Dissociation of Nitric Oxide from Soluble Guanylate Cyclase and Heme-Nitric Oxide/Oxygen Binding Domain Constructs*

Regulation of soluble guanylate cyclase (sGC), the primary NO receptor, is linked to NO binding to the prosthetic heme group. Recent studies have demonstrated that the degree and duration of sGC activation depend on the presence and ratio of purine nucleotides and on the presence of excess NO. We measured NO dissociation from full-length α1β1 sGC, and the constructs β1(1–194), β1(1–385), and β2(1–217), at 37 and 10 °C with and without the substrate analogue guanosine-5′-[(α,β-methylene]triphosphate (GMPCPP) or the activator 3-(5′-hydroxymethyl-3′-furyl)-1-benzylindazole (YC-1). NO dissociation from each construct was complex, requiring two exponentials to fit the data. Decreasing the temperature decreased the contribution of the faster exponential for all constructs. Inclusion of YC-1 moderately accelerated NO dissociation from sGC and β2(1–217) at 37 °C and dramatically accelerated NO dissociation from sGC at 10 °C. The presence of GMPCPP also dramatically accelerated NO dissociation from sGC at 10 °C. This acceleration is due to increases in the observed rate for each exponential and in the contribution of the faster exponential. Increases in the contribution of the faster exponential correlated with higher activation of sGC by NO. These data indicate that the sGC ferrous-nitrosyl complex adopts two 5-coordinate conformations, a lower activity “closed” form, which releases NO slowly, and a higher activity “open” form, which releases NO rapidly. The ratio of these two species affects the overall rate of NO dissociation. These results have implications for the function of sGC in vivo, where there is evidence for two NO-regulated activity states.

Soluble guanylate cyclase (sGC) 4 is the best characterized physiological receptor for the gaseous signaling agent nitric oxide (NO) (1)(2)(3)(4)(5). In response to NO, sGC produces the second messenger cGMP, modulating physiological processes such as neurotransmission and vasodilation (6). The ␣1␤1 sGC heterodimer is activated several hundredfold above the basal level by the binding of NO to the heme of the ␤1 H-NOX domain (7)(8)(9), a conserved domain of unique structure (10 -12). However, it remains unclear how this binding event is translated into increased catalytic activity.
The mechanism by which sGC deactivation occurs has been a focus of much investigation (8,13,14). Initially thought to result from simple dissociation of NO from the heme, the deactivation process has turned out to be more complicated. In fact, regulation of sGC by NO has been shown to involve a complex interplay between binding of NO to the heme and to non-heme sites as well as allosteric regulation by GTP, also a substrate, and ATP, a reporter for the energy status of the cell (15). In integrating the inputs from all these signals, sGC has been shown to have a remarkable attribute, the ability to exist in a stable low activity state or a transient high activity state, both containing NO bound to the heme (16,17). Many questions concerning the existence and characteristics of these two sGC heme-NO species, and how they might be regulated by NO and nucleotides, remain to be addressed.
In addition to responding to cellular inputs, sGC can be activated by a class of small molecules exemplified by 3-(5Ј-hydroxymethyl-3Ј-furyl)-1-benzylindazole (YC-1). These molecules not only activate sGC in the absence of heme ligands but also synergize with the NO-and CO-bound forms of the enzyme to reach maximal activity. There has been much speculation about the binding site and mechanism of action of YC-1 (18 -20); however, experimental results remain inconclusive. By using purified sGC, and sGC heme domain constructs ( Fig.  1) that possess heme characteristics similar to those of the fulllength ␣1␤1 enzyme (21,22), we carried out spectroscopic and kinetic analyses of NO dissociation in the absence and presence of YC-1 and the substrate analogue guanosine-5Ј-[(␣,␤-methylene]triphosphate (GMPCPP), as well as activity studies with YC-1. Dissociation was found to exhibit two exponential phases, the relative contributions of which could be differentially affected by YC-1, GMPCPP, or changes in temperature. We propose a model for dissociation of NO from sGC involving the existence of two 5-coordinate sGC heme-NO species indistinguishable by electronic absorption spectroscopy but clearly different in their ability to release NO from the heme; this model is discussed in context of the current hypothesis for regulation of sGC by NO and nucleotides.
