Organellar gene transcription and early seedling development are affected in the rpoT;2 mutant of Arabidopsis


  • Kyoko Baba,

    1. Umeå Plant Science Centre, Department of Forest Genetics and Plant Physiology, Swedish University of Agricultural Sciences, 901 83 Umeå, Sweden,
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Julien Schmidt,

    1. Umeå Plant Science Centre, Department of Forest Genetics and Plant Physiology, Swedish University of Agricultural Sciences, 901 83 Umeå, Sweden,
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Ana Espinosa-Ruiz,

    1. Umeå Plant Science Centre, Department of Forest Genetics and Plant Physiology, Swedish University of Agricultural Sciences, 901 83 Umeå, Sweden,
    Search for more papers by this author
    • Present address: Centro Nacional de Biotecnología-CSIC, Campus de la Universidad Autónoma de Madrid, Cantoblanco, 28049 Madrid, Spain.

    • These authors contributed equally to this work.

  • Arsenio Villarejo,

    1. Umeå Plant Science Centre, Department of Plant Physiology, Umeå University, 901 87 Umeå, Sweden,
    Search for more papers by this author
  • Takashi Shiina,

    1. Faculty of Human and Environment, Kyoto Prefectural University, Shimogamo, Sakyo-ku, Kyoto 606-8522, Japan, and
    Search for more papers by this author
  • Per Gardeström,

    1. Umeå Plant Science Centre, Department of Plant Physiology, Umeå University, 901 87 Umeå, Sweden,
    Search for more papers by this author
  • Aniruddha P. Sane,

    1. Plant Gene Expression Laboratory, National Botanical Research Institute, Lucknow-226001, India
    Search for more papers by this author
  • Rishikesh P. Bhalerao

    Corresponding author
    1. Umeå Plant Science Centre, Department of Forest Genetics and Plant Physiology, Swedish University of Agricultural Sciences, 901 83 Umeå, Sweden,
    Search for more papers by this author

For correspondence (fax +46 90 7865901; e-mail


An Arabidopsis mutant that exhibited reduced root length was isolated from a population of activation-tagged T-DNA insertion lines in a screen for aberrant root growth. This mutant also exhibited reduced hypocotyl length as well as a delay in greening and altered leaf shape. Molecular genetic analysis of the mutant indicated a single T-DNA insertion in the gene RpoT;2 encoding a homolog of the phage-type RNA polymerase (RNAP), that is targeted to both mitochondria and plastids. A second T-DNA-tagged allele also showed a similar phenotype. The mutation in RpoT;2 affected the light-induced accumulation of several plastid mRNAs and proteins and resulted in a lower photosynthetic efficiency. In contrast to the alterations in the plastid gene expression, no major effect of the rpoT;2 mutation on the accumulation of examined mitochondrial gene transcripts and proteins was observed. The rpoT;2 mutant exhibited tissue-specific alterations in the transcript levels of two other organelle-directed nuclear-encoded RNAPs, RpoT;1 and RpoT;3. This suggests the existence of cross-talk between the regulatory pathways of the three RNAPs through organelle to nucleus communication. These data provide an important information on a role of RpoT;2 in plastid gene expression and early plant development.


Mitochondrial and plastid genes are transcribed by a set of RNA polymerases (RNAPs) distinct from RNAPs, which transcribe nuclear genes (Gray and Lang, 1998; Hess and Börner, 1999). No RNAP genes are found in the mitochondrial genome (Unseld et al., 1997). In yeast, a nuclear-encoded T3/T7 bacteriophage-like RNAP is found to serve as a mitochondrial RNAP (Greenleaf et al., 1986; Kelly et al., 1986). Genes homologous to the yeast mitochondrial RNAP genes are found in a phylogenetically broad range of eukaryotes, including higher plants, suggesting conservation of the mitochondrial transcriptional machinery among most organisms (Cermakian et al., 1996).

Transcription in plastids is considerably more complex than in mitochondria and involves at least two distinct groups of RNAPs (Maliga, 1998; Smith and Purton, 2002; Stern et al., 1997). One of these, designated as the plastid-encoded plastid (PEP) RNAP, is encoded by the plastid genome of higher plants and is similar to the eubacterial multisubunit RNAP found in Escherichia coli (Allison et al., 1996; Igloi and Kössel, 1992; Little and Hallick, 1988; Sugiura, 1992). The PEP is mainly responsible for transcription of genes whose products form the core components of photosystem (PS)II, PSI, the cytb6/f complex and the large subunit of RuBisCO (Hajdukiewicz et al., 1997; Hess and Börner, 1999). These genes are under control of E. coli-type promoters (Hanley-Bowdoin and Chua, 1987) that are recognized by specific sigma factors encoded by the nucleus (reviewed by Allison, 2000).

