Scaling further GPCR summits

There's a nice review on GPCRs and their continuing challenges in the British Journal of Pharmacology this month. The authors focus on both structural and functional challenges in the characterization of this most important class of signaling proteins. As is well-known, drugs targeting GPCRs generate the highest revenue among all drugs. And given their basic roles in signal transduction, GPCRs are also clearly very important from an academic standpoint. Yet there is a wall of obstacles confronting us.

For starters there are the well-known problems with crystallization plaguing all membrane proteins like GPCRs. Until now only four GPCRs- rhodopsin, beta1 and beta2 adrenergic receptors and A2a adenosine receptor- have been crystallized, and the publication of each structure was considered a breakthrough. As the review mentions, the proteins are unstable outside the membrane and conditions for stabilization and crystallization are frequently incompatible; for instance stabilization is often effected by long-chain detergents while the opposite is true for crystallization. To circumvent these problems clever strategies have been adopted and immense trial and error and hard work were required. The rhodopsin and adrenergic receptors were crystallized by point mutations and special techniques; in one case an antibody was tethered to the protein and in another case a fusion protein was attached to stabilize the domain.

It's when we enter the dense jungle of GPCR biology that crystallization problems almost start sounding trivial. GPCRs couple to a variety of ligands including well-known biogenic amines (like adrenaline and serotonin), peptides, proteins and nucleotides. Where is starts to become complex is in the kind of response these ligands elicit, which could be full agonism, partial agonism, inverse agonism and full antagonism.

What structural features distinguish these different responses from each other? This is a key question in GPCR biology. But not only can ligands be agonists or antagonists but they can act in different ways on the same GPCR, activating different pathways. The case of partial agonists is especially interesting and more protein-partial agonist structures would be quite valuable.

The traditional model of protein binding assumes two dominant states, inactive and active. Agonists stabilize the active state, antagonists stabilize both states, and inverse agonists stabilize the inactive state. But, as the authors say, the traditional model is slowly undergoing a revision:

The concept of a receptor existing in a simple pair of active and inactive states (R and R*) is no longer sufficient to explain the observations of pharmacology. Agonists vary considerably in their efficacy and how this relates to the bound conformational states is unclear. A partial agonist with 50% efficacy could fully activate 50% of the receptors or could activate 100% of the receptor by 50%. Alternatively, a partial agonist might stabilize a different form of the receptor to a full agonist state and this different conformation might activate the G protein with a lower efficiency. The study of rhodopsin suggests that activation of the receptor involves the release of key structural constraints within the E/DRY and NPxxY regions. Energy provided by agonist binding must be sufficient to break these constraints and stabilize the new active conformation. In the case of rhodopsin, whether this transition is complete or partial depends on the chemical nature of the ligand (Fritze et al., 2003). The retinal analogue 9-demethyl-retinal is a partial agonist of rhodopsin which only poorly activates G protein in response to light. Spin-labeling studies (Knierim et al., 2008) suggest that in the presence of this ligand, only a small proportion of receptors are in the active conformation equivalent to all-trans-retinal. However, this can also result in a new state that is not formed with the full agonist. Therefore, rhodopsin studies suggest that that partial agonism may result in either a reduced number of fully active receptors or conformations which are not capable of fully engaging the signal transduction process. Structures of other GPCRs in complex with partial agonists are required to determine their effects on conformation.

An example makes the hideous complexity clear. The mu-opioid receptor is activated by several ligands including morphine, etorphine and fentanyl. However, morphine acts only as a partial agonist in effecting a phosphorylation endpoint whereas the other two act as full agonists. But it gets more interesting. While morphine effects phosphorylation of the kinase ERK through activation of PKC (protein kinase C), etorphine also activates ERK but by activation of beta-arrestin. Thus the same endpoint can be effected through different pathways. And it doesn't even stop there. Morphine causes the phosphorylated ERK to stay in the cytoplasm while etorphine causes the ERK to translocate to the nucleus. Not done yet; in addition, morphine can reverse its role and act as a full agonist on the adenylyl cyclase pathway.

