Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

I. Introduction

All plants and algae possess plastids that contain their own genome and a transcription-translation machinery distinct from that of the nucleus-cytoplasm. Plastids evolved from an ancestral cyanobacterium engulfed by a host cell, and therefore the transcription-translation machinery of plastids has a number of prokaryote-like features (Sugiura 1992). For example, plastid ribosomes are 70S in size, plastid rRNAs and tRNAs are very similar in sequence to their Escherichia coli and cyanobacterial counterparts, and plastid mRNAs contain triphosphates at their 5′ ends and lack long poly(A) tails. In addition, many plastid genes of higher plants are co-transcribed as polycistronic pre-RNAs, which are then extensively processed into shorter RNA species. Most components of the plastid genetic system are not encoded by the plastid DNA but are instead nuclear-encoded. Thereby, formation of the plastid genetic system requires the coordinated expression of nuclear and plastid genes (Woodson and Chrory 2008).

Plastids of multicellular plants usually ­differentiate into different types of plastids: undifferentiated proplastids (in meristematic tissues or embryos), etioplasts (in dark-grown seedlings), photosynthesis-performing chloroplasts (in leaves), starch-storing ­amyloplasts (in roots), carotenoid-accumulating colored chromoplasts (in flowers and fruits) and gerontoplasts (Biswal et al. 2003; Pyke 2007). In the photosynthesis-active chloroplasts, photosynthetic genes are actively transcribed and their transcripts accumulate to substantial levels. Although identical copies of the plastid genome are present in all plastid types, the levels and patterns of accumulation of plastid transcripts vary largely among the different plastid types and under different environmental conditions.

Transcription rates and steady-state RNA levels of plastid genes are generally not ­coincident, and many plastid genes are known to be constitutively transcribed, suggesting that post-transcriptional RNA processing of primary transcripts represents an important step in the control of plastid gene expression (Gruissem et al. 1988; Mullet 1988). Post-transcriptional RNA processing steps include endonucleolytic cleavage, 3′-end trimming, cis/trans-splicing and RNA editing (Sugita and Sugiura 1996; Monde et al. 2000; Herrin and Nickelsen 2004). Most regulatory factors required for plastid gene expression are encoded by nuclear genes. To understand the molecular basis of plastid gene expression, many studies have focused on the posttranscriptional regulation in plastids and identified nuclear-encoded regulatory proteins (Nakamura et al. 2004; Schmitz-Linneweber and Small 2008; Falcon de Longevialle et al. 2010). In addition to this research trend, the importance of transcriptional regulation in plastids has been reconsidered and many outstanding findings on transcription have accumulated since the existence of nuclear encoded RNA polymerases were proposed (Falk et al. 1993; Hess et al. 1993).

The first part of this chapter describes the overall structure and gene contents of plastid genomes in higher plants. The second part summarizes the characteristics of plastid gene expression in different types of plastids. Finally, we review our current knowledge of transcriptional regulation in plastid gene expression of higher plants. Transcriptional and posttranscriptional regulations of gene expression in plastids have been discussed in recently published review articles (Maier et al. 2008; del Campo 2009; Tillich et al. 2010; Stern et al. 2010; Lerbs-Mache 2011).

II. Plastid Genome

A. Genetic Information of Plastid DNA in Higher Plants

1. Overall Structure and Gene Content of the Plastid Genome

Plastid DNA (ptDNA) was first extracted by CsCl density gradient centrifugation from the isolated chloroplasts of the green alga Chlamydomonas reinhardtii (Sager and Ishida 1963). The circular molecules of ptDNA were observed in Euglena gracilis (Manning et al. 1971) and then also in higher plants (Herrmann et al. 1975). The first entire nucleotide sequences of plastid genomes were determined from tobacco (Shinozaki et al. 1986) and liverwort (Ohyama et al. 1986). In the past 25 years, 200 plastid genomes were fully sequenced at the end of 2010 (the NCBI organelle genome resources database (http://www.ncbi.nlm.nih.gov/genomes/GenomesGroup.cgi?taxid=2759&opt=plastid)). Extensive genomic changes have been revealed in the plastid genomes, such as a 30-kb or short inversions, gene losses and/or gains, which are common events throughout evolution of the plastid genome. Recent plastid genome ­phylogenomics studies have provided additional evidence for deep-level phylogenetic relationships as well as increased phylogenetic resolution at low-taxonomic levels (Bock 2007; Ravi et al. 2008; Gao et al. 2010).

In land plants, most plastid genomes display a tetrapartite genome organization with a large single copy region (LSC) and a small copy region (SSC) separating two inverted repeat sequences (IRA and IRB). The two IR sequences are identical in their nucleotide sequence containing the ribosomal DNA gene cluster (rrn16-rrn23-rrn4.5-rrn5), several tRNA and protein-coding genes (Fig. 10.1). Pea (Pisum sativum) and black pine (Pinus thunbergii) plastid genomes lack such long IR sequences. Overall, the GC content of ptDNA is 36–42% in vascular plants and 28–37% in non-vascular plants (bryophytes). The plastid genomes of land plants contain 120–160 genes (Sugiura 1992; Bock 2007). In the case of A. thaliana plastids (Sato et al. 1999), the genome contains 79 distinct protein-coding genes including the conserved hypothetical open reading frames (ycf), and 34 RNA-coding genes. Among them 6 protein genes and 11 RNA genes are present in the IR sequences. Plastid-encoded proteins include 35 ­photosynthesis-related, 11 NAD(P)H oxidoreductase subunits, 21 ribosomal proteins, 4 RNA polymerase subunits, and 8 other proteins including AccD, ClpP, and Ycfs. RNA-coding genes include 4 rRNA genes and 30 tRNA genes. In addition to the structural RNA genes, unidentified small RNA-coding genes can be present in the plastid genome. Eighteen genes for 6 tRNAs and 12 genes encoding proteins contain introns as shown in Fig. 10.1. Genes and their products described in this chapter are listed in Table 10.1. The rps12 gene for ribosomal protein S12 is divided into 5′-rps12 (in LSC) and 3′-rps12 (in IRs), and each gene segment can be transcribed independently and the transcripts trans-spliced. Several small open reading frames (ORFs) also are present in the plastid genome. The lack of evolutionary conservation, even among closely related species, is interpreted as evidence for these ORFs that have no functional significance.

