S-Nitrosylation-induced Conformational Change in Blackfin Tuna Myoglobin*

S-Nitrosylation is a post-translational protein modification that can alter the function of a variety of proteins. Despite the growing wealth of information that this modification may have important functional consequences, little is known about the structure of the moiety or its effect on protein tertiary structure. Here we report high-resolution x-ray crystal structures of S-nitrosylated and unmodified blackfin tuna myoglobin, which demonstrate that in vitro S-nitrosylation of this protein at the surface-exposed Cys-10 directly causes a reversible conformational change by “wedging” apart a helix and loop. Furthermore, we have demonstrated in solution and in a single crystal that reduction of the S-nitrosylated myoglobin with dithionite results in NO cleavage from the sulfur of Cys-10 and rebinding to the reduced heme iron, showing the reversibility of both the modification and the conformational changes. Finally, we report the 0.95-Å structure of ferrous nitrosyl myoglobin, which provides an accurate structural view of the NO coordination geometry in the context of a globin heme pocket.

Protein S-nitrosylation, or the formation of an S-NO bond involving the sulfur of a cysteine residue, is an important posttranslational modification. Numerous proteins whose functions are altered by S-nitrosylation have been identified, including enzymes, transcription factors, receptors, channels, and structural proteins (reviewed in Ref. 1). Parallels have been drawn between S-nitrosylation and O-phosphorylation, a ubiquitous biological signal (2,3). Regulation of the function of many proteins by O-phosphorylation is a consequence of structural changes that take place in the protein following phosphorylation. The molecular consequences of protein S-nitrosylation are less well characterized. Enzymes such as caspases that require a cysteine thiol in the active site can be modulated in their activity by S-nitrosylation, which alters the reactive prop-erties of the cysteine (4). S-Nitrosylation can also influence protein-protein interactions, as recently demonstrated with yeast two-hybrid screening (5). Finally, S-nitrosylation may have the ability to modulate the function of a protein via allosteric changes in its structure. Although this mechanism has been hypothesized for several systems, there has thus far been little structural evidence published (6,7).
Another critical role for NO in cell physiology involves its interaction with heme proteins. The first well defined mechanism for a NO-dependent physiological effect was the relaxation of vascular smooth muscle following the binding of NO to the heme iron of guanylate cyclase, which activates this enzyme to convert guanosine triphosphate to the second messenger cGMP (8). The reaction of the NO group (NO, NO ϩ , NO Ϫ ) with heme and thiol is complex (9). However, under various in vitro and in vivo conditions, the NO group is capable of binding at one or the other or exchanging between the two. It has been shown that the bioavailability of NO when bound to Cys␤93 of hemoglobin is dependent upon blood O 2 concentration as well as the allosteric and redox state and that it acts as an important regulator of vessel tone and blood flow (10 -13).
Sperm whale myoglobin (Mb), 2 the first myoglobin whose x-ray structure was determined, lacks a cysteinyl residue. Human myoglobin has a single reactive cysteine at position 110 and is known to form an S-nitrosylated species in vitro and in vascular smooth muscle cell culture when exposed to physiological concentrations of a nitric oxide donor (14,15). Myoglobins with reactive cysteines are seen scattered throughout the vertebrate species, but there seems to be little to suggest a functional pattern. In fishes, the group that includes tunas frequently possesses a single cysteine within the N-terminal helix. It has been suggested that this reactive -SH plays a role in S-nitrosylation in vivo (16), although there are no hard data yet to support this hypothesis.
With the recent advancement in methodologies for the detection and identification of S-nitrosylated proteins in cell culture and tissue samples, such as the S-nitrosylation site identification (17,18) extension of the original biotin switch method (19), an extensive data base of S-nitrosylation sites within proteins is being accumulated. In light of these advancements, a more detailed knowledge of the structural consequences of this modification will be critical for understanding the molecular mechanisms by which S-nitrosylation alters protein function.
In this work, we have demonstrated with atomic resolution the effects of S-nitrosylation on the x-ray crystal structure of blackfin tuna myoglobin. In addition, we have identified conformational changes in this protein that are the direct result of the formation of the S-nitrosothiol group. We also demonstrate that, upon reduction with dithionite, the NO group is transferred from the surface-exposed Cys-10 to the reduced heme iron, showing the reversibility of the S-NO modification and the conformational changes. Finally, we have provided the first atomic resolution structure of nitric oxide bound to the heme iron of a globin protein, which allows the calculation of accurate geometric parameters for this important coordination complex.

