There are a few advantages to using a manned vehicle, such as not needing to put attitude control, propulsion, and TT&C systems on the modules. However, that may not be enough to make up for the greatly increased cost of a Shuttle launch. A Proton launch costs somewhere around $60 million, a Shuttle launch is at least 10 times as much. NASA's fuzzy accounting makes it hard to come up with a per mission cost for the Shuttle.

A heavy lift booster derived from Shuttle (or Energiya) technology and capable of putting well over 100,000 pounds into LEO would greatly reduce the number of assembly missions required. You could launch the propellants on an EELV or two, depending on the amount required. I agree that the most likely propellants, at least in the near term, will be hypergolic. If lunar ice exists and can be harvested, it might be possible to use it to create LH/LOX propellants. Perhaps the transit vehicle could be modular with replacable propulsion systems to handle the change in propellants. You'd still need some hypergolics the the attitude control thrusters, though.