Systematic analysis of cyclic di-GMP signalling enzymes and their role in biofilm formation and virulence in Yersinia pestis

Abstract

Cyclic di-GMP (c-di-GMP) is a signalling molecule that governs the transition between planktonic and biofilm states. Previously, we showed that the diguanylate cyclase HmsT and the putative c-di-GMP phosphodiesterase HmsP inversely regulate biofilm formation through control of HmsHFRS-dependent poly-β-1,6-N-acetylglucosamine synthesis. Here, we systematically examine the functionality of the genes encoding putative c-di-GMP metabolic enzymes in Yersinia pestis. We determine that, in addition to hmsT and hmsP, only the gene y3730 encodes a functional enzyme capable of synthesizing c-di-GMP. The seven remaining genes are pseudogenes or encode proteins that do not function catalytically or are not expressed. Furthermore, we show that HmsP has c-di-GMP-specific phosphodiesterase activity. We report that a mutant incapable of c-di-GMP synthesis is unaffected in virulence in plague mouse models. Conversely, an hmsP mutant, unable to degrade c-di-GMP, is defective in virulence by a subcutaneous route of infection due to poly-β-1,6-N-acetylglucosamine overproduction. This suggests that c-di-GMP signalling is not only dispensable but deleterious for Y. pestis virulence. Our results show that a key event in the evolution of Y. pestis from the ancestral Yersinia pseudotuberculosis was a significant reduction in the complexity of its c-di-GMP signalling network likely resulting from the different disease cycles of these human pathogens.

Fig. 1. GGDEF and EAL domain proteins encoded by Y. pestis KIM10+

(A). The predicted structure and protein IDs of putative GGDEF and EAL domain proteins is shown. For simplicity, only GGDEF, EAL and transmembrane domains are shown. Black bars represent transmembrane domains in HmsT and HmsP as indicated by experimental data (Bobrov et al., 2008). Putative transmembrane domains predicted by TMHMM server v. 2.0 are shown as grey bars. The number of amino acid (aa) changes in conserved GGDEF or EAL residues are noted. (B). Each uncharacterized PDE domain (top) and DGC domain (bottom) from Y. pestis was aligned with characterized PDEs or DGCs from multiple organisms using T-Coffee (EBI server). Identical and similar residues (>80%) in enzymatically active c-di-GMP metabolic enzymes are indicated by black and grey backgrounds, respectively. Amino acid residues critical for activity (Crit. A.a.) are indicated in bold (Rao et al., 2008; Schirmer and Jenal, 2009). In addition to the Y. pestis (Yp) uncharacterized GGDEF and EAL domain proteins, 4 c-di-GMP PDEs (RocR from Pseudomonas aeruginosa [Pa] (Rao et al., 2008), BlrP1 from Klebsiella pneumoniae [Kp] (Barends et al., 2009), YhjH from Escherichia coli [Ec] (Pesavento et al., 2008), and HmsP from Y. pestis (Bobrov et al., 2005) as well as 4 DGCs (PleD from Caulobacter crescentus [Cc] (Paul et al., 2004), TM1163 from Thermotoga maritima [Tm], Slr1143 from Synechocystis sp [Sn] (Ryjenkov et al., 2005), and HmsT from Y. pestis (Simm et al., 2005)) are shown in Fig. 1. For simplicity, the remaining PDEs and DGCs used to generate the sequence alignments, as indicated in the Experimental Procedures, are not shown.

Fig. 2. A putative DGC, Y3730, is functional in Y. pestis

CR plates and a CV staining assay were used to assess Hms EPS production and biofilm formation in the following strains: KIM6-2051+ [hmsT⁻], KIM6-2051+ (pBAD-y3730), KIM6-2051.1+ [hmsT⁻ hmsP⁻ ], KIM6-2051.7+ [hmsT⁻ hmsP⁻ y3730⁻] and KIM6-2051.7+ (pBAD-y3730). Y. pestis KIM6-2118 (ΔhmsR) strain was used as a negative control. Results are from duplicate assays on two independent cultures. Error bars indicate standard deviations.

Fig. 3. Influence of HmsP, HmsT and Y3730 on cellular levels of c-di-GMP in Y. pestis cells

Relative c-di-GMP cellular concentrations in KIM6+ (parent strain), KIM6-2090.1+ (hmsP⁻), KIM6-2051.1+ (hmsT⁻ hmsP⁻) and KIM6-2051.7+ (hmsT⁻ hmsP⁻ y3730⁻) are shown. Results are from duplicate assays performed on two independent cultures. Error bars indicate standard deviations.

