The more senior I've become, the more likely I am to argue against reviewers' comments. It hasn't hurt me, and I think it's actually helped the average quality of the end-product, but I'm also at a point where any individual publication is less likely to affect my career.
If this is your first R&R, do exactly what Mellony said. And if you do decide to not change something, be sure to oversell how much you changed the other things R2 recommended and talk about how the thing you chose not to change isn't really a big deal anyway because the other points improved the quality of the article so much.