Coupling sensor to enzyme in the voltage sensing phosphatase

Abstract

Voltage-sensing phosphatases (VSPs) dephosphorylate phosphoinositide (PIP) signaling lipids in response to membrane depolarization. VSPs possess an S4-containing voltage sensor domain (VSD), resembling that of voltage-gated cation channels, and a lipid phosphatase domain (PD). The mechanism by which voltage turns on enzyme activity is unclear. Structural analysis and modeling suggest several sites of VSD-PD interaction that could couple voltage sensing to catalysis. Voltage clamp fluorometry reveals voltage-driven rearrangements in three sites implicated earlier in enzyme activation-the VSD-PD linker, gating loop and R loop-as well as the N-terminal domain, which has not yet been explored. N-terminus mutations perturb both rearrangements in the other segments and enzyme activity. Our results provide a model for a dynamic assembly by which S4 controls the catalytic site.

Fig. 1. The linker, R loop, gating loop and catalytic site undergo voltage-driven conformational changes and mutation of K252E and D400R strongly affect these movements.

A Representative fluorescence traces from TMRM-labeled G214C of WT Ci-VSP, with K252E, and with D400R at 200 mV (magenta), 80 mV (blue), 20 mV (green) and −100 mV (black). B–H Superposition of representative fluorescence traces from ANAP incorporated at proximal linker 243 (B), distal linker 261 (C), near R-loop 290 (D), gating loop 401 (E) and 409 (F), catalytic site 522 (G) and C2 domain 555 (H). Fluorescence was collected simultaneously from 420–460 nm (red) and 460–500 nm (black) using protocol 2 (See “Methods”). Voltage steps are 200 mV, 160 mV, 120 mV, 80 mV, 40 mV and 0 mV. Only traces at 200 mV are shown for F401ANAP-K252E due to its small amplitude. I Comparison of fluorescence amplitude of ANAP at different sites to that with K252E and D400R from 460–500 nm (top) and 420–460 nm (bottom) (mean ± s.e.m.; n = 7, 5, 4, 7, 13, 17, 3, 15, 10, 9, 10, 19, 7, 8, 6, 16, 6, 14, 8, 14, 12 from left to right respectively). J Comparison of fluorescence ratio change between the 460–500 nm and the 420–460 nm emission channels at different sites to that with K252E or D400R (mean ± s.e.m.; n = 7, 6, 7, 4, 9, 7, 6, 15, 9, 4, 7, 7, 7, 6, 6, 16, 7, 14, 8, 14, 13 from left to right respectively). The cartoon in each panel indicates the individual labeling site. Recordings on individual sites with and without mutations were done on the same batch of oocytes and on the same day, and the same batch of injected oocytes were labeled with TMRM under the same conditions. Each n represents an independent measurement from an individual oocyte. The scales of X and Y axis are the same for each ANAP labeling site and that with K252E and D400R.

Fig. 2. Mutation of K252E and D400R strongly affect voltage dependent S4 conformational change and enzymatic activity.

A Representative fluorescence traces of TMRM-labeled Q208C of WT Ci-VSP, with K252E, with D400R and the ∆PD mutant (truncated after residue 240) at 200 mV (magenta), 80 mV (blue), 20 mV (green) and −100 mV (black). Cartoon indicates labeling scheme. B Comparison of F-V curves obtained from TMRM-labeled Q208C of WT Ci-VSP, with K252E, with D400R and the ∆PD mutant (mean ± s.e.m.; n = 23, 9, 10, 9, respectively). Fluorescence measured at the end of each test voltage was normalized to that obtained at 200 mV. C Schematics of the phosphoinositide pathways and substrates sensed by different FRET probes. D, E Representative FRET traces from F-PLC (D) and F-TAPP (E) co-expressed with WT Ci-VSP, K252E, D400R or ∆PD (under the corresponding labels and TMRM F-V curves in (A) using the 2-s test protocol (See “Methods”) and the comparison of normalized FRET from each condition (mean ± s.e.m.; n = 3, 3, 4, 9 respectively for F-PLC and n = 3, 4, 3, 9 respectively for F-TAPP). Each n represents an independent measurement from an individual oocyte. Some test voltages are shown next to corresponding traces. F, G Representative FRET traces from F-PLC (F) and F-TAPP (G) co-expressed with WT Ci-VSP, K252E and D400R using the long (20 s) test protocol.

Fig. 3. AlphaFold model of Ci-VSP and cryo-EM structure of Dr-VSP in lipid nanodiscs.

A AlphaFold predicted structure of Ci-VSP (with only part of the N-terminus shown). (Inset top left) Superposition of predicted model (cyan) and the crystal structure of Ci-VSP VSD R217E (PDB ID: 4G7V) (white). (Inset bottom left) Zoom in view of the catalytic site with side chain of catalytic residue C363 and gating loop residue 411 shown (top view). (Inset right) Zoom in view of areas with possible interactions between VSD and PD: PD linker (red), R loop (purple), gating loop (blue), and N-terminus (yellow). Important residues were shown. B The predicted structure colored by model confidence score pLDDT from low confidence (orange) to high confidence (blue). C Superposition of AlphaFold-predicted models of Ci-VSP (cyan) and Dr-VSP (gray). RMSD value for superpose is 0.889 Å. Same segments are colored with the same scheme as in (A). D Selected 2D class averages of Dr-VSP in lipid nanodiscs. E, F 6.8 Å cryo-EM map (gray) docked with AlphaFold model of Dr-VSP, viewed from the plane of the membrane. Dr-VSP with rainbow colored from N-terminus (blue) to C-terminus (red).

