Oxygen Binding and Redox Properties of the Heme in Soluble Guanylate Cyclase

Soluble guanylate cyclase is an NO-sensing hemoprotein that serves as a NO receptor in NO-mediated signaling pathways. It has been believed that this enzyme displays no measurable affinity for O2, thereby enabling the selective NO sensing in aerobic environments. Despite the physiological significance, the reactivity of the enzyme-heme for O2 has not been examined in detail. In this paper we demonstrated that the high spin heme of the ferrous enzyme converted to a low spin oxyheme (Fe2+-O2) when frozen at 77 K in the presence of O2. The ligation of O2 was confirmed by EPR analyses using cobalt-substituted enzyme. The oxy form was produced also under solution conditions at −7 °C, with the extremely low affinity for O2. The low O2 affinity was not caused by a distal steric protein effect and by rupture of the Fe2+-proximal His bond as revealed by extended x-ray absorption fine structure. The midpoint potential of the enzyme-heme was +187 mV, which is the most positive among high spin protoheme-hemoproteins. This observation implies that the electron density of the ferrous heme iron is relatively low by comparison to those of other hemoproteins, presumably due to the weak Fe2+-proximal His bond. Based on our results, we propose that the weak Fe2+-proximal His bond is a key determinant for the low O2 affinity of the heme moiety of soluble guanylate cyclase.

Soluble guanylate cyclase (sGC) 2 is a well characterized NO receptor involved in cell-cell signal transduction pathways associated with neuronal communication and vasodilation (1)(2)(3)(4)(5)(6)(7). Mammalian sGC is a heterodimeric (␣␤) hemoprotein (8 -10) in which the ␤ subunit binds a stoichiometric amount of heme via a weak bond between the heme iron and His-104 (11)(12)(13)(14). The binding of NO to the ferrous heme cleaves the weak Fe 2ϩ -proximal His bond, and the resultant NO complex with 5-coordinate NO heme markedly stimulates the enzymic production of cGMP (9, 14 -17). The enzyme-heme also binds carbon monoxide (CO) with moderate stimulation of enzyme activity. It has been thought that the ferrous enzyme-heme in sGC does not exhibit a measurable affinity for O 2 despite having a vacant axial position on the heme (9). This is in contrast to other hemoproteins with a Fe 2ϩ -proximal His linkage, including globins and heme-containing oxygenases. The lack of affinity for O 2 allows sGC to function as a selective NO-sensor even in the presence of high concentrations of O 2 and prevents oxidation of ferrous heme by O 2 .
We have examined the reaction of the enzyme-heme in sGC with external ligands using a rapid scan-stopped flow method as well as EPR, resonance Raman, and an infrared spectroscopy and established the following. (i) A 5-coordinate NO complex is produced via 6-coordinate NO complex in the reaction with NO (14). (ii) The ferric heme of sGC combines N 3 Ϫ to form a unique 5-coordinate high spin complex with a high cyclase activity (14). (iii) YC-1(3-(5Ј-hydroxymethyl-3Ј-furyl)-1-benzylindazole), an allosteric activator, induces the coordination changes in the CO complex from 6-coordinate CO heme to a 5-coordinate CO heme (17). Most of these anomalous heme coordination structures are specific to soluble guanylate cyclase among hemoproteins and seem to be associated with the weak Fe 2ϩ -proximal His bond.
X-ray Structural analyses of the H-NOX (heme-NO/oxygen binding) or SONO (sensor of nitric oxide) domain, which share considerable sequence homology with the sGC heme domain, have been carried out to identify possible key determinants for modulating the O 2 binding ability of sGC (18,19). H-NOX/ SONO have a proximal His residue that is probably involved in the signaling pathway for O 2 and/or NO. Spectroscopic characterization revealed that the H-NOX/SONO heme sensor domain from Thermoanaerobacter tengcongensis, an obligate anaerobe, produced a 6-coordinate NO heme and a stable oxyheme (Fe 2ϩ -O 2 ), reminiscent of globins (19,20). By contrast, the H-NOX/SONO protein from Clostridium botulinum resembles mammalian and insect sGCs (21) and forms a stable 5-coordinate NO heme but not a stable oxy heme (18). The crystal structure of the oxy form of T. tengcongensis H-NOX revealed that a Tyr residue on the distal side of the heme was * This work was supported by Frontier Project "Life Adaptation Strategies to Environmental Changes" of Rikkyo University and grants-in-aid from the Ministry of Culture, Education, Sports, Science, and Technology of Japan (to R. M. involved in hydrogen bond formation with the bound O 2 molecule (19). However, H-NOX proteins from facultative aerobes as well as typical NO-regulated sGCs from mammalian sources possess an Ile residue at the position corresponding to the distal Tyr (19). Replacement of the Tyr residue with Ile markedly reduced the O 2 affinity of the heme-domain of the T. tengcongensis protein, thereby substantiating the crucial