There are a few options for this;
#1-use Elastic Audio and lower all the pitched instruments by 1 or 2 steps(undo this after the vocal is recorded and shift the vocal up to match).
#2-bounce a 2-mix at 48K, then import that to a blank session at 44.1K(
without sample rate conversion) and record the singer. Export the vocal and bring into the 48K session and adjust pitch.
#3-Bounce a 2-mix and import back into the session>use Serato Pitch 'n Time to lower the 2-mix. Mute(or make inactive) the rest of the tracks until you get the vocal recorded.
#4-I'm sure there's at least 4 ways and maybe more
My 2 cents; doing this may deliver varying results. Lowering/raising by 1 or 2 half-steps might be "okay"(depending on how picky you or the singer are) and there are plugins that can alter the formant to get closer to "correct". After 3 half-steps(or more), its likely that the results will not sound good. I often work with inexperienced singers that are not savvy on where their range really is and they try to sing stuff that is just never going to work. In those cases, I recommend that they alter the melody to sit within their actual range(instead of the range they "think" they can sing
)