40 41 D I A M O N D L I G H T S O U R C E A N N U A L R E V I E W 2 0 2 0 / 2 1 D I A M O N D L I G H T S O U R C E A N N U A L R E V I E W 2 0 2 0 / 2 1 Biological Cryo-Imaging Group eBIC Molecular viewof a key protein could helpmake cancer cells more vulnerable to treatments Related publication: Tannous E. A., Yates L. A., Zhang X. & Burgers P. M. Mechanism of auto-inhibition and activation of Mec1ATR checkpoint kinase. Nat. Struct. Mol. Biol. 28 , 50–61 (2021). DOI: 10.1038/s41594-020-00522-0 Publication keywords: Mec1/ATR; DNA damage response; Checkpoint control; Enzyme kinetics; Serine/threonine protein kinase; Cryo-EM structures; Activationmechanism D amage to DNA needs to be repaired quickly, or it can result in defects that eventually cause cancer and ageing, particularly when a cell is replicating. Fortunately, our cells have evolved sophisticated pathways to counter the damage. One key process in the cell is the ‘DNA damage response’, where signalling factors are recruited that coordinate cell cycle progression with DNA repair. In humans, ATR is a key protein involved in the start of the repair process. A teamof scientists at Imperial College London andWashington University School of Medicine in St. Louis used high-resolution cryo-electronmicroscopy (cryo-EM) structures to identify howATR kick-starts the DNA repair process. ATR is sometimes mutated in cancer cells and is a validated drug target for cancer treatment. It and similar proteins are normally turned off (autoinhibited). They are activated when damage is detected. One key question is how ATR is maintained in an auto-inhibited state and how it is activated. The Mec1 yeast protein is essentially the same as human ATR. Using data collected at the Electron Bio-Imaging Centre (eBIC) at Diamond Light Source, the team obtained high-resolution structures of Mec1, in complex with its integral binding partner, Ddc2. In combination with biochemistry and genetics, these structures explain how this protein maintains an inhibited (off) state and the key steps required for its activation. These results allowed the researchers to propose a molecular mechanism for activation. This information helps to rationalise cancer mutations and provides amolecular framework for novel rational drug design of anticancer treatments. ProteinkinasesareenzymesthathydrolyseATPtophosphorylatesubstrates. Unlike conventional ATPases where g -phosphate is released and the chemical energy generated by ATP hydrolysis is often converted into mechanical forces, g -phosphates generated by protein kinases are subsequently transferred onto an amino acid (such as Ser,Thr,Tyr or His), resulting in protein phosphorylation, which is a major post-translational modification utilised to modulate protein functions. PIKKs (phosphatidylinositol 3-kinase-like kinases) are a family of giant protein kinases with conserved domain structures. They are characterised by a variable length of helical repeats (called HEAT repeats) at the amino-terminus followed by another region of helical repeats (called FAT domain) preceding the carboxy-terminal kinase domain. PIKKs are critical for a broad range of cellular processes, for example: DNA-PKc, ATM, ATR are apical kinases in DNA damage response; TRRAP is involved in chromatin remodelling and transcription control; SMG1 plays a key role in messenger RNA decay; and mTOR is the central coordinator in cellular metabolism and cell growth. Due to their large size and the requirement of specific chaperones, these protein giants have been difficult to produce recombinantly, therefore hindering structural and biochemical characterisations. With the cryo-EM resolution revolution, which combines the requirement of relatively small quantities of samples with the ability to separate conformational heterogeneity, we have seen a remarkable increase in structural information for these proteins, revealing their conserved structural domains and functional motifs 1,2 the tumour suppressors ATM and ATR (Tel1 and Mec1 in yeast). All PIKKs have relatively low basal kinase activity, via auto-inhibition, and must be activated by specific factors and/or post-translational modifications. A number of structures have identified key elements that help to maintain some of the PIKKs in auto-inhibited states and several recent studies have started to shed light on how they are activated 2,3 . We present the structure of Tel1 in a nucleotide-bound state. Our structure reveals molecular details of key residues surrounding the nucleotide binding site and provides a structural and molecular basis for its intrinsically low basal activity.We show that the catalytic residues are in a productive conformation for catalysis. ATR is an apical signalling kinase in DNA damage repair, especially in homology-dependentrepair,andakeyplayerinreplicationstressresponseaswell as regulating normal replication 4 . ATR activation results in the phosphorylation of a myriad of downstream proteins, including repair proteins (e.g. CtIP and EXO1), and proteins involved in cell cycle regulation, such as CHK1 leading to cell cycle arrest. ATR signalling can therefore coordinate cell cycle progression with DNA repair 4 . Mutations in ATR have frequently been found in cancer cells and ATR is a validated drug target for cancer treatment. ATR is conserved from yeast to man and Mec1 is the yeast orthologue of ATR. Taking advantage of simpler yeast genetics, a screen of MEC1 mutations in key functional regions identified a single point mutation that renders Mec1 constitutively active. Using data collected at eBIC, cryo-EM structures of Mec1 with its integral partner Ddc2 and in a complex with the ATP analogue AMPPNP were resolved, both in an auto-inhibited state (wild-type at 3.8 Å), and in the constitutively active state (Mec1(F2244L)-Ddc2 at 2.8 Å) (Fig. 1).These structures reveal molecular details of how Mec1 maintains its low basal activity, they correct previous structural models