Fig. 4. y2559 is a pseudogene in Y. pestis

(A).Comparison of genomic region upstream of y2559 in Y. pestis KIM10+ (Y. p.) with Y. pseudotuberculosis PB1/+ ypts_1751 (Y. pst.). ORFs (arrows) are indicated above the nucleotide sequences. The insertion of a G in Y. pestis is highlighted in black and the resulting stop-codon prior to the start of y2559 is boxed. Identical nucleotide sequences in Y. pestis and Y. pseudotuberculosis before the frameshift mutation are shown in grey boxes. (B). Colony color on CR agar and a CV staining assay with KIM6-2051.7+ (hmsT⁻ hmsP⁻ y3730⁻) carrying pBAD30, pBAD-y3730 , pBAD-ypts_1751 and pBAD-y2559 are shown. Results are from duplicate assays performed on two independent cultures. Error bars indicate standard deviations.

Fig. 5. y2909 and y3389 are pseudogenes in Y. pestis

(A) Comparison of the genomic region upstream of y2909 in Y. pestis KIM10+ (Y. p.) with Y. pseudotuberculosis PB1/+ ypts_1400 (Y. pst.) and (B) y3389 (Y. p.) with ypts_3446 (Y. pst.). ORFs are indicated above the partial nucleotide sequences of genes encoding putative c-di-GMP PDEs. The insertion of an A upstream of y2909 and a G to an A transition upstream of y3389 are highlighted in black and the resulting stop-codon prior to the y3389 is boxed. Identical nucleotide sequences before the frameshift mutation are shown in grey boxes. (C) Colony color on CR agar and a CV staining assay with Y. pestis KIM6+ carrying the indicated full-length genes encoding putative c-di-GMP PDEs from Y. pestis and Y. pseudotuberculosis expressed from their own promoter on the low-copy vector pWSK29 are shown. Y. pestis KIM6-2118 (ΔhmsR) was used as a negative control. Results are from duplicate assays performed on two independent cultures. Error bars indicate standard deviations.

Fig. 6. y3841 lacks a consensus RBS and is not expressed

Colony color on CR agar and a CV staining assay with Y. pestis KIM6+ carrying plasmids with y3841 containing a C-terminal 6xtag and its native translation initiation region (pBAD-y3841-6xHis) or an artificial RBS (pBAD-RBS-y3841-6xHis) expressed using an arabinose-inducible promoter. KIM6+ carrying the vector plasmid (pBAD30) serves as a negative control. Results are from duplicate assays performed on two independent cultures. Error bars indicate standard deviations. Western blot analysis comparing Y3841-6xHis protein levels was performed with 6xHis-tag antibodies.

Fig. 7. The EAL domain of HmsP degrades c-di-GMP into linear di-GMP

(A) The effect of Mg²⁺, Mn²⁺, and Ca²⁺ on the c-di-GMP PDE activity of EAL-HmsP is shown. The PDE reaction with Mg²⁺ and Mn²⁺ was performed at pH 9.35 and 8.5, respectively (pH8.5 was used to avoid precipitation of Mn²⁺ complexes at higher pH values). Where indicated, CaCl₂ was added at a final concentration of 10 mM. (B): EAL-HmsP activity was examined at multiple pHs. C: The effect of alanine substitutions in conserved residues of the EAL domain of HmsP on c-di-GMP PDE activity is shown. Values are the averages of two independent experiments. Error bars indicate standard deviations.

Fig. 8. Mutations in hmsP reduce virulence in mouse models of bubonic and pneumonic plague

(A) The percentage of mice surviving a subcutaneous challenge with ~ 10³ cells of KIM5-2093.1 (pCD1Ap)+ [Parent strain] and ~ 10⁴ cells of KIM5-2090.9 (pCD1Ap)+ [hmsP⁻] is shown over a 3 week period. The percentage of mice surviving an intranasal challenge with ~ 10³ (B) or ~ 10⁴ (C) cells of KIM5-2093.1 (pCD1Ap)+ and KIM5-2090.9 (pCD1Ap)+ is shown over a 3 week period.

Fig. 9. Dot-blot analysis of the production of poly-β-1,6-GlcNAc by Y. pestis strains at 26°C and 37°C

Undiluted (U) and diluted (1:2, 1:8, 1:32) samples of crude cell extracts were spotted onto a nitrocellulose membrane and exopolysaccharide was detected with antisera against purified polysaccharide intercellular adhesin of Staphylococcus epidermidis. In (A) and (B), extracts are from different experiments. Strains: KIM6-2093.1+ (Parent); KIM6-2090.9+ (hmsP⁻); KIM6-2012 (hmsR⁻).

Fig. 10. A model of Y. pestis enzootic cycle regulation by c-di-GMP

An increase in c-di-GMP levels, modulated by HmsT, stimulates synthesis of the EPS and biofilm formation at ambient temperatures required for blockage-dependent plague transmission from fleas to mammals. Degradation of HmsT at 37°C reduces synthesis of c-di-GMP while the PDE activity of HmsP degrades the preexisting signal, abrogating the formation of the c-di-GMP- dependent biofilm EPS and allows the development of a lethal infection in mammals.