About this Author
DBL%20Hendrix%20small.png College chemistry, 1983

Derek Lowe The 2002 Model

Dbl%20new%20portrait%20B%26W.png After 10 years of blogging. . .

Derek Lowe, an Arkansan by birth, got his BA from Hendrix College and his PhD in organic chemistry from Duke before spending time in Germany on a Humboldt Fellowship on his post-doc. He's worked for several major pharmaceutical companies since 1989 on drug discovery projects against schizophrenia, Alzheimer's, diabetes, osteoporosis and other diseases. To contact Derek email him directly: Twitter: Dereklowe

Chemistry and Drug Data: Drugbank
Chempedia Lab
Synthetic Pages
Organic Chemistry Portal
Not Voodoo

Chemistry and Pharma Blogs:
Org Prep Daily
The Haystack
A New Merck, Reviewed
Liberal Arts Chemistry
Electron Pusher
All Things Metathesis
C&E News Blogs
Chemiotics II
Chemical Space
Noel O'Blog
In Vivo Blog
Terra Sigilatta
BBSRC/Douglas Kell
Realizations in Biostatistics
ChemSpider Blog
Organic Chem - Education & Industry
Pharma Strategy Blog
No Name No Slogan
Practical Fragments
The Curious Wavefunction
Natural Product Man
Fragment Literature
Chemistry World Blog
Synthetic Nature
Chemistry Blog
Synthesizing Ideas
Eye on FDA
Chemical Forums
Symyx Blog
Sceptical Chymist
Lamentations on Chemistry
Computational Organic Chemistry
Mining Drugs
Henry Rzepa

Science Blogs and News:
Bad Science
The Loom
Uncertain Principles
Fierce Biotech
Blogs for Industry
Omics! Omics!
Young Female Scientist
Notional Slurry
Nobel Intent
SciTech Daily
Science Blog
Gene Expression (I)
Gene Expression (II)
Adventures in Ethics and Science
Transterrestrial Musings
Slashdot Science
Cosmic Variance
Biology News Net

Medical Blogs
DB's Medical Rants
Science-Based Medicine
Respectful Insolence
Diabetes Mine

Economics and Business
Marginal Revolution
The Volokh Conspiracy
Knowledge Problem

Politics / Current Events
Virginia Postrel
Belmont Club
Mickey Kaus

Belles Lettres
Uncouth Reflections
Arts and Letters Daily
In the Pipeline: Don't miss Derek Lowe's excellent commentary on drug discovery and the pharma industry in general at In the Pipeline

In the Pipeline

« If You Want Your Explanations Overnight, It'll Cost You | Main | Walking the Plank at Bristol-Myers Squibb »

September 11, 2006

Enzymes Do Whatever They Want To

Email This Entry

Posted by Derek

It's been a while since I wrote about the neuraminidase inhibitors (Tamiflu and Relenza, oseltamavir and zanamivir). As we start to head into fall, though, I'm sure that avian flu will invade the headlines again, if nothing else (and I hope it's nothing else).

There's an interesting report in Nature (subscriber link) on how these drugs work. Bird flu is a Type A influenza, but there are two broad groups inside that class, which are defined by what variety of neuraminidase enzyme they express. (There are actually nine enzyme variants known, but four of them fall into one group and five into the other).

The drugs were developed against group-2 enzymes, but they're also effective against group-1 influenzas. Since the X-ray crystal structures showed the the drugs bound in the same way to all the group-2 neuraminidases, and since the active sites of all the subtypes across the two groups are extremely similar, no one ever thought that their binding modes would be different. Well, until last month, anyway, when the X-ray crystallographic data came in.

And what it showed was that the active sites of the group-1 enzymes, sequence homology be damned, have a much different structure than the group-2s. As it turns out, though, they can adopt a similar shape when an inhibitor binds to them, which is why the marketed inhibitors still work on them, but they're fundamentally quite different.

I can't resist the urge to use this example to illustrate some of the real problems in our current state of the art for computation and modeling. The differences between these two enzymes are due to their different amino acid residues far away from the active site, which makes modeling them much, much more difficult (and makes the error bars much, much wider when you do). That's why no one realized how far off the group-1 and group-2 neuraminidases were until the X-ray structure was available: modeling couldn't tell you. Any modeling efforts that tried would probably have decided, incorrectly, that the two groups were nearly identical. Why shouldn't they be?

