A designer pore-blocking peptide reveals an electrostatic complementation mechanism for the inhibition-resistant Kv1.5 channel

Figures

• 

  Fig. 1.. Human Kv1.5 cannot be blocked by classical venom peptide blockers. (A) Cartoon depicting the architecture of the Kv1.5 channel and toxin peptide blocker in the pore of Kv1.5. (B) Top view of the Kv1.5 channel predicted using the AlphaFold Server, with the four subunits shown in different colors. The four R487 residues, highlighted as stick representations, are located within the outer pore vestibule. (C) Sequence alignment of Kv1.5, Kv1.1, and Kv1.2 channels. The R487 residue immediately after the channel selective filter can repel the binding of classical peptide blockers that are usually highly positively charged.

• 

  Fig. 2.. Kv1.5 peptide blockers are isolated from the SAK1 toxin phage display library by cell panning. (A) SAK1 toxin library contains ∼1.5 million natural and novel toxin peptides (16). Library phages were sorted on HEK293T and CHO cells expressing Kv1.5-RFP to select ligands for Kv1.5 channels. (B) After five rounds of selection on cells expressing Kv1.5-RFP, phage enrichment was observed compared with control cells without Kv1.5 expression. (C) Sequence alignment of CmTx, BgK, and HmK. A, B, and C domains that were used for SAK1 toxin library construction are underlined. Classical pore occluding residues in BgK (K25) and HmK (K22) were labeled blue, and the corresponding residues in CmTx (E23) were labeled red. The sequences and origins of the three parental peptides that contributed segments to CmTx are shown in fig. S1.

• 

  Fig. 3.. Block of Kv1.5 channels by CmTx. Kv1.5, Kv1.5-L510A, and other Kv channels expressed in HEK293T cells and recorded using the whole-cell patch clamp to evaluate equilibrium inhibition and kinetic blocking parameters. Two protocols were employed. Physiological conditions (A to C) mimicked the firing frequency and pH of healthy cardiac tissue: cells were depolarized to +0 mV for 0.2 s from a holding potential of −80 mV at 1 s intervals, with bath and intracellular pH of 7.4 and 7.2, respectively. Pathological conditions (D to G) used increased firing frequency and acidosis: Cells were depolarized to +20 mV for 0.2 s from a holding potential of −60 mV at 0.5-s intervals, with bath and intracellular pH maintained at 6.0. Values are means ± SEM. Some error bars are smaller than symbols. (A) Representative current traces for Kv1.5 channels at steady state before and in the presence of 2 μM CmTx. Currents were elicited with 0.2-s depolarizing steps in 20 mV increments from −80 to +40 mV at 1 s intervals (pHi = 7.2, pHo = 7.4). (B) Representative current traces for Kv1.5-L510A channels at steady state before and in the presence of 2 μM CmTx using the conditions in (A). (C) Inhibition of…

• 

  Fig. 4.. Effects of Kv1.5 pore mutations on CmTx affinity. CmTx inhibition of Kv1.5 and point mutant channels expressed in HEK293T cells was determined as in Fig. 3D, using 0.2 s depolarizing steps in 20 mV increments from −60 to +40 mV at 0.5 s intervals (pHi = pHo = 6). (A) Ki of CmTx for 11 channel mutants estimated using the Hill equation and assuming the same Hill coefficient as in the WT channel (Fig. 3F). Fractional currents (ICmTx/IWT or mutant) for WT or mutant channels in the presence of 2 μM CmTx were plotted. Changes of free energy of CmTx blockade (ΔΔG) were calculated as described in the Materials and Methods. Values are means ± SEM; n = 9 cells for WT Kv1.5; n = 5 cells each for the N459A, Q460A, G461A, H463A, and R487V mutants; and n = 3 cells each for the remaining six mutants. (B) AlphaFold-predicted tetrameric structure of Kv1.5. Individual subunits are shown in different colors, and five pore residues essential for CmTx binding are highlighted in stick representation on a single subunit.

• 

  Fig. 5.. Screening mutagenesis of T-CmTx. Kv1.5 was expressed in oocytes and inhibition by T-CmTx and its variants were studied by TEVC after coinjection of Kv1.5 and T-CmTx variant cRNAs. (A) T-CmTx was constructed as chimeric fusion protein with an N-terminal secretory signal sequence (SP; gray), the CmTx sequence (blue), a hydrophilic flexible linker (46 residues), and a C-terminal GPI membrane anchor targeting sequence (orange). The colors in the schematic mark T-CmTx components. (B) Representative current traces (0.2-s depolarizing steps from −60 to +60 mV at 0.5-s intervals, pHo = 7.5) for Kv1.5 channels with coinjection of 2 ng of T-CmTx cRNA or 2 ng of T-CTL, which is the T-toxin construct without the CmTx sequence (Materials and Methods). (C) Kv1.5 currents at +20 mV with T-CmTx variants coinjection were measured and normalized to unblocked Kv1.5 currents with T-CTL coinjection (IT-CmTx/IT-CTL), recorded within the same injection batch and day. Values are means ± SEM; n = 10 cells for each variant. (D) Ribbon representations of the AlphaFold-predicted structure of CmTx. Side chains of key residues identified by screening mutagenesis are shown in stick representation. Disulfide…

• 

  Fig. 6.. Structural models of the Kv1.5-CmTx complex predicted using AlphaFold. (A) Surface and ribbon representations of the predicted structure of CmTx. The Cα atoms and side-chain heavy atoms of key charged residues identified through mutagenesis screening (Fig. 5C) are shown as sticks. The electrostatic potential surface is shown alongside the structure. (B) Top view of the conductive Kv1.5 channel with key residues (Fig. 4A) highlighted as sticks (left). Molecular surfaces of WT Kv1.5 (middle) and the R487V mutant (right) are colored according to electrostatic potential, as calculated using ChimeraX (40). The R487V mutation reverses the local electrostatic potential from positive to negative (circled), thereby abolishing CmTx block. (C) Close-up view of the interaction interface between CmTx and conductive Kv1.5 in the structural model. Residues Y22, E23, K24, E27, and E28 of CmTx form extensive electrostatic interactions with R487 and H463 at the extracellular entrance of the Kv1.5 selectivity filter, highlighted by dashed lines. (D) Structural comparison of the conductive and inactivated selectivity filters of Kv1.5. Outward flipping of Y483 in the inactivated state disrupts the…