Drugging the cell's best friend

The tumor suppressor p53 is one of the cell’s very best friends. Just how good a friend it is becomes apparent when, just like in other relationships, this particular relationship turns sour. p53 is the “master guardian angel” of the genome and constitutes the most frequent genetic alteration in cancer. More than 50% of human tumors contain a mutation in the p53 gene. With this kind of glowing track record, p53 would be a prime target for drugs.

It turns out that discovering drugs for p53 is trickier than you think. The protein displays complex structural biology, and the mechanism of inhibitor action is not clear. But p53 malfunction is also characterized by one of the most distinctive physical mechanisms to ever emerge in an oncoprotein- about 30% of mutations in p53 simply lower the melting temperature (Tm) so that the protein becomes unstable and disordered. Thus, potential inhibitors of mutated p53 have been often termed ‘rescuers’ since ideally they would ‘rescue’ the protein from its unstable state.

In a recent paper, a team led by Alan Fersht of Cambridge- one of the world’s foremost protein chemists and p53 experts- explores one frequent mutation in p53 and how its consequences could be suitably exploited for rational drug discovery. The study is a nice example of the value of interdisciplinary research in tackling a complex problem. The rogue mutation is quite simple; it turns a tyrosine on the protein surface to a cysteine. The change to a smaller amino acid opens up a cavity on the protein surface. Is this cavity druggable’, that is, can a small molecule be found that selectively and potently binds to this cavity? This is what the researchers seek to do in this study (The presence of the cysteine makes me wonder if someone has tried a covalent tethering strategy for targeting this site).

The targeted site is an interesting one. It’s not exactly an allosteric site, since it is far away from the functional site but does not seem to affect the functional site. But Fersht and his colleagues have previously found a small molecule that binds to this site and raises the melting temperature. In this report, the authors extent p53 inhibitor discovery for the Y220C mutation further by using a combination of experimental and computational techniques.

They start by screening a fragment library; fragment-based drug discovery is now a stable of rational drug design. To minimize false positives and negatives, they use two complementary techniques: a 1D NMR method and a technique called thermal denaturation scanning fluorimetry, which detects the effect of a ligand using an exogenous dye. The two methods interestingly gave quite diverse hits, indicating the wisdom of combining them. To further confirm the binding of the fragments to the protein, they then use N15/H1 HSQC, an NMR method that detects changes in the proton and N15 chemical shifts when a ligand binds to the protein. By comparing unbound protein shifts to bound ones, one can locate only the amino acids that interact with the ligand. This is a really nice method since one can make out the concerned amino acids just by inspection; in the figure below, the multi-colored areas indicate significantly perturbed amino acids, and it’s very useful to locate the exact binding sites since amino acids outside the site don’t seem to be affected. In this particular case the key residues turned out to be a valine, an aspartate and a few others.




The study proceeded by employing the one method which can confirm ligand binding better than any other- x-ray crystallography. Crystal structures revealed binding modes for a few of the hits. One molecule turned out to bind to the binding cavity in two copies.

What exactly are the hits doing to the protein? To investigate this, the authors used molecular dynamics (MD) simulations in isopropanol. In this case isopropanol is not a solvent mimic but it’s actually a drug mimic. It has a polar and non-polar part and it can approximate a typical protein-binding molecule well. Binding cavities can be detected by looking at the density of isopropanol in the pockets. In this particular example, the highest density is actually found in other sites, but those sites are not relevant for this study.

The most intriguing observation from the MD simulations was that the size of the cavity fluctuates wildly when the ligand is not present. This dynamic flexibility of residues can be characteristic of an unstable region (this was also observed in some computational enzyme designs that I described earlier). To validate this flexibility, the simulations were run with the ligand present; not surprisingly, the size fluctuation in the cavity reduces. Intriguingly, the ligand seems to play the role of the previous tyrosine in packing against the other residues and keeping the site stable.

There is some way to go before we have a bonafide drug that ‘rescues’ p53 from its ignominious fate. But this study is a nice illustration of how only interdisciplinary computational and experimental work can really help us to unravel the mysteries of this ubiquitous and enigmatic Jekyll and Hyde.

Basse, N., Kaar, J., Settanni, G., Joerger, A., Rutherford, T., & Fersht, A. (2010). Toward the Rational Design of p53-Stabilizing Drugs: Probing the Surface of the Oncogenic Y220C Mutant Chemistry & Biology, 17 (1), 46-56 DOI: 10.1016/j.chembiol.2009.12.011

What's an Iodide doing stabilizing a helix?

