Multiple posttranscriptional regulatory mechanisms partner to control ethanolamine utilization in Enterococcus faecalis

Abstract

Ethanolamine, a product of the breakdown of phosphatidylethanolamine from cell membranes, is abundant in the human intestinal tract and in processed foods. Effective utilization of ethanolamine as a carbon and nitrogen source may provide a survival advantage to bacteria that inhabit the gastrointestinal tract and may influence the virulence of pathogens. In this work, we describe a unique series of posttranscriptional regulatory strategies that influence expression of ethanolamine utilization genes (eut) in Enterococcus, Clostridium, and Listeria species. One of these mechanisms requires an unusual 2-component regulatory system. Regulation involves specific sensing of ethanolamine by a sensor histidine kinase (EutW), resulting in autophosphorylation and subsequent phosphoryl transfer to a response regulator (EutV) containing a RNA-binding domain. Our data suggests that EutV is likely to affect downstream gene expression by interacting with conserved transcription termination signals located within the eut locus. Breakdown of ethanolamine requires adenosylcobalamin (AdoCbl) as a cofactor, and, intriguingly, we also identify an intercistronic AdoCbl riboswitch that has a predicted structure different from previously established AdoCbl riboswitches. We demonstrate that association of AdoCbl to this riboswitch prevents formation of an intrinsic transcription terminator element located within the intercistronic region. Together, these results suggest an intricate and carefully coordinated interplay of multiple regulatory strategies for control of ethanolamine utilization genes. Gene expression appears to be directed by overlapping posttranscriptional regulatory mechanisms, each responding to a particular metabolic signal, conceptually akin to regulation by multiple DNA-binding transcription factors.

Fig. 1.

Model for posttranscriptional control of gene expression in the eut locus of E. faecalis. A subset of genes present in the eut locus of E. faecalis are shown (blue boxes). A classical AdoCbl-sensing riboswitch (blue) is schematically represented upstream of genes predicted to encode for cobalamine transport proteins. A novel, AdoCbl-sensing riboswitch subclass described herein is indicated (red) within the intergenic region upstream of eutG. Conserved hairpins predicted to be transcriptional terminators are shown as green stem-loops. The stability of these terminators is postulated to be affected by the presence of ethanolamine, via the EutWV two-component system, or adenosylcobalamin, via a riboswitch. Posttranscriptional mechanisms predicted to increase downstream expression in response to their metabolic stimulus are indicated by a plus sign, whereas those predicted to decrease expression are indicated with a minus.

Fig. 2.

The 2-component system, EutWV, encoded within the eut locus is required for induced expression. (A) Organization of the eut locus in E. faecalis with the inset showing the translational fusions made to lacZ in an E. faecalis shuttle vector. For the predicted functions of the genes, see Table S1. (B) β-galactosidase activity levels observed in wild-type or eutVW E. faecalis strains containing the vectors with the lacZ fusions shown in A.

Fig. 3.

A putative ANTAR recognition motif. (A) Comparative sequence alignment of portions of the intergenic regions upstream of eutG, eutP, eutS, eutV, and eutA from Enterococcus, Listeria, and Clostridium species. A potential base-paired region is shown (red, helical residues) with a conserved primary sequence motif (gray box), which likely constitutes an ANTAR recognition motif. (B) Weblogo representation (38) of the conserved sequence motif. Size of the letter is indicative of frequency of occurrence. (C) A consensus secondary structure model for the base-paired stem and putative ANTAR recognition motif. (D) An electrophoretic mobility shift assay using γ-³²P-radiolabeled RNA that encompasses the eutP 5′ UTR demonstrates an interaction of the RNA with unphosphorylated (Left), but not phosphorylated (Right) EutV.

Fig. 4.

AdoCbl induces structural changes within the eutG 5′ UTR. (A) Representative in-line probing data for the E. faecalis eutG riboswitch region with increasing concentrations of AdoCbl and cyanocobalamin. Controls include nonreacted RNA (NR), partial digestion by RNase T1 (cleavage at G residues) and partial alkaline digestion (cleavage at all residues). Shaded symbols mark bands used for quantitative analyses. (B) Proposed secondary structure model for the E. faecalis eutT-G riboswitch showing nucleotides that are more constrained (red), less constrained (green), or unaffected (yellow) on addition of AdoCbl. Also shown is the consensus secondary structure model of the classical AdoCbl-sensing riboswitch. The gray box highlights the conserved core region (32). (C) Normalized fraction cleaved plotted against AdoCbl concentration is shown for the residues indicated in A. The gray line results from a 4-parameter logistic fit of the data. The results of the nonlinear regression analysis suggest a binding affinity of 540 nM for AdoCbl, and >2 mM for cyanocobalamin.

Fig. 5.

AdoCbl induces antitermination in vitro. (A) In vitro transcription analysis of the eutG 5′ UTR with varying concentrations of AdoCbl. Each individual reaction resulted in transcripts corresponding to premature termination (Lower) and runoff transcription of the DNA template (Upper). (B) The fraction of transcripts resulting from premature transcription termination is shown plotted against [AdoCbl]. The half-maximal change in the fraction termination corresponded to 350 nM AdoCbl.