On the real part of the proton optical model potential