Corante

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: derekb.lowe@gmail.com Twitter: Dereklowe

Chemistry and Drug Data: Drugbank
Emolecules
ChemSpider
Chempedia Lab
Synthetic Pages
Organic Chemistry Portal
PubChem
Not Voodoo
DailyMed
Druglib
Clinicaltrials.gov

Chemistry and Pharma Blogs:
Org Prep Daily
The Haystack
Kilomentor
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
ChemBark
Realizations in Biostatistics
Chemjobber
Pharmalot
ChemSpider Blog
Pharmagossip
Med-Chemist
Organic Chem - Education & Industry
Pharma Strategy Blog
No Name No Slogan
Practical Fragments
SimBioSys
The Curious Wavefunction
Natural Product Man
Fragment Literature
Chemistry World Blog
Synthetic Nature
Chemistry Blog
Synthesizing Ideas
Business|Bytes|Genes|Molecules
Eye on FDA
Chemical Forums
Depth-First
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
FuturePundit
Aetiology
Gene Expression (I)
Gene Expression (II)
Sciencebase
Pharyngula
Adventures in Ethics and Science
Transterrestrial Musings
Slashdot Science
Cosmic Variance
Biology News Net


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


Economics and Business
Marginal Revolution
The Volokh Conspiracy
Knowledge Problem


Politics / Current Events
Virginia Postrel
Instapundit
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

« The Limits of Free Scientist Chow? | Main | Inspiration Is Where You Find It »

February 24, 2009

Structure-Activity: Lather, Rinse, and Repeat

Email This Entry

Posted by Derek

Medicinal chemists spend a lot of their time exploring and trying to make sense of structure-activity relationships (SARs). We vary our molecules in all kinds of ways, have the biologists run them through the assays, and then sit down to make sense of the results.

And then, like as not, we get up again after a few minutes, shaking our heads. Has anyone out there ever worked on a project where the entire SAR made sense? I’ve always considered it a triumph if even a reasonable majority of the compounds fit into an interpretable pattern. SAR development is a perfect example of things not quite working out the way that they do in textbooks.

The most common surprise when you get your results back, if that phrase “common surprise” makes any sense, is to find that you’ve pushed some trend a bit too far. Methyl was pretty good, ethyl was better, but anything larger drops dead. I don’t count that sort of thing – those are boundary conditions, for the most part, and one of the things you do in a med-chem program is establish the limits under which you can work. But there are still a number of cases where what you thought was a wall turns out to have a secret passage or two hidden in it. You can’t put any para-substituents on that ring, sure. . .unless you have a basic amine over on the other end of the molecule, and then you suddenly can.

I’d say that a lot of these get missed, because after a project’s been running a while, various SAR dogmas get propagated. There are features of the structure space that “everybody knows”, and that few people want to spend their time violating. But it’s worth devoting a small (but real) amount of effort to going back and checking some of these after the lead molecule has evolved a bit, since you can get surprised.

Some projects I’ve worked on have so many conditional clauses of this sort built into their SAR that you wonder whether there are any boundaries at all. This works, unless you have this, but if you have that over there it can be OK, although there is that other compound which didn’t. . .making sense of this stuff can just be impossible. The opposite situation, the fabled Perfectly Additive SAR, is something I’ve never encountered in person, although I’ve heard tales after the fact. That’s the closest we come to the textbooks, where you can mix and match groups and substituents any way you like, predicting as you go from the previous trends just how they’ll come out. I have to think that any time you can do this, that it has to be taking place in a fairly narrow structure space – surely we can always break any trend like this with a little imagination.

Another well-known bit of craziness is the Only Thing That Works There. You’ll have whole series of compounds that have to have a a methyl group at some position, or they’re all dead. Nothing smaller, nothing larger, nothing with a different electronic flavor: it’s methyl or death. (Or fluoro, or a thiazole, or what have you – I’ve probably seen this with methyl more than with other groups, but it can happen all over the place). A sharp SAR is certainly nothing to fear; it’s probably telling you that you really are making good close contacts with the protein target somewhere. But it can be unnerving, and sometimes there’s not a lot of room left on the ledge when you have more than one constraint like this.

Why does all this go on? Multiple binding modes, you have to think. Proteins are flexible beasts, and they've got lots of ways to react to ligands. And it's important never to forget that we can't predict their responses, at least not yet and not very well. And of course, in all this discussion, we've just been considering one target protein. When you think about the other things your molecule might be hitting in cells or in a whole animal, and that the SAR relationships for those off-target things are just as fluid and complicated as for your target, well. . .you can see why medicinal chemistry is not going away anytime soon. Or shouldn't, anyway.

