In silico activation of Src tyrosine kinase reveals the molecular basis for intramolecular autophosphorylation☆
Introduction
Cellular Src (c-Src) is the normal cellular counterpart of the product of the first oncogene identified from Rous sarcoma virus, v-Src, and also the first protein characterized as being able to phosphorylate tyrosine residues [1]. Src substrates are found in the cytosol or at the inner face of the plasma membrane, or at cell–matrix or cell–cell adhesions. Tyrosine phosphorylation of these proteins can affect their function directly or, alternatively, the phosphotyrosyl residues can serve as docking sites for the binding of signaling proteins containing SH2 domains. The resulting complexes then initiate pathways that regulate protein synthesis, gene expression, cytoskeletal assembly and many other aspects of cell function [2]. Because Src family members are involved in many signaling pathways by means of which several surface receptors regulate cell growth and proliferation, their catalytic activity is strictly regulated [3], thus offering unique opportunities for modulation by small molecules in the fight against disease.
Src can also phosphorylate itself at tyrosine residues. Although the major in vivo phosphorylation site on platelet Src was shown to be Tyr-527, which acts as a negative regulator of its kinase activity [4] and is absent in v-Src, the major site of autophosphorylation in vitro is Tyr-416, a residue located at the center of the catalytic domain [5]. Phosphorylation of Tyr-416 in both v- and c-Src results in increased activation of the enzyme [6], [7], [8] whereas mutation of this residue to phenylalanine (Y416F) significantly reduces the transforming potential of Src [9]. A tyrosine kinase that specifically phosphorylates Src at Tyr-527, carboxy-terminal Src kinase (Csk), has been identified [10], as have phosphatases that can dephosphorylate this residue [11], highlighting the fact that the phosphorylation status of Tyr-527 is key for in vivo c-Src regulation. However, to the best of our knowledge, no tyrosine kinase specific for Src Tyr-416 has been described yet. Instead, an autophosphorylation mechanism has been proposed based on the finding that the level of phosphorylation of this tyrosine residue correlates with the activity of the enzyme and its mutants [12]. Since autophosphorylation of purified, monomeric, v-Src was shown to be independent of its concentration, an intramolecular reaction was early proposed [13]. Nevertheless, intermolecular autophosphorylation is also possible because Src can be phosphorylated in cells expressing both the Y416F mutant enzyme and another mutant form that can be phosphorylated at Tyr-416 but cannot donate phosphate [14]. Since Tyr-416 lies within a region conserved in all members of the tyrosine kinase family named the activation loop, the phosphorylation of this residue could induce an important conformational change responsible for the activation of the enzyme.
The structural domains of Src kinase (and other members of the family) are, in order from the N-terminus: the SH4 (Src homology 4), SH3, SH2 and SH1 domains. SH1 is the catalytic domain, SH2 and SH3 are both molecular adhesives important for protein–protein interaction, and SH4 plays a role in membrane attachment. SH2 domains bind phosphotyrosine-containing peptides [15], [16] whereas SH3 domains recognize proline-rich peptides [17]. Both SH2 and SH3 domains were early shown to be involved in the regulation of the kinase activity of Src [18], [19]. The complex interactions between different domains of c-Src were better understood when the three-dimensional structures of c-Src [20], and the Src family tyrosine kinase Hck [21], were solved, both in a catalytically inactive conformation.
Each of these crystal structures revealed a finely adjusted nanomachine [2] in which all structural elements, i.e., the SH3 and SH2 domains, the linker between SH2 and SH1 domains, and the carboxy-terminal tail cooperate in order to keep the catalytic domain under control. As in all known kinases, the catalytic domain is made up of a small α/β N-terminal lobe and a large α C-terminal lobe. The linker joining the SH2 and the catalytic domain adopts a polyproline type II helix structure and serves as an adapter to fit together the N-terminal lobe and the SH3 domain. This conformation places the SH2 domain in a suitable position to bind the phosphotyrosine residue at the C-terminal tail (pTyr-527) which extends from the base of the catalytic domain. Binding of pTyr-527 to SH2 acts as a safety catch that locks the whole structure in an inactive conformation [2].
The overall structure of the catalytic domain, shaped as two lobes connected by a short polypeptidic strand, strongly suggests considerable interlobe mobility. In the inactive conformation the two lobes approach each other closely, and the interactions of SH1 with the SH2 and SH3 regulatory domains, the linker, and the C-terminal tail serve to bring the catalytic lobes close together thereby narrowing the cleft and preventing substrate-binding. In an ‘open’ active conformation [22], as that found in the crystal structure of the catalytic domain of Lck (another member of the Src family), the lobes are swung apart and the substrate-binding cleft is easily accessible, thus supporting the importance of structural mobility in Src regulation.
Lobe mobility could be essential for Src activation, not only to facilitate substrate recognition, but also to reorganize the active site during catalysis. This kind of motion is common to many other enzymes and protein molecules [23], in which two or more domains are connected by a few strands of polypeptide chains that can be considered as hinges [24]. The conformational changes responsible for these motions are usually limited to the hinge region in so far as the domains behave as rigid bodies.
