Development of selective inhibitors of phosphatidylinositol 3-kinase C2α.

Phosphatidylinositol 3-kinase type 2α (PI3KC2α) and related class II PI3K isoforms are of increasing biomedical interest because of their crucial roles in endocytic membrane dynamics, cell division and signaling, angiogenesis, and platelet morphology and function. Herein we report the development and characterization of PhosphatidylInositol Three-kinase Class twO INhibitors (PITCOINs), potent and highly selective small-molecule inhibitors of PI3KC2α catalytic activity. PITCOIN compounds exhibit strong selectivity toward PI3KC2α due to their unique mode of interaction with the ATP-binding site of the enzyme. We demonstrate that acute inhibition of PI3KC2α-mediated synthesis of phosphatidylinositol 3-phosphates by PITCOINs impairs endocytic membrane dynamics and membrane remodeling during platelet-dependent thrombus formation. PITCOINs are potent and selective cell-permeable inhibitors of PI3KC2α function with potential biomedical applications ranging from thrombosis to diabetes and cancer.

Activation of GCN2 by the ribosomal P-stalk.

Cells dynamically adjust their protein translation profile to maintain homeostasis in changing environments. During nutrient stress, the kinase general control nonderepressible 2 (GCN2) phosphorylates translation initiation factor eIF2α, initiating the integrated stress response (ISR). To examine the mechanism of GCN2 activation, we have reconstituted this process in vitro, using purified components. We find that recombinant human GCN2 is potently stimulated by ribosomes and, to a lesser extent, by tRNA. Hydrogen/deuterium exchange-mass spectrometry (HDX-MS) mapped GCN2-ribosome interactions to domain II of the uL10 subunit of the ribosomal P-stalk. Using recombinant, purified P-stalk, we showed that this domain of uL10 is the principal component of binding to GCN2; however, the conserved 14-residue C-terminal tails (CTTs) in the P1 and P2 P-stalk proteins are also essential for GCN2 activation. The HisRS-like and kinase domains of GCN2 show conformational changes upon binding recombinant P-stalk complex. Given that the ribosomal P-stalk stimulates the GTPase activity of elongation factors during translation, we propose that the P-stalk could link GCN2 activation to translational stress, leading to initiation of ISR.

Structure of lipid kinase p110ß/p85ß elucidates an unusual SH2-domain-mediated inhibitory mechanism.

Phosphoinositide 3-kinases (PI3Ks) are essential for cell growth, migration, and survival. The structure of a p110ß/p85ß complex identifies an inhibitory function for the C-terminal SH2 domain (cSH2) of the p85 regulatory subunit. Mutagenesis of a cSH2 contact residue activates downstream signaling in cells. This inhibitory contact ties up the C-terminal region of the p110ß catalytic subunit, which is essential for lipid kinase activity. In vitro, p110ß basal activity is tightly restrained by contacts with three p85 domains: the cSH2, nSH2, and iSH2. RTK phosphopeptides relieve inhibition by nSH2 and cSH2 using completely different mechanisms. The binding site for the RTK's pYXXM motif is exposed on the cSH2, requiring an extended RTK motif to reach and disrupt the inhibitory contact with p110ß. This contrasts with the nSH2 where the pY-binding site itself forms the inhibitory contact. This establishes an unusual mechanism by which p85 SH2 domains contribute to RTK signaling specificities.

The intrinsically disordered tails of PTEN and PTEN-L have distinct roles in regulating substrate specificity and membrane activity.

