Picosecond resonance Raman evidence for unrelaxed heme in the (carbonmonoxy)myoglobin photoproduct

An actively and passively mode-locked Nd:YAG laser, producing 30-ps pulses of 1-mJ energy at 532 nm, has been used to photolyze (carbonmonoxy)myoglobin (MbCO) and generate its resonance Raman spectrum, which was recorded with a vidicon multichannel analyzer. The photoproduct spectrum was obtained by subtraction of the MbCO spectrum, obtained at lower incident power levels. Comparison with the spectrum of deoxyMb, obtained with the same apparatus, revealed frequency downshifts of approximately 4 cm-1, for bands at 1604, 1554, and 1542 cm-1, which are identified with porphyrin skeletal modes v10, v19, and v11. These frequencies are known to correlate inversely with the core size of the porphyrin ring, and the shifts imply a larger core size for the photoproduct than for deoxyMb. Similar shifts have been observed for the (carbonmonoxy)hemoglobin (HbCO) photoproduct; in that case, the shifts persist for longer than 20 ns, whereas they are absent in the MbCO photoproduct spectrum within 7 ns of photolysis. The unrelaxed state of the heme group region is therefore suggested to be maintained by protein forces, which relax more rapidly for Mb than Hb. This may reflect a tighter coupling in Hb of the out-of-plane movement of the Fe atom with the proximal histidine-containing F helix.     

Effect of solvent viscosity on the heme-pocket dynamics of photolyzed (carbonmonoxy)hemoglobin

The heme-pocket dynamics subsequent to carbon monoxide photolysis from human hemoglobin have been monitored as a function of glycerol-water solvent composition with time-resolved resonance Raman spectroscopy. Prompt (geminate) ligand recombination rates and the transient heme-pocket geometry established within 10 ns after photolysis appear to be largely independent of solvent composition. The rate of relaxation of the transient geometry to an equilibrium deoxy configuration is, however, quite sensitive to solvent composition. These observations suggest that the former processes result from local, internal motions of the protein, while the relaxation dynamics of the proximal heme pocket are predicated upon more global protein motions that are dependent upon solvent viscosity.     

Resonance assignments in the proton-NMR spectrum of carbonmonoxy hemoglobin by two-dimensional methods

Proton-nuclear-magnetic-resonance spectroscopy is a powerful tool for investigating the solution structure of biopolymers provided that a substantial number of proton resonances are assigned in the spectrum. For large proteins the assignments have usually been made by the comparative one-dimensional NMR investigations of the parent and derivative proteins in different physicochemical conditions. In this paper, we show that the more powerful two-dimensional methods could be successfully applide to proteins of the size of human adult hemoglobin (Mr = 64,500). J-Correlated and NOE-correlated spectroscopy, together with topological relationships in the known crystalline structure, enabled us to assign a large number of resonances. The majority of the assigned resonances correspond to the heme substituents and to amino acids in the heme pockets of the two subunits. These results thus provide an extensive set of intrinsic probes for mapping the conformation of the ligand-binding site and its functional changes. Comparison of the observed ring-current shifts of the assigned resonances with those calculated from the known crystallographic coordinates suggests a close similarity between the heme-pocket tertiary conformation in solution and in the crystalline state. A significant difference was noted for Leu141 in beta subunits which, in solution, appears to have stronger contacts with the heme groups than in the crystalline form. The present results also demonstrate that two-dimensional-NMR methods could be successfully applied to the investigation of the structure of large biomolecules in solution (Mr less than or equal to 65,000).     

Zn(II)-induced dimerization of human carbonmonoxy hemoglobin

Binding of Zn(II) to the carbon monoxide complex of human hemoglobin was shown by equilibrium sedimentation and sedimentation velocity experiments at pH 7.0 to induce the dissociation of liganded tetramers to dimers but not to monomers. These results provide direct confirmation of previous kinetic and gel filtration experiments (R. D. Gray, (1980) J. Biol. Chem.255, 1812–1818) that Zn(II) binding to liganded hemoglobin produces a change in aggregation state of liganded hemoglobin.     

