Breast Cancer and Her2 Kinase: Projects 9104-9114

We are continuing to make a big push into studying cancer. Next up, is work relevant for breast cancer. Specifically, we have started to study the Her2 Kinase, a part of the EGFR family of Tyrosine kinases, responsible for initiating a host of biochemical pathways. These kinases are critical for regulating cell division and thus mutations within the EGFR family have been linked to various types of cancers, including breast and pancreatic cancer.

The aim of projects 9104 to 9114 is to understand the effect of certain mutations in the kinase domain of Her2. We are also hoping to find new druggable states within the system for creating the next generation of targeted cancer therapeutics, as well as to study the effect of mutations, which will give us insight into mutations present in breast cancer tumors.


Model of ATP bound to Her2 Kinase

A discussion of recent FAH work on ab initio nanoreactor

What’s an ab initio nanoreactor for?

In an ab initio nanoreactor, molecules are allowed to react freely with each other over the course of the molecular dynamics simulation, and then we observe what products come out of it and how the products were formed. Besides obeying the fundamental laws of physics, no additional assumptions were imposed to the system, hence ab initio.

The number of reactant molecules used to seed the simulations was small (50-100 molecules) compared to the number of molecules typically used in experimental methods, but is nonetheless very large from the standpoint of quantum chemistry calculations. To make the reactions occur more rapidly, we periodically push the molecules to the center of the ab initio nanoreactor with a virtual piston. What this does is to make the molecules bump into one another more frequently, and also provide the energy required for certain reactions to take place.

The significance of ab initio nanoreator

Traditionally, experimental methods are heavily relied on to discover new molecules and reaction pathways, and computational methods mainly played a supportive role to complement experimental methods. The results of this study prove that computational methods can also play the leading role in discovery, and can help guide experimental methods by posing new hypotheses and suggesting which experiments to do. It’s especially useful for detecting complex chemical reactions where several things happen at the same step during the reaction process that’s hard to detect via experiments.

The potential applications of ab initio nanoreactors are broad. Because of the ab initio approach coupled with some refinement methods and automatic analysis, we can achieve the goal of discovering new molecules, new reaction pathways and mechanisms in many different settings and environments. For instance, it could contribute to out future understanding of the origin of life, birth of stars, means to increase the rate of chemical reactions, earth’s atmosphere, etc.

Results of the study

We carried out two ab initio nanoreactor simulations. The first simulation started with purely acetylene molecules, and we call it acetylene nanoreactor. The second simulation started with a mixture of chemicals postulated to exist in the early earth atmosphere. The second simulation is the computational version of a famous experiment conducted in 1952 (Urey-Miller experiment) that showed complex building blocks of life could form from simple inorganic molecules (1). We call the second simulation Urey-Miller nanoreactor.

For the acetylene nanoreactor, nearly 100 distinct products were formed after ~500 picoseconds simulation time. Many of these product molecules are large (up to over 70 atoms) due to the tendency of acetylene molecules to form long chains and 3D networks. These products are also diverse, for example some have rings some don’t; some are linear some are branched. After comparing our results with those of previous experiments, we found that the acetylene nanoreactor produced not only similar products, but also new products (2, 3).

For the Urey-Miller nanoreactor, the products were relatively small (up to 16 atoms). Among the discovered products, we have amino acids (which are what proteins consist of), urea (participating in metabolism, and the first byproduct of life to be synthesized in the lab) and a bunch of other molecules, all of which have also been detected in meteorites that may have delivered organic molecules to the early earth (4). Many of these molecules are also found in interstellar clouds (5). In addition to the high diversity of products, the Urey-Miller nanoreactor also identified a complex network of reactions (more than 700 distinct reactions). A significant fraction of these reactions are viable in the common environment we live in. Moreover, we found out that water and ammonia allow reactions to proceed faster with less energy for many of these reactions. Last but not least, hydrogen was found rarely involved in the synthesis of a naturally occurring amino acid, glycine, which supports previous proposals that molecules that tend to lose electrons (including hydrogen) don’t participate in biomolecule formations (4).

Method of analysis

To derive insight from a complex network of reactions, we focus on a particular molecule in the network and investigate the reactions it’s involved in, either it’s the product or the reactant. In this way, it allows us to trace the synthetic pathways that lead from the starting molecules. There could be several different pathways to get from the starting material to our molecule of interest. Some intermediate molecules are more common than the others among these distinct pathways.


(1) Miller, S. L. & Urey, H.C. Organic Compound Synthesis on the Primitive Earth. Science 130, 245-251 (1959). Doi: 10.1126/science.130.3370.245

(2) Trout, C.C. & Badding, J.V. Solid State Polymerization of Acetylene at High Pressure and Low Temperature. J. Phys. Chem. A 104, 8142-8145 (2000).

(3) Sakashita, M., Yamawaki H. & Aoki, K. FT-IR Study of the Solid State Polymerization of Acetylene Under Pressure. J. Phys. Chem. 100, 9943-9947 (1996).

