A quantum-chemical picture of hemoglobin affinity.

Understanding the molecular mechanism of hemoglobin cooperativity remains an enduring challenge. Protein forces that control ligand affinity are not directly accessible by experiment. We demonstrate that computational quantum mechanics/molecular mechanics methods can provide reasonable values of ligand binding energies in Hb, and of their dependence on allostery. About 40% of the binding energy differences between the relaxed state and tense state quaternary structures result from strain induced in the heme and its ligands, especially in one of the pyrrole rings. The proximal histidine also contributes significantly, in particular, in the alpha-chains. The remaining energy difference resides in protein contacts, involving residues responsible for locking the quaternary changes. In the alpha-chains, the most important contacts involve the FG corner, at the "hinge" region of the alpha(1)beta(2) quaternary interface. The energy differences are spread more evenly among the beta-chain residues, suggesting greater flexibility for the beta- than for the alpha-chains along the quaternary transition. Despite this chain differentiation, the chains contribute equally to the relaxed substitute state energy difference. Thus, nature has evolved a symmetric response to the quaternary structure change, which is a requirement for maximum cooperativity, via different mechanisms for the two kinds of chains.

Heme-based sensors: theoretical modeling of heme-ligand-protein interactions.

Heme proteins are uniquely adapted to bind the important diatomic molecules O(2), NO and CO. An increasing number of heme proteins are being discovered that sense these molecules and thereby regulate a variety of biochemical responses. The interactions of diatomic molecules with heme, and with the surrounding protein, are therefore of great interest. Recent theoretical modeling, using density functional theory, captures many features of these interactions, as exemplified by the well-characterized heme protein myoglobin. Important details, however, especially the mutual influence of the bound diatomic molecule and the proximal ligand, remain to be clarified.

Carbon monoxide binding by de novo heme proteins derived from designed combinatorial libraries.

Carbon monoxide binding was studied in a collection of de novo heme proteins derived from combinatorial libraries of sequences designed to fold into 4-helix bundles. The design of the de novo sequences was based on the previously reported "binary code" strategy, in which the patterning of polar and nonpolar amino acids is specified explicitly, but the exact identities of the side chains are varied extensively.(1) The combinatorial mixture of amino acids included histidine and methionine, which ligate heme iron in natural proteins. However, no attempt was made to explicitly design a heme binding site. Nonetheless, as reported previously, approximately half of the binary code proteins bind heme.(2) This collection of novel heme proteins provides a unique opportunity for an unbiased assessment of the functional potentialities of heme proteins that have not been prejudiced either by explicit design or by evolutionary selection. To assess the capabilities of the de novo heme proteins to bind diatomic ligands, we measured the affinity for CO, the kinetics of CO binding and release, and the resonance Raman spectra of the CO complexes for eight de novo heme proteins from two combinatorial libraries. The CO binding affinities for all eight proteins were similar to that of myoglobin, with dissociation constants (K(d)) in the low nanomolar range. The CO association kinetics (k(on)) revealed that the heme environment in all eight of the de novo proteins is partially buried, and the resonance Raman studies indicated that the local environment around the bound CO is devoid of hydrogen-bonding groups. Overall, the CO binding properties of the de novo heme proteins span a narrow range of values near the center of the range observed for diverse families of natural heme proteins. The measured properties of the de novo heme proteins can be considered as a "default" range for CO binding in alpha-helical proteins that have neither been designed to bind heme or CO, nor subjected to genetic selections for heme or CO binding.

Assignment of the 1511 cm(-1) UV resonance Raman marker band of hemoglobin to tryptophan.