Fig. 4. Voltage-driven conformational change in N-terminal domain and perturbation by mutations in linker, R loop and gating loop suggest role for N-terminus in VSP activation.

A Labeling scheme and representative fluorescence traces of ANAP incorporated at different N-terminus sites as shown at 200 mV. Fluorescence was simultaneously detected from 460–500 nm (black) and 420–460 nm (red). B Representative fluorescence traces of ANAP incorporated at N-terminus residue 100 with mutation K252E, L284Q, F285Q and D400R at 200 mV. Fluorescence was simultaneously detected from 460–500 nm (black) and 420–460 nm (red). C Comparison of ANAP ∆F from residue 100 and from that with mutation K252E, L284Q, F285Q and D400R recorded from 460–500 nm (left) and 420–460 nm (right) (mean ± s.e.m.; n = 16, 8, 3, 3, 12 from left to right respectively). Each n represents an independent measurement from an individual oocyte.

Fig. 5. N-terminus deletion and D91A/D92A disrupts conformational coupling.

A–E Representative fluorescence traces from ANAP incorporated at proximal linker 243 (A), distal linker 261 (B), gating loop 401 (C), catalytic site 522 (D), C2 domain 555 (E), and the corresponding sites with N-terminus deletion ∆2-105 and with D91A/D92A mutation at 200 mV. Fluorescence was simultaneously recorded from 460–500 nm (black) and 420–460 nm (red). The cartoon in each panel indicates the individual labeling site.

Fig. 6. N-terminus mutations and AlphaFold-predicted interacting partner of D91/D92 alter voltage-driven S4 conformational changes.

A–C Representative fluorescence traces from TMRM-labeled Q208C of WT Ci-VSP and different N-terminus mutations as indicated at 200 mV (magenta), 80 mV (blue), 20 mV (green) and −100 mV (black). D Representative fluorescence traces from TMRM-labeled Q208C of Ci-VSP with mutation of the gating loop residue R397 (the AlphaFold-predicted interacting partner of D91/D92) at 200 mV (magenta), 80 mV (blue), 20 mV (green) and −100 mV (black). E–G Comparison of F-V curves of WT and different N-terminus deletions (E) (mean ± s.e.m.; n = 80, 4, 3, 4, 10, in the top to bottom order shown in the inset), point mutations (F) (mean ± s.e.m.; n = 80, 4, 4, 16, 3, in the top to bottom  order shown in the inset) and mutations of R397 (G) (mean ± s.e.m.; n = 80, 13, 13, in the top to bottom order shown in the inset). Fluorescence measured at the end of each test voltage was normalized to that obtained at 200 mV. Each n represents an independent measurement from an individual oocyte.

Fig. 7. N-terminus mutations that alter conformational coupling and R397D/N also affect enzyme activity.

A Representative fluorescence traces from F-TAPP (top row) or F-PLC (bottom row) co-expressed with WT, ∆2-105, D91N/D92N, D91A/D92A, R397D and R397N mutation using the 2-s test protocol (See “Methods”). Some test voltages are shown next to corresponding traces. B Comparison of the voltage dependence of FRET changes from F-TAPP when co-expressed with different Ci-VSP constructs as indicated. FRET values were measured at the end of the 2-s pulse to each voltage and normalized to that of −100 mV (n = 11, 9, 10, 10, 11, 11 in the listed order respectively). C Averaged fluorescence traces from F-TAPP and F-PLC co-expressed with different Ci-VSP constructs (as indicated and color-coded) using the long (15 s) test protocol. Traces from multiple rounds of experiments were averaged and experiments on WT and mutants were performed on the same batch of oocytes for each round (mean ± s.e.m.; n = 9, 8, 8, 10, 5, 12 (for FPLC) and n = 14, 12, 12, 15, 10, 7 (for FTAPP) for Q208C, ∆2-105, D91N/D92N, D91A/D92A, R397D and R397N, respectively). Each n represents an independent measurement from an individual oocyte.

Fig. 8. Model of voltage-catalysis coupling in VSPs.

A 3-way interaction between the distal VSD-PD linker (orange), the PD gating loop (pink) and the S1-proximal N-terminal domain (black) couples two sequential depolarization-driven outward movements of S4 to rearrangements in the PD that toggle it between three functional states. In the first step, S4 movement reorients the VSD-PD linker and the interacting gating loop to swing glutamate 411 (red) away from the catalytic pocket, thereby opening access for entry of PIP₃ substrate. In the second step, further reorientation of the linker-gating loop engages the membrane-proximal N-terminal domain of the VSD and alters the catalytic site to switch substrate preference to PIP₂.