptional apparatus are encoded separately by plastid and nuclear genomes. The nuclear-encoded components synthesized in the cytoplasm are imported post-translationally into the plastids and assembled with the plastid-encoded components (Martin and Herrmann, 1998).

Plastids of higher (seed) plants contain two distinct DNA-dependent RNA polymerases: the plastid-encoded plastid RNA polymerase (PEP), and the nuclear-encoded plastid RNA polymerase (NEP) (Maliga, 1998; Hess and Börner, 1999). The PEP enzyme comprises a core complex ααββ′β′′ encoded as plastid genes rpoA, rpoB, rpoC1 and rpoC2. Transcription initiation by PEP is required for multiple nuclear-encoded σ factors, which recognize the bacterial-type promoter sequence of the photosynthesis genes while containing canonical –10 and –35 elements (Shiina et al., 2005). PEP activity is severely inhibited by tagetitoxin, an inhibitor of prokaryote RNA polymerase (Mathews and Durbin, 1990). In contrast, NEP preferentially transcribes housekeeping genes such as rpoB, rpl23 and clpP. The NEP promoter resembles the plant mitochondrial promoter sequence and differs completely from the PEP promoter (Kapoor and Sugiura, 1999; Liere and Maliga, 1999). In general, genes for non-photosynthetic components are transcribed by NEP during the early stage of plastid differentiation and development. Subsequent transcription of the genes for photosynthesis-related components is directed by PEP (Hajdukiewicz et al., 1997). Several lines of evidence indicated that a plastid-localized bacteriophage-type RNA polymerase (RpoT) is an NEP (Lerbs-Mache, 1993; Hedtke et al., 1997; Liere et al., 2004). NEP activity is not affected by tagetitoxin.

We have reported previously that the rpoA gene is absent from the plastid genome of the moss Physcomitrella patens, and we have identified a nuclear counterpart, PpRpoA (Sugiura et al., 2003). The loss of rpoA from the plastid genome is a general occurrence in the arthrodontous mosses, suggesting that the rpoA gene was lost from the plastid genome and transferred to the nucleus during the evolutionary history of the mosses (Sugita et al., 2004; Goffinet et al., 2005). However, it is unclear whether the nuclear PpRpoA gene really encodes the α subunit of PEP and is required for the function of PEP. The RNA polymerase core enzyme of Escherichia coli is assembled in the sequence: α→αα→ααβ→ααββ′, indicating that the α subunit plays a key role in the assembly of the core enzyme (Kimura and Ishihama, 1995).

In this study, we identified a second PpRpoA gene (PpRpoA2), and showed that both PpRpoA and PpRpoA2 proteins were fractionated with PEP enzyme activity. The two PpRpoA genes were differently expressed under different light and dark conditions. We disrupted the PpRpoA1 gene by homologous recombination and characterized the disruptant with respect to plastid gene expression. Disruption of the PpRpoA1 gene resulted in the alteration of several chloroplast genes at the transcript level. We discuss this transcription modulation, which appears to be mediated by the two α subunits of PEP in Physcomitrella.


Identification and characterization of the PpRpoA2 gene

To search for PpRpoA paralog(s) we performed a tBLAST search using the query as the amino acid sequence of the PpRpoA against the expressed sequence tag (EST) database at PHYSCObase ( (Nishiyama et al., 2003). End sequences of an EST clone pphf35o11 (DDBJ/EMBL/GenBank accession nos BJ947461 and BJ958286) were found to encode a partial coding sequence homologous to PpRpoA. We then isolated and sequenced the cognate cDNA. The encoded protein comprised 525 amino acid residues, which showed 59.1% identity with PpRpoA. Therefore, the newly identified sequence was designated as PpRpoA2 and the previously identified PpRpoA as PpRpoA1. No other homologous sequences were found in this analysis. Comparison of the PpRpoA2 cDNA and the corresponding genomic sequences in PHYSCObase (gnl|ti|870055012, gnl|ti|870076393, gnl|ti|86233862, gnl|ti|713796403, gnl|ti|692455034 and gnl|ti|846052148) revealed that the PpRpoA2 gene comprises seven exons and six introns (Figure 1a). The first and sixth introns are located in the 5′- and 3′-untranslated regions, respectively. The intron insertion positions are not conserved between PpRpoA1 and PpRpoA2 genes (Figure 1c). As shown in Figure 1b, PpRpoA2 protein has a sequence that is homologous to a part of HSP70 at the N-terminal region. PpRpoA2 is highly homologous to the sequences of known plastid-encoded RpoAs and a cyanobacterium RpoA. Functional amino acid residues were determined in the α subunit of E. coli (Kimura and Ishihama, 1995). These are involved in the dimerization of α subunits (45Arg at residue 45), in the assembly of ααβ (48Leu) and in the core complex formation of ααββ′ (86Lys and 173 Val). The two PpRpoA proteins also have these conserved amino acid residues (Figure 1c), and therefore are predicted be able to assemble the core PEP enzyme.

