Arginine-140 and isoleucine-141 determine the 17beta-estradiol-binding specificity of the sex-steroid-binding protein (SBP, or SHBG) of human plasma.

Arginine-140 and isoleucine-141 were identified as key determinants of 17beta-estradiol (E(2)) binding affinity of the sex-steroid-binding protein (SBP, or SHBG) of human plasma. Amino acid residues that differ between human and rabbit SBP sequences were replaced in the human protein and the products tested for lowered E(2)binding activity as are seen in the rabbit protein. Only mutants containing either R140K or I141L replacements display an E(2) equilibrium dissociation constant (Kd) higher than the wild type, reaching a value of 30 nM when both were present. The 5alpha-dihydrotestosterone (DHT) equilibrium dissociation constant of these mutants was unaffected. The quadruple mutant M107I/I138V/R140K/I141L yielded an E(2) Kd of 65 nM, significantly closer to the 80 nM rabbit SBP E(2) Kd value. Although mutants containing the M107I and I138V replacements in the absence of R140K and I141L had normal E(2) Kds, the presence of the M107I replacement in the quadruple mutant was necessary to obtain an accurate E(2) Kd value by competitive Scatchard analysis. Molecular modeling using coordinates for the recently determined N-terminal domain of human SBP revealed a significant shift of the F56 phenyl ring away from ring A of E(2) in mutant models containing the R140K and I141L replacements. We conclude that R140 and I141 are required for sustaining the right proximity of the phenyl ring of F56 to ring A of 17beta-estradiol, thus optimizing the E(2)-binding affinity of human SBP.

Localization of the C-terminus of rat glutathione S-transferase A1-1: crystal structure of mutants W21F and W21F/F220Y.

Twelve C-terminal residues of human glutathione S-transferase A1-1 form a helix in the presence of glutathione-conjugate, or substrate alone, and partly cover the active site. According to X-ray structures, the helix is disordered in the absence of glutathione, but it is not known if it is helical and delocalized, or in a random-coil conformation. Mutation to a tyrosine of residue 220 within this helix was previously shown to affect the pK(a) of Tyr-9 at the active site, in the apo form of the enzyme, and it was proposed that an on-face hydrogen bond between Tyr-220 and Tyr-9 provided a means for affecting this pK(a). In the current study, X-ray structures of the W21F and of the C-terminal mutation, W21F/F220Y, with glutathione sulfonate bound, show that the C-terminal helix is disordered (or delocalized) in the W21F crystal but is visible and ordered in a novel location, a crystal packing crevice, in one of three monomers in the W21F/F220Y crystal, and the proposed hydrogen bond is not formed. Fluorescence spectroscopy studies using an engineered F222W mutant show that the C-terminus remains delocalized in the absence of glutathione or when only the glutathione binding site is occupied, but is ordered and localized in the presence of substrate or conjugate, consistent with these and previous crystallographic studies. Proteins 2001;42:192-200.

Molecular architecture of the mutagenic active site of human immunodeficiency virus type 1 reverse transcriptase: roles of the beta 8-alpha E loop in fidelity, processivity, and substrate interactions.

Human immunodeficiency virus type 1 (HIV-1) reverse transcriptase (RT) is a putative source of the genomic hypermutation that promotes rapid evolution of HIV-1. To understand the molecular strategies that create a highly mutagenic DNA polymerase active site in HIV-1 RT, we investigated the roles of four residues in the beta 8-alpha E loop. Gln151, which interacts with the sugar of the incoming dNTP, and Lys154, which interacts with the template, yielded site-directed mutants with increased fidelity, suggesting that these residues are directly involved in the mutagenic architecture of the active site. Mutations at Gln151 and Lys154 also reduced processivity. Q151N RT showed enhanced ability to discriminate between TTP and AZT triphosphate, consistent with the observation that the Q151M mutation confers AZT resistance in vivo. Mutations at Gly152 greatly decreased RT activity; molecular modeling suggests that Gly152 is critical for the proper geometric alignment that permits base-pairing of the incoming dNTP with the template. Mutations at Trp153 reduced the expression level, and presumably the stability, of RT proteins in bacteria. These observations support the conclusion that interactions of active site residues in the beta 8-alpha E loop with incoming dNTPs and the template are determinants of the accuracy, processivity, and substrate selectivity of HIV-1 RT.

Thermus aquaticus DNA polymerase I mutants with altered fidelity. Interacting mutations in the O-helix.