Dynamic hinge-bending motions in proteins are not easily amenable to experimental structural studies but molecular dynamics (MD) simulations provide a computational alternative that can help to gain insight into these processes at the atomic level. However, these motions take place at time scales that are more than one order of magnitude longer than those currently achieved by state-of-the-art MD simulations. This is so because the system must surmount the energy barrier that separates the open and the closed forms. Thus, even though the energy-barrier crossing process itself is normally quite fast, the time required for random thermal fluctuations within the system to produce the local atomic momenta required for overcoming the local energy barrier may be of the order of milliseconds or longer. In these cases, a targeted MD (tMD) approach [25] can be used to accelerate the process and simulate the subdomain hinge-bending motion. This approach has been previously used to study conformational changes in several proteins such as the GroEl chaperone [26] and the glutamate receptor ligand-binding core [27], as well as the coupling between SH2 and SH3 domains when Src is forced to adopt the open conformation [28].
In the following, we focus our attention on the dynamic properties of the catalytic domain itself. We first study its molecular architecture to pinpoint the mobile parts responsible for the interlobe motion. We then simulate the hinge-bending motion using tMD to explore how the conformational change in the activation loop is triggered, and in doing so we also probe the feasibility of an intramolecular mechanism for Tyr-416 autophosphorylation.
Section snippets
Determination of the hinge regions of Src catalytic domain
Evidence for a hinge-bending motion associated with the activation of the Src catalytic domain can be gained from comparison of the crystal forms of non-active Src [20] and active Lck catalytic domains [22]. Visual inspection of these structures shows that, irrespective of changes in interlobe orientation and activation loop conformation, the overall structure of each lobe is maintained upon activation (Fig. 1).
Alignment of the catalytic domains of Src and Lck also shows that not only do they
Atomic data
The atomic coordinates of human tyrosine kinase, c-Src [20], and human lymphocyte kinase, Lck [22], were obtained from the protein data bank (PDB codes: 2SRC and 3LCK, respectively). Both proteins were structurally aligned with the DALI software [35]. Differences in dihedral angles between the α-carbon (Cα) traces of open and closed forms were measured, with positive and negative values representing, respectively, clockwise and anticlockwise increments relative to the closed form.
Neither ATP
Conclusions
The overall molecular architecture of the catalytic domain of all protein kinases, which is shaped as two lobes connected by a short polypeptidic strand, reveals a remarkable plasticity [34] and strongly suggests considerable interlobe mobility. Comparison of Src in a closed inactive form and Lck in an open active conformation provides structural evidence that the motion of the lobes is an important event in the activation of the catalytic domain. The lobes behave essentially as rigid bodies,
Acknowledgements
J.M. is the recipient of a research fellowship from Comunidad de Madrid. We thank the University of Alcalá Computing Centre and the CIEMAT (Madrid) for generous allowances of computer time on their SGI servers. Financial support from the Spanish CICYT (SAF2003-07219-C02) and the National Foundation for Cancer Research is gratefully acknowledged.
References (40)
- et al.
Purification and enzymatic characterization of pp60c-src from human platelets
J. Biol. Chem
(1990) - et al.
Effect of autophosphorylation on catalytic and regulatory properties of tyrosine kinase Src
Arch. Biochem. Biophys
(2002) - et al.
A protein tyrosine kinase involved in regulation of pp60c-src function
J. Biol. Chem
(1989) - et al.
Tyrosine phosphorylation regulates the biochemical and biological properties of pp60c-src
Cell
(1987) - et al.
Inter- and intramolecular interactions of highly purified Rous sarcoma protein pp60c-src
J. Biol. Chem
(1985) - et al.
A dynamic model for the allosteric mechanism of GroEL
J. Mol. Biol
(2000) - et al.
Dynamic coupling between the SH2 and SH3 domains of c-Src and Hck underlies their inactivation by C-terminal tyrosine kinase phosphorylation
Cell
(2001) - et al.
The structure of phosphorylated Gsk-3β complexed with a peptide, frattide, that inhibits β-catenin phosphorylation
Structure
(2001) - et al.
Specificity of bovine heart protein kinase for the delta-stereoisomer of the metal–ATP complex
FEBS Lett
(1979) - et al.
The conformational plasticity of protein kinases
Cell
(2002)
Transforming gene product of Rous sarcoma virus phosphorylates tyrosine
Proc. Natl. Acad. Sci. USA
The hunting of Src
Nat. Rev. Mol. Cell Biol
Protein tyrosine kinase structure and function
Annu. Rev. Biochem
Tyr527 is phosphorylated in pp60c-src: implications for regulation
Science
Characterization of sites for tyrosine phosphorylation in the transforming protein of Rous sarcoma virus (pp60v-src) and its normal cellular homologe (pp60c-src)
Proc. Natl. Acad. Sci. USA
Analysis of the sequence of amino acids surrounding sites of tyrosine phosphorylation
Proc. Natl. Acad. Sci. USA
Analysis of mutant forms of the c-src gene product containing a phenylalanine substitution for tyrosine 416
Oncogene Res
Activation of the pp60c-src kinase by middle T antigen binding or by dephosphorylation
EMBO J
Potential positive and negative autoregulation of p60c-src by intermolecular autophosphorylation
Proc. Natl. Acad. Sci. USA
Binding of transforming protein p47gag-crk to a broad range of phosphotyrosine-containing proteins
Science
Cited by (23)
Lck-dependent fyn activation requires C terminus-dependent targeting of kinase-active Lck to lipid rafts
2008, Journal of Biological ChemistryThirty years of molecular dynamics simulations on posttranslational modifications of proteins
2022, Physical Chemistry Chemical PhysicsTyrosine kinases: complex molecular systems challenging computational methodologies
2021, European Physical Journal B
- ☆
Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.jmgm.2004.06.001.
- 1
Present affiliation: Centro de Biologia Molecular “Severo Ochoa” C.S.I.C. Universidad Autonoma de Madrid, E-28049 Cantoblanco, Madrid, Spain.