Figure 1.

 (a) Diagrams of the PpRpoA1 and PpRpoA2 genes. Boxes indicate exons.
(b) Diagrams of the PpRpoA1 and PpRpoA2 proteins. The black boxes indicate the C-terminal domain of the α subunit (αCTD), the gray boxes indicate the N-terminal domain of the α subunit (αNTD) and the white box indicates a part of the HSP70 protein.
(c) Alignment of α subunit homologs of PpRpoA2 (AB293564) with PpRpoA1 (AB110072), Glycine max RpoA (Gm_rpoA; DQ317523), Nicotiana sylvestris (Ns_rpoA; NC007500), Oryza sativa (Os_rpoA; X15901), fern Adiantum capillus-veneris (Ac_rpoA; NC004766), liverwort Marchantia polymorpha (Mp_rpoA; X04465), red alga Cyanidiaschyzon merolae (Cm_rpoA; NC004799) and cyanobacterium Synechococcus elongatus PCC 6301(sy_rpoA; AP008231). The black and white arrowheads indicate the positions of introns in PpRpoA1 and PpRpoA2, respectively. The asterisks indicate the functional amino acid residues, which were identified in the E. coliα subunit (Kimura and Ishihama, 1995).
(d) Localization of PpRpoA-GFP fusion proteins. Physcomitrella protoplasts were transformed with PpRpoA2-GFP fusion plasmids PpRpoA2-gfp (a and b) and PpRpoA1-gfp (c and d). a and c, fluorescence of GFP (green) using the cube U-MNIBA (Olympus,; b and c, fluorescence of chlorophyll (red) using the cube U-MWIG.

PpRpoA2 has the N-terminal extension sequence, which is predicted to specify plastid targeting (a score of 0.863) by the TargetP program for protein sorting (Emanuelsson et al., 2000). To examine the cellular localization of PpRpoA2, we constructed the plasmid PpRpoA2-gfp encoding a chimeric protein of the N-terminal 125 amino acid residues fused to sGFP and introduced it to the moss protonemal protoplasts. Green fluorescence of PpRpoA2-GFP was localized in the chloroplasts (Figure 1d, panel a) as was that of PpRpoA1-GFP (Figure 1d, panel c). This clearly indicates that PpRpoA2 is a plastid-localized protein. This strongly suggests that the two PpRpoA proteins function as an α subunit of the PEP enzyme in the Physcomitrella plastids.

PpRpoA1 and PpRpoA2 proteins are components of PEP enzyme

Amino acid sequence identity and the domain structures of PpRpoA proteins strongly suggested that PpRpoA1 and PpRpoA2 are the α subunit of PEP. To confirm this, we investigated whether PpRpoA1 and PpRpoA2 proteins are contained in PEP active fractions separated by anion exchange column chromatography. As shown in Figure 2a and Figure S1, transcriptional active fractions 21–26 were eluted with about 0.4 m KCl, and these transcriptional activities were severely inhibited by the addition of tagetitoxin, an inhibitor of PEP transcription activity. Western blot analysis showed that PpRpoA1 and PpRpoA2 are detected as 40-kDa and 50-kDa bands, respectively, in the PEP active fraction 25, but not in inactive fractions 11 and 31 (Figure 2b). The cross-reaction of the antibodies against the respective PpRpoA proteins with the recombinant proteins was estimated to be less than 5% (Figure 2c). This result indicates that both PpRpoA1 and PpRpoA2 proteins are bona fide components of the functional PEP enzyme.

Figure 2.

 Preparation of plastid-encoded RNA polymerase active fractions from the isolated chloroplasts and immunoblot detection of PpRpoA1 and PpRpoA2.
(a) Fractionation of transcription activity by anion exchange chromatography. Fractions were eluted with a linear KCl gradient. Transcription activities of each fraction were measured as incorporation of [α-32P]UTP into the synthesized RNA. The broken line indicates the concentration of KCl. Squares indicate transcription activity without tagetitoxin and closed circles indicate transcription activity with tagetitoxin.
(b) Immunoblot detection of PpRpoA1 and PpRpoA2 in the eluted fractions. Aliquots (10 μl) of fractions 11 (lanes 1 and 4), 25 (lanes 2 and 5), and 35 (lanes 3 and 6), were separated using 10% PAGE, and then subjected to immunodetection using anti-PpRpoA1 or anti-PpRpoA2 antisera.
(c) Evaluation of the cross-reaction of PpRpoA1 and PpRpoA2 with the respective anti-PpRpoA antisera. A series of diluted (100, 25 or 5 ng) Thio-PpRpoA1 or His-PpRpoA2 was loaded in 10% PAGE and then subjected to immunodetection.