The other class of RNAPs is nuclear-encoded and designated as nuclear-encoded plastid (NEP) RNAP. Evidence for the existence of NEPs came from studies on Epiphagus (Morden et al., 1991), the plastid ribosome deficient albostrians mutant of barley and heat-bleached rye (Hess et al., 1993), the iojap mutant of maize (Han et al., 1993) and the rpoB (RNA polymerase beta subunit)-deleted mutants of tobacco (Allison et al., 1996). In all these plants, absence of PEP does not prevent transcription of a certain set of genes related mainly to the translational machinery. These genes are characterized by non-E. coli-type promoters and include rps (ribosomal protein S)15, the rpo (RNA polymerase)B/C1/C2 operon, rpl (ribosomal protein L)16, rrn (ribosomal RNA)16, several tRNA genes, the atp (ATP synthase)B/E operon, atpI, etc. (Gruissem et al., 1986; Hajdukiewicz et al., 1997; Hess et al., 1993; Iratni et al., 1994; Neuhaus et al., 1989; Vera and Sugiura, 1995). Biochemical studies revealed the existence of at least two types of NEPs in spinach, including a single-subunit RNAP of a molecular mass of 110 kDa with features similar to the T3/T7 bacteriophage RNAP (Bligny et al., 2000; Lerbs-Mache, 1993). The NEPs resemble the phage-type RNAPs in that they are insensitive to inhibitors such as tagetin and rifampicin (Bligny et al., 2000; Kapoor et al., 1997; Liere and Maliga, 1999). However, unlike phage-type polymerases, they require specific factors to initiate transcription (Bligny et al., 2000). While NEPs have been cloned from a variety of plant species such as Chenopodium album (Weihe et al., 1997), Nicotiana (Hedtke et al., 2002; Kobayashi et al., 2001), maize (Chang et al., 1999), Physcomitrella patens (Richter et al., 2002), and Arabidopsis thaliana (Hedtke et al., 1997, 2000) their role in the regulation of plant growth and development is unclear.

In Arabidopsis, three phage-type RNAPs were identified on the basis of sequence comparison. While RpoT;1 and RpoT;3 were shown to be directed exclusively to mitochondria and chloroplasts, respectively, a third RNAP, RpoT;2, had dual targeting properties (Hedtke et al., 1999, 2000). Homologs of this dual targeted RNAP have also been isolated from Nicotiana (Hedtke et al., 2002; Kobayashi et al., 2001), and the moss P. patens (Richter et al., 2002). Transcriptional activity of recombinant RpoT;2 was demonstrated in vitro by incorporation of UTP using calf thymus DNA as a template (Hedtke et al., 2000). As this RNAP serves two organelles that are very different in function and their mode of gene regulation, there has been considerable interest in determining its role, its specificity in transcription of organellar genes and its regulation in different tissues. However, the presence of multiple RNAPs in the same organelles has made it difficult to distinguish between the contributions of different RNAPs.

In this study, an Arabidopsis mutant was identified in which the T-DNA was inserted into the RpoT;2 gene. Using this mutant, we demonstrate that RpoT;2 is responsible for the light-induced accumulation of several plastid genes transcripts during early seedling development, thereby influencing plant development. Surprisingly, the mutation in RpoT;2 gene appears to have little effect on the transcription of mitochondrial genes examined in this study.


Identification of a T-DNA-tagged short-root mutant

In a screen for aberrant root growth performed using activation-tagged T-DNA lines (Weigel et al., 2000), a mutant (termed line 833) was isolated that exhibited short roots compared to wild type when grown on MS medium (Figure 1b). The mutant plants also displayed a general reduction in growth during the juvenile stage of development in the aerial part (Figure 1a). Additionally, the mutant seedlings displayed a delay in greening when 6-day-old etiolated mutant seedlings were transferred to light (Figure 1c). The leaves of the soil-grown adult plants showed altered morphology, with a more rounded and wrinkled form and a defect in petiole elongation (Figure 1a). A delay of 4–7 days was also observed in flowering time (results not shown). From the 4th week onwards, the mutant recovered gradually and differences between mutant and wild type became less noticeable. In the later stages, the major difference between the mutant and the wild type was the altered leaf shape in the mutant.

Figure 1.

Activation-tagged mutant line (line 833) displaying slow growing phenotype.

(a) Three-week old plants grown in soil. (i) wild type; (ii) the mutant line 833; (iii) the F1 progeny of the cross between the mutant line 833 and the rpoT;2 mutant line from the GABI-Kat T-DNA collection (line 286E07); (iv) The mutant line 833 expressing 35S::RpoT;2. The magnification is the same for all the pictures.

(b) Short-root phenotype of the mutant line 833. WT, wild type; 833, mutant line 833; 8 and 15, mutant line 833 expressing 35S::RpoT;2. Seedlings were grown on a MS medium without sucrose for 7 days.

(c) Light-induced chlorophyll accumulation (µmol g−1 FW) in the dark-grown wild-type and mutant line 833 seedlings. 6D, 6-day dark-grown; 6D1L, 6-day dark 1-day light; 6D6L, 6-day dark 6-day light.

The T-DNA insertion in the mutant line 833 was found after plasmid rescue in the last intron between exons 18 and 19 of the RpoT;2 gene (At5g15700.1, nucleotide 31514 of cosmid F14F8, GenBank sequence AL391144) in chromosome 5, that encodes a phage-type RNAP (Hedtke et al., 2000; Figure 2a). Southern blot analysis using a probe derived from the Bar (BASTA resistance) gene of the T-DNA proved a single T-DNA insertion in the mutant genome (data not shown). A second hybridization using an RpoT;2 probe showed that the 7-kbp band, corresponding to the RpoT;2 gene in the wild-type genome shifts to a higher molecular weight size, confirming the presence of the T-DNA in the RpoT;2 gene in the mutant (Figure 2b). Northern blot analysis of 3-week old wild-type and mutant plants was performed to examine the effect of T-DNA insertion in RpoT;2 expression (Figure 2c) using an RpoT;2 RNA probe corresponding to the 5′ end of the coding sequence. In wild-type plants, a band at approximately 3.7 kbp, which is likely to be the mature mRNA (3.76 kbp), was observed, while in the mutant, a strong band around 3 kbp was detected. These results indicate that the insertion of T-DNA does not result in loss of the RpoT;2 transcript but causes accumulation of truncated mRNA.

Figure 2.

Identification of T-DNA insertion.

