Oncogenic mutations of PIK3CA lead to increased membrane recruitment driven by reorientation of the ABD, p85 and C-terminus

Jenkins, Meredith L.; Ranga-Prasad, Harish; Parson, Matthew A. H.; Harris, Noah J.; Rathinaswamy, Manoj K.; Burke, John E.

doi:10.1038/s41467-023-35789-6

Download PDF

Article
Open access
Published: 12 January 2023

Oncogenic mutations of PIK3CA lead to increased membrane recruitment driven by reorientation of the ABD, p85 and C-terminus

Meredith L. Jenkins¹,
Harish Ranga-Prasad¹,
Matthew A. H. Parson ORCID: orcid.org/0000-0001-6270-559X¹,
Noah J. Harris¹,
Manoj K. Rathinaswamy¹ &
…
John E. Burke ORCID: orcid.org/0000-0001-7904-9859^1,2

Nature Communications volume 14, Article number: 181 (2023) Cite this article

4229 Accesses
11 Citations
2 Altmetric
Metrics details

Subjects

Abstract

PIK3CA encoding the phosphoinositide 3-kinase (PI3K) p110α catalytic subunit is frequently mutated in cancer, with mutations occurring widely throughout the primary sequence. The full set of mechanisms underlying how PI3Ks are activated by all oncogenic mutations on membranes are unclear. Using a synergy of biochemical assays and hydrogen deuterium exchange mass spectrometry (HDX-MS), we reveal unique regulatory mechanisms underlying PI3K activation. Engagement of p110α on membranes leads to disengagement of the ABD of p110α from the catalytic core, and the C2 domain from the iSH2 domain of the p85 regulatory subunit. PI3K activation also requires reorientation of the p110α C-terminus, with mutations that alter the inhibited conformation of the C-terminus increasing membrane binding. Mutations at the C-terminus (M1043I/L, H1047R, G1049R, and N1068KLKR) activate p110α through distinct mechanisms, with this having important implications for mutant selective inhibitor development. This work reveals unique mechanisms underlying how PI3K is activated by oncogenic mutations, and explains how double mutants can synergistically increase PI3K activity.

Nuclear translocation of p85β promotes tumorigenesis of PIK3CA helical domain mutant cancer

Article Open access 13 April 2022

Yujun Hao, Baoyu He, … Zhenghe Wang

Structural basis of phosphatidylinositol 3-kinase C2α function

Article Open access 07 March 2022

Wen-Ting Lo, Yingyi Zhang, … Volker Haucke

PI3K isoforms in cell signalling and vesicle trafficking

Article 20 May 2019

Benoit Bilanges, York Posor & Bart Vanhaesebroeck

Introduction

Activating mutations in the gene PIK3CA, which encodes the class IA phosphoinositide 3-kinase (PI3K) catalytic subunit p110α are among the most common mutations across human cancers^1,2,3,4. Class I PI3Ks are composed of four isoforms (class IA [p110α, p110β, p110δ] and class IB [p110γ]), which together generate the lipid second messenger phosphatidylinositol 3,4,5 trisphosphate (PIP₃) downstream of multiple cell surface receptors, including receptor tyrosine kinases (RTKs), G-protein coupled receptors (GPCRs) and Ras superfamily GTPases. PIP₃ recruits multiple effectors, including kinases and regulators of G protein signalling which play critical roles in regulating growth, survival, proliferation and metabolism^5,6,7.

PI3Kα is a heterodimer composed of a p110α catalytic subunit and a p85 regulatory subunit, of which there are 5 isoforms: p85α, p55α and p50α (encoded by PIK3R1), p85β (encoded by PIK3R2) and p55γ (encoded by PIK3R3). The class IA regulatory subunits play three key roles in regulating class IA PI3Ks: they stabilise and inhibit the p110 catalytic subunits^8,9, and they allow for activation through the direct engagement of the nSH2 and cSH2 domains of regulatory subunits with pYXXM motifs in RTKs¹⁰. In unstimulated cells, the p110α protein is kept in an inactive and stable cytosolic configuration due to its interactions with the regulatory subunit, with full activation and membrane recruitment of the PI3Kα complex requiring binding to both GTP loaded Ras, and engagement of both the nSH2 and cSH2 of regulatory subunits by bis-phosphorylated pYXXM motifs present in RTKs and their adaptors^11,12,13. However, the full molecular mechanisms underpinning activation and membrane binding of class IA PI3K, and how oncogenic mutants alter this are not completely understood.

Extensive biochemical and biophysical studies have revealed how p110α is inhibited by regulatory subunits^{14,15,16,17,18,19,20}. The p110α catalytic subunit is composed of five domains (an Adaptor Binding Domain (ABD), Ras Binding Domain (RBD), C2, helical and a bi-lobal kinase domain)^14,21 (Fig. 1A). The RBD, C2, helical and kinase domains together form the catalytic core of PI3Ks, which is conserved through the class I, II and III PI3Ks^22,23. The ABD of class IA PI3Ks binds irreversibly to the iSH2 coiled coil domain present in all class IA regulatory subunits¹⁵, and also forms an inhibitory intra-subunit interface with the N-lobe of the kinase domain (Fig. 1A, B). The iSH2 also makes inhibitory contacts with the C2 and activation loop of the kinase domain^16,24. The nSH2 domain of regulatory subunits binds reversibly to the C2, helical and kinase domains^15,17, with these contacts broken when the nSH2 binds to pYXXM motifs¹⁹. The cSH2 of regulatory subunits is strictly required for high affinity association with bis-phosphorylated receptors and their adaptors¹¹, and in p110β and p110δ forms an inhibitory interface with the kinase domain, however, this interaction is either absent or transient in p110α^25,26. Recently Cryo-EM analysis of p110α bound to p85α showed that upon binding to phosphopeptides there appeared to be complete disengagement of the ABD and regulatory subunit from the catalytic core²⁷, which was not fully supported by previous HDX-MS analysis of phosphopeptide binding^16,19. However, disengagement of the catalytic core from the ABD and regulatory subunits is consistent with HDX-MS results studying PI3K membrane binding¹².

**Fig. 1: Structure of p110α/p85α and location of *PIK3CA* oncogenic mutations.**

The PIK3CA and PIK3R1 genes encoding p110α and p85α are both frequently mutated in human cancers^1,2,28. The majority of mutations lead to amino acid substitutions (Fig. 1C)²⁹, although, more complex insertions and deletions also occur³⁰. Oncogenic transformation by p85α mutations is driven by activation of the p110α isoform³¹. For p110α the most frequent mutations are located at two hot spots located at the helical-nSH2 interface (E542K, E545K) and the C-terminus of the kinase domain (H1047R) (Fig. 1B). The E545K disrupts the nSH2-helical interface¹⁵ while H1047R alters the membrane binding surface of the kinase domain, and how the kinase domain packs against the activation loop (Fig. 1D, E)^17,32. However, there also are relatively high-frequency mutations at the ABD-kinase interface, the ABD-RBD linker, the C2-iSH2/C2-nSH2 interfaces, the putative membrane interface of the N-lobe with membranes, and in a region of the kinase domain C-terminal to the activation loop, referred to as the regulatory motif (Fig. 1C)³³. We have previously found that oncogenic mutations mimicked and enhanced conformational changes observed in the catalytic cycle of WT PI3Kα¹⁶, with different mutants showing unique conformational changes. Different PIK3CA mutations activating through unique mechanisms is supported by the discovery of tumours harbouring double PIK3CA mutations in cis, with these tumours showing enhanced sensitivity to PI3K inhibition³⁴.

Here, we show using a combination of hydrogen deuterium exchange mass spectrometry (HDX-MS) and kinase/membrane binding assays a molecular model for how oncogenic mutants in p110α activate both kinase activity and membrane binding. This reveals the critical role of disengagement of the inhibitory contacts of the ABD domain and the p85 regulatory subunit, as well as the importance of reorganisation of the membrane binding C-terminus in PI3K activation. Intriguingly, mutations at the C-terminal tail activate PI3K through distinct molecular mechanisms which provides insight into how this might be utilised for design of mutant selective inhibitors. Overall, this work provides unique insight into the molecular mechanisms mediating PI3K activation by oncogenic mutations.