Baculovirus Construction-Rat sGC ␤1 cDNA was inserted between the NotI and XbaI sites of the plasmid pFastBac1 (Invitrogen) to generate pFastBac1/sGC␤1. PCR was used to insert an in-frame C-terminal RGS-H 6 tag in front of the stop codon of the ␣1 cDNA. The forward primer was 5Ј-TGGCGG-CCGCAAGGAGGAAACCAC-3Ј, and the reverse primer was 5Ј-CGTCTAGATTAGTGGTGGTGGTGGTGGTGAGATC-CTCTATCTACCCCTGATGCTTTGCCTAAGAAGTTAG-CGTTTCC-3Ј. PCR products were sequenced to confirm the presence of the desired changes (University of Michigan Biomedical Research Core Facilities). The H 6 sGC ␣1 gene was inserted between the NotI and XbaI sites of pFastBac1 to generate the construct pFastBac1/sGC␣1. The Bac-to-Bac baculovirus expression system (Invitrogen) was used to generate recombinant baculoviruses from pFastBac1/sGC␣1 and pFastBac1/sGC␤1 according to the manufacturer's pro-tocol. High titer stocks of recombinant baculoviruses were prepared by standard methods. Optimization of the amount of each virus used for protein production was carried out as described previously (23).
Cell Culture and Production of Recombinant sGC-Sf9 cells were cultured in Ex-Cell 420 insect serum-free medium (JRH Biosciences) supplemented with 10% fetal calf serum (Hyclone) and 1% antibiotic/antimycotic (Invitrogen) at 28°C. Cultures were grown in 2800-ml Fernbach flasks with shaking at 135 rpm. Cells were subcultured between 0.7 ϫ 10 6 and 5 ϫ 10 6 cells/ml. Cell density and viability were determined by trypan blue exclusion using a hemocytometer. For protein expression, 1-liter cultures of Sf9 cells at a density of 1.5-2 ϫ 10 6 cells/ml in 2800-ml Fernbach flasks were infected with H 6 ␣1 and ␤1 recombinant viruses. Cells were harvested 3 days post-infection by centrifugation, and the pellet was stored at Ϫ80°C.
Purification of Recombinant sGC-All manipulations were carried out at 4°C. Frozen cell pellets from 5-liter expression cultures were thawed on ice and resuspended in buffer A (50 mM KH 2 PO 4 , pH 8.0, 200 mM NaCl, 5 mM ␤-mercaptoethanol, 1 mM imidazole, 1 mM Pefabloc (Pentapharm), 1 mM benzamidine, 5% glycerol) plus Complete EDTA-free protease inhibitor mixture (Roche Applied Science). Resuspended cells were broken with a Bead Beater (BioSpec Products) using 0.1-mm diameter glass beads, and the lysate was centrifuged at 200,000 ϫ g for 2 h. The supernatant was applied to a 2.5-ml column of nickel-nitrilotriacetic acid-agarose (Qiagen) equilibrated with buffer A at a flow rate of 1 ml/min using a BioLogic LP (Bio-Rad). The column was washed with buffer A until the A 280 was stable, and then an aliquot of buffer A (25 ml) was brought to 1.2 M NaCl and applied to the column. The column was washed with 50 ml of 12.5 mM imidazole in buffer A and eluted with 25 ml of 125 mM imidazole in buffer A, collecting 2-ml fractions during the elution. Fractions containing sGC (identified by yellow color) were pooled, concentrated to 1-1.5 ml in a Vivaspin-20 50K filter (Vivascience), and exchanged into buffer B (25 mM triethanolamine, pH 7.4, 25 mM NaCl, 5 mM dithiothreitol)) on a PD-10 column (Amersham Biosciences). The sample was diluted to ϳ7 ml with buffer B and applied to a 2-ml prepacked POROS HQ2 anion-exchange column (Applied Biosystems) at 2 ml/min using a BioLogic Duo Flow (Bio-Rad). The column was washed with 5 ml of buffer B and developed with a 35-ml 120 -285 mM gradient of NaCl in buffer B, collecting 1-ml fractions. Fractions containing purified sGC (exhibiting an A 278 /A 431 Ͻ 1.1) were pooled, concentrated in a Vivaspin-6 50K filter, drop-frozen in liquid N 2 , and stored in liquid N 2 . Protein purity was assessed by SDS-PAGE using pre-cast 10% Tris-glycine gels (Invitrogen) and was routinely greater than 95%. Protein concentrations were determined using the Bradford microassay (Bio-Rad) or calculated from the A 431 using an extinction coefficient of 148,000 M Ϫ1 cm Ϫ1 (8).