(a) Map of the RpoT;2 gene with T-DNA insertions in the mutant lines 833 and 286E07. E, EcoRI sites.

(b) Southern blot on wild-type (WT) DNA and rpoT;2 mutant (mt) DNA (isolated from line 833). Two micrograms of genomic DNA were digested by EcoRI and hybridized with radio-labeled PstI probe indicated in the figure. M, DNA molecular size marker.

(c) Northern blot of RpoT;2. mRNA (1.5 µg) extracted from 3-week old WT and mutant whole Arabidopsis seedlings were loaded on each lane and hybridized with radio-labeled 480-bp-long RpoT;2 RNA probe corresponding to the 5′ end of the coding sequence. The arrow indicates the position of the 3.7 kbp WT mRNA.

Genetic analysis of the rpoT;2 mutant

The homozygous rpoT;2 mutant line 833 was crossed with wild-type Columbia plants. All F1 seedlings displayed a wild-type aerial phenotype, thus indicating the recessive nature of the mutation and ruling out the possibility that the mutant phenotype is caused by overexpression of truncated RpoT;2 mRNA or any other genes because of the presence of the 35S enhancer in the T-DNA. Segregation of the rpoT;2 mutant phenotype was analyzed in the F2 population by measuring the root length of 7-day-old seedlings on plates without sucrose in the growth medium. Of 369 F2 seedlings, 254 exhibited a long-root phenotype and 115 exhibited a short-root phenotype (Table 1). Among them, 100 long-root seedlings and 54 short-root seedlings were transferred to soil to determine the aerial phenotype. None of the long-root seedlings exhibited altered leaf shape, and 31 out of 100 were Basta (AgrEvo, Berlin, Germany) sensitive, while all of 54 selected short-root seedlings exhibited altered leaf shape and were Basta resistant. These results support the co-segregation of the T-DNA insertion with mutant phenotype in line 833. In addition, a second rpoT;2 mutant allele (line 286E07) was obtained from the GABI-Kat (German plant genomic research program-Kölner Arabidopsis T-DNA lines) T-DNA mutant collection (Rosso et al., 2004). This line is likely to be a null mutant as the T-DNA insertion of this allele is found in the first exon of the RpoT;2 gene (Figure 2a). The homozygous mutant line 286E07, as well as the F1 progeny of the cross between the mutant lines 833 and 286E07, displayed short-root phenotype (data not shown), slow growth, and the same altered leaf shape (Figure 1a) as the mutant line 833. Furthermore, a T-DNA carrying the entire coding sequence of the wild-type RpoT;2 cDNA under the control of 35S promoter was introduced into the homozygous rpoT;2 mutant plants (line 833). The T2 plants obtained from the transformed plants exhibited long roots (Figure 1b) and normal growth (Figure 1a) comparable to that of the wild type, which indicates functional complementation of the rpoT;2 mutation by the expression of the wild-type RpoT;2 cDNA. Taken together, these results indicate that the short root, slow growth, and altered leaf shape in the rpoT:2 mutant line 833 result from the insertion of a T-DNA into the RpoT;2 gene.

Table 1.  Result of segregation test of F2 population
Root phenotypeAerial phenotypeBastaresistantBastasensitive
Long (n = 254)
Short (n = 116)

Developmental- and tissue-specific transcript accumulation pattern of the RpoT genes

The rpoT;2 mutant plants showed a defect in chloroplast function. However, this defect was visible only during the early stages of plant development. This suggests that RpoT;2 plays a key role in young photosynthetic organs. To test this possibility, the transcript levels of RpoT;2 were investigated by semiquantitative RT-PCR in young and old leaves and in roots of 4-week-old mature mutant plants. As apparent from Figure 3(b) (middle), RpoT;2 transcripts accumulated in these organs at a constant level. Moreover, this transcript accumulation was age-independent (Figure 3a). Thus, we conclude that RpoT;2 is transcribed ubiquitously, with the regulation of its function probably taking place at the post-transcriptional level.

Figure 3.

Semiquantitative RT-PCR analysis of RpoT gene expressions.

(a) mRNA accumulation of RpoT;2 gene in wild type (WT) at 6, 10, 14, 20, and 30 days after germination. A representative image is shown in the upper part of the figure.

(b) Tissue-specific expression pattern of RpoT;1, RpoT;2, and RpoT;3 genes in the WT, and RpoT;1 and RpoT;3 genes in rpoT;2 mutant (mt; isolated from line 833). Plants are 4 weeks old. R, root; YL, young expanding leaf; OL, old leaf.

(c) RpoT;1, RpoT;2, and RpoT;3 expression time course in 6-day dark-grown (open bar), additional 1-day (gray bar) and 6-day (dark bar) light-grown WT and mt plants.

PCR was performed three times, and averaged results are shown in the graph. Bars represent SD. Total RNA was reverse transcribed and subjected to semiquantitative RT-PCR analysis using 18S rRNA as internal control.

It is possible that RpoT;2 is functionally redundant with RpoT;1 and RpoT;3 in mitochondria and chloroplasts, respectively. Hence, knowledge of the tissue-specific expression of all the three RNAPs is necessary to obtain a better understanding of the RpoT;2 function. The result of semiquantitative RT-PCR of RpoT;1 and RpoT;3 transcripts using total RNAs from roots, young leaves and old leaves from 4-week-old Arabidopsis revealed that in contrast to RpoT;2, RpoT;1, and RpoT;3 transcripts showed distinct tissue-specific accumulation patterns (Figure 3b, top and bottom). The transcript level of RpoT;1 was significantly higher in the roots compared with the leaves while that of the RpoT;3 was higher in the leaves compared to the roots in the wild type. Interestingly, in older leaves, the transcript accumulation of RpoT;1 and RpoT;3 in the mutant was higher compared to that in the wild type, suggesting the possible existence of a transcriptional regulation mechanism to compensate for the loss of RpoT;2 function in the mutant. In contrast, the levels of RpoT;1 and RpoT;3 transcripts were not significantly elevated in young leaves or roots of the rpoT;2 mutant. In fact, a decrease in the levels of RpoT;1 to just about 40% of wild type levels was observed in the mutant roots.