EXPERIMENTAL PROCEDURES
Purification of Myoglobin from Blackfin Tuna-Myoglobin was purified from the red muscle of blackfin tuna (Thunnus atlanticus) caught near Rincón, Puerto Rico. The red muscle was dissected, homogenized, and centrifuged, and myoglobin was precipitated from the supernatant between 65 and 95% ammonium sulfate. Following resuspension of the precipitated myoglobin and dialysis against phosphate-buffered saline, the protein was separated by size exclusion chromatography on a Sephadex G-75 column. Myoglobin-containing fractions were concentrated to ϳ1.2 mM and stored at Ϫ80°C. Myoglobin isolated by this method was nearly 100% ferric aquo (met)myoglobin as judged by UV-visible spectroscopy.
Trans-S-nitrosylation of Myoglobin-Purified myoglobin was S-nitrosylated by reaction of 300 M myoglobin with 3 mM S-nitrosocysteine in borax buffer containing 500 M diethylenetriamine pentaacetate at pH 9. S-Nitrosocysteine was removed by passing the sample through two sequential Micro Bio-Spin 6 size exclusion columns (Bio-Rad). The amount of S-nitrosylation in purified myoglobin samples was quantitated using the Saville (20) and Griess (21) reactions in acidified aqueous solution as described previously (22)(23)(24), which results in the production of an azo dye product that can be quantitated spectrophotometrically using an extinction coefficient of 50,000 M Ϫ1 cm Ϫ1 at 545 nm. A calibration curve was generated for this assay using S-nitrosoglutathione, which was quantitated using an extinction coefficient of 767 M Ϫ1 cm Ϫ1 at 337 nm (25). A chemiluminescence nitric oxide analyzer (Sievers) was also used to verify production of S-nitrosylated myoglobin, again using S-nitrosoglutathione or S-nitrosocysteine for calibration.
Crystallization and Data Collection-Purified myoglobin, at a concentration of 1.2 mM in phosphate-buffered saline, was crystallized by hanging-drop vapor diffusion at room temperature by mixing 2 l of protein solution with 2 l of a precipitant solution containing 0.2 M potassium/sodium tartrate, 0.1 M sodium citrate (pH 5.6), and 2.0 M ammonium sulfate. Crystals grew as clusters of large plates in 2-7 days. Crystals were cryoprotected by soaking for 30 s in the precipitant solution containing either 30% ethylene glycol or 30% glycerol and rapidly cooled to 100 K in a gaseous nitrogen stream for data collection. Crystals of Mb-SNO were grown and cryoprotected in the same way as unmodified myoglobin.
Crystals of ferrous nitrosyl myoglobin were prepared by soaking of metmyoglobin crystals in deoxygenated precipitant solution containing 2.2 M ammonium sulfate, 0.2 M potassium/ sodium tartrate, 0.1 M sodium citrate, pH 5.6, and 20 mM sodium dithionite for 20 min followed by anaerobic transfer to precipitant solution supplemented with 3 mM nitric oxide donor 2-(N,N-diethylamino)-diazenolate-2-oxide (kindly supplied by Dr. Katrina Miranda). After 10 min, the crystals were quickly transferred to 3.5 M ammonium sulfate, 0.16 M potassium/sodium tartrate, 0.1 M sodium citrate, pH 5.6, and flash-frozen in liquid N 2 .
Data were collected either on a rotating copper anode home x-ray source equipped with a Saturn 92 CCD detector (Rigaku/ Molecular Structure Corporation) or at the Advanced Photon Source synchrotron beamlines indicated in Table 1. Diffraction data were integrated and scaled using either d*TREK (26) from within the CrystalClear software suite (Rigaku/Molecular Structure Corporation) or HKL2000 (27). Data collection statistics are provided in Table 1.
