To ∼P or Not to ∼P? Non-canonical activation by two-component response regulators

Abstract

Bacteria sense and respond to their environment through the use of two-component regulatory systems. The ability to adapt to a wide range of environmental stresses is directly related to the number of two-component systems an organism possesses. Recent advances in this area have identified numerous variations on the archetype systems that employ a sensor kinase and a response regulator. It is now evident that many orphan regulators that lack cognate kinases do not rely on phosphorylation for activation and new roles for unphosphorylated response regulators have been identified. The significance of recent findings and suggestions for further research are discussed.

Fig. 1

Conserved residues required for RR phosphorylation. The N-terminal receiver domain of OmpR is represented modelled on the CheY structure (PDB 4qyw). The side chains of conserved residues are depicted including: aspartate 11 and 12, aspartate 55 (site of phosphorylation), threonine 83, lysine 102 and tyrosine 105. The side chain of tyrosine 105 exists in both outward (depicted here) and inward positions. Alpha helices are illustrated in cyan, beta strands in red and loops are in grey.

Fig. 2

Unphosphorylated response regulators bind DNA and bend DNA to de-repress H-NS and relieve gene silencing. The upper panel shows a three dimensional reconstruction of SsrB (yellow) binding to the csgD (green) regulatory sequence (for details see Desai et al., 2016) on which the repressor, H-NS (blue), has formed a rigid nucleoprotein complex. Binding of SsrB introduces a bend in the csgD regulatory sequence, resulting in destabilization of the H-NS filament and a relief of transcriptional silencing, as shown in the lower panel. We propose that binding of other SsrB homologues (DegU, AlgR and RcsB-BglJ) to upstream gene sequences, may also lead to similar conformational changes and transcriptional activation by anti-silencing.

Fig. 3

Strategies employed by RRs that do not involve phosphorylation. A. Binding and bending DNA de-represses H-NS by driving it off the DNA or remodelling how it interacts with DNA. For example, unphosphorylated SsrB binds and bends the DNA and that relieves H-NS silencing at csgD. B. Dimerization creates an active interface that is capable of DNA binding. This can involve homo- or hetero-dimerization. C. RR sequestration via protection of the active surface (preventing dimerization) by small molecule binding or protein binding. A modification of this theme in the case of KaiA/KaiC involves KaiA protein binding to KaiC, stimulating KaiC autophosphorylation. For AmiR, AmiC binds and prevents AmiR interaction with RNA. Peptide binding to AmiC releases AmiR, allowing AmiR dimerization and activation.

Fig. 4

Structure of an active HP1043 RR dimer. Electrostatic and hydrophobic interactions of the dimeric interface are shown. A network of ionic interactions is formed between Glu83 (α4), Asp93 (β5) and Arg108 (α5). The core interactions of the hydrophobic patch consisting of Val84 (α4), Phe87 (α4), Ala104 (α5), Ala107 (α5) and Ala111 (α5) are also shown between the dimeric interface. Reprinted with permission from Hong et al., 2007).

Fig. 5

Alignment of RRs, highlighting the absence of key catalytic residues. Sequences of the N-terminal receiver domains of RRs or the first 120 amino acids in the case of AmiR, GlnR and RamR, were obtained from the UniProtKB database (www.uniprot.org/uniprot/) or the NCBI Gene database (http://www.ncbi.nlm.nih.gov/). Alignment of these sequences using the software Expresso (tcoffee.crg.cat) is shown. The ‘Expresso’ colour codes indicate the consistency in pairwise structural alignments. The group of conserved residues in the receiver domain of OmpR are boxed in turquoise, while the (α/β)₅ topology is shown as black brackets for β-strands and magenta brackets for α-helices on top of the OmpR sequence.