The alteration of transcript levels of RpoT genes following the exposure of dark-grown wild-type and rpoT;2 mutant seedlings to light was examined (Figure 3c). While the transcript level of RpoT;1 increased continuously upon transfer of the dark-grown seedlings to light, RpoT;3 transcripts accumulated after 1 day of illumination and then remained constant. In contrast, the transcript level of RpoT;2 did not show any significant change throughout the experiment. Also, RpoT;1 and RpoT;3 transcript levels were not altered by the lack of functional RpoT;2 gene during the early stage of seedling development (Figure 3c).

Plastid and mitochondrial gene transcription in the rpoT;2 mutant

RpoT;2 is a nuclear-encoded organellar RNAP, and its polymerase activity has been demonstrated by in vitro UTP incorporation (Hedtke et al., 2000). However, the requirement of the RpoT;2 protein in vivo for mitochondrial or plastid transcription is not clear especially in view of the presence of two additional NEPs in mitochondria and plastids. To investigate whether the steady-state transcript levels of organellar genes are affected in the rpoT;2 mutant, we performed a dot-blot analysis of 28 plastid genes and 11 mitochondrial genes. As the phenotypic difference between mutant and wild type was most significant during the early stage of greening in the seedlings, we compared the accumulation levels of organellar gene transcripts between mutant and the wild type grown in dark for 6 days or upon transfer to light for 1 and 6 days (Figure 4).

Figure 4.

Macroarray analysis of organellar gene expression. (a) Plastid genes. (b) Mitochondria genes.

Total RNAs were reverse transcribed in the presence of [α-32P] dATP, and the resulting cDNAs were hybridized to a dot-blot membrane containing three replicas of each gene. The obtained data are normalized against a spiking control and shown in the graphs. Open bar, 6-day dark-grown; gray bar, 6-day dark, 1-day light; dark bar, 6-day dark, 6-day light-grown plants. WT, wild type; mt, rpoT;2 mutant.

According to Hajdukiewicz et al. (1997), plastid genes in tobacco are categorized into three classes: PEP, NEP + PEP, and NEP based on whether plastid- (PEP) or plastid- and nuclear- (PEP + NEP) or nuclear- (NEP) encoded polymerases mainly transcribe these genes (classes I, II, and III, respectively, by Hajdukiewicz et al., 1997). Evidence of whether these classes also hold true for Arabidopsis is not yet available although it is unlikely to be significantly different from what has been reported in other systems, and therefore the same classification is used here. In the wild type, there was a rapid increase of 2–16-fold in all the plastid gene transcripts examined following illumination of dark-grown seedlings for 1 day (Figure 4a). Between day 1 and day 6, in contrast, little enhancement of transcript levels for PEP and NEP genes was observed in light and even a reduction was observed in case of transcript levels for some of the PEP + NEP genes (Figure 4a). This result is consistent with those obtained with Hordeum vulgare (Baumgartner et al., 1993; Klein and Mullet, 1990), Vigna aconitifolia (Kelkar et al., 1993), and Pisum sativum (DuBell and Mullet, 1995) even though a direct comparison is difficult because of different experimental conditions used. In the rpoT;2 mutant seedlings, the increment of transcript accumulation after 1 day of illumination was considerably lower than in the wild type for all of the tested genes. For some of the PEP genes such as psb (photosystem II subunit genes)A, psbB, and psbC, as well as for genes such as rrn23, there was actually a reduction in the transcript levels following illumination. While the levels of most PEP transcripts in the mutant were not significantly different from those of the wild type after 6-day illumination, the levels of NEP and NEP + PEP genes were much lower in the rpoT;2 mutant compared to the wild type. Interestingly, transcripts of NEP genes in mutant plants seemed to accumulate at a higher level than in the wild type at 0-day light (dark-grown), but much lower at 6-day light as the transcript level was not altered by illumination (Figure 4a). Transcript levels of eight other genes involved in transcription and translation in the plastids and whose promoters have not been investigated to date were also examined (Figure 4a, undefined). In the wild type, rpoA, rps4, trn (tRNA for valine)V, and ycf (hypothetical chloroplast frame)3 showed constant increases in transcript accumulation similar to NEP and PEP genes, whereas rps2, trnDYE, trnG, and trnL displayed a time course similar to that of the PEP + NEP genes. The effect of the mutation on these genes was generally the same as that of NEP and NEP + PEP genes.

In the case of the mitochondrial genes, only 6-day-dark and additional 6-day-light samples were taken (Figure 4b). In contrast to that observed for plastid genes, there was little effect of light on the examined mitochondrial transcript levels following transfer of dark-grown seedlings to light in the wild type. In the case of the mutant, the only significant difference was observed in the atp1 transcript level, which decreased in the mutant upon transfer to light, to about half of the corresponding level in the wild type. For the other genes investigated, there was no significant difference in transcript accumulation between the mutant and the wild type.

