Molecular Dynamics Use in Personalized Cancer Medicine

Example of MET Y501C Mutation

Igor F. Tsigelny

1,2,3

, Razelle Kurzrock

Åge Aleksander Skjevik

2,4

Valentina L. Kouznetsova

1,2

and Sadakatsu Ikeda

Moores Cancer Center, Univesity of California at San Diego, La Jolla, California, U.S.A.

San Diego Supercomputer Center, Univesity of California at San Diego, La Jolla, California, U.S.A.

Department of Neurosciences, Univesity of California at San Diego, La Jolla, California, U.S.A.

Department of Biomedicine, University of Bergen, Bergen, Norway

Keywords: Personalized Cancer Medicine, Molecular Dynamics, Structure of Proteins, Sema Domain, MET, c-Met.

Abstract: We explored a possible new method for prediction of activating mutations in cancer-related proteins. This

method is based on elucidation of flexibility of proteins associated in activating complexes. Based on the

theory of intermediate binding complexes, the binding process is not only related to the three-dimensional

structure of proteins, but also to the four-dimensional set of possible conformations allowed by the flexible

regions of the involved members of the associated complex. Using molecular dynamics simulations, we found

that an Y501C mutation in the MET gene might activate it. Using this information, a specific drug that

functioned as a potent MET inhibitor was prescribed and had a salutary impact on the tumor.

1 INTRODUCTION

Elucidation of possible changes in protein activity

based on an aberration in its sequence is one of the

most important tasks of personalized medicine

(Tsigelny et al., 2015). One of the most powerful

databases in cancer medicine is Cosmic (Forbes et al.,

2011). Even if one would extract only the 50 most

frequent mutations in cancers in 600 proteins (which

are used most frequently for cancer diagnostics), it

would give 30000 aberrations. In some cases, such

aberrations are described in the literature and related

databases. Nevertheless, more than 90 percent of

them are not covered. Here, we address the problem

of how to elucidate possible activity changes that

occur due to these mutations. Replacement of a

residue in a pdb file may help if we have a simple case

when the aberration effect is obvious: change of a

charged residue to an oppositely charged residue in a

salt bridge, or insertion of a hydrophobic residue

instead of a hydrophilic residue that participates in

hydrogen bonding etc. At the same time, the effect of

a residue substitution might not be that obvious in

other cases. Here molecular dynamics (MD)

simulation might help elucidate the changes in the

ensemble of conformers that could affect the activity

of the protein or protein complex.

2 MET STRUCTURE AND

FUNCTION

The tyrosine kinase MET is a receptor for the ligand–

hepatocyte growth factor (HGF). It is known to be

involved in cancerogenesis. When activated (in many

cases because of a single amino acid replacement

mutation), it affects cells in a number of organs

creating invasive cancers (Stamos et al., 2004,

Montesano et al, 1991). The MET receptor has

significant structural similarity to Ron and Sea

receptors (Ronsin et al., 1993; Huff et al., 1993,

Stamos et al., 2004).

In order to be activated, MET requires binding of

its ligand—hepatocyte growth factor (HGF) that is

active only after proteolytic conversion to a two chain

configuration (Stamos et al., 2004; Hartmann et al.,

1992). Direct binding sites show that HGF-beta

chains bind to the extracellular domain of MET with

a K

of about 90 nM. Accordingly, binding of HGF is

crucial to activation of MET (Figure 1). Analysis of

the MET–HGF interface shows a set of moderately

complementary side chains on both sides.

Tsigelny, I., Kurzrock, R., Skjevik, Å., Kouznetsova, V. and Ikeda, S.

Molecular Dynamics Use in Personalized Cancer Medicine - Example of MET Y501C Mutation.

DOI: 10.5220/0005959500710074

In Proceedings of the 6th International Conference on Simulation and Modeling Methodologies, Technologies and Applications (SIMULTECH 2016), pages 71-74

ISBN: 978-989-758-199-1

Figure 1: Interaction of the extracellular Sema domain of

Met (green) with the HGF-beta chain (brown).

2.1 Sema Domain

The Sema domain of MET forms a conformation

sometimes referred to as “seven-bladed beta-

propeller” in the shape of a funnel with an inner

diameter of 25 Å and a total diameter of around 50 Å.

The blades of this propeller are formed by antiparallel

beta-strands. The Sema domain is stabilized by the

interactions between the C- and N-terminal residues,

and the beta-propeller structure is stabilized by the

seven disulphide bridges also found in a number of

proteins with notable homology in the amino acids

included in these domains. The HGF beta-chain

associates with the Sema domain at the bottom face

of the propeller with at least seven electrostatic pas

interactions between the two proteins (Stamos et al.,

2004).

In the studied case, there is a mutation Y501C in

the Sema domain of MET. This residue is located at

the interface between the C- and N-terminal of the

Sema domain. From the general point of view used in

elucidation of possible activity changes of the MET–

HGF complex, a substitution of tyrosine to cysteine

would not make any changes in activity unless

cysteine is involved in a disulphide bond (not in this

case). Another possibility is that this mutation

happens in the N-C-terminal interface of the Sema

domain and affects its flexibility.