Phe(667) in the conserved O-helix of Thermus aquaticus (Taq) DNA polymerase I (pol I) is known to be important for discrimination against dideoxy-NTPs. We show here that Phe(667) is also important for base selection fidelity. In a forward mutation assay at high polymerase concentration, wild type pol I catalyzed frequent A --> T and G --> T transversions and -1 frameshifts at nonreiterated sites involving loss of a purine immediately downstream of a pyrimidine. The mutants F667L and A661E,I665T,F667L exhibited large decreases in A --> T and G --> T transversions, and the triple mutant displayed reduction in the aforementioned -1 frameshifts as well. Kinetic analysis showed that the F667L and A661E,I665T,F667L polymerases discriminated against synthesis of A:A mispairs more effectively and catalyzed less extension of A:A mispairs than the wild type enzyme. These data indicate that Phe(667) functions in maintaining the error frequency and spectrum, and the catalytic efficiency, of wild type pol I. We also found that the strong general mutator activity conferred by the single A661E substitution was entirely suppressed in the A661E, I665T,F667L polymerase, exemplifying how interactions among O-helix residues can contribute to fidelity. We discuss the mutator and anti-mutator mutations in light of recently obtained three-dimensional structures of T. aquaticus pol I.

Identification and functional characterization of human soluble epoxide hydrolase genetic polymorphisms.

Human soluble epoxide hydrolase (sEH), an enzyme directing the functional disposition of a variety of endogenous and xenobiotic-derived chemical epoxides, was characterized at the genomic level for interindividual variation capable of impacting function. RNA was isolated from 25 human liver samples and used to generate full-length copies of soluble epoxide hydrolase cDNA. The resulting cDNAs were polymerase chain reaction amplified, sequenced, and eight variant loci were identified. The coding region contained five silent single nucleotide polymorphisms (SNPs) and two variant loci resulting in altered protein sequence. An amino acid substitution was identified at residue 287 in exon 8, where the more common arginine was replaced by glutamine. A second variant locus was identified in exon 13 where an arginine residue was inserted following serine 402 resulting in the sequence, arginine 403-404, instead of the more common, arginine 403. This amino acid insertion was confirmed by analyzing genomic DNA from individuals harboring the polymorphic allele. Slot blot hybridization analyses of the liver samples indicated that sEH mRNA steady-state expression varied approximately 10-fold. Transient transfection experiments with CHO and COS-7 cells were used to demonstrate that the two new alleles possess catalytic activity using trans-stilbene oxide as a model substrate. Although the activity of the glutamine 287 variant was similar to the sEH wild type allele, proteins containing the arginine insertion exhibited strikingly lower activity. Allelic forms of human sEH, with markedly different enzymatic profiles, may have important physiological implications with respect to the disposition of epoxides formed from the oxidation of fatty acids, such as arachidonic acid-derived intermediates, as well in the regulation of toxicity due to xenobiotic epoxide exposures.

Tolerance of 5-fluorodeoxyuridine resistant human thymidylate synthases to alterations in active site residues.

Fluoropyrimidines, such as 5-fluorouracil (5-FU), are used extensively in cancer therapy. In the cell, 5-FU is metabolized to 5-fluorodeoxyuridylate (5-FdUMP), a tight binding covalent inhibitor of thymidylate synthase (TS). In order to create 5-FdUMP resistant enzymes to protect chemosensitive normal cells and further understand mechanisms of 5-FdUMP resistance, we have randomized four residues within the active site of TS. Our previous studies identified alterations in residues which produce active TS with enhanced resistance to 5-fluorouridine (5-FdUR). By remutagenizing a subset of the 13 previously targeted residues (A197, L198, C199 and V204), an unbiased random library can be created allowing for extensive testing of all possible amino acid substitutions at each of the sites. Using genetic complementation and selection in Escherichia coli, we identified the spectrum of substitutions that yield active TS as well as those that resulted in 5-FdUR resistant mutants of TS. The 5-FdUR resistant TS were found to share several structural features including hydrophobic substitutions at residue 197, retention of the wild-type leucine 198, the alteration C199L (present in 64% of the drug-resistant library), and polar alterations of valine 204. The catalytic activity of mutants with these features was approximately equal to that of the wild-type TS.

Enhancement of tumor ablation by a selected HSV-1 thymidine kinase mutant.

