Map and sequence of the rpoS region. (A) Partial restriction map of the DNA encompassing the nlpD and rpoS open reading frames of E. coli, which consist of 321 and 330 codons, respectively. The function of nlpD (encoding a lipoprotein) is not related to that of rpoS. Arrows indicate the orientation of the two open reading frames and the known promoters. The upstream promoter cluster (P_nlpD), which is thought not to be regulated, serves both nlpD and rpoS. The downstream promoter (P_rpoS), is regulated and serves only rpoS; its transcript includes an untranslated leader of 564 nt. The first nucleotide of the transcript from P_rpoS (as specified in the text) is taken as the basis for numbering used here (nt 1). The numbers used previously by us (7, 10) can be converted to this system by adding an offset of 341 nt. Most lac fusions used in this study have lacZ placed to form either an operon or protein fusion at the EagI site within rpoS (codon 73). Other fusions, to codon 8 of rpoS, are so indicated in the text. (B) Two possible secondary structures for RNA including the end of the nlpD coding sequence and the short intergenic region up to the AUG start codon of rpoS. The UAA stop codon terminating nlpD lies at nt 500 to 502. The top structure is that proposed previously by us (7) and includes an upstream antisense element with three stems that can pair with a complementary sequence within the RBS, directly upstream of the rpoS AUG start codon. The Shine-Dalgarno (S.D.) sequence complementary to 16S rRNA is also indicated. The lacUV5 promoter was used to drive expression of various constructs of two general types, as detailed in the text. The end points of sequence derived from nlpD-rpoS included in these fusions can be described with reference to this figure as follows. Fusions for which the promoter is shown as rpoS include DNA starting from the ClaI site (A). Other fusions contain substitutions of the lacUV5 promoter followed by DNA starting from nt 1 of the P_rpoS transcript, the KpnI site, Δ1 (nt 344), Δ2 (nt 454), Δ3 (nt 477), or Δ4 (nt 541). For Δ2 through Δ4, included leader DNA sequences are shown in panel B.

Nature of the ppGpp^o regulatory defect. Various rpoS-lac fusion strains were grown and assayed for β-galactosidase as reported in Table 2. The stationary-phase induction ratio (expression at S + 3 divided by that at E1) is plotted in bar format for both mutant and ppGpp^o versions of each fusion. The ratio (value in the wild type divided by that in the ppGpp^o mutant) of the stationary-phase induction ratios is shown above the bar for each wild-type strain. The strains used are listed in Table 1.

Additive effect of ppGpp^o and a dsrA mutation. β-Galactosidase activities of exponential-phase (OD₆₀₀ = 0.25) and stationary-phase (S + 3 h) cultures are given. The activity of the ppGpp^o dsrA strain was below the limit of detection (<0.2 U) under both conditions. Stationary-phase induction ratios were as follows: wild type, 35-fold; dsrA, 23-fold; and ppGpp^o, 20-fold. The strains used are listed in Table 1.

Expression of rpoS-lac [pr] as a function of growth phase. (A) Growth curve of the wild-type strain carrying rpoS-lac [pr] (TE8197, solid squares) and its ppGpp^o mutant derivative (TE8199, open squares) in LB medium at 37°C. As described in the text, stationary phase (S) was defined as 1 h past the time at which the OD₆₀₀ reached 0.5. In this experiment, for the wild-type strain, S = 222 min, and for the ppGpp^o strain, S = 234 min. (B) Cultures were grown as in panel A and sampled at various times for assay of β-galactosidase. The first point for each curve corresponds to the time at which the OD₆₀₀ was 0.25. The x axes of the plots have been shifted slightly to align the points corresponding to an OD₆₀₀ of 0.5 at the 3-h mark. Each subsequent sample was taken at 1-h intervals.

Deletion mapping of sequences required for ppGpp^o and dksA mutation effects on rpoS expression. Various rpoS-lac fusion strains, all driven by the lacUV5 promoter but with various amounts of the upstream rpoS leader, were assayed for β-galactosidase during exponential phase and stationary phase (S + 3 h). Stationary-phase induction ratios were calculated and are plotted here. The underlying data for the ppGpp^o strain are from Table 3. The strains used are listed in Table 1.

Stationary-phase regulation requires the wild-type rpoS RBS. Various rpoS-lac fusion strains driven by the lacUV5 promoter were grown and assayed for β-galactosidase. The stationary-phase induction ratio (expression at S + 3 divided by that at E1) is plotted in bar format for protein (pr) and operon (op) versions of each fusion. Fusions marked +1 contained the entire 564-nt rpoS leader; those marked 541 contained only 24 nt upstream of the AUG start codon. The pr control transcript sequence is compared to the pr 541 transcript below the bar graph. In each fusion, codon 8 of the rpoS sequence is joined to lac. The sequence between the AUG initiation codon and the joint to lac was AUG AGU CAG AAU ACG CUG AAA GUU CAG GAT CCG (non-rpoS nucleotides underlined). The strains used are listed in Table 1.