56
M. Stein
calculations when a μ-hydroxide (OH
− ) in the bridging position between the nickel
and iron atoms is assumed (see Table 4). QM calculations allow to probe different
binding scenarios and then to elucidate the conformation of the hydroxide. The
1 HENDOR transition rotational dependence can only be explained when the proton
of the hydroxide is in the ‘endo’ position (see Fig. 7). In the ‘exo’ position, the
proton would be in a different orientation with respect to the d z2 orbital of the Ni,
give rise to different
1 H hyperfine coupling signs, different hyperfine tensor principal
axes orientations and different rotation angle-dependent ENDOR spectra. Here, the
calculations do reproduce not only the
1 H hyperfine couplings in magnitude but
also resolve experimental ambiguities in terms of the directionality of magnetic
interactions and the orientation of the protonated bridging ligand. Table 4 summarizes
the comparison of experimental and calculated
1 H hyperfine coupling constants.
BP86 calculation using ZORA perform equally well as those with the B3LYP hybrid
functional and the mean-field SOMFI operator. As a matter of fact, the isotropic
hyperfine interactions from ZORA are in even better agreement with experiment,
which might be due to the use of Slater-type basis sets and their correct description
of the electron cusp close to the nucleus.
4.3.3 Manganese Zero-Field Splitting Parameters in Protein Single
Crystals
The enolase enzyme catalyzes the reversible dehydration of D-2-phosphoglycerate
to phosphopyruvate in the glycolytic pathway. The native enzyme uses Mg
2+ in
two binding sites I and II for the reaction but Mn
2+ /Zn
2+ substitutions also display
catalytic activity in vivo. The enolase-PhAH inhibitor complex (PhAH phosphonoacetohydroxamate) from yeast was investigated in manganese/zink-substituted
protein single crystals using high-frequency W-band EPR and
1 H-ENDOR experiments [83]. The crystal structure of P2 1 symmetry was refined to a resolution of
1.54 Å. The rotational angle-dependent EPR spectra were recorded in two different
protein crystal orientations parallel and perpendicular to a morphological crystal axis
[83]. The
1 H-ENDOR results resolved the location of two water molecule ligands
and showed their involvement in a hydrogen bonding network connecting the metal
ion and the inhibitor. The Mn
2+ (S 5/2) zero-field splitting (ZFS) parameters D
and D/E were determined experimentally and their assignment was based on DFT
calculations which included spin-dependent two-electron second order contributions
[84]. The D value is a very sensitive measure for any deviation from cubic symmetry
and accurate structural models are required. The BP86/CP(PPP), TZVP, SV(P) calculations on a partially constrained cluster model with 81 atoms (see Fig. 8) and the
spin–orbit mean field approximation significantly overestimated the experimental
ZFS parameters by 50–70% and were then scaled by a factor of 2/3. The calculated
high rhombicity E/D for site I was in agreement with its low coordination symmetry
in the crystal structure. While the sign of D could not be determined experimentally,
the DFT calculations predicted a negative zero-field splitting D and a larger absolute
value of |D|. The calculated orientations of the principal axes of the ZFS for Mn
2+
M. Stein
calculations when a μ-hydroxide (OH
− ) in the bridging position between the nickel
and iron atoms is assumed (see Table 4). QM calculations allow to probe different
binding scenarios and then to elucidate the conformation of the hydroxide. The
1 HENDOR transition rotational dependence can only be explained when the proton
of the hydroxide is in the ‘endo’ position (see Fig. 7). In the ‘exo’ position, the
proton would be in a different orientation with respect to the d z2 orbital of the Ni,
give rise to different
1 H hyperfine coupling signs, different hyperfine tensor principal
axes orientations and different rotation angle-dependent ENDOR spectra. Here, the
calculations do reproduce not only the
1 H hyperfine couplings in magnitude but
also resolve experimental ambiguities in terms of the directionality of magnetic
interactions and the orientation of the protonated bridging ligand. Table 4 summarizes
the comparison of experimental and calculated
1 H hyperfine coupling constants.
BP86 calculation using ZORA perform equally well as those with the B3LYP hybrid
functional and the mean-field SOMFI operator. As a matter of fact, the isotropic
hyperfine interactions from ZORA are in even better agreement with experiment,
which might be due to the use of Slater-type basis sets and their correct description
of the electron cusp close to the nucleus.
4.3.3 Manganese Zero-Field Splitting Parameters in Protein Single
Crystals
The enolase enzyme catalyzes the reversible dehydration of D-2-phosphoglycerate
to phosphopyruvate in the glycolytic pathway. The native enzyme uses Mg
2+ in
two binding sites I and II for the reaction but Mn
2+ /Zn
2+ substitutions also display
catalytic activity in vivo. The enolase-PhAH inhibitor complex (PhAH phosphonoacetohydroxamate) from yeast was investigated in manganese/zink-substituted
protein single crystals using high-frequency W-band EPR and
1 H-ENDOR experiments [83]. The crystal structure of P2 1 symmetry was refined to a resolution of
1.54 Å. The rotational angle-dependent EPR spectra were recorded in two different
protein crystal orientations parallel and perpendicular to a morphological crystal axis
[83]. The
1 H-ENDOR results resolved the location of two water molecule ligands
and showed their involvement in a hydrogen bonding network connecting the metal
ion and the inhibitor. The Mn
2+ (S 5/2) zero-field splitting (ZFS) parameters D
and D/E were determined experimentally and their assignment was based on DFT
calculations which included spin-dependent two-electron second order contributions
[84]. The D value is a very sensitive measure for any deviation from cubic symmetry
and accurate structural models are required. The BP86/CP(PPP), TZVP, SV(P) calculations on a partially constrained cluster model with 81 atoms (see Fig. 8) and the
spin–orbit mean field approximation significantly overestimated the experimental
ZFS parameters by 50–70% and were then scaled by a factor of 2/3. The calculated
high rhombicity E/D for site I was in agreement with its low coordination symmetry
in the crystal structure. While the sign of D could not be determined experimentally,
the DFT calculations predicted a negative zero-field splitting D and a larger absolute
value of |D|. The calculated orientations of the principal axes of the ZFS for Mn
2+