Identification of mutations that alter the gating of the Escherichia coli mechanosensitive channel protein, MscK

Abstract

Mechanosensitive channels allow bacteria to survive rapid increases in turgor pressure. Substantial questions remain as to how these channels sense and respond to mechanical stress. Here we describe a set of mutants with alterations in their MscK channel protein. The mutants were detected fortuitously by their enhanced ability to modify the accumulation of quinolinic acid. Some amino acid changes lie in the putative pore region of MscK, but others affect sequences that lie amino-terminal to the domain aligning with MscS. We demonstrate that the alterations in MscK cause the channel to open more frequently in the absence of excessive mechanical stress. This is manifested in changes in sensitivity to external K(+) by cells expressing the mutant proteins. Single-channel analysis highlighted a range of gating behaviours: activation at lower pressures than the wild type, inability to achieve the fully open state or a modified requirement for K(+). Thus, the dominant uptake phenotype of these mutants may result from a defect in their ability to regulate the gating of MscK. The locations of the substituted residues suggest that the overall gating mechanism of MscK is comparable to that of MscS, but with subtleties introduced by the additional protein sequences in MscK.

Fig. 1

Organization of MscK and MscS. Hydrophobicity plots for E. coli MscK and MscS proteins, using the Kyte–Doolittle algorithm (Kyte and Doolittle, 1982) with a window of 19 residues, were created using the DNAstar suite of programs. A. Full-length MscK. B. The MscK region corresponding most closely to MscS, showing the positions of the different mutations tested in this study; those originally identified from Salmonella MscK (below line); and that from E. coli strain RQ2 (above line). C. MscS, showing the positions of the three transmembrane (TM) helices and the carboxy-terminal domain deduced from the crystal structure (Bass et al., 2002).

Fig. 2

Expression of MscK mutant proteins and complementation of the nadB phenotype. A. Mutations were created using pTrcMscK as template and Quickchange (Stratagene) mutagenesis. Membrane preparations derived from MJF465 expressing wild-type MscK or MscK mutants were collected after induction with 0.3 mM IPTG for 30 min. Expression was detected by Western blot analysis using anti-His₆ antibodies (Sigma). Lanes 1–9 correspond to: vector pTrc99A, wild-type MscK, L565Q, G922S, double mutant L565Q/G922S, W909R, L866Q, G924S and R792P respectively. B. Salmonella typhimurium strain TT22824 (nadB499::mudJ, mscK267::Tn10dTc) was transformed with plasmids as described in Experimental procedures and after overnight growth in minimal medium with nicotinic acid, the cultures were serially diluted onto agar plates containing either nicotinic acid (NA; 1 μg ml⁻¹) or quinolinic acid (QA; 10 mM or 0.1 mM) and photographed after 24 h growth for NA and 10 mM QA and after 72 h for 0.1 mM QA.

Fig. 3

Effects of MscK mutations on growth of strain RQ2 in high-K⁺ or -Na⁺ medium in the presence of betaine. RQ2 cells expressing the MscK constructs were grown in K₂₀ medium to exponential phase, diluted 20-fold into K₂₀ containing either 0.6 M K⁺ or Na⁺ plus 50 μM IPTG and 1 mM betaine and the OD₆₅₀ of the cultures measured every 30 min. Specific growth rates were calculated and are compared (mean ± SD) for growth under high-K⁺ (filled columns) or high-Na⁺ (open columns) conditions (n = 3 for each). RQ2 strain, E. coli mutant previously described (McLaggan et al., 2002); WT, pTrcMscK; L565Q, R792P, L866Q, G922S, G924S and W909R, mutations created in pTrcMscK as described in Experimental procedures; pMAJ1, E. coli mscK expressed from its own promoter (McLaggan et al., 2002). Negative values show slow progressive reduction in light scattering indicating a complete lack of growth caused by induction of the cloned mscK gene.

