Dynamics-Driven Allostery in Protein Kinases

Abstract

Protein kinases have very dynamic structures and their functionality strongly depends on their dynamic state. Active kinases reveal a dynamic pattern with residues clustering into semirigid communities that move in μs-ms timescale. Previously detected hydrophobic spines serve as connectors between communities. Communities do not follow the traditional subdomain structure of the kinase core or its secondary structure elements. Instead they are organized around main functional units. Integration of the communities depends on the assembly of the hydrophobic spine and phosphorylation of the activation loop. Single mutations can significantly disrupt the dynamic infrastructure and thereby interfere with long-distance allosteric signaling that propagates throughout the whole molecule. Dynamics is proposed to be the underlying mechanism for allosteric regulation in protein kinases.

Keywords: allostery; community analysis; protein dynamics; protein kinases.

Figure 1

Internal architecture of the protein kinase core. (A) Bilobal structure of protein kinase core shows ATP imbedded in the deep cleft between the smaller N-lobe (white surface) and the larger C-lobe (olive surface). (B) The N-lobe has five β-strands (teal) and a large α-C helix (red). (C) The C-lobe is predominantly helical with a large Activation Segment between β8 and αF (yellow). (D) Two “hydrophobic spines” are assembled in active kinases: The R-spine (red surface) that contains four residues (RS1 through RS4) is bound to the αF-helix via conserved hydrogen bond and is correctly assembled only in the active kinase. The C-spine (yellow surface) is completed by adenine ring of ATP. The spines connect the two lobes and span across the molecule from the large αF-helix in the C-lobe to the rigid β-sheet in the N-lobe. (E) Glycine-rich Loop (G-loop) lies between β1 and β2 and coordinates the phosphate groups of ATP. β2-strand also contains the conserved Val57 that is a part of the C-spine. (F) β3-strand contains two universally conserved residues, Ala70 and Lys72. The former is a part of the C-spine and the latter coordinates the α- and β-phosphates of ATP and binds to the invariant Glu91 from the αC-helix, stabilizing the R-spine (red surface) and ordering the αC-helix for catalysis. (G) Residues that are directly involved in the phosphotransfer are in the Catalytic loop between β6 and β9 (red) with Asp184 coming from the Magnesium Positioning Loop (yellow).

Figure 2

G-loops and P-loops. The G-loop of PKA (Protein Data Bank[105] identifier (PDBID): 1ATP) is compared to the P-loop of Phosphoenolpyruvate Carboxykinase (PCK) (PDBID: 1AQ2). (A) The G-loop connects two β-strands and contains three glycines (red spheres). The residue that follows the last glycine points away from the ATP phosphates. Right after this residue there is a highly conserved valine that is a part of the C-spine (tan surface). (B) P-loop connects a β-strand and an α-helix and contains two conserved glycines (red spheres) that are followed by lysine and serine/threonine. Unlike Arg56 and Val57 in PKA Thr255 and Lys254 in PCK contribute directly to metal and phosphate binding.

Figure 3

Multiple ways to stabilize the αC-helix. (A) N-terminal part of the αC-helix is often bound to the phosphorylated residue in the Activation Loop, while the middle part of the helix is stabilized by the conserved Lys72-Glu91 salt bridge. (B) αC-helices of PKA and BRAF shown as a wheel highlighting their interactions. While interactions of the αC-helix with the kinase core is usually conserved, other surfaces of the helix have different binding partners. (C) C terminus of the αC-helix in all kinases is anchored to the kinase core by the αC-β4 Loop and via integration of its RS3 residue with the R-spine. (D) Conserved pockets flank the surface of the αC-helix as detected by Thompson et al. [46]. One pocket is located at the N terminus (tan surface) and two pockets at the C terminus (yellow and orange surfaces).