Structure Solution, Modeling, and Refinement-The blackfin tuna myoglobin structure was solved by molecular replacement using the program Phaser (28) with yellowfin tuna myoglobin as a search model (Protein Data Bank (PDB) code 1MYT) (29). The structure was rebuilt and refined using Coot (30) and Ref-mac5 programs within the CCP4 suite (31), respectively. Although the amino acid sequence of blackfin tuna myoglobin is not available in any data bases, the sequence was tentatively assigned based on the excellent quality electron density maps and appears to be identical to bluefin and bigeye tuna myoglobins, except for a His 3 Gln substitution at position 112, which is also present in other tuna species. Positive difference electron density at the N terminus was modeled as an N-acetyl group, given that this modification is known to occur in myoglobins from other tuna species (32). Electron density at position 10 in crystals of myoglobin trans-S-nitrosylated with S-nitrosocysteine was interpreted as two conformations of Cys S-NO ( Fig.  1, A and B) and one conformation of unmodified Cys-10 ( Fig.  1C). Geometric restraints for the refinement of the Cys-10 S-NO group were generated based on a small molecule crystal structure of S-nitroso-L-cysteine ethyl ester (33). For the ferrous nitrosyl myoglobin structure, no restraints were placed on the Fe-NO bond lengths or angles during the final stages of refinement. During the later stages of refinement of the atomic resolution Mb structures, anisotropic B-factors were refined for all atoms, and hydrogens were added in their riding positions. Refinement statistics are provided in Table 1.

Structure of Blackfin Tuna
Myoglobin-The structure of myoglobin isolated from blackfin tuna was determined to 0.91 Å resolution by molecular replacement. The overall structure is very similar to that of myoglobins from other species, especially yellowfin tuna (29), and will not be discussed further here. Detailed descriptions of various heme oxidation and ligation states of blackfin tuna myoglobin from high-resolution crystal structures will be published separately.
Generation and Stability of S-Nitrosomyoglobin-Blackfin tuna myoglobin was successfully S-nitrosylated at Cys-10 via reaction with excess S-nitrosocysteine or S-nitrosoglutathione. Although the characteristic absorption band indicative of S-nitrosothiols at ϳ330 nm (extinction coefficient ϳ600 -800 M Ϫ1 cm Ϫ1 ) was masked by the much stronger heme absorption at that wavelength (extinction coefficient ϳ20,000 M Ϫ1 cm Ϫ1 ), the presence of S-nitrosomyoglobin was confirmed using the Saville-Griess assay as well as chemiluminescence after exten-sive desalting of the sample to remove S-nitrosocysteine and free cysteine. Not all of the myoglobin Cys-10 was S-nitrosylated under these reaction conditions, with an average yield of 50 -70% Mb S-NO. The amount of S-nitrosothiol in the S-nitrosomyoglobin sample did not decrease significantly over the course of a month at room temperature when stored in the dark in a metal ion-free buffer.
Structure of S-Nitrosomyoglobin-Myoglobin trans-S-nitrosylated by reaction with excess S-nitrosocysteine crystallized in the same form as unmodified myoglobin. Data were collected to 1.09 Å resolution using a rotating copper anode x-ray generator. Attempts to collect data to higher resolution using a synchrotron source led to complete removal of the S-NO group. The x-ray sensitivity of protein S-NO groups has been observed in other systems as well (34). Slight loss of the S-NO group was also observed using a rotating anode source when long exposure times were used; therefore, exposure times were minimized to prevent degradation of the S-NO. The electron density map around position 10 clearly indicated a chemical modification of the sulfur of Cys-10 as well as multiple conformations of the Cys-10 side chain (Fig. 1). The final model included three conformations of Cys-10, two that had been S-nitrosylated (Fig. 1, A and is the ith measured diffraction intensity and ͗I (hkl) ͘ is the mean of the intensity for the miller index (hkl).
B), and one that remained unmodified (Fig. 1C). This interpretation gave the best fit to the electron density map and was only possible because of the high resolution of the data. The unmodified Cys-10 conformation closely matched that seen in the unmodified myoglobin structure, and was present at ϳ30% occupancy. The ϳ70% S-nitrosylation of Cys-10 observed in the crystal matched our synthetic yields of Mb-SNO formation in solution. The major conformation of S-nitrosylated Cys-10 ( Fig. 1A) is rotated by only 25°around 1 (rotation around the C␣-C␤ bond) from the unmodified rotamer (Fig. 1). The second modified conformation (Fig. 1B) has a substantially different 1 rotamer, rotated by 115°, but the attached NO moiety occupied nearly the same space as that of conformation A. Both Cys-10 SNO conformations are present as the cis dihedral, with C-S-N-O dihedral angles (3) close to 0°.