Fig. 10.1.
figure 1

Physical map of the Physcomitrella patens plastid genome. The large inverted repeat sequences (IRA and IRB) are separated by large single copy (LSC) and small single copy (SSC) regions. Genes inside the circle are transcribed clockwise and genes outside the circle are transcribed counterclockwise. Genes with related functions are shown in the same color. Asterisks indicate intron-containing genes; introns are depicted as open boxes. For gene products and their functions, compare with Table 10.1. (This figure was modified from Sugiura et al. (2003).)

Full size image
Table 10.1. Plastid-encoded genes described in this chapter
Full size table

As shown in Fig. 10.1, most plastid genes of land plants are organized in clusters and are co-transcribed as polycistronic pre-RNAs, which are then extensively processed into shorter RNA species (Sugita and Sugiura 1996). Transcription rates and steady-state RNA levels of plastid genes are generally not coincident as described below, and all plastid genes are transcribed by either one or two distinct plastid RNA polymerases, depending on the different stages of plastid development or fluctuation of environmental conditions.

2. Non-coding RNAs in Plastids

In bacteria, regulatory non-coding RNAs (ncRNAs) are widespread. For example, to date, over 80 regulatory ncRNAs have been identified in E. coli (Gottesman 2005) and several ncRNAs in cyanobacteria (Axmann et al. 2005; Dühring et al. 2006; Nakamura et al. 2007; Ionescu et al. 2010). In plastids, tscA RNA (430 nt) was first identified as ncRNA required for trans-splicing of exons 1 and 2 of psaA pre-mRNAs in the green alga C. reinhardtii (Goldschmitz-Clermont et al. 1991). Thereafter, SpRNA (218 nt) was found as an ncRNA in tobacco plastids (Vera and Sugiura 1994). Disruption of the SpRNA gene (sprA) revealed that this RNA is dispensable for cell viability under normal growth conditions and its function is still unknown (Sugita et al. 1997). Furthermore, several small ncRNAs were identified in tobacco and Arabidopsis chloroplasts.

At least 7 small non-messenger RNA (snmRNA) ranging from 50 to 290 nt are encoded by the Arabidopsis ptDNA (Marker et al. 2002). Among these, five snmRNAs are located to the intergenic regions, one is located to the intron of the petD gene, and the remaining one is located to the 3′ end of the rbcL gene. All are encoded by the same strand of their neighboring genes. Likewise, Lung et al. (2006) identified 12 ncRNAs (20–500 nt) from tobacco chloroplast cDNA libraries. Most plastid ncRNAs are located to intergenic regions and one ncRNA is encoded by the antisense strand of neighbouring genes. Another ncRNA, Ntc-5 RNA, is transcribed in opposite (antisense) orientation to the 3′-untranslated region (UTR) of the atpE gene (located on the complementary strand). This might function in analogy to an antisense RNA or miRNA.

Recently, a long antisense RNA (asRNA, 650 nt) was identified in the chloroplasts in Arabidopsis, tobacco and poplar (Georg et al. 2010). asRNA maps in an antisense orientation to ndhB and is predominantly accumulated in young leaves and at physiological growth temperatures. Interestingly, the correlation between the accumulation of asRNA and RNA editing of the ndhB transcript appeared weak in a temperature shift experiment, suggesting involvement of this RNA in RNA maturation or in the control of RNA stability. Posttranscriptional regulation via ncRNAs would provide a fast and efficient mechanism for the rapid adjustment of organellar gene expression to changing environmental conditions and/or metabolic demands of the cell. Further studies need to address their function in plastid biogenesis.

B. Conformation and Copy Number of Plastid DNA in Plastids

1. Conformation of ptDNA and Plastid Nucleoids

Plastid DNA (ptDNA) is usually depicted as a circular molecule. However, conformation of ptDNA seems to be rather complex because some fractions of ptDNAs were observed as circular dimer or multimers (Herrmann et al. 1975; Deng et al. 1989) as well as in linear molecules (Bendich 1991; Lilly et al. 2001). Herrmann et al. (1975), observed using an electron microscopy, that a large fraction (80%) of the ptDNA extracted from isolated chloroplasts from young leaves of four plant species (Antirrhinum majus, spinach, Oenothera hookeri, and Beta vulgaris) was circular including supertwisted circles and circular dimers. Thereafter, in-gel procedures by pulsed-field gel electrophoresis revealed that at least four distinct forms of monomers to tetramers exist in isolated spinach leaf chloroplasts (Deng et al. 1989). In contrast, 25% (in pea) to 45% (in tobacco) of the ptDNA within developing leaf tissue was observed as circular forms and 22–35% are linear forms using DNA fiber-based fluorescence in situ hybridization (fiber-FISH) (Lilly et al. 2001). Linear molecules are presumably produced by breakage of the circles during extraction of ptDNA or are intermediates of replicating DNA molecules. Even now, the functional relevance of most of the different conformations of the ptDNA is still unclear. Nevertheless, we cannot exclude the possibility that conformational changes of the ptDNA are involved in the regulation of plastid gene expression. Of consequence, fluctuations of ptDNA topology are tightly correlated with changes in transcriptional activity of C. reinhardtii plastid psaB, rbcL, atpA and atpB genes (Salvador et al. 1998).

ptDNAs are organized into so-called plastid nuclei or nucleoids that are composed of DNA, proteins and RNAs (Kuroiwa 1991; see also Liere and Börner, Chap. 11). Number, shape, and size of the nucleoids vary depending on the plant species and the stage of plastid development. Likewise, ptDNA copy number per nucleoid is variable between plant species and independent of plastid differentiation (Kuroiwa 1991). Proplastids often contain only one small nucleoid whereas mature chloroplasts contain several or even dozens of nucleoids. The plastid nucleoids are localized in the inner membrane in young leaves but are localized in the matrix between thylakoid membrane and/or grana stacks in mature leaves (Kuroiwa 1991). Various enzymatic activities for DNA synthesis, transcription, DNA condensing, and Rubisco degradation are retained in the plastid nucleoids (Sakai 2001; Sekine et al. 2002; Kato et al. 2004). Proteomic analysis of a Triton X-100-insoluble 30,000  ×  g pellet (equivalent to plastid nuclei) from purified pea chloroplasts identified over 30 proteins including DNA gyrases and proteins required for transcription and translation (Phinney and Thelen 2005). Likewise, a proteomic analysis of Arabidopsis and mustard plastid transcriptionally active chromosome (pTAC) identified 35 components involved in transcription, DNA replication, DNA topology, DNA binding, detoxification, or protein modification (Pfalz et al. 2006). These analyses will provide a clue in the understanding of how ptDNAs are organized in a higher ordered configuration and how plastid nucleoids contribute to the regulation of plastid gene expression in plastid development.