Thus, the same ligand adopts different roles when activating different pathways. To begin with it's not even clear which pathway is activated under what circumstances. And the problem is only accentuated by the participation of different G proteins in inducing different responses.

Another dense layer of complexity is added by the fact that GPCRs have been found to dimerize and oligomerize. Crystallography can often be misleading in studying these dimers since there are several documented reports of dimers being formed as misleading artifacts of the crystallization conditions.

Apart from the stated problems, there are even more differences in further downstream signaling and receptor internalization induced by oligomerization. It's clearly a jungle out there. No wonder the design of drugs targeting GPCRs needs a measure of faith. For instance consider the various drugs targeting CNS proteins. CNS drug discovery has long been considered a black box for a good reason. Once a drug enters the brain, one can imagine it not only targeting a diverse subset of GPCRs (and even other classes of proteins) but, given the above complexities, also acting separately as agonist and antagonist at the various receptors. We clearly have a long way to go before we can prospectively design a CNS drug that will do all this on cue.

It would be a tall order trying to explain all these differences simply through structural modifications induced by the ligands. Yet whatever signal is eventually transmitted to the G proteins must begin with a crucial structural movement. It seems that elucidating the differences in helix and loop movements induced by partial and full agonists, inverse agonists and antagonists is a tantalizing part of the GPCR puzzle.

Since crystal structure data on GPCR is lacking, modeling approaches especially based on homology modeling have proved especially fruitful. Earlier attempts were all based on the single rhodopsin template. Since then the higher resolution adrenergic and adenosine receptor structures have provided significant insight. But here again numerous caveats abound. Modeling the helices is relatively easy since all GPCRs share the same general 7TM helix topology which is highly conserved, but modeling the fine differences between helices that lead to structural changes upon ligand binding is harder. And most difficult and important of all is modeling the extracellular loops which actually bind the ligands. Subtle changes in loop movement, salt-bridge breakage, hydrophobic effects and interaction of loops with helices is difficult to model. Often a change in conformation of a single residue can be enough to throw the modeling off balance. Nonetheless, the paucity of structural data means that modeling when done right will continue to be valuable. In the absence of structural data, computational ligand-based approaches which search for ligands similar to known compounds could be useful.

We have made a lot of progress in understanding the structure and function of these key proteins. But investigations seem to have unearthed more questions than answers. Which is always good for science since then it can have more choice fodder for contemplation.

Congreve, M., & Marshall, F. (2009). The impact of GPCR structures on pharmacology and structure-based drug design British Journal of Pharmacology DOI: 10.1111/j.1476-5381.2009.00476.x

Zheng, H., Loh, H., & Law, P. (2010). Agonist-selective signaling of G protein-coupled receptor: Mechanisms and implications IUBMB Life DOI: 10.1002/iub.293

Linkland

1. A crystal structure of the important PI3 kinase delta form which may provide insight into designing selective inhibitors of this key kinase implicated in cancer.

2. The discovery that a class of sandalwood odorants targets both the traditional GPCR odorant receptors but also the unexpected estrogen receptor (ER). The targeting of these two apparently functionally unrelated proteins may suggest roles for ORs different from olfaction.

3. And speaking of smell, the identification of odorant receptors in malaria-causing mosquitoes. The researchers identify receptors responding to specific body odors that could help the insects home in. Maybe they can also identify dietary components that cause/eliminate these?

4. And finally, the world's first 1GHz NMR in France. Pacemakers beware.

Indian village has unusually low rates of Alzheimer's disease

This caught my attention recently.

As the sun breaks through the morning mist in Ballabgarh, the elders of the village make their way to their regular meeting spot to exchange stories and share a traditional hookah pipe.

These men are in their sixties and seventies, while their faces bear the evidence of years of hard work in the fields, their minds are still sharp.

In other parts of the world, people of their age would be at some risk of developing dementia. But here, Alzheimer's disease is rare. In fact, scientists believe recorded rates of the condition in this small community are lower than anywhere else in the world.

Apparently the villagers here, mostly farmers, were tested for the ApoE4 gene which has been indicated as a risk factor for Alzheimer's. ApoE4 frequency was the same as in a population of farmers in rural Pennsylvania. Unfortunately the explanations suggested (vegetarian diet, lack of obesity, low cholesterol levels, physically fit farmers) does not seem to be unique to this farming community.