Structural Changes upon S-Nitrosylation-Superposition of the structure of S-nitrosylated myoglobin with the unmodified structure reveals that there are significant differences in several portions of the structure. The differences are primarily localized to the structural elements surrounding Cys-10, including helix A, the N terminus of helix H, loop EF, and loop GH ( Fig.  2A). Comparison of the C␣ traces of S-nitrosylated and unmodified myoglobin yielded a root mean square deviation value of 1.2 Å for portions of the structure surrounding Cys-10 (residues 1-10, 71-83, and 115-127, 34 C␣ atoms), whereas the rest of the structure had a root mean square deviation of only 0.2 Å (residues 11-70, 84 -114, and 128 -146, 109 C␣ atoms).
The NO group of S-nitrosylated Cys-10 packs between the side chain of Leu-117 within loop GH and the backbone carbonyl oxygen of Ala-6 from helix A (Fig. 2B). Superposition of the unmodified myoglobin structure revealed that the S-nitrosothiol group could create steric clashes with Leu-117 and Ala-6 if these portions of the structure did not move. Although two conformations of S-nitrosylated Cys-10 exist in our structure, these observations hold true for both. Therefore, helix A and loop GH were forced in opposite directions, or wedged apart, as a direct consequence of S-nitrosylation at Cys-10. We also observed secondary effects of the movement of helix A. The N terminus of helix H, which was packed against helix A, shifted to occupy the void created by the movement of helix A. Additionally, loop EF was shifted away from helix A. Although loop EF exhibited both static and dynamic disorder in both structures and therefore appeared to be flexible, at least one static conformation of this loop in the unmodified structure was displaced by steric clashes created upon movement of helix A following S-nitrosylation. In contrast to the structural changes surrounding Cys-10 following S-nitrosylation, no significant changes were observed in the vicinity of the heme prosthetic group, the site of oxygen binding in myoglobin.
NO Cleavage from Cys-10 and Rebinding to Reduced Heme-Following trans-S-nitrosylation with S-nitrosocysteine and desalting, a UV-visible absorption spectrum showed that the heme group of isolated blackfin tuna myoglobin remains in the ferric aquo (met) state (Fig. 3A, black spectrum). The addition FIGURE 2. Conformational change in myoglobin upon S-nitrosylation. A, stereo view of the superimposed structures of unmodified myoglobin (yellow) and also following trans-S-nitrosylation (magenta). Black arrows depict the direction of movement of various portions of the structure upon S-nitrosylation at Cys-10. Helices and loops discussed in the text are labeled. The two modeled conformations of loop EF are shown for each structure. B, stereo view a close-up of structural elements surrounding the S-nitrosocysteine that undergoes a conformational change. The S-nitrosocysteine at Cys-10, the side chain of Leu-117, and the backbone carbonyl of Ala-6 are shown as sticks with the same color scheme as A. Semitransparent van der Waals spheres demonstrate the steric clash that would occur between the S-nitrosocysteine group and Leu-117 and Ala-6 if no conformational change occurred. These clashes are also represented as dashed red lines. In each panel, only the major conformation of each residue is shown for clarity. of the reducing agent sodium dithionite to this sample resulted in a shift of the Soret and Q bands of the heme absorption spectrum to wavelengths indicative of the ferrous nitrosyl (Fe II -NO) derivative (Fig. 3A, red spectrum). Spectral deconvolution indicated that 40% of the sample was converted to the ferrous nitrosyl form, 37% became unliganded ferrous deoxy, and 23% remained ferric met. Furthermore, the addition of dithionite to solutions of either S-nitro-glutathione or S-nitrocysteine resulted in the release of nitric oxide that could be detected by a chemiluminescence-based nitric oxide monitor. 3 This leads us to conclude that, upon reaction of Mb-SNO with dithionite, the NO group is cleaved from the Cys-10 sulfur and can subsequently bind to the reduced heme iron. As expected, treatment of unmodified metmyoglobin with dithionite resulted in primarily the unliganded ferrous deoxy form of the heme (Fig. 3B). Spectral deconvolution showed 0% ferrous nitrosyl, 77% ferrous deoxy, and 23% ferric met.