Results

To investigate the role of the ABD domain/p85 regulatory subunit in controlling PI3K enzyme activity, we needed a construct that allowed us to interrogate the dynamic effects of full ABD disengagement. We engineered and purified the catalytic core of p110α (106–1068, referred to as the p110α core) along with the full-length complex of p110α-p85α (full set of all constructs purified in this manuscript shown in Supplementary Fig. 1, and SDS-page gels in Supplementary Fig. 2). Initial attempts to purify the p110α core construct were unsuccessful, and we were only able to successfully purify this construct when it contained a kinase dead mutation (D915N). To validate that the D915N p110α construct did not cause any significant changes in protein conformation or membrane binding, we carried out HDX-MS experiments on WT and kinase dead p110α-p85α complexes bound to a PDGFR bis-phosphorylated pY peptide (referred to going forward as pY) composed of PDGFR residues 735–767 with phosphorylation present at Y740 and Y751 in the presence and absence of membrane vesicles (5% PIP₂, 30% PS, 65% PE, referred to afterwards as PIP₂/PS/PE). HDX-MS is a technique that measures the exchange rate of amide hydrogens, and as the rate is dependent on the presence and stability of secondary structure, it is an excellent probe of protein conformational dynamics, and we have extensively used it to study PI3K activation^{11,12,16,19,25}. There were no significant changes in conformation in the kinase dead p110α compared to wild type, and both had equivalent membrane binding to lipid membranes (Supplementary Fig. 3), highlighting the suitability of this mutant for membrane binding and HDX-MS studies.

We carried out HDX-MS experiments to identify conformational changes that occur in the catalytic core of p110α upon removal of the ABD and p85α regulatory subunit. We compared H/D exchange differences for both the p110α core and the full-length complex of p110α-p85α. The full details of HDX-MS data processing are in Table S1, with all raw HDX-MS data for all time points available in the source data. For the p110α core construct there was significantly increased exchange (for all HDX-MS experiments this is defined as differences at any time point >5%, >0.4 Da, and a p-value <0.01 for a two-tailed t-test) in the ABD-RBD linker (114–119), C2 domain (347–355 and 444–475), helical domain (524–551), N-lobe of the kinase domain (691–697, 735–744), the activation loop (930–937) and the C-lobe of the kinase domain (1002–1013) (Fig. 2A–D and S4A, B). All significant differences in the catalytic core of p110α were in regions that were either in contact with ABD or p85α. This dataset allowed us to compare changes to those we previously observed upon membrane recruitment by both pY, and membrane bound Ras of the full-length p110α-p85α complex (Fig. 2B)¹². Intriguingly, we find that the region of the ABD-RBD linker in contact with the ABD domain (114–119) had similar increases in exchange between WT p110α-p85α and either the p110α core or upon pY/HRas mediated membrane recruitment (Fig. 2D). There is a small increase in exchange in this region upon pY binding alone, but this was minor compared to the effect upon membrane binding^16,19. The increase in region 114–119 was greater when bound to pY/HRas membranes compared to pY-mediated membrane binding alone, suggesting this increase is dependent on the amount of membrane binding (Fig. 2D). This is also the case for increases in exchange observed upon membrane binding in the N-terminal and C-terminal ends of the iSH2 domain that are in contact with the C2 domain (470–476, 556–570, Supplementary Fig. 4E), showing the clear link between increases in the ABD-RBD linker and at the C2-iSH2 interface.

**Fig. 2: Conformational changes in p110α core compared to full-length p110α- p85α, and comparison to changes upon pY/Ras membrane recruitment.**

This data comparing the full-length heterodimer vs p110α core allowed us to define the effect of ABD removal on the contact site at the ABD-RBD linker. This region still is protected from exchange at early time points, suggesting presence of secondary structure, however, it is much more dynamic in the absence of the ABD. Comparing this to previous HDX-MS experiments examining pY-Ras membrane recruitment of p110α -p85α¹², showed that the exchange rate of the core is similar to the p110α-p85α membrane bound state, suggesting a correlative ABD disengagement occurring with membrane binding. This is supported by our previous observation of increased membrane binding for oncogenic mutants at the C2-iSH2 or ABD interfaces (N345K, G106V and G118D) that would be expected to promote ABD / iSH2 disengagement¹⁶. An important note is that this data does not support complete dissociation of the p110-p85 complex (due to the extremely high affinity interaction of the ABD to the iSH2³⁵), but instead the ABD-p85 becoming mobile relative to the p110α catalytic core.

Enhanced membrane binding of p110α catalytic core

Our hypothesis that disengagement of the ABD and the regulatory subunit p85α subunit is required for membrane binding suggested that there should be differential membrane binding of the p110α core compared to full-length p110α/p85α. We used protein-lipid Fluorescent Resonance Energy Transfer (FRET) assays (Fig. 3A) to compare membrane recruitment of p110α core to full-length p110α-p85α in the presence and absence of 1 μM pY. This assay was carried out on two different lipids: one optimised for maximal PI3K recruitment (5% PIP₂, 10% Dansyl-PS, 25% PS and 60% PE, Fig. 3B), and another roughly mimicking the plasma membrane (5% PIP₂, 10% Dansyl PS, 15% PS, 40% PE, 15% PC, 10% cholesterol and 5% sphingomyelin, Supplementary Fig. 4E). While pY was required for robust binding of full-length p110α-p85α, it was dispensable for p110α core association, which is expected due to the lack of SH2 domains required for pY binding (Fig. 3B). The p110α core showed increased membrane recruitment for both lipid mixtures compared to the pY activated p110α/p85α complex. To determine the role of free p85α in PI3K membrane recruitment, we also purified recombinant free p85α and analysed the protein-lipid FRET signal. There was a weak FRET signal for p85α alone, with only a small change upon pY binding (Fig. 3B and Supplementary Fig. 4F). This signal was significantly lower than the p110α/p85α complex, indicating that the p110α catalytic core is fully capable of membrane binding in the absence of p85α.

**Fig. 3: Enhanced membrane binding of p110α core compared to full-length p110α/p85α, and mapping of the p110α membrane binding interface.**

Defining the membrane binding surface of p110α core

We have extensively characterised the membrane binding of the p110α/p85α complex using HDX-MS, however, the disengagement of the ABD and p85 from the catalytic core has likely complicated the analysis of membrane binding regions. We carried out HDX-MS experiments of p110α core in the presence and absence of 5% PIP₂/PS/PE membranes to fully understand the molecular underpinnings of p110α membrane binding. We observed protection in the ABD-RBD linker (114–119), C2 (343–355), N-lobe kinase domain (713–734,735–744,799–811 and 850–859), Activation loop (930–961) and the C-terminus of the C-lobe kinase domain (1039–1055 and 1056–1068) (Fig. 3A–C). The largest differences occurred in the C-terminus, and N-lobe, with only minor differences in the C2 domain. However, the region of the C2 domain that interacts with membrane has limited secondary structure (see Supplementary Fig. 4D), which can make tracking transient membrane differences using HDX challenging. Previous HDX-MS experiments testing N345K p110α-p85α binding to membranes showed this same region being protected by membranes¹⁶. Overall, this supports a model where p110α binds membrane at a surface composed of the C2 domain, the kα1–kα2 helices (720–744) and the 859–872 region of the N-lobe, the activation loop, along with the C-terminal tail. Intriguingly, the kα1–kα2 helices in the N-lobe interact with the ABD, and the 343–355 region of the C2 domain binds the N + C termini of the iSH2 domain, which provides a putative molecular explanation for why disengagement of the ABD and p85 leads to increased membrane association.

When comparing our data to the full set of missense oncogenic mutations in the ABD, ABD-RBD linker, C2, helical and the N-lobe of the kinase domain we find that all mutations found in >30 tumours except one (E726K) are located at either the ABD or p85 interfaces. We had previously observed that mutations in the ABD-RBD linker caused similar conformational changes to those located at the C2-iSH2 interface, with this being explained by both mutations leading to disengagement of the ABD and p85 from the catalytic core¹⁶.

Biochemical analysis of oncogenic mutations in the inhibitory C-terminus

While the disengagement of the ABD and p85 being involved in membrane binding provides a molecular rationale for activation by oncogenic mutations in the ABD, C2, and helical domains, it does not fully explain the molecular mechanism of mutations in the C-lobe of the kinase domain. Our previous HDX-MS analysis of the kinase domain mutant, H1047R, showed increased exposure throughout regions of the N-lobe and C-lobe of the kinase domain, with this resulting in a dramatic increase in membrane binding¹⁶. In high-resolution structures the C-terminus of p110α makes extensive contacts with helices that make up the regulatory motif (kα8 and kα11) and sits on top of the C-terminus of the activation loop (Fig. 1E)^36,37,38,39. One of the primary interactions is an extensive interface between Trp1051 of the C-terminus and His1047 as well as hydrophobic residues lining kα8 and kα11 of the kinase domain (Met1043, Phe977,980 and 1039). This region is also in contact with Leu956 in the C-terminal end of the activation loop. This orientation positions the membrane binding WIF motif (residues 1057–1059) distant from the membrane binding interface. When comparing these WT p110α structures to the H1047R crystal structure (3HHM), there is a significant reorganisation of the C-terminus, with all contacts between the C-terminus and kα8 and kα11 disrupted, and the WIF motif oriented at the putative membrane surface (Fig. 1D, E)¹⁷. However, many features of this structure are likely artefactual, and driven by crystal packing, as recent Cryo-EM structures did not observe this difference³².

