Here is my latest approach. It is more suited to a truss than a tube, for reasons that will become apparent.
The back of the mirror simply sits on six teflon sliders spaced at 58% (those grey buttons). The side supports are bearings spaced at 90 degrees, and placed on the lateral centre of gravity of the mirror (those white things to the right of the picture). You can calculate where the lateral centre of gravity is using Houdart's calculator:
http://www.cruxis.com/scope/mirroredgecalculator.htm
I also have a restraining post at the upper edge of the mirror with felt inside, which just clears the side of the mirror. This prevents the mirror sliding around the mirror box while in transport.
The obvious limitation of this approach is that you can't invert the scope, else the mirror will fall out, & that would not be good. Since a tube is usually transported horizontally, that rules it out for a sonotube dob. Also, I wouldn't be quite so confident using it for a small mirror (this is a 14"). But it has been working well for me & holds collimation. You can see there's plenty of ventilation around the back. I also have a fan on the rear of the scope (not attached in this photo).
Phil