Comments (40) + TrackBacks (0) | Category: Drug Assays | In Silico | Life in the Drug Labs


COMMENTS

1. A-nony-muse on February 24, 2009 9:39 AM writes...

Derek: Haven't you just made the argument in support of why medicinal chemistry SHOULD go away, and soon? The absence of a better alternative to drug discovery does not necessarily justify continuing to spend billions of dollars in an irrational manner. Unless the US government is involved, of course, then it becomes trillions and is always justifiable.

Permalink to Comment

2. Derek Lowe on February 24, 2009 9:46 AM writes...

No, I don't think so. We still need drugs, and this is still the only way we know how to make them. Med-chem should only go away when we have something better, and I'm not holding my breath for that.

I take your point about the government being involved, though. That, apparently, is quite different, and A Good Thing For Everyone.

Permalink to Comment

3. biotechchap on February 24, 2009 9:57 AM writes...

Excellent intro to SAR and the challenges involved in small molecule drug discovery to a biologist like me. Thanks...

Permalink to Comment

4. Matthew on February 24, 2009 9:58 AM writes...

"We vary our molecules in all kinds of ways, have the biologists run them through the assays, and then sit down to make sense of the results."

I love it... the biologists sound like lackeys for the chemists, running their mundane "biological" experiments to validate the chemistry. You made me smile this morning. :)

Permalink to Comment

5. Joannes on February 24, 2009 10:20 AM writes...

One premisse (basically wrong, but we can't do better) in every (Q)SAR discussion is assuming that all compounds bind in exactly the same way, i.e. the aromatic rings of which you are studying the substitution pattern all perfectly overlap. A substituent in a "wrong" position will cause the whole molecule to move a little bit, or maybe even to flip around. Result: another SAR analysis that goes down the drain.

Permalink to Comment

6. NH_chem on February 24, 2009 10:38 AM writes...

Good chemists will make good molecules. The computational folks are getting better but not long ago, I was given a structure that was to be nM compound and it was crap. Funny how that works.

It is hunches and luck most of the time. That is why protein structures make me laugh and why you need to look at these things in solution (i.e. the power of NMR has helped this). A long way to go still...........

Permalink to Comment

7. Robert Bruce Thompson on February 24, 2009 10:38 AM writes...

When I think of medicinal chemists, I think of people who spend their days building keys and then go off to search the earth for a lock that their shiny new key fits.

Permalink to Comment

8. milkshake on February 24, 2009 10:38 AM writes...

the current industry model (of how drugs are being discovered) does not work well but the rot is in the prevailing management methods and with the dishonest, mercenary business practices - not in the science sector.

The medchem research is the less costly part of the drug discovery/development process - and it would be great to keep it at good distance from the corporate management. I would not mind to be on a medchem project that (if and when it produces a candidate) ends up with Pfizer - as long as I am not directed by their appartchiks and our research organization cannot get acquired and then disemboweled.

Permalink to Comment

9. Philip on February 24, 2009 10:39 AM writes...

What you just did was write an obituary for QSAR.

May it "Rest in Peace".

Permalink to Comment

10. Jose on February 24, 2009 10:57 AM writes...

Was QSAR ever really alive, except in cherrypicked post-hoc, dog and pony shows? Was there ever a full blown, fully operational QSAR constellation during a project?

Permalink to Comment

11. emjeff on February 24, 2009 11:14 AM writes...


"It's not what you don't know, but what you know that ain't so." Sounds like Will Rogers was a medicinal chemist...

Permalink to Comment

12. milkshake on February 24, 2009 11:22 AM writes...

I have seen recently with a kinase project (where we got multiple co-crystal structures with dissimilar ligands from HTS) that one class of our ligands ripped up a new hole in the ATP binding site - a new hole that was not there before, big enough to accommodate a phenyl group. This created very isoform-selective class of compounds because only one kinase isoform has this flexibility, the other would reject the substituent. (In unbound state the ATP binding sites of these isoforms look exactly the same - the AA sequence differences are buried further back).

Also, a Rho-kinase frug Fasudil has a major active metabolite, hydroxyfasudil that binds in the same active site - only upside down... It must have been great fun to work on a project with such a mixed SAR like this.

Permalink to Comment

13. medicnman on February 24, 2009 12:49 PM writes...

When I was introduced to SAR as an undergraduate by my PI, I realized it holds several roles for chemists.

Firstly, it does allow for quantification of predicted activity in many cases. Conversely, in many more cases, this information can be detrimental by excluding other molecular scaffolds through dead-leads that have lost their luster due to a SAR study.