Time-resolved Raman spectroscopy with subpicosecond resolution: vibrational cooling and delocalization of strain energy in photodissociated (carbonmonoxy)hemoglobin

A Raman spectrometer that provides both subpicosecond resolution and independent, tunable pump and probe pulses is described. The spectrometer is employed to obtain time-resolved spectra of (carbonmonoxy)hemoglobin (HbCO) at times from 0.2 to 95 ps subsequent to ligand photodissociation. The spectra are interpreted in terms of a vibrationally hot heme that cools substantially in 10 ps. Concomitant with the proposed vibrational cooling is a slower relaxation, which we suggest results from a protein response to heme doming induced by ligand detachment. Results and interpretations are discussed in the context of current models of the heme photophysics and of hemoglobin reactivity.     

Response of the local heme environment of (carbonmonoxy)hemoglobin to protein dehydration

The effects of protein dehydration upon the equilibrium and dynamic properties of the heme active site in human hemoglobin (HbA) have been probed by resonance Raman scattering. Spectra of equilibrium carbonmonoxy-HbA and the photolytic heme transient species generated within 10 ns of ligand photolysis have been obtained from thin films of protein in various stages of dehydration. These data provide detailed information concerning the response of the heme and its bonding interactions with both the proximal histidine and carbon monoxide as a function of protein hydration. For protein hydration levels of 0.4-1.0 g of H2O/g of protein, our results indicate that the C = O stretching mode of carbonmonoxy-HbA is dramatically affected by protein hydration levels, thus corroborating the infrared results of Brown et al. [Brown, W. E., Sutcliffe, J. W., & Pulsinelli, P. D. (1983) Biochemistry 22, 2914-2923]. However, we find that both heme skeletal modes and the Fe-C bond strength are largely insensitive to dehydration. Moreover, the proximal pocket geometry (as reflected in the behavior of the Fe-proximal histidine stretching mode) immediately following ligand photolysis was found to be very similar to that of R-state solution hemoglobin. At protein hydration levels below the theoretical monolayer limit, small changes in the resonance Raman spectra of both equilibrium HbCO and the transient heme species generated subsequent to ligand photolysis are detected. These include broadening of the Fe-C stretching mode in equilibrium HbCO and a small shift to lower frequency of the Fe-His mode in the photolytic transient species.(ABSTRACT TRUNCATED AT 250 WORDS)     

Resonance Raman analysis of the mechanism of energy storage and chromophore distortion in the primary visual photoproduct

The vibrational structure of the chromophore in the primary photoproduct of vision, bathorhodopsin, is examined to determine the cause of the anomalously decoupled and intense C(11)=C(12) hydrogen-out-of-plane (HOOP) wagging modes and their relation to energy storage in the primary photoproduct. Low-temperature (77 K) resonance Raman spectra of Glu181 and Ser186 mutants of bovine rhodopsin reveal only mild mutagenic perturbations of the photoproduct spectrum suggesting that dipolar, electrostatic, or steric interactions with these residues do not cause the HOOP mode frequencies and intensities. Density functional theory calculations are performed to investigate the effect of geometric distortion on the HOOP coupling. The decoupled HOOP modes can be simulated by imposing approximately 40 degrees twists in the same direction about the C(11)=C(12) and C(12)-C(13) bonds. Sequence comparison and examination of the binding site suggests that these distortions are caused by three constraints consisting of an electrostatic anchor between the protonated Schiff base and the Glu113 counterion, as well as steric interactions of the 9- and 13-methyl groups with surrounding residues. This distortion stores light energy that is used to drive the subsequent protein conformational changes that activate rhodopsin.     

The enigma of the liganded hemoglobin end state: a novel quaternary structure of human carbonmonoxy hemoglobin