With the advent of gene therapy, herpes simplex virus type I (HSV-1) thymidine kinase (TK) has garnered much interest as a suicide gene for cancer ablation. As a means to improve the overall efficacy of the prodrug-gene activation approach, as well as to reduce ganciclovir-mediated toxicity, a large library of mutant thymidine kinases was generated and screened for the ability to enhance in vitro cell sensitivity to the prodrugs, ganciclovir (GCV) and acyclovir (ACV). Enzyme kinetics of one thymidine kinase mutant from this library that contains six amino acid substitutions at or near the active site reveals a distinct mechanism for providing enhanced prodrug-mediated killing in mammalian cells. In in vitro rat C6 cell prodrug sensitivity assays the TK mutant (mutant 30) achieves nanomolar IC50 values with GCV and ACV, in contrast to IC50values of 30 microM and >100 microM, respectively, for wild-type TK. In a mouse xenograft tumor model, growth of mutant 30 expressing tumors is restricted by ganciclovir at a dose at least 10- fold lower than one that impedes growth of wild-type TK-expressing tumors. Furthermore, in the presence of GCV a substantial bystander effect is observable when only 20% of the tumor cells express mutant 30 whereas no restriction in tumor growth is seen in tumors bearing the wild-type TK under the same conditions. The enhanced sensitization to prodrugs conferred by mutant 30 is apparently due to a 35-fold increase in thymidine Km which results in reduced competition between prodrug and thymidine at the active site. This provides mutant 30 a substantial kinetic advantage despite very high Kms for both ganciclovir and acyclovir. Molecular modeling of the mutations within the active site suggests that a tyrosine substitution at alanine 168 (A168) alters thymidine and prodrug interactions by causing catalytically important residues to move. The use of mutant 30 in place of the wild-type TK should provide a more effective gene therapy of cancer.

New human immunodeficiency virus, type 1 reverse transcriptase (HIV-1 RT) mutants with increased fidelity of DNA synthesis. Accuracy, template binding, and processivity.

Infidelity of DNA synthesis by human immunodeficiency virus, type 1 reverse transcriptase (HIV-1 RT) is a presumptive determinant of HIV-1 hypervariability and is incompletely understood at the mechanistic and structural levels. Amino acid substitution at only three residues, including Asp-76 (Kim, B., Hathaway, T. R., and Loeb, L. A. (1996) Biochemistry 37, 5831-5839), is known to increase fidelity. We report here that substitution at Arg-78 can also increase accuracy. Mutant R78A RT showed reduced primer extension in misincorporation assays lacking a complementary dNTP and exhibited a 9-fold decrease in mutation frequency in the M13mp2 lacZ forward mutation assay. Previous structural studies indicate that Arg-78 and Asp-76 lie in a region that interacts with template nucleotides. Interestingly, R78A RT exhibited 6- to 8-fold decreases in binding affinity (K(d)) for RNA and DNA templates relative to wild type RT. In contrast, D76V RT, which also increases fidelity (Kim et al., 1996), showed a 6- to 7-fold increased affinity. The processivity of R78A RT on both RNA and DNA templates was substantially reduced relative to wild type RT, whereas the processivity of D76V RT was increased. We discuss relationships of fidelity, template binding, and processivity in these and other HIV RT mutants.

Site-directed mutations within the core "alpha-crystallin" domain of the small heat-shock protein, human alphaB-crystallin, decrease molecular chaperone functions.

Site-directed mutagenesis was used to evaluate the effects on structure and function of selected substitutions within and N-terminal to the core "alpha-crystallin" domain of the small heat-shock protein (sHsp) and molecular chaperone, human alphaB-crystallin. Five alphaB-crystallin mutants containing single amino acid substitutions within the core alpha-crystallin domain displayed a modest decrease in chaperone activity in aggregation assays in vitro and in protecting cell viability of E. coli at 50 degrees C in vivo. In contrast, seven alphaB-crystallin mutants containing substitutions N-terminal to the core alpha-crystallin domain generally resembled wild-type alphaB-crystallin in chaperone activity in vitro and in vivo. Size-exclusion chromatography, ultraviolet circular dichroism spectroscopy and limited proteolysis were used to evaluate potential structural changes in the 12 alphaB-crystallin mutants. The secondary, tertiary and quaternary structures of mutants within and N-terminal to the core alpha-crystallin domain were similar to wild-type alphaB-crystallin. SDS-PAGE patterns of chymotryptic digestion were also similar in the mutant and wild-type proteins, indicating that the mutations did not introduce structural modifications that altered the exposure of proteolytic cleavage sites in alphaB-crystallin. On the basis of the similarities between the sequences of human alphaB-crystallin and the sHsp Mj HSP16.5, the only sHsp for which there exists high resolution structural information, a three-dimensional model for alphaB-crystallin was constructed. The mutations at sites within the core alpha-crystallin domain of alphaB-crystallin identify regions that may be important for the molecular chaperone functions of sHsps.