Dissociation of NO from the Heme of ␤1 , ␤1(1-385), ␤2 , and sGC-The dissociation of NO from the heme of each H-NOX domain construct and sGC was measured at 37 and 10°C using the CO/dithionite trapping method described previously (16,24). The trapping solution was prepared as follows: a solution of sodium dithionite (Na 2 S 2 O 4 ) in 50 mM HEPES, pH 7.4, 50 mM NaCl was prepared in a Teflon-sealed Reacti-Vial (Pierce) using an anaerobic chamber (Coy Laboratory Products). The solution was removed from the anaerobic chamber and saturated with CO by bubbling the gas through the solution for 10 min. Protein-NO complexes were formed by incubation with excess DEA/NO (in 10 mM NaOH) at 25°C in 50 mM HEPES, pH 7.4, 50 mM NaCl for 10 min. Complete conversion to the nitrosyl species was verified by following the shift in the Soret maximum from 431 to 399 nm. Stock solutions of YC-1 were made in Me 2 SO. When present, YC-1 concentrations ranged from 0.96 to 96 M, and the final concentration of Me 2 SO was 1%. Experiments with GMPCPP contained 10 -1000 M nucleotide (added before DEA/NO addition) and included 5 mM MgCl 2 , which alone had no effect on NO dissociation. Proteins were placed in a septum-sealed anaerobic cuvette and deoxygenated using an oxygen-scavenged gas train. A small amount of DEA/NO (ϳ3 eq) was added just before deoxygenation to maintain the nitrosyl species (any remainder was subsequently destroyed by the large excess of Na 2 S 2 O 4 in the trapping solution). The head space of the anaerobic cuvette was replaced with CO, and the cuvette and trap solutions were equilibrated at assay temperature for 1 min. The reaction was initiated by addition of CO/dithionite solution to the anaerobic cuvette with a Hamilton gas-tight syringe and mixing. The final concentration of Na 2 S 2 O 4 in the reaction mixture was 30 mM. Final protein concentrations were 1.9 -2.5 M for ␤1 , ␤1(1-385), and ␤2(1-217), and 0.88 -2.5 M for sGC. Data collection was initiated ϳ10 s after trap addition. The reaction was monitored by electronic absorption spectroscopy using a Cary 3E spectrophotometer equipped with a Neslab RTE-100 temperature controller. Data were collected over the range of 380 -450 nm at 909 nm/min with a 1.5-nm data point interval. Spectra were recorded every 18 s for 5 min, every 1 min for 10 min, and every 2 min thereafter for a total of 3 h, or until the reaction was complete. A buffer base line was subtracted from each spectrum, and spectra were corrected for base-line drift by normalization to an isosbestic point at ϳ410 nm. For data obtained in the absence of YC-1 or GMPCPP, difference spectra were obtained by subtraction of the time 0 spectrum from all subsequent spectra. To obtain difference spectra for data acquired in the presence of YC-1 or GMPCPP, a time 0 spectrum from a reaction carried out in the absence of either compound and containing an identical amount of protein was subtracted from all subsequent spectra, and all time points were offset by an amount corresponding to the mixing time for the experiment. Values for the change in absorbance at 423 nm (⌬A 423 ; ␤1  and ␤1(1-385)) or 424 nm (⌬A 424 ; sGC and ␤2(1-217)) were extracted from the difference spectra and plotted versus time to obtain dissociation time courses for each experiment. Dissociation time courses were obtained in duplicate or triplicate, and each experiment was repeated 2-5 times over several days. Generally, because of the relative difficulty in obtaining large amounts of purified sGC, ⌬A 424 values for full-length sGC, which are proportional to the experimental protein concentrations, were smaller than for the heme domain constructs.
YC-1 Activation of the sGC-NO Complex-End point assays were performed in triplicate at 10 and 37°C as described previously (25). Stock solutions of DEA/NO (10 mM) were prepared in NaOH (10 mM). Stock solutions of YC-1 (15 mM) were pre-pared in Me 2 SO. Assay mixtures contained 0.2 g of sGC in 50 mM HEPES, pH 7.4, 2 mM dithiothreitol, and 150 M YC-1 where indicated. sGC was incubated with DEA/NO (100 M) for 10 min at 25°C and equilibrated at assay temperature for 1 min. Assays were initiated by addition of GTP and MgCl 2 to 1 and 3 mM, respectively. Final assay volumes were 100 l and contained 2% Me 2 SO, which did not affect enzyme activity. Reactions were quenched after 3 min by the addition of 400 l of 125 mM Zn(CH 3 CO 2 ) 2 and 500 l of 125 mM Na 2 CO 3 . cGMP quantification was carried out using a cGMP enzyme immunoassay kit, format B (Biomol), per the manufacturer's instructions. Each experiment was repeated four times to ensure reproducibility.
