In the last fifteen years, the important roles of small non-coding RNAs in regulation in all organisms have been recognized. Our laboratory, in collaboration with others, undertook two global searches for non-coding RNAs in E. coli, contributing significantly to the more than 80 regulatory RNAs that are now identified. A large number of these small RNAs (sRNAs) bind tightly to the RNA chaperone Hfq. We and others have shown that every RNA that binds tightly to Hfq acts by pairing with target mRNAs, regulating stability and translation of the mRNA, either positively or negatively. Our lab has studied a number of these sRNAs in detail. Each sRNA is regulated by different stress conditions, and that the sRNA plays an important role in adapting to stress. We have also examined the mechanism by which Hfq operates to allow sRNAs to act. The lab continues to investigate the in vivo roles of small RNAs, identifying the regulatory networks they participate in and their roles in those networks, an Hfq-binding RNA, named MgrR, is regulated by PhoP and PhoQ, a two-component system important for Salmonella virulence. PhoP and PhoQ activate synthesis of the RNA under low Magnesium and low calcium conditions; the small RNA inactivates an enzyme for modification of the cell surface lipopolysaccharide, eptB, affecting the cells sensitivity to antimicrobial peptides such as polymyxin. This is the first example of regulation of an LPS modifying enzyme by sRNAs. In collaborative work, we have confirmed the effect of the sRNA in regulating LPS modification. In addition, we find that the gene for the LPS modification enzyme is positively regulated by the specialized sigma factor Sigma E, leading to expression under conditions of periplasmic stress, when this LPS barrier may be particularly important. A second small RNA regulator of the eptB gene was also identified, linking regulation to a switch between aerobic and anaerobic growth. Many targets of sRNAs are regulated by more than one sRNA, and our earlier approaches have not allowed us to identify all of the sRNAs regulating a given target. In addition, regulatory proteins expressed at low levels or under special conditions may not be identified as targets for a given sRNA if their transcripts are not reasonably abundant. We have developed and made use of an approach to dealing with these limitations. Translational fusions can easily be created to possible target genes, identified based on the ability of their mRNA to bind Hfq, among other criteria, and the fusions can be rapidly screened with a library of plasmids, each expressing a different sRNA. Bacterial behavior can similarly be screened. We have applied this approach to the study of bacterial motility and biofilm formation, and have also examined the sRNAs regulating a set of important transcriptional regulators and proteins involved in stress responses. We find that multiple sRNAs regulate bacterial motility, many of them by regulating a critical transcriptional activator of flagellar synthesis, flhDC. Two sRNAs positively regulate motility, while at least four down-regulate motility. These provide unexpected new inputs to the well-studied regulation of flagellar synthesis. Bacteria such as E. coli are motile under some circumstances, but in some growth conditions form non-motile biofilms. Not surprisingly, we find that sRNAs play important roles in biofilm formation as well. Our results suggest that both flhDC, the central regulator of motility, and rpoS, encoding the stress sigma factor, act as nodes for regulation by multiple sRNAs. In our screen of transcriptional regulators for sRNA regulation, we find that only a subset of regulators, including Lrp and SoxS, are subject to sRNA regulation, and we are investigating the physiological significance of this extra level of regulation. In another study, a small RNA was found to negatively regulate tolC, the core of multiple drug efflux pumps in E. coli. The impact of this regulation on drug resistance is being investigated. The action of these small RNAs depends on the RNA chaperone Hfq, a protein with homology to the Lsm and Sm families of eukaryotic proteins involved in RNA splicing and other functions. Hfq binds both to sRNAs and to mRNAs, and stimulates pairing, but exactly how it does this is not entirely clear. We have now undertaken a careful in vivo dissection of Hfq, in collaboration with G. Storz (NICHD). Interesting hfq alleles were analyzed in vivo with multiple sRNA:mRNA reporters; the results demonstrate that some mutants are defective only for specific sRNA/mRNA pairs, suggesting that there are multiple modes for Hfq to bind and act to stimulate pairing. In addition, we have found that the Hfq-dependent sRNAs fall into two classes, defined by their behavior in different Hfq mutants. All of these sRNAs depend on the known sRNA binding site on the proximal face of Hfq for in vivo stability. The larger group, called Class I, is rapidly degraded when used, most likely dependent upon pairing. Mutations in the distal face of Hfq, which disrupt target mRNA binding, stabilize the Class I sRNAs, and mutations in the rim of Hfq, implicated in pairing, destabilize the sRNAs. Class II sRNAs are generally more stable than Class I sRNAs, are not destabilized by the rim mutants, but are by the distal site mutants. These results suggest at least two distinct modes of sRNA binding to Hfq. We have identified the distal site binding motifs in Class II sRNAs, and find that deleting these motifs renders the Class II sRNA stability more like that of Class I. We suggest that the roles of Class I and Class II sRNAs in the cell are somewhat different, reflecting their different stabilities, and are testing this by creating mutants and chimeric sRNAs that alter the behavior of the sRNAs. In order to determine if factors other than Hfq are necessary for the action of these sRNAs, a genetic selection was developed to select for failure of two sRNAs to act. The mutants isolated fell into two classes; the first, expected class, had changes in conserved and essential amino acids in hfq. The second class were loss of function mutations in pnp, encoding polynucleotide phosphorylase. Polynucleotide phosphorylase (PNPase) is a 3' to 5' endonuclease that associates with the RNA degradosome, an RNAse known to be involved in degradation of sRNAs as well as their target mRNAs. pnp mutations lead to increased instability and decreased levels of multiple sRNAs, and this decreased accumulation may be sufficient to explain their failure to act. Our genetic analysis suggests that PNPase may play an unexpected role in protecting sRNAs from degradation, probably by regulating the activity of the RNA degradosome. This proposal has now been confirmed by in vitro work from B. Luisi and students at the U. of Cambridge, and we are collaborating with them to further dissect how PNPase, Hfq, and the degradosome interact. Further genetic dissection of PNPase demonstrates that the active site of PNPase is critical for protection of sRNAs; this will also be investigated in vitro. This project is being pursued by Dr. DeLay, the postdoctoral fellow who carried out the work on PNPase, in his own laboratory. An unexpected additional direction for our study of translational regulation came from studies of the effect of aceE mutants on expression of RpoS, discussed as part of our proteolysis project. In addition to stabilizing RpoS, the aceE mutants greatly increase synthesis of RpoS, independent of the promoter and known sRNA regulators. Overall, we have developed highly efficient in vivo tools for studying sRNAs and the networks they reside in. Our focus is increasingly on the role of the sRNAs in complex bacterial behavior and investigations into the mechanism of sRNA function.