I'm thinking of using a aerobraking and engine thrust as necessary to reenter LEO. Aerobraking alone could be pretty tricky since the vehicle would be coming back at roughly 25,000 MPH. I remember from the Apollo era hearing that their approach window was only about a degree wide. If they came in too steep, they'd burn up in the atmosphere. If they came in too shallow, they'd skip off the atmosphere and be lost in space. I suspect a direct aerobraking approach would have an even tighter window, so perhaps combining some aerobraking with thruster firings as necessary might be a better tactic. It might be possible to use that inflatable heat shield technology the Russians tested last year but that would only work if the shield is very stable. If the shield flexed during the aerobraking phase, it could cause the vehicle to go off course.

My original idea germinated when the story about possible lunar ice first appeared. If the ice is in a form that can be economically harvested, I'd like to launch it to LEO and/or lunar orbit for processing into propellant.

There are some issues with using cyrogenic propellants like H2/LOX to power a long space mission. Probably the biggest problem is storability in space. Cyrogenic liquids tend to boil off unless it's kept in a Dewar (Thermos type container) under pressure, making for some pretty heavy propellant tanks. That's one of the reasons why the Apollo Service Module and Lunar Module used storable hypergolic propellants. Gemini, Apollo, and the Shuttle use H2/LOX to power fuel cells for missions lasting up to two weeks or so but the quantity of cyrogenics is small compared to the amount of propellants you'd need to fly the lunar missions, so the tank weight wasn't as severe an issue.

Another reason in favor of hypergolics is their reliability. All you have to do is open the propellant valves and the propellants will ignite on contact in the combustion chamber. That's why the Shuttle still uses hypergolic propellants for its OMS and attitude control thrusters. Hypergolic propellants have a lower specific impulse than H2/LOX but the engineers have decided that it makes for a more viable solution.

Resolving these issues might take some time so it's possible the first version of the vehicles might have to use hypergolics. This would require launching the propellant from the Earth's surface (expensive and somewhat dangerous). However, it'd take time to set up orbital fuel processing facilities and lunar ice harvesting facilities so this might not be as big an issue as it first appears. I wonder how well the higher specific impulse from cyrogenic propellants would offset the substancially heavier tank weight. If the ice can be harvested economically then the issue may be moot.

I don't know how much Shuttle derived hardware would be useful for the translunar vehicle. You might be able to adapt the crew compartment, the avionics (reprogrammed, of course), auxillary propulsion, electrical, etc. for the new design. Trying to convert an actual orbiter simply wouldn't work because there's too much dead weight in the structure, wings, landing gear, etc. In the end, I think what we'd need to do is to look at the design with the goal of maximum system reuse but not reuse of any of the structural components.