Sialic Acid Derivatives Inhibit SiaT Transporters and Delay Bacterial Growth

Abstract

Antibiotic resistance is a major worldwide concern, and new drugs with mechanistically novel modes of action are urgently needed. Here, we report the structure-based drug design, synthesis, and evaluation in vitro and in cellular systems of sialic acid derivatives able to inhibit the bacterial sialic acid symporter SiaT. We designed and synthesized 21 sialic acid derivatives and screened their affinity for SiaT by a thermal shift assay and elucidated the inhibitory mechanism through binding thermodynamics, computational methods, and inhibitory kinetic studies. The most potent compounds, which have a 180-fold higher affinity compared to the natural substrate, were tested in bacterial growth assays and indicate bacterial growth delay in methicillin-resistant Staphylococcus aureus. This study represents the first example and a promising lead in developing sialic acid uptake inhibitors as novel antibacterial agents.

Figure 1

(A) Neu5Ac, the most abundant sialic acid in humans. (B) Catabolic pathway and molecular mimicry: the two different pathways employed by bacteria after sialic acid uptake. (C)Four known bacterial sialic acid transporter systems.

Figure 2

PmSiaT (green) and SaSiaT (blue) binding Neu5Ac, outside view of the two transporters. The binding sites are enlarged to display all the relevant residues and the differences between the two proteins. Positions 4, 5, and 9 of Neu5Ac are highlighted in pale red. The H-bond between 9-OH and Gln82 and Asp83, for PmSiaT and SaSiaT, respectively, is shown as a dashed line.

Figure 3

(A) General reaction conditions. (i) ArCH₂Br, Ag₂O, TBAI, and dry ACN; (ii) LiOH at r.t, followed by Amberlyst 15 H⁺ form at 100 °C in ACN-H₂O; (iii) RCOCl, TEA, dry MeOH, and 0 °C; (iv) Amberlyst 15 H⁺ form, ACN-H₂O 100 °C; (v) Nu in dimethylformamide or ACN, followed by either acidic (Amberlyst 15 H⁺ form, H₂O, 95 °C) or basic (NaOH or LiOH) hydrolysis. (B) nanoDSF results for all the synthesized compounds. The ΔT_m reported are given in °C and relative to Neu5Ac. Therefore, compounds with a positive ΔT_m indicate a thermal stabilization greater than Neu5Ac, while negative ΔT_m is associated with a smaller one. Compounds displaying ΔT_m values in the range of the control experiments (i.e., <−2.0 °C for PmSiaT and <−5 °C for SaSiaT) are to be considered nonbinders. All the experiments were performed in triplicates (n = 3) and the error bars represent ±s.d.

Figure 4

Representative MD snapshots at 195 ns of: (A) compound 3a; (B) compound 3e; and (C) compound 3f in complex with PmSiaT (pdb ID 5NV9). The bromo to Ser66 backbone carbonyl contact for compound 3f is indicated with a turquoise dashed line in panel (C).

Figure 5

Kinetic analysis of the inhibition of recombinant SiaT from P. mirabilis (A,B) and S. aureus (C,D) reconstituted in proteoliposomes and [³H]Neu5Ac efflux in proteoliposomes (E,F). (a–d) Data were plotted according to Lineweaver–Burk as a reciprocal transport rate vs reciprocal Neu5Ac concentration. The transport rate was measured, in 5 min, by adding [³H] Neu5Ac at the indicated concentrations to proteoliposomes containing 20 mM K⁺-gluconate in the presence of valinomycin as described in the Supporting Information. [(A,B) P. mirabilis SiaT]; (A) compound 3f 0.25 μM (●) or 0.75 μM (□) was added in comparison to samples without inhibitor (○). (B) Compound 3e 12 μM (●) or 40 μM (□) was added in comparison to samples without an inhibitor (○). [(C,D) S. aureus SiaT]. (C) Compound 3f 80 μM (●) or 200 μM (□) was added in comparison to samples without an inhibitor (○). (D) Compound 3e 5 μM (●) or 12 μM (□) was added in comparison to samples without an inhibitor (○). Results are mean ± s.d. from four independent experiments (n = 4). In (E), the efflux of [³H]Neu5Ac was measured from proteoliposomes harboring PmSiaT in the absence of external substrate (○), or in the presence of 0.1 mM of external Neu5Ac (■) or 0.1 mM compound 3e (□) or 0.1 mM compound 3f (●) at the indicated times. In (F), the efflux of [³H]Neu5Ac was measured from proteoliposomes harboring SaSiaT in the absence of an external substrate (○), or in the presence of 0.1 mM of external Neu5Ac (■) or 0.1 mM compound 3e (□) or 0.1 mM compound 3f (●) at the indicated times. Data are calculated as the percent of residual activity with respect to the control sample (efflux time zero). The results are the means ± s.d. of three independent experiments.

Figure 6

Inhibition of bacterial growth. (A) P. mirabilis HI4320 grown in defined carbon-limited MM supplemented with Neu5Ac (3.2 mM) and no inhibitor (dotted line), compound 3e (0.5 mM, solid line) and compound 3f (0.5 mM, dashed line). (b) S. aureus COL grown in defined carbon-limited MM supplemented with Neu5Ac (3.2 mM) and no inhibitor (dotted line), compound 3e (0.5 mM, solid line) and compound 3f (0.5 mM, dashed line). Average values of five independent biological replicates (n = 5) are presented, including a total of 13 technical replicates are presented. The horizontal dashed line indicates the end of the lag phase corresponding to the time for the initial population density to increase twofold.