(4) Danger, G., Plasson, R. & Pascal R. Pathways For the Formation and Evolution of Peptides in Prebiotic Environments. Chem. Soc. Rev. 41, 5416-5429 (2012).

(5) Menten, K. M. & Wyrowski, F. in Sterstellar Molecules: Their Laboratory and Interstellar Habitat (eds Yamada, K. M. T. & Winnewisser, G.) 27-42 (Springer Tracts in Modern Physics 241, Springer, 2011).

Everything else described here is from Wang, L. P., Titov, A. McGibbon, R., Liu, F., Pande, V. S. & Martinez, T. J. Discovering Chemistry With An ab initio Nanoreactor. Nature Chemistry. 2014. Doi: 10.1038/nchem.2099. The article can also be read about on Nov 10th issue of C&E News:

Ab initio nanoreactor discovers new reaction pathways

Some very exciting research by Pande Group members Lee-Ping Wang and Robert McGibbon in collaboration with the Martinez Lab was recently published in Nature Chemistry. They report the development and application of the ab initio nanoreactor—a highly accelerated first-principles molecular dynamics simulation of chemical reactions that discovers new molecules and mechanisms.

Using the nanoreactor, they showed new pathways for the amino acid glycine’s synthesis from primitive compounds proposed to exist on the early Earth, which provide new insight into the classic Urey–Miller experiment. These results highlight the emergence of theoretical and computational chemistry as a tool for discovery.

The nanoreactor simulations were made possible by GPUs and the TeraChem quantum chemistry software; these technologies accelerate the calculation over conventional CPU codes by 10-100x.

Below is the nanoreactor simulation of the classic Urey–Miller experiment.

Bryostatin and Projects 9000-9015

Steven Ryckbosch, a graduate student in the Pande Group recently presented his work on Bryostatin. Folding@home projects 9000-9015 are running simulations to help answer the questions he has about it’s structure and function.

Bryostatin is a naturally occurring marine molecule that shows promising and unique activity against several diseases (most notably, cancer, HIV/AIDS, HIV latency, and Alzheimer’s). Its main target, protein kinase C (PKC), is a signaling protein central to many cellular functions. In its active form, PKC binds its ligand and is associated with the cell membrane, but we currently lack structural information about this complex in its membrane microenvironment.

The simulations performed on FAH will help to provide a structure to the PKC-ligand-membrane complex. This is complicated by the fact that while other compounds such as the phorbol esters also bind to PKC, they exhibit extremely different effects in cells and organisms. The structure and dynamics of this complex would allow us to understand bryostatin and other ligands’ binding mode and thus how to modify and tune it’s structure to improve function or even create new functions as needed for new therapies in the clinic.

Some questions Steven and the group are trying to answer:

How can we use simulation to find protein-membrane structures?
How can ligands modulate protein-membrane interaction?
How are membranes affecting bryostatin function?
How can this inform our design of new bryostatin analogues?

Below is a molecular simulation model of bryostatin bound to PKC’s active site.


Bowman lab update on vision

About four months ago, we started a new set of projects to understand the dynamics of some of the key proteins involved in vision.  Now, we have about 600 microseconds of simulation and have begun some preliminary analysis of the data.  Excitingly, it appears we may already have captured the conformational change we were targeting.  More data will be needed to improve the statistical significance of our results, but we are increasingly confident that we’ll be able to begin understanding some of the conformational changes required for vision and, eventually , how mutations lead to various blinding diseases.

Project 10470 and T4 Lysozyme

(Guest post by Kyle Beauchamp from the Chodera lab.)

In the Chodera lab, we’d like to understand how drugs bind to proteins, particularly for challenging diseases such as cancer or Alzheimer’s. To get to this point, however, will require a lot of hard work on simple systems—systems where we already “know the answer”.

T4 Lysozyme has been a key model system for understanding protein stability (Matthews, 2010). A version of T4 Lysozyme—with mutation L99A—binds a number of greasy molecules like benzene (see picture, PDBID 3DMX). Our hope is that a better understanding of how T4 Lysozyme L99A binds various molecules could lead us to better models for drug binding. (Mobley, 2007).

Project 10470 simulates T4 Lysozyme mutant L99A. These simulations will be used to improve models for ligand binding.


New projects to help design selective inhibitors of protein methyltransferases

The Chodera lab has teamed up with Luo lab at MSKCC to study another important class of cancer targets: protein methyltransferases.

These are protein-modifying enzymes that catalyze the transfer of methyl groups to lysine or arginine residues as part of complex regulatory programs. While a number of cancers have alterations in protein methyltransferases, making them appealing targets for new anticancer therapeutics, it is not yet possible to fully understand their role in disease because of the current limited repertoire of compounds available to selectively inhibit these enzymes.