3 FLEXIBILITY OF BINDING

PROTEINS IMPROVES THEIR

ASSOCIATION

As we pointed above, the activation of the MET–HGF

complex significantly depends on interaction

between these two proteins. As was shown by Levi

and colleagues (2005), who studied more than 100

protein–protein complexes, the flexibility of the

binding partners is one very important feature that

often defines the process of protein–protein

association. In other words, the binding process is not

only related to the three-dimensional structure of

proteins, but also to the four-dimensional set of

possible conformations adopted by means of the

flexible regions of the involved proteins. It is

interesting to note that so-called transition-state

conformational ensembles for general folding of

proteins and their binding have similar characteristics

(Levi et al., 2005). Taking into consideration the

abovementioned concept, we hypothesized that

increasing flexibility of the Sema domain of MET

would increase its interactions with HGF and

consequently improve binding between these proteins

and thereby increase activation of the entire complex.

4 MUTATION Y501C AFFECTS

THE FLEXIBILITY OF THE

SEMA DOMAIN

In order to elucidate possible changes in flexibility of

the Sema domain, we conducted 300 ns MD for the

wild-type and mutated versions of the protein. Our

results show significant increase in flexibility of

several important parts of the Sema domain (Figure

2). The violet rectangles encompass the most flexible

regions of the mutant structure. It is interesting to note

that the maximum flexibility changes occur in the

Figure 2: Superposition of the 300 ns conformers, with the

wild-type represented by brown ribbons of the alpha-trace

and the Y501C mutant of the Sema domain of MET shown

as blue ribbons.

SIMULTECH 2016 - 6th International Conference on Simulation and Modeling Methodologies, Technologies and Applications

regions of the Sema domain that are in direct contact

with the HGF protein when bound (Figure 3). Note

that the regions around residues 150 and 210 of the

Sema domain are in direct contact with HGF residues

during its binding. These results suggest that the

mutation Y501C of MET leads to a significant

increase in the flexibility of the MET Sema domain

in the regions contacting HFG. Such changes may

improve the binding and consequently the activity of

the MET–HGF complex.

Figure 3: Flexibility (defined by B-factor values) of the

Sema domain residues in the wild-type (brown) and Y501C

mutant (blue) conformers as calculated from a 300 ns MD

trajectory.

5 PATIENT TREATMENT BASED

ON THE MD SIMULATIONS

A patient was diagnosed with hepatocellular

carcinoma (HCC), with the MET Y501C (tyrosine to

cysteine) missense mutation that was elucidated from

circulating tumor DNA. Based on the results of MD

simulation, we suggest that this mutation activates

MET kinase and consequently has oncogenic effects.

The patient received cabozantinib—a MET inhibitor.

This drug administration caused significant (65%)

reduction of alphapheto protein (AFP), a tumor

marker for HCC.

6 CONCLUSIONS

Molecular dynamics simulations can be used to

elucidate the four-dimensional ensembles of possible

conformers and flexibility of the binding partners in

protein–protein complexes and can help in decision

making for physicians in cancer therapy.

7 METHODS

In order to generate the Y462C Sema domain mutant,

Tyr462 in the wild-type protein was replaced by a

cysteine while the rest of the protein structure

remained unchanged. Disulfide bridges between the

relevant cysteine pairs in each protein were

generated. The SEMA domain of MET protein has

been extracted from the complex with heparin

(Stamos et al., 2004) pdb ID 1shy.

We conducted molecular dynamics (MD)

simulations for both the wild-type and mutated

versions of Sema domain of MET. The two proteins

were each placed in an octahedral water box

consisting of about 48,500 TIP3P water molecules

and 9 neutralizing K

ions modelled by

Joung/Cheatham ion parameters (Joung and

Cheatham, 2008). The simulations were conducted

using the GPU/CUDA-accelerated version of

PMEMD implemented in the AMBER14 software

suite (Case et al, 2014; Goetz et al., 2012; Salomon-

Ferrer et al., 2013A, B), with the protein described by

AMBER ff14 SB parameters. Each of the two protein

systems were subjected to the following

minimization/simulation steps: i) Unrestrained

minimization for 10,000 steps; ii) Gradual constant

volume heating from 0 to 100 K over 5 ps with

restraints applied to the protein backbone; iii) Gradual

constant pressure heating to 310 K over 100 ps with

restraints applied to the protein backbone; iv) 300 ns

unrestrained constant pressure simulation at 310 K.

The Langevin thermostat (Loncharich et al.,

1992) was applied for regulation of temperature with

a 1.0 ps

-1

collision frequency, and the pressure was

regulated isotropically during the second heating step

and the production simulation by means of the

Berendsen barostat (Berendsen et al., 1984) at a

reference pressure of 1.0 bar. Bond lengths for bonds

involving hydrogen were constrained using the

SHAKE algorithm (Rycjaert et al., 1984), allowing

for a time step of 2 fs. Periodic boundary conditions

were applied, and the particle mesh Ewald (PME)

method (Roe et al., 2013) was used for the evaluation

of electrostatics.

