Pybluez-0.16 still depends on bluez-libs, althought bluez-4.* has been unmasked, causing a conflict.
DEPEND="|| ( net-wireless/bluez >=net-wireless/bluez-libs-2.10 )" I don't see the problem here.
Please get back to us and reopen with the information requested.
Sorry for the very long delay, I've been away... I'm sorry, it seems that I made a blunder here... I based this bug report on what portage told me, and the output of equery depends bluez-libs the latter of which, in this situation, gives false positives.