• N&PD Moderators: Skorpio | someguyontheinternet

Compounds With the Most Frustrating Syntheses

3DQSAR

Bluelighter
Joined
Oct 27, 2024
Messages
395
I'm guilty of always asking people to provide a facile synthesis for the novel compounds sometimes posted. It's not to be contrary but because if you can't produce it, you cannot confirm activity and safety profiles. But at least as importantly, it's key to becoming a competent medicinal chemist. Yes, ultimately the head of a team may only provide in-silico models (or properly calculated Drieding models) BUT only after they have spent decades at the 'coalface' finding practical, scalable production methodologies.

The compound that has haunted me for the last ohhh, thirty years is the 'reversed ester' of nortilidine. I used inverted commas because in fact the cyclohexene ring of (nor)tilidine is substituted for a cyclopentane. Neat fact - the cyclohexane homologue if nortilidine is far, far less potent than the parent. I strongly suspect one key reason for the development of tilidine was it's simple synthesis. Yes, one of the precursors is very costly on the open market BUT I presume that the makers have telescoped the synthesis. Coupled to the fact that tilidine is a rare example of the trans-pair having advantages of the pure enantiomer, thus clawing back some of the costs. In addition, the unwanted cis-pair can be epimerized so nothing is wasted (yield goes up). Dextilidine (or rather it's metabolite - nortilidine) is an opiod and DRI while laevotilidine has NMDA activity.

US4291059A Cycloaliphatic compounds, analgesic compositions thereof and method of use thereof as analgesics

It's worth a quick look at the parent and metabolites of both.


As you can see, of the four possible enantiomers, it's the trans-pair that is used medically but as mentioned, the cis-pair can easily be isolated and racemized so it doesn't get wasted. Obviously the optimized route isn't in the public domain.


Forgive the cyclopentane rings not being quite to scale but the KEY elements are that the benzene ring perfectly overlays the A-ring of morphine, the ester function overlays the E (dihydrofuran) ring of morphine and the basic nitrogen or, more specifically, the N: lone-pair overlays that of morphine. The cyclohexene or cyclopentane rings do not appear to overlay any of the rings seen in morphine. No, it merely acts as a scaffold to place the three key moieties in the appropriate relative spatial positions.

So at first glance, the reversed-esters don't appear any more complex than tilidine itself. But like quite a few medicines, researchers homed in on a Diels-Alder reaction possibly because it's regarded as the 'Mona Lisa' of chemical reactions but more pragmatically, because it's such a powerful tool.

But as you see in the patent, that reversed ester is a pain. Yes, there are references to the selective O-acetylation of other compounds BUT in this case it's esterification of a tertiary hydroxyl, propionylation rather than acetylation is required and in such systems it appears that a large excess of the carboxylic acid derivative are required.

As I've previously mentioned, while acetic anhydride, ethane-1,1-diyl diacetate and propanoic anhydride are all controlled chemicals, the mixed anhydrides and mixed diyls are yet to be controlled. Still imperfect BUT feasible? Well, things only get worse from that shaky start. I figured that the patents choice of (cyclopent-1-en-1-yl)benzene is still likely the best option but it's a rare and costly beast. That's going to be 4 steps to the raecemic produce. About the only upside is that most routes will produce almost purely the trans isomers.

I've been down some frankly hideous pathways in my mind. One was the direct aminohydroxylation using N-(2,4-dinitrophenoxy)methanamine... but try finding THAT material.

It seems like every time I go down a new path, some obstacle gets in the way.

It's not as if isonortilidine is a particularly potent compound. It's my opinion that the researchers at Glaxo were hoping that like the phenylpiperidine class, reversing the ester function might increase activity in a similar manner. For a compound x10 M, it might have been a practical proposition but tests suggest that it's no more potent that the parent.

Now as far as I know, nobody has ever studied derivatives in which the cyclohexene or cyclopentane rings were further substituted. I suggest the reason for that is the sheer COST of doing so would make it impractical.

