5.4.2) Lrp regulation by MicF and DsrA
MicF is a 90 nt sRNA that was discovered as a negative regulator oflrp expression (Figure 4a ). The repressive activity of MicF on lrp was experimentally validated by analysing strains lacking or overexpressing MicF. In addition, the base pairing between MicF and lrp that were predicted by the Mfold programme was experimentally validated by (Holmqvist et al. , 2012). The conserved 5’-end of MicF represses the lrp mRNAs by seed pairing and binding to a region overlapping the AUG start codon and the early CDS region of lrp, and the binding process requires Hfq (Figure 4d ) (Corcoran et al. , 2012; Lee and Gottesman, 2016). Thus, Lrp translation is repressed by the formation of a complex that obstructs the proper 30S positioning on TIRs in lrp mRNAs (Holmqvist et al. , 2012). The repression of Lrp synthesis by MicF indirectly regulates downstream genes in the Lrp regulon (Holmqvistet al. , 2012).
DsrA has also been predicted and demonstrated to directly bindlrp mRNA (Lee and Gottesman, 2016). Regulation of lrp by DsrA was only demonstrated in strains overexpressing DsrA, while deletion of dsrA results in little or no change in the translational expression of lrp (Lee and Gottesman, 2016). This result suggested either a possible repression of lrp under low temperature and acid stress (Repoila and Gottesman, 2001; Repoila and Darfeuille, 2009; Bak et al. , 2014); or no regulatory role for DsrA exists in physiological condition. DsrA pairs early in thelrp ORF using the same region known to repress hns mRNA (Figure 4e ) (Lease, Cusick and Belfort, 1998). The role of this interaction is still under debate and it is expected to be intricate, considering that the binding site of DsrA on lrp overlaps with MicF binding site (Figure 4b ). In summary, if lrp mRNA is the target of several bindings of sRNA, the effectiveness of their regulation in vivo still needs to be studied in depth.
In this study, we delineated the manner by which GcvB and MicF regulate Lrp at the post-transcriptional level. Nevertheless, Lrp has also been shown to regulate GcvB and MicF at the transcriptional level (Figure 4a ) (Ferrario et al. , 1995; Modi et al. , 2011), implying the existence of a double-negative feedback loop. Considering that GcvB and MicF are highly expressed in fast-growing conditions (nutrient-rich medium), whereas Lrp is present in nutrient-poor environments, it could be suggested that MicF and GcvB repress lrp in nutrient-rich media and Lrp represses GcvB and MicF in nutrient-poor medium (Figure 4a ). Since most of the Lrp regulon contains proteins involved in amino acid biosynthesis, this regulatory circuit is physiologically reasonable. In accordance, deletion of gcvB affects lrp in LB medium but not in minimal medium when GcvB is poorly expressed (Lee and Gottesman, 2016), and the overexpression of MicF in a minimum medium showed a severe growth defect similar to the one observed in a lrp depletion strain (Holmqvist et al. , 2012). GcvB repression of lrp was also found to be effective under oxidative stress conditions (Figure 4a ), whereas only a modest increase in mRNA lrp stability and Lrp proteins was observed after oxidative stress in strains deleted for micF, indicating a modest repression oflrp by MicF under these conditions (Lee and Gottesman, 2016). Therefore, a double-negative feedback loop involving mainly GcvB could also occur under oxidative stress. This regulatory circuit also seems probable since Lrp regulates genes involved in oxidative stress (Kroner, Wolfe and Freddolino, 2019). Therefore, E. coli may use this dual repression scheme to promote a switch for adequate Lrp-dependent adaptation to nutrient availability and to oxidative stress. Further work is needed to determine exactly how this loop translates changes in MicF and GcvB abundance into changes in Lrp abundance. Considering that the binding sites of MicF and GcvB do not overlap, an additive repressive effect of both sRNAs on lrp mRNA stability needs to be tested.