Data Analysis and Statistics-Curve fitting, data analysis, and figure generation were carried out using Kaleidagraph (Synergy Software). The data from each dissociation experiment were fit to single or double exponentials as shown in Equations 2 and 3 under "Results" to obtain observed rate constants. To determine whether a single exponential or two exponentials best fit the data, the residuals from each fit were compared. Additionally, for each set of dissociation data, the fit to a single exponential was compared with the fit to two exponentials using the F test. A two-exponential fit was considered better than a one-exponential fit if p Ͻ 0.0001. For dissociation experiments, rates are expressed as means Ϯ S.D. For activity assays, significant differences between the means were determined using the two-tailed t test; the significance level used was 0.05.

RESULTS
Kinetic Considerations; Developing a Model for NO Dissociation from sGC-The binding of NO to the sGC heme has been shown to proceed through initial formation of a 6-coordinate intermediate, followed by rupture of the iron-histidine bond to form the final 5-coordinate ferrous nitrosyl complex (7,26,27). The simplest mechanism for dissociation of NO from the 5-coordinate sGC heme-NO complex would be the reverse of NO binding as follows: rebinding of the proximal histidine ligand to form a 6-coordinate heme-NO intermediate followed by dissociation of NO to form a 5-coordinate histidyl-ligated heme, as shown in Scheme 1.
The values of k Hr and k Hd are the rate constants for rebinding and dissociation, respectively, of the proximal histidine ligand; k on is the rate constant for binding of NO to the heme, and k off is the rate constant for the dissociation of NO from the 6-coordinate heme-NO complex. In this study, we employed an NO trap, consisting of a sodium dithionite (Na 2 S 2 O 4 ) solution saturated with CO (CO sat ), to obtain rates for dissociation uncomplicated by NO rebinding. This system has been used previously to determine the NO dissociation rates for a number of heme proteins (16,24,28) and functions by destroying dissociated NO through reaction with dithionite and by preventing NO rebinding by blocking the open heme coordination site with CO. At the concentrations of dithionite and CO used in these experiments, the reaction of NO with dithionite and the binding of CO to the vacated heme coordination site are not ratelimiting (Refs. 24 and 29 and data not shown). Thus, under our NO-trapping conditions, the above reaction scheme simplifies to Scheme 2, where k off is the rate constant for the dissociation of NO from the 6-coordinate heme-NO complex to form the CO complex, which is irreversible because of the CO sat /dithionite trap. Assuming a steady-state equilibrium between heme-NO 5C and heme-NO 6C , Equation 1 can be derived for the observed reaction rate after mixing (derivation of the first-order rate constant k obs for mechanisms similar to that described in Scheme 2 has been discussed in detail (29 -33)), A single exponential increase in the concentration of heme-CO is expected when starting from a uniform population of either heme-NO 5C or heme-NO 6C , as described by Equation 2, where ⌬A t is the change in signal amplitude at time t; ⌬A T is the total change in signal amplitude, and k 1 is the observed reaction rate constant. Importantly, when starting from a mixture of heme-NO 5C and heme-NO 6C , if k off is faster than k Hr , a twoexponential increase as described by Equation 3 is predicted, where ⌬A t is the change in signal amplitude at time t; ⌬A 1 and ⌬A 2 are the contributions of each exponential process to the total change in signal amplitude, and k 1 and k 2 are the observed rate constants for each process. That is exactly what is observed for dissociation of NO from several prokaryotic H-NOX domains, for which the heme-NO complexes have been demonstrated to exist as an equilibrium mixture of 5-and 6-coordinate states (29). However, previous studies using electronic absorption and resonance Raman spectroscopy have demonstrated that for sGC and the sGC H-NOX constructs studied in this work, the heme-NO complex is exclusively 5-coordinate (21,34,35). From these observations, it can be inferred that there is no appreciable amount of heme-NO 6C in a solution of NO-bound sGC, relegating heme-NO 6C to the status of a transient intermediate in the dissociation reaction pathway. Thus, in order to accommodate any observed two-exponential dissociation of NO, Scheme 2 must be expanded to include an additional 5-coordinate heme-NO species, as shown in Scheme 3, where heme-NO 5C ‫ء‬ and heme-NO 5C are two 5-coordinate heme-NO species that are spectroscopically identical but kinetically distinct. In this mechanism, the 5-coordinate heme-NO complex of sGC exists as an equilibrium mixture of 5-coordinate heme-NO species (heme-NO 5C ‫ء‬ and heme-NO 5C ), with slow interconversion between the two forms compared with NO dissociation. A two-exponential NO dissociation time course would be observed if, upon mixing with the NO trap, NO dissociated from the heme-NO 5C fraction with a k obs according to the reaction in Scheme 2, with the remainder of the dissociation reaction proceeding from the heme-NO 5C ‫ء‬ fraction, as indicated in Scheme 3, and dependent on the rate of interconversion between 5-coordinate species (k O Ϫ k C ). Thus, the observation of two exponentials versus one in the dissociation of NO from a sGC H-NOX domain would indicate that NO dissociates from a mixture of 5-coordinate heme-NO species, with ⌬A 1 and ⌬A 2 proportional to the amount of each species at the start of the dissociation reaction.