role of the distal Tyr in the discrimination of O 2 binding in the H-NOX proteins (20). Mammalian sGC contains Ile-145 at the position homologous to the distal Tyr. Boon et al. (20) converted the Ile-145 of sGC ␤-subunit homodimer to Tyr and found that the mutant homodimer produced a stable oxy form, although the affinity for O 2 was extremely low. Hence, it was hypothesized that the absence of a hydrogen-bonding residue in the distal heme pocket is essential for O 2 exclusion by sGC. Martin et al. (22) have tested the hypothesis by employing a complete human sGC heterodimer. However, substitution of Ile-145 with Tyr in the ␤-subunit did not facilitate the binding of O 2 to the enzyme. This unexpected finding may be due to the inappropriate orientation of the Tyr phenolic OH group relative to the ligand. Indeed, a recent publication revealed that an additional mutation, I149E, in the distal pocket enabled the enzyme-heme to react with O 2 (23). The I149E mutation probably induces a repositioning of the phenolic OH group of the introduced Tyr toward the bound O 2 , facilitating the formation of a hydrogen bond. Although the above mutational study demonstrates that a hydrogen bond in the distal pocket is one of the main factors responsible for the stabilization of bound O 2 , the oxy form of the mutant enzyme was still unstable and only detected as a transient species. This finding implies that an additional factor(s) might be involved in the mechanism to control the reactivity of the heme for O 2 . Despite the important implication, detailed experiments to examine the reaction of sGC with O 2 have not been reported. In the present paper we describe the detection and characterization of the oxy form of sGC.
When ferrous sGC was frozen at 77 K, we found that the enzyme-heme converted to a new species with an optical spectrum similar to that of oxyhemoglobin. The new species was also produced under fluid conditions at Ϫ7°C, but the amount was remarkably small, suggesting an extremely low O 2 affinity of the ferrous heme. This species was assigned to be an oxy form by the spectral similarity with oxymyoglobin, by the inhibitory action of isocyanide for the new species formation, and by EPR characterization of the corresponding form of the Co 2ϩ -porphyrin-substituted enzyme. EXAFS analyses revealed that upon binding O 2 , the ferrous iron in the out-of-plane position moved toward the heme plane without rupture of the Fe 2ϩproximal His bond. These results indicate that the oxy form is in a 6-coordinate state and that the low affinity for O 2 is not caused by cleavage of the Fe 2ϩ -proximal His bond. Electrochemical analyses revealed that the enzyme-heme had the most positive midpoint potential (ϩ187 mV) among high spin protoheme-containing hemoproteins. This result strongly suggests that the electron density on the ferrous heme in sGC is significantly lowered relative to the ferrous heme of other hemoproteins. The decrease in the electron density, which may be due to the weak Fe 2ϩ -proximal His bond, weakens the Fe 2ϩ -O 2 bond strength as a result of diminished electron donation from iron to O 2 . Our results suggest the weak Fe 2ϩ -proximal His bond is a critical factor that could account for the lower O 2 affinity of sGC. By contrast, the distal protein effect, comprising steric hindrance in the distal heme pocket, had no significant impact on the ligand binding of sGC.

EXPERIMENTAL PROCEDURES
Enzyme Purification-Fresh bovine lung (5 kg) was minced and homogenized using a Waring blender in 15-liters of 50 mM potassium phosphate buffer, pH 7.4, containing a mixture of protease inhibitors (1 mM phenylmethylsulfonyl fluoride, 1 mM benzamidine, and 1 mM EDTA) and 55 mM ␤-mercaptoethanol (14,17,24). Protease inhibitors and ␤-mercaptoethanol were included in all the buffers throughout the purification unless stated otherwise. The successive purification steps of the enzyme were the same as those described earlier (24). The purified enzyme preparations were stored in liquid nitrogen until use.
Spectral Measurements-The formation of the oxy form was examined under the solution conditions at 3 and Ϫ7°C. The experiments at subzero temperature were carried out in 40 mM TEA buffer, pH 7.5, containing 10% (v/v) ethylene glycol as antifreeze. The temperature of the cuvette holder was maintained at Ϫ7°C or 3°C by thermomodule elements. The fully reduced sGC was added to the anaerobic buffer solution in a septum-sealed anaerobic cuvette that was kept in an anaerobic state by flushing with N 2 gas. The optical absorption spectra under the conditions were recorded on a PerkinElmer Life Sciences Lambda 18 spectrophotometer, and 5ϳ6 scans were averaged to improve the signal to noise ratio. After the spectra of the ferrous enzyme were collected, the solution was kept in two atmospheres of O 2 gas introduced via the septum. Optical spectra were recorded (average of 5ϳ6 scans) after carefully shaking to equilibrate with O 2 .