Effect of rpoT;2 mutation on plastid and mitochondrial protein levels and PSII activity

The effect of altered transcript levels was investigated in the rpoT;2 mutant on the accumulation of mitochondrial and plastid proteins (Figure 5). For several components of the PSII and PSI complex such as D1, PsaD (photosystem I D subunit) and PsbO (photosystem II O subunit), a clear delay in accumulation was observed in the mutant. After 1 day of illumination, the levels of the four polypeptides were drastically reduced in mutant seedlings. Even after 6-day illumination, the amount of these proteins was lower in the mutant seedlings compared to the wild type, especially for the D1 polypeptide. It is important to note that a similar delay was observed in case of both nuclear- (PsbO and PsaD) and plastid-encoded (D1) proteins. In the same set of experiments, we also investigated the level of CF1, a plastid-encoded component of the thylakoid ATPase complex. The level of this protein was almost identical in the mutant and the wild-type seedlings at all the selected times were in agreement with that observed for its transcript levels.

Figure 5.

Immunoblot analysis of total cell proteins from wild type (WT) and rpoT;2 mutant (833).

Total protein was isolated from the identical samples used for the dot-blot and probed with antibodies specific for the indicated proteins. The lanes were loaded with 10 µg of protein. 0 d, 6-day dark-grown; 1 d, 6-day dark-, 1-day light-grown; 6 d, 6-day dark-, 6-day light-grown.

In contrast, the polypeptide level of the mitochondrial bc1 complex was not altered in the mutant. The protein was detected in all the samples, including those of dark-grown seedlings, and at a similar level in both mutant and the wild type. These results indicate that the rpoT;2 mutation may differentially affect the accumulation of chloroplast and mitochondrial proteins during the early stage of seedling development, leading to a dramatic reduction in the accumulation of chloroplast proteins.

To investigate the effect of the mutation of RpoT;2 on photosynthesis, the activity of PSII was compared between wild-type and rpoT;2 mutant seedlings. Chlorophyll a fluorescence measurements at room temperature in the mutant revealed alterations in the characteristics of PSII (Table 2). The variable fluorescence yield/maximum fluorescence yield ratio (FV/Fm) of the dark-adapted plants, a parameter that is regarded as reflecting changes in PSII photochemical efficiency (Maxwell and Johnson, 2000), was reduced in 3-week-old mutant plants compared to wild-type plants of the same age. This alteration was no longer observed in 4- and 5-week-old mutant plants, which showed FV/Fm values similar to those measured in wild-type plants.

Table 2. FV/Fm values for leaves from different aged mutant and wild-type plants
Lines3 weeks4 weeks5 weeks
  • The photosynthetic parameter was measured in 10 different leaves obtained from three different experiments (n = 15 ±SD).

  • *

    P < 0.01.

Wild type0.84 ± 0.0050.82 ± 0.0070.83 ± 0.006
rpoT;2 mutant0.79 ± 0.01*0.81 ± 0.0090.83 ± 0.005


We have investigated the function and physiological significance of RpoT;2, a phage-type RNAP targeted to both plastids and mitochondria, using a T-DNA-tagged Arabidopsis mutant. In the mutant, the T-DNA insertion was found to be present in the last intron of the RpoT;2 gene. The insertion resulted in the truncation of the RpoT;2 transcript, and this truncated transcript was found to accumulate at much higher levels in the rpoT;2 mutant (Figure 2c), probably because of the presence of the 35S enhancer element in the inserted T-DNA. Such enhancement of transcript levels has been reported to occur even when the enhancer element is inserted at the 3′ end of the gene (Weigel et al., 2000). However, the mutant phenotype is not due to a dominant negative effect resulting from the expression of the mutant RpoT;2 gene because our data indicate that the mutation in the line 833 is recessive. Furthermore, it is highly unlikely that a functionally active RpoT;2 is produced in the 833 mutant as the C-terminal part of the protein, which is essential for the function of this type of RNAPs (Patra et al., 1992), is modified because of the insertion of the T-DNA. Moreover, the same phenotype is also observed with another mutant allele 286E07 obtained from the GABI-Kat collection, which is likely to be a null mutant. These data taken together indicate that the rpoT;2 mutant investigated here is a null mutant.

RpoT;2 homologs showing dual targeting have been identified in Nicotiana (Hedtke et al., 2002; Kobayashi et al., 2001) and in the moss P. patens (Richter et al., 2002). Conservation of this dual targeted RNAP from mosses to higher plants would indicate an indispensable role for this RNAP in both mitochondria and chloroplasts. Our studies demonstrate that RpoT;2 is indeed very important for early plant development. However, at least under the conditions investigated here, RpoT;2 is not essential for survival of the plant.

Comparison of transcript levels of 11 mitochondrial genes between the wild type and the mutant indicate that with the exception of atp1 in light, the levels of none of the other investigated transcripts are affected in the rpoT;2 mutant at any of the stages studied (Figure 4b). Thus, it is likely that the loss of RpoT;2 could be compensated for by RpoT;1 in the mitochondria. In contrast, a clear effect of rpoT;2 mutation was detected in plastid gene transcription (Figure 4a). While there is clear distinction between the roles of NEP (RpoT;2 and RpoT;3) and PEP (eubacterial RNAP) in plastids based on differential promoter recognition properties (Maliga, 1998), the distinction between the relative roles of RpoT;2 and RpoT;3 in plastid gene expression is not clear. Our results suggest that one major effect of the rpoT;2 mutation is the lack of induction of several plastid genes in dark-grown mutant seedlings immediately upon illumination (Figure 4a), the 2–16-fold increase in transcript levels of several plastid genes in the wild type being not observed in the mutant. This stage is characterized by a rapid increase in transcription of most of the plastid genes in response to greening and probably requires the activity of all the RNAPs. However, it is important to note that the RpoT;3 transcription is not affected significantly in the rpoT;2 mutant at this stage of development (Figure 4b). This observation strongly suggests that the reduction of the plastid transcripts in the mutant is a consequence of lack of functional RpoT;2. Plastid transcription machinery has an ability to alternate transcription start in case of a deficiency in one of the RNAPs as observed in PEP-deficient tobacco (Legen et al., 2002), but deficiency in RpoT;2 was not entirely complemented for in our experimental conditions. These results indicate that during the early stage of seedling development and during the de-etiolation process, RpoT;2 plays a key role compared to RpoT;3 in transcribing plastid genes. However, an aberrant post-transcriptional regulation of RpoT;3 in the rpoT;2 mutant that results in reduced transcript levels for some of the plastid genes cannot be ruled out.