Phosphatase and tensin homologue deleted on chromosome 10 (PTEN) is a lipid and protein phosphatase, and both activities are necessary for its role as a tumour suppressor. PTEN activity is controlled by phosphorylation of its intrinsically disordered C-terminal tail. A recently discovered variant of PTEN, PTEN-long (PTEN-L), has a 173-residue N-terminal extension that causes PTEN-L to exhibit unique behaviour, such as movement from one cell to another. Using hydrogen/deuterium exchange mass spectrometry (HDX-MS) and biophysical assays, we show that both the N-terminal extension of PTEN-L and C-terminal tail of PTEN affect the phosphatase activity using unique mechanisms. Phosphorylation of six residues in the C-terminal tail of PTEN results in auto-inhibitory interactions with the phosphatase and C2 domains, effectively blocking both the active site and the membrane-binding interface of PTEN. Partially dephosphorylating PTEN on pThr(366)/pSer(370) results in sufficient exposure of the active site to allow a selective activation for soluble substrates. Using HDX-MS, we identified a membrane-binding element in the N-terminal extension of PTEN-L, termed the membrane-binding helix (MBH). The MBH radically alters the membrane binding mechanism of PTEN-L compared with PTEN, switching PTEN-L to a 'scooting' mode of catalysis from the 'hopping' mode that is characteristic of PTEN.

Molecular determinants of PI3Kγ-mediated activation downstream of G-protein-coupled receptors (GPCRs).

Phosphoinositide 3-kinase gamma (PI3Kγ) has profound roles downstream of G-protein-coupled receptors in inflammation, cardiac function, and tumor progression. To gain insight into how the enzyme's activity is shaped by association with its p101 adaptor subunit, lipid membranes, and Gßγ heterodimers, we mapped these regulatory interactions using hydrogen-deuterium exchange mass spectrometry. We identify residues in both the p110γ and p101 subunits that contribute critical interactions with Gßγ heterodimers, leading to PI3Kγ activation. Mutating Gßγ-interaction sites of either p110γ or p101 ablates G-protein-coupled receptor-mediated signaling to p110γ/p101 in cells and severely affects chemotaxis and cell transformation induced by PI3Kγ overexpression. Hydrogen-deuterium exchange mass spectrometry shows that association with the p101 regulatory subunit causes substantial protection of the RBD-C2 linker as well as the helical domain of p110γ. Lipid interaction massively exposes that same helical site, which is then stabilized by Gßγ. Membrane-elicited conformational change of the helical domain could help prepare the enzyme for Gßγ binding. Our studies and others identify the helical domain of the class I PI3Ks as a hub for diverse regulatory interactions that include the p101, p87 (also known as p84), and p85 adaptor subunits; Rab5 and Gßγ heterodimers; and the ß-adrenergic receptor kinase.

Oncogenic mutations mimic and enhance dynamic events in the natural activation of phosphoinositide 3-kinase p110α (PIK3CA).

The p110α catalytic subunit (PIK3CA) is one of the most frequently mutated genes in cancer. We have examined the activation of the wild-type p110α/p85α and a spectrum of oncogenic mutants using hydrogen/deuterium exchange mass spectrometry (HDX-MS). We find that for the wild-type enzyme, the natural transition from an inactive cytosolic conformation to an activated form on membranes entails four distinct events. Analysis of oncogenic mutations shows that all up-regulate the enzyme by enhancing one or more of these dynamic events. We provide the first insight into the activation mechanism by mutations in the linker between the adapter-binding domain (ABD) and the Ras-binding domain (RBD) (G106V and G118D). These mutations, which are common in endometrial cancers, enhance two of the natural activation events: movement of the ABD and ABD-RBD linker relative to the rest of the catalytic subunit and breaking the C2-iSH2 interface on binding membranes. C2 domain mutants (N345K and C420R) also mimic these events, even in the absence of membranes. A third event is breaking the nSH2-helical domain contact caused by phosphotyrosine-containing peptides binding to the enzyme, which is mimicked by a helical domain mutation (E545K). Interaction of the C lobe of the kinase domain with membranes is the fourth activation event, and is potentiated by kinase domain mutations (e.g., H1047R). All mutations increased lipid binding and basal activity, even mutants distant from the membrane surface. Our results elucidate a unifying mechanism in which diverse PIK3CA mutations stimulate lipid kinase activity by facilitating allosteric motions required for catalysis on membranes.

ESCRT-III binding protein MITD1 is involved in cytokinesis and has an unanticipated PLD fold that binds membranes.

