I set up two model reactions yesterday. Both had 2 equivalents of Boc-Ala-OSu and one equivalent of Boc-Orn-OH. One had 10% DMAP. The Boc-Ornithine was not entirely soluble in THF at 0.15 M, probably owing to the presence of the carboxylic acid. I followed the reaction using TLC (85:15 EtOAc/methanol) with detection using either phosphomolydate or ninhydrin. I also followed by taking 1 µL of reaction into 999 µL of pH 7 buffer and scanning and recording the absorbance at 260 nm. I started the reactions late in the afternoon yesterday and ran overnight.
The ornithine derivative stayed at the baseline and was detectable with ninhydrin and PMA. The alanine active ester was also ninhydrin-positive, which I tentatively attribute to loss of the BOC group upon heating, which would produce a free amine capable of reacting with ninhydrin. By ninhydrin each reaction produced a spot that was slower than Boc-Ala-OSu, but ahead of each spot there was a weaker spot with a slightly lower Rf than the starting material. By PMA similar spots showed up for the standards and for the products. There was an additional, relatively slow spot that showed up in each product lane that might be N-hydroxysuccinimide (I don't have a standard). The only surprising result is that I expected to see a more obvious spot for Boc-Ala-OSu in the two product lanes, given that it was in excess. I did see an extra spot in each product lane, and perhaps there is enough error in the Rf values to allow the conclusion that this is leftover starting material.
The initial time points in the A260 experiment both showed strong absorbance, which was unexpected (there is probably a contaminant, possibly NHS itself). Absorbance increased at nearly the same rate, based on time points taken roughly 5 hours into the reaction and overnight. I plan to take one more time point each, to see whether or not the absorbance increases further. Although I have not seen difference between (+)DMAP and (-)DMAP, this is just one experiment. I chose pH 7 so that most of the N-hydroxysuccinimide would be in its conjugate base form.
I like the idea of increasing the number of equivalents of Boc-Ala-OSu from 1.0 to 1.5-2.0, provided that the remainder can be separated chromatographically. I think that the use of absorbance over time to look for completeness has promise in some circumstances (assuming that one has access to a spectrophotometer), but it needs a little experimental tweaking. If I had it to over again, I would probably have used benzylamine as my amine, but I would have had to adjust the absorbance experiment for the background absorption due to benzylamine.
Assuming that all of the initial (time zero) absorbance that I saw was from contaminating N-hydroxysuccinimide (NHS), its concentration was about 62 mM. I have not worked out the mole fraction yet, but I would guess it to be about 20-25%, relative to Boc-Ala-OSu.