PpRpoA1 and PpRpoA2 genes are differentially expressed in the protonemata

To investigate the transcript levels of PpRpoA1 and PpRpoA2 in the protonemata, we performed Northern blot analysis. The transcript levels of PpRpoA1 accumulated at substantial levels and showed low-amplitude fluctuation for all RNA samples (Figure 3), as reported previously (Ichikawa et al., 2004). The transcript level of PpRpoA2 decreased significantly after 24 h in the dark, and its transcript declined to 20% of its level in constant light (Figure 3, 4 L versus 3LD). Upon transfer of the moss protonemata back into light, these transcripts accumulated until they were restored to the control level (Figure 3, 3LDL). This profile was similar to that of Lhcb2 (encoding light-harvesting chlorophyll a/b binding protein), a known light-responsive gene. This result indicates that expression of the PpRpoA2 gene is differentially regulated in a light-dependent manner.

Figure 3.

 Transcript levels of the PpRpoA1 and PpRpoA2 genes in the moss protonemata. Total RNA (15 μg) was separated on a 1.2% formaldehyde-containing agarose gel and subjected to northern blot analysis. RNA was extracted from 4-day-old protonemata grown under continuous illumination (4 L), from 4-day-old protonemata treated for 1 day in darkness before harvesting (3LD) and 3LD protonemata treated for a further 1 day under light conditions (3LDL). Ethidium bromide-staining gel was shown as a loading control (rRNA). Sizes of detected transcripts are indicated on the right of the panels.

Figure 4.

 Generation of the PpRpoA1 gene disruptant and phenotype of the disruptant.
(a) The genomic structure of the wild-type and the PpRpoA1 disruptant. The expected fragment sizes after EcoRV digestion of genomic DNA for the DNA-blot analyzed are shown. Primers (P-F, forward and P-R, reverse) and the expected fragment sizes for PCR analysis are also shown.
(b) DNA-blot analysis of G418-resistant mosses. Total DNA from wild-type (WT) and eight independent G418-resistant mosses (#11 to #24) were digested with EcoRV and hybridized with the PpRpoA1 cDNA or ntpII probes. PCR analysis showed that 1.9-kb and 3.9-kb fragments were derived from the wild-type and G418-resistant mosses, respectively.
(c) Total RNA (15 μg) from 4-day-old protonemata of the wild type or transgenic line #22 was subjected to Northern blot analysis with the PpRpoA1 cDNA probe.
(d) Total cellular protein (50 μg) from wild type and line #22 was subjected to Western blot analysis. To control for loading, antiserum detected against tobacco chloroplast RNA-binding protein cp28 was used. The PpRpoA1 protein (40 kDa) was detected in WT but not in line #22.
(e) Morphology of the wild type and the PpRpoA1 disruptant protonemal colonies under continuous light conditions. Scale bars = 1 cm (protonema colony) and 1 mm (leafy shoot). Colonies were grown for 18 days and average diameters and SD of 10 colonies are plotted. The length of the leafy shoots from the wild type and disruptant #22 was also measured and the SD are shown.

Targeted disruption of the PpRpoA1 gene

To investigate further the precise role of PpRpoA1 and PpRpoA2 gene products, we attempted to generate either PpRpoA1 or PpRpoA2 knock-out mosses. Targeted disruption of the PpRpoA1 gene was achieved by inserting an nptII cassette into the HindIII site within exon 2 of the plasmid (Figure 4a). As shown in Figure 4b, we isolated eight G418-resistant mosses and performed Southern blot analysis to verify the targeted disruption. Probing with PpRpoA1 cDNA detected the predicted 6.1-kb EcoRV signal in the wild-type moss. In contrast, an 8.1-kb signal appeared in the G418-resistant moss lines (#14, #22, #23 and #24), corresponding to the 2.0-kb nptII cassette integrated into the PpRpoA1 locus. A uniform population of the transformed moss genome in the transgenic moss was verified further by PCR analysis (Figure 4b). In contrast, all transformants for construction of the PpRpoA2 disruptant possessed both the wild-type and the nptII cassette-inserted PpRpoA2 gene, representing heteroplasmic moss lines (Figure S2). Among the PpRpoA1-disruptant mosses, the #22 transgenic moss was selected as the representative PpRpoA1 disruptant and was characterized further.

Both the transcript and the gene product of PpRpoA1 were not detected by RNA-blot and immunoblot analyses in the #22 transgenic moss (Figures 4c,d). This result clearly indicates that PpRpoA1 protein is absent from the PpRpoA1 disruptant. The PpRpoA1 disruptant displayed the green phenotype like the wild-type mosses, but showed slightly retarded growth (Figures 4d,e). In the continuous light condition, the colony size was smaller in the PpRpoA1 disruptant than in the wild type. The mean length of the leafy shoot of the PpRpoA1 disruptant was the same as that of wild type until the 34-day-old adult gametophore stage. Thereafter, the disruptant leafy shoots grew slowly and were somewhat smaller than those of the wild-type moss. Thus, the phenotypic characters did not differ significantly between the disruptant and wild-type mosses.

