Which compounds should we be synthesising next?

[alintheopen]

The team need to decide which compounds to synthesise next and it would be good to have 10-20 to focus on.

ESAC made some recommendations:

• Aim for log P between 1.5 and 3.5
• Use more heteroaryls, or saturated groups
• Revisit other cores with good metabolic profiles, e.g. MMV669846
• Use more sp3 carbons, and de-planarize through chirality at C-5 or in substitutions at C-2 or C-5

and

• Additional chemical biology approaches to MoA: mutants, ko, kd, pull down
• Use predictive tools such as MetaSite for predicting sites of metabolism, vs knowns

[alintheopen]

@JoannaUbels and I had a short brainstorm before our Skype meeting tomorrow @mattodd and @tscmacdonald https://www.dropbox.com/s/y8w1rtvjadvcx1u/Compound%20Suggestions%20July.cdxml

Perhaps we can refine ideas with this as a starting point.

[alintheopen]

Following a Skype meeting with @mattodd, the Sydney chemists are firming up some new targets. Scaffolds are posted below and more detailed compound suggestions are coming from the individual chemists (@tscmacdonald and @JoannaUbels and myself) very shortly..

[mattodd]

Nice summary! It's a good range of "well we ought to make this compound and we can" through to "we ought to investigate this new chemistry to access fundamentally new structures." Looking forward to the specific shortlist, plus cLogPs, and who's going to make what. We ought to be aiming for mailing off a new batch of compounds in 5-6 weeks, so let's plan with that approximate timescale for now. i.e. that might mean finishing some of the easier compounds, and making solid progress on methodology on the others.

[JoannaUbels]

Below is the shortlist of compounds that I am planning to make.

[PatrickThomson]

@JoannaUbels - something maybe worth bearing in mind - compound 5 contains a homobenzylamine moeity, which we had problems buying in the past because it's isostructural with a bunch of controlled amphetamines; however, whilst typing this comment, I realised that it's also the alcohol version of phenylalanine; you can probably buy (both enantiomers of) the amino-protected alcohol straight off the shelf with no problems at great expense, or use LAH on (Boc-Phe-OMe)[http://www.sigmaaldrich.com/catalog/product/aldrich/421707?lang=en&region=GB].

(update) It's early, I'm not quite firing on all cylinders - you probably chose 5 because it's an amino acid derivative ...

[alintheopen]

Here are some compounds I'd like to make, @tscmacdonald will focus on variation of the traditional amides and changing the right hand side of the molecule in his post.

[tscmacdonald]

Firstly, here's a very interesting patent I found yesterday: http://worldwide.espacenet.com/publicationDetails/biblio?CC=WO&NR=2012173521A3&KC=A3&FT=D

In particular, the use of this scheme (p. 47, schemes IV - VII) They don't make any triazolopyrazines, but if this works for triazolopyridines and triazolopyridazines it's probably worth investigating. Also, the fact that they carry an ester all the way through is interesting: this may or may not work for esters adjacent to the nitrogen being cyclised to, but is probably worth trying. Adapting this, if successful, would give us the ability to Pd-couple to the bromine (the patent diversifies with Sonogashiras), hydrolyse the ester, then diversify. Luxury. I'd like to order some triethyl orthoformate, and try this for a triazolopyrazine (maybe an amide, and an ester) asap.

Anyway, on to desireable compounds:

Lowering of CLogP has, again, largely been attempted by throwing nitrogens at everything. I've sketched out some 3-trifluoromethoxy pyridines (25-27), largely based on the potency of MMV670944. These are attractive, but the corresponding aminopyridines are quite expensive. Also worth investigating (and cheaper) are the three unsubstituted pyridine amides 21-24; the lack of these seems an obvious gap. Removing the trifluoromethoxy group also drops CLogP by ~1. Other amides include reduces amino acids (amino alcohols?) 28, 29 to reduce planarity, and perhaps trifluoro-terminated alkyl chains (31). I've also thrown propargyl amide 30 in there for no particular reason other than the lower CLogP: thoughts on that? Worth trying, or a dead end?

When it comes to the top right group, nitrogen heterocycles are again in vogue. Formylpyridines are fairly cheap and are definitely worth exploring in 33-35, while formylpyrazine-derived 36 slightly more difficult (formylpyrazine is expensive; I'd need to reduce some pyrazine carboxylic acid with LAH). The unsubstituted triazolopyrazine 32 (hopefully accessible by the reaction from the paper above) is interesting both as a building block for further funcitonalisation, and in its own right.