To understand the regulatory mechanisms underlying the inhibitory interface with the C-terminus we analysed the most frequent oncogenic mutants that occur at or near this interface. While H1047R/L is the most frequent mutation (>6000 from the COSMIC database), there are multiple frequent missense mutations in this region, including M1043L/I/V (>300), G1049R/S (>150), and N1044K/S (>80). In addition, we analysed an activating frameshift variant that alters the C-terminus by replacing the terminal N with a KLKR extension, which was recently identified in tumour samples (N1068KLKR, referred to afterwards as N1068fs)³⁰. To understand if these mutants were activated in a similar way to H1047R, we purified the four p110α mutant complexes (M1043L, H1047R, G1049R, and N1068fs) all bound to full-length p85α.

We characterised the intrinsic ATPase activity of each p110α mutant (Fig. 4A + B), and while this assay does not measure biologically relevant PIP₃ activity, it can measure intrinsic differences in PI3K activity independent of membrane binding. H1047R, G1049R and M1043L had significantly increased ATPase activity compared to WT (Fig. 4A + B). However, the N1068fs mutants showed no increase in ATPase activity compared to WT. This suggested a possible conformational difference between M1043L, H1047R and G1049R compared to N1068fs.

**Fig. 4: Biochemical analysis of C-terminal *PIK3CA* mutations and their effect on membrane binding and ATPase assays.**

For these mutants, we had difficulty in obtaining sufficient yield of the proteins for extensive biophysical analysis. To circumvent this, we used the kinase dead variants to characterise their membrane binding using protein-lipid FRET using both PM mimic and optimised binding lipids. Membrane binding was enhanced for mutants upon addition of pY, with greater binding for the optimised lipids over the PM mimic (Supplementary Fig. 5D). H1047R, G1049R and N1068fs all showed significantly increased membrane binding over WT both with and without pY (Fig. 4C), while the M1043L mutant only showed a slight increase in membrane binding without pY for the PM mimic vesicles (Fig. 4C). This clustered the mutants into three groups, H1047R and G1049R which increased basal ATPase and membrane binding, M1043L which increased basal ATPase with only a limited effect on membrane binding, and N1068fs which did not alter basal ATPase, but did increase membrane binding.

Conformational changes in oncogenic mutations C-terminus of the kinase domain

To test if C-terminal mutations worked by disrupting the inhibitory interaction with the C-terminus, we carried out HDX-MS studies on six constructs of full-length p110α (WT, M1043L, H1047R, G1049R, N1068fs, and a construct lacking the C-terminus (1–1048, referred to afterwards as ΔCter), all bound to full-length p85α. The use of the ΔCter construct allowed for a direct comparison of removal of the C-terminus, versus a possible reorientation upon oncogenic mutation. We used kinase active M1043L, H1047R, G1049R, and a kinase dead variant of N1068fs for this comparison.

HDX-MS experiments were carried out for 4–5 timepoints of exchange (3 s at 1 °C, 3, 30, 300, and 3000 s at 20 °C) for each complex. The full set of all peptides analysed for both p110α and p85α are shown in the Source data, with peptide exchange data presented in Supplementary Fig. 5. The changes observed for H1047R matched almost exactly our previous studies¹⁶, but this experiment expanded the HDX time course, allowing for a more in-depth analysis of perturbations in conformation.

The H1047R, G1049R, and the ΔCter constructs showed similar significant increases compared to the WT in the kinase domain (Fig. 5A–C). These included regions covering 850–858 (hinge between the N and C lobes), the activation loop (930–956) and helices spanning the regulatory arch (1014–1021 in kα10, 1021–1038 in kα11; Fig. 5A–C). Many of these regions correspond directly to the contact site between the C-terminus and kα11. This validates the inhibitory interaction observed in the 4OVU crystal structure, with this interface stabilising helix kα11 and the activation loop. One region in kα8 (962–980) showed a significant change between H1047R, G1049R and WT, which was not observed comparing WT and ΔCter. This helix contacts both the C-terminus and kα11. A possible mechanism explaining this difference is the rotation that occurs in kα11 upon mutation of H1047R. Structures of free p110α lacking the C-terminus do not show this same rotation of the kα11 helix. The rotation of kα11 would alter the interface between kα8 and kα11, leading to increased exchange in kα8. This could be driven by a unique orientation of the C-terminus packing against kα7 in the H1047R or G1049R mutant, which would be lost in the absence of the C-terminus as would occur in the ΔCter protein. Overall, this data suggests that H1047R and G1049R lead to activation through disruption of the inhibitory conformation of the C-terminal tail, which reorients the lipid binding WIF motif at the membrane binding surface. Both M1043L and N1068fs showed no significant differences compared to WT, suggesting these mutants do not disrupt the inhibitory C-terminal conformation (Fig. 5D + E).

**Fig. 5: Structural difference between various c-terminal mutants of p110α compared to WT p110α/p85α, and mapping of the p110α membrane binding interface.**

We also compared HDX-MS differences in full-length p110α-p85α between WT, H1047R and ΔC in the presence and absence of pY (Supplementary Fig. 6). The binding of pY led to significant increases for all three constructs at interfaces that have been previously described (ABD-RBD linker, C2-iSH2, nSH2-helical). However, intriguingly there were unique differences upon pY binding for the ΔCter and H1047R constructs at the interface of the regulatory arch with the nSH2 domain (1014–1021, kα10). This is intriguing as it suggests that the opening of this portion of the regulatory motif only occurs upon both disengagement of the nSH2 domain (mediated by pY binding) and disengagement of the inhibitory C-terminus (Supplementary Fig. 6).

Discussion

Understanding how p110α is regulated and how oncogenic PIK3CA mutants alter this regulation is essential in the development of novel PI3K therapeutics. Due to the critical homeostatic roles that WT p110α plays in growth, development and metabolism⁴⁰, as well as the cell intrinsic^41,42 and systematic⁴³ negative feedback loops that oppose PI3K pharmacological inhibition, there are extensive advantages to selectively targeting mutant p110α over WT p110α. Therefore, defining the exact molecular basis underlying mutant specific conformational changes may reveal opportunities for mutant-selective drug design. Oncogenic mutants in p110α span multiple domains of the catalytic subunit, with foci occurring at the ABD, ABD-RBD linker, C2-iSH2 interface, nSH2-helical interface, and the N + C lobes of the kinase domain⁴⁴. While most mis-sense oncogenic mutations discovered in tumours are at hot-spot locations (E542K, E545K and H1047R), more than 25% of all mutations in PIK3CA occur outside the hotspots. Extensive biochemical¹⁶ and cellular experiments^45,46 have defined that PIK3CA mutations activate lipid kinase activity by different mechanisms. This together with the recent discovery that multiple mutations in PIK3CA occur in cis³⁴, where hotspot and rare mutations are found in the same gene, highlights the need to fully understand how multiple oncogenic mutations can synergically activate PI3K. Here using a suite of biochemical and biophysical approaches we propose a unifying molecular model for how all PIK3CA mutations can lead to increased kinase activity, as well as how different oncogenic mutations in cis can increase oncogenicity.

We wanted to define how extensive oncogenic gain of function PIK3CA mutations at both the ABD interface and the C2-iSH2 interface activate PI3K. The critical role of the ABD in regulating p110α kinase activity was originally suggested by its direct interaction with the N-lobe of the kinase domain^15,47. We previously observed exposure occurring at the ABD interface (ABD-RBD linker) upon phosphopeptide pY binding^16,19, and more extensive exposure at the ABD interface upon membrane binding^12,16. Cryo EM analysis of p110α with p85α revealed disengagement of the ABD and p85 subunit from the catalytic core of p110α upon binding phosphorylated peptide²⁷. To fully define the relative role of ABD disengagement in PI3K activation on membranes we generated the p110α catalytic core and directly compared its exchange to the full-length p110α-p85α. Increased exposure was seen at all contact interfaces with both p85 and the ABD. Intriguingly, we found that the p110α catalytic core (p110α core) showed H/D exchange rates at the ABD-RBD linker that almost exactly matched those observed upon Ras and pY mediated membrane binding of WT p110α-p85α¹². Importantly, we did not observe similar rates when comparing phosphopeptide bound p110α-p85α, suggesting that complete ABD disengagement only occurred when p110α is membrane bound which is in contrast from what was suggested from the cryo-EM model of PI3Kα in the presence of pY peptides.

