Novel compounds that specifically bind and modulate MscL: insights into channel gating mechanisms

Abstract

The bacterial mechanosensitive channel of large conductance (MscL) normally functions as an emergency release valve discharging cytoplasmic solutes upon osmotic stress. Opening the large pore of MscL inappropriately is detrimental to the cell, and thus it has been speculated to be a potential antibiotic target. Although MscL is one of the best studied mechanosensitive channels, no chemical that influenced bacterial growth by modulating MscL is known. We therefore used a high-throughput screen to identify compounds that slowed growth in an MscL-dependent manner. We characterized 2 novel sulfonamide compounds identified in the screen. We demonstrated that, although both increase MscL gating, one of these compounds does not work through the folate pathway, as other antimicrobial sulfonamides; indeed, the sulfonamide portion of the compound is not needed for activity. The only mode of action appears to be MscL activation. The binding pocket is where an α-helix runs along the cytoplasmic membrane and interacts with a neighboring subunit; analogous motifs have been observed in several prokaryotic and eukaryotic channels. The data not only demonstrate that MscL is a viable antibiotic target, but also give insight into the gating mechanisms of MscL, and they may have implications for developing agonists for other channels.-Wray, R., Iscla, I., Kovacs, Z., Wang, J., Blount, P. Novel compounds that specifically bind and modulate MscL: insights into channel gating mechanisms.

Keywords: antibiotic; mechanosensitive; osmoregulation.

Figure 1

Compound 011 specifically inhibits growth of MscL expressing bacteria. A) Minimal inhibitory concentration curves are shown for compound 120 (green) and 011 (red). Growth of the E. coli strain MJF612 (ΔmscL, ΔmscS, ΔmscK, and ΔybdG) carrying an empty plasmid (black) or expressing MscS (open squares) or Eco-MscL (solid squares) was measured as the OD₆₀₀ of the cultures. Values are expressed as the percentage decrease in growth in the presence of compound 120 or 011, relative to the nontreated samples (n = 3–8). B) Representative traces showing the effects on MscL channel activity of compounds 120 (upper traces) and 011 (lower traces). The schemes on the right illustrate 2 different configurations where the patch is presented with the compound from the periplasmic side (pipette) of the cytoplasmic side (bath). Channel activity was recorded by patch clamp experiments of native membranes, in a single patch before (left traces) and after (right traces) treatment with the compounds, at the same indicated pressure.

Figure 2

The mode of action compound 011. A) Minimal inhibitory concentration curves are shown for compounds 120 or 011 in cultures of the sulfonamide resistant strain PB121 (open markers) and its parental MJF612 strain (closed markers). Bacterial cultures carrying an empty plasmid control (black) or expressing Eco-MscL, were treated with compound 120 (green) or compound 011 (red). Values represent the percentage decrease in growth (OD₆₀₀) vs. nontreated. B) Minimal inhibitory concentration curves are shown for the MJF612 strain treated with compound 011 or 011A. Note that 011A is similar to 011 but lacking the sulfonamide group. Bacterial cultures carrying an empty plasmid (control) are shown in black, treated with 011 (closed squares) or 011A (open squares); cultures expressing Eco MscL treated with 011 (red) or 011A (blue) are shown. Values represent the percentage decreased in growth in the presence of compound 011 or 011A vs. nontreated. C) The reduction in viability of stationary cultures treated with 011 (red) or 011A (blue), expressed as the percent reduction of colony forming units (CFUs) vs. no treatment (n = 3–5). ***P < 0.0005, ****P < 0.00005, Eco WT vs. empty plasmid (2-tailed, 2-sample homoscedastic t test).

Figure 3

Compounds 011 and 011A block the binding of MTS-PEG5000 to specific cysteine mutants in a dose-dependent manner. A) The Eco MscL structure from molecular modeling, based on the Mycobacterium tuberculosis MscL crystal structure, is shown in a side view (left), with the approximate location of the bilayer indicated with horizontal bars; a single subunit is darkened for clarity. The location of the affected cysteine mutants is shown in a single subunit (middle) with residues E6 in green, V16 in blue, I96 in magenta, and K97 in red. A view from the cytoplasmic side (right) is also shown. B) Western blot analysis of MscL cysteine mutants after MTS-PEG5000 vs. compound 011 competition assay. The absence (−) or presence (+) of 50 μM MTS-PEG5000 (MTS-PEG), as well as the absence or different concentration of compound 011 used, are indicated in the table at the bottom of C. Top band: protein that has been PEGylated; this band disappears as the concentration of 011 is increased (1.5–5 mM). Both E6 and V16 dimmers can be seen slightly higher than the PEGylated protein band (PEG) (n = 3–7). Original magnification, ×2. C) Negative controls for compound 011: Eco-MscL with a mutation near the pore, I25C, shows a PEGylated band that is not inhibited or shifted by compound 011 (left), and WT Eco-MscL, which does not have any naturally occurring cysteine, shows no upper band with MTS-PEG5000 (right; n = 2). D) Western blot analysis after MTS-PEG5000 vs. compound 011A competition assay. The absence (−) or presence (+) of 50 μM MTS-PEG5000 (MTS-PEG) and the absence (−) or different concentration of compound 011A used, are indicated in the table at the bottom of E. Top band: protein that has been PEGylated; this band disappears as the concentration of 011A is increased (1.5–5 mM) (n = 3–5). E) Negative controls for compound 011A are shown: Eco-MscL I25C, shows a PEGylated band that is not inhibited or by compound 011A (left), and WT Eco-MscL, shows no top band with MTS-PEG5000 (right; n = 2). F) Docking site for 011A using Glide software. The position within the complex is shown on the left, and a close-up showing relative position to residues E6, F10, and K97 (K369 = K97, subunit 3).

Figure 4

Mutational analysis provides additional evidence that residue K97 in Eco-MscL contributes to the binding site of compound 011A. A) Sequence alignment of 2 MscL homologs from Eco-MscL and B.sub-MscL showing the difference at amino acid 97 (boxed). B) Inhibition of growth (OD₆₀₀) of E. coli strain MJF455 cultures, in the presence of compound 011A, expressed as the percentage of the nontreated. Note that changing the Eco-MscL K97 to R decreases 011A sensitivity, whereas changing B.sub-MscL R88 to K increases sensitivity (n = 3–6). ***P < 0.0005, ****P < 0.00005, mutations vs. WT for either Eco- or B.sub-MscL (2-tailed, 2-sample homoscedastic t test). C) Representative traces of the channel activity of Eco-MscL, B.sub-MscL, R88K B.sub-MscL, and K97R Eco-MscL, recorded from giant spheroplasts derived from the MJF612 strain. The effects of compound 011A on channel activity of MscL orthologs and mutants was recorded in the same patch before (left traces) and after (right traces) treatment with the 011A compound in the bath, at the indicated.