Optical spectra at 77 K were measured on a Shimadzu MPS-2450 spectrophotometer (Shimadzu, Kyoto, Japan) equipped with a homemade low temperature attachment consisting of a liquid N 2 Dewar and twin cuvettes (light path, 1 mm) for sample and reference solutions, as reported previously (25). To obtain a well balanced spectrum at 77 K, the buffer containing 5% ethylene glycol was employed.
EPR spectra were measured on a Varian E-12 X-band EPR spectrometer (Varian, Palo Alto, CA) with a 100-kHz field modulation. The sample temperature was controlled with an Oxford ESR-900 cryostat as described previously (14,17).
The iron K-edge EXAFS measurements were performed using synchrotron radiation at station BLC12C of Photon Factory in the National Laboratory of High Energy Physics (Tsukuba, Japan). EXAFS data were collected at 80 K as fluorescence spectra using a 13-element germanium array detector. The results presented in this paper are the average of multiple scans. Data analyses were carried out as described previously (26 -28). EXAFS, which were extracted by subtracting the background absorption, was converted to electron momentum k space, where k is a photoelectron wave vector. The resultant curve was then multiplied by k 3 to equalize the oscillation amplitude in the k space. Curve fittings were performed using the non-linear least squares program EXCURV92 on the raw data weighted by k 3 . Other details, including refinement of the data, are described elsewhere (26 -28).
Resonance Raman spectra were measured with a JASCO NR-1800 spectrophotometer equipped with a cooled chargecoupled device detector (Princeton Instruments, Trenton, NJ). The excitation wavelength was 413.1 nm from a krypton ion laser (Coherent, Innova 90). The sample was directly mounted on an aluminum sample holder and frozen with liquid N 2 . The sample holder was then inserted into the cryostat (Oxford DN1704), and the temperature of the sample was kept at 85 K with a temperature controller (Oxford ITC502).
Stopped-flow Measurements-The binding of alkyl isocyanides to the ferrous sGC was followed by a DX-18MV stoppedflow apparatus (Applied Photophysics, Leatherhead, UK). The anaerobic sGC solution was mixed with an anaerobic solution containing a desired amount of t-butyl or isopropyl isocyanide in the stopped-flow instrument at 20°C. The reaction was performed under pseudo-first order conditions, and the rate constants were determined by fitting to a single exponential function using built-in software. The association rate constant (k on ) and dissociation rate constant (k off ) were determined from the slope and the y axis intercept, respectively, in the plot of the observed rates versus isocyanide concentrations.
Activity Measurements-End-point assays were performed in the cooling bath maintained at Ϫ7°C by a Eulabo F 13 temperature controller. The assay mixture contained 470 M GTP, 7 mM MgCl 2 , 50 mM NaCl, 10% ethylene glycol, and an appropriate amount of the enzyme solution in 150 l of 40 mM TEA buffer. When desired, 104 M YC-1 or 37 M BAY41-2272 was added to the reaction mixture supplemented with 4% DMF to maintain the solubility of YC-1 or BAY41-2272. The mixtures were equilibrated in a septum-sealed anaerobic reaction vial with 2 atmospheres of O 2 or N 2 . Reactions were started by the addition of 1.5 M native or deuteroheme-substituted sGC and conducted at Ϫ7°C for 30 min. The reaction was terminated by the addition of 5 l of 30% acetic acid. The amount of cGMP formed was quantified by analysis on a C18 high performance liquid chromatography column equilibrated with 20 mM potassium phosphate buffer containing 10% methanol at a constant flow rate of 1 ml/min.
Oxidation-Reduction (Redox) Potential Measurements-Spectroelectrochemical apparatus originally designed by Tsujimura et al. (29) was modified to enable direct monitoring of the redox potential with a platinum indicator electrode. An anaerobic 1-cm path length cuvette was used with a septumcapped port for injection and a female ground glass joint at the top that fitted to a male joint of a micro combined electrode, EA 234 (Metrohm, Herisau, Switzerland). A platinum mesh electrode (52 mesh, 8 ϫ 10 mm) and platinum wire electrode (0.6-mm diameter) were also fixed on the inside walls of the cuvette to act as working and counter electrodes, respectively, in three-electrode potentiostat system. The combined electrode comprised a platinum indicator and Ag ϩ /AgCl reference electrode that was connected to a pH meter, model Accumet AR15 (Fisher), to directly monitor the electrode potentials. After the buffer solution containing mediators (1.6 ml) was introduced into the spectroelectrochemical cuvette, the com-bined electrode was attached to the cuvette by fitting the joints, thereby making an air-tight seal. The solution was purged with purified N 2 for 10 min. Then, the concentrated protein sample was introduced into the cuvette through the rubber septum cap. The cuvette was then placed in a temperature-controlled cell holder. The solution was constantly stirred with a magnetic stirrer during data collection. The desired redox levels of the protein sample were maintained by coulometric generation of mediator-titrant that was controlled by the three-electrode potentiostat system. The potential control by the three-electrode system was achieved by using a potentiostat, model HA-151A (Hokuto Denko Co., Tokyo, Japan). Redox potentials are quoted relative to the normal hydrogen electrode. The mediators used were 33 M Ru(NH 3 ) 6 Cl 3 , 33 M p-benzoquinone, 10 M toluylene blue, and 20 M 3Ј-chloroindophenol. Dithiothreitol included in the stored sGC solution was removed by passing through a Superdex 200HR column (GE Healthcare) to avoid undesired redox reactions.