Effect of PpRpoA1 disruption on the plastid gene expression

To examine the effect of PpRpoA1 disruption on the expression of PpRpoA2, the transcript level of PpRpoA2 was measured by Northern blot analysis (Figure 5a). The PpRpoA2 transcript level in the PpRpoA1 disruptant was twice that in the wild type. To further examine the steady-state transcript levels of the plastid genes in the PpRpoA1 disruptant, we performed plastid DNA microarray analysis. In the 4-day-old protonemata grown under constant light conditions, most plastid genes including psaA, psbA, psbD and rrn16 were expressed at similar levels in the wild type and in the PpRpoA1 disruptant, but some tRNA levels increased in the disruptant (Table 1).

Figure 5.

 Transcript levels of the PpRpoA2 gene and plastid genes in the wild type and the PpRpoA1 disruptant. Total RNA was extracted from the wild type (WT) or the PpRpoA1 disruptant (ΔA1) of 4-day-old protonemata grown under continuous illumination (4 L) and 4-day-old protonemata treated for 1 day in darkness before harvesting (3LD).
(a) Transcripts of three nuclear genes, PpRpoA1, PpRpoA2 and Lhcb2 were detected by Northern blot analysis.
(b)–(e) Transcripts of plastid genes were detected using plastid gene-specific probes. Positions of RNA markers were indicated on the right of panels.

To confirm the microarray analysis results, we performed RNA blot hybridization (Figure 5). The transcripts of psaA, psbA, psbD, chlN, atpF, ycf4, trnL-UAG, trnfM-CAU and rrn16 accumulated at similar levels in the wild type and in the PpRpoA1 disruptant under both light and dark conditions (Figure 5b). This result was consistent with that of the array analysis. In addition, the transcripts of six genes (atpB, ycf2, matK, rpoC1, chlB and psaM) accumulated greatly and at similar levels in the wild type and in the disruptant grown under constant light, whereas their transcripts declined to faint levels in dark conditions (Figure 5c). In contrast, petN and ycf3 transcript levels decreased to 40% and 20% of the wild-type level, respectively, under constant light conditions (Figure 5d, 4L lanes). In addition, psbZ transcript level in the disruptant decreased to 25% of wild-type level under dark conditions (Figure 5d, 3LD lanes), although it was unchanged under constant light conditions (4 L lanes). The six tRNAs accumulated at 2–10-fold higher levels in the disruptant than in the wild-type protonemata grown under constant light conditions (Figure 5e). In dark-treated protonemata, however, their transcript levels were the same in the disruptant and in the wild type (Figure 5e).

Disruption of PpRpoA1 does not affect the expression of plastid σ factor genes

The effect of PpRpoA1 gene disruption on the plastid gene expression may be caused by the modulation of the expression of plastid σ factors. To investigate this possibility, the transcript levels of three genes, PpSig1, PpSig2 and PpSig5, encoding the plastid σ factor were measured by reverse transcriptase-polymerase chain reaction (RT-PCR) in the wild type and the PpRpoA1 disruptant (Figure 6). The three σ factor genes were highly expressed in the light (4 L and 3LDL) and were low in the dark (3LD). This expression profile was similar to those of the PpRpoA1 and PpRpoA2 genes. Although there could be other differences at the protein level, at least the transcript levels of the three PpSig genes were not significantly different between the wild type and the PpRpoA1 disruptant.

Figure 6.

 Semi-quantitative reverse transcriptase-polymerase chain reaction analysis of PpRpoA1, PpRpoA2 and plastid σ factor transcript levels in the wild type and the PpRpoA1 disruptant. Total RNA was extracted from the wild type or the PpRpoA1 disruptant of 4-day-old protonemata grown under continuous illumination (4 L), 4-day-old protonemata treated for 1 day in darkness before harvesting (3LD) and 3LD protonemata treated for a further 1 day under light conditions (3LDL), reverse-transcribed, and amplified by PCR using a primer set specific to each PpRpoA, PpSig or PpActin3 gene.


In this study, we identified the second nuclear gene PpRpoA2 encoding the PEP α subunit. Both PpRpoA2 and PpRpoA1 proteins were detected immunologically in the tagetitoxin-sensitive PEP active fractions. This biochemical property confirms that the nuclear-encoded RpoA constitutes the core PEP enzyme.