In contrast with lives in Pennsylvania and other parts of the world, the people of Ballabgarh are unusually healthy. It is a farming community, so most of them are very physically active and most eat a low-fat, vegetarian diet. Obesity is virtually unheard of.

Life in this fertile farming community is also low in stress, and family support is still strong, unlike in other, more urban parts of India.

"It all leads to a happy body, and a happy mind and hopefully a happy brain," says Dr Chandra. "Cholesterol levels here are much lower. We believe that is what is protecting the community."

There must surely be other farming communities in India and other places whose residents have a "happy body and a happy brain". I need to look up the original reference.

Simulations long enough to...put you to sleep

For something as widely used for as long as general anesthetics (GAs), one would think that their molecular mechanism of action would have been fairly understood. Far from it.

From Linus Pauling's theory of gases like xenon acting at high concentrations by forming clathrates to more recent theories of GA action on lipids and now on proteins, tantalizing clues have emerged, but speculation remains rife.

In a recent Acc. Chem. Res. review, a group of researchers explains some recent studies on GA action. Now there's a field that to me seems primed for computational studies. This is for two reasons.

Firstly, experimental information on GAs is hard to come by. Consider their chemical features; halogenated, apolar small molecules lacking polar hydrogen bonding and other interactions, binding to their targets with low affinity (it's interesting that halogenation seems to be a key criterion for GA action). In addition most GAs do not bind to a highly specific active site but instead influence protein action indirectly. Such features make any kind of NMR or x-ray structure determination an enormous challenge.

Secondly, molecular dynamics simulations (MDS) have come of age. With recent programs augmented by tremendous gains in hardware and software, microsecond to millisecond simulations have gradually become a reality. This particular field seems to provide a classic and worthy challenge for MDS, since GAs seem to interact indirectly and subtly with proteins by influencing their local and global dynamics rather than binding to well-defined active pockets. Such dynamic perturbations would fail to be captured during the pico to nanosecond timescales typically sampled by MDS. For instance, the most prevalent belief for GAs right now is that they interact with ligand-gated ion channels like the GABA and NMDA receptor and with potassium ion channels. One hypothesis for the mode of action of halothane is that it binds to the open conformation of a potassium ion channel. The channel stays open for milliseconds, thus thwarting experimental study. However, a millisecond transition provides a robust and respectable challenge for long time-scale MD simulations.

At the same time, caveats abound in the field. For instance it's easy to infer that a GA molecule binds to a certain site and obstructs the motion of a tyrosine residue, thus providing support to fluorescence quenching and other studies. But the results of such studies as well as the all-important site-directed mutagenesis studies are notoriously hard to interpret; indirect influences on protein motion may be construed as direct binding to particular sites. Plus, it seems to me that one can read too much into the mere, rather obvious observation that a molecule binding to a protein site inhibits the motion of some residues; whether that observation translates into a realistic phenomenon may be much harder to glean.

So yes, it seems that GA action provides a fertile field for computer simulation. Long MD simulations generally seem to me to be a solution looking for problems; after all most interesting molecular interactions in the body take place on the order of micro to milliseconds. There is a huge number of important problems waiting to be tackled with such tools. However, interpretations of the results will always have to be guided by the sure hand of experiment, with the always important caveat that when it comes to interpretation, one computational study and one experiment can have several offspring.

Vemparala, S., Domene, C., & Klein, M. (2010). Computational Studies on the Interactions of Inhalational Anesthetics with Proteins Accounts of Chemical Research, 43 (1), 103-110 DOI: 10.1021/ar900149j

When radiation misfires...literally

The New York Times has a rather chilling account of how radiation overdose in the treatment of some cancer patients caused deadly side effects leading to death. The entire sobering article deserves to be read. In one case a man's tongue was going to be selectively irradiated; instead his whole face received a blast of radiation that led to a horrible, slow death. Scott Jerome-Parks's story makes for very painful reading. In another case, misguided radiation beams literally cut out a hole in a woman's chest that gradually killed her. This was Alexandra Jn-Charles. Both Mr. Jerome Parks and Ms. Jn Charles died within a month of each other in 2007.