One of the most important- and least understood- effects dealing with biomolecular structure concerns the effects of salts on protein conformation. The famous Hofmeister Series for ions that either 'salt-in' or 'salt-out' proteins is well known, but the mechanism through which the ions act is controversial and probably involves not one mechanism but different ones under different circumstances.

In an interesting single-author JACS paper, Joachim Dzubiella studied the effects of different salts of sodium and potassium on the structure of alpha helices in solution. Even something as common and widely studied as the alpha helix is still an enigma. For example, the simple question "What contributes to the stability of an alpha helix?" is controversial and not fully answered yet. In this context I will refer the reader to an excellent perspective written by Robert Baldwin at Stanford that tries to answer the rather simple question: "How much energetic stability does a peptide bond in a helix contribute?". Baldwin looks at two approaches to understand the problem. One is the 'hydrogen bond inventory' approach which simply lists the bonds broken and formed on each side when an amide group desolvates and forms a peptide bond. Based on this approach, the mean figure for peptide h-bond energy has been estimated as 0.9 kcal/mol/h-bond. Even though this quantity is small, a 100 residue protein where 70 residues form hydrogen bonds is clearly going to lead to a very substantial net stabilization. The second approach that Baldwin considers is the electrostatic solvation enthalpy or free-energy method, where one uses the Born equation to estimate the strength of a h-bond. Using this approach Baldwin gets a very different answer- 2.5 kcal/mol. Clearly there is still some way to go toward estimating how much peptide h-bonds contribute to stability. One important factor not considered by Baldwin is the entropy of the water. Another important factor that he does consider is the preferential desolvation for helix formation that depends on the exact residues involved. We have ourselves encountered desolvation issues in continuing work on amyloid beta-sheets.

But back to Dzubiella's paper. Dzubiella uses MD simulations to study the dynamics of helix-salt interaction. He considers helices where a i---> (i+4) salt bridge between the side chains of a glutamate and lysine has stabilized the conformation. He looks at which salts stabilize helices and which ones destabilize them. From these detailed simulations he gains some valuable insight into the radically different behavior of rather similar ions. For example, K+ ions are much less able to destabilize helices than Na+ ions. This is due to preferential interaction of carboxylate groups involved in salt-bridge formation by Na+. Due to its smaller size, Na+ is better able to interact with carboxylates than K+.

However, we have to remember that Na+ or K+ or any of the other ions have to compete with water when interacting with amino acids in the peptide. Water is in great excess and water also efficiently interacts with carboxylates (1). The MD simulations reveal that a curious and unexpected helper comes to the aid of the Na+ ions- I- ions. Iodide interestingly interacts with the non-polar parts of the peptide, thus "clearing" water away and paving the way for Na+ to access the carboxylates and carbonyls. This unexpected observation again sheds light on the different properties of iodine compared to the rest of the halogens (2). Iodide is much bigger, has a diffuse charge and is therefore much more polarizable. Apparently it is so electronically watered down that even carbon thinks it is harmless and can preferentially interact with it.

This curious observation tells us that we know less about the elements than we think. From the observation of weak hydrogen bonds and halogen bonds to the unexpected non-polarity of iodide, surprises await us in the realm of biomolecular structure and indeed in all of chemistry. It is also thanks to tools like MD that we can now gain insights into the details of such molecular interaction.

Notes:
(1) In fact water can interact so well that it might steal a few h-bonds from the peptide and destabilize the helix. That's why trifluoroethanol (TFE) or hexafluoroacetone are so good at stabilizing helices (these lead to "Teflon-coated peptides"), because the fluorine cannot steal h-bonds from the peptide backbone.

(2) For example iodide most efficiently forms halogen bonds with oxygen, a phenomenon now well-accepted.

References:
Joachim Dzubiella (2008). Salt-Specific Stability and Denaturation of a Short Salt-Bridge-Forming α-Helix Journal of the American Chemical Society DOI: 10.1021/ja805562g

R. L. Baldwin (2003). In Search of the Energetic Role of Peptide Hydrogen Bonds Journal of Biological Chemistry, 278 (20), 17581-17588 DOI: 10.1074/jbc.X200009200

Elegant switches in an E. Coli ion channel

Molecular dynamics simulations comprise one of the most important tools in the armamentarium of chemists and biologists. Initially a curiosity for theoretical scientists, MD is now an explanatory and predictive tool in chemistry, biology, materials science, engineering and even weather prediction. In the field of biology, some masters such as Martin Karplus have honed this tool to the status of an art. While great leaps have been made in the context of methods, hardware, software and applications in this field, much remains to be still done. One of the reasons is that even now, running microsecond MD is computationally quite expensive. But many important biological events involving biomacromolecules take place on this time scale, thus making the achievement of efficient microsecond MD simulations important.