Northern blot analysis (Figure 3) indicated that expression of the PpRpoA2 gene is regulated tightly in a light-dependent manner, as is the light-responsive gene Lhcb2, whereas PpRpoA1 is expressed constitutively in both light and dark conditions. This suggests that the two PpRpoA proteins play different roles in the transcription of plastid genes. Although disruption of the PpRpoA1 gene resulted in the slightly retarded growth of protonemal colonies, expression of most plastid genes, including psbA, rrn16 or psaA, was not affected by disruption of PpRpoA1. This implies that PpRpoA1 is dispensable to plastid function and that PpRpoA2 plays a central role in plastid transcription. An alternative possibility is that PpRpoA2 merely complements the loss of PpRpoA1 function. We prefer the first suggestion because PpRpoA1 disruptants were obtained easily, but PpRpoA2 was not disrupted.

The most interesting finding is that the expression of petN, psbZ, ycf3 and several tRNA genes was altered in the PpRpoA1 disruptant (Figure 5). Of these genes, three (petN, psbZ and ycf3) can be categorized as PpRpoA1-PEP-dependent genes. Thus the modulation of transcription may be mediated by two α subunits of PEP in the moss Physcomitrella. In the wild-type moss chloroplasts the PEP enzyme comprises PpRpoA1 (α1 subunit) or PpRpoA2 (α2 subunit), or both, together with ββ′β′′, and may exist as three isoforms (α1α1ββ′β′′, α1α2ββ′β′′ and α2α2ββ′β′′). In contrast, in the PpRpoA1 disruptant, PEP exists presumably as a uniform complex of α2α2ββ′β′′. The lower petN, psbZ and ycf3 transcript levels in the PpRpoA1 disruptant indicates that the three genes are transcribed predominantly by PpRpoA1-PEP (α1α1ββ′β′′). Interestingly, transcript levels of some tRNA genes were very low in the light and high in the dark (Figure 5e). The transcript levels of those tRNA genes were enhanced significantly by PpRpoA1 disruption even under light conditions (Figure 5e). This might be caused by overexpression of PpRpoA2. Perhaps their transcript levels may be modulated by different stability under diurnal day and night control in the disruptant.

In E. coli the α subunit is required for transcription activation by protein factors, and for interaction with DNA-activation elements (Kimura and Ishihama, 1995). Sigma factors are the most important determinants for the selection and initiation of transcription of plastid genes (Shiina et al., 2005), and recognize the promoter consisting of –10 and –35 elements. PpRpoA proteins must interact with some σ factor bound to the promoter. To compare the promoter sequences of the PpRpoA1-dependent or -independent genes, we used a primer extension experiment to identify putative transcription initiation sites of the moss plastid genes (Figure S3). As shown in Figure 7, PpRpoA1-independent genes have canonical –35 and –10-like elements, and PpRpoA1-dependent genes (petN and ycf3) also have a canonical –35 element and an extended –10 sequence, GAT(G/A)TATATA(T/A)AT. The other putative PpRpoA1-dependent gene, psbZ, has a sequence TCGGCCA that is also found in the upstream region of the ycf3 − 253. Among the three Physcomitrella plastid σ factor genes, the PpSig1 and PpSig2 genes are expressed throughout the day, and their fluctuations suggest very low amplitude diurnal rhythms of mRNA levels. In contrast, PpSig5 mRNA showed a very high amplitude diurnal rhythm with peaks observed in the light phases (Ichikawa et al., 2004). Although similar results were also observed in this study (Figure 6), these plastid σ factors are unlikely to be responsible for the alteration of plastid gene expression in the disruptant. However, we cannot exclude the possibility that alternation of protein levels or phosphorylation states of PpSigs influences the steady-state transcript levels of plastid genes in the disruptant. Therefore, we hypothesize that different combinations of the two α subunits with a certain σ factor (rather than PpSig1, PpSig2 and PpSig5), or that some transcription factors interact with an upstream element or promoter of each plastid gene to modulate the strength of transcription activity. PpRpoA2 possesses a portion of DnaK/HSP70, which may provide an additive function to PpRpoA2 as an α subunit. We speculate that PpRpoA2 protein is able to facilitate or interfere with the σ or other transcription factors.

Figure 7.

 Alignment of nucleotide sequences of the region upstream of the 5′-ends of PpRpoA1-independent and -dependent gene transcripts. The underlines indicate the position corresponding to the 5′-end of the major transcript. The arrowheads indicate the initiation nucleotide of mature tRNAs. The σ70-type promoter in Escherichia coli comprising TTGACA (–35 element) and TATAAT (–10 element) are indicated by the gray box. The conserved element specific to the PpRpoA1-dependent promoter is boxed.

