Experimental and computational analysis of transcriptional start sites in the cyanobacterium Prochlorococcus MED4

Abstract

In contrast to certain model eubacteria, little is known as to where transcription is initiated in the genomes of cyanobacteria, which are largely distinct from other prokaryotes. In this work, 25 transcription start sites (TSS) of 21 different genes of Prochlorococcus sp. MED4 were determined experimentally. The data suggest more than one TSS for the genes ftsZ, petH, psbD and ntcA. In contrast, the rbcL-rbcS operon encoding ribulose 1,5-bisphosphate carboxylase/oxygenase lacks a detectable promoter and is co-transcribed with the upstream located gene ccmK. The entire set of experimental data was used in a genome-wide scan for putative TSS in Prochlorococcus. A -10 element could be defined, whereas at the -35 position there was no element common to all investigated sequences. However, splitting the data set into sub-classes revealed different types of putative -35 boxes. Only one of them resembled the consensus sequence TTGACA recognized by the vegetative sigma factor (sigma70) of enterobacteria. Using a scoring matrix of the -10 element, more than 3000 TSS were predicted, about 40% of which were estimated to be functional. This is the first systematic study of transcription initiation sites in a cyanobacterium.

Figure 1

RACE mapping of cyanobacterial primary mRNA 5′ ends. Amplification of mRNAs for ftsZ, psbD, NtcA and rbcLS is shown. In most cases a single TAP-specific PCR product was obtained (lane T+). RNA control samples included: untreated (U), 5′ dephosphorylated (C) and mock-treated in TAP-buffer but without TAP (T–). Total DNA served as yet another control for amplification of reverse transcription-specific cDNAs (lane D).

Figure 2

Alignment of –10 boxes of 25 promoter sequences of experimentally determined promoters (left) and counted nucleotides at each position of the 25 aligned –10 boxes (right). The positions –12 to –7 were found for the majority of TSSs, except that it was –13 to –8 for ntcA2, groES, rbcL, petH1 and coaT and –14 to –9 for atpB (see Tables 1 and 2).

Figure 3

Scheme of promoter prediction using the kaiB gene as an example. The strategy consisted of three steps: (i) raster, a regular expression search based on biological features; (ii) score, determining the weights of the potential boxes with a scoring matrix; (iii) filter, using a filter (cut-off value) to reduce the false positive rate of the predicted boxes. The numerical values have been rounded.

Figure 4

Cross-species comparison of the 5′-UTR of ccmk-rbcLS from marine cyanobacteria. The experimentally identified TSS (labelled by arrows above the sequence), 5′-Gaacat for Prochlorococcus MED4 (this work) and 5′-gAacat for Synechococcus WH 7803 (21), are part of a conserved nucleotide motif at the mRNA 5′ end. A strong conservation of this motif and of the suggested –10/–35 regions from MED4 is not only obvious in the two species but also apparent in Synechococcus WH 8102 and Prochlorococcus MIT9313, for which no experimental data exist.