The liganded hemoglobin (Hb) high-salt crystallization condition described by Max Perutz has generated three different crystals of human adult carbonmonoxy hemoglobin (COHbA). The first crystal is isomorphous with the "classical" liganded or R Hb structure. The second crystal reveals a new liganded Hb quaternary structure, RR2, that assumes an intermediate conformation between the R form and another liganded Hb quaternary structure, R2, which was discovered more than a decade ago. Like the R2 structure, the diagnostic R state hydrogen bond between beta2His97 and alpha1Thr38 is missing in the RR2 structure. The third crystal adopts a novel liganded Hb conformation, which we have termed R3, and it shows substantial quaternary structural differences from the R, RR2, and R2 structures. The quaternary structure differences between T and R3 are as large as those between T and R2; however, the T --> R3 and T --> R2 transitions are in different directions as defined by rigid-body screw rotation. Moreover, R3 represents an end state. Compared to all known liganded Hb structures, R3 shows remarkably reduced strain at the alpha-heme, reduced steric contact between the beta-heme ligand and the distal residues, smaller alpha- and beta-clefts, and reduced alpha1-alpha2 and beta1-beta2 iron-iron distances. Together, these unique structural features in R3 should make it the most relaxed and/or greatly enhance its affinity for oxygen compared to the other liganded Hbs. The current Hb structure-function relationships that are now based on T --> R, T -->R --> R2, or T --> R2 --> R transitions may have to be reexamined to take into account the RR2 and R3 liganded structures.     

Ultraviolet Raman spectroscopy indicates fast (less than 7 ns) R----T-like motion in hemoglobin

Raman spectra of oxy- and deoxyhemoglobin obtained with 218 and 200 nm pulsed (7 ns) laser excitation show changes (loss of 880 cm-1 tryptophan band intensity, increase in the 830/850 cm-1 tyrosine doublet intensity ratio) which are attributed to the aromatic contacts (Trp beta 37-Tyr alpha 140 and Tyr alpha 42-Asp beta 99) that are specific to the T quaternary structure. At high concentration (2 mM in heme) HbCO shows the same spectral signatures as HbO2. As the HbCO concentration is decreased, however, the spectra approach those shown by deoxy-Hb. This dilution effect is attributable to photolysis, which increases with decreasing concentration. The results imply that the HbCO photoproduct shows the same aromatic environments as does deoxy-Hb. Thus, T-like contacts are apparently formed at the alpha 1 beta 2 interface within 7 ns of photolysis, a time short compared to the spectral alterations of the heme group (approximately 100 ns, approximately 1 microsecond, and approximately 20 microseconds) which have previously been attributed to tertiary and quaternary relaxations.     

Resonance Raman studies of CO and O2 binding to elephant myoglobin (distal His(E7)----Gln)

Carbon monoxide and dioxygen were employed as resonance Raman-visible ligands for probing the nature of the heme-binding site in elephant myoglobin, which has glutamine in the distal position (E7) instead of the usual histidine. The distal histidine (E7) residue has been thought to be responsible for weakening carbon monoxide binding to hemoproteins. It is of interest to see how the His(E7)----Gln replacement affects such parameters as nu(Fe-N epsilon), nu(Fe-CO), delta(Fe-C-O), nu(C-O), delta(Fe-O-O), and nu(O-O) vibrational frequencies and relative intensities. Elephant myoglobin has a CO affinity approximately 6 times higher than that for human/sperm whale myoglobin (Mb). If this enhanced affinity were solely due to the removal of some of the steric hindrance that normally tilts the CO off the heme axis, one would expect the nu(Fe-CO) frequency to decrease and the nu(C-O) frequency to increase relative to the corresponding values in sperm whale Mb. However, the opposite was found. In addition, strong enhancement of the Fe-C-O bending mode was observed. These results suggest that the Fe-C-O linkage remains distorted. In elephant Mb, new interactions resulting from the conformational change accompanying ligand binding may be responsible for the increased CO binding. Similar spectra were obtained for elephant and sperm whale oxymyoglobin. This suggests that the interactions of bound O2 are not markedly affected by the glutamine replacement.     

Resonance Raman characterization of the heme prosthetic group in eosinophil peroxidase

The resonance-enhanced Raman spectrum of eosinophil peroxidase (EPO) from horse and human eosinophils is reported. Based upon the spectral energies, distribution and depolarization ratios of the high-frequency skeletal modes and upon the presence of weak bands assignable to vinyl substituent groups, we conclude that the heme prosthetic group is high-spin, 6 coordinate protoporphyrin. The Raman spectrum reveals clear differences from lactoperoxidase (LPO), an enzyme which appears nearly structurally isomorphous by other physical techniques; the data indicate a stronger axial 6th ligand in EPO. Mechanistic implications are discussed in relation to LPO and myeloperoxidase, an enzyme present in neutrophils and monocytes which contains a unique functional active-site chlorin.     