Reagents-GTP, t-butyl isocyanide, and isopropyl isocyanide were purchased from Sigma). YC-1 was purchased from ALEXIS (San Diego, CA). Other chemicals, purchased from Wako Chemicals Co. (Tokyo, Japan), were of the highest commercial grade.

Detection and Characterization of the Oxy Form at Low
Temperature-We observed that the yellowish-red-colored sGC preparation in the air-saturated buffer changed to brilliant red upon freezing in liquid N 2 , suggesting formation of a low spin heme. We analyzed the temperature-dependent change by low temperature optical spectroscopy at 77 K. Under anaerobic conditions, in which dissolved O 2 was removed by Na 2 S 2 O 4 , the enzyme exhibited a spectrum corresponding to a high spin heme with the Soret band at ϳ430 nm and a broad band centered at 560 nm in the visible region at 77 K (Fig. 1a). By contrast, the spectrum of sGC in air-saturated buffer (i.e. in the absence of Na 2 S 2 O 4 ) at 77 K displayed well resolved ␣ and ␤ bands at 543 and 577 nm, respectively, accompanied by blue shift of the Soret band to 421 nm (Fig. 1b). Such distinct O 2 -dependent spectral changes were not observed at ambient temperature (298 K) where sGC exhibited a spectrum typical of ferrous high spin heme both in the presence and absence of Na 2 S 2 O 4 at 298 K (inset of Fig. 1). The new species did not exhibit EPR signals assignable to ferric heme iron at either 15 and 5K. These findings together indicate that the species is in a ferrous low spin state and may be assigned to an oxy form of sGC based on the spectral similarities with that of oxyhemoglobin and the absolute requirement of O 2 for its formation. Hereafter, we interpret experimental results by assuming that the ferrous heme iron in the enzyme is capable of binding O 2 like the cobalt-substituted enzyme.
Although we have no available data to argue the mechanism of the formation, it is clear that the oxy form is produced in a course of freezing. The putative oxy form exhibited a somewhat broad Soret band with an obvious shoulder at around 430 nm and with a significant red-shifted Soret peak position in comparison with the Soret peak (417 nm) of oxyhemoglobin (Fig.  1b). These spectral properties suggested that the formation of the oxy form was incomplete because of the low affinity of the enzyme-heme for O 2 . Likewise, complete formation of the oxy form was not observed for deuteroheme-substituted sGC despite the large increase in O 2 affinity by heme substitution as shown for myoglobin and hemoglobin (30,31) (Fig. 1e). Therefore, it is unlikely that the incomplete O 2 occupation is due to the low O 2 affinity of the enzyme-heme. The complete formation of the oxy form was achieved by the addition of YC-1, as shown by a blue shift of the Soret band and intensified ␣ and ␤ bands (Fig. 1, c and f). These findings seem to indicate that YC-1 converts the O 2 -insensitive conformation to an O 2 binding conformation.
We attempted to detect O 2 ligation at the axial position of the heme by resonance Raman spectroscopy at 80 K. Oxyhemoglobin used as control exhibited a 4 Raman band at 1380 cm Ϫ1 and a 3 band at 1511 cm Ϫ1 characteristic of 6-coordinate low spin oxy heme (data not shown). Unlike hemoglobin, sGC frozen at 80K in the air-saturated buffer exhibited only 5-coordi-nate high spin bands at 1358 and 1476 cm Ϫ1 without showing any 6-coordinate low spin Raman bands either in the presence and absence of YC-1 (data not shown). We interpreted these spectral characteristics to be caused by conversion of the 6-coordinate oxy heme to 5-coordinate heme as a result of photodissociation of the bound O 2 . Thus, it was not possible to identify O 2 ligation by resonance Raman spectroscopy. Next, we tried to detect ligation of O 2 at the metal center using Co 2ϩprotoporphyrin-substituted enzyme.
We have reported that the Co 2ϩ -protoporphyrin-substituted sGC has a weak Co 2ϩ -His bond and produces 5-coordinate NO complex like the native sGC (14). The present experiments further established that the metal substitution has no effect on the structure as well as catalytic properties of the enzyme, because there are no significant differences in the catalytic and the nucleotide binding properties and in the subunit structure and protein surface charges between the native and the Co-substituted enzymes (supplemental Figs. 1 and 2).
