replying to J. Wang et al. Nature Communications https://doi.org/10.1038/s41467-025-60689-2 (2025)

In our published work1, we found that the increase of cellular c-di-GMP levels in Salmonella enterica serovar Typhimurium can relieve H-NS-dependent transcriptional silencing of T6SS genes. We further showed that the binding of c-di-GMP to H-NS prevents binding of H-NS to DNA, thus relieving H-NS-imposed gene silencing. In addition, we found that the K107A variant of H-NS maintains unaffected DNA binding activity but loses response to c-di-GMP in vivo.

In their Matters Arising2, Wang et al. raise concerns regarding the interaction between c-di-GMP and H-NS reported in our article. After NMR analysis, ITC and EMSA experiments, Wang et al. conclude that c-di-GMP does not bind H-NS and does not prevent H-NS from binding DNA. We note that a c-di-GMP binding protein PilF159-302 has been included in their NMR analysis as a positive control, but not in their ITC experiments. In our published work, we incorporated multiple controls to validate experiments, including ITC analysis, UV-crosslinking and EMSA assays. In the UV-crosslinking experiments, the YcgR protein, a well-known c-di-GMP receptor3, was used as a positive control to ensure the reliability of the experimental results. In the ITC experiments, interactions of H-NS with c-di-GMP, c-di-AMP and cGMP were detected under the same experimental conditions to confirm the binding specificity of H-NS for c-di-GMP. In the EMSA experiments, control DNA was used in the protein-DNA binding analysis, and c-di-AMP and cGMP were used in the c-di-GMP-mediated interference of protein-DNA binding. Furthermore, the first author of our work, Shuyu Li, has repeated the ITC analysis and UV-crosslinking assay, which confirm our conclusion that c-di-GMP binds H-NS with high affinity. Furthermore, a recent article has reported that Lsr2, a xenogeneic silencer in mycobacteria, acts as a c-di-GMP receptor4. Although Lsr2 is quite different from H-NS, and its DNA-binding ability was reported to be enhanced after binding c-di-GMP, the study by Ling et al.4 and our study suggest a general pattern of nucleoid-associated proteins sensing alarmones.

Wang et al.2 also questioned the ability of c-di-GMP to prevent H-NS from binding DNA. Contrary to the EMSA results by Wang et al., our ITC and EMSA results have shown that c-di-GMP interferes with the binding of H-NS to DNA. As for the binding ability of the variants of H-NS for c-di-GMP and DNA, docking analysis was used to propose potential c-di-GMP-binding sites on H-NS, and the DNA-binding sites of H-NS were obtained from the structure models of H-NSCtd/DNA by Gordon et al.5. In our work, ITC analysis and/or EMSA assays showed that the point mutations Y99A, D101A, K107A and T115A result in reduction in the c-di-GMP binding affinity of H-NS, among which T115A reduces its DNA binding affinity while K107A does not affect its DNA binding. To support our findings, the ITC data of the H-NS variants Y99A, K107A and T115A titrated with c-di-GMP as well as K107A and T115A titrated with the promoter sequence of the clpV gene were provided (Fig. 1). Our results showed that the K107A variant of H-NS maintains unaltered DNA binding activity but loses response to c-di-GMP.

Fig. 1: ITC analysis of binding of H-NS and/or its variants for c-di-GMP and the promoter sequence of clpV.
figure 1

a–c c-di-GMP binds to the H-NS variants Y99A, K107A and T115A with reduced affinity. d–f K107A showed similar binding affinity, while T115A showed reduced affinity with the promoter sequence of clpV as compared to wild type H-NS. Data shown are one representative of three independent experiments with similar results, with Kd and complex stoichiometry (n) presented as mean ± SD.

Full size image

Regarding the inquiry about protein tags raised by Wang et al., we acknowledge their concerns. Indeed, in our experiments, the H-NS protein was produced using the pET-28a expression system, which leaves 17 non-native amino acid residues at the N-terminus after thrombin cleavage. However, our ITC results demonstrate a ~63–178-fold difference in c-di-GMP binding affinity between wild-type H-NS and its alanine-substituted mutants Y99A, K107A and T115A (Fig. 1a–c), all of which harbor the identical N-terminal tag. Furthermore, our previous study showed that the thrombin-cleaved His6-tag InvF with the identical 17 non-native N-terminal residues does not bind c-di-GMP in the ITC assays6. These results demonstrate that the high-affinity interaction between wild-type H-NS and c-di-GMP cannot be attributed to the additional 17 non-native N-terminal residues. In contrast, Wang et al. added 8 non-native residues LEHHHHHH at the C-terminus of the DNA-binding domain of H-NS. Based on the 3D structure of the DNA-binding domain of H-NS (PDB ID: 2L93; residues 91–137 of H-NS and 2 additional non-native C-terminal residues LE), we wonder whether the 8 non-native residues LEHHHHHH may sterically occlude some critical c-di-GMP-binding residues such as Y99 and K107, if the 8-residue C-terminal segment extends in some direction. In our study, the 17 non-native residues added at the N-terminus of full-length H-NS are distant from both the DNA-binding AT-hook-like motif and our proposed c-di-GMP binding site at the C-terminus of H-NS (according to the 3D structure of full-length H-NS predicted by the AlphaFold Protein Structure Database, AFDB accession number AF-P0A1S2-F1; https://alphafold.ebi.ac.uk/entry/P0A1S2). We therefore think unlikely that this 17-residue N-terminal segment will sterically occlude the DNA-binding AT-hook-like motif or our proposed c-di-GMP binding site. It remains to be tested if the 8 non-native C-terminal residues added by Wang et al. may modulate H-NS properties, including its interactions with c-di-GMP.

In conclusion, the conclusions published in our article1 were validated through multiple in vitro and in vivo experiments. We hope that other research groups will verify whether c-di-GMP binds H-NS and inhibits H-NS binding to DNA in the near future.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.