New UV resonance Raman (UVRR) data provide convincing evidence that a characteristic 1511 cm(-1) band in the T - R difference spectra of hemoglobin is due to the overtone of the Trp W18 fundamental at 756 cm(-1). Measured isotope shifts for 2-H and 15-N substitution at the indole NH group are twice as large for the 1511 cm(-1) band as for W18, and the 1511 cm(-1) intensity scales with that of W18 in the difference spectrum. Moreover, the UVRR excitation profile of the 1511 cm(-1) band tracks that of another tryptophan band, W16. Both are redshifted in hemoglobin, relative to aqueous tryptophan, reflecting H bonding within a hydrophobic environment in the protein. The 2xW18 assignment had been thrown into question by the observation of remnant 1511 cm(-1) intensity in the T - R spectra of hemoglobin labeled with tryptophan-d(5), a substitution that shifts W18 over 50 cm(-1). However, reexamination of the data suggests that this remnant intensity may result from a subtraction artifact arising from the downshift of another difference band, W3, from 1549 cm(-1) in unlabeled protein to 1522 cm(-1) in labeled protein. Restoration of the 2xW18 assignment establishes that the 1511 cm(-1) difference band, which is a useful indicator of the extent of T-state formation in hemoglobin, arises from the same residue, Trpbeta37, that gives rise to essentially all of the T - R signal from tryptophan.

Is the CO adduct of myoglobin bent, and does it matter?

Early reports of a severely bent CO adduct in myoglobin inspired the idea that heme proteins discriminate against CO, relative to O(2), via steric hindrance imposed by a distal histidine residue. Recent results showing that the bound CO is only slightly distorted do not by themselves overthrow the steric hypothesis, because the steric energy could be stored in displacements of the protein. However, experimental data on site-directed mutants show that the main determinant of ligand affinity changes is the polarity of the binding pocket and that H-bonding by the distal histidine accounts for about 85% of the O(2)/CO discrimination while steric hindrance accounts for the remaining 15%.

Selective enhancement of resonance Raman spectra of separate bacteriopheophytins in Rb. sphaeroides reaction centers.

Tunable dye laser excitation of carefully prepared samples of Rb. sphaeroides reaction centers provides richly detailed resonance Raman (RR) spectra of the bacteriopheophytins, H, and the accessory bacteriochlorophylls, B. These spectra demonstrate selective enhancement of the separate bacteriopheophytins on the active (H(L)) and inactive (H(M)) sides of the reaction centers. The spectra are assigned with the aid of normal coordinate analyses using force fields previously developed for porphyrins and reduced porphyrins. Comparison of the H(L) and H(M) vibrational mode frequencies reveals evidence for greater polarization of the acetyl substituent in H(L) than H(M). This polarization is expected to make H(L) easier to reduce, thereby contributing to the directionality of electron transfer from the special pair, P. In addition, the acetyl polarization of H(L) is increased at low temperature (100 K), helping to account for the increase in electron transfer rate. The polarizing field is suggested to arise from the Mg(2+) of the neighboring accessory bacteriochlorophyll, which is 4.9 A from the acetyl O atom. The 100 K spectra show sharpening and intensification of a number of RR bands, suggesting a narrowing of the conformational distribution of chromophores, which is consistent with the reported narrowing of the distribution in electron transfer rates. Excitation at 800 nm produces high-quality RR spectra of the accessory bacteriochlorophylls, and the spectral pattern is unaltered on tuning the excitation to 810 nm in resonance with the upper exciton transition of P. Either the resonance enhancement of P is weak, or the bacteriochlorophyll RR spectra are indistinguishable for P and B.

Electronic absorption, EPR, and resonance raman spectroscopy of CooA, a CO-sensing transcription activator from R. rubrum, reveals a five-coordinate NO-heme.

Electronic absorption, EPR, and resonance Raman spectroscopies revealed that CooA, the CO-sensing transcriptional regulator from Rhodospirillum rubrum, reacts with NO to form a five-coordinate NO-heme. NO must therefore displace both of the heme ligands from six-coordinate, low-spin Fe(II)CooA in forming five-coordinate Fe(II)CooA(NO). CO, in contrast, displaces a single heme ligand from Fe(II)CooA to form six-coordinate Fe(II)CooA(CO). Of a series of common heme-binding ligands, only CO and NO were able to bind to the heme of wild-type CooA; imidazole, azide anion, and cyanide anion had no effect on the heme absorption spectrum. Although NO binds to the heme and displaces the endogenous ligands, NO was not able to induce CooA to bind to its target DNA. The mechanism of CO-dependent activation of CooA is thus more complex than simple displacement of a ligand from the heme iron since NO does not trigger DNA binding. These observations suggest that the CooA heme site discriminates between NO and the biologically relevant signal, CO.

UV resonance Raman spectra reveal a structural basis for diminished proton and CO2 binding to alpha,alpha-cross-linked hemoglobin.