Previous studies of tobacco plants have demonstrated that photosynthesis genes are transcribed by PEP, that some plastid genes such as atpB and rrn16 are transcribed by both PEP and NEP, and that most housekeeping genes are transcribed by NEP (Hajdukiewicz et al., 1997). A relative increase in NEP activity was observed in PEP-deficient tobacco (Krause et al., 2000; Legen et al., 2002). We did not construct PEP-deficient mosses in this study, and therefore we cannot conclude whether NEP exists in the P. patens chloroplasts. Double knock-out mutants of PpRpoA1 and PpRpoA2 are needed to address this question. Alternatively, disruption of plastid genes rpoB, rpoC1 or rpoC2 may also help address this issue. We attempted to construct plastid rpo gene knock-out mosses, but we obtained only heterotransplastomic lines, which possessed both the wild-type plastomes and the rpo gene-disrupted plastome (unpublished data). In the moss P. patens, the haploid gametophyte dominates the life cycle, represented by the filamentous protonema (juvenile gametophytes) and the leafy moss plant (adult gametophyte). Plastid ontogeny in mosses differs distinctly from that in vascular plants (Reski, 1998). This may be the reason why this moss developed such a unique system for plastid transcription that is unlike higher plants.

Experimental procedures

Plant material

Physcomitrella patens (Hedew.) Bruch & Schimp subsp. patens Tan was grown at 25°C under continuous illumination at 30 μmol m−2 sec−1 on the minimal medium (BCD medium) supplemented with 0.5% glucose, 1 mm CaCl2 and 5 mm diammonium (+)-tartrate agar plate as described previously (Sugiura and Sugita, 2004).

Isolation and sequence analysis of cDNA

cDNA encoding PpRpoA2 was prepared using primers cA2.F (5′-TCTCTCCTGCAGGCCTCTTCACCTCTAC-3′) and cA2.R (5′-GCCTGTCAGGCTCCATCTCTAAGTGGTTTC-3′) designed from sequences of an EST clone (pphf35o11). Sequencing was performed with an ABI PRISM 3100 sequencer and the DYEnamic ET Terminator Cycle Sequencing Kit (GE Healthcare, using appropriate sequencing primers. Alignments of amino acid sequences were constructed by the ClustalX program, version 1.81 (Thompson et al., 1994).

Construction of the PpRpoA2-GFP fusion gene and moss transformation

A DNA fragment encoding the N-terminal 125 amino acid residues of PpRpoA2 was amplified from the cDNA as above with primers A2GFP.F (5′-CCCGTCGACCACCATGGCAACTGTCATGGGCGC-3′) and A2GFP.R (5′-ACGTGTCGACGGCCTTTTCTGCAGCTTCTGTAA-3′). The PCR product was digested with SalI and inserted into the SalI-cleaved CaMV35S-sGFP(S65T)-nos3′ (Chiu et al., 1996) to create the PpRpoA2-gfp construct. The reporter construct was introduced into the protoplasts prepared from the 3-day-old protonemata. As a positive control, the PpRpoA-gfp construct was also used for transformation (Sugiura et al., 2003).

Chloroplast isolation and anion exchange chromatography

Intact moss chloroplasts were isolated from 4-day-old protonemal cells as described previously (Kabeya and Sato, 2005). To prepare the chloroplast lysate, intact chloroplasts were resuspended in buffer 1 (10 mm Tris–HCl, pH 8.0, 1 mm EDTA, 5 mm DTT, 1 m KCl) and incubated for 15 min on ice. An equal volume of buffer 2 (50 mm Tris–HCl, pH 8.0, 10 mm MgCl2, 2 mm DTT, 20% sucrose, 50% glycerol) was added to the lysed plastids, and (NH4)2SO4 was added to the plastid lysates to a final concentration of 10%. The mixture was stirred for 30 min on ice and then centrifuged at 50 000 g for 1 h. (NH4)2SO4 was added to the supernatant to a final concentration of 60%. The pellet was resuspended in buffer 3 (20 mm HEPES, pH 8.0, 100 mm KCl, 12.5 mm MgCl2, 2 mm EDTA) and dialyzed with 10 volumes of buffer 3 containing protease inhibitor cocktail (Roche, for 16 h. The supernatant was loaded onto a Mini-Q FPLC column (GE Healthcare). The column was washed with buffer 3, eluted with a 30-ml liner gradient of 0.1–0.4 m KCl, and then to 1.0 m KCl, and fractions (1 ml) were collected.

Measurement of transcription activity

Transcription activity was measured as incorporation of [α-32P]UTP. The reactions were carried out in a total volume of 60 μl containing transcription buffer (60 mm Tris–HCl, pH 8.0, 10 mm MgCl2), 80 U RNase inhibitor (Takara,, 2% glycerol, 2 mm DTT, 0.5 mm ATP, 0.5 mm CTP, 0.5 mm GTP, 5 μm [α-32P]UTP (at a specific radioactivity 0.16 GBq μmol−1), 6 μg denatured calf thymus DNA as a template and 10 μl dialyzed supernatants of each fraction, as described above. The reaction mixtures were incubated in the presence or absence of tagetitoxin (10 μm) for 30 min at 25°C and then 10-μl aliquots were spotted onto DEAE paper (DE-81; Whatman, After successive washing with 5% Na2PO4, water and ethanol, the radioactivity was determined by scintillation counting.