Surprisingly, in addition to the lack of the light-induced increase in plastid transcripts, the levels of many of the PEP gene transcripts such as psbA, psbB, and psbC decreased 1 day after transfer to light in the rpoT;2 mutant (Figure 4a) although the transcript levels of rpoA and rpoB (components of PEP involved in the transcription of psbA, psbB, and psbC) after 1-day light were similar for the mutant and the wild type. Possible explanations for the reduction in the transcript levels of PEP-dependent genes in the mutant are an altered stability of the transcripts and/or a reduced transcription. The reduction of the PEP gene transcript levels coincides well with the reduction of the transcript level of rrn23. It is known that the stability of plastid gene transcripts is affected by the ribosome association efficiency (Mayfield et al., 1995). The reduced level of ribosomal RNA (rrn23) in the rpoT;2 mutant may result in reduced polysome formation, which could cause a reduction in the stability of some PEP gene transcripts. At the same time, a reduction in the translation of RpoA and RpoB transcripts may occur because of the reduced levels of trnA, trnI, and trnL in the mutant at 1-day illumination. Yet another explanation for this phenomenon could be an indirect effect resulting in altered promoter preference or lower levels of the active form of PEP. Promoter selectivity of PEP by alternative sigma factors (Allison, 2000; Mullet, 1993) has been observed for AtSigE (sigE/sig5 of Arabidopsis), which specifically recognize psbD promoter (Shiina, unpublished data). Support for this suggestion also comes from the observation in mustard that PEPs exist in heterogeneous forms whose polypeptide composition depends on the light conditions (Pfannschmidt et al., 2000).

Upon transfer of 6-day dark-grown seedlings to light for 6 days, NEP-transcribed genes and some genes involved in translation such as rps and rrn genes, trnDYE, trnG, and trnI showed a 40–50% decreased transcript level in the mutant (Figure 4a). This reduction is likely to affect the protein synthesis in the plastids of the mutant, leading to reduced levels of polypeptides (Figure 5) and PSII activity (Table 2). On the other hand, most of the transcripts accumulated at 1.2–3-fold higher levels in the mutant compared with the wild type at 0-day light. The elevation was more significant for the NEP-dependent genes, where 7 out of 10 transcripts accumulated more than twofold compared to the wild type. It is known that transcript levels of NEP-dependent genes are increased in Δrpo mutants of tobacco (Allison et al., 1996; Hajdukiewicz et al., 1997; Krause et al., 2000; Legen et al., 2002), as well as the sig (sigma factor)2 mutant of Arabidopsis (Kanamaru et al., 2001). These results along with our observations imply that the NEP genes could be upregulated in response to disruption of any plastid RNAP either through increased transcriptional activity of other RNAPs or by stabilization of transcripts. The latter is more likely, as we could not detect any upregulation of RpoT;3 transcript accumulation in the rpoT;2 mutant seedlings by semiquantitative RT-PCR (Figure 3c).

It is important to note that the accumulation of nuclear-encoded proteins, PsaD and PsbO, were also delayed in the mutant (Figure 5). As the levels of PsbO and other PSII extrinsic proteins have been suggested to be independent of their functional association with an assembled PSII unit (Eisenberg-Domovich et al., 1995), PsbO polypeptides are unlikely to be degraded in the absence of properly assembled PSII complex in the rpoT;2 mutant. Alternatively, delayed plastid development may affect nuclear transcription. The developmental status of chloroplasts has been shown to control the expression of nuclear-encoded chloroplast genes via organelle–nuclear communication (Surpin et al., 2002). As chloroplast development is likely to be delayed in the rpoT;2 mutant, it is probable that this affects nuclear gene expression leading to a reduced levels of PsaD and PsbO in the mutant.

Despite the fact that the major role of RpoT;2 is immediately after chloroplast development starts, RpoT;2 transcripts can be detected at all stages in plant development and at nearly constant levels in different tissues in light (Figure 3). In this respect, the role of RpoT;2 needs to be studied in relation to expression of the other two organelle-directed RpoTs. In 6-day-old seedlings, RpoT;1 and RpoT;3 transcription is induced by 1 and 6 days of illumination, which is not the case for RpoT;2 (Figure 3c). This implies that RpoT;2 may be the key RNAP transcribing organellar genes in the early stage of seedling growth, and the other RNAPs are necessary at a later stage of development. Furthermore, as shown in Figure 3(b), in mature leaves in the mutant, both RpoT;1 and RpoT;3 transcripts accumulate to more than twofold of wild-type levels. It would thus appear as if the mutant plants compensate for the lack of functional RpoT;2, by an organelle to nucleus signaling that could increase the transcript levels of both RpoT;1 and RpoT;3 in older mutant leaves. This enhancement may then be sufficient to sustain normal organellar activities, thus explaining the absence of mutant phenotype during the later stages of plant development. Young leaves, in contrast, represent a stage where the compensatory increase in levels of RpoT;1 and RpoT;3, may not have been attained. Hence, the steady-state levels of these two RNAPs, in absence of RpoT;2, are probably not sufficient to compensate for the organellar transcription in young leaves, resulting in a visible phenotype at this stage. In the mutant roots, the compensation for the lack of RpoT;2 by an increase in the expression of the other RNAPs does not appear to occur. In fact, RpoT;1 transcript levels are reduced to 40% of their wild-type levels. Thus, the reduction in the root length in the rpoT;2 mutant may not simply be because of the lack of RpoT;2 but also because of a reduction of RpoT;1 levels.