UV resonance Raman difference spectra between ligated and deoxyhemoglobin contain tryptophan and tyrosine signals which arise from quaternary H-bonds in the T state, which are broken in the R state. These H-bonds are unaffected by bis(3,5-dibromosalicyl) fumarate cross-linking at the Lys alpha 99 residues, which prevents dissociation of Hb tetramers to dimers. However, when the pH is lowered from 9.0, or when NaCl is added, intensity is diminished for the tyrosine Y8 and tryptophan W3 bands of cross-linked deoxyHb, but not of native deoxyHb. This effect is attributed to weakening of tertiary H-bonds involving Tyr alpha 140 and Trp alpha 14, when the T state salt bridge between Val alpha 1 and Arg alpha 141 is formed via protonation of the terminal amino group and anion binding. The Tyr alpha 140-Val alpha 93 H-bond connects the Arg alpha 141-bearing H helix with the Lys alpha 99-bearing G helix. Weakening of the H-bond reflects a tension between the fumarate linker and the salt-bridge. This tension inhibits protonation of the Val alpha 1 amino terminus, thus accounting for the diminution of both proton [Bohr effect] and CO2 binding in the T state as a result of cross-linking.

Identification by UV resonance Raman spectroscopy of an imino tautomer of 5-hydroxy-2'-deoxycytidine, a powerful base analog transition mutagen with a much higher unfavored tautomer frequency than that of the natural residue 2'-deoxycytidine.

UV resonance Raman spectroscopy was used to detect and estimate the frequency of the unfavored imino tautomer of the transition mutagen 5-hydroxy-2'-deoxycytidine (HO5dCyt) in its anionic form. In DNA, this 2'-deoxycytidine analog arises from the oxidation of 2'-deoxycytidine and induces C --> T transitions with 10(2) greater frequency than such spontaneous transitions. An imino tautomer marker carbonyl band (approximately 1650 cm-1) is enhanced at approximately 65 degrees C against an otherwise stable spectrum of bands associated with the favored amino tautomer. This band is similarly present in the UV resonance Raman spectra of the imino cytidine analogs N3-methylcytidine at high pH and N4-methoxy-2'-deoxycytidine at pH 7 and displays features attributable to the imino form of C residues and their derivatives. The fact that the imino tautomer of HO5dCyt occurs at a frequency consistent with its high mutagenic enhancement lends strong support to the hypothesis that unfavored base tautomers play important roles in the mispair intermediates of replication leading to substitution mutations.

New light on allostery: dynamic resonance Raman spectroscopy of hemoglobin kempsey.

On the basis of static and time-resolved resonance Raman spectroscopy of HbA and of a mutant, HbK (Dalpha99N), a specific reaction coordinate is proposed for the allosteric transition in human hemoglobin. The heme is held between proximal (F) and distal (E) helices, whose orientation is responsive to forces generated by ligation and deligation. The E and F helices are in turn tethered via H-bonds to the A and H helices. These outer helices follow the E-F motion, thereby repositioning the N- and C-termini, which form the intersubunit salt bridges in the T quaternary structure. When the T state interface is weakened by Asp --> Asn substitution at a quaternary H-bond (HbK), the Fe-His bond is relaxed and becomes responsive to allosteric effectors. The same E-F motion is observed in HbK, but the A-H following motion is delayed, relative to HbA, as is the Asn H-bond formation.

Resonance Raman evidence for a novel charge relay activation mechanism of the CO-dependent heme protein transcription factor CooA.

Resonance Raman spectra of the CO-responsive transcription factor CooA from Rhodospirillum rubrum provides evidence on the nature of heme ligation and its CO activation mechanism. The Fe(III) form gives standard low-spin heme spectrum, while the Fe(II) form is low spin for wild-type (WT) CooA and mixed spin for a CooA variant, H77Y, with an His77Tyr substitution. The Fe(II) porphyrin skeletal mode nu11 is at a value (1541 cm-1) indicative of a neutral donor ligand for the H77Y variant but is at an unusually depressed frequency (1529 cm-1) for the WT protein, indicating a strongly donating ligand. This ligand is proposed to be His77 imidazolate, formed by proton transfer to a nearby acceptor. The WT CO adduct has FeCO and CO stretching frequencies that indicate a neutral trans ligand and negative polarity in the CO binding pocket, while the CO adduct of His77Tyr has both 6- and 5-coordinate signals and a nonpolar CO environment. Photolysis of the WT CO adduct by the Raman laser produced a low-spin product at steady state, indicating fast recombination of the displaced ligand. These data suggest a novel YH- - -His- charge relay mechanism for CooA activation by CO. In this mechanism, His77 is reprotonated upon ligand displacement by CO; CO displaces either His77 or the trans ligand, X. The resulting charge on Y- may induce the protein conformation change required for site-selective DNA binding.