The optical spectrum of Co 2ϩ -proto sGC measured at 77 K in the presence of Na 2 S 2 O 4 showed a Soret band at 403 nm and 524-and 557-nm bands in the visible region ( Fig. 2A). When frozen in the air-saturated buffer, the 557-nm band of Co 2ϩprotoporphyrin sGC was significantly decreased in intensity with an appearance of a new band at 574-nm and a significant red shift of the Soret band (Fig. 2B). The addition of YC-1 augmented the O 2 -dependent spectral change (Fig. 2C). In the O 2 -saturated buffer supplemented with YC-1, the spectrum converted to that of a single species with 420, 542, and 574 bands (Fig. 2D). These spectral features essentially agree with those of oxy Co-proto myoglobin (32). The species showed a free radical type EPR signal with hyperfine structure at g 1 ϳ 2.08, which results from a coupling of unpaired electron to the Co nucleus (inset of Fig. 2). This is conclusive evidence for the formation of the oxy form (Co 3ϩ -O 2 Ϫ ) of cobalt-porphyrin (33,34). Co 2ϩ -mesoporphyrin sGC also exhibited similar O 2dependent spectral changes (supplemental Fig. 3).
The coordination structures of the enzyme-heme in the unliganded, O 2 -bound, and CO-bound states were examined by the iron K-edge EXAFS. The k 3 -weighted EXAFS and the corresponding Fourier transforms are summarized in Fig. 3. To fit the experimental data, His, His and O 2 , and His and CO were employed as the axial ligands for the ferrous, the O 2 -bound, and the CO-bound hemes, respectively. The atomic coordination of corresponding myoglobin derivatives were employed as a starting model of the curve fitting. The structural parameters that satisfy the raw data by these approaches are summarized in Table 1. EXAFS of the unliganded ferrous sGC, obtained in the presence of Na 2 S 2 O 4 (spectrum A in the left panel Fig. 3), was similar to that of deoxymyoglobin (28,35). The iron in the unliganded ferrous sGC was displaced relative to the heme plane (Fe 2ϩ -Ct, distance between Fe 2ϩ and porphyrin plane center) by 0.56 Å, like deoxymyoglobin ( Table 1). The displacement is characteristic of high spin heme, because the high spin heme iron cannot be forced into the porphyrin plane due to the large covalent radius of high spin iron.
It is generally accepted that the Fe 2ϩ -His bond strength of sGC is weaker than that of myoglobin due to strain at the heme center (13, 14, 17, 36). Although the difference in the Fe 2ϩ -His In traces c and f, 4% (v/v) DMF was supplemented in the above buffer to maintain the solubility of YC-1. The addition of DMF did not changed the spectral features both in anaerobic and aerobic conditions. In the inset, optical spectra of native sGC at 298 K in the presence and absence of Na 2 S 2 O 4 are shown. MAY 6, 2011 • VOLUME 286 • NUMBER 18 bond strength between sGC and myoglobin was thought to be reflected in the Fe 2ϩ -His bond distance, the Fe 2ϩ -His bond distance observed in unliganded sGC essentially agreed with that of deoxymyoglobin ( Table 1). The discrepancy may be explained by the tilting of the Fe 2ϩ -imidazole nitrogen bond from the heme normal, because such a distortion also weakens the Fe 2ϩ -His bond strength probably by decreasing -bond interaction between iron d-orbital and imidazole nitrogen p-orbital.

Oxy Form of Guanylate Cyclase
EXAFS of sGC in the presence and absence of O 2 were different, indicating binding of exogenous ligand at the axial position (spectrum B in the left panel of Fig. 3). Placing O 2 at the 6th position of the heme using the parameters listed in Table 1, the heme iron was found to still bind the proximal His with a bond distance of 2.13 Å ( Table 1). The iron displacement from the heme plane was much reduced upon placing O 2 on the heme iron (0.09 Å) ( Table 1). These results showed that the proximal His moved toward the heme plane along with the movement of the heme iron to produce the 6-coordinate oxy form, and therefore, the low affinity for O 2 may be not caused by cleavage of the Fe 2ϩ -His bond. In the case of the CO complex, iron displacement of the CO complex was 0.13 Å. The value was somewhat larger than that of the corresponding derivative of myoglobin FIGURE 2. Low temperature optical spectra of cobalt protoporphyrinsubstituted sGC. Trace A, Co 2ϩ -protoporphyrin-substituted sGC in the anaerobic buffer in which dissolved O 2 was removed by Na 2 S 2 O 4 was frozen at 77 K, and then the spectrum was taken. Trace B, the spectrum of Co 2ϩprotoporphyrin-substituted sGC in the air-saturated buffer is shown. Trace C, the spectrum of Co 2ϩ -protoporphyrin sGC in the air-saturated buffer containing YC-1(104 M) is shown. Trace D, the spectrum of Co 2ϩ -protoporphyrinsubstituted sGC in the O 2 -saurated buffer containing YC-1(104 M) is shown. The buffer used was 40 mM TEA, pH 7.5, containing 5% (v/v) ethylene glycol and 50 mM NaCl. Inset, X-band (9.22 GHz) EPR spectrum of Co 2ϩ -protoporphyrin-substituted sGC in the O 2 -saturated buffer was taken at 35 K and by 100 K Hz field modulation with 1-millitesla width. Other details including refinement of data are described elsewhere (26 -28).

