Identification of small molecules that antagonize diguanylate cyclase enzymes to inhibit biofilm formation

Abstract

Bacterial biofilm formation is responsible for numerous chronic infections, causing a severe health burden. Many of these infections cannot be resolved, as bacteria in biofilms are resistant to the host's immune defenses and antibiotic therapy. New strategies to treat biofilm-based infections are critically needed. Cyclic di-GMP (c-di-GMP) is a widely conserved second-messenger signal essential for biofilm formation. As this signaling system is found only in bacteria, it is an attractive target for the development of new antibiofilm interventions. Here, we describe the results of a high-throughput screen to identify small-molecule inhibitors of diguanylate cyclase (DGC) enzymes that synthesize c-di-GMP. We report seven small molecules that antagonize these enzymes and inhibit biofilm formation by Vibrio cholerae. Moreover, two of these compounds significantly reduce the total concentration of c-di-GMP in V. cholerae, one of which also inhibits biofilm formation by Pseudomonas aeruginosa in a continuous-flow system. These molecules represent the first compounds described that are able to inhibit DGC activity to prevent biofilm formation.

Fig 1

Summary of lead hits from chemical screens. The numbers of hits with the given IC₅₀s isolated from the small-molecule screen are shown.

Fig 2

Representative enzyme inhibition assays. Inhibition of the DGCs VC2370(142)-D484E (circles) and WspR-R242A (squares) and of CIP (triangles) at various inhibitor concentrations is shown for DI-18 and DI-19.

Fig 3

Chemical structures of the DGC inhibitors.

Fig 4

Abilities of the seven DGC inhibitors to reduce biofilm formation in V. cholerae. (A) Biofilm formation of the V. cholerae ΔVC1086 mutant analyzed using an MBEC assay with and without 100 μM direct inhibitors. The ΔvpsL mutant of V. cholerae (vpsL), which cannot form biofilms, was a negative control. All strains/conditions were statistically significantly different from the ΔVC1086 mutant treated with DMSO (n = 6; P < 0.012). (B) Representative false-color flow cell image depicting the biofilm depth of V. cholerae, untreated or grown in 100 μM DI-3 and DI-10. (C) The biofilm biomass was determined by averaging nine separate images for each flow cell. The experiment was repeated 3 to 5 times for each treatment, and the graph displays the average biofilm biomasses with the associated standard deviations. *, P < 0.05; **, P < 0.001. 3, DI-3; 10, DI-10.

Fig 5

Abilities of the seven DGC inhibitors to reduce biofilm formation in P. aeruginosa. (A) Biofilm formation by P. aeruginosa strain PAO1 with and without 100 μM direct inhibitors. The pel fliA mutant of P. aeruginosa, which cannot form biofilms, was the negative control. No treatments were statistically significantly different from the DMSO-treated control. (B) Representative false-color flow cell image depicting the depths of P. aeruginosa biofilms, untreated or grown in 100 μM DI-3 and DI-10. (C) The biofilm biomass was determined by averaging nine separate images for each flow cell. The experiment was repeated 3 to 5 times for each treatment, and the graph displays the average biofilm biomasses with the associated standard deviations. * = P < 0.03). 3, DI-3; 10, DI-10.

Fig 6

DI-3 and DI-10 significantly reduce the intracellular concentration of c-di-GMP. The intracellular concentrations of c-di-GMP in the WT, the ΔVC1086 mutant, and the ΔVC1086 mutant grown with 100 μM each inhibitor were determined by UPLC-MS-MS. The data are normalized to the untreated ΔVC1086 strain. The error bars indicate the standard deviations. *, P < 0.02 compared with the DMSO-treated control.

Fig 7

Concentration response curve for DI-3. The IC₅₀ for the inhibition of V. cholerae biofilm formation in an MBEC assay by DI-3 was determined to be 26.2 μM, with a 95% confidence interval of 15.1 to 45.6 μM. The concentration response curve was generated in triplicate, and each point represents the mean value and standard deviation. The line is the best-fit curve as generated by the software Prism.

Fig 8

DI-3 is not toxic to mammalian cells. THP-1 macrophages were treated as indicated, and viability was measured at 8 h by trypan blue staining. The cells were killed by addition of 0.025% glutaraldehyde. The error bars indicate standard deviations.