H-bonding maintains the active site of type 1 copper proteins: site-directed mutagenesis of Asn38 in poplar plastocyanin.

Type I Cu proteins maintain a trigonal N2S coordination group (with weak axial ligation) in both oxidation states of the Cu2+/+ ion, thereby reducing the reorganization energy for electron transfer. Requirements for maintaining this coordination group were investigated in poplar plastocyanin (Pcy) by mutation of a conserved element of the type 1 architecture, an asparagine residue (Asn38) adjacent to one of the ligating histidines. The side chain of this asparagine forms an active site clasp via two H-bonds with the residue (Ser85) adjacent to the ligating cysteine (Cys84). In addition, the main chain NH of Asn38 donates an H-bond to the thiolate ligand. We have investigated the importance of these interactions by mutating Asn38 to Gln, Thr, and Leu. The mutant proteins are capable of folding and binding Cu2+, but the blue color fades; the rate of fading increases in the order Gln < Thr < Leu. The color is not restored by ferricyanide, showing that the protein is modified irreversibly, probably by oxidation of Cys84. The more stable mutants N38Q and N38T were characterized spectroscopically. The wild-type properties are slightly perturbed for N38Q, but N38T shows remarkable similarity to another type 1 Cu protein, azurin (Azu) from Pseudomonas aeruginosa. The Cu-S(Cys) bond is longer in Azu than in Pcy, and the NH H-bond to the ligating S atom is shorter. Molecular modeling suggests a similar effect for N38T because the threonine residue shifts toward Ser85 in order to avoid a steric clash and to optimize H-bonding. These results demonstrate that H-bonding adjacent to the type 1 site stabilizes an architecture which both modulates the electronic properties of the Cu, and suppresses side reactions of the cysteine ligand.

Variable forms of soluble guanylyl cyclase: protein-ligand interactions and the issue of activation by carbon monoxide.

Soluble guanylyl cyclase (sGC) is known to be activated by NO binding to the heme moiety; previous studies have shown that CO does not activate sGC to the same extent as NO. Resonance Raman spectroscopy reveals different heme pocket structures for soluble guanylyl cyclase prepared by alternate methods, all of which display activation by NO. In our preparation, and in the expressed protein sGC1, the resting Fe(II) state is mainly 6-coordinate and low-spin, and the CO adduct has vibrational frequencies characteristic of a histidine-heme-CO complex in a hydrophobic environment. In contrast, the protein sGC2 is 5-coordinate, high-spin in the resting state, and the CO adduct has perturbed vibrational frequencies indicative of a negatively polarizing residue in the binding pocket. The differences may result from the need to reconstitute sGC1 or different isolation procedures for sGC1 versus sGC2. However, both sGC1 and sGC2 are activated by the same mechanism, namely displacement of the proximal histidine ligand upon NO binding, and neither one is activated by CO. If CO is an activator in vivo, some additional molecular component is required.

Fourier transform infrared evidence against Asp beta 99 protonation in hemoglobin: nature of the Tyr alpha 42-Asp beta 99 quaternary H-bond.