Alternatively, there could be a hitherto uncharacterized post-transcriptional regulatory mechanism that modulates the levels and/or the activity of the RpoT;2 protein through its interaction with other factors in a developmental or tissue-specific manner. Bligny et al. (2000) have isolated a NEP in spinach that interacts differentially with CDF (chloroplast DNA-binding factor)2 transcription factor depending on growth conditions. Such a differential interaction would lead to a modification of RpoT;2 activity depending on the developmental stages of plants. Another possibility is that the alternative translation from one of the two initiation codons at the 5′ end of RpoT;2 may differentially regulate its localization (Kobayashi et al., 2001). Similar case has been reported recently for AtSig5, whose truncated form from the second methionine is targeted to mitochondria, and the translation is likely to be regulated in tissue-specific manner (Yao et al., 2003).

In conclusion, the changes in early developmental stages of plastid development in the mutant support the model of a control of plastid transcription, which suggests an early role for NEP followed by PEP transcription. Our results presented here are indicative of the involvement of RpoT;2 in controlling the early phase of chloroplast development. Further studies on knock out mutants of RpoT;1 and RpoT;3 and their double mutants, as well as post-transcriptional regulatory factors of NEPs will be highly informative for elucidating the complex nature of organellar gene transcription.

Experimental procedures

Plant growth, screening conditions and mutant lines

Plants were routinely grown in a growth chamber at 22°C under long-day conditions (16 h light, 8 h darkness) in soil or on plates containing 0.5× MS (Murashige and Skoog, 1962), 1% agar, and 0.5% sucrose except where otherwise stated. Approximately 50 000 T-DNA-tagged lines of A. thaliana, Columbia ecotype (Weigel et al., 2000) were sown on 0.5× MS medium and grown vertically for 10 days. Seedlings with abnormal root development were selected and further characterized in the next generation. For the greening experiment, seeds were sown on plates and were kept for 24 h under light to enhance germination, then wrapped in aluminum foil and kept for 6 days. After 6 days, plates were unwrapped and samples were taken after 0, 1, and 6 days of continuous light. The second rpoT;2 mutant allele (termed line 286E07) was obtained from the GABI-Kat collection (; Rosso et al., 2003) of T-DNA insertion lines. The T-DNA insertion in the RpoT;2 gene in the mutant line 286E07 was checked with PCR using the primer pair 5′-TCAAGGTAATGTGCATAAGTTGCT-3′ and 5′-CCCATTTGGACGTGAATGTAGACAC-3′.

Plasmid rescue and Southern blot analysis

Total DNA was extracted from mature plants using the DNeasy Plant Maxi Kit (Qiagen, Hilden, Germany) according to the manufacturer's instructions. Two micrograms of DNA was digested by EcoRI, re-ligated, and used for plasmid rescue as described by Weigel et al. (2000). The plasmids of the resulting clones were sequenced and compared with the GenBank, PDB, Swissprot, SPupdate, and PIR databases using the blast tool (Altschul et al., 1997). The same DNA, digested by EcoRI, was used for the Southern blot analysis.

Analysis of gene expression

For Northern blot analysis, mRNA was extracted from 3-week light-grown wild-type or mutant plants using Dynabeads Oligo (dT)25 (Dynal, Oslo, Norway) according to the manufacturer's protocol. For RNA probe preparation, a 480-bp fragment of the RpoT;2 coding region was PCR amplified using primers 5′-TGGTTATTGTTCTGGTTTAT-3′ and 5′-AACAATTCATCTACCTCAGG-3′, cloned into pGEM-T Easy vector (Promega, Madison, WI, USA), linearized, and used as a template for in vitro transcription (Ambion, Austin, TX, USA). For Semi-quantitative RT-PCR, total RNA was extracted using the RNeasy Mini Kit (Qiagen), and 2 µg of the RNA was transcribed into cDNA with the first strand cDNA synthesis kit (Amersham Pharmacia Biotech). PCR amplification was carried out with each cDNA and primer pairs 5′-GTTTGTTTCTAATCGAACGG-3′ and 5′-TTCTGCTTCACATTTCATGG-3′ for RpoT;1, 5′-TGGTTATTGTTCTGGTTTAT-3′ and 5′-AACAATTCATCTACCTCAGG-3′ for RpoT;2, and 5′-TTCATTGGAAGACCAATACC-3′ and 5′-TTTTCAAAGTCTGATCAACC-3′ for RpoT;3. Quantum 18S RNA internal standard kit (Ambion) served as internal control.