TABLE 1 Iron-ligand distances for soluble guanylate cyclase and myoglobin in their unliganded, O 2 , and CO forms estimated by EXAFS
The parameters used to simulate the K-edge EXAFS are principally the same as those by Binsted et. al. (26). Distances (R) and Debye-Waller terms of the ligand atom coordinating iron (2 2 ) are included in the table. The abbreviations used are: Fe 2ϩ -Ct, distance of iron from the porphyrin plane center (magnitude of iron displacement); Fe 2ϩ -Npyr, bond distance between iron and porphyrin nitrogen atom; Fe 2ϩ -Nim, bond distance between iron and proximal His imidazole nitrogen atom; Fe 2ϩ -L, bond distance between iron and sixth ligand.  Fig. 3 and Table 1), but these values fall in the range of a low spin iron.
To characterize the distal heme pocket structure, we focused on the bond distance and angle of Fe 2ϩ -ligand coordination, which were obtained with multiple scattering analyses of a limited narrow k range EXAFS (3-13 Å Ϫ1 ) (28). The geometries of sGC-ligand complexes are Fe 2ϩ -O 2 bond distance ϭ 1.89 Å and Fe 2ϩ -O-O bond angle ϭ 120°for the oxy form, and Fe 2ϩ -CO bond distance ϭ 1.85 Å and Fe 2ϩ -C-O bond angle ϭ 171°for the CO form. The coordination geometry for myoglobin obtained by the present curve fitting procedure was 1.85 Å (bond distance) and 104°(bond angle) for the O 2 form and 1.81 Å and 149°for the CO form. These values for myoglobin essentially agreed with those obtained by x-ray crystallographic analyses (37,38). The present method for data analyses is useful for estimating iron-ligand geometry, although there is uncertainty in bond angle determination of ϳ10°, including absolute and fitting errors (28,39). Therefore, the EXAFS studies on sGC indicate that the CO moiety of the Fe 2ϩ -CO unit binds to the heme iron with a nearly linear geometry. Such geometry implies no steric protein effect on the distal side of the heme. Conversely, O 2 seems to accommodate on the distal pocket with intrinsically bent Fe 2ϩ -O 2 structure.
To establish whether the ligation of O 2 affects the cyclase activity, we attempted to detect the oxy form under fluid conditions. In these experiments the formation of the oxy form was followed by the difference spectra against the spectrum in the presence of N 2 . At Ϫ7°C, the difference spectra exhibited a 410-nm peak and a 434-nm trough in the Soret region and peaks at 539 and 578 nm in the visible region (trace a in Fig. 4A). The intensity of the 410-nm peak in the difference spectrum significantly increased in the presence of YC-1 (trace b in Fig.  4A). The spectral species formed under an atmosphere of O 2 was assignable to an oxy form, because these peak and trough positions were essentially the same as those in the difference spectrum at 77 K (Fig. 4B). The amount of oxy form produced at Ϫ7°C was estimated to be only 3% of total protein even in the presence of YC-1, based on the pure oxy form obtained arithmetically as described below. The formation of oxy form was temperature-dependent and significantly decreased upon raising the temperature to 3°C (data not shown).
Deuteroheme substitution has been known to increase O 2 affinity (30,31). As anticipated, the degree of oxy form formed at Ϫ7°C was greater than that of the native enzyme (trace c in Fig. 4A). Nevertheless, the yield of oxy form for the deuteroheme-substituted enzyme was still only ϳ4%. YC-1 significantly increased the formation of the oxy form in the deuteroheme-substituted enzyme to approximately ϳ7% of total protein (trace d in Fig. 4A).
