Cryptic Retraction, Uncovered

Earlier today, a curious Twitter tipster wondered aloud about an "obtuse retraction notice" in JACS:

"The structure of compound 1, the major compound, of the manuscript was mistakenly assigned. As a result the authors withdraw this manuscript."

You heard it right, folks: An entire (published) manuscript, all down to one set of spectra.

So, what went wrong here? Here's the carbon-13 spectrum, from the SI:

Source: Jang group | JACS 2008

Whoa! That's a lot of carbons for that relatively simple product. I count 39 signals, aside from solvent, despite the compound's formula - and the authors' peak lists - only accounting for 26.

Another tweet (thanks, Neil!) clued me in to this Organometallics paper, in which they prepare the same compound. Compare the spectrum above to this one:

Source: Hor group | Organometallics 2011

I count 26 major signals, about as many as should be there, given the slight magnetic inequivalency of the benzyl carbons.

So, what went wrong? One clue might be solvent; the first spectrum's taken in a highly polar solvent (d6-acetone), whereas #2 uses ol' NMR stand-by deuterated chloroform. Given the highly polar nature of the first compound, along with the extra signals (and perhaps a second benzyl group in the proton NMR), I'm guessing that spectrum #1 actually shows a quaternary ammonium salt, which might result from "over-benzylation" of the cinchonine starting material.

The real bummer here? I've looked through the rest of the SI, and most compounds appear spot on.

Certainly, the authors managed to perform a challenging radical addition with high selectivity. Even more curiously, the ammonium salt used to effect the transformation (1a) looks correct!

Tough pill to swallow. Kudos to the authors for making the right (tough) choice here, voluntary or not.

Update, 8/8/13: Over at Reddit, stop_chemistry_time has staged a fantastic, ongoing debate with me in the comments. Here's the link.