Research Areas

The Kuriyan lab seeks to understand the mechanisms that enable protein complexes to respond allosterically to inputs. The lab studies signaling proteins, particularly protein kinases and small GTPases, and nucleotide-dependent motors in DNA replication.

EGF Receptor

We made a breakthrough in understanding how the epidermal growth factor receptor (EGFR) is activated, leading to major shifts in how the field views kinase activation by dimerization. We showed that activation of EGFR involves the formation of an asymmetric dimer between the kinase domains of the receptor, in which one kinase domain serves as an activator of the other, analogous to the activation of cyclin-dependent kinases by the cell cycle control proteins known as cyclins (Zhang et al., Cell 2006 ; Zhang et al., Nature 2007). This new paradigm for the activation of a receptor tyrosine kinase explained why two human EGFR family members, HER2 and HER3, are particularly potent in combination even though HER3 lacks catalytic activity. The interaction between HER2 and HER3 underlies many cancers, including glioblastomas, and the results from our group provided the first clear mechanistic insights into how these two receptors form an active signaling complex.

We also demonstrated that the cytoplasmic module of EGFR is capable of dimerizing and activating on its own, because the juxtamembrane segment of the receptor is able to dimerize the kinase domains (Jura et al., Cell 2009; this was also shown independently by Mark Lemmon and colleagues, University of Pennsylvania). Data from our group are consistent with a mechanism in which the extracellular domains prevent the intrinsic ability of the transmembrane and cytoplasmic domains to dimerize and activate, with ligand binding releasing this block. Recently, we used cell-based analyses to work out the nature of the coupling between the extracellular domains of the receptor and the kinase domains, demonstrating that the extracellular binding of ligands to the receptor is transmitted to the intracellular kinase domains through a specific conformational coupling of the transmembrane helices (Endres et al., Cell 2013).

Towards the goal of defining how EGFR transduces signals, we determined the structure of the transmembrane and juxtamembrane segments of the receptor in a lipid bilayer. This led to a model for how ligand binding activates the intracellular kinase domains. We studied the stoichiometry of the receptor and showed that activation converts monomeric EGFR not only to dimers, but also to higher-order multimers . We developed a molecular model for these multimers, to understand how they boost the EGF response(Huang et al 2016). More recently we determined the CryoEM structure of the EGF receptor. Our results show how EGFR can couple the binding of different ligands to differential modulation of this proximity, thereby suggesting a molecular mechanism for the generation of ligand-sensitive differential outputs in this receptor family (Huang et al 2021).

SH2 domains and non-receptor tyrosine kinases

The activation of receptor tyrosine kinases results in the recruitment of signaling proteins, principally those containing SH2 domains, leading to the onward transmission of the signal. Our group was the first to report the structure of an SH2 domain bound to a phosphopeptide (Waksman et al., Nature 1992). This was followed by the elucidation of the structure of the auto-inhibited form of a Src-family tyrosine kinase, Hck (Sicheri et al., Nature 1997). This landmark structure, and a similar one determined at the same time in the laboratory of Stephen C. Harrison, revealed how the two targeting domains of the Src family kinases, the SH2 and SH3 domains, control the catalytic activity of the kinase domain when the SH2 domain docks on a phosphotyrosine residue in the C-terminal tail. We worked out how the tyrosine kinase Csk recognizes Src-family kinases with high specificity and inactivates them by phosphorylation (Levinson et al., Cell 2008). Another important advance concerning SH2 domains was the first report of the structure of the STAT transcription factor bound to DNA, showing how the STATs are activated by a clasping interaction between the SH2 domains and their phosphorylated C-terminal tails (Chen et al., Cell 1998).

Our structural analysis of the Abl tyrosine kinase has been instrumental in explaining the unexpected specificity of the cancer drug Gleevec, an effective treatment for chronic myelogenous leukemia. Our definition of the conformational states of the Abl kinase have laid the intellectual framework for understanding the molecular basis for resistance to Gleevec and for the development of next generation CML treatments by others (Schindler et al., Science 2000; Nagar et al., Cell 2003). Our work also provided the first view of the autoinhibited structure of the T-cell tyrosine kinase, ZAP-70, explaining how the tandem SH2 domains of this kinase regulate activity (Deindl et al., Cell 2007).