The Tyr alpha 42-Asp beta 99 intersubunit H-bond stabilizes the T quaternary structure in hemoglobin (Hb) tetramers. We had proposed that Tyr alpha 42 acts as an acceptor in this H-bond, because the tyrosine Y8a/8b and Y7a' UVRR (ultraviolet resonance Raman) bands shift in directions opposite to those expected if tyrosine is an H-bond donor. If Asp beta 99 is the H-bond donor, then it must be protonated in the T state, and would be a previously unrecognized contributor to the Bohr effect. This implication was strengthened by the discovery that an R-minus-T difference FTIR (Fourier transform infrared) band at 1693 cm-1, which might be a signal from protonated carboxylate, is missing in Hb Kempsey, a mutant in which Asp beta 99 is replaced by Asn. However, we now find that this FTIR signal is insensitive to 13C-labeling of the aspartate residues in Hb, and cannot arise from protonated Asp beta 99. There are no other difference signals in the 1700 cm-1 region at a sensitivity of one COOH group. We conclude that Asp beta 99 is not protonated, and that the anomalous UVRR shifts must arise from compensating polarization of the Tyr alpha 42 OH. Candidates for this compensation are the H-bond donated by the Asp beta 94 backbone NH, and the nearby positive charge of Arg beta 40.

Structure changes in hemoglobin upon deletion of C-terminal residues, monitored by resonance Raman spectroscopy.

Loss of C-terminal residues in hemoglobin raises oxygen affinity and reduces both cooperativity and the Bohr effect. These functional changes are expected from the loss of C-terminal salt bridges, which are seen crystallographically to stabilize the T quaternary structure. Ultraviolet resonance Raman (UVRR) difference spectroscopy confirms that the strength of the T state contacts is diminished when the C-terminal and also the penultimate residues are removed chemically. Deoxy minus CO difference signals arising from the Trpbeta37-Aspalpha94 and Tyralpha42-Aspbeta99 H bonds at the alpha1 beta2 subunit interface are diminished, and at pH 9, the difference spectra reveal a shift to the R quaternary structure. These effects are small for desHisbeta146 Hb and large for desArgalpha141 Hb, consistent with the order of functional changes. In addition, the H bond between the A and E helices is strengthened by removal of Argalpha141 and is further strengthened when the effector molecule IHP (inositol hexaphosphate) is added to deoxy-desArgalpha141 Hb or when its pH is lowered to 5.8. This effect is attributed to the loss of the C-terminal anchor of the alpha chain H helix, which supports the F and A helices. The beta chain is not as sensitive because it has extra F-H interhelix H bonds. Removal of both Hisbeta146 and Tauyrbeta145 produce UVRR changes which are intermediate between desHisbeta146 and desArgalpha141 Hb, although the functional consequences are greater than for desArgalpha141 Hb. Removal of Tyralpha140 as well as Argalpha141 abolishes cooperative binding as well as the Bohr effect, and the UVRR difference signals are also lost, suggesting that quaternary constraints are removed in both the T and the R states. When the approximately 220 cm-1 iron-histidine stretching vibration of the deoxy-proteins is examined, using Raman excitation in resonance with the heme Soret band, the frequency is observed to diminish toward that of deoxyHb A (215 cm-1) as the pH is lowered and IHP is added and to increase toward a completely relaxed value (223 cm-1) as the pH is raised to 9. The relaxation is in the same order as the functional perturbations: desHisbeta146 < desArgalpha141 < desHisbeta146-Tyrbeta145 < desArgalpha141-Tyralpha140. However, even desArgalpha141-Tyralpha140 Hb shows significant reduction in the Fe-His frequency as IHP is added at low pH. The Fe-His frequency is sensitive to both tertiary and quaternary structure changes and is a global indicator of forces at the heme. The order of affinity changes can be understood on the basis of the number of stabilizing H bonds between the F and H helices. Titration curves of the Fe-His frequency against pH are not sigmoidal, consistent with a multiplicity of contributions to the Bohr effect.

Alternative modes of substrate distortion in enzyme and antibody catalyzed ferrochelation reactions.

Both an antibody that catalyzes metal insertion into porphyrins and the corresponding enzyme, ferrochelatase, are shown by resonance Raman spectroscopy to induce distortion in the bound porphyrin substrate. It was found that the enzyme-induced distortion is different from that induced by the antibody; the catalytic antibody produces a distortion which is similar to the one present in the hapten, N-methylmesoporphyrin IX (N-MeMP). Activation of specific out-of-plane vibrational modes reveal that the antibody induces an alternating up-and-down tilting of the pyrrole rings, while ferrochelatase induces tilting of all four pyrrole rings in the same direction (doming). Both distortions are effective in catalyzing metal insertion. The distortion induced in the enzyme is only seen when an inhibitory metal ion is also bound. This observation suggests an allosteric mechanism, in which a conformational change which distorts the porphyrin toward the transition state geometry, is induced by metal binding at an adjacent site. In contrast, the antibody does not have a metal binding site and appears to function largely through binding interactions with the porphyrin.