Figure 4

αC-β4 Loop is a pivot point for the αC-helix movement. (A) The αC-β4 Loop (white) connects aC-helix (red) and β4 (teal). It is anchored to the αE-helix (olive) via interaction with conserved Tyr156. The geometry of the loop is highly conserved with the β-turn between Phe100 and Leu103 and three water molecules (red and white spheres). The assembled R-spine is shown as a red surface. Val104 is shown as a white surface. (B) The αC-β4 Loops of five inactive kinases (transparent cartoons) in comparison with the αC-β4 Loop of active PKA (red and teal). The structures were aligned by their C-lobes. The following structures were used: PKA (PDBID:1ATP), BRAF (PDBID:1UWH), Src (PDBID:2SRC), IRK (PDBID:1IRK), CDK2 (PDBID:1B39), MSK1 (PDBID:1VZO).

Figure 5

Changes in slow dynamics in PKA and ERK2. (A) Nuclear spin relaxation rate (R_ex), that reflects amino acid dynamics in µs-ms timescale, is mapped onto the PKA structure. Relatively low signal is detected in PKA with no nucleotide bound. (B) Binding of the ATP-analog AMP-PNP causes a significant increase of R_ex throughout the whole molecule. (C) Binding of peptide inhibitor (PKI) leads to significant decrease of the slow dynamics. (Adapted from [26]) (D) Relaxation dispersion analysis of [¹³C]-labeled methyl groups provides information on dynamics in the µs-ms range in ERK2. Not phosphorylated form of ERK2 reveals diverse dynamic properties of hydrophobic sidechains in the range of 1–2kHz. (E) Double phosphorylation of the Activation Loop causes a significant decrease in the dynamics rate and leads to a more uniform dynamic profile of the kinase core. (Adapted from [84])

Figure 6

Community map analysis. (A) Nine communities detected in PKA with ATP and two magnesium ions bound mapped onto PKA structure (left) and presented as a graph (right). Vertex size is proportional to the number of residues in the community. Edge widths reflect the degree of communication between communities. (B) Community map for PKA bound to ATP and one magnesium ion. (C) Community map for the Apo form of PKA (D) Community map of Y204A mutant of PKA bound to ATP and one magnesium ion. Adapted from [94] and [72].

Figure 7

Detailed view of major communities in active PKA. (A) Community A (red) includes most of the rigid β-sheet from the N-lobe and adjacent parts of N- and C-tails. Community B (orange) includes the end of the C-tail, the αB-helix, a part of the αC-helix and two loops the G-loop and β4-β5. (B) Close-up of the community assignments around the G-loop. The three conserved glycines are shown as spheres. Most of the G-loop is associated with Community B with Phe54 facing the αB-helix. ATP binding is mostly controlled by the residues from Community A: V57 and Ala70 from the C-spine and Lys72 from the β3, but the β- and γ-phosphates of ATP, are positioned by Community B. (C) Community C (yellow) includes most of the αC-helix, a large portion of the α-E helix, the β6-β9 sheet and the N-tail of PKA with the αA-helix. (D) Community C contains elements that are the most critical for active conformation: the DFG-motif, RS3 from the αC-helix and the β6-β9 sheet. (E) Community D (green). Adenine ring and ribose of ATP are supported by the β7-β8 sheet including three C-spine residues on β7 (Leu172, Leu173 and Ile174). Community D also includes the most of important residues for catalysis and magnesium coordination, the main chain of Asp184 from the DFG-motif and Asn171 from the Catalytic Loop. (F) Most of the C-spine (white surface) is a part of the Community D. (G) Community F. This community includes the EPK-specific Activation Segment, the Catalytic Loop, most of the αF-helix and the αH-αI Loop. It is built around two conserved structural elements (red circles): the phosphorylated residue in the Activation Loop (Thr197) and the buried salt bridge between Glu208 from the APE-motif and Arg280 from the αH-αI Loop.

Figure 8. Key figure

Protein dynamics define communication lines inside the kinase core. (A) Two hydrophobic spines span the kinase core and connect major communities that are responsible for catalysis and regulation: A, C, D and F. (B) The amended central dogma of biology where the connection between protein structure and its function is mediated by protein dynamics. (C) The conserved αC-β4 Loop appears to be an important “communication hub” that connects Community A from the N-lobe to three C-lobe communities: Community D, which plays an important role in catalysis, regulatory Community C and Community E. It includes the two R-spine residues in the N-lobe (RS3 and RS4) and anchors them to the two R-spine residues from the C-lobe when the active conformation is assembled.