We previously observed that mutations at the ABD interface exposed the C2-iSH2 interface, and vice versa¹⁶. As the ABD will always remain tightly bound to the iSH2 domain, its disengagement would likely occur simultaneously with disruption of the C2-iSH2 interface. For these mutants, full disruption of the ABD and C2-iSH2 interfaces occurred only upon removal of the nSH2 (through pY binding), with no requirement for membrane. This reveals an unexpected inhibitory role of the nSH2 domain, whereby packing against the helical, kinase, and C2 domains of p110α and the iSH2 of p85α, it can stabilise the interface with the iSH2 domain, which prevents ABD disengagement from the catalytic core of p110α. Intriguingly, ABD disengagement is likely involved in membrane recruitment for all class IA PI3Ks, as HDX-MS experiments on membranes showed exposure in the iSH2 and the ABD-RBD linker of p110β⁴⁸ and p110δ^25,49. These changes were also seen in primary immunodeficiency mutations of p110δ⁵⁰ and in activating PIK3R1 truncations at the N + C termini of the iSH2 involved in immunodeficiencies and oncogenic transformation^11,49. An interesting implication in this model is that unique class IA regulatory subunits may have distinctive propensities for ABD-p85 disengagement, which may partially explain a differential role for regulatory subunits in oncogenic transformation⁵¹. This will require further study to investigate regulatory subunit isoform differences in disengagement. Together this highlights the key role of ABD disengagement in PI3K activation, and how oncogenic mutants can alter this regulation.

We next wanted to understand how ABD and p85 disengagement is involved in membrane binding of PI3K. We hypothesised that the free catalytic core of p110α may more efficiently bind to lipid membranes. Our protein-lipid FRET experiments with the p110α core showed enhanced membrane binding compared to the pY activated p110α- p85α heterodimer, which was similar to our previous studies on the catalytic core construct of p110δ²⁵. To understand the mechanism for how this occurs we mapped the membrane binding interface of p110α core using HDX-MS. We observed decreased exchange at the C2 and the N-lobe and C-Lobe of the kinase domain. Many of these changes were not observed in the full-length heterodimer upon binding membranes, with these regions located at either the ABD (kα1–kα2 helices of N-lobe) or iSH2 (C2 domain) interfaces. This suggests that some of the membrane binding regions of p110α are shielded by the ABD or p85 subunit, and disengagement of these regions allows for efficient membrane binding. This model also suggests that the role of phosphopeptide binding in membrane recruitment is two-fold as it both breaks an inhibitory contact between the nSH2 and the regulatory motif of the kinase domain, and breaks the nSH2 contact with the helical domain, which weakens the interface of the ABD and p85 subunits with the catalytic core, exposing membrane binding surfaces of p110α (Fig. 6A).

**Fig. 6: Molecular mechanism of activation of WT PIK3CA, and how oncogenic mutations mimic this process.**

One of the largest decreases in exchange observed upon membrane binding was in the C-terminus of p110α, specifically at the membrane binding WIF motif. In high resolution structures of inhibited p110α, the C-terminus is pointed away from the putative membrane binding surface³⁶, with the C-terminus pointed towards the membrane surface in the X-ray structure of H1047R p110α¹⁷ due to a ~180 rotation at the end of the kα11 helix. The C-terminus in the WT structures makes a set of contacts the regulatory motif helices (kα8- kα11) located after the activation loop and contacts the C-terminus of the activation loop. The regulatory motif helices and the C-terminus encompass the activation loop and make extensive contacts proposed to maintain an inactive conformation. The regulatory motif is also directly in contact with the nSH2 domain of regulatory subunits. Sequestering the C-terminus in a inhibited conformation is a conserved aspect of regulation across all class I PI3Ks, with the inhibited conformation of the C-terminus of p110β and p110δ binding to the cSH2 of regulatory subunits^25,26, and p110γ having its C-terminus inhibited by an inhibitory Tryptophan lock which does not require regulatory subunits³³. Removal of the C-terminus causes a substantial rearrangement of the activation loop, and a rotation of the kα11 helix⁵². Most of the frequent mutations in the C-terminus of p110α would be expected to disrupt this inhibitory contact, either through steric hindrance (M1043I/L, H1047R, G1049R) or disruption of key hydrogen bonds (N1044S). One of the only frequent oncogenic mutations that is not at either the regulatory C-terminus or an interface with the ABD or p85 is located at the loop between kα1–kα2 in the N-lobe (E726K) in a location at the membrane binding surface. This mutant likely drives increased lipid kinase activity through enhanced membrane binding of negatively charged lipids through this charge reversal mutant.

To test if conformational changes in H1047R were caused by disruption of the inhibitory contacts with the C-terminus, we compared the H/D exchange rates of H1047R and a deletion of the C-terminus to the WT protein. The H1047R and C-terminal deletion showed similar increases in exchange in the kα11 helix in contact with the inhibited C-terminus. This suggests a key role of the H1047R mutant is to disrupt inhibitory contacts between the C-terminus and orient the C-terminus in a productive conformation for membrane binding. The H1047R and G1049R, mutants also led to increased activity towards ATP hydrolysis. Therefore, we propose that mutations that disrupt this inhibitory C-terminal contact (H1047R, G1049R) activate PI3K by two unique mechanisms: they reorient the WIF motif towards the membrane binding surface, increasing membrane binding, while also causing the activation loop to adapt a catalytically competent conformation.

There was no increase in ATPase activity for N1068fs compared to wild-type, but it did lead to increased membrane binding. A putative mechanism explaining this is that the 1068 fs replaces the terminal Asn, with a set of positive and hydrophobic residues (KLKR), which all would contribute enhanced membrane binding. So, while the 1068fs mutant does not reorient the membrane binding interface, it likely interacts more extensively, increasing membrane residency time, and increasing PIP₃ production. The M1043L mutant has a much smaller effect on membrane binding but did increase basal ATPase rate. Comparing the structure of WT p110α to H1047R p110α shows a rotation of the kα11 helix in the mutant, with the M1043 residue rotating into where the W1051 from the C-terminus is located in the WT structure¹⁷. We expect that reorientation of the C-terminus is required for full activity, and therefore we propose that the enhanced membrane binding of M1043L (and potentially M1043I/V) is driven by the added steric bulk of the branched chain amino acids, which may more efficiently reorient the C-terminus upon membrane binding. Further structural studies of the M1043 mutants on membrane surfaces will be required to fully access the molecular basis of activation.

Overall, integrating our experimental efforts with the extensive previous biophysical and biochemical analysis of p110α mutants we can propose a unifying model for how p110α is inhibited by regulatory subunits, and how it can be activated both by activating partners and oncogenic mutants (Fig. 6B–D). Activation of p110α/p85α by bis-phosphorylated receptors and their adaptors leads to breaking of the inhibitory nSH2 contacts with the p110α subunit. Helical domain mutants (E542K, E545K) mimic this conformational change (red equilibrium arrows between Fig. 6A-B). The removal of the nSH2 partially destabilises the iSH2-C2 and ABD-kinase interfaces, but in the wild-type complex the equilibrium is mainly towards a closed p85-ABD engaged complex (black arrows). The pYXXM binding surface of the nSH2 exposed in helical mutants leads to greatly enhanced recruitment to RTKs and their adaptors⁵³. Oncogenic mutations that occur at the ABD or iSH2 interfaces in p110α (red equilibrium arrows between Fig. 6A–C) shifts the equilibrium towards the disengaged state, leading to increased affinity for membranes. Full membrane recruitment still depends on the opening of the C-terminus of p110α. In the wild-type complex, membrane recruitment requires additional signalling inputs from Ras GTPases. Oncogenic mutants in the kinase domain lead in increased membrane binding either through reorientation of the WIF binding motif and the activation loop (i.e. H1047R, G1049R), or by altering residues at the membrane interface that can more extensively interact with negatively charged membranes (i.e. E726K, 1068 fs) (red equilibrium arrows in Fig. 6D). Together, this model can account for the putative mechanism of activation for >98% of all PIK3CA mutants reported in the COSMIC database, and for why double PIK3CA mutants lead to increased oncogenicity (Supplementary Fig. 7).

Extensive efforts have focused on the development of p110α selective inhibitors. PIK3CA mutant selective inhibitors would likely have major advantages in cancer treatment as they can evade feedback mechanisms that counteract pharmacological inhibition of WT PIK3CA. Mutant selective inhibitors have been discovered that lead to selective degradation of E545K and H1047R over WT⁵⁴, along with recent reports of H1047R selective inhibitors. Our data showing very similar conformational changes in H1047R and G1049R suggest that 1047R/L selective small molecules may also be useful in targeting G1049R/S. There also could be advantages of generating tool compounds that activate PI3K activity, with small molecules that disrupt the ABD or p85 interfaces likely acting as activators. Our findings provide a molecular framework for the future development of the next generation of PI3K modulators as therapeutics.

