Studies of oxygen transport in pregnant mammals have shown that the O2-saturation curves of fetal and maternal blood are markedly different when measured under the same conditions. Fetal erythrocytes contain a structural variant of hemoglobin, HbF, consisting of two γ and two β subunits (γ2β2), whereas maternal erythrocytes contain HbA (α2β2).

When all the BPG is carefully removed from samples of HbA and HbF, the measured O2-saturation curves (and consequently the O2 affinities) are displaced to the left. However, HbA now has a greater affinity for oxygen than does HbF. When BPG is reintroduced, the O2-saturation curves return to normal.
What is the effect of BPG on the O2 affinity of hemoglobin and why are HbA and HbF impacted differently by BPG?