But if we'd had that X-ray data from the start, modeling would very likely have told you, incorrectly, that there was little chance that either Relenza or Tamiflu would work on the group-1 enzyme variants. Why should they? The "induced fit" binding modes, where the enzyme changes shape significantly as the ligand binds, are understandably very difficult to model. There are just too many possibilities, too many of which are within each other's computational error bars.

Now, it's true that this latest work isn't based on molecular modeling at all. (You have to wonder how close these guys got, though). But plenty of projects that are using it are just as much in the dark as a neuraminidase team would have been, and they may not even realize it. Most molecular modelers are well aware of these limitations, but not all of them - or all of the managers over them - are willing to accept them. And when you get out to investors or the general public, it's all too easy for modelers or managers to act as if things are perfectly under control, when in reality they're lurching around in the dark. Like the rest of us. . .

Comments (11) + TrackBacks (0) | Category: In Silico | Infectious Diseases


1. secret milkshake on September 12, 2006 12:57 AM writes...

Over-reliance on computer modeling often leads to a lousy drug design.

Permalink to Comment

2. kiwi on September 12, 2006 6:53 AM writes...

it happens the other way too, a couple of enzymes i'm familar with have very low block sequence homologies, down in the noise. but when the x-ray data came out, voila! the active site residues overlay like carbon copies. imho there are serious issues with in silico screening, especially related to fit- induced fit tyoe mechanisms. i guess your resaults at the end of the day are only as good as your assumptions.

Permalink to Comment

3. MolecModeler on September 12, 2006 8:28 AM writes...

Speaking as a modeler, I would say that having unrealistic expectations of what modelling can do can lead to lousy drug design. But as we all know, there are many more factors that lead to bad drug design (i.e. taking stupid hits from HTS). Few modelers truly understand how terrible most modern programs are, and managers at senior levels have NO clue at all. But guess who looks bad if things don't work out? Yeah, us.

Permalink to Comment

4. Derek Lowe on September 12, 2006 8:46 AM writes...

MM, I've known others in your situation. Once you start dealing with a level of management that hasn't had a good deal of math/physical chemistry/comp sci experience, which in some places you reach pretty quickly, your ability to explain the limits of simulations is gone.

All they see is a wonderful package that the brochures say will help them discover drugs more cheaply - and who, outside of someone with a suspiciously negative attitude - could be opposed to that?

Permalink to Comment

5. MolecularGeek on September 12, 2006 10:40 AM writes...

I think the real problem is that there really aren't that many easy targets left out there. Sure, it's easy to recreate methotrexate from the DHFR crystal structure, but hindsight is almost always 20/20. Everyone remembers those lovely QSAR models that showed that all that mattered in inhalation anesthetics was logP.

The low hanging fruit has already been plucked. As an erstwhile modelling type, I think that looking for a magic bullet for the drug design problem is pointless. Use any and all technologies possible, but be skeptical of all of them until someone has actually made the compound and (in the best of all possible worlds) tested them in two independently designed assays.

I guess that's good advice not just for the scientists, but also for management in general. You would think that the people who start out in sales and marketing would learn to recognize their own snake oil.


Permalink to Comment

6. MolecModeler on September 12, 2006 10:57 AM writes...

The final test of any model/structure-based suggestion is the compound being made and tested; nothing less will do.

One of the real bad things about being a modeler (which I didn't appreciate before I got a job in industry), is that our performance is really hard to evaluate. We rely on our synthetic colleagues to make compounds, and so our success is intimately tied to their time/willingess/co-operation. It's no fun knowing that your future is not entirely in your hands.

Our bane is protein flexibility. "New" scoring functions/force-fields/whatever are a waste of time and should be soundly ignored. This is why a good modeler has to have a good understanding of protein flexibility as well as chemical sense to know what is reasonable to propose. Sometimes you need to make suggestions which are not based on any calculation, like for example trying to nudge Met/Phe/Tyr residues aside (which are of course rigid in docking simulations by and large).