We are currently studying the regulation of the cytoplasmic tyrosine kinases that transmit signals from T cell and B cell receptors. These include Lck (a Src-family kinase; Lyn in B cells), ZAP-70 (Syk in B cells), and Itk (a Tec-family kinase; Btk in B cells). We are following a strategy that uses reconstituted systems to monitor function, in correlation with high-throughput mutagenesis to survey the effects of sequence variation. We have developed a method to determine the specificity profiles of tyrosine kinases, based on bacterial surface display of peptides and deep sequencing. Using this, we discovered an unexpected degree of specificity in ZAP-70 and Lck. Our results modify the view that tyrosine kinase specificity arises primarily from localization effects. For the Tec family of tyrosine kinases, we determined the mechanism by which the lipid-binding PH-TH module regulates the activity of Btk, an important drug target in cancer.

Ras and SOS

The link between tyrosine kinases and Ras activation is provided by a nucleotide exchange factor known as Son-of-Sevenless (SOS). Our group provided a near complete structural description of SOS, in collaboration with Dafna Bar-Sagi (NYU), and we made the unexpected discovery that SOS has two binding sites for Ras, and is itself dependent on Ras for its own activation (Boriack-Sjodin et al, Nature 1998; Margarit et al., Cell 2003). This unanticipated feedback loop in Ras activation by SOS was later shown by others to underlie the sharp response of T-cells to the strength of input stimuli. By reconstituting the Ras activation reaction on lipid vesicles, we have shown that SOS integrates multiple signals at the membrane before activating Ras, in collaboration with the group of Jay Groves (Gureasko et al., Nat Struct Mol Biol 2008).

The mechanisms by which the activity of SOS is controlled are unexpectedly complex, and different from those of related nucleotide exchange factors. This diversity in the regulation of related signaling proteins is common, and runs counter to our expectation that important mechanisms would be conserved. In an “Insight” article in Nature (2007), Dr. Kuriyan and David Eisenberg (UCLA) point out that colocalization, such as at the membrane, can amplify the effect on one protein of random mutations in another protein and can therefore, through natural selection, lead to random interactions between proteins being strengthened and to a startling variety of complex allosteric controls.

We determined the structural mechanisms underlying the interaction of Ras with one class of nucleotide exchange factors, the RasGRP proteins. A new understanding of the sensitivity of Ras to mutations has emerged from studies using deep mutagenesis, which allowed us to map the mutational landscape of Ras comprehensively. Moving forward, we have reconstituted Ras with activators and effectors on membranes for mechanistic studies and cryoEM analysis. Our data on the mutational sensitivity of Ras indicate that the protein gained tighter allosteric control in the vertebrate lineage.

CAMKII

Our lab provided the first structural descriptions of the calcium/calmodulin-dependent protein kinase II (CaMKII). Interest in CaMKII arises because of its ability to respond not just to the amplitude but also the frequency of calcium spikes, which makes it one of the critical molecular components underlying long term potentiation in neurons. The action of CaMK-II is balanced by that of the protein phosphatase known as PP1, the structure of which our group was also the first to determine (Goldberg et al., Nature 1995).

CaMKII is unique among protein kinases for its dodecameric assembly and we recently determined the first crystal structure of an intact fully auto-inhibited human CaMKII holoenzyme assembly (Chao et al., Cell 2011). The structure shows a tight integration of all 12 kinase domains around the central hub and, combined with SAXS measurements, explains how access to calmodulin is controlled and how nature tunes the spike frequency threshold for the activation of different isoforms.

Structures of variant forms of CaMKII hubs, some of which were captured in open-spiral, or cracked-ring forms (Bhattacharyya et al., 2016) suggest possible mechanisms for subunit exchange. Our studies showed that the CaM-binding regulatory segment of activated CaMKII can act as a wedge by docking at intersubunit interfaces in the hub (Karandur et al. 2020). This could either break the hub, or convert it into an unstable spiral form. Our data indicate a competition for the CaM-binding element, between the kinase domain, calmodulin, and the hub interface, with phosphorylation biasing the CaM-binding element towards the hub.

DNA Polymerase Clamp Loaders

Highly processive DNA polymerases are tethered to ring-shaped sliding clamps that encircle DNA, allowing rapid movement of the replication fork (Kelch et al., 2012). With Mike O’Donnell (Rockefeller University) we establish the structural principles underlying the clamp-loader complexes that load the sliding clamps onto DNA. These machines have ATPase subunits that respond allosterically to ATP, to each other, and to DNA. These subunits are related evolutionarily to Ras, providing a conceptual connection to our projects on signaling.

Previously, we had determined structures of clamp-loader complexes in various states of assembly, but not in complete form. In a new direction for this project, we used deep mutagenesis and protein engineering to gain a definitive understanding of how ATP hydrolysis is coupled to DNA recognition, with allosteric communication via a critical hydrogen-bonded junction (Subramanian et al eLife 2021). The ultimate goal of the project is to relate clamp-loader mechanism to that of DNA translocases.