• N&PD Moderators: Skorpio | thegreenhand

  • Neuroscience & Pharmacology Discussion Welcome Guest
    Posting Rules Bluelight Rules
    Recent Journal Articles Chemistry Mega-Thread
    FREE Chemistry Databases! Self-Education Guide

modeling the 5-HT 2a receptor

N

nirvus

Bluelighter
Joined
Feb 24, 2011
Messages
70
I was learning to use programs that dock small molecules to macromolecules. Then I thought, wouldn't it be nice to have the structure of the 5ht2a receptor! There are so many known (and interesting) ligands: LSD, mescaline, psilocin, serotonin, just about everything in PiHKaL and TihKAL, haloperidol, ketanserin, clozapine, and newer ones: FLY, D-FLY, NBOMe, and TCB2 would all be expected to bind to 5ht2a. Only problem was, the structure for 5ht2a is not publicly available, so I had to make my own. Started with gene sequence for human 5ht2a, ended up with a 3D model that binds known ligands in the appropriate active site, to the amino acid residues that are known to be involved in binding, with binding energies that correlate with published data. In other words, the model is probably very close to the real deal.

So now I can use software to auto-dock molecules to the receptor, and then study the binding interactions, render images, etc. And so I thought I'd share some of the images I've been producing of hallucinogenic 5ht2a agonists interacting with the serotonin 2a receptor. Hope you enjoy!

First up, good old lysergic acid diethylamide (LSD-25), shown as space-filling spheres (the balls), bound to the serotonin 2a receptor, shown as a ribbon model.
lsd.png


Stay tuned, more to come. If there are questions, feel free to ask.
 
Here's the same interaction shown in a different way: LSD is now pink, and only the amino acid side chains involved in binding are shown for 5ht2a receptor. Note those amino acid numbers in the pic, those are the amino acids known to be involved in agonist binding for this receptor.
lsd_flexx_67.png
 
And here's 5-MeO-DMT:
5meo-dmt2.png

that's all i have time to post right now. but i've got more where that came from, any requests? hope the mods are cool with this, i think it's in the right place. thanks to all BL'ers. Have been learning from you guys for years, now maybe i can give a lil' something back.

edit: ok so that one's kinda BIG! I'll try to keep 'em within reason.
 
Last edited:
What did you use as the template? Your poses look a bit off, LSD seems to form a pi-stacking interaction with Phe340, and a H-bond between the indole N and Ser242 (I think). The H-bonding partner of the amide, if there is one, is still up for debate I think. 5-MeO-DMT forms the same pi-stacking interaction, and H-bonds between the indole N and Ser242, and the 5-MeO and Ser239.

Bradens thesis has a lot of mutagenesis data that can be used to infer likely bonding interactions.
 
Requesting ketanserin or other 5-HT2A antagonists. Thanks in advance.
 
thanks for taking the time to do this!

requesting any NBOMe-, Ibogaine, and TCB-2 :)
 
I hate to burst your bubble, but a model of the active 5HT-2a receptor was already published earlier this year in J. Chem. Inf. Model. 2011, 51, 315–325. They used molecular dynamics to get the activated conformation, since the beta-adrenergic receptor structure used for homology modeling is of an antagonist complex. If you download the supplementary information file, there is included a pdb file with the coordinates.
 
Thanks for the support/interest! I used several automated protein 3D structure predicting servers from various institutions to get a lot of basic models and then sifted through those till I found one that had correct overall structure (ie., 7 transmembrane helices, correct disulfide bond, visually similar to other published models, correct binding affinities, etc). Yes, the turkey beta-adrenergic structure was used as template, you're right, this is not modeled on an active configuration, it's a pretty good rough draft. I was working from the Karuppiah Kanagarajadurai paper (www.rsc.org/molecularbiosystems ) and chose an LSD binding mode that matched theirs for that pic, I don't think their model was based on active structure either. But I did get very good agreement with their binding data for the eight ligands they docked.
nirvusbindingsmall.png
 
Once I had the model in hand, I used Autodock Vina (via the PyRx GUI) for docking. It's currently the academic standard for docking, and it's available for educational use as freeware. Then the binding modes were saved as sdf and fed into PyMol for image rendering (say cheese!). PyMol is also available as freeware for educational use. ChemSketch (also freeware) was used to build and 3D optimize ligands. So there, a pretty decent drug binding workstation for free on my Win7/i5 laptop. If I've omitted anything critical or there are other questions, fire away. BIG HINT: check out qsar.org for links to many interesting programs, some free some not. Anyone else out there have good freeware setups?
 
@tryp2fun: thanks for the tip. Pretty sad if i'm too busy to keep up with the literature on this fascinating field! I'm gonna try to find that paper with the prize inside (pdb file). Will be interesting to compare their model, I wasn't able to find a published one, I know Dr. Nichols has one and I was gonna beg but I ended up making my own for what it's worth. And now if any of you wants to try modeling you don't have to butter me up to get mine! (joking of course . . .but not really)

@Astavats: I have done ketanserin and a couple other antagonists and inverse agonists (listed in the bar graph above). Lemme dig out the pic and post it. Also, just yesterday afternoon I got bored and docked the antagonist epivanserin. This is not for commercial use is it? haha

@the night watch: I've done some NBOMe's and TCB2 already, lemme dig up a pic. Ibogaine! How could I have overlooked that? Nice one. I'll do it.
 
@ bleh123: For the level of accuracy of these models, relative binding energies seem to have some validity. At least that's what I gathered is the state of the art for homology modeling/binding studies. So the answer is yes, it would be very interesting to see the relative binding energies of RC's and compare that to subjective effects, although you'd have to use caution for predicting effects from modeled binding energies. So what were you thinking, 2c-x's, trypt's, maybe DOx? I think it'd be most fruitful to look at analog series like that first, then maybe compare across the board. What did you have in mind?
 
nirvus said:
@Astavats: I have done ketanserin and a couple other antagonists and inverse agonists (listed in the bar graph above). Lemme dig out the pic and post it. Also, just yesterday afternoon I got bored and docked the antagonist epivanserin. This is not for commercial use is it? haha
Click to expand...

None, I have no intentions aside personal interest. Pharmacodynamics is most interesting to me, less so the psychological/behavioral which entails. It's probably reversed for most others.

Thanks again!
 
You are modeling both agonists and antagonists binding to the same structure. However, the general model in pharmacology is that they bind to different conformations. If you used the beta2-adrenergic receptor structure for your homology modeling, that is the structure of an antagonist complex and thus is not in an active conformation. The recent paper used molecular dynamics to find the active conformation. A few years back I made a homology model using the rhodopsin structure, similar to Nichols' model, but I wasn't very happy with the results.
 
Astavats said:
Thank you.
Click to expand...
my pleasure.

@tryp2fun: thanks for schoolin' me on that article, for real. super sick. i've got their active state model now, so as much as I love the one I made (for sentimental reasons), I'll def do some docking and modelling with that one. it will be nice to compare mine to (intellectual d***-sizing right?). joking aside, that is some real good stuff on H-bonding and rotamer switching. i've gotten to know the active site residues pretty intimately as of late and i can't wait to dig in to that article.
 
You must log in or register to reply here.
Top