The effect of O 2 on cyclase activity was examined in the presence of YC-1 at Ϫ7°C (Fig. 4C). The addition of NO resulted in 120-fold activation of the native enzyme, whereas the presence of O 2 did not appear to enhance the cyclase activity in the absence and presence of YC-1. Similar results were also observed for the deuteroheme-substituted enzyme, where the presence of O 2 did not enhance the cyclase activity of the substituted-enzyme in the presence of YC-1 and its derivative, BAY 41-2272 (3-(5Ј-hydroxymethyl-3Ј-furyl)-1-benzylindazole) (Fig. 4C). In this connection it is interesting to note that the cyclase activity catalyzed by Gcy-88E from Drosophila was not stimulated by CO, NO, and O 2 (40). This enzyme formed stable 6-coordinate NO and O 2 complex, unlike typical sGC. These results together with the present data suggest that the liganddependent stimulation may be closely coupled with the Fe 2ϩ -His bond strength, although the detailed mechanism remains to be elucidated. To ask whether the exogenous heme-ligand can prevent O 2 binding to the enzyme-heme, we examined the inhibitory effect on the O 2 binding. As shown in Fig. 5A, the O 2 -dependent spectral change in the presence of BAY 41-2272 yielded the largest change observed so far (compare the solid line with the dotted line). In contrast, the O 2 -dependent spectral change was noted to be much reduced when isopropyl isocyanide was included in the mixture (compare trace a with trace b in the inset of Fig. 5A). The difference spectrum (trace b) does not agree with trace a in the entire spectral region. Although the trace b displayed the trough at 420 nm similar to that of trace c, the bandwidth of trace b was much larger than that of trace c, suggesting the displacement of the bound isocyanide by O 2 . The nearly identical result was also observed for native sGC in the reaction of the t-butyl isocyanide adduct with O 2 (data not shown). These spectral features agreed with the view that isocyanide competed with the same site as O 2 .
The spectrum of the complete oxy form could be arithmetically obtained by subtracting the ferrous-enzyme spectrum multiplied by a factor from the spectrum in the O 2 -saturated buffer. In the presence of YC-1 and BAY 41-2272, as shown in Fig. 5A, the subtractions of the ferrous enzyme multiplied by 0.97 and 0.89 could satisfactorily eliminate the residual ferrous enzyme for the native and the deuteroheme enzyme, respectively. The resultant arithmetic spectra (Fig. 5B) agree with those of the oxy form obtained at 77K (Fig. 1). On the basis of these results together with EXAFS and other optical spectral studies described in this paper, we finally conclude that the heme in sGC is capable of binding O 2 .
Protein Effects on Ligand Binding-Using the systematic kinetic data, Mims et al. (41) assessed the protein effects on the ligand binding of myoglobin. The effects are summarized as (i) distal steric hindrance to restrict the approach of the ligand to its final position and (ii) a protein proximal effect that controls successive bond formation of the iron with ligand on the distal side. When the bond between iron and ligand (such as O 2 ) is generated, the distal His residue forms a hydrogen bond with the bound O 2 to stabilize the O 2 complex. These mechanisms, which were originally formulated for myoglobin, provide key clues in understanding the ligand binding of other hemoproteins.
We have assessed the distal steric protein effect by analyzing the reaction of the ferrous heme iron with alkyl isocyanides. As shown in Fig. 6A, careful titration with t-butyl isocyanide showed that the unliganded ferrous sGC converted to the isocyanide adduct through one set of isosbestic points. The binding of isopropyl isocyanide also yielded nearly identical species (data not shown). The optical spectra of these alkyl isocyanide adducts were essentially the same as those of the corresponding isocyanide adducts of myoglobin, demonstrating that the isocyanide adducts of sGC were in a six-coordinate low spin state (42). The dissociation constants (K d ) calculated were 46 and 90 M for t-butyl isocyanide and isopropyl isocyanide (inset of Fig.  6), respectively, indicating that the affinity for t-butyl isocyanide is significantly higher by comparison to that for isopropyl isocyanide. In general, the affinities of isocyanide for the hemoproteins decreased with increasing size of alkyl group in the isocyanide molecule (41). In contrast, sGC exhibited a higher affinity for isocyanide with a larger alkyl group. The anomalous property has also been noted for human hemeoxygenase (43). The association rates of isocyanides with the ferrous sGC decreased with increasing size of alkyl group in the isocyanide molecule. The rates are 10-fold faster than the formation of the corresponding isocyanide adducts of myoglobin ( Fig. 6B and Table 2) and nearly equivalent to those of leghemoglobin (44), which are the highest among the globin family. Taking into consideration these findings, we propose that there is no substantial resistance to the binding of bulky isocyanides with sGC, in contrast to myoglobin, supporting the previous result (45).
In accordance with a previous report (22), sGC binds CO with an association rate constant of 3.6 ϫ 10 4 M Ϫ1 s Ϫ1 , which is particularly slow among 5-coordinate high spin hemoproteins ( Table 2). The most striking feature is that the association rate The buffer used was 40 mM TEA buffer, pH 7.5, containing 10% ethylene glycol, 50 mM NaCl, 4% DMF, and 37 M BAY 41-2272. The difference spectra were summarized in the inset, where trace a is a difference spectrum by subtracting the spectrum in the N 2 -saturated buffer from that in the O 2 -saturated buffer, and trace b was obtained by subtracting the spectrum in the N 2 buffer from that in the O 2 buffer, of which buffers contained isopropyl isocyanide (170 M). Trace c is the difference spectrum by subtracting the spectrum of the isopropyl isocyanide adduct from that of the ferrous enzyme, which is illustrated in the 1 ⁄10 scale. In B, the spectrum of the complete oxy form was arithmetically obtained by subtracting the ferrous spectrum multiplied by a factor from the spectrum in the O 2 -saturated buffer. The data used for the native enzyme were the absolute spectra (in N 2 and O 2 ) used to generate trace b in Fig. 4A and for the deuteroheme-substituted enzyme the spectra in Fig. 5A. The subtractions of the ferrous enzyme multiplied by 0.97 and 0.89 could satisfactorily eliminate the residual ferrous enzyme for the native and the deuteroheme enzymes, respectively.