2. Copy Number of ptDNA

A single leaf mesophyll cell contains 50–100 chloroplasts, each of which harbors 100 or more ptDNA copies (Bendich 1987). Accordingly, the ploidy of leaf cells reaches 10,000 ptDNA. The copy number might change during plastid differentiation or leaf development. The multiplicity of ptDNA can be explained by genome partition during plastid division, plastid partition during cell division, and increased gene dosage (Bendich 1987).

DNA reassociation kinetics measurements showed that the percentage of pea ptDNA in total cellular DNA ranges from 0.4% in roots, 1.4% in embryos and etiolated tissues to 12% in fully greened leaves (Lamppa and Bendich 1979). This accounts for 244 copies per plastid in young leaves and 174 in fully greened leaves. An early study using the DNA-specific 4′,6-diamidino-2-phenylindole (DAPI) fluoro­chrome showed that the level of ptDNA increases 26-fold (500–13,000 copies per cell) from the basal to mature regions of spinach leaves (Lawrence and Possingham 1985). Quantification by Southern dot blot hybridization indicated that plastids isolated from the basal primary barley leaves of 4-days-old seedlings contain 130 copies of ptDNA in plastids and that the DNA copy number increased to 210 in plastids of the developing leaf, declining gradually with increasing cell age, reaching 50 copies per plastid in the oldest cells of leaves (Baumgartner et al. 1989). In contrast, Southern blot analysis of the amount of ptDNA in Arabidopsis and tobacco plants shows that ptDNA copy number remains remarkably constant during leaf development and even in the senescent leaf (Li et al. 2006). This indicates that during leaf development, plastid gene expression in higher plants is not significantly regulated at the level of genome copy number. A similar result was reported by Zoschke et al. (2007). They showed that ptDNA copy numbers varied from only 1,000 to 1,700 per cell during development from young to old rosette leaves of Arabidopsis plants. In contrast, the transcriptional activity and steady-state transcript levels of plastid genes were significantly reduced in older rosette leaves (Zoschke et al. 2007). Thus, once plastid differentiation is complete, ptDNA copy number remains constant and does not vary significantly with leaf age or the plant’s developmental stage. This topic is addressed in detail by Liere and Börner (Chap. 11).

III. Plastid Gene Expression in Different Types of Plastids

A. Dynamics of Gene Expression During Plastid Development

Proplastids, etioplasts, and amyloplasts retain the ability to develop into photosynthetic chloroplasts when exposed to light. Although identical copies of the plastid genome are present in all plastid types, ­plastid genes are differentially expressed depending on plastid types. In the photosynthetically active chloroplasts, photosynthesis genes are actively transcribed and their transcripts accumulate at higher levels. In contrast, photosynthesis genes are rarely or not transcribed while non-­photosynthesis genes are constantly or actively transcribed in non-photosynthetic plastids.

1. Plastid Gene Expression During Amyloplast Formation

Root cells contain starch-storing amyloplasts. The steady-state transcript levels of photosynthetic genes psbA, psaA and rbcL are 200–900-fold lower in roots than in leaves, whereas they are constitutively transcribed at rates similar to those in root amyloplasts and leaf chloroplasts (Deng and Gruissem 1988). When cultured tobacco BY-2 cells in the stationary phase were transferred to auxin-depleted but cytokinin-supplemented culture medium, leucoplast-like plastids were converted to amyloplasts within two days. The number of plastids and the relative amount of ptDNA per cell remained nearly constant throughout conversion of leucoplasts to amyloplasts while overall transcriptional activity of the isolated plastid-nuclei rapidly decreased (Sakai et al. 1999). This suggests that the transcriptional activity of plastid genes is downregulated during plastid conversion.

2. Plastid Gene Expression During Chloroplast Development

In monocots such as barley and wheat, leaves are composed of cells containing different types of plastids in a developmental gradient ranging from proplastids in the meristematic cells at the base of the leaf to either fully mature chloroplasts or etioplasts in the cells at the tip. Baumgartner et al. (1989) showed that transcriptional activity is low in pro­plastids near the base of 4-day-old dark-grown or illuminated seedlings and increases tenfold in the region just above the base of the leaf. By contrast, transcriptional activity rapidly declines in older apical cells during light dependent differentiation of etioplasts to chloroplasts. Krupinska and Apel (1989) showed that the relative transcription rates of the various plastid DNA fragments were almost identical in etioplasts and chloroplasts. This suggests that etioplasts are already well prepared to differentiate into photosynthetically active chloroplasts.

DNA copy number per plastid increased only 1.6-fold during early leaf development and remained constant with increasing cell age of seedlings. A similar result was observed in spinach chloroplasts (Deng and Gruissem 1987). Thus, the general concept being accepted is that the plastid transcriptional activity per plastid and DNA copy number increase early in chloroplast development and that transcriptional activity per DNA template varies during leaf biogenesis (Mullet 1993).

