Good info here, but I'm still a little confused. There's a regulator by the USB but it's 3.3V. There also seems to be a regulator onboard the MK22 user processor, input power on pin 8 (Vregin) and 3.3V out on pin 7 (Vout33). That output seems to be connected directly to the output of the discrete 3.3 regulator- in the schematics, there's a bunch of 0-ohm "DNI" resistors which I'm assuming are "uninstalled jumpers" on board? It's hard to determine what's going where, because the layout of my physical board doesn't match the schematic, and I'm not sure it the DNIs really open circuit and/or the non-DNI resistors are shorting resistors actually installed onboard.
Anyway, I'm guessing the user processor and the cell modem get their power from the "Vbatt" (i.e. unregulated USB/battery supply), and the 3.3V regulator (which is supplied from Vbatt/USB) supplies the "peripheral" circuits (serial interfaces, reset buttons, etc).
If that's the case, feeding with 3.3 might work (well, considering the evidence, it does work lol), but there will be a drop for the aforementioned peripheral circuits- although maybe not enough to cause problems. Until the modem kicks in, but that's another story!
Well, it's back to the bench- here's to hoping I don't release the magic smoke...