for isopropyl isocyanide was approximately one order of magnitude faster than that for CO binding, irrespective of the larger size of the isocyanide by comparison to CO (Table 2). These findings strongly suggest that sGC permits easy access for small ligands, such as CO and O 2 , to the coordination position. Therefore, the formation of the coordinate bond dominates the kinetics of association, accentuating the significance of the proximal protein effect.
Based on quantum mechanical and molecular mechanical analyses, it was proposed that the degree of electron density on the heme iron might be controlled by the protein proximal effect (46). This effect on the proximal side modulates the Fe 2ϩligand bond formation on the distal side. For instance, when the Fe 2ϩ -proximal His bond is weakened by strain imposed on the bond, the electron density on the heme iron is reduced relative to that of the Fe 2ϩ -His bond without strain. This might reduce the electron donation from the iron to ligand such as O 2 , resulting in the weakening of the Fe 2ϩ -O 2 bond strength (46). Thus, a weak Fe 2ϩ -His bond correlates with a weak Fe 2ϩ -O 2 bond and vice versa. Such a protein effect, referred to as a positive trans effect, accounts for the unique character of the CO and O 2 binding characteristics of the iron-porphyrin complexes (47)(48)(49)(50). The proximal effect derived by strain on the Fe 2ϩ -His bond probably affects redox potential of the heme, because the decrease in electron density at the ferrous heme makes it more difficult to remove an electron (51). For example, T-state hemoglobin with a more strained Fe 2ϩ -His bond and lower O 2 affinity relative to R-state hemoglobin exhibited significantly higher midpoint potentials of the heme than R-state hemoglobin (52). The redox potential measurements of sperm whale myoglobin and sGC indicated that the electrochemical titration curves fitted to a Nernst equation with n ϭ 1 in both hemoproteins (Fig. 7). The ferric-ferrous couple of myoglobin gave the midpoint potential of ϩ58 mV, in reasonable agreement with the reported value. The midpoint potential of sGC, ϩ187 mV, was considerably higher than that FIGURE 6. Equilibrium binding and kinetic analyses of alkyl isocyanide binding. A, shown are changes in the absorption spectra of sGC during titration with t-butyl isocyanide. B, k obs values obtained from stopped-flow traces were plotted against the final concentration of isocyanide after mixing. Closed circles (F) are data from t-butyl isocyanide, and closed squares (f) are data from isopropyl isocyanide. In the inset, isocyanide binding curves are plotted as the fractional saturation versus the effective concentration of isocyanide. Solid lines depicted simulated lines obtained by nonlinear regression analyses. Equilibrium and kinetic binding experiments were done at 20°C. Other details are described under "Experimental Procedures."

TABLE 2
Ligand binding properties of soluble guanylate cyclase, sperm whale myoglobin, and soybean leghemoglobin of myoglobin. It should be noted that the measured value was the most positive among high spin protoheme-containing hemoproteins, presumably because the electron density of the heme in sGC is significantly reduced by the protein proximal effect. Based on the above considerations and EXAFS data, the significance of the proximal protein effect on the ligand binding was noted as summarized in Fig. 8.
In summary, we have detected and characterized the oxy form of sGC. To assess the crucial determinant(s) for the low O 2 affinity of sGC, we have analyzed the coordination structure by EXAFS. Our results indicate that the low affinity for O 2 is not caused by cleavage of the Fe 2ϩ -proximal His bond. Among protein effects to regulate the reactivity of heme, the distal steric hindrance could be excluded based on the kinetic studies. The critical factor that may contribute to the low O 2 affinity is the protein proximal effect, a regulatory effect caused by the weak Fe 2ϩ -proximal His bond. Measurement of the midpoint potential of the heme also highlights the significance of the protein proximal effect in terms of the unique low O 2 affinity of sGC. Based on the findings described in this paper, we to propose that the weak Fe 2ϩ -proximal His bond is a key factor in regulating the reactivity of the heme in sGC along with the hydrogen bond interaction in the distal pocket, which was identified previously by mutational analysis (20).  (Table 1). Closed circles illustrated in the center of heme denote ferrous iron atom. The weak Fe 2ϩ -His bond in the ferrous enzyme may decrease charge density on the heme iron, thereby decreasing the O 2 affinity and elevating the redox potential of the heme.