The endosomal sorting complexes required for transport (ESCRT) proteins have a critical function in abscission, the final separation of the daughter cells during cytokinesis. Here, we describe the structure and function of a previously uncharacterized ESCRT-III interacting protein, MIT-domain containing protein 1 (MITD1). Crystal structures of MITD1 reveal a dimer, with a microtubule-interacting and trafficking (MIT) domain at the N terminus and a unique, unanticipated phospholipase D-like (PLD) domain at the C terminus that binds membranes. We show that the MIT domain binds to a subset of ESCRT-III subunits and that this interaction mediates MITD1 recruitment to the midbody during cytokinesis. Depletion of MITD1 causes a distinct cytokinetic phenotype consistent with destabilization of the midbody and abscission failure. These results suggest a model whereby MITD1 coordinates the activity of ESCRT-III during abscission with earlier events in the final stages of cell division.

Structural basis for selective recognition of ESCRT-III by the AAA ATPase Vps4.

The AAA+ ATPases are essential for various activities such as membrane trafficking, organelle biogenesis, DNA replication, intracellular locomotion, cytoskeletal remodelling, protein folding and proteolysis. The AAA ATPase Vps4, which is central to endosomal traffic to lysosomes, retroviral budding and cytokinesis, dissociates ESCRT complexes (the endosomal sorting complexes required for transport) from membranes. Here we show that, of the six ESCRT--related subunits in yeast, only Vps2 and Did2 bind the MIT (microtubule interacting and transport) domain of Vps4, and that the carboxy-terminal 30 residues of the subunits are both necessary and sufficient for interaction. We determined the crystal structure of the Vps2 C terminus in a complex with the Vps4 MIT domain, explaining the basis for selective ESCRT-III recognition. MIT helices alpha2 and alpha3 recognize a (D/E)xxLxxRLxxL(K/R) motif, and mutations within this motif cause sorting defects in yeast. Our crystal structure of the amino-terminal domain of an archaeal AAA ATPase of unknown function shows that it is closely related to the MIT domain of Vps4. The archaeal ATPase interacts with an archaeal ESCRT-III-like protein even though these organisms have no endomembrane system, suggesting that the Vps4/ESCRT-III partnership is a relic of a function that pre-dates the divergence of eukaryotes and Archaea.

Structural insights into the activation of ataxia-telangiectasia mutated by oxidative stress.

Ataxia-telangiectasia mutated (ATM) is a master kinase regulating DNA damage response that is activated by DNA double-strand breaks. However, ATM is also directly activated by reactive oxygen species, but how oxidative activation is achieved remains unknown. We determined the cryo-EM structure of an H2O2-activated ATM and showed that under oxidizing conditions, ATM formed an intramolecular disulfide bridge between two protomers that are rotated relative to each other when compared to the basal state. This rotation is accompanied by release of the substrate-blocking PRD region and twisting of the N-lobe relative to the C-lobe, which greatly optimizes catalysis. This active site remodeling enabled us to capture a substrate (p53) bound to the enzyme. This provides the first structural insights into how ATM is activated during oxidative stress.

Paradoxical dominant negative activity of an immunodeficiency-associated activating PIK3R1 variant.

PIK3R1 encodes three regulatory subunits of class IA phosphoinositide 3-kinase (PI3K), each associating with any of three catalytic subunits, namely p110α, p110ß or p110δ. Constitutional PIK3R1 mutations cause diseases with a genotype-phenotype relationship not yet fully explained: heterozygous loss-of-function mutations cause SHORT syndrome, featuring insulin resistance and short stature attributed to reduced p110α function, while heterozygous activating mutations cause immunodeficiency, attributed to p110δ activation and known as APDS2. Surprisingly, APDS2 patients do not show features of p110α hyperactivation, but do commonly have short stature or SHORT syndrome, suggesting p110α hypofunction. We sought to investigate this. In dermal fibroblasts from an APDS2 patient, we found no increased PI3K signalling, with p110δ expression markedly reduced. In preadipocytes, the APDS2 variant was potently dominant negative, associating with Irs1 and Irs2 but failing to heterodimerise with p110α. This attenuation of p110α signalling by a p110δ-activating PIK3R1 variant potentially explains co-incidence of gain-of-function and loss-of-function PIK3R1 phenotypes.