We were also able to observe this NO group transfer in two structures derived from a single crystal. Rapid data collection to minimize the x-ray dose on a crystal of S-nitrosomyoglobin produced a structure identical to that described earlier (Figs. 1 and 2), with Cys-10 modified by NO (Fig. 3C) and the heme in the ferric aquo state (Fig. 3D). Subsequent soaking of the same crystal in a solution containing 200 mM dithionite and the collection of another dataset showed that Cys-10 was no longer modified (Fig. 3E) and that the heme was primarily ligated by NO (Fig. 3F). This structure likely represents a mixture of the ferrous nitrosyl, ferrous deoxy, and ferric met forms observed upon dithionite reduction in solution, as described previously, with ferrous nitrosyl as the major species. Importantly, conformational changes observed in the structure after dithionite soaking show that the myoglobin reverts to the same conformation it had prior to S-nitrosylation. It is worth noting that FIGURE 3. NO group transfer from Cys-10 to the heme iron upon reduction by dithionite. A, UV-visible spectra of trans-S-nitrosylated ferric aquo myoglobin used for crystallization (black) following the addition of dithionite (red) and after the addition of dithionite and excess Cys-NO to saturate the ferrous nitrosyl form (blue). B, the same series of UV-visible spectra as in A except for unmodified myoglobin. The red spectra in A and B illustrate that reduction of trans-Snitrosylated ferric aquo myoglobin by dithionite produces mostly the ferrous nitrosyl form, whereas dithionite reduction of the unmodified myoglobin generates primarily the deoxy form. C, the 2F o Ϫ F c electron density map at 1.5 Å resolution, calculated with residue 10 and the upper axial heme ligand omitted from the model and contoured at 1.0 , is shown as blue mesh around the S-nitrosocysteine at position 10. D, the same electron density map as in C, contoured at 3 , is shown around the ferric aquo heme. His-60 and His-89 are shown, but other amino acid side chains were omitted from the figure for clarity. E and F, the 2F o Ϫ F c electron density map for data to 1.35 Å resolution from the same crystal as in C and D following a soak in dithionite-containing solution. The portions of the model displayed and electron density map calculations are the same as C and D. C and E show that the NO group is lost from Cys-10 following treatment with dithionite. D and F illustrate the conversion of the heme group from ferric aquo to primarily ferrous nitrosyl after soaking the crystal in dithionite. Note that the conformational change in myoglobin resulting from S-nitrosylation of Cys-10, which is illustrated in Fig. 2, is observed in these two structures obtained from the same myoglobin crystal. JULY 6, 2007 • VOLUME 282 • NUMBER 27 this conformational change is accommodated within the context of a single crystal. We wish to make clear that the reactions of Mb-SNO with dithionite do not necessarily shed light on the chemistry that may take place in vivo. Dithionite was used in these studies as a means to demonstrate the reversibility of the S-nitrosylation of Mb at Cys-10 and was convenient for the rapid soaking of crystalline samples. Nonetheless, it is interesting to note that if a suitable reductant were present in vivo, the myoglobin heme would efficiently capture the released NO.

S-Nitrosylation-induced Conformational Change
Ultrahigh Resolution Structure of Fe II -NO Myoglobin-Crystals of the Fe II -NO form of blackfin tuna myoglobin were prepared by anaerobic reduction and treatment with the nitric oxide donor 2-(N,Ndiethylamino)-diazenolate-2-oxide, and a dataset was collected to 0.95 Å resolution using synchrotron radiation. At this resolution, individual atoms of the heme-NO complex were resolved in the electron density map (Fig. 4), allowing highly accurate bond lengths and angles to be calculated. Geometrical parameters of the heme Fe-NO ligation from this structure are summarized in Table 2. The heme group was nearly planar, as demonstrated by the small distortion parameters in Table 2, calculated using normal coordinate structural decomposition calculations as described previously (35).