Second, it is one tool in the box and not the "end all, be all" or holy grail for chemists; that is why scientists cannot simply plug in a target molecule and push the "make me a a molecule button" on a computer.

Third, and rounding out the beauty of SAR, is this: more information can be gathered through an SAR study through outliers than by any other models being used. My PI loved tight SAR lines just as any other scientist would, what he loved more were plucking the outliers from the data and determining the potential reason for its horrible fit. The lesson taken by myself from this was, "To have your data fit the model is useful, to have an outlier and understand why something does not fit the model is a chance to learn something new."

Since my comment may seem plebian by an experienced medicinal chemist, I will add a disclaimer: I am a novice in medchem by anyone's standards, but I do love the science.

Permalink to Comment

14. Bill on February 24, 2009 1:42 PM writes...

Many SAR investigations have involved chemical modification of known natural products. I recall a seminar presented a number of years ago by, if I recall correctly, John Clardy, pointing out how difficult it was to improve upon nature. This may explain lack of success. However, what is the alternative? Combinatorial chemistry doesn't seem to be the answer( can anyone tell me of any drugs on the market developed by combinatorial chemistry?). Are we left wwith serendipity?

Permalink to Comment

15. Lucifer on February 24, 2009 1:48 PM writes...

Good idea or bad idea? PD332,334 was somewhat innovative in that it did target an ion channel other than GABA. It might have other uses, who knows?

Chlorpromazine was a failed antihistamine used to reduce pre-operative anxiety, Iproniazid was developed an anti-mycobacterial drug, Chlordiazepoxide was a synthesis gone wrong. You get the point..
_________
NEW YORK--(BUSINESS WIRE)--Pfizer Inc announced today that, following a review of the development and commercial portfolios in the Primary Care Business Unit, it is terminating Phase 3 development programs for the investigational compounds esreboxetine for fibromyalgia and PD 332,334 for generalized anxiety disorder (GAD).
The decision to terminate these programs will enable the Business Unit to allocate additional resources to higher-potential development programs as part of its continuing effort to deliver greater value to patients and Pfizer shareholders.

Permalink to Comment

16. CMC Guy on February 24, 2009 1:52 PM writes...

Derek as always you well express that frequently experiments do not provide absolute clarity and must push forward the best one can.

#13 medicnman you/PI principles do not seem far off in the theory however, particularly for point three on outliers, there is often an overriding time pressure of hard deadlines on a project to select a lead candidate so do not get a chance to really fully consider anything outside the main funnel. Only on a couple rare occasions have I seen "odd balls" being further elaborated as potential back up compounds.

#4 Matthew you have it reversed as in most cases the medchemists are subservient to the biologists for getting the data. Its only when they work and communicate well together that can sort through the difficulties, especially when SAR not precise.

Permalink to Comment

17. Retread on February 24, 2009 2:04 PM writes...

Another source of the problem is probably that we don't understand protein structure very well (as proteins actually exist in the complicated and very concentrated chemical soup of the cell).

For years after the first Xray crystallographic results on proteins, people spoke about THE structure of a protein. That proteins have multiple conformations was apparent with the very first one to be solved (sperm whale myoglobin) -- there was no way for O2 to get inside the protein to the iron in the porphyrin ring given the Xray structure -- so we knew the protein had to 'breathe.'

Recently the battle between induced fit (Koshland) and lock and key has been joined once again. I'm not sure who is ahead at this point. What is the current thinking?

Permalink to Comment

18. Wavefunction on February 24, 2009 3:11 PM writes...

Induced fit vs lock and key; I guess it depends on the protein. Some proteins like urokinase are like rocks. Others are all around the place. We have only now begun to be able to somewhat model induced fit effects. Multiple crystal structures of proteins bound to different ligands are helping, but it's still a long way to go before modeling large-scale loop and domain motions that may enhance ligand binding. MD is gradually starting to handle such millisecond-long events. For instance see:

"A conserved protonation-dependent switch controls drug binding in the Abl kinase"
PNAS 2009 106:139-144
December 24, 2008, doi:10.1073/pnas.0811223106

and

"Dynamic control of slow water transport by aquaporin 0: Implications for hydration and junction stability in the eye lens"
PNAS 2008 105:14430-14435
September 11, 2008, doi:10.1073/pnas.0802401105

Permalink to Comment

19. DLIB on February 24, 2009 3:29 PM writes...

Does anyone have a suggestion as to how to change SAR into DSAR ( D = Deterministic :-) That is, what knowledge is required to be more predictive? Does anyone know?