Dissociation of NO from sGC and sGC H-NOX Domain Constructs at 37°C-In the presence of the NO trap (CO sat /dithionite), dissociation of NO from the heme of sGC and sGC H-NOX domain constructs resulted in an increase in absorbance at 423-424 nm because of formation of the heme-CO complex. A representative set of difference spectra for dissociation at 37°C from sGC, ␤1(1-194), ␤1(1-385), and ␤2(1-217) is shown in Fig. 2. The corresponding plots of ⌬A 423 or ⌬A 424 against time, shown in the top panels of Fig. 3, A-D, yield a dissociation time course for each construct. For each dissociation time course, the fits to both single and double exponentials (Equations 2 and 3) are shown. The residuals for each fit are plotted above each time course. In each case, examination of the residuals suggested that a two-exponential fit provided a better model for the data than a single exponential fit. A comparison of the one-and two-exponential fits using the F test supported the two-exponential fit as the better model in each case (p Ͻ 0.0001). The two rate constants obtained for each construct (a faster constant k 1 and a slower constant k 2 , averaged from 2-4 experiments per construct) and the amplitudes of each corresponding phase (as a percent of the calculated total) are shown in Table 1. The observed data are consistent with a model where dissociation proceeds from an initial equilibrium mixture of two 5-coordinate heme-NO complexes, as outlined in Scheme 3. Accordingly, we propose that k 1 corresponds to dissociation of NO from the heme-NO 5C conformation at a rate equal to k obs in Equation 1, whereas k 2 represents the observed rate of reaction, corresponding to k O Ϫ k C , that is limited by the slower conversion from heme-NO 5C ‫ء‬ to heme-NO 5C .
To further investigate the possibility that the 5-coordinate heme-NO complex of sGC can exist in two conformations, we examined the effect of the small molecule activator YC-1 (36) on NO dissociation from sGC, ␤1(1-194), ␤1(1-385), and ␤2(1-217) at 37°C (Fig. 3 and Table 1). YC-1 caused no significant change in the amplitudes or rate constants for NO dissociation from ␤1  or ␤1(1-385). However, for dissociation of NO from full-length sGC in the presence of YC-1, k 1 increased ϳ15-fold, k 2 increased roughly 5-fold, and ⌬A 1 , the fractional amplitude due to k 1 , modestly increased from 35 to 66% (Fig. 3C). The increase in both rate constants and the doubling of ⌬A 1 indicate that NO dissociation from sGC is significantly accelerated by YC-1. A similar effect was observed for ␤2(1-217) as follows: k 1 doubled, k 2 tripled, and ⌬A 1 increased from 40 to 66% of the total amplitude change (Fig. 3D). Thus, at 37°C, YC-1 appears to accelerate the dissociation of NO from sGC and ␤2(1-217), but not from ␤1  or ␤1(1-385), by increasing k 1 , k 2 , and the fraction of sGC-NO with the fast dissociation rate.
Dissociation of NO from sGC and sGC H-NOX Domain Constructs at 10°C-The temperature dependence of the observed rate constants and amplitudes for each exponential in the NO dissociation time course was examined at 10°C in the absence and presence of YC-1. Plots of NO dissociation time courses at 10°C for each construct are shown in Fig. 4, and the derived rate constants and amplitudes are shown in Table 2. In the absence of YC-1, the values of the observed rate constant k 1 for ␤1  and ␤1(1-385) are similar to those obtained at 37°C, whereas those of full-length sGC and ␤2(1-217) are slightly increased. However, ⌬A 1 , the fraction of the calculated total amplitude described by k 1 , is markedly diminished for all constructs, dropping from 25,29,35, and 41% at 37°C to 6, 7, 5, and 15% at 10°C for ␤1 , ␤1(1-385), sGC, and ␤2(1-217), respectively (compare Tables 1 and 2). Furthermore, although the slower rate constant k 2 for each construct appears to double for every 10°C increase in temperature (exhibiting an Arrhenius temperature dependence), the faster rate constant k 1 changes very little from 10 to 37°C.