Production and Stability of S-Nitrosylated Blackfin Tuna
Myoglobin-We were able to S-nitrosylate myoglobin isolated from the red muscle of blackfin tuna by in vitro trans-S-nitrosylation using either S-nitrosocysteine or S-nitrosoglutathione. S-Nitrosylation of the myoglobin was confirmed using chemiluminescence, the Saville-Griess colorimetric assay, and by release of the NO group from Cys-10 and recapture at the heme iron following dithionite reduction monitored by UV-visible spectroscopy and x-ray crystallography (Fig. 3). The SNO group at Cys-10 was quite stable in the dark and in the absence of metal ions and produced highly ordered crystals quite readily, providing an ideal system for the study of the effects of S-nitrosylation on protein structure. Factors previously proposed to affect the formation or stability of S-nitrosothiols include the steric bulk surrounding the thiol (25), redox environment (36,37), reactivity of the thiol, and pH. The sulfur atom of Cys-10 in blackfin tuna myoglobin is partially solvent-exposed and is located between the side chains of an aspartate (Asp-118) and a  Ϫ0.037 Propellering (A 1u ) 0.049 a The 0.95-Å ferrous nitrosyl blackfin tuna myoglobin structure described in this work. b The 1.35-Å structure of the Mb-SNO crystal following a soak in a 200 mM dithionite solution. As demonstrated by UV-visible spectroscopy in solution (Fig. 3), this structure likely represents a mixture of Fe II -NO, Met, and deoxy myoglobin, with Fe II -NO as the major species. c X-ray crystal structure of nitrosyl-␣,␤,␥,␦-tetraphenylporphinato(1-methylimidazole)iron(II) (Fe(TPP)(NO)(1-MeIm)) (45). Values in parentheses are for a minor secondary conformation of NO observed in the crystal structure. d Optimized structure from density functional theory calculations on the NO-Fe II (por)-Im model system (46). lysine (Lys-9) at the surface of the protein. It has been proposed that similar "acid-base" motifs could be an important determinant of cysteine S-nitrosylation (1). The charged functional groups of the Asp-118 and Lys-9 side chains are located 7-9 Å from the Cys-10 sulfur, however, and therefore may not have a strong influence over its reactivity. In the S-nitrosylated structure, the sulfur-bound NO group is also partially surface-exposed and is surrounded primarily by hydrophobic amino acid side chains from the protein core and backbone carbonyl oxygens. No hydrogen bonds were observed to the NO group. Hydrophobic regions of proteins were also proposed to enhance S-nitrosylation by increasing the local formation of N 2 O 3 , a potent nitrosylating agent (38). This may play a role in the formation of the Mb-SNO, given that S-nitrosylated Cys-10 orients toward the hydrophobic protein interior in our structure. SNO Geometry-In the atomic resolution structure of S-nitrosylated blackfin tuna myoglobin, we observed two conformations of the modified cysteine side chain. Both were present as the cis conformer ( Fig. 1), with C-S-N-O dihedral angles of ϳ0°. Interestingly, two conformations of this group were similarly observed in the small molecule crystal structure of S-nitroso-L-cysteine ethyl ester (33), implying that some degree of static disorder may be an intrinsic feature of this chemical substituent. Quantum mechanical calculations demonstrate that two energy minima exist for the C-S-N-O dihedral angle, at 0 and 180°, due to delocalization of the N ϭ O electrons over the S-N bond (a phenomenon that may be familiar from the planar peptide bond within the backbone of proteins). This results in a slightly shorter S-N bond and a 12-kcal/mol energy barrier for rotation of the dihedral between 0 and 180° (39). Both conformations of Cys-10 -SNO that we observed in myoglobin closely matched the predicted and observed geometric parameters for S-nitrosothiols.