Fig. 4

Effects of expression of MscK mutants on the growth of MJF465. MJF465 transformed with the MscK plasmids were grown overnight in K₁ or K₂₀ medium supplemented with ampicillin and then diluted 100-fold into K₁ or K₂₀ medium respectively. IPTG was added to a final concentration of 50 μM when the OD₆₅₀ reached 0.1 and growth was followed. Representative growth curves are shown for cells grown in K₁ medium (top) and cells grown in K₂₀ medium (bottom). Dotted line indicates time of IPTG addition. Data are shown for pTrcMscK (closed squares), L866Q (open triangles), W909R (open squares) and G922S (open circles). Data for R792P were almost identical to those for L866Q and data for G924S were identical to those for G922S (and are omitted for clarity).

Fig. 5

Assessment of the opening frequency of the MscK mutant channels using complementation of the K⁺ transport defect of a triple K⁺-uptake mutant. A. Single colonies of transformants of MJF603 (TK2309, mscK::Kan) carrying each mutant plasmid were streaked onto minimal medium plates containing 5–115 mM K⁺, incubated at 37°C for 24 h and then photographed. Data are shown for plates containing either 5 mM K⁺ (left) or 40 mM K⁺ (right): A, double mutant L565Q/G922S; B, G924S; C, G922S; D, W909R; E, L866Q; F, R792P; G, L565Q; and H, wild-type MscK. B. Minimal medium was inoculated with a culture that had been grown to exponential phase in K₄₀, such that the final concentration of K⁺ was either 9 mM (closed bars) or 40 mM (open bars) and the cells were grown in a shaking incubator at 37°C (n = 3). The OD₆₅₀ was measured at intervals and the specific growth rate determined from log-linear plots. Control, parent strain without any plasmid; WT, pTrcMscK; other mutations, created in pTrcMscK as described in Experimental procedures.

Fig. 6

Electrophysiological recordings of MscK channels. The wild-type and mutant MscK constructs were expressed in E. coli strain MJF429 and protoplasts prepared, as described in Experimental procedures. Inside-out excised patches were used to record single-channel activity. Representative traces are shown for all proteins that stably expressed in the membrane and gated with reproducible characteristics. Note that channels bearing the G922S mutation were unable to sustain the open conformation, thus exhibiting the appearance of flickery activity, and some mutant channels opened spontaneously, examples shown for the double mutant L565Q/G922S and the single mutant G924S (inset) where activity has occurred in the absence of pressure.

Fig. 7

Alignment of the putative pore region of MscK with the pore region of MscS. A. Identical and conserved residues are marked with asterisks and dots respectively. Residues in MscK that have led to gain-of-function phenotypes when mutated are indicated in larger font. Similarly, equivalent residues in MscS that have yielded gain-of-function channels (Miller et al., 2003b; Edwards et al., 2005) are also highlighted in larger font. B. Two helices from MscS (left) and MscK (right) are depicted to indicate the interface residue pairing for MscS (derived from the crystal structure) (Bass et al., 2002) and suggested interface pairing for MscK. The pairings for MscK have been created by analogy with the MscS structure on the assumptions that (i) MscK forms a heptamer and (ii) the helix crossing angles are not changed relative to MscS.

Fig. 8

Expression and electrophysiological recordings of the engineered pore mutants. A. MJF465 cells expressing either wild-type or the mutant MscK plasmids were induced with 0.3 mM IPTG for 30 min and harvested for membrane preparation. Expression was detected by Western blot using anti-His₆ antibodies (Sigma). Lanes 1–5 correspond to: vector pTrc99A, wild-type MscK, A916G, S920A and G926A respectively. B. Representative single-channel recordings of the three mutant channels after expression in strain MJF429 and preparation of protoplasts. Note that S920A MscK channels opened at pressures close to those required to open MscL channels (see text) and as such, a MscL single-channel opening is present in this trace (denoted by an asterisk).