MD of. One of the chain can be done if only one of them is involved in  the functional (catalysis/reaction) part of the protein. If protein can function without the assistance of 2nd chain of homodimer then you can do MD of single chain of the homodimer. However if conformational change of both chains lead to the catalysis or help the protein to do its function. Then you have to do MD of homodimer. To check this, consult the literature and read about the functional domain of protein and the chain that participate in the enzyme kinetics 

"In other words, how will the dynamics of a monomer differ from a dimer for homodimers particularly in the context of side chain conformations & dimer interface ?"

A good question that can be answered by performing long MD of the monomer and dimer :)

It's really hard to otherwise answer the question. If I was to review this work I would definitely not be happy with including only the monomer and will raise this problem, asking to do more MD. As Indu suggested, consulting the literature to justify this step is the best idea.

Thank U indukumari & trzaskowsi for ur valuable inputs.

This is not the conventional MD but a fast dynamics for understanding near native structures of proteins.

......This fast MD of 10ns is just to understand how a terminal missing loop folds...... I did MD of the 2 chains separately & as both chains are exactly identical they gave same kind of molecular motions, the structures only differing by their translations in space. I combined both of them together after preserving dimer interface contacts through modeller and again subjected this to an Energy minimization. I am subjecting the resulting structure for all downstream studies...

Any comments if U were to review this ???