[alintheopen]

Hi @PaulWillisMMV, @mattodd and anyone else who's looking at this feed....Would you be able to give us some feedback on the compounds when you have a some spare time? Cheers Alice

[MFernflower]

Is my proposal worth producing? http://www.chemicalize.org/structure/#!mol=FC%28%3DCC1%3DCN%3DCC%3D2N1C%28%3DNN2%29C2%3DCC%28%3DNC%3DC2%29C%23N%29CC2%3DCC%28%3DCC%3DC2%29F&source=calculate

[tscmacdonald]

@MFernflower That looks like a plausible variation on the kind of thing @JoannaUbels will be making once she has a reliably working scheme to make the fluoroalkenes. Something else I sent your way on twitter a few weeks back (but I'm not sure if you still look at your account): rather than suggesting analogues (which there's nothing really wrong with, but the hard part is the actual synthesis, and unless you have a very good reason as to why your ideas are especially good, people will tend to make their own things), perhaps a more productive use of your time would be working on actually generating new information for the project. You were talking about Autodock the other day: playing around with plausible receptors, seeing if you could replicate our observed biological activities, and then suggesting new ligands based on those results would be beautiful. Some things to read over: http://malaria.ourexperiment.org/pharmacophore_modelling_ http://malaria.ourexperiment.org/in_silico_prediction

Without access to a wet chemistry lab (I assume this is correct?), I'd think that computer-based analyses would be the most productive use of your time. I don't really know enough to give more detailed suggestions, but I'm sure you could do some reading.

Of course, none of this is related to this Issue, so if you'd like to talk more about this perhaps make a new issue, start a twitter/g+ conversation, etc

[MFernflower]

What is the most potent tp molecule to date? Any ideas of what these target? I'll gladly do as you say. Provided I have some guidance

[drc007]

@PaulWillisMMV, @mattodd, @alintheopen A few thoughts. Looking through the compounds you have made to date in the Ether series it looks like whilst phenyl ethers and phenethyl ethers have been well explored, I could only find one example of a benzyl ether (MMV672688) which has a naphthyl ring so not really a representative example, that said finding data leaves a little to be desired ;-) and I may have missed some. Whilst a minor modification, the loss of a methylene reduces the logP and in my experience benzyl ethers are metabolically more stable than phenethyl ether. Perhaps a couple of halo-subtituted benzyl ethers might be interesting?

Is there any evidence for hydrophobic collapse in these compounds, is there any evidence for intramolecular aryl-aryl interaction between the two pendant rings? You can sometimes see this in the NMR (http://www.cambridgemedchemconsulting.com/resources/molecular_interactions.html).

[MFernflower]

I found out that replacing the difluoromethoxy group in OSM-S-202 with a thiol (I know im going to get killed for this!) reduces logp by a good amount and the fact that the product doesnt have a DFM group should save time on synthesis! Is this worth looking into, Or have I barked up the wrong and smelly tree?!

[drc007]

@MFernflower In general thiols are unpopular because they tend to dimerise and are prone to oxidation. They also react with biomolecules which can lead to severe unwanted side-effects. I don't think there are any drugs that contain a thiophenol. You could alkylate and then oxidise to the sulphone?

[PaulWillisMMV]

The compounds are PfATP4 inhibitors, you should find details of the most potent on the project wiki

Sent from my iPhone

On 16 Jul 2014, at 09:49, "Mandrake Dolan Fernflower (IGN: TheDoctor15555)" notifications@github.com<mailto:notifications@github.com> wrote:

What is the most potent tp molecule to date? Any ideas of what these target? I'll gladly do as you say. Provided I have some guidance

Reply to this email directly or view it on GitHubhttps://github.com/OpenSourceMalaria/OSM_To_Do_List/issues/232#issuecomment-49130108.

---

This email and any attachments may contain confidential information. If you are not the intended recipient, please delete the email and any attachments and notify the sender immediately.

---

[MFernflower]

@drc007 Brillant! Replacing the difluoromethoxy group with a methylsulfonyl reduces logp to a stunning 1.06 (1.57 with ethanesulfonyl)!