A 2-deoxyribonolactone-containing nucleotide: isolation and characterization of the alkali-sensitive photoproduct of the trideoxyribonucleotide d(ApCpA)

It has been reported that ACA sequences in DNA are mutagenic hot spots in UV-induced mutagenesis and are sites of an alkali-sensitive lesion produced by UV irradiation. In order to characterize the UV-induced lesion of an ACA site, chemically synthesized trideoxyribonucleotide d(ApCpA) was irradiated with UV light and the alkali-sensitive photoproduct was isolated. The structure of this photoproduct was characterized as the trinucleotide containing 2-deoxyribonolactone at the internal residue by 2D DQF-COSY, FT-IR, FAB-MS, and chemical properties. It is known that this lesion is also produced by gamma-irradiation, neocarzinostatin, the 1,10-phenanthroline-copper complex, and hydrogen peroxide and is highly mutagenic because of its resistance to cleavage by certain apurinic/apyrimidinic (AP) endonucleases. Thus, 2-deoxyribonolactone may be one of the lethal DNA lesions induced by UV irradiation to organisms and one of the intermediates of UV-induced DNA strand breaks because the DNA strand is cleaved at this site with beta- and subsequent delta-elimination mechanisms.     

Heme-bound tyrosine vibrations in hemoglobin M: Resonance Raman,crystallography, and DFT calculation

Hemoglobins M (Hbs M) are human hemoglobin variants in which either the α or β subunit contains a ferric heme in the α₂β₂ tetramer. Though the ferric subunit cannot bind O₂, it regulates O₂ affinity of its counterpart ferrous subunit. We have investigated resonance Raman spectra of two Hbs, M Iwate (α87His → tyrosine [Tyr]) and M Boston (α58His → Tyr), having tyrosine as a heme axial ligand at proximal and distal positions, respectively, that exhibit unassigned resonance Raman bands arising from ferric (not ferrous) hemes at 899 and 876 cm^-1. Our quantum chemical calculations using density functional theory on Fe-porphyrin models with p-cresol and/or 4-methylimidazole showed that the unassigned bands correspond to the breathing-like modes of Fe³⁺-bound Tyr and are sensitive to the Fe-O-C(Tyr) angle. Based on the frequencies of the Raman bands, the Fe-O-C(Tyr) angles of Hbs M Iwate and M Boston were predicted to be 153.5° and 129.2°, respectively. Consistent with this prediction, x-ray crystallographic analysis showed that the Fe-O-C(Tyr) angles of Hbs M Iwate and M Boston in the T quaternary structure were 153.6° and 134.6°, respectively. It also showed a similar Fe-O bond length (1.96 and 1.97 Å) and different tilting angles.     

Resonance Raman spectroscopic characterization of compound III of lignin peroxidase

Resonance Raman (RR) spectra of several compounds III of lignin peroxidase (LiP) have been measured at 90 K with Soret and visible excitation wavelengths. The samples include LiPIIIa (or oxyLiP) prepared by oxygenation of the ferrous enzyme, LiPIIIb generated by reaction of the native ferric enzyme with superoxide, LiPIIIc prepared from native LiP plus H2O2 followed by removal of excess peroxide with catalase, and LiPIII* made by addition of excess H2O2 to the native enzyme. The RR spectra of these four products appear to be similar and, thus, indicate that the environments of these hexacoordinate, low-spin ferriheme species must also be very similar. Nonetheless, the Soret absorption band of LiPIII* is red-shifted by 5 nm from the 414-nm maximum common to LiPIIIa, -b, and -c [Wariishi, H., & Gold, M.H. (1990) J. Biol. Chem. 265, 2070-2077]. Analysis of the iron-porphyrin vibrational frequencies indicates that the electronic structures for the various compounds III are consistent with an FeIIIO2.-formulation. The spectral changes observed between the oxygenated complex and the ferrous heme of lignin peroxidase are similar to those between oxymyoglobin and deoxymyoglobin. The contraction in the core sizes in compound III relative to the native peroxidase is analyzed and compared with that of other heme systems. EPR spectra confirm that the high-spin ferric form of the native enzyme, with an apparent g = 5.83, is converted into the EPR-silent LiPIII* upon addition of excess H2O2. Its magnetic behavior may be explained by anti-ferromagnetic coupling between the low-spin FeIII and the superoxide ligand.(ABSTRACT TRUNCATED AT 250 WORDS)     