Permalink to Comment

20. test on February 24, 2009 4:37 PM writes...

test

Permalink to Comment

21. Desolvator on February 24, 2009 4:44 PM writes...

I do not believe what you said:

"You’ll have whole series of compounds that have to have a a methyl group at some position, or they’re all dead. Nothing smaller, nothing larger, nothing with a different electronic flavor: it’s methyl or death..."

It's just impossible from basic physical principles. Think about what the maximal free energy for a methyl group could contribute to the binding? one log unit, or even two log units?

Please list a couple of published data, then I would think it over again :-)

Permalink to Comment

22. Jose on February 24, 2009 4:59 PM writes...

"It's just impossible from basic physical principles."

I think you should realize that proteins *appear* to violate basic physical principles on a pretty regular basis. Yes, there is likely to be some prefect methyl isostere out there, but is it synthetically tractable in under 30 steps? Does it have good PK/PD?

Permalink to Comment

23. Hap on February 24, 2009 6:08 PM writes...

Sounds like another beautiful theory slain by ugly reality.

Permalink to Comment

24. weirdo on February 24, 2009 6:17 PM writes...

"methyl or death" is a fact of life that most (not all) medicinal chemists have observed.

However, it has more to do with our definition of "death" than anything. If active is 100 nM, and "dead" is 10 micromolar, there is your two log units.

I can't show you data for a pretty simple reason: "dead" molecules tend not to show up in patent applications-- thus (in my company) they do not show up in publications.

Permalink to Comment

25. Anonymous BMS Researcher on February 24, 2009 6:41 PM writes...

Reading this and the previous post leads me to suggest doing some SAR-type experiments on how quickly food-like items placed in break rooms vanish.

Permalink to Comment

26. Anonymous BMS Researcher on February 24, 2009 6:44 PM writes...

And to respond to a scientific theme in the comments on this posting: MANY MANY MANY times at working group meetings I have heard the chemists say "we just cannot get away from a foobar group over here no matter what we try," anybody with actual experience in drug discovery has heard this.

Permalink to Comment

27. Vittorio Montanari on February 24, 2009 7:52 PM writes...

I had once the privilege to make a chemical having some anticancer (leukemia) activity, and selectivity (the other cell lines were not killed) as a total shot in the dark. The NIH was accepting any substance, if it was new, for testing. My rationale for submitting was that the thing was doable in one step (fact), and uncommonly lipophilic for its size and structure (my guess). Later, an NIH senior investigator even volunteered his time to rationalize the result, but his databases gave no rationale at the time. There is value to doable. I don't know enough to discuss QSAR, but let us please test them if we can make them? For example, the Harvard Neurodiscovery Center is inviting new drug-like substances for testing.

Permalink to Comment

28. desolvator on February 24, 2009 8:08 PM writes...

Certainly I could creat many examples of "methyl then death": things such as adding a methyl to H-bond donor (like OH or NH) since it disrupts a key interaction, but changing methyl into H usually does not have that big effect, so I would bet $5 for anything equal to or great than 2 log units :-)

Permalink to Comment

29. A Nonie Mouse on February 24, 2009 8:25 PM writes...

More often than not, said methyl group either induces some conformational change (or at least limits flexibility) or screens some polar group from something. Regardless, it happens all the time. It probably has been observed in almost every project that I have ever worked on.

Permalink to Comment

30. Mat Todd on February 25, 2009 4:08 AM writes...

I teach 8 hours of introductory medchem to some postgrad students who have little chemistry background. We cover lopP, sigma, Taft, Hansch equations i.e. basic stuff about what molecules are and how they might behave. We then go on to talk about the predictive ability of computers, and CoMFA analysis. We cover how to find drugs, and some combichem. I end with a discussion of an example JACS paper where several of these things are combined in one story.

But I'm very dissatisfied with this syllabus, and am redoing it from scratch. I don't even believe what I'm saying when I talk about Taft - I keep it as an exercise, illustrating only that it is possible to be quantitative about the physical chemistry of a series of analogs.

So: as practising medicinal chemists, what would you like students to know? i.e. if they came in to your company lab, day 1.

Permalink to Comment

31. petros on February 25, 2009 4:09 AM writes...

with regard to the methyl effect. I remember an article in Chemistry in Britain some 20 years ago on that topic. There must be moe out there

Permalink to Comment

32. Ex UK-Pharma on February 25, 2009 4:52 AM writes...

#4 "I love it... the biologists sound like lackeys for the chemists, running their mundane "biological" experiments to validate the chemistry. You made me smile this morning. :)"

Speaking as a former biologist in the "here's new compounds, screen please" game, I see a certain amount of truth in this statement ;) The SAR stuff was handled on the Chem side of things at my company - once assays were optimised and settled in, I tended to have a closer relationship with the chemist leads on projects than with the biology ones.

