Molecular basis for plasma membrane recruitment of PI4KA by EFR3

Abstract

The lipid kinase phosphatidylinositol 4 kinase III α (PI4KIIIα/PI4KA) is a master regulator of the lipid composition and asymmetry of the plasma membrane. PI4KA exists primarily in a heterotrimeric complex with its regulatory proteins TTC7 and FAM126. Fundamental to PI4KA activity is its targeted recruitment to the plasma membrane by the lipidated proteins EFR3A and EFR3B. Here, we report a cryogenic electron microscopy structure of the C terminus of EFR3A bound to the PI4KA-TTC7B-FAM126A complex, with extensive validation using both hydrogen deuterium exchange mass spectrometry, and mutational analysis. The EFR3A C terminus undergoes a disorder-order transition upon binding to the PI4KA complex, with an unexpected direct interaction with both TTC7B and FAM126A. Complex disrupting mutations in TTC7B, FAM126A, and EFR3 decrease PI4KA recruitment to the plasma membrane. Multiple posttranslational modifications and disease linked mutations map to this site, providing insight into how PI4KA membrane recruitment can be regulated and disrupted in human disease.

Fig. 1.. Cryo-EM analysis of EFR3 binding to the PI4KA complex.

(A) Domain schematics of the full-length PI4KA complex and EFR3A constructs used in this paper. Constructs used in this paper are PI4KA/TTC7B/FAM126A ΔC (referred to as PI4KA complex) and EFR3A (721 to 791) referred to as EFR3A C terminus. (B) Size exclusion chromatography traces of (L) PI4KA complex/EFR3A C terminus, (M) PI4KA complex, and (R) EFR3A C terminus, with corresponding SDS-PAGE gels to show protein present in the highlighted size exclusion chromatography peaks. (C) Cryo-EM density map of the PI4KA complex bound to EFR3A C terminus. EFR3A C terminus was bound to both sides of the heterotrimer. (D) Molecular model of the triple helical V in the C terminus of EFR3A (residues 724 to 787) bound to the PI4KA complex, with a zoom-in highlighting EFR3A contacts with both TTC7B and FAM126A.

Fig. 2.. HDX-MS analysis of the interaction of EFR3A with TTC7B and FAM126A.

(A) Significant differences in deuterium exchange (defined as >5%, >0.45 Da, and P < 0.01 in an unpaired two-tailed t test at any time point) upon PI4KA complex binding to EFR3A. (B) Sum of the number of deuteron differences of PI4KA complex upon binding with EFR3A analyzed over the entire deuterium exchange time course for the PI4KA complex. Each point is representative of the center residue of an individual peptide. Peptides that met the significance criteria described in (A) are colored red. Error is shown as the sum of SDs across all four time points (SD) (n = 3). (C) Size exclusion chromatography traces of (L) TTC7B-FAM126A ΔC-EFR3A, (M) TTC7B-FAM126A ΔC, and (R) EFR3A, with corresponding SDS-PAGE gels of the indicated peaks. (D) Significant differences in deuterium exchange in TTC7B-FAM126A upon EFR3A binding [significance criteria described in (A)]. (E) Sum of the number of deuteron differences of TTC7B and FAM126A upon complex formation with EFR3A analyzed over the entire deuterium exchange time course for the dimer. Peptides that met the significance criteria described in (A) are colored red. Error is shown as the sum of SDs across all five time points (SD) (n = 3). (F) Significant differences in deuterium exchange in MBP-EFR3A upon TTC7B-FAM126A binding [significance criteria described in (A)]. (G) Sum of the number of deuteron differences of EFR3A upon complex formation with TTC7B-FAM126A analyzed over the entire deuterium exchange time course for the EFR3A. Peptides that met the significance criteria described in (A) are colored red. Error is shown as the sum of SDs across all four time points (SD) (n = 3). Individual deuterium exchange curves for significant differences for all conditions are shown in fig. S4.

Fig. 3.. Molecular basis of EFR3A binding to TTC7B.