Spurred by recent encouraging results from the Luo lab in developing sinefungin scaffolds to selectively target key methyltransferases, we are working with them to better understand the origin of selectivity of these compounds, and to help them design new compounds that will allow researchers to better understand the roles these enzymes play in cancer and, eventually, develop potent new anticancer therapeutics.

Projects 10474, 10475, and 10476 study key protein methyltransferases NSD1, NSD2, and SETD2.


protein methyltransferase NSD1 (PDB ID 4h12)

Protein methyltransferase NSD1 (pdbid 4h12)

Investigating conformations accessible by Abl kinase- drug target for chronic myelogenous leukemia

Guest post by Sonya Hanson, postdoc in the Chodera lab.
(Project 10472)

We’re working our way through the kinase family here at the Chodera lab. You may have seen Danny’s post about EGFR earlier this year, and now we’ve started simulations of Abl kinase. Abl kinase has a special place in the history of cancer therapeutics, ‘dispelling the long-held myth that it was not feasible to develop selective inhibitors of key cell-signaling molecules as safe and effective medicines.’ Novartis’ development of the drug imatinib (or Gleevec commercially) to treat chronic myelogenous leukemia (CML) specifically targets a mutant Abl kinase that results from a chromosomal abnormality called ‘the Philadelphia chromosome’. There is even a book out recently that chronicles the development of Gleevec (The Philadelphia Chromosome).

While the success of imatinib was remarkable, many patients develop resistance to it and regress. A more recent drug targeting Abl kinase, ponatinib of Ariad Pharmaceuticals (Iclusig commercially), has been developed that overcomes some of these resistance mutations. However, now even ponatinib has been found to be susceptible to resistance mutants. With these simulations of Abl kinase, we are hoping to begin to understand a structural basis for the development of resistance mutants so we can develop drugs that anticipate and overcome them before the patient even has to experience regression. But to do this we will need many long timescale trajectories of Abl and later its mutants to achieve this deeper understanding of the development of resistance in CML. Additionally, this knowledge could inform models of resistance development in other cancers that result from kinase mutations or kinase up-regulation.

A new project to study early folding events in apomyoglobin

In a new NSF-funded collaboration, the Voelz Lab is working with the Roder Lab at Fox Chase Cancer Center to study early folding events in apomyoglobin.

Apomyoglobin (myoglobin without the heme group) is an extremely well-studied protein. In fact, mygolobin was the first protein to have its structure solved by x-ray crystallography (John Kendrew, 1958). At low pH, apomyoglobin assumes a “molten globule” state that is compact and only partially structured. Seminal experiments by Jennings and Wright (1993) showed that when apomyoglobin folds at normal pH, it goes through an early intermediate that closely corresponds to the low-pH molten globule state.

Now, more recent experiments from the Roder lab have revealed even more details of early folding events in myoglobin (Xu et al. 2012). Using Trp fluorescence spectroscopy in a continuous-flow fast mixer, the Roder lab have resolved the formation of up to four different conformational states, on timescales ranging from microseconds to milliseconds.

The Voelz Lab is working toward using molecular simulation to characterize these conformational states in atomic detail. Both the size of the protein (153 residues) and the timescale of early folding (~200 µs) make this a challenging problem to tackle, but we hope that simulations on Folding@home (coming soon!) combined with Markov State Model approaches will enable us to construct a highly detailed model of the early folding reaction, and new level of quantitative connection between simulations and experiments. In the years to come, this work will lead to new ways to combine computation and experiment to understand and fight human diseases.

Combining simulation and experiments to solve molecular structures

The Voelz lab has been making progress on combining simulation and experiments to solve molecular structures.

Most molecules do not have a single rigid structure in solvent. Instead, they exist in a range of different conformations. Stable proteins exist mostly in the folded conformation, but there is always a small fraction of population that is unfolded. Other molecules may have a very heterogeneous set of conformations, which can make determining their structure in solvent difficult. NMR experiments, for example, can be used for this, but structural information often gets “washed out” due to motional averaging.

Our new method, called BICePs (Bayesian inference of conformational populations) is a robust method to infer the populations of conformational states, using a combination of high-resolution computer modeling and information from experiments. We think BICePs will be very useful for determining the extent to which proteins and other molecules are well-structured in solution. In the future we plan to use it as a tool for designing well-structured mimics of proteins, called peptidomimetics. Our paper describing the new BICePs algorithm has been published in the latest issue of the Journal of Computational Chemistry.

Add your computer's power to over 327,000 others that are helping us find cures to Alzheimer's, Huntington's, Parkinson's and many cancers ...

... in just 5 minutes.

Step 1.

Download protein folding simulation software called



Step 2.

Run the installation. The software will automatically start up and open a web browser with your control panel.

Step 3.

Follow the instructions to Start Folding.

Stanford University

will send your computer a folding problem to solve. When your first job is completed, your computer will swap the results for a new job.

Download the protein folding simulation software that fits your machine.


Installation guide