Temperature-dependent ultraviolet resonance Raman spectroscopy of the premelting state of dA.dT DNA.

Poly(dA).poly(dT) and DNA duplex with four or more adenine bases in a row exhibits a broad, solid-state structural premelting transition at about 35 degrees C. The low-temperature structure is correlated with the phenomena of "bent DNA." We have conducted temperature-dependent ultraviolet resonance Raman measurements of the structural transition using poly(dA).poly(dT) at physiological salt conditions, and are able to identify, between the high and low temperature limits, changes in the vibrational frequencies associated with the C4 carbonyl stretching mode in the thymine ring and the N6 scissors mode of the amine in the adenine ring of poly(dA).poly(dT). This work supports the model that the oligo-dA tracts' solid-state structural premelting transition is due to a set of cross-stand bifurcated hydrogen bonds between consecutive dA. dT pairs.

Tyrosine and tryptophan structure markers in hemoglobin ultraviolet resonance Raman spectra: mode assignments via subunit-specific isotope labeling of recombinant protein.

Phenyl-deuterated tyrosine (Tyr-d4) and indole-deuterated tryptophan (Trp-d5) have been selectively incorporated into hemoglobin (Hb) by expressing the gene in auxotrophic strains of Escherichia coli. Ultraviolet resonance Raman (UVRR) spectra, using 229-nm excitation, show that difference features characteristic of the Hb quaternary R --> T transition are not perturbed by the incorporation of the isotopes. All the UVRR bands between 800 and 1700 cm-1 are assigned to either Tyr or Trp except for the 1511 cm-1 band, which had been thought to arise from the Trp 2 x W18 overtone. This band does not shift upon Trp or Tyr labeling but does shift 5 cm-1 in D2O, suggesting assignment to a histidine (His) residue. Its intensification in the T-state is consistent with His protonation. The alpha- and beta-subunits were selectively labeled, by reconstitution of labeled subunits with unlabeled subunits, to make isotope hybrids. Selective Tyr labeling identified the alpha subunits as the locus of the Y8a upshift observed in Hb, supporting the previous inference that this shift is associated with the T-state H-bond involving the interfacial Tyr alpha42 [Rodgers, Su, Subramaniam, & Spiro (1992) J. Am. Chem. Soc. 114, 3697]. Selective Trp labeling showed the Trp alpha14 contributions to the T - R difference spectrum to be negligible and confirmed Trp beta37 as the locus of the W3 difference signal, and probably of the remaining Trp signals as well. The observed downshift of W17 and upshift of Wd5 in the T-state are consistent with a stronger T-state H-bond between Trp beta37 and Asp alpha94; the resulting excitation profile red shift accounts for the dominance of the Trp beta37 contribution to the T - R difference UVRR spectrum.

De novo heme proteins from designed combinatorial libraries.

We previously reported the design of a library of de novo amino acid sequences targeted to fold into four-helix bundles. The design of these sequences was based on a "binary code" strategy, in which the patterning of polar and nonpolar amino acids is specified explicitly, but the exact identities of the side chains is varied extensively (Kamtekar S, Schiffer JM, Xiong H, Babik JM, Hecht MH, 1993, Science 262:1680-1685). Because of this variability, the resulting collection of amino acid sequences may include de novo proteins capable of binding biologically important cofactors. To probe for such binding, the de novo sequences were screened for their ability to bind the heme cofactor. Among an initial collection of 30 binary code sequences, 15 are shown to bind heme and form bright red complexes. Characterization of several of these de novo heme proteins demonstrated that their absorption spectra and resonance Raman spectra resemble those of natural cytochromes. Because the design of these sequences is based on global features of polar/ nonpolar patterning, the finding that half of them bind heme highlights the power of the binary code strategy, and demonstrates that isolating de novo heme proteins does not require explicit design of the cofactor binding site. Because bound heme plays a key role in the functions of many natural proteins, these results suggest that binary code sequences may serve as initial prototypes for the development of large collections of functionally active de novo proteins.