Writing in Science, scientists from D. E. Shaw company, Columbia and the Hebrew University of Jerusalem have a lovely paper documenting the application of the new and innovative MD program Desmond to the dynamics of a bacterial ion channel that transports Na+ ions using the electromotive force generated by proton transport. Desmond is supposed to enable efficient microsecond MD. It is going to be interfaced with the Maestro interface developed by Schrodinger and is due to be released this year I believe. Currently the fastest MD program on a single processor is GROMACS. While head to head comparisons of GROMACS and Desmond have not been reported to my knowledge, Desmond is supposed to be very fast on multiple processors, a facility that many can now afford to have.

In the Science paper, the researchers apply Desmond to understand the transport mechanism of the Na+/H+ antiport ion channel in E. Coli. This protein is crucial for E. Coli to survive harsh conditions of pH, alkalinity, and ionic lithium environments. The authors basically focus on the protonation state of certain key aspartates and find something pretty interesting- two crucial aspartates essentially act as switches that decide whether Na+ ions would be transported to the cytoplasm or to the periplasm. Using many long MD simulations involving different protonation states, the authors discovered that one of the carboxylates always has to be protonated. This acts like a "master aspartate" switch. Once this switch's state is set, it's the state of the other switch that decides the direction of transport- protonated leads to expulsion of the Na+ into the periplasm, while deprotonated leads to expulsion into the cytoplasm.

The observation reminded me of a high-school "staircase lighting" electricity experiment. A master switch had to be always on for the assembly to work. The On/Off state of another switch would then govern whether current flowed or not.

Using this discovery as the basis for exploring further conformational changes related to it, the authors come up with an elegant stepwise mechanism for the transport of Na+ and H+ ions that accounts for the observed stoichiometry of one Na+ ion for every two H+ transported. Using free energy perturbation binding affinity calculations, the authors also rationalize the channel's observed selectivity for Na+ over K+, and slightly for Li+ over Na+.

There is also a a very intriguing explanation for the pH sensitivity of this ion channel. The crystal structure of the channel is solved at pH 4, and it is inactive at this pH. How does the channel get activated at higher pH? To explore this, the authors do something simple but quite clever. They first identify all the key aspartates lining the channel and determine their pKa values. Perhaps not surprisingly, the pKa values of these are abnormally high- not an uncommon observation for amino acids in the unusual environments in protein interiors. They then selectively deprotonate one aspartate keeping all others protonated and do MD on the resulting structures. If there is a key "pH sensing" aspartate, its protonation state will likely govern a conformational change from inactive-active. Indeed, one aspartate, D133, turns out to modulate a conformational change involving two helices when it is deprotonated. Crucially, this results in the "master aspartate" noted above to move away from the Na+ entry/exit pathways. Thus it can no longer bind the ion, resulting in an inactive channel. Mutagenesis studies also support the observations.

A neat conclusion from a beautiful set of experiments. A fascinating example of how nature essentially and surprisingly uses high-school chemistry to modulate movements in complex proteins. And a highly successful and inspiring example of how efficient, long MD simulations can shed light on these crucial processes.

Arkin, I.T., Xu, H., Jensen, M.O., Arbely, E., Bennett, E.R., Bowers, K.J., Chow, E., Dror, R.O., Eastwood, M.P., Flitman-Tene, R., Gregersen, B.A., Klepeis, J.L., Kolossvary, I., Shan, Y., Shaw, D.E. (2007). Mechanism of Na+/H+ Antiporting. Science, 317(5839), 799-803. DOI: 10.1126/science.1142824

MD? Check...at least in this case

I have often talked on this blog about the fallacy of deriving single average conformations for molecules from average NMR data. How can a molecule with ten rotatable bonds (and frequently more) have any single or dominant conformation in solution? And yet, one repeatedly comes across publications dealing with single average conformations derived for organic molecules.

Peptides, being favourite targets for solution conformational analysis, are often subjected to such analysis. One comes across many instances where people want an alpha helix, or a 3/10 helix, or a gamma turn. They then design a peptide which they think is constrained to form that motif. They get their average NMR data (coupling constants and distances from NOE data), and then do a constrained molecular dynamics run, and see their favourite motif on the screen. But of course, what you put in is what you get out. If you are constraining the molecule to fit the NMR data in the first place, it's not going to stray away from it, and not surprisingly you see what you want to see (Isn't this a general theme and trap for so many of us...to be tempted to see what we want to see?).

