Comprehensive identification of PIP3-regulated PH domains from C. elegans to H. sapiens by model prediction and live imaging

Abstract

Phosphoinositide 3-kinase (PI3K) and its product phosphatidylinositol(3,4,5)-trisphosphate (PIP3) control cell growth, migration, and other processes by recruiting proteins with pleckstrin homology (PH) domains and possibly other domains to the plasma membrane (PM). However, previous experimental and structural work with PH domains left conflicting evidence about which ones are PIP3 regulated. Here we used live-cell confocal imaging of 130 YFP-conjugated mouse PH domains and found that 20% translocated to the PM in response to receptor-generated PIP3 production. We developed a recursive-learning algorithm to predict PIP3 regulation of 1200 PH domains from different eukaryotes and validated that it accurately predicts PIP3 regulation. Strikingly, this algorithm showed that PIP3 regulation is specified by amino acids across the PH domain, not just the PIP3-binding pocket, and must have evolved several times independently from PIP3-insensitive ancestral PH domains. Finally, our algorithm and live-cell experiments provide a functional survey of PH domains in different species, showing that PI3K regulation increased from approximately two C. elegans and four Drosophila to 40 vertebrate proteins.

Figure 1. Genome-wide Live-Cell Imaging Identifies a Subset of Mouse PH Domains Regulated by Receptor Stimulation

(A) In the initial set of experiments, 132 YFP-tagged mouse PH domain constructs were transfected into NIH 3T3 cells and imaged. (B) Examples of PH domains that translocate to the PM in response to the addition of PDGF (5 nM final concentration). (C) Examples of PH domains that are constitutively localized to the PM (top row) and that remain cytosolic even after PDGF receptor stimulation (bottom row). Scale bars are 20 µm.

Figure 2. Control Experiments and Characterization of Two Translocators: SH3BP2-PH and FGD6-PH

(A) Inhibition of PI3K using 50 µM LY294002 reversed the PDGF receptor-triggered PM translocation of PH domains in NIH 3T3 cells. Cells were treated with PDGF for 3 min prior to LY294002 treatment. Images were taken 5 min after the treatment with LY294002. See also Figures S9A and S9B. (B) Inhibition of PI3K using a dominant-negative PI3K construct (DN-p85). (C) Time course of SH3BP2-PH domain translocation to the PM in response to sequential addition of PDGF and LY294002. The y-axis is the ratio of cytosolic over PM fluorescence intensity. (D) PDGF receptor-triggered PM translocation of the PH domain in a protein resulted in translocation of the full-length protein. FGD6 and SH3BP2 full-length proteins were conjugated with YFP. Scale bars, 20 µm.

Figure 3. In Vitro Binding Selectivity of PH Domains to PI Lipids and Lipid Controls

In vitro PI-lipid binding specificity of PH domains. Extracts of cells with overexpressed PH domains were added to PIP strips (A) and lipid arrays (B). (C) Constitutive PM targeting and receptor-triggered PM targeting partially correlate with in vitro binding to PI(4,5)P2 and PI(3,4,5)P3, respectively. PH domains with observed receptor-triggered PM translocation are shown in red. PH domains with constitutive, unregulated PM localization are shown in blue, while PH domains that remained constitutively cytosolic are marked in black without labels. A summary of all measured lipid blot binding interactions can be found in Table S2, and Figure S2 contains bar graphs showing selected lipid blot binding values.

Figure 4. Applying a Previously Published PIP3-Binding Motif to Explain the Translocation Data

(A) Sequence comparison of the three variable loops between β strands 1–2, 3–4, and 6–7 of the PIP3-regulated PH domains shows only minimal sequence similarity. Amino acids that have been shown in known structures to interact with the 1, 3, 4, and 5 phosphates of PIP3 are highlighted in red. The motif from Isakoff et al. (1998) is shown in bold black type. (B) Phylogenetic tree with scaled branches showing distinct subclasses of PIP3-binding PH domains.

Figure 5. Development of a Recursive-Learning Strategy for Predicting PIP3 Regulation

(A) Schematic description of the RFC algorithm. (B) Restricting the RFC algorithm to the three variable loop regions between the 1–2 β strands, the 3–4 β strands, and the 6–7 β strands showed only partial predictive capabilities. Rank-ordered values are shown (x-axis) for the RFC score of the 130 initially tested mouse PH domains. Translocating PH domains are shown in red, and non-translocating ones are shown in blue. All tested mouse PH domains are used in the bar diagram. (C) Allowing all amino acid positions in the PH domain to contribute to the RFC score enables exact separation of PIP3-regulated from non-PIP3-regulated PH domains. Same color bar representation as in (B).

Figure 6. Prediction of PIP3-Regulated PH Domains in Other Species

(A) High-scoring predicted human and mouse PH domains translocated to the PM. (B) High-scoring C. elegans and Drosophila PH domains translocated to the PM. (C) Example of PH domains that scored near or below the cutoff score and did not translocate. Distribution of PIP3-regulated PH domains in mouse (D), human and zebrafish (E), and C. elegans and Drosophila (F). The cumulative representations in (D) and (E) show the PIP3 score of each PH domain in a particular species on the x-axis and the number of PH domains that have a specific PIP3 score or higher on the y-axis. This representation enables one to visualize the distribution of the PIP3 scores and number of PH domains in a particular species. The tested PH domains that were PIP3 regulated are marked in red, and those that were not PIP3 regulated are marked in blue. Table S6 lists all S_RFC scores for the different species. Figure S7 shows images of all tested PH-domains used for validation of the RFC algorithm. Figure S8 shows time courses of mammalian translocating PH domains.

Figure 7. Structural and Evolutionary Insights into PIP3-Regulated Proteomes Derived from the RFC Algorithm

(A) Amino acid positions that contribute to PIP3 regulation are distributed across the PH domain, not just the first β strand loop. A two-dimensional plot of the RFC-matrix used to score Figures 6D–6F is plotted at the top. Amino acid positions contributing the most to PIP3 regulation are marked with a star (*) in the bar graph on the bottom. (B) Amino acid positions marked in (A) with a star were overlaid on the crystal structure of AKT1-PH. The headgroup of P1(3,4,5)P3 is shown in red. (C) Number and percentage of PIP3-regulated PH domains in different model organisms. (D) Radial phylogenetic tree with scaled branches showing nonregulated PH domains from different species that are homologs of five PIP3-regulated mammalian PH domains. The first letter before the gene name specifies the species where the gene is from: y, S. cerevisiae; p, S. pombe; c, C. elegans; d, D. melanogaster; z, zebrafish; and h, human. The branches are colored black if the gene is not PIP3 regulated and red if the gene is PIP3 regulated. The green circles mark where PIP3-regulated genes have branched off from a non-PIP3-regulated ancestor.