Methods

Plasmid generation

Genes of interest were inserted into the pFastBac1 vector to allow baculovirus expression in Spodoptera frugiperda (Sf9) cells. The full list of all plasmids and reagents utilised in this manuscript are shown in Table 1. The plasmids containing p110 and free p85α isoforms also expressed N-terminal to the protein a 10X histidine tag, followed by a 2X Strep tag, followed by a Tobacco Etch Virus protease cleavage site. Single substitution mutations (D915N, H1047R, M1043L, G1049R and N1068fs) and p110 truncations (p110α core and ΔC) were generated using site-directed mutagenesis according to published commercial protocols (QuickChange Site-Directed Mutagenesis, Novagen). DNA oligonucleotides spanning the desired region and either containing the altered nucleotides (single substitutions) or lacking the truncated region were ordered (Sigma). PCR reactions were performed on the WT p85α, and PCR purified (Q5 High-Fidelity 2X MasterMix, New England Biosciences #M0492L; QiaQuick PCR Purification Kit, Qiagen #28104). Single colonies were grown overnight and purified using QIAprep Spin Miniprep Kit (Qiagen #27104). Plasmid identity was confirmed by sanger sequencing (Eurofins Genomics and Plasmidsaurus).

Table 1 Source and identifier for reagents/resources used

Full size table

Virus generation and amplification

The plasmids harbouring p110α WT and other variants and p85α were transformed into DH10MultiBac cells (MultiBac, Geneva Biotech) containing the baculovirus viral genome (bacmid) and a helper plasmid expressing transposase to transpose the expression cassette harbouring the gene of interest into the baculovirus genome. Bacmids with successful incorporation of the expression cassette of pFastBac/ pACEBac1 into the viral genome was identified by blue-white screening and were purified from a single white colony using a standard isopropanol-ethanol extraction method. Briefly, colonies were grown overnight (∼16 h) in 3–5 mL 2xYT (BioBasic #SD7019). Cells were pelleted by centrifugation and the pellet was resuspended in 225 μL P1 Buffer (Qiagen MiniPrep Kit, #27106), chemically lysed by the addition of 225 μL Buffer P2, and the lysis reaction was neutralised by addition of 300 μL Buffer N3. Following centrifugation at 21130 rcf and 4 °C (Rotor #5424 R), the supernatant was separated and mixed with 600 μL isopropanol to precipitate the DNA out of solution. Further centrifugation at the same temperature and speed pelleted the bacmid DNA, which was then washed with 500 μL 70% ethanol three times. The Bacmid DNA pellet was then dried for 1 min and re-suspended in 50 μL Buffer EB.

Purified bacmid was then transfected into Sf9 cells. 2 mL of Sf9 cells between 0.3–0.5 × 10⁶ cells/mL were aliquoted into the wells of a six-well plate and allowed to attach, creating a monolayer of cells at ∼70–80% confluency. Transfection reactions were prepared by the addition of 2–10 μg of bacmid DNA to 100 μl 1xPBS and 12 μL polyethyleneimine (PEI) at 1 mg/mL (Polyethyleneimine “Max” MW 40,000, Polysciences #24765, USA) to 100 μL 1xPBS. The bacmid-PBS and the PEI-PBS solutions were mixed, and the reaction occurred for 20–30 min before addition drop-by-drop to an Sf9 monolayer containing well. Transfections were allowed to proceed for 5–7 days before harvesting virus containing supernatant as a PI viral stock.

Viral stocks were amplified by adding P1 viral stock to suspension Sf9 cells between 1–2 × 10⁶ cells/mL at a 1/100 volume ratio. This amplification produces a P2 stage viral stock that can be used in final protein expression. The amplification proceeded for 4–5 days before harvesting, with cell shaking at 120 RPM in a 27 °C shaker (New Brunswick). Harvesting of P2 viral stocks was carried out by centrifuging cell suspensions in 50 mL Falcon tubes at 2281 RCF (Beckman GS-15), collecting the supernatant in a fresh sterile tube, and adding 5–10% inactivated foetal bovine serum (FBS; VWR Canada #97068-085).

Expression and purification of recombinant proteins

All PI3K constructs were purified by expressing the catalytic subunit and the regulatory subunit using the pFASTBAC/ pACEBac1 expression system in Sf9 cells. After expressing the cells at 27 °C (42 h for Kinase active variants and 55 h for Kinase dead), the cells were harvested at 1739 × g at 4 °C using Eppendorf Centrifuge 5810R and the cells were flash frozen using liquid nitrogen and stored in −80 °C.

The frozen pellets were resuspended in lysis buffer containing 20 mM Tris pH 8, 10 mM Imidazole, 100 mM NaCl, 5% glycerol [v/v], 2 mM βME, protease inhibitor [Protease Inhibitor Cocktail Set III, Sigma]) and sonicated for 2 min (15 s on, 15 s off, level 4.0, Misonix sonicator 3000). Triton-X 100 was added to the lysate at a final concentration of 0.1% and then clarified by spinning at 15,366 × g for 45 min (Beckman Coulter JA-20 rotor). The supernatant was loaded onto a 5 ml crude Ni-NTA column (GE healthcare) equilibrated in NiNTA A buffer containing 20 mM Tris pH 8, 100 mM NaCl, 10 mM Imidazole and 5% glycerol [v/v]. The column was washed using high salt buffer containing 20 mM Tris, 1 M NaCl, 10 mM Imidazole, 5% Glycerol [v/v] followed by NiNTA buffer wash (20 mM Tris pH 8, 100 mM NaCl, 21 mM Imidazole and 5% Glycerol). The protein was eluted using 100% NiNTA B buffer (20 mM Tris pH 8, 100 mM NaCl, 200 mM Imidazole and 5% Glycerol). The elute from the nickel column was loaded onto Streptavidin column (GE healthcare) and subjected to buffer wash using Hep A buffer (20 mM Tris pH 8, 100 mM NaCl, 5% Glycerol and 0.5 mM tris(2-carboxyethyl) phosphine [TCEP]). The column was incubated on ice for 3 h in the presence of TEV protease and then eluted by a wash with HEP A buffer. The eluent was loaded onto Q column equilibrated with HEP A buffer. The column was washed with HEP A buffer to remove TEV protease and the final PI3K was eluted by passing Hep B elution buffer (20 mM Tris pH 8, 325 mM NaCl, 5% glycerol [v/v] and 0.5 mM TCEP). The protein was exchanged to the final buffer containing 20 mM HEPES pH 7.5, 100 mM NaCl, 10% Glycerol [v/v] and 0.5 mM TCEP using a desalting column. The protein was concentrated to 1 mg/ml using a 50,000 MWCO Amicon Concentrator (Millipore), flash frozen and stored at −80 °C.

For HDX experiments involving WT, p110α core, H1047R, M1043L, ΔCter and free p85α were further subjected to gel filtration using Superdex^TM 200 10/300 GL Increase from GE healthcare. After gel filtration, the protein was concentrated, aliquoted, frozen and stored at −80 °C.

Lipid vesicle preparation

To measure membrane recruitment using Protein-Lipid FRET two sets of lipid vesicles were prepared: PE/PS/PIP₂ containing 5 % brain PIP₂, 65% egg yolk PE, 25 % brain PS and 10% Dansyl-PS (Avanti, #810225 C) and PM mimic consisting of 5% brain PIP₂, 20% brain PS, 10% Dansyl-PS, 45% egg yolk PE, 15% egg yolk phosphatidylcholine (PC) (Avanti #840051C), 10% cholesterol (Sigma Aldrich, #47127-U) and 5% egg yolk sphingomyelin (Sigma Aldrich, #S0756). To generate vesicles the lipid mixtures were combined in organic solvent. The mixture was then evaporated using a stream of argon gas followed by desiccation under vacuum for 45 min. The lipids were resuspended in a lipid buffer (25 mM HEPES pH 7, 100 mM NaCl, 10% Glycerol [v/v]) and the solution was subjected to sonication for 15 mins. The vesicles were subjected to five freeze thaw cycles and extruded 11 times through a 100 nm filter (T and T Scientific: TT-002–0010). The extruded vesicles were sonicated again for 5 min, aliquoted and stored at −80 °C.