[mattodd]

OK, lots of content here. First up - @tscmacdonald

Tom: very nice catch on the patent. That would provide a means to vary the northeast portion along the lines we were talking about when we Skyped last week. We should certainly build this in to synthetic design - this could make the assembly of analogs with variation in the northeast usefully divergent.

Tom: Your variations seem fine. It would be so great if we could have a more divergent synthetic approach here too, meaning you’d be able to make each of these from the carboxylic acid. So that means I’d pick a sub-set of 4 or so for now to cap the simple number of reactions you’re doing. Your choice, depending on commercial availability of materials. The F-subst pyridines are likely to be better for us if we can do them since they are, I believe, slightly more resistant to aldehyde oxidase - note that those F-groups tend to be next to the aromatic N’s for that reason. For the amino acid analogs and others, I’m not sure. Perhaps @PaulWillisMMV can comment on those? I can’t believe that the propargyl is a good idea, but I can’t refute it, and it’s good to be a little off-the-wall. Make sure you’ve done the face-palm searches on things like this, i.e. both Google and academic databases for, essentially, “drug propargyl toxic” etc. Again, it would be trivial to make all these if we had access to your carboxylic acid. I would say focus on these kinds of analogs more - more sp3 centers, some chirality. For the moment I’d leave 33-36. Only one of us at a time should be addressing that area and you’ll have your hands full with this and carbonylation.

So my feeling would be to aim for 4 of 21-27, then 3-4 sp3/less planar/less aromatic versions of variation in that position. Rest of the time: carbonylation.

Tom - @drc007 asks about aryl-aryl interactions assessed by NMR. That’s perhaps something you can check out? More so if you make an analog with something other than an aromatic ring at one or other point as a control, but check out the link he provides.

[mattodd]

Everyone: can we report logP’s to one decimal place? I know Chemdraw does 5 decimal places by default, but if there’s a way to reduce the bogus accuracy..? The analogs here are all fine, since they are under 3.5, which is now our absolute cut off. (those above 3.5 I know are being suggested as targets of synthetic interest). I’d like to be strict about 3.5 - so 3.6 is no good. As you can see from the ESAC recommendations, 3.5 is on the upper limit of the Window of Acceptability.

[mattodd]

@MFernflower : Tom’s right. It’s great you’re providing new inputs, but you need to always provide reasons for suggestions - why is your suggestion better than Tom’s - because there are always a gazillion possibilities for what can be made. Is it more synthetically accessible, or more likely to be interesting based on what’s known? And yes, read through the wiki to get up to speed on what’s known of this series so far. I don’t like the look of that thiol, but yes, sulfones are possible. Could this be included in an existing synthetic scheme without increasing the synthetic steps - e.g. if introducing that group is known to be synthetically easy, or if there are materials that can be bought that would make the route simpler?

[mattodd]

@drc007 - that’s very interesting, and something that had completely passed me by. @JoannaUbels - take note, benzyl ethers seem to be underpopulated… Anything we can do there straight out of the box?

[tscmacdonald]

@drc007: interesting points about the aryl-aryl interactions; it's something I've thought about, but not in great detail. If you have any resources that go into more detail than that link, I'd be interested to read them. As Mat says, without a control spectrum to compare what I have to, it's hard to say if we're seeing these interactions. Is there a reason beyond curiosity you ask about this? I suppose it could help explain the dramatic sensitivity to substitution on the aryl amide we've seen, but I'd be surprised if the difference in potential energy were enough to be a huge factor here.

[drc007]

@tscmacdonald @mattodd - I was interested in the possibility of aryl-aryl interactions when thinking about the preferred 3D conformation. This would be important when trying to overlay any alternative structural classes. It would also help when thinking about the pharmacophore and how it might bind to the target protein. We did a lot of work on this when looking at NK1 antagonists (albeit with fantastic molecular biology support). Nature 362,350-353 (1993) The Journal of Biological Chemistry, 269, 14957-14961 (1994) Canadian Journal of Physiology and Pharmacology, 1995, 73(7): 860-865 Molecular Pharmacology April 1995 vol. 47 no. 4 660-665

You might see evidence in the nmr by comparing the final compound with uncoupled side-chain?

[drc007]

@tscmacdonald @mattodd - "I can’t believe that the propargyl is a good idea, but I can’t refute it" Propargyl amines are known MAO irreversible inhibitors (see Pargyline), I suspect the mechanism is initial oxidation followed by reaction with the protein. The amide should be more metabolically stable but I'm not aware of any drugs bearing this functionality.