Bipartite binding and partial inhibition links DEPTOR and mTOR in a mutually antagonistic embrace.

The mTORC1 kinase complex regulates cell growth, proliferation, and survival. Because mis-regulation of DEPTOR, an endogenous mTORC1 inhibitor, is associated with some cancers, we reconstituted mTORC1 with DEPTOR to understand its function. We find that DEPTOR is a unique partial mTORC1 inhibitor that may have evolved to preserve feedback inhibition of PI3K. Counterintuitively, mTORC1 activated by RHEB or oncogenic mutation is much more potently inhibited by DEPTOR. Although DEPTOR partially inhibits mTORC1, mTORC1 prevents this inhibition by phosphorylating DEPTOR, a mutual antagonism that requires no exogenous factors. Structural analyses of the mTORC1/DEPTOR complex showed DEPTOR's PDZ domain interacting with the mTOR FAT region, and the unstructured linker preceding the PDZ binding to the mTOR FRB domain. The linker and PDZ form the minimal inhibitory unit, but the N-terminal tandem DEP domains also significantly contribute to inhibition.

Structural basis for VPS34 kinase activation by Rab1 and Rab5 on membranes.

The lipid phosphatidylinositol-3-phosphate (PI3P) is a regulator of two fundamental but distinct cellular processes, endocytosis and autophagy, so its generation needs to be under precise temporal and spatial control. PI3P is generated by two complexes that both contain the lipid kinase VPS34: complex II on endosomes (VPS34/VPS15/Beclin 1/UVRAG), and complex I on autophagosomes (VPS34/VPS15/Beclin 1/ATG14L). The endosomal GTPase Rab5 binds complex II, but the mechanism of VPS34 activation by Rab5 has remained elusive, and no GTPase is known to bind complex I. Here we show that Rab5a-GTP recruits endocytic complex II to membranes and activates it by binding between the VPS34 C2 and VPS15 WD40 domains. Electron cryotomography of complex II on Rab5a-decorated vesicles shows that the VPS34 kinase domain is released from inhibition by VPS15 and hovers over the lipid bilayer, poised for catalysis. We also show that the GTPase Rab1a, which is known to be involved in autophagy, recruits and activates the autophagy-specific complex I, but not complex II. Both Rabs bind to the same VPS34 interface but in a manner unique for each. These findings reveal how VPS34 complexes are activated on membranes by specific Rab GTPases and how they are recruited to unique cellular locations.

ESCRT-II, an endosome-associated complex required for protein sorting: crystal structure and interactions with ESCRT-III and membranes.

ESCRT-I, -II, and -III protein complexes are sequentially recruited to endosomal membranes, where they orchestrate protein sorting and MVB biogenesis. In addition, they play a critical role in retrovirus budding. Structural understanding of ESCRT interaction networks is largely lacking. The 3.6 A structure of the yeast ESCRT-II core presented here reveals a trilobal complex containing two copies of Vps25, one copy of Vps22, and the C-terminal region of Vps36. Unexpectedly, the entire ESCRT-II core consists of eight repeats of a common building block, a "winged helix" domain. Two PPXY-motifs from Vps25 are involved in contacts with Vps22 and Vps36, and their mutation leads to ESCRT-II disruption. We show that purified ESCRT-II binds directly to the Vps20 component of ESCRT-III. Surprisingly, this binding does not require the protruding N-terminal coiled-coil of Vps22. Vps25 is the chief subunit responsible for Vps20 recruitment. This interaction dramatically increases binding of both components to lipid vesicles in vitro.

Evolution and assembly of ESCRTs.