The presence of YC-1 had negligible effect on the dissociation of NO from ␤1(1-194) and ␤1(1-385) at 10°C, with little change in k 1 , k 2 , ⌬A 1 , or ⌬A 2 (Fig. 4, A and B). Furthermore, the acceleration of NO dissociation from ␤2(1-217) by YC-1 at 37°C was not detected at 10°C; k 1 , k 2 , ⌬A 1 , and ⌬A 2 were unaffected by the presence of YC-1 (Fig. 4D). In contrast, the acceleration of the dissociation of NO from sGC by YC-1 was even more pronounced at 10°C than at 37°C (Fig. 4C). Although the presence of YC-1 resulted in a moderate increase in both k 1 and k 2 , ⌬A 1 increased from 5 to 73% of the calculated total amplitude at 10°C, a 15-fold change, compared with an ϳ2-fold change for sGC at 37°C. Thus, at 10°C, although YC-1 did not significantly alter the dissociation of NO from ␤1(1-194), ␤1(1-385), or ␤2(1-217), it drastically accelerated the dissociation of NO from sGC, primarily by increasing ⌬A 1 , the fraction of the calculated total amplitude described by k 1 .
The Effect of YC-1 on the Exponentials Observed in Dissociation of NO from sGC-To gain more information about the YC-1-induced acceleration of NO dissociation from sGC, we carried out dissociation reactions at 10°C in the presence of increasing amounts of YC-1. As shown in Fig. 5A, ⌬A 1 increased as the concentration of YC-1 was increased. ⌬A 1 values for each curve were plotted against log[YC-1] and fit to a concentration dependence function (Equation 4), yielding an approximate EC 50 value of 4 M for the ability of YC-1 to increase ⌬A 1 at 10°C (Fig. 5B).
Together, these observations suggest that YC-1 shifts the equilibrium ratio of 5-coordinate sGC-NO species from heme-NO 5C ‫ء‬ toward heme-NO 5C (as shown in Scheme 3), leading to an acceleration of NO dissociation. YC-1 has been shown to potentiate the NO stimulation of sGC (37); to examine whether the YC-1-induced changes in the fraction of the calculated total amplitude described by k 1 have any implications for enzyme activation, the basal and NO-stimulated activities of sGC in the absence and presence of YC-1 were measured at 37°C and 10°C (Fig. 5C). As expected, in each case sGC activity was lower at 10°C than at 37°C. However, the fold-activation by NO was much lower at 10°C (1.5-fold) than at 37°C (107-fold), qualitatively mirroring the much smaller ⌬A 1 at 10°C (5%) versus 37°C (35%) (Fig. 5D). Furthermore, YC-1 potentiated the NO-stimulated activity of sGC to a much greater extent at 10°C (7.1-fold over NO alone) than at 37°C (1.4-fold over NO alone), which correlates well with the increase in ⌬A 1 caused by YC-1 at 10°C (5 to 73%) versus 37°C (35 to 66%) (Fig. 5E). These results are consistent with the observation that YC-1 increases ⌬A 1 , the fraction of NO dissociation from sGC that occurs via k 1 , and suggest that the heme-NO 5C form of the enzyme exhibits higher activity than does the heme-NO 5C * form.

The Effect of Substrate on the Exponentials Observed in NO
Dissociation from sGC-The presence of substrate GTP has been reported to greatly accelerate the dissociation of NO from sGC (16,38). However, the presence of two phases in the NO dissociation time course was not examined. To further investigate the effect of substrate on the dissociation of NO from the sGC heme, experiments were carried out with increasing concentrations of GMPCPP, a noncyclizable GTP analogue, at 10°C (Fig. 6). In the presence of 1 mM GMPCPP (or GTP, data not shown; the noncyclizable analogue was used to ensure that the nucleotide concentration remained constant throughout the course of the experiment), NO dissociation was extremely rapid, as reported (16). Furthermore, time courses for NO dissociation in the presence of GMPCPP exhibited two exponential phases. The data were fit to Equation 3 to obtain values of 0.178 Ϯ 0.009 s Ϫ1 for k 1 , 0.00154 Ϯ 0.00011 s Ϫ1 for k 2 , and a ⌬A 1 of 83% (Table 3). Compared with data obtained in the  absence of GMPCPP at 10°C, k 1 increased 22-fold; k 2 increased 8-fold; and ⌬A 1 , the fractional amplitude attributed to k 1 , increased 6-fold, from 13 to 83% (Table 3). As shown in Fig. 6, the increase in ⌬A 1 was found to be dependent on the concentration of GMPCPP, and fitting the data to Equation 4 yielded an EC 50 of 9 M for the ability of GMPCPP to increase ⌬A 1 at 10°C. The observed substrate-mediated increase in dissociation of NO from sGC appears to occur by a mechanism similar to that of YC-1, namely an increase in the fraction of the reaction that occurs via the fast phase. Although YC-1 and GMPCPP have similar potency (EC 50 values of 4 and 9 M, respectively) and efficacy (⌬A 1 values of 84 and 75%, respectively), GMPCPP increases k 1 and k 2 to a greater extent than YC-1 (Table 3), leading to faster overall NO dissociation. GMPCPP or GTP caused no significant change in the amplitudes or rates for NO dissociation from ␤1(1-194), ␤1(1-385), or ␤2(1-217) (data not shown), indicating that the acceleration of NO dissociation by GTP for the full-length enzyme must be mediated by domains or interactions not present in the constructs used here.