Preparation of recombinant PpRpoA proteins and Western blot analysis

A 1068-bp DNA fragment encoding the amino acid residues 90–450 of PpRpoA1 were amplified from the PpRpoA1 cDNA (Sugiura et al., 2003) using specific primers rA1.F (5′-GACGTACTAGCTTGGACAAAAGCT-3′) and rA1.R (5′-ATTGCATAATGGATTGTTCTCAG-3′). The PCR product was inserted into the expression vector pBAD/Thio-TOPO (Invitrogen, and the resultant plasmid pBAD-A1 was obtained. A 1146-bp DNA fragment encoding the amino acid residues 143–525 of PpRpoA2 was amplified from the PpRpoA2 cDNA (this study) using specific primers rA2.F (5′-AGTAGATCTACTACCACTGCGGACGGACCCATG-3′) and rA2.R (5′-AGTGTCGACCTACTACGTCCTGCAGTGACTTTGCAG-3′). The PCR product was digested with SalI and BglII, and inserted into BamHI and SalI sites of the expression vector pQE-30 (Qiagen, and the resultant plasmid pQE-A2 was generated. pBAD-A1 and pQE-A2 was transformed into E. coli LMG194 and M15/pRep4 cells, respectively. The overexpression and purification of thioredoxin-tagged PpRpoA1 protein (Thio-PpRpoA1) or His-tagged PpRpoA2 protein (His-PpRpoA2) with Ni-NTA agarose was carried out according to the manufacturer’s instructions. Purified recombinant proteins were used to immunize rabbits. Polyclonal antisera were obtained and used in the immunoblot analysis. For immunoblot analysis, sodium dodecylsulphate–polyacrylamide-gel electrophoresis and blotting were carried out using a 10% polyacrylamide gel as described previously (Kabeya et al., 2002). Anti-tobacco chloroplast RNA-binding protein cp28 antibody was used as the control (Nakamura et al., 1999).

Isolation and gel-blot analysis of DNA and RNA

Total DNA and RNA were isolated from the protonemata as described previously (Hattori et al., 2007). RNA was extracted from the protonemata grown under different light and dark conditions, as indicated. For Southern blot analysis, DNA was digested with restriction enzymes, separated on 1% agaraose gel, and blotted onto a Hybond N+ membrane (GE Healthcare). The membrane was hybridized for 15 h at 65°C with a [32P]-labeled DNA probe and washed at 65°C. For RNA gel-blot analysis, total RNA (15 μg) was subjected to electrophoresis in 1.2% formaldehyde-containing agarose gel, and transferred to Hybond N+ membranes. Hybridization and detection were carried out as described using digoxigenin-labeled DNA probes (Kabeya et al., 2002). As plastid gene-specific DNA probes, DNA fragments spotted on the moss plastid DNA microarray (Nakamura et al., 2005) were used. A PpRpoA1 cDNA probe (a NotI-digested 1414-bp 5′ RACE-cDNA clone; Sugiura et al., 2003) and an nptII gene cassette probe (a HindIII-digested fragment from pMBL6, a gift from Dr Jesse Machuka of the Physcomitrella EST Programme) were used. A PpRpoA2 probe was amplified with the primers rA2.F and rA2.R, and a Lhcb2 probe was prepared using PCR with primers Lhcb2.F (5′-TAACGGTGAGTTCGCTGGTGAC-3′) and Lhcb2.R (5′-GTTCATGTCAATAGTCTAGTTC-3′).

Construction for PpRpoA gene disruption and moss transformation

A 2511-bp region containing the PpRpoA1 gene was amplified from P. patens genomic DNA with A1_KO.F (5′-GTTAACAAAACATACAATGTAAAG-3′) and A1_KO.R (5′-AATGCGGTGGTAAACTGGTCTCTG-3′), and cloned into the pGEM-T Easy (Promega, to generate pYK-ΔA1. The chimeric nptII gene cassette from pMBL6, which consists of the cauliflower mosaic virus 35S promoter, the nptII coding region and the 35S terminator, was excised as a 1961-bp HindIII fragment. The nptII cassette was inserted into the HindIII site in the exon 2 of the PpRpoA1 gene in pYK-ΔA1, either with the same (pYK-ΔA1-1) or opposite (pYK-ΔA1-2) direction of transcription as PpRpoA1. For construction of the PpRpoA2 disruptant, a 4440-bp DNA fragment (AB293563) was amplified with A2_KO.F (5′-ATGCGGCCGCTTGTAGATGATAATACCTCAATTCCGA-3′) and A2_KO.R (5′-ATGCGGCCGCGCAAATAGTAAGACGTCCAGTAAGA-3′). The PCR fragment was digested with NotI and cloned into the NotI site of pBlueScriptII SK+ to generate pYK-ΔA2. The nptII cassette was inserted into the NcoI sites in the exon 2 to exon 4 of the PpRpoA2 gene in pYK-ΔA2.