3. Plastid Gene Expression During Chromoplast Formation and Maturation

Chromoplasts are carotenoid-accumulating plastids conferring color to many flowers and fruits. Chromoplast differentiation proceeds from preexisting plastids, most often chloroplasts (Egea et al. 2010). During the differentiation process the plastid genome is essentially stable and transcriptional activity is limited (Kahlau and Bock 2008; Egea et al. 2010). Piechulla et al. (1985) first ­analyzed the transcript levels of several plastid photosynthesis-related genes during tomato fruit ripening. Their transcript levels relative to cytoplasmic rRNA dramatically decreased during fruit ripening. A similar observation was reported in chromoplasts from bell pepper (Gounaris and Price 1987). Thus, plastid gene expression can be generalized to be down-regulated in nonphotosynthetic cells and tissues. To further address this issue, Kahlau and Bock (2008) carried out a systematic transcriptomics and translatomics analysis of the tomato plastid genome during fruit development and ­chloroplast-to-chromoplast conversion. The data showed that photosynthesis genes are much more strongly downregulated than the genetic system genes, including matK and the genes encoding ribosomal proteins, the subunits of plastid-encoded plastid RNA polymerase (PEP). Interestingly, the transcript abundance of accD encoding a subunit of the acetyl-CoA carboxylase increased during chloroplast-to-chromoplast conversion (Kahlau and Bock 2008). Such a global decrease of transcript levels during tomato fruit development may be attributed to a dramatic down-regulation of transcription by PEP that directs the transcription of photosynthesis-related genes. RbcL and PsbD protein levels dramatically decreased while the AccD protein accumulated to substantial levels in the plastids during this process. AccD is involved in fatty acid biosynthesis required for fruit development to provide the membrane lipids that accommodate the storage carotenoids in ripening fruits. In addition, splicing and RNA editing of most gene transcripts investigated did not significantly change during ripening whereas splicing of ndhB transcripts and RNA editing of ndhB, ndhD, and ndhF transcripts was significantly reduced or absent from chromoplasts. Thus, expression of many plastid genes is negatively controlled at transcriptional and posttranscriptional levels in chromoplast formation and maturation. This study provides new evidence that NEP-dependent genes (accD, matK, clpP, etc.) are actively transcribed in chromoplasts.

B. Cell Type-Specific Expression of Plastid Genes in C4 Plant Leaves

In maize, a C4 plant, the leaves are mostly composed of two types of cells, bundle sheath (BS) and mesophyll (M) cells. M cells are responsible for the linear light reactions of photosynthesis and BS cells complete the Calvin cycle and perform cyclic Photosystem I (PS I) reactions (Kanai and Edwards 1999). M and BS cells contain anatomically and biochemically distinct chloroplasts, M and BS chloroplasts, respectively. Both chloroplast types have the same proplastid origin. M chloroplasts develop thick granal stacks and BS chloroplasts develop mostly parallel lamellae with diffuse grana and accumulated starch. Rubisco activity is detected in BS cells but not in M cells (Kanai and Edwards 1973). The absence of Rubisco from M cells is accounted for by the absence of rbcL mRNA (Link et al. 1978). Thereafter, many studies on the control of C4 photosynthesis gene expression have been described (Sheen 1999).

Cahoon et al. (2008) performed DNA microarray analysis to quantify the levels of 62 plastid transcripts in the yellow base region (etioplasts) and the green tip (mature BS and M chloroplasts) of maize leaves. They showed that 51 plastid genes are expressed >twofold higher in the leaf tip than the leaf base and 10 genes, mostly ribosomal protein genes, showed less than a twofold difference between the base and the tip. Only clpP was more than twofold higher in the base. However, this analysis could not reveal ­differences in the transcript levels of respective plastid genes between BS and M chloroplasts.

Sharpe et al. (2011) carried out qRT-PCR to compare the transcript abundance of 18 plastid genes between young and mature BS and M cells of maize leaves. The middle region contained young BS and M while the tip section had mature BS and M chloroplasts. Among the 18 transcripts, the following 10 changed abundance in significant cell type-specific patterns. Photosystem II (PS II) transcripts (psbA, C, D, and K) accumulated at 8–19-fold higher in M cells than BS cells in the middle section, but these differences were diminished in the tip section due to an increase of transcript abundance in BS. rbcL mRNA was 17-fold higher in BS in middle tissue and 240-fold higher in BS in mature tip tissue. Expression of PS I (psaA and B) and cytochrome b 6 f complex (petA) genes was enhanced in the BS of the leaf tip. Although NAD(P)H dehydrogenase complex mRNAs (ndhA and ndhC) had a similar expression pattern as rbcL, the abundance of ndhA and ndhC mRNA in BS increased only 2- and 28-fold, respectively as the leaf matured. Thus, changes in transcript abundance followed no single pattern, suggesting that transcription and/or transcript turnover for plastid genes is tissue-, cell-, and gene-specific. Further, quantitative analyses are needed to obtain general conclusions on this issue.

IV. Transcription by Two RNA Polymerases in Plastids

During the endosymbiotic evolution of mitochondria from proteobacteria and of plastids from cyanobacteria-like prokaryotes, the two organelles established a unique internal gene expression system. In consequence, genes on the ptDNA are, therefore, semi-autonomously expressed with well-defined machineries consisting of a combination of nuclear encoded and plastid encoded gene products.

Until the 1980s, the importance of transcriptional regulation of plastid genes had not been emphasized that much, because meaningful transcriptional differences between etioplasts and chloroplasts or between dark-adapted and light-illuminated chloroplasts in seedlings had not been found due to limited observations (Deng and Gruissem 1987; Mullet and Klein 1987). However, the next two decades clarified the dynamics of plastid transcription in response to environmental stresses and in the process of chloroplast development (Allison 2000). It is now widely recognized that transcriptional regulation in plastids, particularly sigma factor-dependent transcription in chloroplasts, is primarily important (Lysenko 2007) and that subsequent post-transcriptional controls, including RNA editing and splicing, are also critical for correct and full gene expression from ptDNA (del Campo 2009).

Plastids have two different types of DNA-dependent RNA polymerases from lower to higher plants. One is PEP (plastid-encoded RNA polymerase) and the other is NEP (nuclear-encoded RNA polymerase) (Maliga 1998). Genes on the ptDNA are classified into three groups based on the transcriptional dependency on PEP and NEP in chloroplasts. Class-I genes such as photosynthetic psbA, psbD and rbcL are mainly transcribed by PEP. Class-III genes such as rpoB and accD are exclusively transcribed by NEP. Class-II genes such as rrn and clpP are mutually transcribed by both PEP and NEP (Hajdukiewicz et al. 1997).

A. Plastid-Encoded RNA Polymerase (PEP)

1. Transcriptional Activity and Physiological Roles of PEP

PEP is a bacterial type multisubunit enzyme consisting of catalytic core subunits and a promoter recognition sigma factor (Allison et al. 1996; Serino and Maliga 1998). Most ptDNAs, even in non-photosynthetic plastids of parasitic plants, have four rpo genes (rpoA, rpoB, rpoC1, and rpoC2) encoding the PEP core subunits (αββ′ and β″, respectively). However, it should be noted that the rpoA gene in the moss Physcomitrella patens had been deleted from its ptDNA, and two paralog genes exist on the nuclear DNA (Kabeya et al. 2007). PEP is indispensable for chloroplast development because disruptants of each plastid-encoded rpo gene no longer develop chloroplasts (Krause et al. 2000). However, the autonomy of PEP activity is on the side of nuclear DNA because the promoter recognition sigma factors of PEP had been deprived from ptDNA and relocated to the nuclear DNA during symbiotic evolution. In chloroplasts, PEP exists as a much larger protein complex with more than 13 accessory proteins (Bülow and Link 1987; Loschelder et al. 2004; Pfalz et al. 2006; Bollenbach et al. 2009), although the physiological roles of these accessory proteins have been only elucidated in part (Schröter et al. 2010).