In general, molecular dynamics is not a very efficient method of searching conformational space. Plus, every MD program uses some force field, and force fields are well-known to give fairly accurate geometries if well parametrized, but not good energies. So there is no guarantee that all the "low-energy" conformations of any molecule will be sampled. But in theory, if one could actually search all of conformational space with a good force field, then one would explore the whole ensemble of structure that a molecule adopts in solution. Even with faster computers though, that goal is going to unrealizable for a general situation.

However, MD can be a good way of exploring conformational space if the molecule is highly constrained in the first place. In such a case, it can serve as a good way of deconvoluting the average NMR data. And this is prcisely what Nikiforovich and Marshall, two researchers who do seem aware of the general problem of assigning multiple conformations, have done (J. Med. Chem, DOI: 10.1021/jm070084n). They have taken a cyclic pentapeptide, Pro-Ala-Ala-Ala-Ala, and done long 100 nanosecond (which is a huge amount of time for a MD run) on the molecule. Most importantly, their MD runs are unconstrained. After running the MD with two different force fields and different starting velocities, they consistently see the same results; a major conformer without a gamma turn, and a minor one (about 6%) with a gamma turn, a usually forbidden motif. They also gratifyingly note that the average distances from their ensemble of MD structures match well with the NMR data.

This analysis was feasible and realistic because of a couple of reasons; firstly, the molecule is small (although it still has 8-10 rotatable bonds) and one can actually do a 100 ns MD run on this. Secondly, they used many different conditions that converged to the same two conformational profiles, and thirdly they used unconstrained MD to remove any bias. I am not completely satisfied with it though, because the molecule still does have 10 rotatable bonds and is quite flexible. But it is a ring, and some of the bonds are not as flexible, so in general it is constrained and I think the conclusions are quite believable.

For once, it gave me a satisfied feeling to see an analysis that gave results that are quite likely to be true, complete, and sweet.

P.S. Garland Marshall also has one of the more memorable quotes in science that I have heard:

"Nature does not shave with Ockham's Razor!"

Schrodinger's equation

A friend of mine just returned from a conference in New York organised by Schrodinger, and I have to say that Schrodinger really seems to be poised to be the one-stop shop for all things computational.

They already have some great programs in their Maestro suite, including Glide for docking, which you find folks in industry using more and more these days. In their next revisions, they are going to introduce a program named PrimeX for doing crystallography, which will perform analysis similar to CNS, which will be groovy if it brings such analysis to the desktop. They are also going to introduce electron-density fitting for loop refinement in proteins. Right now, loop refinement of, say a 10 residue loop takes forever. But with PrimeX and friends, one can have constraints effected by electron density to restrict conformational searching, thus greatly speeding up the process.

Other products include the very impressive new Glide XP docking protocol. I have been glued to their site ever since they published their admirable paper in 2006. I have already written about the capabilities of GlideXP. This is really the best of computational chemistry applied to docking, where you find chemists trying to include as many experimental parameters as they can in a program. Schrodinger is definitely one company whose chemists have a firm and steady hand on experimental variables.

A very important development is going to be the interfacing of William Jorgensen's MCPRO, a program for doing free energy perturbation (FEP) calculations. FEP calculations are as close as you can come to accurately reproducing experimental binding free energies, one of the holy grails of computational methodology. While GlideXP astoundingly claims to also be able to do that, it would be super to have a GUI and easy operability for a good FEP program at your fingertips. Admittedly, FEP works only for ligand which differ little in their structure (eg. Me vs H). But that's also the phenomenon which we understand the least, how "similar" ligands can have great differences in binding affinity, something which FEP should help us understand.

Other improvements will include better parameters in standard docking, and a new force field, OPLS 2008, which will be "better than MMFF". Considering that the force behind this field is Tom Halgren, the same guy who meticulously crafted MMFF, I would be looking forward to it. There is also talk of a new MD program comparable to Gromacs, AMBER etc. which can do millisecond MD efficiently. That would probably complete the list of capabilities in one program that almost any computational chemist could want.

What I like best about Schrodinger is that it has people at its helm who are among the best that computational chemistry has to offer, most importantly Richard Friesner and Tom Halgren. Looking at their papers, it's clear that like ideal computational chemists, they thoroughly understand experimental data, and clearly know what the limitations of their programs are.