DISCUSSION
Over the years, the interaction of NO with sGC has been investigated with a variety of methods, leading to several models for the regulation of activation and deactivation of sGC by NO. Deactivation of the enzyme has been linked to dissociation of NO from the heme, and to gain insight into this process, NO dissociation has been studied in some detail (8,16,24,38). The simplest model for the dissociation of NO from the sGC heme involves the rebinding of the proximal histidine to form a 6-coordinate heme-NO complex and subsequent dissociation of NO to form the 5-coordinate histidyl-ligated enzyme. Because studies employing electronic absorption spectroscopy have suggested that the sGC heme-NO complex is exclusively 5-coordinate (8,35,39), the model predicts that NO dissociation should occur as a single exponential process (Equation 2) with   (1-385), and ␤2(1-217) are best described by two exponential processes (Equation 3), and we propose a model for sGC-NO dissociation (Fig. 7A) that invokes an equilibrium between two ferrousnitrosyl enzyme species indistinguishable by electronic absorption spectroscopy.
Our model is analogous to one proposed to explain multiple exponentials observed for ligand binding to hexacoordinate globins, and to the CO sensor CooA, by invoking an equilibrium between "closed" and "open" protein conformations (30,40). In both the globin and CooA models, the closed conformation was proposed to undergo a conformational shift to an open form prior to ligand binding, with the equilibration occurring on a slower time scale than subsequent ligand binding to the open conformation. An analogous mechanism has also been discussed in the context of hydrogen exchange kinetics, where a protein exists in an equilibrium between two conformations, one of which is capable of undergoing H-D exchange (32). In that case, two exponentials are also observed. Based on the two-exponential NO dissociation time courses revealed in this work (Figs. 3 and 4), we propose that the 5-coordinate sGC heme-NO complex exists as an equilibrium mixture of closed (heme-NO 5C ‫ء‬ ) and open (heme-NO 5C ) conformations. The observations that the rate constants and relative fractional amplitudes of each exponential can be shifted by changes in temperature, or by addition of small molecules such as YC-1 or GMPCPP, provide further support for this hypothesis.
Ligand-induced conformational changes in proteins are common, and the idea that the binding of NO to sGC induces a conformational change is well accepted. The hypothesis that the sGC-NO complex can exist in multiple conformations has also been discussed previously. Based on a flash-photolysis ) to obtain an EC 50 of 9 M for the GMPCPP-induced increase in ⌬A 1 , a value similar to that found for YC-1. The Mechanism of Action of YC-1 and the Effect of GTP-YC-1 weakly activates unliganded sGC (4 -8-fold) and synergistically activates CO-bound sGC to levels similar to those caused by NO (36). Significantly, YC-1 has also been reported to potentiate activation of sGC by NO (36,37). In one study it was found that the deactivation rate of sGC-NO was markedly reduced by the presence of YC-1. The authors proposed that YC-1 potentiation of NO-stimulated sGC activity is because of decreased dissociation of NO from the sGC heme in the presence of YC-1, based on the assumption that deactivation of NO-stimulated sGC is a proxy for NO dissociation (37). However, recent studies have shown that deactivation does not always correlate with dissociation (16). Moreover, YC-1 accelerates overall NO dissociation from full-length sGC (Table 1), an effect magnified by lowering the temperature from 37 to 10°C (Table  2). Thus, the acceleration of NO dissociation by YC-1 reported here indicates that the potentiation of NO-stimulated activity by YC-1 cannot be a result of a reduction in the rate of NO dissociation.