Most of the Class-I and -III genes have characteristic upstream sequences similar to bacterial sigma70-type promoter consisting of −35 (TTGACA) and −10 (TATAAT) consensus sequences (Harley and Reynolds 1987; Weihe and Börner 1999). In bacteria, sigma factors recognize the promoter sequences and initiate transcription with the core subunits (Gruber and Gross 2003). Thus the bacterial promoter-like sequences on the ptDNA should be similarly recognized by plant sigma factors to initiate the PEP-dependent transcription. It is noteworthy that additional cis elements such as a bacterial extended −10 region (Sato et al. 1999) and a eukaryotic TATA box-like TATATAAgt element between the −35 and −10 regions are also found in some plastid genes (Eisermann et al. 1990).

2. Structure and Multiplicity of Sigma Factors

The nuclear-encoded sigma factors are synthesized as precursor proteins having an N-terminal transit peptide that consists of 32–80 amino acids in length. These precursors are transported into chloroplasts and then the N-terminal domain is cleaved by the stromal processing peptidase (SPP). The C-terminal region of plant sigma factors is well conserved and contains the regions 1.2–4.2 as well as those assigned in bacterial sigma factors (Hakimi et al. 2000). In contrast, the N-terminal halves of these mature sigma ­factors do not share significant homology with the bacterial orthologs (Schweer 2010).

Plant sigma factor genes were cloned from unicellular red algae for the first time (Liu et al. 1996; Tanaka et al. 1996). So far, sigma factors were also identified and characterized in a wide variety of photosynthetic organisms, including a moss P. patens (Hara et al. 2001), monocotyledonous rice (Tozawa et al. 1998; Kasai et al. 2004; Kubota et al. 2007), maize (Tan and Troxler 1999; Lahiri and Allison 2000), dicotyledonous mustard (Kestermann et al. 1998), tobacco (Oikawa et al. 2000), and A. thaliana in which a set of six sigma factor genes named AtSIG1 to 6 are encoded by the nuclear genome (Tanaka et al. 1997; Isono et al. 1997; Kanamaru et al. 1999; Fujiwara et al. 2000).

From a comparison of the conserved C-terminal domain, AtSIG2 and its orthologs in other plant species are most closely related to bacterial sigma70. Interestingly, AtSIG4 seems to be unique because no ortholog of this sigma factor has been identified so far (Schweer 2010).

3. Expression and Gene Specificity of Sigma Factors

PEP controls chloroplast development and its functional maintenance is primarily mediated by variation of sigma factors (i.e. sigma heterogeneity). In the last decade, the expression profile of Arabidopsis sigma factors (AtSIGs) and their gene specificity have been extensively studied. The latter topic has been mainly promoted by the use of T-DNA insertion (knockout) mutants of each sigma factor gene, except for AtSIG1 (Fig. 10.2).

Fig. 10.2.
figure 2

Transcriptional control in chloroplasts. NEP (RPOTp and RPOTmp) transcribes some primary genes, including rpo genes for plastid-encoded PEP core subunits, at the beginning of plastid differentiation into chloroplast. The PEP core enzyme binds to a promoter recognition sigma factor (SIG1 to 6) and transcribes proper genes for the establishment and maintenance of chloroplast function. Switching from NEP to PEP during the chloroplast development is achieved by NEP inactivation. Plastidial tRNA-Glu and thylakoid membrane-localized NIPs directly bind to RPOTp and RPOTmp, respectively, and repress their transcriptional activity. Among sigma factors, at least SIG1 and SIG6 are phosphorylated by kinase(s) such as cpCK2. The phosphorylated sigma factor would change its transcriptional activity and/or promoter specificity. Chloroplast-localized DG1 and SIB1 have been identified as sigma factor-binding proteins.

Full size image

Although very little is known about the function of AtSIG1, it appears as if the transcript is induced by blue and red light (Onda et al. 2008). An in vitro transcription assay using the psbA or rbcL promoter showed that AtSIG1 was less active than AtSIG2 or AtSIG3 (Privat et al. 2003). It is noteworthy that a disruptant of the rice SIG1 homolog (OsSIG1), whose gene products accumulate at a late stage during leaf development, was obtained (Tozawa et al. 2007). The mature leaves of the OsSIG1 mutant accumulated only a third of the chlorophyll content compared to the wild type. Since the level of the transcripts of the psaA operon was markedly reduced in the mutant, OsSIG1 seems to play an important role in the maintenance of PS I components in rice, and this is possibly similar in Arabidopsis.

The AtSIG2 knockout mutant was first reported by Shirano et al. (2000). The mutant showed reduced chlorophyll and aberrant organization of thylakoid membranes even in the mature leaves. AtSIG2 primarily regulates transcription of a part of housekeeping tRNA genes, including the trnEYD operon and two trnV genes (Kanamaru et al. 2001; Hanaoka et al. 2003). trnE, encoding tRNA-Glu, is particularly important because tRNA-Glu has at least three physiological functions indispensable for plastidial translation, tetrapyrrole synthesis, and the switch from NEP to PEP during chloroplast development (Hanaoka et al. 2005) as described below. AtSIG2 is also involved in the transcription of psbD initiated from one of multiple promoters (Hanaoka et al. 2003), psaJ (Nagashima et al. 2004a) in vivo, and rRNA operons initiated from the P2 promoter in vitro (Privat et al. 2003).

An AtSIG3 knockout mutant did not show a significant visual phenotype, but the expression of psbN was selectively decreased and indeed photosynthetic electron transfer was affected (Zghidi et al. 2007). It is notable that the AtSIG3 protein has a membrane-bound form (Privat et al. 2003).

An AtSIG4 knockout mutant also did not show any visual phenotype but the transcription of ndhF and possibly ndhG was significantly decreased (Favory et al. 2005). The functions of AtSIG3 and AtSIG4 may be redundant or minor compared to those of other sigma factors, otherwise their transcriptional activity is weaker than other sigma factors and consequently they rather function as a competitive “low gear” for plastid gene expression.

