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.

No abstract