A cyclic-di-GMP receptor required for bacterial exopolysaccharide production

Abstract

Bis-(3',5')-cyclic-dimeric-guanosine monophosphate (c-di-GMP) has been shown to be a global regulatory molecule that modulates the reciprocal responses of bacteria to activate either virulence pathways or biofilm formation. The mechanism of c-di-GMP signal transduction, including recognition of c-di-GMP and subsequent phenotypic regulation, remain largely uncharacterized. The key components of these regulatory pathways are the various adaptor proteins (c-di-GMP receptors). There is compelling evidence suggesting that, in addition to PilZ domains, there are other unidentified c-di-GMP receptors. Here we show that the PelD protein of Pseudomonas aeruginosa is a novel c-di-GMP receptor that mediates c-di-GMP regulation of PEL polysaccharide biosynthesis. Analysis of PelD orthologues identified a number of conserved residues that are required for c-di-GMP binding as well as synthesis of the PEL polysaccharide. Secondary structure similarities of PelD to the inhibitory site of diguanylate cyclase suggest that a common fold can act as a platform to bind c-di-GMP. The combination of a c-di-GMP binding site with a variety of output signalling motifs within one protein domain provides an explanation for the specificity for different cellular responses to this regulatory dinucleotide.

Fig. 1

Phenotypic comparison of ΔretS mutation or overexpression of diguanylate cyclase on pellicle production and pel mRNA levels. A and B. Pellicle formation in static culture as revealed by crystal violet staining for (A) PA14, PA14ΔpelA, PA14ΔpelD and PA14ΔpelE harbouring vector control or induced for expression of diguanylate cyclases PA1107, PA1120 or PA3702; (B) PA14ΔretS, PA14ΔretSΔpelA, PA14ΔretSΔpelD and PA14ΔretSΔpelE. C. Expression of pel was determined by measuring β-galactosidase levels using a pelA–lacZ reporter construct in strains PA14, PA14ΔpelA, PA14ΔpelD and PA14ΔpelE harbouring vector control or induced for expression of diguanylate cyclases PA1107, PA1120 or PA3702. D. Quantification of PEL polysaccharide by binding to Congo red was normalized to total protein for strains PA14 pMMB-PA3702, PA14ΔpelD pMMB-PA3702, PA14ΔretS and PA14ΔretSΔpelD.

Fig. 2

Identification of c-di-GMP-binding protein within the pel operon. A. Coomassie-stained SDS-PAGE of purified MBP fusions to PelA and domains of PelC, PelD, PelE and PelF lacking transmembrane segments. B. Binding of [³²P]-c-di-GMP to each of the purified MBP-Pel fusion proteins. C. Specificity of c-di-GMP binding to MBP-PelD by competition with indicated unlabelled nucleotides.

Fig. 3

Role of conserved residues in PelD in c-di-GMP binding. A. Sequence alignment of PelD homologues. Similar residues are shaded in light grey and conserved residues (> 75%) are indicated by bold lettering. Residues conserved in all homologues are indicated by blue asterisks. Locations of point mutations introduced into PelD are indicate by red asterisks. B. Cartoon of the locations of introduced point mutations in the PelD sequence. C. Coomassie-stained SDS-PAGE of purified MBP-PelD with indicated point mutations. D. Binding of [³²P]-c-di-GMP to each of the purified MBP-PelD fusion proteins.

Fig. 4

Affinity of PelD for c-di-GMP. MBP-PelD is chemically cross-linked on gold surface and probed with either c-di-GMP at 5 μM (brown), 10 μM (purple) or 20 μM (cyan) or GTP at 5 μM (yellow), 10 μM (pink) or 20 μM (blue) and washed away at times indicated by the arrows. A. MBP-PelD. B. MBP-PelD (R367A). C. MBP-PelD (D370A). D. MBP-PelD (R402A).

Fig. 5

Secondary structure prediction and consensus sequence of PelD and PleD. A. Secondary structure predication was made using the web-based ProteinPredict program hosted at http://www.predictprotein.org/ for PelD and PleD. The confidence level of the prediction for each amino acid is located below the prediction. α-Helix and β-sheet elements predicted with a confidence level of greater than 5 is depicted on the diagram. Conserved residues for PelD and PleD are shown above the diagram and residues required for binding are shown in red. RxxD motif is indicated by the red box and GGEEF motif is in purple and is underlined. B. Consensus sequence of PelD and PleD was based on Phyre alignment of 20 and 100 homologues respectively. Residues required for c-di-GMP binding is indicated in red, GGEEF motif is in purple and is underlined and other conserved residues are in green. RxxD motif is indicated by the red box.

Fig. 6

Residues required for c-di-GMP binding are required for PEL polysaccharide synthesis. Complementation of the ΔpelD mutation in PA14ΔpelD pDN19-PA3702 with pMMB containing full-length pelD and indicated point mutations. Quantification of PEL polysaccharide binding to Congo red was normalized to total protein for each strain. Pellicle formation for each stain was tested in static culture and revealed by crystal violet staining. PelD and mutant derivatives were fused at their C-termini to HA tags. Total-cell extracts were prepared, loaded equally and separated on SDS-PAGE, transferred to PVDF and PelD-HA was detected with anti-HA antibody and chemiluminescence.