Identification of a guanylyl cyclase-activating protein-binding site within the catalytic domain of retinal guanylyl cyclase 1.

Regulation of cAMP and cGMP production is a fundamental step in a broad range of signal transduction systems, including phototransduction. To identify regions within photoreceptor guanylyl cyclase 1 (GC1) that interact with GC-activating proteins (GCAPs), we synthesized the intracellular fragment of GC1, residues 491-1110, as a set of 15 amino acid long, partially overlapping peptides on the surface of individual pins arranged in a microtiter plate format. This pin assay identified 8 peptides derived from different regions of the GC1 intracellular domain that bind GCAPs. Peptide variants containing these sequences were synthesized as free peptides and tested for their ability to inhibit GC1 stimulation by GCAPs. A free peptide,968GTFRMRHMPEVPVRIRIG, from the catalytic domain of GC1 was the strongest inhibitor of GCAP1/GCAP2-mediated activation. In native GC1, this polypeptide fragment is likely to form a loop between alpha-helix 3 and beta-strand 4. When this region in GC1 was replaced by the corresponding sequence of GCAP-insensitive GC type A, GCAPs did not stimulate the GC1 mutant. The corresponding loops in related adenylyl cyclase (AC) are involved in the activating and inhibiting interactions with Gs alpha and Gi alpha, respectively. Thus, despite interacting with different activating proteins, both AC and GC activity may be modulated through their respective regions within catalytic domains.

1.8 A crystal structure of the major NAD(P)H:FMN oxidoreductase of a bioluminescent bacterium, Vibrio fischeri: overall structure, cofactor and substrate-analog binding, and comparison with related flavoproteins.

We have solved the crystal structure of FRase I, the major NAD(P)H:FMN oxidoreductase of Vibrio fischeri, by the multiple isomorphous replacement method (MIR) at 1.8 A resolution with the conventional R factor of 0.187. The crystal structure of FRase I complexed with its competitive inhibitor, dicoumarol, has also been solved at 2.2 A resolution with the conventional R factor of 0.161. FRase I is a homodimer, having one FMN cofactor per subunit, which is situated at the interface of two subunits. The overall fold can be divided into two domains; 80% of the residues form a rigid core and the remaining, a small flexible domain. The overall core folding is similar to those of an NADPH-dependent flavin reductase of Vibrio harveyi (FRP) and the NADH oxidase of Thermus thermophilus (NOX) in spite of the very low identity in amino acid sequences (10% with FRP and 21% with NOX). 56% of alpha-carbons of FRase I core residues could be superposed onto NOX counterparts with an r.m.s. distance of 1.2 A. The remaining residues have relatively high B-values and may be essential for defining the substrate specificity. Indeed, one of them, Phe124, was found to participate in the binding of dicoumarol through stacking to one of the rings of dicoumarol. Upon binding of dicoumarol, most of the exposed re-face of the FMN cofactor is buried, which is consistent with the ping pong bi bi catalytic mechanism.

Structure of nitrite bound to copper-containing nitrite reductase from Alcaligenes faecalis. Mechanistic implications.

The structures of oxidized, reduced, nitrite-soaked oxidized and nitrite-soaked reduced nitrite reductase from Alcaligenes faecalis have been determined at 1.8-2.0 A resolution using data collected at -160 degrees C. The active site at cryogenic temperature, as at room temperature, contains a tetrahedral type II copper site liganded by three histidines and a water molecule. The solvent site is empty when crystals are reduced with ascorbate. A fully occupied oxygen-coordinate nitrite occupies the solvent site in crystals soaked in nitrite. Ascorbate-reduced crystals soaked in a glycerol-methanol solution and nitrite at -40 degrees C remain colorless at -160 degrees C but turn amber-brown when warmed, suggesting that NO is released. Nitrite is found at one-half occupancy. Five new solvent sites in the oxidized nitrite bound form exhibit defined but different occupancies in the other three forms. These results support a previously proposed mechanism by which nitrite is bound primarily by a single oxygen atom that is protonable, and after reduction and cleavage of that N-O bond, NO is released leaving the oxygen atom bound to the Cu site as hydroxide or water.

