Chloroplast and etioplast RNA polymerase preparations each consist of a multi‐subunit core and a set of three sigma‐like transcription factors, SLF67, SLF52 and SLF29. Despite this structural similarity, the enzymes from either plastid type are functionally distinct, as is reflected by their different promoter usage and the tight core‐SLF association in the etioplast but not the chloroplast holoenzyme. We tested whether these differences are related to phosphorylation. Treatment of the chloroplast enzyme with protein kinase converted it to an etioplast‐type form and vice versa, treatment of the etioplast enzyme with phosphatase generated chloroplast‐type properties. Although both the core enzyme and the SLF polypeptides were phosphorylation targets, only the SLFs seem to confer plastid‐type‐specific DNA binding characteristics. Methylation interference and DNase I footprint patterns in the psbA promoter region were found to correlate with the phosphorylation state of the chloroplast and etioplast enzymes.

A chloroplast protein fraction with σ‐like activity [Bülow, S. & Link, G. (1988) Plant Mol. Biol. 10, 349–357], was further purified and characterized. Chromatography on heparin‐Sepharose, DEAE‐Sepharose and Sephacryl S‐300 led to the separation of three σ‐like factors (SLF) polypeptides with Mr 67000 (SLF67), 52000 (SLF52) and 29000 (SLF29). None of these polypeptides bind to DNA itself, but each one confers enhanced binding and transcriptional activity when added to Escherichia coli RNA‐polymerase core enzyme and DNA fragments carrying a chloroplast promoter. SLF67, SLF52, and SLF29 differ in their ionic‐strength requirements for activity. They each mediate the binding to promoters of the chloroplast genes psbA, trnQ, and rps16, with different efficiencies. It is suggested that chloroplast transcription in vivo might be controlled at least in part by these functionally distinct factors.

The psbA gene which is differentially expressed in vivo in chloroplasts and etioplasts has an unusual promoter, containing both prokaryotic‐type ‘‐35’ and ‘‐10’ elements and a sequence motif that resembles the nuclear TATA box. Single base pair substitutions were introduced into the mustard psbA promoter and the mutants were tested in transcription and DNA binding experiments, using extracts from either chloroplasts or etioplasts. Positions within the ‘‐35’ region appear to play an essential role in the chloroplast but not in the etioplast system. Altering the first or second position of the ‘TATA box’‐like region led to decreased psbA in vitro transcription in either plastid extract. These two mutations, however, did not affect binding of extracts to the (linear) psbA promoter fragment in gel retardation assays. Fragments carrying two other plastid promoters effectively competed psbA promoter binding of the etioplast extract, but more weakly that of the chloroplast extract. Lambda exonuclease mapping shows that the 5′ border of the binding region is more upstream with the etioplast than with the chloroplast system, whereas the 3′ border appears to be the same. Hence, protein(s) of the two plastid types seem to interact differently with the mustard psbA promoter in vitro and perhaps also in vivo.

Three proteins resembling bacterial sigma factors were previously isolated from mustard chloroplasts (K. Tiller, A. Eisermann and G. Link, Eur J Biochem 198: 93-99, 1991). These sigma-like factors (SLFs) confer DNA-binding and transcription specificity to a system consisting of Escherichia coli core RNA polymerase and cloned DNA regions that carry a chloroplast promoter. Sigma-like activity was now isolated also from etioplasts and could be assigned to three polypeptides of M(r) 67,000 (SLF67), 52,000 (SLF52) and 29,000 (SLF29), i.e. the same sizes as for the chloroplast SLFs. The purification scheme for the factors from either plastid type included an initial heparin-Sepharose and a final gel filtration step. For the etioplast factors, however, an additional phosphocellulose step was required to release these polypeptides from the RNA polymerase. The etioplast SLFs have similar, but not identical, salt requirements for DNA binding as compared to their chloroplast counterparts. Under conditions of maximum binding activity there is overall preference of etioplast SLFs for the psbA promoter over the trnQ and rps16 promoters.

The major RNA polymerase from mustard chloroplasts is a multi-subunit enzyme consisting of core components and associated factors. Among the latter is a heterotrimeric factor named PTK (plastid transcription kinase) because of its serine/threonine-type protein kinase activity. PTK activity itself depends on its phosphorylation state. In addition, we show that it responds to glutathione but not to other redox-reactive reagents that were tested, and both glutathione and phosphorylation act antagonistically. Using a homologous in vitro system, we find that PTK selectively phosphorylates subunit(s) of plastid RNA polymerase and is involved in determining the level of faithful transcription from the chloroplast psbA promoter. Together, these results establish a role for phosphorylation and redox state in the regulation of plastid gene expression.