Permalink to Comment

7. Ashutosh on September 12, 2006 11:55 AM writes...

Largely I have come across two kinds of synthetic chemists; those who underestimate the capability of what we do, and those who overestimate it. However, one thing I can say, which has been also iterated by other computational chemists I have met, is that synthetic chemists many times don't think about conformation and structure as much as they think about configuration, and end up reaching unrealistic conclusions about conformation. I am not blaming them, because when you are doing a total synthesis, you often don't care what the solution conformation is. But the problem really begins when top synthetic chemists publish a SINGLE conformation in solution for some flexible molecule. This phenomenon has been seen time and time again. It's actually obvious if we think about it, that a molecule with any degree of flexibility simply cannot exist as one single conformation, or even as one dominant conformation (dominant being more than 50%). In fact, the single conformation deduced from NMR data is the average, often 'virtual' and high energy conformation, which is completely unrealistic. Conformational deconvolution, which combines good NMR data with a good conformational search, is a must to find out the ensemble of structures for a solution conformation, and this has worked well in many cases. It's a little surprising how synthetic chemists just go ahead and publish a single conformation for some long chain polyketide.

As far as modeling protein flexibility goes, one needs to make a general assesment rather than trusting one specific parameter or result. For example, using one of these new induced fit modeling programs, one simply cannot trust scores or the ranking of poses. However, if a pose repeatedly shows up in the top 20 or 30 cases, it has a good chance of being close to the real one. I appreciate that the problem is that there are many exceptions, but benchmarking studies help, and some recent methodologies like the MMGBSA docking methodology have worked well for many cases. I do agree that it is wrong to blindly put your faith in almost any kind of calculation except perhaps very high level quantum chemical theory for small molecules. But sometimes modeling definitely can provide very good leads, although it is hard to see this except in retrospect.
In the case of neuraminidase, I remember a study of some very similar inhibitors with strikingly different potencies against neuraminidase, that Peter Kollman had done. After some pretty taxing computational work, he could rationalize the subtle differences in nonpolar contacts that the inhibitors had with the active site as contributing to their potencies.

J. Med. Chem.; (Article); 2003; 46(26); 5628-5637. DOI: 10.1021/jm030060q

If the interaction is allosteric, yes, I don't think we have the computational ability to predict such differences. But then, can we even do it experimentally with aplomb?

Permalink to Comment

8. Kent G. Budge on September 13, 2006 5:23 PM writes...

Not an organic chemist, but enjoy the articles. This one brought to mind some pop culture references:

"It's My Reaction, And I'll Catalyze It If I Want To."

"Enzymes Just Want To Have Fu-un..."

Hope they lighten your day.

Permalink to Comment

9. KinasePro on September 13, 2006 7:05 PM writes...

I'm a synthetic chemist with an advanced degree in Pymol~ I've used it to great effect, and yah the 'ol modeler / chemist debate is a good one alrighty.

Imnsho, a good understanding for the synthetic chemistry, the physical chemistry, the existing SAR, knowledge of the competive landscape, and a healthy dose of computational chemistry can all come together to make for some good 'ol fasion medicinal chemistry.

but umm... LEADS..... from MODELING?????

heh, good -->> luck

Permalink to Comment

10. tom bartlett on September 15, 2006 10:54 AM writes...

"Imnsho, a good understanding for the synthetic chemistry, the physical chemistry, the existing SAR, knowledge of the competive landscape, and a healthy dose of computational chemistry can all come together to make for some good 'ol fasion medicinal chemistry."

SOOOO right!. A med chemist is a master chef; you need JUST the right smidgens of ingredients.

Permalink to Comment

11. Don Vanselow on July 30, 2007 7:41 PM writes...

And another thing. There is very little or no evidence, for any particular protein, that the native structure is the same as the X-ray crystal structure in all respects. No account is taken of flexibility or subunit re-arrangement. See my exploration of new native structures of neuraminidases at

Permalink to Comment


Remember Me?


Email this entry to:

Your email address:

Message (optional):

The Last Post
The GSK Layoffs Continue, By Proxy
The Move is Nigh
Another Alzheimer's IPO
Cutbacks at C&E News
Sanofi Pays to Get Back Into Oncology
An Irresponsible Statement About Curing Cancer
Oliver Sacks on Turning Back to Chemistry