And all this mainly because of computer errors that were not detected by human beings, errors that caused the radiation to be overdosed or misdirected. Seems like one of those classic "technology is a double-edged sword" kind of scenarios with the whole system just becoming too complex for human understanding. In one instance, a wedge in a linear accelerator delivering the radiation was supposed to focus the beam in the "in" position. But the computer that used Varian software- the same software that I used in grad school for operating the NMR spectrometer built by the same company- made a mistake and instead pivoted the wedge to the "out" position, removing the radiation shielding. The mistake was not detected 27 times, leading to acute radiation overdoses in the wrong parts of the body. In the case of the man whose tongue was supposed to be treated, an error in the software failed to save the critical settings for the accelerator which would have focused the radiation to the right parts. The computer repeatedly crashed, leading to the collimator beams being left wide open, and nobody noticed this.

The statistics unearthed by the Times are startling. From 2001 to 2009, more than 600 cases of improper radiation treatment were reported. Out of those, 255 were related to an overdose, while 284 were related to the wrong parts of the body being exposed to radiation. Even in its idealized form radiation has side-effects, so one would assume that doctors and technicians would be deathly serious about operating these protocols. These statistics were collected for New York State, which is apparently supposed to have some of the strictest radiation standards in the country.

What is even more shocking is the lack of transparency due to "privacy laws". Names of the culprits have been withheld, and some of them seem to have been let off the hook with a simple reprimand. St. Vincent's hospital and University Hospital of Brooklyn, where the two accidents had happened, were simply fined a thousand dollars by the city of New York. Some doctors who have participated in the treatments refused to talk to the journalists. There also does not seem to be a single agency responsible for these radiation safeguards. On top of it all there seem to be scant ways for patients to pick beforehand which hospital they would like to receive radiation treatment in, since records of mistakes are not available to the public. The whole shebang sounds appalling.

Now I understand that 600 cases in 8 years is probably small potatoes compared to the total number of cases in which radiation has worked successfully. Nonetheless, the factors responsible for the lapses and the horrendous consequences deserve scrutiny (seriously, death due to "computer error" sounds like something out of a bad science fiction horror movie). For something as serious as radiation treatment for cancer, one would assume that the same kinds of safeguards, fail-safe mechanisms and backup checks would be in place as are used in nuclear reactor safety. What boggles my mind is that there exist no fail safe mechanisms which would simply shut down the system when they detect an overdose. It simply seems that shoddy training, computer error, and lack of accountability are dealing out death and enormous physical and psychological suffering to patients and their families.

Go, learn some linear algebra

When I was taking math classes in college I enjoyed topology, differential equations, calculus and combinatorial math, but somehow could not bring myself to drum up enthusiasm for linear algebra.

If I had picked physics as my major (which I almost did), I would not probably have escaped from the clutches of linear algebra while learning quantum mechanics. As it happened I picked chemistry, and most of the quantum chemistry that was served to me after that was sans linear algebra.

On his blog luysii has an excellent set of notes on linear algebra from a QM class that he audited. As I mentioned on his blog, it's interesting how much one can get away with in QC without linear algebra. Thus, take a look at some classic textbooks- Levine, McQuarrie and the classic Pauling and Wilson- and one can go a long way with very little LA. The only things that you are really required to know are eigenvalues and eigenfunctions, but even then the Dirac notation is usually skipped in elementary QC. About the only QC book I know which utilizes large doses of LA is the sophisticated book by Szabo and Ostlund.

Yet LA matters and as luysii demonstrates, there is a generality and elegance to it. There is at least one key LA theorem which is mandatory knowledge in QC. When you are learning about the variational principle (which is used to find approximations to the ground state energy of a system), you derive the so-called secular equation by utilizing a very important LA theorem; that a set of linear homogeneous equations has a non-trivial (non-zero) solution if and only if the determinant of the coefficients is zero. Further on, matrices also come into play in important ways when you are learning about the calculation of transition states, normal modes, and energy minima in molecular mechanics. In the latter exercise you have to calculate the Hessian matrix and then diagonalize this monstrosity (thank god for computer programs).