Protein–lipid FRET assay

Protein-lipid FRET experiments were carried out either at a saturating protein concentration (Fig. 4C, also Supplementary Fig. 4E, F) or as a dose response with PI3K (Supplementary Fig. 3B). Protein lipid FRET assays with saturating PI3K were initiated by mixing 2.5 µL PI3K (final concentration of 1 µM) with 2.5 µL of pY peptide diluted in protein buffer [20 mM HEPES pH 7.5, 100 mM NaCl, 10 % Glycerol [v/v]] (final concentration of 1 µM) for 15 mins at 20 °C. 5 µL of lipid vesicles (either PIP₂/PS/PE or PM mimic vesicles, both at final conc. 16.65 µg/mL) was added to the protein-pY mixture and were incubated for 15 mins at 20 °C. Dose experiment was carried out exactly the same, except PI3K amount varied from 0.015 µM to 1 µM. The plate was then read using a SpectraMax M5 plate reader using a 280-nm excitation filter with 350-nm and 520- nm emission filters to measure Trp and Dansyl-PS FRET emissions, respectively. The FRET signal shown in the figure has I-I_o along the Y axis where I is the intensity of 520 with protein and Io is the intensity of lipid alone.

ATPase assay

All ATPase assays used the Transcreener ADP2 Fluorescence Intensity (FI) assay (Bellbrook labs) which measures formation of ADP. In all, 2 μL of a kinase solution (final concentration 200nM-0.186 nM) at 2X final concentration was mixed with 2 μL substrate solution containing ATP, and the reaction was allowed to proceed for 60 min at 37 °C. The reaction was stopped with 4 μL of 2X stop and detect solution containing EDTA (chelates Mg²⁺, stopping kinase activity) along with 8 nM ADP Alexa Fluor 594 Tracer and 93.7 μg/mL ADP2 Antibody IRDye QC-1 (Bellbrooks lab kit, 3013-1 K), and was allowed to incubate for 60 min. The fluorescence intensity was measured using a SpectraMax M5 plate reader at excitation 590 nm and emission 620 nm. This data was normalised against a 0–100% ADP window made using conditions containing a final concentration of 100 μM ATP or ADP. % ATP turnover was interpolated from an ATP standard curve obtained from performing the assay on 100 μM (total) ATP/ADP mixtures with increasing concentrations of ADP.

HDX-MS analysis: sample preparation

HDX-MS experiments for all conditions were conducted as follows: reactions comparing full-length p110α/p85α and p110α core were conducted in 13.5-μl reaction volumes with a final PI3K amount of 10 pmol. Prior to HD exchange, 3.5 μl of either protein was incubated with 3.5 μl of PIP₂/PS/PE lipid (5 % brain PIP₂, 65% egg yolk PE, 30 % brain PS) at 2.5 mg/ml final or lipid buffer for 2 min at room temperature. To initiate HD exchange, a mixture of either 6.5 μl of the same lipid at 0.76 mg/ml final or lipid buffer and 36.5 μl of D₂O buffer [20 mM HEPES (pH 7.5), 100 mM NaCl, 94.3% D₂O (v/v)] was added to the protein-lipid/buffer mix (final D₂O concentration of 69%). Exchange was carried out for 3, 30 and 300 s at 20 °C.

HDX reactions comparing full-length p110α/p85α WT against M1043L, H1047R and ΔC were conducted in 20 μl reaction volumes with a final PI3K amount of 9 pmol. Prior to HD exchange, 3 μl of protein was incubated with 1 μl of 50 μM pY peptide or protein buffer and allowed to incubate for 15 min on ice. Exchange was initiated by the addition of 16 μl of D₂O buffer to the protein + /‒ pY mixture (final D₂O concentration of 75%). Exchange was carried out for 3, 30, 300 and 3000 s at 20 °C and 0.3 s (3 s on ice).

HDX reactions comparing full-length p110α/p85α WT against G1049R were conducted in 20 μl reaction volumes with a final PI3K amount of 15 pmol. HD exchange was initiated by the addition of 16 μl of D₂O buffer to 4 μl of protein (final D₂O concentration of 75%). The reaction proceeded for 3, 30, 300 and 3000 s at 20 °C.

HDX reactions comparing full-length p110α/p85α WT against N1068fs were conducted in 20 μl reaction volumes with a final PI3K amount of 11 pmol. HD exchange was initiated by the addition of 17.2 μl of D₂O buffer to 2.8 μl of protein (final D₂O concentration of 81%). The reaction proceeded for 3, 30, 300 and 3000 s at 20 °C.

HDX reactions comparing full-length p110α/p85α WT kinase active vs kinase dead were conducted in 50 μl reaction volumes with a final PI3K amount of 10 pmol. Protein was then incubated with 1 μM pY peptide and allowed to incubate for 15 min on ice. Prior to HD exchange, 4 μl of either protein was incubated with 10 μl of PIP₂/PS/PE lipid (5% brain PIP₂, 65% egg yolk PE, 30 % brain PS) at 1 mg/ml or lipid buffer for 2 min at 20 °C. HD exchange was initiated by the addition of 36 μl of D₂O buffer (final D₂O concentration of 69%). The reaction proceeded for 3, and 300 s at 20 °C. All conditions and timepoints were generated in independent triplicate. All exchange reactions were terminated by the addition of ice-cold quench buffer to give a final concentration 0.6 M guanidine-HCl and 0.9% formic acid. Samples were flash frozen in liquid nitrogen immediately after quenching and stored at −80 °C until injected onto the ultra-performance liquid chromatography (UPLC) system for proteolytic cleavage, peptide separation, and injection onto a QTOF for mass analysis, described below.

HDX-MS analysis: protein digestion and tandem MS data collection

Protein samples were rapidly thawed and injected onto an integrated fluidics system containing an HDx-3 PAL liquid handling robot and climate-controlled (2 °C) chromatography system (LEAP Technologies), a Dionex Ultimate 3000 UHPLC system, and an Impact HD QTOF mass spectrometer (Bruker). The protein was run over one immobilised pepsin column (Trajan; ProDx protease column, 2.1 mm × 30 mm PDX.PP01-F32) at 200 μl/min for 3 min at 10 °C. The resulting peptides were collected and desalted on a C18 trap column [Acquity UPLC BEH C18 1.7 mm column (2.1 × 5 mm); Waters 186003975]. The trap was subsequently eluted in line with an ACQUITY 1.7 µm particle, 100 × 1 mm² C18 UPLC column (Waters 186002352), using a gradient of 3–35% B (buffer A, 0.1% formic acid; buffer B, 100% acetonitrile) over 11 min immediately followed by a gradient of 35–80% B over 5 min. MS experiments acquired over a mass range from 150 to 2200 mass/charge ratio (m/z) using an electrospray ionisation source operated at a temperature of 200 °C and a spray voltage of 4.5 kV.

HDX-MS analysis: peptide identification

Peptides were identified from the nondeuterated samples of p110α/p85α complex for WT and other mutants using data-dependent acquisition following tandem MS (MS/MS) experiments (0.5-s precursor scan from 150 to 2000 m/z: 12 0.25 s fragment scans from 150 to 2000 m/z). MS/MS datasets were analysed using PEAKS7 (PEAKS), and peptide identification was carried out by using a false discovery–based approach, with a threshold set to 1% using a database of purified proteins and known contaminants found in SF9 cells⁵⁵. The search parameters were set with a precursor tolerance of 20 parts per million, fragment mass error 0.02 Da, and charge states from 1 to 8.

HDX-MS analysis: mass analysis of peptide centroids and measurement of deuterium incorporation

HD-Examiner Software (Sierra Analytics) was used to automatically calculate the level of deuterium incorporation into each peptide. All peptides were manually inspected for correct charge state, correct retention time, and appropriate selection of isotopic distribution. Deuteration levels were calculated using the centroid of the experimental isotope clusters. HDX-MS results are presented with no correction for back exchange shown in the Source data, with the only correction being applied correcting for the deuterium oxide percentage of the buffer used in the exchange (69% for p110α core experiments, 69% for WT vs M1043L, H1047R and ΔC, 75% for WT vs G1049R, 81% for WT vs N1068fs and 69% for WT kinase active vs kinase dead). Attempts to generate a fully deuterated class I PI3K sample were unsuccessful, which is common for large macromolecular complexes. Therefore, all deuterium exchange values are relative.

Differences in exchange in a peptide were considered significant if they met all three of the following criteria: ≥5% change in exchange, ≥0.4 Da difference in exchange, and a P-value of <0.01 using a two-tailed Student’s t test. The raw HDX data are shown in two different formats.

