Assuming you have a structure in detergent micelles or something similar, one method is

http://opm.phar.umich.edu/server.php

Lomize, M. A.; Lomize, A. L.; Pogozheva, I. D.; Mosberg, H. I.OPM: Orientations of proteins in membranes database Bioinformatics 2006, 22, 623– 625

Lomize, A. L.; Pogozheva, I. D.; Lomize, M. A.; Mosberg, H. I.Positioning of proteins in membranes: A computational approach Protein Sci. 2006, 15, 1318– 1333

Comments:

1. The actual orientation may change with the thickness of the membrane. The program uses the average thickness of inner and outer bacterial membranes (30 and 23.6 Å) which may be different then what you want depending on the localization of the protein.

2. You will have to likely have to remove the linker loop to the Extra-Cellular Domain to get it to work. It is not a molecular dynamics program, so the loop will be fixed in position which may cause problems.

Hi Tobias, the trans-membrane (TM) protein is composed of a long N terminal part located at the extra-cellular (EC) domain and it is directly (covalently) connected to the first TM helix of the TM domain. We don't know if there is any special secondary structure for the linker. What I've seen so far from the literature is that it seems the linker is a long and flexible loop with no special secondary structure. I have to remind that again these information regarding the linker is just coming from some people who tried to model the protein and there is no structural data available on that. Recently one x-ray structure came out for the TMD of one of the members of the family who my TM protein belongs to. Therefore I can model the TM part of my protein using this template, moreover we already knew about the ECD of the protein for which the x-ray structure is available. But the main problem is the relative orientation of the the ECD with respect to the TMD.

Hi Jeffrey, Thanks for introducing the server but I'm not sure if it solves my problem as what I'm looking for is one step before embedding the TM protein in the membrane. In fact I have problem with determining the structure of the TM protein and that is how to optimize the relative orientation of ECD and TMD.

Try to use modeler with multi-template modeling script. I modeled the structure of PfATPase protein with the help of same script. PfATPase is also having same extra cellular and Trans-Membrane Domain.

A recipe for modeling of helical membrane proteins

Given that the previous suggestions should all be taken into account, in my experience these steps are fundamental:

1) Modeling of the protein with a good modeling software. On the model should be performed a molecular dynamics up to equilibrium in the presence of water at controlled temp., salts and pH to verify its structural stability

2) Modeling of the membrane, with suitable phospholipids for the two layers (suitable script is needed) then also on the membrane should be performed the usual molecular dynamics in water.

3) The model of membrane should be prepared to accept the model of the protein (another suitable script is needed here)

4) The membrane-protein complex should run in MD up to equilibrium in a water box (often between 4,000 and 5,000 water molecules) at controlled temp., pH and salts. This will usually require a very fast and powerful machine (at least 8 - 16 processors and a professional graphic-board with 4000 - 5000 GPUs) because you have to proceed even well above 100 ns..

Of course this is the general logic to follow. The problems that will surely come, need to be addressed from time to time. It takes many many months for a project of this kind.

Good luck.