[modeller_usage] Post - problem with modelling a chain using another chain as template

Is it possible to re-orient chains in a pentameric way after using one of

those chains as template to build others?

I have a pentameric protein. The protein and the nearest template miss

some residues at a particular position in all the chains, except one chain.

So, I carried out homology modeling using modeller9v3 as follows:

1. I modelled the pentamer at a time using the nearest available pdb file as template (~85% similarity).
2. Each of the chains are about 70% similar to each other. I built a model of chain C Using chain A of available pdb file
3. Then, I overlapped chain C from step 3 and step 1, using VMD "RMSD tool". Saved both pdb's in this orientation. Copied the "region of missing residues (14) only" to chain C of step 1.
4. Similarly I did for chain B.
5. When, I reassembled chains B and C with other 3 subunits from step 1,

Chains B and C are overlapping with chain A. Imagine, its like a trimer now, with 3 chains overlapping at one place.
I tried both VMD, InsightII using Chains from the pentamer as reference and overlay the modelled chains. But, there is no change in orientation. Is there any way I can do this? Is this a plausible way to build missing residues, when there is not at all any template available for that region in the entire family of this protein?