But every year or so I go back to the papers, try out some more ideas and give up again. It's by no means with the intent of producing the compound, it's merely to see if the application of more modern reactions make it a facile target.

BTW why not try overlaying isonortilidine with cypenamine.
 
Last edited:
Can you get

1-hydroxy1-phenyl-2-methylaminocyclohex-3-ene.png


1-hydroxy1-phenyl-2-methylaminocyclohex-3-ene?
 
Well, that's an intermediate mentioned in the patent - but how do you synthesize 1-hydroxy1-phenyl-2-methylaminocyclohex-3-ene? Or, more specifically, how do you make the trans-pair because isolation of the cis-pair is possible but it's not clear if it can be epimerized.
 
If It's Not Listed In The Experimental Procedure, It Should Be Referenced To Another Scientific Paper's Experimental Procedure.

Have You Tried Emailing The Authors?

I Would Imagine Its Synthesis Involves At Least One Grignard Reagent And Probably Starts With Benzoquinone.

The Easiest Outcome Would Be To Simply Buy The Intermediate.
 
Last edited:
Yes, I contacted Derek P Reynolds years ago. But the syntheses in the patent ate the syntheses used. But how would you add the alkene unless the ring is formed using something like a Diels-Alder?

It's entirely possible that one reason the compounds never reached market was BECAUSE the synthesis made it uneconomica.
 
My point entirely.

That's why I keep looking at recent literature to see if something new would make it a facile target.

One that on one hand is awesome, on another downright scary is N-(2,4-dinitrophenoxy)methanamine which is able to perform the trans-aminohydroxylation of an alkene which would have potentially provided for a 2 step synthesis...

But the MSDS of N-(2,4-dinitrophenoxy)methanamine makes it clear that it's not something that's practical to scale if you value your own life.
 
1-phenyl-1-carboethoxy-2-methylaminocyclohex-3-ene.png


^-- Is Still 4 Steps Removed From The Starting Alcoholic Alkaloid, So It Just Really Boils Down To The Price Of The Precursor.

Even Then, Cyanide Is Necessary.

I Bet It Would Feel Awesome, Though.

I Would Imagine Nortilidine Would Be An Awesome Drug To Try. It Is A Pharmie In Some Countries, Still Being Used.

I Think It Could Lose The Double Bond In The Cyclohex-1-ene. Though, I Would Probably Prefer BOUNCE (andrewamine).

1-carbomethoxy-2-methylamino-1-phenyl-propane.png


1-carbomethoxy-2-methylamino-1-phenyl-propane
 
Last edited:

BTW for anyone wondering, N-(2,4-dinitrophenoxy)methanamine.

Now, even to the untrained eye, the oxygen-balance of the compound is going to make it rather hazardous to work with. Fine if one is using a microscale reaction, not something to be used at scale.

The logical step would be to find a catalyst that would perform the O-amination thus allowing only a small quantity of the material to do the job. But as it is, the research on this class of compound isn't well established.

But a more simple example would be it's ability to synthesize (ring substituted) PPA derivatives. Since it produces only the trans-pair, it would allow a simplified route to ring-substituted 4-MAR derivatives.
 
1-phenyl-1-carboethoxy-2-methylaminocyclohex-3-ene.png


^-- Is Still 4 Steps Removed From The Starting Alcoholic Alkaloid, So It Just Really Boils Down To The Price Of The Precursor.

Even Then, Cyanide Is Necessary.

I Bet It Would Feel Awesome, Though.

I Wgould Imagine Nortilidine Would Be An Awesome Drug To Try. It Is A Pharmie In Some Countries, Still Being Used.

I Think It Could Lose The Double Bond In The Cyclohex-1-ene. Though, I Would Probably Prefer BOUNCE (andrewamine).