The raw data for all analysed peptides is available in the source data. The differences in deuterium exchange are visualised in different ways. To allow for visualisation of differences across conditions, we used number of deuteron difference (#D) plots (Figs. 2C, 3D, 5A–E, Supplemental Fig. 3C–E, Supplemental Fig. 4A + C, Supplemental Fig. 5A, and Supplemental Fig. 6D–F). These plots show the total difference in deuterium incorporation over the entire HDX time course, with each point indicating a single peptide. These graphs are calculated by summing the differences at every time point for each peptide and propagating the error. For a selection of peptides, we are showing the %D incorporation over a time course, which allows for comparison of multiple conditions at the same time for a given region (Figs. 2D + 3E, Supplemental Fig. 3F, Supplemental Fig.4B + D, Supplemental Fig.5B + C and Supplemental Fig.6G). Samples were only compared when they were set at the same time and were never compared to experiments completed with a different final D₂O level. The data analysis statistics for all HDX-MS experiments are in Supplemental Table 1a–d according to published guidelines. The HDX-MS proteomics data generated in this study have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository⁵⁶ with the dataset identifier PXD031080.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability

The data that support this study are available from the corresponding authors upon reasonable request. The HDX-MS proteomics data generated in this study have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository⁵⁶ with the dataset identifier PXD031080. All data generated or analysed during this study are included in the Source Data file. Structures used are available at the following accession codes: 4OVU [https://doi.org/10.2210/pdb7VF9/pdb] and 3HHM [https://doi.org/10.2210/pdb7VF9/pdb]. Source data are provided with this paper.

References

Samuels, Y. et al. High frequency of mutations of the PIK3CA gene in human cancers. Science 304, 554 (2004).
Article CAS Google Scholar
Lawrence, M. S. et al. Discovery and saturation analysis of cancer genes across 21 tumour types. Nature 505, 495–501 (2014).
Article ADS CAS Google Scholar
Bailey, M. H. et al. Comprehensive characterization of cancer driver genes and mutations. Cell 173, 371–385.e18 (2018).
Article CAS Google Scholar
Goncalves, M. D., Hopkins, B. D. & Cantley, L. C. Phosphatidylinositol 3-kinase, growth disorders, and cancer. N. Engl. J. Med. 379, 2052–2062 (2018).
Article CAS Google Scholar
Burke, J. E. Structural basis for regulation of phosphoinositide kinases and their involvement in human disease. Mol. Cell 71, 653–673 (2018).
Article CAS Google Scholar
Burke, J. E. & Williams, R. L. Synergy in activating class I PI3Ks. Trends Biochem. Sci. 40, 88–100 (2015).
Article CAS Google Scholar
Madsen, R. R. & Vanhaesebroeck, B. Cracking the context-specific PI3K signaling code. Sci. Signal 13, eaay2940 (2020).
Article CAS Google Scholar
Yu, J. et al. Regulation of the p85/p110 phosphatidylinositol 3’-kinase: stabilization and inhibition of the p110alpha catalytic subunit by the p85 regulatory subunit. Mol. Cell. Biol. 18, 1379–1387 (1998).
Article CAS Google Scholar
Yu, J., Wjasow, C. & Backer, J. M. Regulation of the p85/p110alpha phosphatidylinositol 3’-kinase. Distinct roles for the n-terminal and c-terminal SH2 domains. J. Biol. Chem. 273, 30199–30203 (1998).
Article CAS Google Scholar
Carpenter, C. L. et al. Phosphoinositide 3-kinase is activated by phosphopeptides that bind to the SH2 domains of the 85-kDa subunit. J. Biol. Chem. 268, 9478–9483 (1993).
Article CAS Google Scholar
Dornan, G. L. et al. Defining how oncogenic and developmental mutations of PIK3R1 alter the regulation of class IA phosphoinositide 3-kinases. Structure 28, 145–156.e5 (2020).
Article CAS Google Scholar
Siempelkamp, B. D., Rathinaswamy, M. K., Jenkins, M. L. & Burke, J. E. Molecular mechanism of activation of class IA phosphoinositide 3-kinases (PI3Ks) by membrane-localized HRas. J. Biol. Chem. 292, 12256–12266 (2017).
Article CAS Google Scholar
Buckles, T. C., Ziemba, B. P., Masson, G. R., Williams, R. L. & Falke, J. J. Single-molecule study reveals how receptor and Ras synergistically activate PI3Kα and PIP3 signaling. Biophys. J. 113, 2396–2405 (2017).
Article ADS CAS Google Scholar
Huang, C.-H. et al. The structure of a human p110alpha/p85alpha complex elucidates the effects of oncogenic PI3Kalpha mutations. Science 318, 1744–1748 (2007).
Article ADS CAS Google Scholar
Miled, N. et al. Mechanism of two classes of cancer mutations in the phosphoinositide 3-kinase catalytic subunit. Science 317, 239–242 (2007).
Article ADS CAS Google Scholar
Burke, J. E., Perisic, O., Masson, G. R., Vadas, O. & Williams, R. L. Oncogenic mutations mimic and enhance dynamic events in the natural activation of phosphoinositide 3-kinase p110α (PIK3CA). Proc. Natl. Acad. Sci. USA 109, 15259–15264 (2012).
Article ADS CAS Google Scholar
Mandelker, D. et al. A frequent kinase domain mutation that changes the interaction between PI3Kalpha and the membrane. Proc. Natl. Acad. Sci. USA 106, 16996–17001 (2009).
Article ADS CAS Google Scholar
Chakrabarti, M., Gabelli, S. B. & Amzel, L. M. Allosteric activation of PI3Kα results in dynamic access to catalytically competent conformations. Structure 28, 465–474.e5 (2020).
Article CAS Google Scholar
Burke, J. E. & Williams, R. L. Dynamic steps in receptor tyrosine kinase mediated activation of class IA phosphoinositide 3-kinases (PI3K) captured by H/D exchange (HDX-MS). Adv. Biol. Regul. 53, 97–110 (2013).
Article CAS Google Scholar
Gkeka, P. et al. Investigating the structure and dynamics of the PIK3CA wild-type and H1047R oncogenic mutant. PLoS Comput. Biol. 10, e1003895 (2014).
Article Google Scholar
Walker, E. H., Perisic, O., Ried, C., Stephens, L. & Williams, R. L. Structural insights into phosphoinositide 3-kinase catalysis and signalling. Nature 402, 313–320 (1999).
Article ADS CAS Google Scholar
Rostislavleva, K. et al. Structure and flexibility of the endosomal Vps34 complex reveals the basis of its function on membranes. Science 350, aac7365 (2015).
Article Google Scholar
Rathinaswamy, M. K. et al. Structure of the phosphoinositide 3-kinase (PI3K) p110γ-p101 complex reveals molecular mechanism of GPCR activation. Sci. Adv. 7, eabj4282 (2021).
Article ADS CAS Google Scholar
Wu, H. et al. Regulation of class IA PI 3-kinases: C2 domain-iSH2 domain contacts inhibit p85/p110alpha and are disrupted in oncogenic p85 mutants. Proc. Natl. Acad. Sci. USA 106, 20258–20263 (2009).
Article ADS CAS Google Scholar
Burke, J. E. et al. Dynamics of the phosphoinositide 3-kinase p110δ interaction with p85α and membranes reveals aspects of regulation distinct from p110α. Structure 19, 1127–1137 (2011).
Article CAS Google Scholar
Zhang, X. et al. Structure of lipid kinase p110β/p85β elucidates an unusual SH2-domain-mediated inhibitory mechanism. Mol. Cell 41, 567–578 (2011).
Article CAS Google Scholar
Liu, X. et al. Cryo-EM structures of PI3Kα reveal conformational changes during inhibition and activation. Proc. Natl. Acad. Sci. USA 118, e2109327118 (2021).
Article CAS Google Scholar
Jaiswal, B. S. et al. Somatic mutations in p85α promote tumorigenesis through class IA PI3K activation. Cancer Cell 16, 463–474 (2009).
Article CAS Google Scholar
Dornan, G. L. & Burke, J. E. Molecular mechanisms of human disease mediated by oncogenic and primary immunodeficiency mutations in class IA phosphoinositide 3-kinases. Front. Immunol. 9, 575 (2018).
Article Google Scholar
Spangle, J. M. et al. PIK3CA C-terminal frameshift mutations are novel oncogenic events that sensitize tumors to PI3K-α inhibition. Proc. Natl. Acad. Sci. USA 117, 24427–24433 (2020).
Article ADS CAS Google Scholar
Sun, M., Hillmann, P., Hofmann, B. T., Hart, J. R. & Vogt, P. K. Cancer-derived mutations in the regulatory subunit p85 of phosphoinositide 3-kinase function through the catalytic subunit p110. Proc. Natl. Acad. Sci. USA 107, 15547–15552 (2010).
Article ADS CAS Google Scholar
Liu, X. et al. Cryo-EM structures of cancer-specific helical and kinase domain mutations of PI3Kα. Proc. Natl. Acad. Sci. USA 119, e2215621119 (2022).
Article CAS Google Scholar
Rathinaswamy, M. K. et al. Disease-related mutations in PI3Kγ disrupt regulatory C-terminal dynamics and reveal a path to selective inhibitors. Elife 10, e64691 (2021).
Article CAS Google Scholar
Vasan, N. et al. Double PIK3CA mutations in cis increase oncogenicity and sensitivity to PI3Kα inhibitors. Science 366, 714–723 (2019).
Article ADS CAS Google Scholar
Fry, M. J. et al. Purification and characterization of a phosphatidylinositol 3-kinase complex from bovine brain by using phosphopeptide affinity columns. Biochem. J. 288, 383–393 (1992).
Article CAS Google Scholar
Furet, P. et al. Discovery of NVP-BYL719 a potent and selective phosphatidylinositol-3 kinase alpha inhibitor selected for clinical evaluation. Bioorg. Med. Chem. Lett. 23, 3741–3748 (2013).
Article CAS Google Scholar
Barsanti, P. A. et al. Structure-based drug design of novel potent and selective tetrahydropyrazolo[1,5-a]pyrazines as ATR inhibitors. ACS Med. Chem. Lett. 6, 37–41 (2015).
Article CAS Google Scholar
Han, W. et al. Discovery of imidazo[1,2-a]-pyridine inhibitors of pan-PI3 kinases that are efficacious in a mouse xenograft model. Bioorg. Med. Chem. Lett. 26, 742–746 (2016).
Article CAS Google Scholar
Miller, M. S. et al. Structural basis of nSH2 regulation and lipid binding in PI3Kα. Oncotarget 5, 5198–5208 (2014).
Article Google Scholar
Fruman, D. A. et al. The PI3K pathway in human disease. Cell 170, 605–635 (2017).
Article CAS Google Scholar
Sun, S.-Y. et al. Activation of Akt and eIF4E survival pathways by rapamycin-mediated mammalian target of rapamycin inhibition. Cancer Res. 65, 7052–7058 (2005).
Article CAS Google Scholar
Chakrabarty, A., Sánchez, V., Kuba, M. G., Rinehart, C. & Arteaga, C. L. Feedback upregulation of HER3 (ErbB3) expression and activity attenuates antitumor effect of PI3K inhibitors. Proc. Natl. Acad. Sci. USA 109, 2718–2723 (2012).
Article ADS CAS Google Scholar
Hopkins, B. D. et al. Suppression of insulin feedback enhances the efficacy of PI3K inhibitors. Nature 560, 499–503 (2018).
Article ADS CAS Google Scholar
Tate, J. G. et al. COSMIC: the catalogue of somatic mutations in cancer. Nucleic Acids Res. 47, D941–D947 (2019).
Article CAS Google Scholar
Zhao, L. & Vogt, P. Helical domain and kinase domain mutations in p110alpha of phosphatidylinositol 3-kinase induce gain of function by different mechanisms. Proc. Natl. Acad. Sci. USA 105, 2652–2657 (2008).
Article ADS CAS Google Scholar
Gymnopoulos, M., Elsliger, M.-A. & Vogt, P. K. Rare cancer-specific mutations in PIK3CA show gain of function. Proc. Natl. Acad. Sci. USA 104, 5569–5574 (2007).
Article ADS CAS Google Scholar
Huang, C. et al. The structure of a human p110alpha/p85alpha complex elucidates the effects of oncogenic PI3Kalpha mutations. Science 318, 1744–1748 (2007).
Article ADS CAS Google Scholar
Dbouk, H. A. et al. G protein-coupled receptor-mediated activation of p110β by Gβγ is required for cellular transformation and invasiveness. Sci. Signal. 5, ra89–ra89 (2012).
Article Google Scholar
Dornan, G. L. et al. Conformational disruption of PI3Kδ regulation by immunodeficiency mutations in PIK3CD and PIK3R1. Proc. Natl. Acad. Sci. USA 114, 1982–1987 (2017).
Article ADS CAS Google Scholar
Takeda, A. J. et al. Novel PIK3CD mutations affecting N-terminal residues of p110δ cause activated PI3Kδ syndrome (APDS) in humans. J. Allergy Clin. Immunol. 140, 1152–1156.e10 (2017).
Article CAS Google Scholar
Vallejo-Díaz, J., Chagoyen, M., Olazabal-Morán, M., González-García, A. & Carrera, A. C. The opposing roles of PIK3R1/p85α and PIK3R2/p85β in cancer. Trends Cancer 5, 233–244 (2019).
Article Google Scholar
Zhang, M., Jang, H. & Nussinov, R. Structural features that distinguish inactive and active PI3K lipid kinases. J. Mol. Biol. 432, 5849–5859 (2020).
Article CAS Google Scholar
Swaney, D. L. et al. A protein network map of head and neck cancer reveals PIK3CA mutant drug sensitivity. Science 374, eabf2911 (2021).
Song, K. W. et al. RTK-dependent inducible degradation of mutant PI3Ka drives GDC-0077 (Inavolisib) efficacy. Cancer Discov. https://doi.org/10.1158/2159-8290.CD-21-0072 (2021).
Dobbs, J. M., Jenkins, M. L. & Burke, J. E. Escherichia coli and Sf9 contaminant databases to increase efficiency of tandem mass spectrometry peptide identification in structural mass spectrometry experiments. J. Am. Soc. Mass Spectrom. 31, 2202–2209 (2020).
Article CAS Google Scholar
Perez-Riverol, Y. et al. The PRIDE database and related tools and resources in 2019: improving support for quantification data. Nucleic Acids Res. 47, D442–D450 (2019).
Article CAS Google Scholar
Hon, W. C., Berndt, A. & Williams, R. L. Regulation of lipid binding underlies the activation mechanism of class IA PI3-kinases. Oncogene 31, 3655–3666 (2012).
Article CAS Google Scholar