The AtSIG5 gene was preferentially ­activated under various environmental conditions such as light, salt and cold/heat stresses and its knockout mutant was surely less resistant under salt stress and strong light (Nagashima et al. 2004b). Since AtSIG5 was involved in the transcription of psbDC from the blue light responsive promoter (BLRP) that is dependent on the upstream AAG-box (Baba et al. 2001; Nagashima et al. 2004b; Tsunoyama et al. 2004), this stress-­responsive sigma factor is very likely to be required for rapid reconstruction of the photosynthetic reaction center (Kanamaru and Tanaka 2004).

Two AtSIG6 knockout lines have been isolated from different mutant pools. One has pale-green cotyledons at the beginning of germination (Ishizaki et al. 2005), while another allele showed a more severe “almost white” phenotype in cotyledons (Loschelder et al. 2006). However, both mutants recovered the early-stage deficiency and exhibited normal green color and growth after a week. AtSIG6 is a specific sigma factor functioning at the early-stage of chloroplast development in seedlings. Thus, AtSIG6 as well as AtSIG2 is a major sigma factor in chloroplast development. Furthermore, AtSIG6 may have an additional role in the transcription of two operons, atpBE and ndhC-psbG-ndhJ, even after the early stage (Schweer et al. 2006).

In addition to the temporal dynamics of sigma factors during chloroplast development, the expression profile of each plant sigma factor gene in organs and cell types is also diverse, and is probably correlated with the major function of each sigma factor (Homann and Link 2003). AtSIG5 proteins exist almost equally in both cotyledons and true leaves. AtSIG1 and 2 were shown to be more abundant in cotyledons, whereas AtSIG3 was abundant in true leaves (Privat et al. 2003). Any sigma factor gene significantly expressed in roots has not been reported yet. In monocotyledonous wheat leaves, PEP in the young cells along the expanding leaf axis showed different activity and specificity than in the mature cells at the leaf tip (Satoh et al. 1999). Maize has six sigma factors, ZmSIG1AB, 2AB, 3, and 6. ZmSIG1A and 1B are abundant in mature chloroplasts at the leaf tip, while the other four factors primarily accumulated in immature plastids at the leaf base. In vitro observation indicated that ZmSIG1A and SIG1B recognized characteristic ­promoters containing an extended −10 region raising the idea that transcription from such promoters is more active in leaf tips but less active at the leaf base (Lahiri and Allison 2000).

4. Regulation of Sigma Factors by Phosphorylation and Binding Proteins

In 2010 several exciting papers related to Arabidopsis sigma factors were published (Schweer et al. 2010; Shimizu et al. 2010; Chi et al. 2010). They suggest that PEP-dependent gene expression is controlled not only by the variation of sigma factors (i.e. sigma heterogeneity) but also by the phosphorylation of sigma factors and by direct binding of proteins with sigma factors (Fig. 10.2).

Mature sigma factors without a transit peptide have a unique variable extra region at the N-terminal portion that is not found in bacterial sigma factors. This plant specific sigma domain would confer functional differences not observed with bacterial sigma factors (Schweer 2010). In addition, it has been demonstrated that nuclear encoded plastid-targeted casein kinase 2 (cpCK2) is closely related to eukaryotic kinase rather than to the bacterial kinase and can phosphorylate plant sigma factors in vitro (Baginsky et al. 1999; Ogrzewalla et al. 2002; Jeong et al. 2004; Schliebner et al. 2008).

Schweer et al. (2010) transformed an AtSIG6 knockout mutant with a series of mutated AtSIG6 genes in order to prove the in vivo importance of phosphorylation of the sigma factor. Although they have not shown direct evidence of the AtSIG6 phosphorylation in vivo yet, their data indicate that the state of multiple phosphorylation of AtSIG6 by cpCK2 (encoded by At2g23070) affects the plastidial transcription profile and the visible phenotype. Furthermore, they propose that prephosphorylation of the Ser-177 residue of AtSIG6 by an unidentified “pathfinder” kinase is critical for generating a functional cpCK2 substrate site. The kinase may make the cpCK2 possible to phos­phorylate the Ser residue of AtSIG6 located at the highly variable region of plastid sigma factors.

On the other hand, Shimizu et al. (2010) showed in vivo phosphorylation of AtSIG1, the most abundant factor among the AtSIGs, and proposed a novel molecular mechanism by which the phosphorylation of AtSIG1 would cause a change of its promoter specificity and adjust gene expression of PS I and PS II components. Since the ­stoichiometry of PS I and II is important for photosynthetic electron transfer that depends on the redox state of chloroplasts, its ­adjustment should enhance photosynthetic efficiency (Link 2003; Steiner et al. 2009). They suggest that phosphorylation of AtSIG1 was stimulated when the plastoquinone pool was oxidized and inhibited. Phosphorylation of AtSIG1 inhibited the transcription of the psaA gene encoding a major PS I protein. In addition, results from transgenic Arabidopsis plants expressing AtSIG1 with or without the putative phosphorylation site indicate the involvement of the Thr-170 residue. By use of the yeast two hybrid system Puthiyaveeti et al. (2010) showed a possibility that Arabidopsis chloroplast sensor kinase (CSK encoded by At1g67840), a bacterial-type sensor histidine kinase, could directly interact with both AtSIG1 and cpCK2 (alternatively named PTK). They also showed a direct interaction between AtSIG1 and cpCK2 in yeast.

Piecing together these findings indicate that the “pathfinder” kinase proposed by Schweer may be CSK, if it can phosphorylate not only unidentified site(s) in cpCK2 but also the initial site in AtSIG6 (Ser-177) and possibly AtSIG1 under oxidative conditions of plastoquinone; in this case, phosphorylated-cpCK2 would give rise to further phosphorylation of these AtSIGs, and this phospho-relay should cause the transcriptional repression of PS I. It seems to be true that at least one of the plant sigma factors, if not all, are phosphorylated and a consequent conformational change should alter their transcriptional activity or specificity in response to developmental states and/or environmental conditions.

