[Babcock_Hall]

Taking one of the aromatic protons as 1.0, the peak at 5.00 ppm has an integral of 0.8.  The material did not bind to the C-25 sephadex column that we tried, and for that reason I am not sure about its form.  Given that the pH was about 8, I am presuming that it is a zwitterion.

[pgk]

Then, what you see in the 1H-NMR spectrum, is obviously the enol form of pyridiniumpyruvate zwitterion. Thus, it’s not a bad idea to repeat the 1H-NMR spectrum up to 15 ppm and see if any other peak appears beyond 10 ppm and if it disappears by exchanging with a deuterated solvent; as well as observing carefully the base line between 5 ppm and the aromatic zone and look for a possibly very small peak (whose integration could permit a rough estimation of the keto/enol equilibrium, though integral of 0.8/1 proton usually means one proton).
Please note that the enol form of pyridiniumpyruvate zwitterion, is obviously more thermodynamically stable as having a more extended conjugation than the keto form.
Furthermore, please also note that neighboring with a carboxylate anion, leads to upfield shift change.

[Babcock_Hall]

We synthesized this molecule in the hope that it would enolize easily, but I did not consider the possibility that the enol form might predominate.  We will continue collecting data.

[wildfyr]

Why not trap the enol with TMS and see what happens to that peak?.

P.S. I've always wondered if TBDMS would work just as well too for that reaction, just in terms of ease of handling. x10000 more hydrolytically stable. I've never had to do that silyl enol ether chemistry and check it.

Edit: Yes, and using pretty ordinary conditions (imidazole) http://reag.paperplane.io/00000479.htm