Download references

Acknowledgements

J.E.B. is supported by the Cancer Research Society (CRS, 843232), and the Michael Smith Foundation for Health Research (MSFHR, scholar 17686).

Author information

Authors and Affiliations

Department of Biochemistry and Microbiology, University of Victoria, Victoria, BC, V8W 2Y2, Canada
Meredith L. Jenkins, Harish Ranga-Prasad, Matthew A. H. Parson, Noah J. Harris, Manoj K. Rathinaswamy & John E. Burke
Department of Biochemistry and Molecular Biology, The University of British Columbia, Vancouver, BC, V6T 1Z3, Canada
John E. Burke

Authors

Meredith L. Jenkins
View author publications
You can also search for this author in PubMed Google Scholar
Harish Ranga-Prasad
View author publications
You can also search for this author in PubMed Google Scholar
Matthew A. H. Parson
View author publications
You can also search for this author in PubMed Google Scholar
Noah J. Harris
View author publications
You can also search for this author in PubMed Google Scholar
Manoj K. Rathinaswamy
View author publications
You can also search for this author in PubMed Google Scholar
John E. Burke
View author publications
You can also search for this author in PubMed Google Scholar

Contributions

M.L.J., H.R.P. and N.J.H. carried out all biochemical analysis in the manuscript. M.L.J., H.R.P., M.A.H.P., M.K.R. and J.E.B. carried out and analysed all HDX-MS data. M.L.J., H.R.P. and J.E.B. designed the experiments. All authors contributed to writing and editing.

Corresponding author

Correspondence to John E. Burke.

Ethics declarations

Competing interests

J.E.B. reports personal fees from Olema Pharmaceuticals (San Francisco, USA) and Scorpion Therapeutics (Boston, USA). The remaining authors declare no competing interests.

Peer review

Peer review information

Nature Communications thanks Neil Vasan and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Peer reviewer reports are available.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information

Peer Review File

Reporting Summary

Source data

Source Data

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and permissions

About this article

Cite this article

Jenkins, M.L., Ranga-Prasad, H., Parson, M.A.H. et al. Oncogenic mutations of PIK3CA lead to increased membrane recruitment driven by reorientation of the ABD, p85 and C-terminus. Nat Commun 14, 181 (2023). https://doi.org/10.1038/s41467-023-35789-6

Download citation

Received: 23 March 2022
Accepted: 29 December 2022
Published: 12 January 2023
DOI: https://doi.org/10.1038/s41467-023-35789-6

This article is cited by

Predictive, preventive, and personalized medicine in breast cancer: targeting the PI3K pathway
- Muhammad Tufail
- Jia-Ju Hu
- Ning Li
Journal of Translational Medicine (2024)

Comments

By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.