Resonance Raman characterization of biotin sulfoxide reductase. Comparing oxomolybdenum enzymes in the ME(2)SO reductase family

Resonance Raman spectroscopy has been used to define active site structures for oxidized Mo(VI) and reduced Mo(IV) forms of recombinant Rhodobacter sphaeroides biotin sulfoxide reductase expressed in Escherichia coli. On the basis of (18)O/(16)O labeling studies involving water and the alternative substrate dimethyl sulfoxide and the close correspondence to the resonance Raman spectra previously reported for dimethyl sulfoxide reductase (Garton, S. D., Hilton, J., Oku, H., Crouse, B. R., Rajagopalan, K. V., and Johnson, M. K. (1997) J. Am. Chem. Soc. 119, 12906-12916), vibrational modes associated with a terminal oxo ligand and the two molybdopterin dithiolene ligands have been assigned. The results indicate that the enzyme cycles between mono-oxo-Mo(VI) and des-oxo-Mo(IV) forms with both molybdopterin dithiolene ligands remaining coordinated in both redox states. Direct evidence for an oxygen atom transfer mechanism is provided by (18)O/(16)O labeling studies, which show that the terminal oxo group at the molybdenum center is exchangeable with water during redox cycling and originates from the substrate in substrate-oxidized samples. Biotin sulfoxide reductase is not reduced by biotin or the nonphysiological products, dimethyl sulfide and trimethylamine. However, product-induced changes in the Mo=O stretching frequency provide direct evidence for a product-associated mono-oxo-Mo(VI) catalytic intermediate. The results indicate that biotin sulfoxide reductase is thermodynamically tuned to catalyze the reductase reaction, and a detailed catalytic mechanism is proposed.     

Resonance Raman evidence for an unusually strong exogenous ligand-metal bond in a monomeric nitrosyl manganese hemoglobin

Resonance Raman spectroscopy has been employed to determine the vibrational modes of monomeric nitrosyl manganese Chironomus thummi thummi hemoglobin (CTT IV). This insect hemoglobin has no distal histidine. By applying various isotope-labeled nitric oxides (14N16O, 15N16O, 14N18O), we have identified the Mn11-NO stretching model at 628 cm-1, the Mn11-N-O bending mode at 574 cm-1 and the N-O stretching mode at 1735 cm-1. The results suggest a strong Mn11-NO bond and a weak N-O bond. The vinyl group substitution does not influence the nu (Mn11-NO), delta (Mn11-N-O) and nu (N-O) vibrations. The Mn11-NO stretching frequency is insensitive to distal histidine interactions with NO, whereas the N-O stretching frequency is sensitive. Nitric oxide also binds to Met manganese CTT IV to form an Mn111. NO complex which undergoes a slow but complete autoreduction resulting in the Mn11.NO species. In manganese meso-IX CTT IV, the Mn111. NO Mn11. NO conversion alters the intensities of the porphyrin ring modes at 342, 360, 1587 and 1598 cm-1, but shifts the frequencies at 1504 and 1633 cm-1 (in Mn111.NO) to 1497 and 1630 cm-1 (in Mn11. NO), respectively. The unshifted marker line at 1378 cm-1 reflects the fact that the pi electron densities of the porphyrin ring are the same in the two complexes.     

Resonance Raman evidence for Fe(IV) in compound II of horseradish peroxidase

Resonance Raman spectra have been obtained for Compound II of horseradish peroxidase. Its prophyrin vibrational frequencies are consistent with a planar low-spin heme containing Fe(IV). The oxidation-state marker band is found at the unprecedentedly high value of 1382 cm^?1. This band was also observed in solutions of myoglobin and cytochrome c peroxidase to which H₂O₂ had been added. No evidence was found for an actual FeO double bond in Compound II.