Is there a reason you want the whole protein? Why not choose a few regions on the outside, put them in a tagged expression vector, express, purify, and continue with that...

Thanks for your answer. But the antibody representing the whole protein is not my necessity. I want to exclude TM regions and make a construct having rest of the sequence. Since my sequence upon analysis shows many of TM regions,and if i start trimming them, i am left with many pieces of sequence.

You can of course order a synthetic gene with all the transmembrane domains excluded from the sequence and express it as a recombinant protein. You will then have a mixture of extracellular and intracellular domains that will probably not have the original structure and be a rather artificial construct.

The conventional way to raise an antibody would be to make a construct with the C-terminus fused to large affinity tags like Gst or maltose binding protein to make it bigger and more immunogenic. You have then a linear fragment that is more likely to adopt the proper conformation.