Transformation was performed essentially according to the procedure of Nishiyama et al. (2000). NotI-BstXI-digested pYKΔ-1 or NotI-digested pYKΔ-2 (30 μg) was incubated at 45°C for 5 min with protonemal protoplasts in polyethleneglycol (PEG) 6000. PEG-treated protoplasts were incubated at 25°C in the dark, and then in protoplast regeneration medium for 3 days under continuous light. Regenerated protoplasts were transferred to BCDATG medium containing 50 μg ml−1 geneticin (G418) to select transformants. For verification of PpRpoA1 disruptants, PCR was performed using P-F (5′-GTGAGAGGATTGAGACTGGTG-3′) and P-R (5′-TAGCCATAGATCAATAAAACAACC-3′). For verification of PpRpoA2 disruption, PCR was performed using A2C.F (5′-TAAGAGGAATTCGACTGTAGTTGCG-3′) and A2C.R (5′-CGTTTGTGTGATCAATCATCCACG-3′).

Microarray analysis

Plastid DNA microarray analysis was performed as described previously (Nakamura et al., 2005). Fluorescence cDNA probes were generated by direct incorporation of Cy3- or Cy5-dUTP (GE Healthcare) during reverse transcription. Briefly, 10 μg of plastid RNA from wild-type and PpRpoA1 disruptant mosses was annealed with a mixture of 216 plastid gene primers (1 pmol each) at 70°C for 5 min. Reverse transcription, purification of cDNA probes, hybridization and washing were performed as described by Nakamura et al. (2005). Fluorescent images were visualized and analyzed with GenePix 4000 and accompanying software (Axon Instruments,

Semi-quantitative RT-PCR

DNA-free total cellular RNA (5 μg) was reverse-transcribed using SuperScript II (Invitrogen) with oligo(dT)15 primer, and first-strand cDNA was synthesized as previously described (Aoki et al., 2004). PCR was performed using the first-strand cDNA and appropriate primer set as follows: A1-RT.F (5′-TGGTCTCTGCTATAGAAGGTTCGAATTCTA-3′) and A1-RT.R (5′-CTCATCAGAGTCACTGCGGATCACAAGCTC-3′) for PpRpoA1, A2-RT.F (5′-TGAAGGACAGTCAATCCGTACTGAGGCTTA-3′) and A2-RT.R (5′-AATGGAGATACGGCAAATCGAGCGTAATGC-3′) for PpRpoA2. The primers for PCR analysis of PpSig (Ichikawa et al., 2004) were as follows: 5′-AAATCCGGCAGTCCGTCTGCTCGT-3′ and 5′-ACTGATGCTCTCTAGTGACA-3′ for PpSig1, 5′-GTTGAATTGGATACAGAGGCT-3′ and 5′-GCTCCTGAACCAGCATTCGCTTTG-3′ for PpSig2, 5′-CAAGTGGCTGAGGATCAGCAAGT-3′ and 5′-TTGGCGCGTTGGATATTCACTCT-3′ for PpSig5. As a control, an actin3 gene sequence was amplified using primers (5′-CGGAGAGGAAGTACAGTGTGTGGA-3′ and 5′-ACCAGCCGTTAGAATTGAGCCCAG-3′, Aoki et al., 2004). A cycle of PCR consisted of 30-s denaturation at 94°C, 30-s annealing at 55°C and 40-s extension at 72°C. The optimal cycle number of PCR was 30 cycles for PpRpoA1 and PpRpoA2, 32 for PpSig1, 35 for PpSig2 and PpSig5, and 26 for PpActin3. PCR products were subjected to 2% agarose gel electrophoresis.


We thank Dr Jesse Machuka for pMBL6 as part of the Physcomitrella EST Programme at the University of Leeds and Washington University (St Louis, MO, USA). We also thank Dr Shin-ya Miyagishima for kindly supporting this work, and Dr Takahiro Nakamura and Dr Setsuyuki Aoki for valuable discussions. This work was supported by a Grant-in-aid from the Ministry of Agriculture, Forestry and Fisheries (Bio-Design Project), and by a Research Grant from the DAIKO FOUNDATION (Nagoya). YK was a recipient of a Japan Society for Promotion of Science (JSPS) Post-doctoral Fellowship.