Permalink to Comment

33. anon the II on February 25, 2009 9:13 AM writes...

This is probably a cold thread but to # 30.

The industry has been pretty clear on this for a number of years and I think you may have passed by it in your first sentence. What I want is a lot of chemistry background. The term "Introductory Medchem" really doesn't make much sense. It's like "Introductory Surgery" or "Introductory Commercial Aviation".

Actually, it's not clear the industry want's anyone much anymore.

Permalink to Comment

34. Wavefunction on February 25, 2009 9:45 AM writes...

For a nice example of how a c-Kit kinase inhibitor was reengineered to improve potency and abrogate off-target cardiotoxic effects by the addition of a single methyl, see the following:

J. Clin. Invest. 117:4044–4054 (2007). doi:10.1172/JCI32373.

Permalink to Comment

35. Mat Todd on February 25, 2009 9:57 AM writes...

@Wavefunction - nice, thanks. I use 10.1021/ja960897t - older, but a good story for serendipity vs design.

Permalink to Comment

36. milkshake on February 25, 2009 2:50 PM writes...

Mat: Most of what you teach is probably of little value in real medchem project but it makes for impressive JACS articles. A practicing chemist does not need to calculate logP and polar surface area - just one look at the structure tell that the compound is too greasy or has too many NH groups. By the way more realistic stuff is to be found in J Med Chem.

I would suggest you teach several case histories of succesful drugs like Gleevec, Fluconazole, Lyrica, Haloperidol, Fentanyl, Zoloft, Lipitor, Celebrex. In many cases these histories were summarized. I would emphasize the irrational and the arbitrary part of the story (dodging someones patents, doing the simplest chemistry first, etc.). Then I would discuss then the mechanism of their side-effects, when known. You don't wish to give your students exaggerated confidence in computer-assisted rational design. Most of the time the actual project is just methodic derivatomania drudgery with few inspired guesses (often based on incorrect rationale) that turned out to work.

Permalink to Comment

37. CMC Guy on February 25, 2009 3:48 PM writes...

milkshake although I agree experience factor goes a long it takes a while to develop (and typically by working closely with those who can teach you the facts of life) so getting early exposure to some basic concepts and theoretical approaches can be of value IMO as long as tempered with practical examples as suggested. I have spent more time in development and can tell when medchemists ignore logP considerations it can create headaches down the pipe including eventual compound failures. The ability to communicate with biologists (and others) is probably the most critical thing which is another thing not sure translates well to academic courses.

#33 you seem to be advocating dominate myth that "Synthetic Jocks" are best source of medchemists which is largely based on premise that they can churn out new molecules (so start out as "hands"). Didn't most start out in "Introductory Organic"? It a part of education and people with formal Medchem training can provide useful perspectives and insights that have to learned by Syntheticers. Only thing that seems clear is that Industry want Cheapest so means continues push to outsourcing discovery/medchem jobs regardless of training.

Permalink to Comment

38. BACE on February 25, 2009 5:49 PM writes...

The only people who have misplaced confidence in CADD are those who don't understand it well. All these tools are supposed to complement each other, not supplement or take over. DD will always remain a combination of rational and irrational design and luck. In DD you don't want to inspire your students' confidence in anything too much.

Permalink to Comment

39. milkshake on February 25, 2009 6:21 PM writes...

Unfortunately my experience is that ambitious graduates with protein modeling background tend to drink their own Kool-Aid too much. That's even more true for companies that develop the software, and company managers are quite receptive to the wishful thinking propaganda.

Permalink to Comment

40. Wavefunction on February 26, 2009 5:49 PM writes...

It depends. Personally I don't trust QSAR. Docking is pretty bad for scoring but can be good for pose prediction on validated targets. Sometimes it can prevent chemists from going down blind alleys in the absence of a crystal structure. Long, millisecond MD can shed light on loop movements. In the end, no academic subject can substitute for chemical intuition gained through experience. Unfortunately this is a rare luxury. What counts finally is knowledge of general chemistry. A lot of times it's the basics that come to your rescue.

Permalink to Comment

POST A COMMENT




Remember Me?



EMAIL THIS ENTRY TO A FRIEND

Email this entry to:

Your email address:

Message (optional):




RELATED ENTRIES
XKCD on Protein Folding
The 2014 Chemistry Nobel: Beating the Diffraction Limit
German Pharma, Or What's Left of It
Sunesis Fails with Vosaroxin
A New Way to Estimate a Compound's Chances?
Meinwald Honored
Molecular Biology Turns Into Chemistry
Speaking at Northeastern