The AAA (ATPase associated with various cellular activities) proteins participate in membrane trafficking, organelle biogenesis, DNA replication, intracellular locomotion, cytoskeletal remodelling, protein folding and proteolysis. The AAA Vps (vacuolar protein sorting) 4 is central to traffic to lysosomes, retroviral budding and mammalian cell division. It dissociates ESCRTs (endosomal sorting complexes required for transport) from endosomal membranes, enabling their recycling to the cytosol, and plays a role in fission of intraluminal vesicles within MVBs (multivesicular bodies). The mechanism of Vps4-catalysed disassembly of ESCRT networks is unknown; however, it requires interaction between Vps4 and ESCRT-III subunits. The 30 C-terminal residues of Vps2 and Vps46 (Did2) subunits are both necessary and sufficient for interaction with the Vps4 N-terminal MIT (microtubule-interacting and transport) domain, and the crystal structure of the Vps2 C-terminus in a complex with the Vps4 MIT domain shows that MIT helices alpha2 and alpha3 recognize a (D/E)XXLXXRLXXL(K/R) MIM (MIT-interacting motif). These Vps2-MIT interactions are essential for vacuolar sorting and for Vps4-catalysed disassembly of ESCRT-III networks in vitro. Electron microscopy of ESCRT-III filaments assembled in vitro has enabled us to identify surfaces of the Vps24 subunit that are critical for protein sorting in vivo. The ESCRT-III-Vps4 interaction predates the divergence of Archaea and Eukarya. The Crenarchaea have three classes of ESCRT-III-like subunits, and one of these subunits interacts with an archaeal Vps4-like protein in a manner closely related to the human Vps4-human ESCRT-III subunit Vps20 interaction. This archaeal Vps4-ESCRT-III interaction appears to have a fundamental role in cell division in the Crenarchaea.

Architecture of human Rag GTPase heterodimers and their complex with mTORC1.

The Rag guanosine triphosphatases (GTPases) recruit the master kinase mTORC1 to lysosomes to regulate cell growth and proliferation in response to amino acid availability. The nucleotide state of Rag heterodimers is critical for their association with mTORC1. Our cryo-electron microscopy structure of RagA/RagC in complex with mTORC1 shows the details of RagA/RagC binding to the RAPTOR subunit of mTORC1 and explains why only the RagAGTP/RagCGDP nucleotide state binds mTORC1. Previous kinetic studies suggested that GTP binding to one Rag locks the heterodimer to prevent GTP binding to the other. Our crystal structures and dynamics of RagA/RagC show the mechanism for this locking and explain how oncogenic hotspot mutations disrupt this process. In contrast to allosteric activation by RHEB, Rag heterodimer binding does not change mTORC1 conformation and activates mTORC1 by targeting it to lysosomes.

Structural studies of phosphoinositide 3-kinase-dependent traffic to multivesicular bodies.

Three large protein complexes known as ESCRT I, ESCRT II and ESCRT III drive the progression of ubiquitinated membrane cargo from early endosomes to lysosomes. Several steps in this process critically depend on PtdIns3P, the product of the class III phosphoinositide 3-kinase. Our work has provided insights into the architecture, membrane recruitment and functional interactions of the ESCRT machinery. The fan-shaped ESCRT I core and the trilobal ESCRT II core are essential to forming stable, rigid scaffolds that support additional, flexibly-linked domains, which serve as gripping tools for recognizing elements of the MVB (multivesicular body) pathway: cargo protein, membranes and other MVB proteins. With these additional (non-core) domains, ESCRT I grasps monoubiquitinated membrane proteins and the Vps36 subunit of the downstream ESCRT II complex. The GLUE (GRAM-like, ubiquitin-binding on Eap45) domain extending beyond the core of the ESCRT II complex recognizes PtdIns3P-containing membranes, monoubiquitinated cargo and ESCRT I. The structure of this GLUE domain demonstrates that it has a split PH (pleckstrin homology) domain fold, with a non-typical phosphoinositide-binding pocket. Mutations in the lipid-binding pocket of the ESCRT II GLUE domain cause a strong defect in vacuolar protein sorting in yeast.