Furthermore, Chi et al. (2010) reported a pentatricopeptide repeat (PPR) protein named DELAYED GREENING 1 (DG1 encoded by At5g67570) directly bound to AtSIG6. PPR proteins are known as site-specific RNA-binding proteins in various enzyme complexes involved in RNA editing, processing, splicing and translation (del Campo 2009). Chi et al. (2010) suggest that direct binding occurs between the C-terminal region of DG1 and the N-terminal unconserved region of AtSIG6. The sig6 dg1 double mutant showed a more severe chlorotic phenotype and a greater decrease of PEP-dependent transcription than each single mutant. In contrast, overexpression of AtSIG6 complemented the chlorophyll deficiency in dg1 cotyledons. Interestingly, the genetic defect of dg1 caused an “increase” of AtSIG6-dependent gene transcripts in vivo.

DG1 is not the first example of a sigma factor-binding protein, because a chloroplast-localized sigma factor-binding protein 1 (SIB1) had already been reported as an AtSIG1 binding protein in vitro; however, the physiological function of SIB1 in plastidial transcription has not been shown (Morikawa et al. 2002). Concerning this point, the expression of SIB1 appeared to be induced by infection with a pathogenic bacterium. Furthermore, a SIB1 knockout mutant caused an apparent decrease in the expression of some nuclear-encoded defense genes, triggered by pathogen infection, salicylic acid and jasmonic acid. Plastid gene transcription by PEP was, however, not altered under these conditions (Narusaka et al. 2008; Xie et al. 2010). Therefore, DG1 is likely to be the first protein whose binding to a plant sigma factor and its physiological involvement in the regulation of plastid gene expression has been shown.

B. Nuclear-Encoded RNA Polymerase (NEP)

1. Transcriptional Activity and Physiological Roles of NEP

During extensive studies of the plastidial transcription machinery, it came to light that a part of plastid-encoded genes were still transcribed in PEP-deficient mutants or under PEP-inactive conditions; for example, the iojap mutant of maize (Han et al. 1992), plastid ribosome-deficient mutants of barley (Falk et al. 1993; Hess et al. 1993), a deletion mutant of rpoA in tobacco (Allison et al. 1996), and plants treated with PEP-specific inhibitors like tagetin and rifampicin (Kapoor et al. 1997; Liere and Maliga 1999; Bligny et al. 2000). These facts suggest the existence of a second RNA polymerase in plastids (Bünger and Feierabend 1980). This enzyme was expected to be a T3/T7 phage-type single-subunit RNA polymerase (RPOT) for two reasons. First, T3/T7 phage-type promoter sequences were found just upstream of the PEP-independent transcription initiation sites (Allison et al. 1996; Hajdukiewicz et al. 1997). Second, a single-subunit 110 kD RNA polymerase purified from spinach initiated transcription from the T7 promoter but not from a typical PEP promoter of rbcL gene in vitro (Lerbs-Mache 1993). The nuclear-encoded T7 phage-type plastidial RNA polymerase named NEP has been isolated from Arabidopsis, tobacco, spinach, rice, barley, Nuphar advena, and the moss P. patens (Hedtke et al. 1997, 2000, 2002; Kabeya et al. 2002; Richter et al. 2002; Emanuel et al. 2004; Kusumi et al. 2004; Liere et al. 2004; Azevedo et al. 2006; Yin et al. 2010). Dicotyledonous plants have two NEP proteins, chloroplast-targeted RPOTp and chloroplast/mitochondria-targeted RPOTmp (Kobayashi et al. 2001), whereas monocotyledonous plants have only chloroplast-targeted NEP (Chang et al. 1999; Kusumi et al. 2004). Both dicotyledonous and monocotyledonous plants have another T7 phage-type RNA polymerase named RPOTm targeted only to mitochondria (Tan et al. 2010).

NEP promoters have been identified from various genes such as accD, atpB operon, atpI, clpP, rpoB operon, rRNA operon, rps15 and rpl16. These promoters were classified into Type-I and -II. Type-I was sub-classified into Ia and Ib (Liere et al. 2004). Most NEP promoters contain a core YRTA motif (Type-Ia) that is also found in mitochondria promoters (Allison et al. 1996; Liere and Maliga 1999). Some of the conserved NEP promoters have an additional GAA-box motif upstream of the core motif (Type-Ib). Type-II promoters have been identified upstream of clpP, atpB, and the rps2-atpIHFA operon in tobacco, and the rRNA operon (Kapoor and Sugiura 1999). They have neither the YRTA-motif nor any other consensus motif. A Type-II promoter in the clpP gene overlapping with the transcribed region (−5 to +25) is highly conserved even in moss and is active in mature chloroplasts (Sriraman et al. 1998). The plastid rRNA operon (rrn) in Arabidopsis and mustard has three promoters. The P1 promoter depends on PEP, whereas P2 and PC promoters are Type-I and -II, respectively (Pfannschmidt and Link 1997). Some tRNA genes have another nonconsensus-type NEP promoter in the coding sequence (Wu et al. 1997). It is also noteworthy that all plastid genes, including the photosynthesis Class-I genes, are transcribed by NEP in PEP-deficient tobacco, although the transcriptional level was very low (Legen et al. 2002). Thus, the structure of NEP promoters is diverse; however, it is still unknown how higher plant plastids established these NEP promoters in the process of symbiotic evolution.

Both RpoTp and RpoTmp are expressed in young leaf cells, although the tissue specificity of these NEP proteins differs: RpoTp is dominantly accumulating in primary cortex cells of the stem and sepals. On the other hand, RpoTmp is apparently present in meristematic cells, leaf veins, companion cells around the phloem, stipules and root distal cells. Expression of RpoTmp precedes that of RpoTp in developing seedlings (Emanuel et al. 2004).

T-DNA insertion mutants of both RpoTp and RpoTmp as well as of sigma factor genes shed light on their functional difference and relevance. RpoTp T-DNA insertion mutants were isolated as “leaf development” mutants. The SCABRA3 mutants exhibit severe defects in chlorophyll biosynthesis and plant growth (Hricová et al. 2006). Despite having almost normal epidermal cells, the mesophyll cells were found to be irregular and expanded in the mutant. The transcripts of the NEP-dependent genes, including accD, rpoB, rpoC1, and clpP, were markedly reduced in the mutant. This is consistent with the observation that overexpression of RPOTp in tobacco enhanced transcription from a distinct subset of Type-I NEP promoters (Liere et al. 2004). The transcriptional compensation system of RpoTp by RpoTmp is unlikely because nuclear RpoTm and RpoTmp expression in the mutant were not altered at an early developmental stage. However, functional compensation of RPOTp by RPOTmp is likely because the RpoTp RpoTmp double mutant showed a severe developmental arrest of germination (Swiatecka-Hagenbruch et al. 2008). Thus, RPOTp-mediated gene expression is critical for an early stage of chloroplast development, and it affects not only mesophyll cell proliferation but also leaf morphogenesis.

