It is definitely a possibility. That particular protein may be more susceptible to sonication than other proteins. A quick check would be to prepare your purified protein in the same manner (even though it is already purified). You could potentially take 100ng of your purified protein and add the lysis buffer and sonicate. Run your samples on a dot blot and see if this destroyed the antibody recognition site. I only suggest doing a dot blot because they can be completed within 3hrs as compared to a two day western blot (depending on your protocol).
Have you tried just adding lysis buffer to your Cos-7 cells, mildly vortexing, and incubating on ice for 30 minutes, spinning down and removing the supernatant? This is significantly milder than sonication and may or may not solve your problem.
If your protocol was given to you and sonication has worked in the past, maybe you should double check your output settings.
I haven't worked with Cos-7 cells, but I do work with HEK cells quite frequently.
Edit - Is this protein expressed from a plasmid that you transfected? If it is, maybe you should take a step back and see if you have mRNA expression via a quick PCR reaction. You don't make it clear as to whether you purified the protein or if the protein is a manufacturer recombinant protein.