Hemoglobin residues Thr alpha 67 and Ser beta 72 have been mutated to Val and Ala, respectively, to test the hypothesis that tertiary H-bonds involving these residues play a key role in the allosteric reaction path between the R and the T state. The H-bonds are donated by the indole side chains of Trp alpha 14, and Trp beta 15; they bridge the outer A helices to the inner E helices, which line the distal side of the heme pocket. The mutants fold properly (CD measurements) and form native-like T state contacts, as revealed by UVRR (RR = resonance Raman) difference spectra between deoxyHb and KbCO, and by the Fe-N (histidine) stretching band in the visible RR spectra of deoxyHb. However, the UVRR intensity of tryptophan bands is diminished in the mutants. This is the expected effect of H-bond elimination, because H-bonding shifts the tryptophan excitation profiles to longer wavelengths, raising the intensity; at 229 nm, the wavelength employed in this study. Consistent with this interpretation, the intensity loss for the W3 band is found exclusively at 1558 cm(-1), the position of Trp alpha 14 and Trp beta 15, and not at 1548 cm(-1), the position of the interfacial residue Trp beta 37: The intensity loss is greater for T alpha 67V than for S beta 72A, consistent with crystallographic data showing a shorter N ... O distance for the H-bond from Trp alpha 14 than from Trp beta 15. The H-bond augmentation of the W3 intensity is calculated to be almost a factor of 2 greater for the former than the latter. UVRR difference spectra obtained 150 ns after photolysis of HbCO reveal negative Tyr and Trp bands for the mutants which are similar to those obtained for native Hb, and are attributed to the first protein intermediate on the allosteric reaction path, R-deoxy. However, the Trp intensity loss is diminished for the mutants, supporting the hypothesis that the R-deoxy Trp signals arise from weakening of the Trp alpha 14 and Trp beta 15 H-bonds, as a result of increased separation between the A and E helices. This separation is proposed to result from rotation of the EF "clamshell" resulting from F helix displacement away from the heme plane, due to the Fe displacement upon deligation, and E helix motion toward the heme plane as the ligand departs the heme pocket.