Site-directed mutants of pseudoazurin: explanation of increased redox potentials from X-ray structures and from calculation of redox potential differences.

In order to understand the origins of differences in redox potentials among cupredoxins (small blue type I copper-containing proteins that reversibly change oxidation state and interact with redox partners), we have determined the structures of the native and two mutants (P80A and P80I) of pseudoazurin from Alcaligenes faecalis S-6 in oxidized and reduced forms at resolutions of 2.2 A in the worst case and 1.6 A in the best case. The P80A mutation creates a surface pocket filled by a new water molecule, whereas the P80I mutant excludes this water. Distinct patterns of change occur in response to reduction for all three molecules: the copper position shifts, Met 7 and Pro 35 move, and the relative orientations of residues 81 to 16, 18 to the amide planes of 77 and 86, all change. Systematic changes in the weak electrostatic interactions seen in the structures of different oxidation states can explain the Met 7/Pro 35 structural differences as well as some fluctuating solvent positions. Overall displacement parameters increase reversibly upon reduction. The reduced forms are slightly expanded over the oxidized forms. The geometries of the mutants become more trigonal in their reduced forms, consistent with higher redox potentials (+409 mV for P80A and +450 mV for P80I). Calculations of the differences in redox potentials, using POLARIS, reveal that a water unique to the P80A mutant is required (with correctly oriented hydrogens) to approximate the observed difference in redox potential. The POLARIS calculations suggest that the reduced forms are additionally stabilized through changes in the solvation of the copper center, specifically via the amides of residues 16, 39, 41, 79, and 80 which interact with either Phe 18, Met 86, or Cys 78. The redox potential of P80A is increased largely due to solvation effects, whereas the redox potential of P80I is increased largely due to geometrical effects.

Structural comparison of cupredoxin domains: domain recycling to construct proteins with novel functions.

The three-dimensional structures of the copper-containing enzymes ascorbate oxidase, ceruloplasmin, and nitrite reductase, comprised of multiple domains with a cupredoxin fold, are consistent with having evolved from a common ancestor. The presence or absence of copper sites has complicated ascertaining the structural and evolutionary relationship among these and related proteins. Simultaneous structural superposition of the enzyme domains and their known cupredoxin relatives shows clearly that there are at least six cupredoxin classes, and that the evolution of the conserved core of these domains is independent of the presence or absence of copper sites. Relationships among the variable loops in these structures show that the two-domain ancestor of the blue oxidases contained a trinuclear-copper interface but could not have functioned in a monomeric state. Comparison of the sequence of the copper-containing, iron-regulating protein. Ferrous transport (Fet3) from yeast to the structurally defined core and loop residues of the cupredoxins suggests specific residues that could be involved in the ferroxidase activity of Fet3.

Displacement-parameter weighted coordinate comparison: I. Detection of significant structural differences between oxidation States.

A method, displacement-parameter weighted coordinate comparison, for comparing closely related structures is developed. A 'probability of similarity' is derived from the overlap between the most probable volumes occupied by two analogous atoms in a comparison. The distribution of distances between atom pairs (difference distances), where each distance is weighted by the probability of similarity, is examined. The subset of atom pairs with normally distributed difference distances is used to estimate errors in the difference distances and to calculate the probability that the difference is significant for each atom-pair. These probabilities of difference are shown to correlate with features observed in difference maps for diffraction data from oxidized and reduced forms of the cupredoxin, pseudoazurin. The pair-wise probability of difference also is shown to be correlated with regions in previously published plastocyanin structures which differ upon copper oxidation state change, and which differ in a similar manner when temperature is changed, or pH is changed.

Site-directed mutagenesis of azurin from Pseudomonas aeruginosa enhances the formation of an electron-transfer complex with a copper-containing nitrite reductase from Alcaligenes faecalis S-6.