(A) Cartoon (L) of a biolayer interferometry (BLI) experiment showing binding of immobilized His-MBP-EFR3A (721 to 791) to TTC7B-FAM126A. Association and dissociation curves (M) for the binding of His-MBP-EFR3A (721 to 791) to TTC7B-FAM126A (10 to 2500 nM). The experiment was carried out in duplicate, with all data shown. Normalized BLI response versus concentration of TTC7B-FAM126A (R), with K_D estimated by one site specific nonlinear regression. Each data point is shown (n = 2). (B) Multiple sequence alignment of EFR3A from H. sapiens, M. musculus, X. laevis, D. rerio, and D. melanogaster. EFR3A secondary structures of α1- and α3-helices are annotated above the alignment. Contact residues within 6 Å of TTC7B are annotated using arrows. Contact residues with BSA > 30 Ų are annotated using red arrows. (C) Multiple sequence alignment of TTC7B from H. sapiens, M. musculus, X. laevis, D. rerio, D. melanogaster. TTC7B secondary structure is annotated above the alignment. Contact residues within 6 Å of EFR3A are annotated using arrows. Contact residues with BSA > 30 Ų are annotated using red arrows. (D) Zoomed in cartoon view of the EFR3A α1-TTC7B interface with putative interfacial residues labelled. EFR3A and TTC7B are colored according to in-figure text. (E) Zoomed in cartoon view of the EFR3A α3-TTC7B interface with putative interfacial residues labelled. EFR3A and TTC7B are colored according to in-figure text. (F) Maximum BLI response of various EFR3A and TTC7B mutants compared to wild type (WT). Error is shown as SD (n = 3) with two-tailed p values indicated as follows: **P < 0.001 and ***P < 0.0001; not significant (ns) > 0.01. (G) Raw BLI association and dissociation curves of EFR3A and TTC7B mutants compared to WT. His-EFR3A was loaded on the anti–penta-His tip at 200 nM and dipped in TTC7B-FAM126A at 500 nM. Raw BLI curves of all mutants in fig. S5.

Fig. 4.. Mutational analysis validates the EFR3A-FAM126A binding interface.

(A) Multiple sequence alignment of EFR3A from H. sapiens, M. musculus, X. laevis, D. rerio, and D. melanogaster. EFR3A secondary structures of α2- and α3-helices are annotated above the alignment. Contact residues within 5 Å of FAM126A are annotated using arrows. Contact residues with BSA > 30 Ų are annotated using red arrows. (B) Multiple sequence alignment of FAM126A from H. sapiens, M. musculus, D. rerio, D. melanogaster, and C. elegans. FAM126A secondary structure is annotated above the alignment. Contact residues within 5 Å of EFR3A are annotated using arrows. Contact residues with BSA > 30 Ų are annotated using red arrows. (C) Zoomed in cartoon view of the EFR3A-FAM126A interface with putative interfacial residues labeled. EFR3A and FAM126A are colored according to in-figure text. (D) Maximum BLI response of various EFR3A and FAM126A mutants compared to WT. Error is shown as SD (n = 3) with p values indicated as follows: *P < 0.01; **P < 0.001; ***P < 0.0001; not significant (ns) > 0.01. (E) Raw BLI association and dissociation curves of EFR3A and FAM126A mutants compared to WT. His-EFR3A was loaded on the anti-penta-His tip at 200 nM and dipped in TTC7B-FAM126A at 500 nM. Raw BLI curves of all mutants in fig. S5.

Fig. 5.. The C terminus of EFR3 mediates PM recruitment of PI4KA.

(A) Cartoon depicting the quantitative BRET-based PI4KA recruitment assay. Briefly, the PM-anchored BRET acceptor (L10-mVenus) will only be near the nLuc-tagged PI4KA BRET-donor, and thereby efficiently increase the relative BRET signal, if co-assembled with the requisite molecular partners for PM targeting. (B) Normalized BRET signal measured from HEK293A cell populations (~0.75 x 10⁵ cells per well) expressing a fixed amount of the PM-PI4KA^BRET biosensor (L10-mVenus-tPT2A-NLuc-PI4KA; 500 ng per well) together with increasing amounts (100, 500, or 1000 ng per well) of constructs consisting of either an empty vector (black), WR control (red; EFR3B-P2A-TTC7B-T2A-FAM126A), or the indicated mutants of the PI4KA complex components [blue, EFR3B ΔC-term(1 to 716)-P2A-TTC7B-T2A-FAM126A; green, EFR3B^L726A,F751A-P2A-TTC7B-T2A-FAM126A; gray, EFR3B-P2A-TTC7B^A702R-T2A-FAM126A; and magenta, EFR3B-P2A-TTC7B-T2A-FAM126A^A103R]. For all treatments, BRET values were normalized relative to an internal basal BRET control (PM-PI4KA^BRET biosensor expressed without the empty vector or EFR3B-P2A-TTC7B-T2A-FAM126A plasmid added) and are presented as the summary of triplicate wells measured for each treatment condition in three independent biological replicates (n = 3, with normalized signals from nine total wells averaged). Normalized BRET signals were then expressed and graphed as fold increases relative to the 100 ng per well pcDNA3.1 control. Error is shown as the SEM across the replicates. (C) Representative confocal images of live HEK293A cells coexpressing a fixed amount (250 ng) of EGFP-PI4KA (grayscale) together with the indicated constructs (500 ng/each). The sequence and color coding of the experimental treatments are matched with the BRET measurements in (B).