The acceleration of NO dissociation by YC-1 is reminiscent of that caused by the presence of GTP (16,38) and suggests that YC-1 and GTP might both accelerate NO dissociation through a similar mechanism. Indeed, we found that NO dissociation in the presence of the substrate analogue GMPCPP exhibits two exponentials, and the large increase in NO dissociation caused by GMPCPP, like that caused by YC-1, is because of an increase in k 1 , k 2 , and ⌬A 1 (Fig. 6). Together, these data imply that GTP and YC-1 may affect NO dissociation from sGC via the same kinds of conformational changes, perhaps involving overlapping binding sites. Like YC-1, the presence of GTP has no effect on NO dissociation from ␤1  or ␤1(1-385), which suggests that these molecules may either bind to the catalytic domains or require their presence to have an effect on dissociation. Recent studies have proposed that both YC-1 and GTP bind to an allosteric site in the catalytic domains of sGC (19,43,44), and it will be interesting to see if that turns out to be the case. It remains a possibility that the N-terminal region of the ␣1 subunit may influence NO dissociation from the ␤1 H-NOX domain or the effects of GTP and YC-1. Unfortunately, we were unable to examine this possibility, as our attempts to express constructs containing the N terminus of the ␣1 subunit were hindered by aggregation (data not shown).
Because both GTP and YC-1 accelerate overall NO dissociation from sGC by increasing the fractional amplitude due to k 1 (⌬A 1 ), and the presence of either GTP or YC-1 has been reported to increase the activity of NO-stimulated sGC (16,17,37), we predicted that the fraction of open sGC-NO species present would correlate with the degree of potentiation of enzyme activity by YC-1. This turned out to be the case; conditions that decreased ⌬A 1 , such as lowering the temperature from 37 to 10°C, resulted in diminished fold-activation by NO (a measure of relative NO stimulation) (Fig. 5D). Similarly, conditions that increased the fractional amplitude due to k 1 , such as the inclusion of YC-1, increased fold-activation by NO. Significantly, the potentiation of NO-stimulated activity by YC-1 was greatest at 10°C (7.1-fold versus 1.4-fold at 37°C), where the YC-1-induced increase in the fractional amplitude due to k 1 was also the greatest (from 5 to 73% at 10°C versus from 35 to 66% at 37°C) (Fig. 5E). Thus it is likely that the potentiation of NO activation of sGC by YC-1 is because of an ability to shift the equilibrium between the different 5-coordinate sGC-NO species from one with low activity to one with high activity.
Physiological Significance of Two Conformations for the 5-Coordinate sGC Heme-NO Complex-The existence of two species of sGC heme-NO has important implications for sGC function, especially in light of recent findings tying spectrally similar sGC heme-NO species with different levels of activation and NO-binding behavior in the presence of allosteric nucleotide modulators to two physiologically important NO signaling modes, tonic and acute (16,17). It was shown that incubation with GTP prior to addition of NO, or with excess NO, is required for maximal activation of the sGC heme-NO complex, an effect that is blocked by ATP. Similarly, preincubation of sGC with GTP was found to accelerate NO dissociation from the heme, an effect that is also blocked by ATP. In the presence of both ATP and GTP, the sGC heme-NO complex was demonstrated to exist as a low activity species, and NO in excess of the heme was needed for full activity. Importantly, YC-1 also converted the low activity sGC heme-NO complex to a fully active species. The results reported in this work provide mechanistic details explaining the effects of GTP and of YC-1 on activation of sGC by NO (Fig. 7B); binding of these molecules causes a conformation change from a low activity species that binds NO tightly to a high activity species with a weaker affinity for NO. Furthermore, this work indicates that the effects of GTP and YC-1 require the presence of the sGC catalytic domains, suggesting a role for these domains in regulating the conformational transition from open to closed and providing support for proposed allosteric binding sites for nucleotides and YC-1 on sGC (19,43,44).
In this work we also studied NO dissociation from the H-NOX domain of the ␤2 isoform of sGC. We found that the ␤2 H-NOX domain is similar to the ␤1 H-NOX domain in that it also exhibits a two-exponential NO dissociation time course with similar rates and amplitudes for each exponential. However, unlike for the ␤1 H-NOX domain constructs, NO dissociation from ␤2(1-217) was significantly accelerated by YC-1, a feature more like sGC. On the other hand, the acceleration of NO dissociation from ␤2(1-217) by YC-1 was reduced by lowering the temperature, which is the opposite of what was observed for full-length sGC. Together, these observations indicate that the ␤2 isoform of sGC, the H-NOX domain of which is 43% identical to the ␤1 H-NOX domain (21), has dif-ferent NO-binding characteristics. These differences might reflect the more specialized physiological localization and function of the ␤2 isoform, which is expressed primarily in the kidney (45) and has been proposed to function as a homodimer (46).