In contrast to the RpoTp mutant, the RpoTmp T-DNA insertion mutant was isolated as a “short root” mutant (Baba et al. 2004). The mutant also showed delayed growth of young seedlings and slow greening of etiolated seedlings that almost disappeared over time. The RpoTmp mutants did not result in a drastic change in plastidial transcription; however, there was specific and transient disordered transcription from the Type-II PC promoter of the rrn operon during seed imbibition and germination (Courtois et al. 2007). In addition, the mutants showed reduced transcription of specific mitochondrial genes, but no change in promoter utilization (Kühn et al. 2009).

An in vitro transcription assay using three Arabidopsis T7 phage-type RNA polymerases overexpressed in E. coli approached the nature of their transcriptional activities and promoter specificities (Kühn et al. 2007). The recombinant RPOTm protein recognized precisely most of mitochondrial promoters as expected. However, the recombinant RPOTmp that exerted high activity when supercoiled DNA was used as a template did not show the ability to recognize promoters correctly, except for mitochondrial atp6 promoters. The RPOTp protein also could not initiate precise transcription from any well-characterized plastid NEP-dependent promoters in rpoB, accD, and clpP genes, although it recognized some mitochondrial promoters precisely. The inconsistency of in vivo and in vitro transcriptional specificity of RpoTmp and RpoTp may be due to the requirement of some additional factors for correct and full function of NEP in vivo.

2. Molecular Mechanisms of Switching from NEP to PEP

In an early stage of chloroplast development, the amount of NEP-dependent transcripts is promptly elevated for the construction of PEP core enzyme and initiation of fatty acid synthesis (Hess and Börner 1999). PEP binding to an adequate sigma factor accelerates transcription of photosynthesis genes including tRNAs; meanwhile, the NEP-dependent transcripts are progressively reduced to a minimum level. This NEP to PEP switch is likely to be critical for correct progression of the development of chloroplasts (Fig. 10.2), although a low level of PEP has already been stored in cells from the start of germination (Demarsy et al. 2006).

Plastid-encoded tRNA-Glu preferentially transcribed by PEP in a SIG2- and SIG6-dependent manner is very likely one of the switching molecules. The absence of functional AtSIG2 or SIG6 causes derepressed and continuous accumulation of NEP-dependent gene transcripts in vivo (Kanamaru et al. 2001; Loschelder et al. 2006). Then purified tRNA-Glu, but not other plastid tRNAs, directly bind to recombinant RPOTp and repress RPOTp-dependent accD transcription in vitro (Hanaoka et al. 2003). Increasing tRNA-Glu molecules in a SIG2- and SIG6-dependent manner at the early-to-middle stage of chloroplast development should cause enzymatic inhibition of the major NEP, RPOTp. The plastid tRNA-Glu has a unique third function as an inhibitor of RPOTp-dependent transcription in addition to its well-known fundamental functions in translation and as indispensable cofactor for 5-aminolevulinate synthesis (Schön et al. 1986).

The difference in turnover between NEP- and PEP-dependent mRNAs was proposed to contribute to the developmental switch (Cahoon et al. 2004). The amount of NEP-dependent mRNAs in maize was almost constant between leaf base and tip, corresponding to the early and late stages of chloroplast development, respectively. This constancy is probably caused by a decreased stability and increased transcription of the NEP-dependent mRNAs. tRNA-Glu bound RPOTp may not only be inactive but also unstable in nature, because the amount of intraplastidial RPOTp decreases as chloroplast development proceeds despite an increase in its transcriptional activity. On the other hand, the level of PEP-dependent mRNAs increased as leaves/chloroplasts mature due to both their increased transcription and constant (or increased) stability.

In addition to the switching of RPOTp to PEP, RPOTmp is also developmentally ­regulated at the level of its sub-plastidial localization and activity. RPOTmp proteins in spinach and Arabidopsis are apparently associated with thylakoid membranes because intrinsic thylakoid membrane proteins fix RPOTmp on the stromal side of the membrane. The NEP interacting proteins (NIPs) have three N-terminal transmembrane domains and a C-terminal ring finger domain. Light-dependent expression of NIPs at an early developmental stage may determine the membrane association of RPOTmp and down-regulate plastid rrn transcription, consequently functioning as an “inactivate” switch of RPOTmp (Azevedo et al. 2006, 2008).

3. Possible Accessory Proteins and Extended Function of NEP

In mammals, some accessory proteins of mitochondrial T7 phage-type RNA polymerase POLRMT have been isolated and characterized (Sologub et al. 2009). Although no homologous proteins have been isolated in plants, nuclear-encoded factors having similar function may be translocated into plastids to fully activate or modulate NEP-dependent transcription. It is also noteworthy that human POLRMT itself contains two PPR motifs at its N-terminus (Asin-Cayuela and Gustafsson 2007). The N-terminal sequences of NEP proteins are different from that of POLRMT, but there might be some specific PPR proteins that bind directly to NEP and modify its activity. At least, it has been already shown that plastids utilize a PPR, DG1, in concert with PEP (Chi et al. 2010).

In humans and rodents, the mitochondrial POLRMT has an alternative transcript producing a single-polypeptide nuclear RNA polymerase (spRNAP-IV). The shorter gene product loses the N-terminal domain necessary for targeting mitochondria. The nuclear-localized protein recognizes some promoters that differ from those for RNA polymerase II (Kravchenko et al. 2005). However, any similar phenomenon has not been found in plants thus far.

During the last decade, genetic information and physiological features of both PEP and NEP have been clarified from lower to higher plants. However, the molecular and biochemical details of the transcription system in plastids are still unclear. How do PEP and NEP mechanically recognize a specific sequence and initiate transcription? How many proteins, including PPRs, function closely with PEP and NEP? What are their functions? How is NEP activity linked to high-order cellular dynamics, for example, morphological control or environmental responses? How is transcription controlled in other plastids besides chloroplasts? What is the true transcriptional feature of wild type expressing NEP and sigma factors in just the adequate proportion? In the next decade, these questions will surely be clarified.