derived from lower resolution cryo-EM reconstructions, and they reconcile a previous crystal structure of Ddc2 with biochemical data. The structures reveal the conformation of the activated state and the conformational changes required for activation when compared to the auto-inhibited structure. Three key features are critical for maintaining the low basal activities of the protein: (1) elements of the ATP binding pocket, including the conserved glycine-rich loop, are not positioned for optimal nucleotide binding, resulting in a reduced affinity for ATP. Biochemical data confirmed that wildtype Mec1-Ddc2 has a lower affinity for ATP than the constitutively active mutant form; (2) In the autoinhibited state the ATP is too far away from the active site for catalysis; (3)The catalytic residues in the active site are in an unproductive conformation. The most notable key residues in the active site are the conserved DFDmotif. Weshow inthisstudythatthetwoAspresidues inDFDare involved inthebinding of twoMg 2+ ions - a prerequisite for catalysis. In the auto-inhibited state, the DFD does not have a defined conformation. This is due to the constraints imposed by interactions with a region in the protein called the PRD-I (PIKK-regulatory domain insert), which was shown in other PIKKs to restrict substrate access, thereby inhibiting kinase activity. In the activated state of Mec1-Ddc2, the Phe residue in DFD motif, which we mutated to Leu, is inserted into a hydrophobic pocket, helping the two Asp residues to orientate in correct conformations. This local change in the active site coincides with a set of larger conformational changes including the retraction of the PRD-I region from the active site, the rotation of the FAT domain relative to the kinase domain, the closure of the N-lobe and glycine-rich loop of the kinase domain, and a shift in ATP position.The range of changes result in the correct orientation and positions of all the catalytic elements. These observations of the active F2244L mutant structure corroborate cellular studies, showing that the constitutively active form of Mec1 no longer requires an activator. These studies provide unprecedented molecular details on the exact conformation of an activated Mec1/ATR and the range of changes that are required to overcome the inhibited state and to convert to the activated state – a process that may be conserved in other PIKKs. Importantly, structural, genetic and biochemical data, when taken together, suggest amechanismthat involves the release of inhibition imposed by the PRD-I coupled to re-configuration and closure of the ATP-binding pocket (Fig. 2). Both cellular and biochemical studies of additional mutant Mec1 proteins support the molecular details observed in the cryo-EM structures, highlighting the necessity to reconfigure the kinase domain in order to establish the activated state of the enzyme. Genetical and biochemical data of mutant proteins support the idea that the enzyme exists in an equilibrium between the auto-inhibited state and the activated state, with the inhibited state dominating in basal conditions. Activator binding drives the equilibrium toward the activated state. Significantly, the genetic and biochemical data indicate that different levels of kinase activity are required for optimal health of the cell during different cellular processes such as normal cell growth, DNA replication and DNA repair. Overall, this study illustrates the greater insight gained from combining functional and structural studies to understand a complex biological system.The next steps are to understand how these enzymes are recruited to the damage sitesandhowtheyengagewithactivatorproteinsto inducethechangesrequired for their activation. References: 1. Williams R. M. et al. Structures and regulations of ATM and ATR, master kinases in genome integrity. Curr. Opin. Struct. Biol. 61 , 98–105 (2020). DOI: 10.1016/j.sbi.2019.12.010 2. Yates L. A. et al. Cryo-EM Structure of Nucleotide-BoundTel1ATMUnravels the Molecular Basis of Inhibition and Structural Rationale for Disease- Associated Mutations. Structure 28 , 96-104.e3 (2020). DOI: 10.1016/j. str.2019.10.012 3. Yang H. et al. Mechanisms of mTORC1 activation by RHEB and inhibition by PRAS40. Nature 552 , 368–373 (2017). DOI: 10.1038/nature25023 4. Williams R. M. et al. Roles of ATM and ATR in DNA double strand breaks and replication stress. Prog. Biophys. Mol. Biol. (2020). DOI: 10.1016/j. pbiomolbio.2020.11.005 Funding acknowledgement: This work was funded in part by theWellcomeTrust 210658/Z/18/Z (to X. Z.) and GM118129 from the National Institutes of Health (to P. M. B.). Corresponding authors: Prof. Xiaodong Zhang, Department of Infectious Disease, Imperial College London,
[email protected] ; Dr Luke A.Yates, Department of Infectious Diseases, Imperial College London,
[email protected] Figure 1: The structure of the Mec1-Ddc2 complex, which is formed of two Mec1 proteins bound to two Ddc2 proteins. Two orthogonal views of the Mec1–Ddc2 AMP-PNP-bound complex resolved to 3.8 Å (a); and two orthogonal views of the constitutively active Mec1(F2244L)–Ddc2–AMP-PNP mutant complex resolved to 2.8-Å resolution (b). Structurally distinct domains are coloured; red, kinase; pink, C-terminal-FAT; yellow, middle-FAT; blue, N-terminal-FAT; navy, bridge; green, C-terminal HEAT; teal, N-terminal HEAT. The Ddc2 subunit is coloured purple, and the nucleotide is coloured bright green. The second Mec1–Ddc2 heterodimer is shown in grey (Mec1 ʹ ) and dark purple (Ddc2 ʹ ). Figure 2: (a) Comparison of the kinase domains between the auto-inhibited and active mutant form; (b) Kinase domain motion between autoinhibited (transparent states) and active state (opaque). Nucleotide (coloured by heteroatom) and Catalytic features are shown. The catalytic features are coloured; orange, catalytic loop; blue, activation loop; green, PRD-I; yellow, G-loop; purple, KαC; hot pink, FATC; C-FAT, light pink.