Kinetic analysis of electron transfer between azurin from Pseudomonas aeruginosa and copper-containing nitrite reductase (NIR) from Akaligenes faecalis S-6 was carried out to investigate the specificity of electron transfer between copper-containing proteins. Apparent values of kcat and Km of NIR for azurin were 300-fold smaller and 172-fold larger than those for the physiological redox partner, pseudoazurin from A. faecalis S-6, respectively, suggesting that the electron transfer between azurin and NIR was less specific than that between pseudoazurin and NIR. One of the major differences in 3-D structure between these redox proteins, azurin and pseudoazurin, is the absence and presence of lysine residues near their type 1 copper sites, respectively. Three mutated azurins, D11K, P36K, and D11K/P36K, were constructed to evaluate the importance of lysine residues in the interaction with NIR. The redox potentials of D11K, P36K, and D11K/P36K azurins were higher than that of wild-type azurin by 48, 7, and 55 mV, respectively. As suggested by the increase in the redox potential, kinetic analysis of electron transfer revealed reduced ability of electron transfer in the mutated azurins. On the other hand, although each of the single mutations caused modest effects on the decrease in the Km value, the simultaneous mutations of D11K and P36K caused significant decrease in the Km value when compared to that for wild-type azurin. These results suggest that the introduction of two lysine residues into azurin facilitated docking to NIR.

Application of known X-ray phases in the crystallographic study of a small protein.

Phase information, assumed known from three-beam diffraction experiments, has been used successfully as input to direct methods to re-determine the crystal structure of rubredoxin. With data at 1.54 A resolution, starting sets containing 26 single phases, or alternatively 45 triplet phases, in both cases known with a mean error of +/-22.5 degrees, were sufficient to initiate solution of the structure. Conventional figures of merit were employed in the early stages to reject the majority of the incorrect phase models. The presence of a FeS(4) cluster in the structure was used in the interpretation of the initial maps. Phase sets including 2500 E's with a mean single phase error approximately 70 degrees or a mean triplet phase error approximately 80 degrees, both relative to the model from the crystallographic refinement, could be refined and expanded by Fourier recycling using the SAYTAN formalism. Several parameters have been varied to study their influence on phase expansion and refinement.

Studies on protein-protein interaction between copper-containing nitrite reductase and pseudoazurin from Alcaligenes faecalis S-6.

Site-directed mutagenesis of a copper-containing nitrite reductase (NIR) from Alcaligenes faecalis S-6 was carried out to identify the amino acid residues involved in interaction with its redox partner, pseudoazurin, in which four positively charged residues were previously shown to be important in the interaction. Ten negatively charged residues located on the surface of NIR were replaced independently by alanine or serine. All the altered NIRs showed CD spectra and optical spectra identical to those of wild-type NIR, suggesting that all the replacements caused no gross change in the overall structure or in the environment of type 1 copper site. Kinetic analysis of electron transfer between pseudoazurin and altered NIRs revealed that the replacement of Glu-118, Glu-197, Asp-201, Glu-204, or Asp-205 by Ala caused a significant increase in the Km value for pseudoazurin compared with that of wild-type NIR. Furthermore, the simultaneous replacement of three of these residues (Glu-118, Glu-197, and Asp-201) caused a further increase in the Km value. These results suggested that the negatively charged residues are involved in electrostatic interaction with pseudoazurin. Kinetic analyses of the altered NIRs (E118A, E197A, or D201A) with altered pseudoazurins (K10A, K57A, or K77A) implicate specific pairs of the charged residues that are involved in electrostatic interaction between NIR and pseudoazurin.

A taste of copper.

New structures of copper containing proteins with cupredoxin-like folds confirm earlier predictions, and reveal electron-transfer routes in cytochrome oxidase, while a new fold for amine-oxidase reveals a new use for copper in forming self-derived quino-cofactor.

The structure of copper-nitrite reductase from Achromobacter cycloclastes at five pH values, with NO2- bound and with type II copper depleted.

High resolution x-ray crystallographic structures of nitrite reductase from Achromobacter cycloclastes, undertaken in order to understand the pH optimum of the reaction with nitrite, show that at pH 5.0, 5.4, 6.0, 6.2, and 6.8, no significant changes occur, other than in the occupancy of the type II copper at the active site. An extensive network of hydrogen bonds, both within and between subunits of the trimer, maintains the rigidity of the protein structure. A water occupies a site approximately 1.5 A from the site of the type II copper in the structure of the type II copper-depleted structure (at pH 5.4), again with no other significant changes in structure. In nitrite-soaked crystals, nitrite binds via its oxygens to the type II copper and replaces the water normally bound to the type II copper. The active-site cavity of the protein is distinctly hydrophobic on one side and hydrophilic on the other, providing a possible path for diffusion of the product NO. Asp-98 exhibits thermal parameter values higher than its surroundings, suggesting a role in shuttling the two protons necessary for the overall reaction. The strong structural homology with cupredoxins is described.