you can biotinylate the proteins with a derivative of biotin that doesn't pass through the cell membrane. Then you can purify the biotinylated proteins with streptavidin beads and then run a Western and detect a protein that would be on the side of the cell that is facing outward.
I think you might find all information you've required in the following elegant paper of Nai-Jung Chianga, Sheng-Nan Wuc, and Li-Tzong Che, just recently published in the Journal of Pharmacological Sciences (Volume 127, Issue 4, Pages 404–413, 2015). The paper's title is: "The potent activation of Ca2+-activated K+ current by NVP-AUY922 in the human pancreatic duct cell line (PANC-1) possibly independent of heat shock protein 90 inhibition".
They shouldn't be inside out! The bud directly from the surface of membrane. You could try to look for the orientation of the transmembrane proteins. Using a monoclonal antibody with a high fidelity target should do. This is the easiest method.