[JoannaUbels]

@mattodd @drc007 Yes I can definitely make some benzyl ethers. We have 3,4-difluorobenzaldehyde on hand which I can reduce to (3,4-difluorophenyl)methanol and couple on. I'm about to search the inventory for any others on hand and can have a look for cheap interesting ones to buy too.

@mattodd I also have a much faster route to the fluoroalkene thanks to the paper you sent me, you can have a look here http://malaria.ourexperiment.org/the_osm_blog/10443/. I placed an order for CFBr3 and Pd(dppb)Cl2 yesterday after consulting Alice and Peter.

[mattodd]

@JoannaUbels OK, so you’re dealing with: 1) The prototypical ether 2) Benzylic ethers - do maybe two of those or so? 3) How easy is your molecule 5, with the extra chain length? Would be a nice one that exploits your ether-making methodology. I’d worry about the benzylic position. Can you block that with anything?

The combination of these means you’re nicely exploring the northwest position quite fully as a theme. Are there any alcohols we could easily try that have pendant polar rings that are not aromatic? We should be trying to, perhaps, de-aromatise the structure a little if that’s do-able, through introduction of more sp3 centers. The morpholino compounds have been crappy though, according to the wiki data. Maybe just stick to bulk in the tethering chain - i.e. substitutions there. Benzylic ethers with something else on the benzylic position, that sort of thing.

4) The fluoroalkene as a major synthetic endeavor. Very nice new route you’ve plotted there! I’d be tempted to try the initial coupling with phenylboronic acid - the pyrazinyl version will behave very differently. Your proposal of coupling the boron-analog of the fluoro fragment with your chlorinated triazolpyrazine must be the way to go. Nicely convergent synthesis if that works. One step at a time - see if you can get that F,Br double bond made.

So for the moment, I’d leave the pyridine in the northeast. Interesting, but lower down the list.

[mattodd]

OK, so @alintheopen you're happy to take the amino acid-derived compounds and the inverse amides in addition to your work on the enantioenriched compounds, correct? Let me know if this basically scopes out the general areas of work for everyone. If this is all OK, it'd be good if we could translate to a shortlist of targets and approximate timescales for completion, so that we can tee up the next round of biology way in advance. Could everyone please think of specific molecules for the shortlist - at least 5-6 each? Commit to those, create issues and tag so that our "being synthesised now" is current?

[alintheopen]

All sounds good. I started making three ether compounds just before the last meeting as I also noticed that there were benzyl ethers with good activity in the MMV data but with high ClogP (they were in the slides for the last online meeting). So, maybe Jo will just have a look at the 3,4 difluorobenzyl ether compound initially. I'm happy with the other targets so yes, shortlist on its way.

[alintheopen]

Just need a little clarification. If we are sticking rigidly to the Log P criteria then we need to define either CLogP or LogP as in some cases the values are wildly different on ChemDraw (e.g. 3.0 vs 4.6 for #236). For this compound I want to make it anyway for purposes of SAR compared to other analogues but it would be good to know which value is preferred. Cheers

[drc007]

@mattodd @alintheopen Do you have experimental logP measurements for any of the compounds? It might be useful for putting the calculated results in perspective.

I should have added, I notice some of the targets have basic substituents, will you be using LogD as a selection criterion?

[MFernflower]

@mattodd I dont plan on keeping the thiol there as planned. Attaching a methyl to it and then treating the resulting phenyl-thioether with H2O2 to generate the sulfone is a better idea, as pointed out by drc007.

[mattodd]

@drc007 yes, logD is of interest - data were obtained on 8 members of Series 4: http://malaria.ourexperiment.org/biological_data/8408 along with pKa and solubility. You'll find the data in the physicochemical evaluation file posted there. Any red flags? (you can still comment on this issue though I'm closing it, or you can start another Issue or a discussion on G+)

For planning, we're using cLogP as a guide at the moment. I guess we ought to be consistent. I don't understand the origin or significance of the large differences we're seeing on Chemdraw between logP and cLogP.

Closing this long and productive Issue. Have installed links on wiki back to this page in 3 places so we do not lose it. Please refer back to this page in any ELN entries for molecules being synthesised as a result of this planning process. Anything still unresolved about logP or aryl-aryl, or any analogs you think need revisiting for discussion, please just start up new issues.