BLAST alignments of representative Ig regions. The indicated Ig regions were used in a BLAST search of the NCBI Unfinished Microbial Genomes database. Each panel shows the summary figure provided by the BLAST program for matches to Salmonella enteritidis, Salmonella paratyphi A, Salmonella typhi, Salmonella typhimurium LT2, and Klebsiella pneumoniae; three contain known sRNA genes (rprA, csrB, and oxyS), and four contain sRNA candidates (#14, #17, #52, and #36; see Table 1). For each panel, the center numbered line represents the length of the full Ig region; the orientation of flanking genes is given by > (clockwise) or < (counterclockwise). The top red line in each panel is the match to Escherichia coli (full identity throughout the Ig). The other red or magenta lines resulted from the closest matches, and the other lines indicate additional less homologous matches. Location of the conserved region with respect to the borders of the Ig region also was a criterion used for the selection of our candidates; conservation 3′ to an ORF or far from the 5′ start of an ORF was considered more likely to encode an sRNA. Note that the conservation within the Ig region encoding oxyS might be interpreted as a leader sequence based on location relative to the start of the flanking gene (oxyR). However, because the conservation extends for 185 nt, candidate regions in our search in which the conservation was near the start of an ORF but was longer than 150 nt were considered further.

Expression profile across high-density oligonucleotide arrays for representative Ig regions. Probe intensities are shown for the indicated Ig regions (red) and the flanking ORFs (blue), calculated from the perfect match minus the mismatch intensities. All negative differences were set to zero. The data shown are for one experiment using cDNA probes, but similar results were seen in the duplicate experiment and with directly labeled RNA probes. The Ig regions and each flanking gene generally contain 15 interrogating probes. Upward bars correspond to genes transcribed on the Watson (W, clockwise) strand, and downward bars correspond to genes transcribed on the Crick (C, counterclockwise) strand. The C strand signal for the CsrB Ig region corresponds well with the known location of the csrB gene. Similarly for the RprA Ig region, the W strand signal corresponds with the location of the rprA gene, but only one probe is positive. The W strand signal for #14 and the C strand signal for #17 overlap well with the conserved regions shown in the BLAST analysis in Figure 1. #36 was chosen for further analysis because of the strong C strand signal; both flanking ORFs are on the W strand. For #52, low levels of expression were seen on both strands; the very low level for probes in the middle of the Ig on the C strand overlapped best with the conserved region found by the BLAST searches (Fig. 1).

Detection of novel sRNAs by Northern hybridization. Northern hybridization using strand-specific probes for each candidate was done on RNA extracted from MG1655 cells grown under three different growth conditions: (E) exponential growth in LB medium, (M) exponential growth in M63-glucose medium, and (S) stationary phase in LB medium. Five micrograms of total RNA was loaded in each lane. Exposure times were optimized for each panel for visualization here, therefore the signal intensity shown does not indicate relative abundance between sRNAs. Oligonucleotide probes were used for #12, #22, #55-I, #55-I, and #61; RNA probes were used for all other panels. DNA molecular weight markers (5′-end-labeled MspI-digested pBR322 DNA) were run with each set of samples for direct estimation of RNA transcript length. One lane of DNA molecular weight markers is shown for comparison, but these are approximate sizes because there was slight variation in the running of gels.

Coimmunoprecipitation of sRNAs with the Hfq protein. (A) Immunoprecipitations using extract from MG1655 cells grown in LB medium in exponential growth (OD₆₀₀ = 0.4) were done using no antibody (lane 1); 5 μL of preimmune serum (lane 2); or 0.5, 1, 5, or 10 μL of hfq antisera (lanes 3–6). Selected RNAs were fractionated on a 10% polyacrylamide urea gel after 3′-end labeling. Asterisks mark RNA bands present in the anti-hfq precipitated samples but not in the preimmune control samples and therefore represent Hfq-interacting RNAs. (B) Immunoprecipitations were done using extract from MG1655 cells grown under three different growth conditions: (E) exponential growth in LB medium, (M) exponential growth in M63–glucose medium, and (S) stationary phase in LB medium. Immunoprecipitations were carried out with 5 μL of preimmune sera (lane 1) or 5 μL of Hfq antisera (lane 2) and compared to total RNA from 1/10 extract equivalent used in the immunoprecipitations (lane 3). RNAs were fractionated on 10% polyacrylamide urea gels and analyzed by Northern hybridization using RNA probes to previously known sRNAs or our novel RNAs as indicated.