1-carbomethoxy-2-methylamino-1-phenyl-propane.png


1-carbomethoxy-2-methylamino-1-phenyl-propane

Well, I have been unable to find a precursor better than the one used in the patent i.e. (cyclopent-1-en-1-yl)benzene (AKA phenyl cyclopropene) CAS 825-54-7. I strongly suspect that the high price is simply a function of it finding no commercial application thus it's likely to be made only when someone orders it.

After all, it's produced using a Grignard between phenylmagnesium bromide and cyclopentanone. BOTH of those are common enough.

Now IF one follows the patent, research HAS found a much better way to synthesize the epoxide intermediate. Potassium peroxymonosulfate (Oxone) is relatively benign, dirt cheap and there are even chiral catalysts!


Hard to know what the ee would be in the specific case BUT if it potentially doubled the potency of the final product, it's something that I for one would consider worthy of extra time and effort involved in finding and optimizing a enantioselective methodology.

Other routes exist, but all of them use unusual nay custom chemicals.
 
I used to think JNJ-7925476 had a synthesis that is somewhat difficult.

However I discovered that 2-Phenylpyrrolidine is commerically available then all one needs to purchase is the styrene oxide.

JNJ-7925476 itself is more complicated than this but one can consider analogs such as Mcn-5292 [105234-89-7]. Certainly that's the easier option in the short-run then there is the para-methoxy analog or para-bromo analog, etc.

Other (named) analogs include: McN-5558, McN5652, McN 4612-z, Mcn 5707, McN-5847 & McN-5908.

McN-5908 is the para-amino and may be the most potent psychostimulants ever discovered.
 
Last edited:
Styrene oxide? OK, I will bite - how do you build the ring. The patent used phenylcyclopentene which is then converted to it's epoxide, but the point is - the cyclopentane is in place.

Those medicines ARE complex to make but are legal most everywhere and IF affinity is any guide, might be facile. I mean, if the dose is in single0figure mgs, it's possibly worth the expense.
 
Sorry I went off topic and started discussing a completely different compound.

You have the patent for the compound you are interested in. Rather than attempting to do something no man has ever done before, you could just follow the recipe in the patent. Of course it may still be a bitch to make, and might require a deeper level of insight than simply following a recipe.
 
Nah i wanna discuss Mcn-5292 [105234-89-7] (McShane). 3DQSAR everything you seem to be discussing is an opioid for some reason. Actually you did discuss a few different ideas such as GHB analogs but i've not had a discussion with you on stimulants such as SNDRI for example.
 
Well, I would begin by comparing the patent routes. These are quite recent compounds so what new applicable syntheses have come to light since the patents.
 
It tells you how to make it here:

Maryanoff BE, McComsey DF, Gardocki JF, Shank RP, Costanzo MJ, Nortey SO, et al. (August 1987). "Pyrroloisoquinoline antidepressants. 2. In-depth exploration of structure-activity relationships". Journal of Medicinal Chemistry. 30 (8): 1433–54. doi:10.1021/jm00391a028. PMID 3039136.

Mcn-5292 [105234-89-7] is the analog with para-chloro is the best place to start with.

4-Chlorostyrene oxide [2788-86-5] is expensive from UK suppliers but my guess if you went straight to China you would be able to buy it cheaper.

According to the journal it is made via the mandelic acid proecdure though that it an extra reduction step.
 

Well funnily enough, I already pointed out that oxidation of styrene HAS seen a useful improvement and even allows for CHIRAL oxirane rings to be formed. Oxone is dirt cheap and you only need a catalytic amount of the chiral ligand. Oxone is dirt cheap.
 
BTW one of the main reasons I mentioned those compounds was because to the best of my knowledge, para-ethynyl amphetamine/cathinone is legal in most places but strikes me as the best of the as yet untested novel amphetamine/cathinone analogues.

Because THAT is a facile target.
 
Xenon. Literally impossible to synthesize as it's just an element. (Would it be possible with atomic work?) As far as shit that I would be comfortable attempting (wouldnt really be comfortable with super complex shit or RCs), THCV seems difficult imo.
 
Top