Perhaps it's not surprising that QC can go a long way without much linear algebra. QC is an application of QM to problems of chemical interest, and the whole reason why the Schrodinger formulation of quantum theory became hugely more popular than the equivalent Heisenberg matrix formulation was that it was more tractable to applications (essentially plug in the correct expression for the potential energy) and couched in the more familiar 19th century language of differential equations. If you wish to know about matrix mechanics take a look at Max Born's excellent book "Atomic Physics"; I had to give up on that particular section.

But even the great Erwin's celebrated paper introducing his equation was titled "Quantization as an Eigenvalue Problem". Maybe it is worth even for a "quantum engineer" (as the late Wolfgang Pauli once somewhat derisively called Enrico Fermi) to learn some linear algebra.

How much chemistry can we wring out of the universe?

Chemiotics II (luysii) had a very interesting post on his blog about the number of proteins of a given length that can be constructed from the entire mass of the earth. Comparing the masses of amino acids to the mass of the earth, he demonstrated that all the earth's mass will be pretty much exhausted with all combinations of a protein that's only 41 amino acids long, which is peanuts as far as your typical protein goes. Such calculations have great relevance for the origin of life if we are to understand the design and evolution of biomolecules.

One can ask similar questions about crystals or small organic molecules. For the latter one can similarly show that the number is much more than the number of atoms in the universe. But most naturally occurring organic molecules have a preponderance of certain fragments like benzene rings. Similarly, there are only a certain rather small number of symmetry groups for crystals. Therefore it seems that in reality, we are dealing with modular units which are much smaller in number (although still quite large) rather than the bare individual units which compose proteins/small molecules/crystals. Thus once these modular units evolved, natural selection probably worked on them instead of trying out possible combinations of their individual atoms. Also remember that natural selection can work on a population of individuals- any kind of individuals- if one of them shows even the slightest advantage with respect to replication. In case of sequences of amino acids, such replicative advantages could arise from several features; stability, charge distributions that could serve to protect the sequences from aqueous hydrolysis or attract one sequence to another, or conformational flexibility that could serve to effect flexibility in the functions of the sequence. Any one of these features could serve to "fix" a particular sequence or group of sequences in a pool of sequences.

In case of proteins for instance, one should ponder how many of the many possible sequences considered could be energetically favored. Some sequences that pit bulky or similarly charged amino acids next to each other could be disfavored by steric and electrostatic factors. Also in case of proteins, the conservation of 3D structure relative to sequence must have been a boon for natural selection. For instance, there's an enormous number of sequences that can fold up into alpha helices (although certain amino acids are favored and others are disfavored) or sheets (where amino acid preferences are not as pronounced). Thus one gets the feeling that natural selection could have some flexibility in designing sequences that would fold into energetically favored secondary structural motifs. However this would not work as well for the active sites of enzymes, where very specific amino acids need to be located in very specific positions in order to effect catalysis. But even here, certain amino acids such as histidine and lysine are interchangeable in terms of their acid-base catalysis roles.

A particularly interesting case that comes to my mind is that of amyloid. Once thought to be the province of only proteins like ß-amyloid, it has now been extensively shown (most notably by Christopher Dobson of Cambridge University, for instance see Nature Chemical Biology 5, 15 - 22 2009, doi:10.1038/nchembio.131 ) that virtually any protein can form amyloid under the right conditions. Amyloid may have been evolution's dream, since it could have tremendous flexibility in picking sequences and coercing them to form amyloid-like structures under the right conditions. As work in which I participated demonstrated (Biochemistry, 2008, 47 (38), pp 10018–10026, DOI: 10.1021/bi801081c), the simplest of changes in conditions like temperature and pH are enough to drastically modulate the architecture of amyloid assemblies.

Thus, while there was potentially an infinite pool of possibilities to design proteins from, as evolution proceeded, I think that the funnel of possibilities became narrower and narrower as the units needed to achieve optimum design became more tailored and building-block like. It's a very interesting question to contemplate the details of this matter.