Enzyme Characterization of Pro-virulent SntA, a Cell Wall-Anchored Protein of Streptococcus suis, With Phosphodiesterase Activity on cyclic-di-AMP at a Level Suited to Limit the Innate Immune System

Streptococcus suis and Streptococcus agalactiae evade the innate immune system of the infected host by mechanisms mediated by cell wall-anchored proteins: SntA and CdnP, respectively. The former has been reported to interfere with complement responses, and the latter dampens STING-dependent type-I interferon (IFN) response by hydrolysis of bacterial cyclic-di-AMP (c-di-AMP). Both proteins are homologous but, while CdnP has been studied as a phosphohydrolase, the enzyme activities of SntA have not been investigated. The core structure of SntA was expressed in Escherichia coli as a GST-tagged protein that, after affinity purification, was characterized as phosphohydrolase with a large series of substrates. This included 3′-nucleotides, 2′,3′-cyclic nucleotides, cyclic and linear dinucleotides, and a variety of phosphoanhydride or phosphodiester compounds, most of them previously considered as substrates of E. coli CpdB, a periplasmic protein homologous to SntA and CdnP. Catalytic efficiency was determined for each SntA substrate, either by dividing parameters k(cat)/K(M) obtained from saturation curves or directly from initial rates at low substrate concentrations when saturation curves could not be obtained. SntA is concluded to act as phosphohydrolase on two groups of substrates with efficiencies higher or lower than ≈ 10(5) M(–1) s(–1) (average value of the enzyme universe). The group with k(cat)/K(M) ≥ 10(5) M(–1) s(–1) (good substrates) includes 3′-nucleotides, 2′,3′-cyclic nucleotides, and linear and cyclic dinucleotides (notably c-di-AMP). Compounds showing efficiencies <10(4) M(–1) s(–1) are considered poor substrates. Compared with CpdB, SntA is more efficient with its good substrates and less efficient with its poor substrates; therefore, the specificity of SntA is more restrictive. The efficiency of the SntA activity on c-di-AMP is comparable with the activity of CdnP that dampens type-I IFN response, suggesting that this virulence mechanism is also functional in S. suis. SntA modeling revealed that Y530 and Y633 form a sandwich with the nitrogen base of nucleotidic ligands in the substrate-binding site. Mutants Y530A-SntA, Y633A-SntA, and Y530A+Y633A-SntA were obtained and kinetically characterized. For orientation toward the catalytic site, one tyrosine is enough, although this may depend on the substrate being attacked. On the other hand, both tyrosines are required for the efficient binding of good SntA substrates.