[Babcock_Hall]

The intended product is ornithine-δ-isonicotinamide, alkylated on the pyridine nitrogen with a 5-carbon chain terminating in a carboxylic acid.  We have been trying to remove a 4-methoxybenzylester used to protect the carboxylic acid on this chain (the carboxylic acid of the ornithine moiety is unprotected).  In the deprotection reaction excess phenol was the scavenger (TFA was present also), and we are not certain whether or not the group was removed.  The next step was to extract away the phenol, and my student mistakenly extracted using 1 M KOH versus ethyl acetate.  Our presumed product is in the aqueous layer, and it is the last product of a six-step synthesis.  I have thought of several things we might do.

1.  Add a large excess of Dowex-50 (not that we have a great deal in lab right now).  Once we added sufficient excess, we would bind all of the potassium ion to the Dowex resin, and we would then elute our product with ammonia.
2.  Try to purify using some sort of reverse-phase.  The KOH should go right through (as long as it stays soluble in water -organic mixtures), but our compound of interest, having some hydrophobic regions, should bind.
3.  Go back one or two steps and try again.

[Arkcon]

A very small dialysis rig?

[Babcock_Hall]

Can dialysis be performed on such a small molecule successfully?  I am very familiar with dialysis of proteins, but the difference in molecular size is much smaller here.  BTW it occurred to me that gel filtration might also work, but the same caveat applies.

[ziqquratu]

Reverse phase should work nicely. A friend was doing some work where he needed to deprotect an amino acid ester, somewhat similar in size to yours, with NaOH, and it cleaned up nicely by RP-HPLC.

[Babcock_Hall]

We only have a small HPLC column and the present volume of the sample is 20 mL.  What about using a Sep Pak cartridge?

[ziqquratu]

Unfortunately I'm not familiar with them, so I couldn't offer any opinion - although someone else may be able to offer an opinion!

[Archer]

We only have a small HPLC column and the present volume of the sample is 20 mL.  What about using a Sep Pak cartridge?

That should work fine but getting your loading & washing volumes correct is going to be key to success with a cartridge.

Best way to check, in my experience, is to use a small C18 SPE cartridge and determine the optimum loading and washing on a few μl at a time.

[Doc Oc]

That pyridine is not going to be neutral, it will be pyridinium, and those can be very unstable.  You essentially create a leaving group out of it, the only way I was able to isolate one for any period of time was to crystallize the pyridinium salt from some organic solvent (DCM in my case).

[Babcock_Hall]

Yes, I should have shown the charge on the pyridinium nitrogen.  It is one of the reasons that our compound has some water-solubility.  What we ended up doing was to find a very old bottle of Amberlite IR-120 (which we rinsed with water first) and added a large portion of the contents to our solution.  Our hope is to bind the amino acid and the potassium ions to the resin, then to release the amino acid with ammonia.  We will see how it works.  I felt that it was important to go ahead and to soak up the base, which I was afraid would degrade our compound.

[Doc Oc]

The point about the pyridinium is not a matter of correcting your drawing.  It's to point out that pyridinium is highly unstable.  I couldn't even filter it, I had to use the crystalline material directly in my next reaction.  I very strongly doubt it survived your aqueous condition, it's almost certainly fallen apart.

[Babcock_Hall]

I am not following you.  Unstable to what conditions?  I would think that the hydroxide ion would be the most serious concern, and that is why I tried to neutralize the hydroxide with Amberlite in the the sulfonic acid form.

We have made compounds of this form (both nicotinamide- and isonicotinamide-based) before and purified them by ion exchange columns, although sometimes with difficulty.  We characterized them by NMR.  NAD has a positively charged nitrogen atom in its nicotinamide ring, but IIRC it is labile to strong base.  We patterned our synthesis after similar work in the literature.  For example:
Harjani et al., Green Chemistry 11 (2009) p. 83
Sambrook et al., J. AM. CHEM. SOC. 2005, 127, 2292-2302 (in the supporting information)

[Doc Oc]

Think of pyridinium being like a halide.  What would happen to an alkyl halide in the presence of a strong base?  I'm guessing there's no product left in your reaction mixture.

I think the more important thing here is to have a talk with your student.  Yes, going back and repeating 6 steps is annoying.  But switching 1M KOH for EtOAc in the extraction is a pretty big error.  Not uncommon, but obviously significant in this case.  I would have a talk with them to make sure they're comfortable asking questions if they're unsure, and then spend a little more time supervising them until you feel good about letting them work independently.

[Babcock_Hall]

Clearly there is a potential for hydroxide to displace the pyridine-like ring, but the question is how fast it happens, and I don't have much data on that.  I am having trouble remembering, but I think one day elapsed between the addition of base and the addition of the cation exchange resin.  The student in question has done the same procedure correctly before (ethyl acetate vs. water in the extraction, as opposed to ethyl acetate vs. hydroxide), and he has worked on this project for more than one year.  However, I may have accidentally confused him by mentioning that when the product of interest is in the organic layer, phenol can be removed from it by extraction versus base.  Live and learn.

[408]

I must be missing something here based on the suggestions above as my solution is relatively simplistic. 
Neutralize the base with 1M HCl in an ice bath.  Evaporate down giving a mix of KCl and product.  Extract product from the KCl using whatever organic solvent this stuff dissolves in that KCl does not. What am I missing?

[Babcock_Hall]

The final product is charged on the pyridinium ring.  At pH greater than about 3, one or both carboxylic acid groups may also be charged.  At pH values less than about 9-10 the ammonium group will be charged.  Therefore, many organic solvents will be not polar enough.  On the other hand, perhaps we could get the product into methanol, and leave some or all of the KCl behind.  I have not looked up the solubility of KCl in methanol yet.  I'll do that today.

As Doc Oc pointed out, we may have degraded our compound already.  I need to look up the stability of a model pyridinium compound today also...