5-N-linked pyrazine side chains

[david1597]

As part of a larger investigation into the 5 vs 8 substitution of nucleophiles onto our core I am attempting to make compounds with N-linked pyrazine side chains at the 5-position. This issue is to summarise these findings.

[david1597]

The first experiment is in and it looks like an 8-substituted product - DGS 84-1. Reacting benzylamine with the core gave the main product in around 60% yield. Two further products were isolated in around 5% yield each: mass spec is a good match with the structures indicated although the 1H and 19F NMR spectra were not good enough to confirm. By mass spec and tlc, an amount of the 8-hydroxyl product was also obtained.

[MFernflower]

Seems like NaH causes things to just react into unmanageable goop
@david1597
I wonder if a really weak basic system like LiOH in crown-ether/toluene would allow you to get more 5 product pertinent to @mcoster 's proposed base driven mechanism

[david1597]

60% for the main product is better (around double) than the previous reported (using KOH/crown ether method) yield for an amine plus Cl-core reaction. And I've accounted for the majority of the material (I didn't get a mass for the isolated -OH product which was my error and would have been super useful, but that would have taken us close to 100% recovery with four compounds). This definitely wasn't goop - see the attempts at directly reacting the triol to make the OHOH for goop.

Not sure LiOH/crown would work. We use 18-crown-6 as it is a good match for potassium. I didn't isolate any 5-product in this reaction so it's not a case of fine-tuning to improve selectivity. I think we need a different approach. Trying this next by using the uncyclised core intermediate for the substitution step.

[mcoster]

@MFernflower - as @david1597 points out, LiOH/crown ether seems unlikely to work. Hydroxide is not strong enough to deprotonate an amine, so those conditions would rely on nucleophilic attack by the neutral amine, and hydroxide would only come in at a later stage. I don't know if the neutral amine is reactive towards the chloro compound without strong base?

If you have time and inclination David, the experiment I would be really interested in is deprotonation of the amine with n-BuLi, and reaction of this lithium amide anion with the chlorotriazolopyrazine.

It's interesting that attack by azide seems to occur exclusively at the 5-position.

[MFernflower]

@mcoster Once again thank you mark! I am learning more and more topics I organic chemistry when I end up talking to you!

The lithium salt route seems like a logical step but if that too fails and gives 8 position could always try Buchwald–Hartwig for coupling to already made tp-core

I still working on the hydrazone (david's idea) will be the most cost effective and most time effective idea

[mcoster]

Oh, I just noticed that David pointed out that he did get product from amine + KOH/18-crown-6, albeit in a lower yield. This suggests that the neutral amine is reactive towards the chloro compound to some extent. I would be interested to know whether stoichiometric hydrogen is evolved from benzylamine + NaH. The pKa's of H₂ and amines are fairly similar.

[maratsydney]

@mcoster your message reminded me that I actually tried to couple this substrate with amine by just mixing and heating using 3 eq of amine. And this conditions delivered 60% yield on 100 mg scale as it did in an example given in #39 for which xray structure was shown. Tlc indicate presence of some other products in the mixture but fractions after main compound appear as a complicated mixture from NMR so I didn't try to investigate it. ELN entry

[MFernflower]

@maratsydney Have you observed the same percentage of 8 product formation regardless of temperature? Just wondering if there may be hope to skew the reaction towards 5 product by playing with temperature?

[maratsydney]

@MFernflower at lower temperature reaction is much slower but still deliver same product based on NMR of reaction mixture and nothing else.

[david1597]

The attempt at performing the substitution on the uncyclised core intermediate gave no reaction. Starting material was recovered. Eventual conditions were NaH with heating at 110 C for 2 days.

[MFernflower]

Perhaps it is time to break out the palladium and JohnPhos? @david1597

[MedChemProf]

We did not try and optimize the reaction, but check out the Pd coupling that one of my students did on a related system. Earlier experiments shown in her notebook indicate that direct nucleophilic displacement of the second chloride without resorting to the Palladium Chemistry did not work.

https://mynotebook.labarchives.com/share/MCPHS%2520MedChem/NjYxLjd8MzY3Ny81MDkvVHJlZU5vZGUvMjI5ODc3NTI4M3wxNjc5LjY5OTk5OTk5OTk5OTg=

I am on my phone at the moment and could not paste in a picture of the reaction sequence shown in her notebook.

[MFernflower]

@MedChemProf @david1597 Good prior results with the slightly cheaper JohnPhos + Pd2(dba)3 (see marat's Buchwald–Hartwig couplings)

[mattodd]

Here's the scheme @MedChemProf was talking about. Nice lead.

[david1597]

I've got our first 5-N-linked compound.

I followed the above prep - thanks @MedChemProf and students! Yield is still low (around 10%) although that's unoptimised and partly due to a tricky purification. Still, I have around 40 mg which is sufficient for our purposes. I'm currently making a few more of these compounds to send together in the next batch.

NMR from an hour ago below. You can see the singlets for the pyrazine protons (and not doublets as we had for the 8-linked compounds). I'm currently updating the ELN with all the details.

[MFernflower]

Fantastic work! What other ethanamines are you planning on using? @david1597

[david1597]

• the parent phenyl
• the shorter one carbon and zero carbon linked phenyls. These will correlate well with the ether and thio-ether compounds, and wrap up investigation of chain length.
• the two enantiomers of the benzylic alcohols. Investigating as it looks important in the ether cases, and the enantiopure starting materials are readily available in this amine case.

That'll be it for now. It may be that they are all non-potent. But if not, should give some info to work on for the future.

[MFernflower]

I know this sounds crazy but do you have access to amphetamine? I know it's sometimes used as an ochem reagent despite the fact it's a drug of abuse. I would be curious to see what that methyl so close to the amine linkage would do potency wise @david1597

[david1597]

We don't have it. And the likely paperwork that would be involved isn't worth it.

We'll wait on results back from the above set before deciding any further compounds. No point wasting time if it turns out they're all inactive, but the above selection should be enough to make a decision.