Very recently, structures of trans-S-nitrosylated human thioredoxin (PDB entries 2HXK, 2IFQ, and 2IIY) revealed S-nitrosylation at multiple cysteines, both buried and accessible. The observed modifications conform primarily to the cis dihedral, with partial occupancy of the trans-conformation observed at one site (34). A structure of the nitrophorin protein (PDB entry 1Y21) from Cimex lectularius revealed S-nitrosylation of a proximal heme ligand cysteine, also present as the cis conformer, with a C-S-N-O dihedral close to 0° (40). Only two other PDB entries are currently annotated as containing a protein S-nitrosocysteine. A structure of human hemoglobin treated with NO gas revealed modification of Cys␤93 (PDB entry 1BUW) (41) but was later concluded to be the thionitroxide radical C-S-NH-O⅐ based on comparison between the observed C-S-N-O dihedral angle (ϳ90°) and quantum mechanical calculations of several possible sulfur-bound nitrogen oxide species (39,42). Finally, an atomic resolution structure of a mammalian dimethylarginine dimethylaminohydrolase treated with S-nitroso-L-homocysteine (PDB entry 2CI1) showed a two-atom modification of the surface-exposed Cys-83. A C-S-N-O dihedral angle of Ϫ69.5°was observed, despite the fact that no steric clashes would occur if the group were present with a cis dihedral angle of 0°. Additionally, the S-N and N-O bond lengths deviated substantially from those observed in small molecule crystal structures (33) and predicted from high level calculations (39), suggesting that it may also be a more reduced sulfur-bound nitrogen oxide species.
S-Nitrosylation-induced Protein Conformational Change-Despite the evidence that S-nitrosylation affects the function of a wide variety of proteins, there is currently very little data directly demonstrating that S-nitrosylation can alter the structure of a protein. NMR experiments with p21 Ras GTPase demonstrated that S-nitrosylation at Cys-118 results in small changes in the H N chemical shift of residues surrounding Cys-118 in 1 H-15 N-HSQC spectra (7). The structural details of these small changes were not elucidated, however. S-Nitrosylation of the Escherichia coli redox-responsive transcription factor OxyR at Cys-199 led to small but significant changes in its circular dichroism spectra, implying a conformational change (6). In other systems, there has been speculation that S-nitrosylation may induce conformational changes in a protein to help explain the alteration of its function. For example, it was proposed that S-nitrosylation of Cys-399 of the N-methyl-D-aspartate receptor allosterically affects its quaternary organization and results in receptor desensitization (43), but no direct biophysical evidence is available to show the nature of these conformational changes.
Ultrahigh Resolution Fe II -NO Myoglobin Structure-Several crystal structures of the Fe II -NO adducts of myoglobins from sperm whale and horse heart have been determined previously and show significant variation in the Fe-N-O angle (summarized in Ref. 44). At resolutions lower than the N-O bond length (ϳ1.2 Å), the electron density feature for the NO group appears egg-shaped (Fig. 3F), making accurate refinement of the NO orientation difficult. At 0.95 Å resolution, individual peaks in the electron density map are present for the nitrogen and oxygen atoms (Fig. 4), which produces more confidence in their refined positions. The geometry of the NO-ligated heme in the 0.95 Å ferrous nitrosyl blackfin tuna myoglobin structure is very similar to that observed in both small molecule crystal structures of model compounds (45) and in high level theoretical calculations (46), with a Fe-NO bond length of 1.74 Å and a Fe-N-O angle of 134°( Table 2). The only other atomic resolution protein crystal structures that feature a nitric oxide-and histidine-ligated heme are of the nitrophorin 4 (NP4) protein from Rhodnius prolixus, a non-globin heme protein. The coordination geometry of the ferrous form of this protein is very similar to those described in Table 2, but it contains a significantly more distorted heme porphyrin, a consequence of the surrounding protein environment and a feature that is almost certainly important for its function (47,48). A detailed understanding of nitric oxide ligation by protein-bound heme has implications for the mechanism of soluble guanylate cyclase, which uses a heme to sense and become activated by NO. Currently, little structural information is available about soluble guanylate cyclase, and several mechanistic models exist regarding how NO binding to the heme group within this protein dramatically changes its activity (49).
Future Directions-Using x-ray crystallography, we were able to observe specific changes in the structure of blackfin tuna myoglobin as a direct result of S-nitrosylation at a surface-exposed cysteine residue (Fig. 2). Although there are technical challenges to the biophysical characterization of S-nitrosylation-induced conformational changes, such as the instability and x-ray sensitivity of some S-nitrosothiols, we hope this work will serve as a starting point to begin to understand the molecular effects of this modification on protein structure and function. Future experiments with blackfin tuna myoglobin will examine whether S-nitrosylation at Cys-10 has a functional effect on the oxygen-binding properties of this protein. It will also be important to determine whether S-nitrosylation of myoglobin occurs as a normal physiological process in the muscle tissue of the blackfin tuna.