Inter-lobe Motions Allosterically Regulate the Structure and Function of EGFR Kinase

Highlights

• Comprehensive analysis of all 206 crystal structures of EGFR kinase was performed.

• Enhanced sampling simulations were also implemented to see the intrinsic structures.

• Lobe arrangement is a determinant for the activation of EGFR kinase via dimerization.

• Catalytically important motifs are highly regulated by the inter-lobe arrangement.

Abstract

Protein kinases play important roles in cellular signaling and have been one of the best-studied drug targets. The kinase domain of epidermal growth factor receptor (EGFR) is a receptor tyrosine kinase that has been extensively studied for cancer drug discovery and for understanding the unique activation mechanism by dimerization. Here, we analyzed all available 206 crystal structures of the EGFR kinase and the dynamics observed in molecular simulations to identify how these structures are determined. It was found that the arrangement between the N- and C-terminal lobes plays a key role in regulating the kinase structure by sensitively responding to the intermolecular interactions, or the crystal environment. A whole variety of crystal forms in the database is thus reflected in the broad distribution of the inter-lobe arrangement. The configuration of the catalytically important motifs as well as the bound ATP is closely coupled with the inter-lobe motion. When the intermolecular interactions are those of the activating asymmetric dimer, EGFR kinase takes the open-lobe arrangement that constructs the catalytically active configuration.

Introduction

Protein kinases play essential roles in cellular signaling process and have been one of the best-studied drug targets [[1], [2], [3], [4], [5]]. The structural information has been accumulated abundantly in the crystal structures of complex with various chemical compounds, revealing the details of protein–ligand interactions [[6], [7], [8]]. These crystal structures also provide another important piece of information for the structural variation in a single protein kinase [[9], [10], [11]] (Figure 1), which is observed mainly in the activation loop (A-loop) that contains the DFG motif, the C-helix, and the inter-lobe arrangement between the N-terminal and C-terminal lobes (N-lobe and C-lobe). Protein kinase utilizes its intrinsic flexibility for switching on and off the phosphorylation activity by changing its structure to respond to the stimulus (usually via trans phosphorylation) coming from the upstream of the signaling pathway. Likewise, the flexibility produces a variety of crystal structures of the kinase molecule in such a way that a snapshot structure is chosen from the structural ensemble of the kinase, depending on the conditions in the crystal environment, such as phosphorylation, mutations, ligand binding, and crystal packing.

In this study, we focus on the kinase domain of the epidermal growth factor receptor (EGFR) [12,13]. EGFR kinase is a receptor tyrosine kinase that is activated by dimerization, rather than phosphorylation [14]. The purpose of this article is to elucidate the structural basis of the unique activation mechanism of EGFR kinases based on the comparative analysis of available crystal structures in various conditions taken from the compilation of the kinase database KLIFS (ver. Nov. 2019) [6,7]. The list of the crystal structures, given in Table S1, contains 148 PDB entries, 206 domain structures, 12 different space groups, 88 kinds of bound ligands, and 15 types of mutations. This enormous amount of structural information covers the entire configurational space of EGFR kinases that occur during the activation/deactivation process. As a complemental way of studying the dynamics, we also performed molecular dynamics (MD) simulations using an enhanced sampling technique, namely, multiscale enhanced sampling (MSES) [15,16], which allows us to sample the entire configurational space in which the 206 kinase structures are distributed.

There have been comparative studies on the available protein kinase structures [10,11], which are based on the analyses of the local structures, the DFG motif and αC-helix. It is reasonable to focus on the local structures for protein kinases regarding phosphorylation-driven activation, because the conformational changes of the activation loop due to phosphorylation are the most important structural features occurring upon activation [17]. On the contrary, in the case of the EGFR kinase, the formation of the asymmetric dimer (Figure 1(c)) is the cause of activation, and thus, activation occurs in an allosteric manner that involves the entire protein molecule. As shown in Figure 1(a), an exceptionally large-scale inter-lobe motion was observed upon activation of the EGFR kinase. Therefore, it is necessary to understand the allosteric mechanism that connects the motion of the entire molecule to the rearrangement of the catalytically important motifs.

The flexibility of the EGFR kinase is immediately reflected in the susceptible response to the protein–protein interactions that occur in crystal packing, and results in a large variation of the inter-lobe arrangement, corresponding to diverse crystal forms. Thus, the present analyses on the crystal structures place emphasis on crystal packing, and simulation studies have been conducted to examine the structural relaxation that occurs in the transfer from the crystal environment to the solution environment.