REFERENCES

Berendsen, H. J. C., Postma, J. P. M., van Gunsteren, W. F.,

DiNola, A., Haak, J. R., 1984. Molecular dynamics

100

150

200

250

300

350

400

125 225 325

B‐factor

Residuenumber

Flexibilityoftheintialand

mutatedSemadomains

Molecular Dynamics Use in Personalized Cancer Medicine - Example of MET Y501C Mutation

with coupling to an external bath J. Chem. Phys., 1984,

81: 3684–3690.

Case, D. A., Babin, V., Berriman, J. T., et al., 2014. AMBER

14, University of California, San Francisco.

Darden, T., York, D., Pedersen, L., 1993. Particle mesh

Ewald: an Nlog(N) method for Ewald sums in large

systems, J. Chem. Phys., 98: 10089–10092.

Forbes, S. A., Bindal, N., Bamford, S., Cole, C., Kok Chai

Yin, Beare, D., Jia, M., Shepherd, R., Leung, K.,

Menzies, A., Teague, J. W., Campbell, P. J., Stratton,

M. R., Futreal, P. A., 2011. COSMIC: mining complete

cancer genomes in the Catalogue of Somatic Mutations

in Cancer, Nucleic Acids Res. (England), 39 (Database

issue): D945–D9450.

Götz, A. W., Williamson, M. J., Xu, D., Poole, D., Le

Grand, S., Walker, R. C., 2012. Routine microsecond

molecular dynamics simulations with AMBER on

GPUs. 1. Generalized Born, J. Chem. Theory Comput.,

8: 1542–1555.

Hartmann, G., Naldini, L., Weidner, K. M., Sachs, M.,

Vigna, E., Comoglio, P. M., Birchmeier, W., 1992. A

functional domain in the heavy chain of scatter

factor/hepatocyte growth factor binds the c-Met

receptor and induces cell dissociation but not

mitogenesis, Proc Natl Acad Sci USA, 89: 11574–

11578.

Huff, J. L., Jelinek, M. A., Borgman, C. A., Lansing T. J.,

Parsons JT, 1993. The protooncogene c-sea encodes a

transmembrane proteintyrosine kinase related to the

Met/hepatocyte growth factor/scatter factor receptor,

Proc Natl Acad Sci USA, 90: 6140–6144.

Joung, I. S., Cheatham, T. E., III, 2008. Determination of

alkali and halide monovalent ion parameters for use in

explicitly solvated biomolecular simulations, J. Phys.

Chem. B, 112: 9020–9041.

Levy, Y., Cho, S. S., Onuchic, J. N., Wolynes, P. G., 2005.

A survey of flexible protein binding mechanisms and

their transition states using native topology based

energy landscapes, 346: 1121–1145.

Loncharich, R. J., Brooks, B. R., Pastor, R. W., 1992.

Langevin dynamics of peptides: the frictional

dependence of isomerization rates of N-acetylalanyl-

N'-methylamide, Biopolymers, 32: 523–535.

Montesano, R., Matsumoto, K., Nakamura, T., Orci, L.,

1991. Identification of a fibroblast-derived epithelial

morphogen as hepatocyte growth factor, Cell, 67: 901–

908.

Roe, D. R., Cheatham, T. E., III, 2013. PTRAJ and

CPPTRAJ: software for processing and analysis of

molecular dynamics trajectory data, J. Chem. Theory

Comput., 9: 3084–3095.

Ronsin, C., Muscatelli, F., Mattei, M. G., Breathnach, R.,

1993. A novel putative receptor protein tyrosine kinase

of the met family. Oncogene, 8: 1195–1202.

Ryckaert, J.-P., Ciccotti, G., Berendsen, H. J. C., 1977.

Numerical integration of the Cartesian equations of

motion of a system with constraints: molecular

dynamics of n-alkanes, J. Comput. Phys., 23: 327–341.

Salomon-Ferrer, R., Case, D. A., Walker, R. C., 2013. An

overview of the Amber biomolecular simulation

package, WIREs Comput. Mol. Sci., 3: 198–210.

Salomon-Ferrer, R., Götz, A. W., Poole, D., Le Grand, S.,

Walker, R. C., 2013. Routine microsecond molecular

dynamics simulations with AMBER on GPUs. 2.

Explicit solvent particle mesh Ewald, J. Chem. Theory

Comput., 9: 3878–3888.

Stamos, J., Lazarus R. A., Yao, X., Kirchhofer, D.,

Wiesmann, C., 2004. Crystal structure of the HGF b-

chain in complex with the Sema domain of the Met

receptor, EMBO, 23: 2325–2335.

Tsigelny, I. F., Wheler, J. J., Greenberg, J. P., Kouznetsova,

V. L., Stewart, D. J., Bazhenova, L., Kurzrock, R. 2015.

Molecular determinants of drug-specific sensitivity for

epidermal growth factor receptor (EGFR) exon 19 and

20 mutants in non-small cell lung cancer, Oncotarget,

6: 6029–6039.

SIMULTECH 2016 - 6th International Conference on Simulation and Modeling Methodologies, Technologies and Applications