Robotic nanolitre protein crystallisation at the MRC Laboratory of Molecular Biology.

We have set up high-throughput robotic systems to screen and optimise crystallisation conditions of biological macromolecules with the aim to make difficult structural biology projects easier. The initial screening involves two robots. A Tecan Genesis liquid handler is used to transfer commercially available crystallisation reagents from 15 ml test tubes into the reservoirs of 96-well crystallisation plates. This step is fully automated and includes a carousel for intermediate plate storage, a Beckman plate sealer and a robotic arm, which transfers plates in between steps. For adding the sample, we use a second robot, a 17-tip Cartesian Technologies PixSys 4200 SynQuad liquid handler, which uses a syringe/solenoid valve combination to dispense small quantities of liquid (typically 100 nl) without touching the surface of the plate. Sixteen of the tips are used to transfer the reservoir solution to the crystallisation wells, while the 17th tip is used to dispense the protein. The screening of our standard set of 1440 conditions takes about 3 h and requires 300 microl of protein solution. Once crystallisation conditions have been found, they are optimised using a second Tecan Genesis liquid handler, which is programmed to pipette gradients from four different corner solutions into a wide range of crystallisation plate formats. For 96-well plates, the Cartesian robot can be used to add the sample. The methods described are now used almost exclusively for obtaining diffraction quality crystals in our laboratory with a throughput of several thousand plates per year. Our set-up has been copied in many institutions worldwide.

Characterization of Atg38 and NRBF2, a fifth subunit of the autophagic Vps34/PIK3C3 complex.

The phosphatidylinositol 3-kinase Vps34 is part of several protein complexes. The structural organization of heterotetrameric complexes is starting to emerge, but little is known about organization of additional accessory subunits that interact with these assemblies. Combining hydrogen-deuterium exchange mass spectrometry (HDX-MS), X-ray crystallography and electron microscopy (EM), we have characterized Atg38 and its human ortholog NRBF2, accessory components of complex I consisting of Vps15-Vps34-Vps30/Atg6-Atg14 (yeast) and PIK3R4/VPS15-PIK3C3/VPS34-BECN1/Beclin 1-ATG14 (human). HDX-MS shows that Atg38 binds the Vps30-Atg14 subcomplex of complex I, using mainly its N-terminal MIT domain and bridges the coiled-coil I regions of Atg14 and Vps30 in the base of complex I. The Atg38 C-terminal domain is important for localization to the phagophore assembly site (PAS) and homodimerization. Our 2.2 Å resolution crystal structure of the Atg38 C-terminal homodimerization domain shows 2 segments of α-helices assembling into a mushroom-like asymmetric homodimer with a 4-helix cap and a parallel coiled-coil stalk. One Atg38 homodimer engages a single complex I. This is in sharp contrast to human NRBF2, which also forms a homodimer, but this homodimer can bridge 2 complex I assemblies.

Structures of PI4KIIIß complexes show simultaneous recruitment of Rab11 and its effectors.

Phosphatidylinositol 4-kinases (PI4Ks) and small guanosine triphosphatases (GTPases) are essential for processes that require expansion and remodeling of phosphatidylinositol 4-phosphate (PI4P)-containing membranes, including cytokinesis, intracellular development of malarial pathogens, and replication of a wide range of RNA viruses. However, the structural basis for coordination of PI4K, GTPases, and their effectors is unknown. Here, we describe structures of PI4Kß (PI4KIIIß) bound to the small GTPase Rab11a without and with the Rab11 effector protein FIP3. The Rab11-PI4KIIIß interface is distinct compared with known structures of Rab complexes and does not involve switch regions used by GTPase effectors. Our data provide a mechanism for how PI4KIIIß coordinates Rab11 and its effectors on PI4P-enriched membranes and also provide strategies for the design of specific inhibitors that could potentially target plasmodial PI4KIIIß to combat malaria.