For generating the probes for dot-blot analysis, DNA fragments were amplified from total DNA of A. thaliana by means of PCR and purified using Qiaquick PCR Purification Kit (Qiagen). The list of genes blotted on the membrane is shown in Table 3. As a probe for a spiking control, a heat-shock promoter fragment of Glycine max (GmHSPm) was used. The DNA fragments were denatured in 0.2 m NaOH and then blotted on Hybond N+ membrane (Amersham Pharmacia Biotech) with the help of a vacuum manifold. Probes were blotted in triplicates with 100 ng DNA per spot. For target preparation, total RNA was extracted using TRI reagent (Sigma-Aldrich, Steinheim, Germany) from 0.1 g of Arabidopsis seedlings. Spike RNA was prepared by in vitro transcription with T7 RNAP and GmHSPm cloned in pBKS+, which had been linearized at the end of the insertion. All RNA preparations were treated with DNAse I (1 U for 10 µg RNA) and purified with the RNeasy Mini Kit (Qiagen). Two micrograms of total RNA was mixed with 70 pg of spike RNA and 0.4 pmol of each 3′ primer used for probe DNA amplification, and incubated in the presence of Superscript II reverse transcriptase (Invitrogen, Paisley, UK) and 50 µCi of [α-32P] dATP (>3000 Ci mmol−1). After purification using the NICK column (Amersham Pharmacia Biotech), 1 000 000 cpm of each labeled cDNA was used for hybridization to the dot-blot membrane. The hybridized membrane was exposed to a GS-525 Molecular Imager (Bio-Rad, Hercules, CA, USA), and the signal strength was quantified with QuantArray (PerkinElmer Life Sciences, Zaventem, Belgium). The data were normalized to the background and the spike control, averaged among the replicas, and compared.

Table 3.  List of genes blotted on the membrane
  1. su, subunit; pt, plastid; mc, mitochondria.

accDAcetyl-CoA carboxylase beta supt
atpAATP synthase alpha supt
clpPATP-dependent protease proteolytic supt
ndhGNADH dehydrogenase 49 kDa supt
petACytochrome fpt
psaAPSI P700 apoprotein A1pt
psbAPSII 32 kDa (D1) proteinpt
psbBPSII 47 kDa protein (CP47)pt
psbCPSII 44 kDa protein (CP43)pt
psbDPSII D2 proteinpt
psbEFLJPSII subunitspt
rbcLRuBisCO large supt
rpl33Ribosomal protein L33pt
rpoARNAP alpha supt
rpoBRNAP beta supt
rps2Ribosomal protein S2pt
rps4Ribosomal protein S4pt
rrn1616S rRNApt
rrn2323S rRNApt
rrn55S rRNApt
trnDYEtRNAAsp(GUC), tRNATyr(GUA), tRNAGlu(UUC)pt
ycf3orf (PSI accumulation)pt
atp1ATP synthase su 1mc
atp6-1ATP synthase su 6mc
ccb203Cytochrome c biogenesis orf 203mc
ccb206Cytochrome c biogenesis orf 206mc
cobApocytochrome Bmc
cox1Cytochrome c oxidase su 1mc
cox3Cytochrome c oxidase su 3mc
rpl2Ribosomal protein L2mc
rps3Ribosomal protein S3mc
Orf153b mc

Complementation analysis of rpoT;2 mutant

A full-length cDNA of RpoT;2 was amplified from Arabidopsis cDNA using the primer pair 5′-GTTGGATCCATGTCCAGTGCTCAAACC-3′ and 5′-CTTTCTAGATCAGTTGAAGAAATAAGG-3′. The resulting DNA fragment was cloned into the BamHI–SalI site of the p11/2 binary vector (a kind gift from B. Reiss, MPI, Cologne, Germany) under the control of 35S promoter, then introduced into A. tumefaciens strain GV3101 (pMP90RK). Homozygous rpoT;2 mutant plants (line 833) were transformed by dipping according to Clough and Bent (1998). The transformants were selected for hygromycin resistance. The presence of 35S::RpoT;2 in the genome, as well as the homozygous mutant background of the T2 plants was confirmed by PCR.

Chlorophyll and PSII activity measurements

Plants were subjected to 1 h of dark adaptation prior to the measurement and either the whole plants or 1.5-cm leaf disks were used. Measurements of the variable fluorescence yield/maximum fluorescence yield ratio (FV/Fm) from leaf tissue were performed using a modulated fluorometer (PAM Fluorometer; Walz, Effeltrich, Germany) with the PAM 103 accessory and two Schott lamps (model KL 1500; Schott, Mainz, Germany) providing saturating flashes and actinic illumination. Chlorophyll content was measured by direct dimethylformamide extraction according to Porra et al. (1989).

Protein extraction and Western blot analysis

Proteins were isolated using TRI reagent (Sigma-Aldrich). The final precipitates in ethanol were re-suspended in 2× SDS–PAGE sample buffer (65 mm Tris–HCl, pH 6.8, 10% (v/v) glycerol, 4% (w/v) SDS, 0.005% (w/v) bromophenol blue, and 5% (v/v) β-mercaptoethanol), heated at 90°C for 5 min, centrifuged, and the supernatant was used for SDS–PAGE.

Proteins were separated on 12% SDS–polyacrylamide gels (Laemmli, 1970) and blotted onto a 0.2-µm nitrocellulose membrane (Trans-Blot, Bio-Rad). Horseradish peroxidase-labeled secondary antibodies and enhanced chemiluminescence (Amersham Pharmacia Biotech) were used to detect the antibody–antigen conjugate.


We thank Ms Gunilla Malmberg and Ms Karin Degerman for technical assistance. We thank Dr Catherine Bellini for helping with the genetic analysis, Dr Åsa Strand and Dr Alan Marchant for carefully reading the manuscript. The 286E07 mutant line was generated in the context of the GABI-Kat program and provided by Dr Bernd Weisshaar (MPI for Plant Breeding Research; Cologne, Germany). This work was funded by Vetenskapsrådet, EU grants and Kempe foundation to R.P.B. A.E.-R. is a recipient of a postdoctoral fellowship from Spanish Ministerio de Educación, Culturay Deportes.