- Email: [email protected]
A proximal tryptophan in NO synthase controls activity by a novel mechanism
Journal of Inorganic Biochemistry 83 (2001) 301–308 www.elsevier.nl / locate / jinorgbio
A proximal tryptophan in NO synthase controls activity by a novel mechanism Subrata Adak, Dennis J. Stuehr* Department of Immunology, Lerner Research Institute, Cleveland Clinic Foundation, 9500 Euclid Avenue, Cleveland, OH 44195, USA Received 15 March 2000; received in revised form 15 June 2000; accepted 17 July 2000
Abstract The heme of neuronal nitric oxide synthase (nNOS) participates in O 2 activation but also binds self-generated NO, resulting in reversible feedback inhibition. We utilized mutagenesis to investigate if a conserved tryptophan residue (Trp409), which engages in p-stacking with the heme and hydrogen bonds to its axial cysteine ligand, helps control catalysis and regulation by NO. Mutants W409F and W409Y were hyperactive regarding NO synthesis without affecting cytochrome c reduction, reductase-independent N-hydroxyarginine oxidation, or Arg and tetrahydrobiopterin binding. In the absence of Arg electron flux through the heme was slower in the W409 mutants than in wild-type. However, less NO complex accumulated during NO synthesis by the mutants. To understand the mechanism, we compared the kinetics of heme–NO complex formation, rate of heme reduction, k cat prior to and after NO complex formation, NO binding affinity, NO complex stability, and its reaction with O 2 . During the initial phase of NO synthesis, heme–NO complex formation was three and five times slower in W409F and W409Y, which corresponded to a slower heme reduction. NO complex formation inhibited wild-type turnover 7-fold but reduced mutant turnover less than 2-fold, giving mutants higher steady-state activities. NO binding kinetics were similar among mutants and wild type, although mutants also formed a 417 nm ferrous-NO complex. Oxidation of ferrous-NO complex was seven times faster in mutants than in wild type. We conclude that mutant hyperactivity primarily derives from slower heme reduction and faster oxidation of the heme–NO complex by O 2 . In this way Trp409 mutations minimize NO feedback inhibition by limiting buildup of the ferrous–NO complex during the steady state. Conservation of W409 among NOS suggests that this proximal Trp may regulate NO feedback inhibition and is important for enzyme physiologic function. 2001 Elsevier Science B.V. All rights reserved. Keywords: Neuronal notric oxide synthase; heme–NO complex; ferrous–NO complex
1. Introduction Nitric oxide (NO) is synthesized from Arg by hemecontaining monooxygenases known as NO synthases (NOSs). The reaction requires NADPH and O 2 as cosubstrates, and occurs in two steps with generation of NhydroxyArg (NOHArg) as an intermediate [1,2]. Calmodulin binding activates NO synthesis in the NOSs by triggering heme reduction by the reductase domain. Although NOSs and cytochrome P-450s both utilize a cysteine thiolate-ligated heme to bind and activate O 2 , they have quite different heme site and tertiary structures [3–5]. In neuronal NOS (nNOS) (Fig. 1), the essential cofactor 6R-tetrahydrobiopterin (H4B) binds at the NOS heme edge perpendicular to its plane. Two conserved aromatic res*Corresponding author. Tel.: 11-216-445-6950; fax: 11-216-4449329. E-mail address: [email protected] (D.J. Stuehr).
idues Phe-584 and Trp-409 stack with the heme on its distal and proximal sides, respectively. This aromatic stacking is absent in cytochrome P450s but is present in peroxidases and influences their electronic and catalytic properties [6]. The NOS crystal structures predict that the indole nitrogen of Trp-409 forms a strong hydrogen bond with the cysteine thiolate heme ligand in NOS. In cytochrome P450s, hydrogen bonding to the coordinating thiolate is often minimal, but stronger hydrogen bonding is present in the thiolate-ligated hemeprotein chloroperoxidase [7]. Thus, it is important to determine how these structural features impact NO synthesis in NOS. The NOS heme also functions under unique circumstances. Consider that during aerobic steady-state NO synthesis approximately 60–90% of nNOS is present as a ferrous–NO complex due to binding self-generated NO [8]. Ferric and ferrous NOSs both form stable six-coordinate NO complexes in the presence of H4B or substrate and regenerate fully active ferric NOS upon breakdown of
0162-0134 / 01 / $ – see front matter 2001 Elsevier Science B.V. All rights reserved. PII: S0162-0134( 00 )00176-8
302
S. Adak, D. J. Stuehr / Journal of Inorganic Biochemistry 83 (2001) 301 – 308
Fig. 1. Some unique aspects of the nNOS heme environment. The figure is based on the mouse iNOS oxygenase domain crystal structure (3). The view looks up from the proximal side of the heme to highlight aromatic stacking between the heme and F584 and W409 in nNOS. Thin lines indicate hydrogen bonds between a W409 indole nitrogen and the cysteine thiolate, and between H4B and a heme propionate.
the complex [9–13].This distinguishes NOSs from the cytochrome P450s which typically form unstable ferrous NO complexes that irreversibly inactivate the enzyme [14,15]. In cytochrome P450s, formation of a five-coordinate ferrous–NO complex typically occurs after NO binds and leads to formation of cytochrome P420 upon air oxidation, loss of heme in some cases, and irreversible loss of activity [14–18]. This inactivation links NO synthesis with down-regulation of tissue P450 activity during inflammation [19–21]. In NOSs, reversible NO complex formation allows NO to act as a feedback inhibitor of activity during the steady state and makes the rate of NO synthesis by the enzyme proportional to the O 2 concentration across the physiologic range [22,23]. We have mostly studied heme–NO complex formation in nNOS. After initiating NO synthesis, nNOS molecules quickly partition between a catalytically active form and an inactive ferrous–NO complex [8]. This creates two cycles that have their own rate-limiting step and together determine the observed rate of NO synthesis during the steady-state (Fig. 2). The rate-limiting step in the active cycle that generates NO has been suggested to be either heme reduction or product release [24–27]. If a nNOS molecule forms a ferrous–NO complex, it is placed in the inactive cycle and the rate-limiting step becomes the O 2 dependent decay of the heme–NO complex [22]. Because NO binding to the heme regulates NOS catalysis, it is important to identify the structural features that allow NOS to function in this capacity. To address this issue, we have used site-directed mutagenesis to investigate what role Trp-409 has in regulating nNOS activity. Recently we found that this proximal tryptophan controls its activity by regulating NO feedback inhibition [28]. Mutation of Trp-409 to Phe or Tyr in nNOS resulted in less heme–NO complex buildup and
Fig. 2. Proposed mechanism for autoinhibition of nNOS by NO under steady-state catalysis. nNOS cycles between an active, NO-generating cycle (Left) and an inactive cycle (Right) that involves formation and O 2 -dependent decay of a ferrous–NO complex. Product NO may bind to ferric or ferrous nNOS which partitions nNOS molecules to the inactive cycle.
hyperactivity with regard to NO synthesis. This report summarizes our past and more recent studies on W409 mutants that show hyperactivity results from changes in electron flux and heme–NO complex oxidation.
2. Materials and methods
2.1. Materials Oxygen and NO gas were purchased from Liquid Carbonic Company and Matheson Inc., respectively. All other reagents and materials were obtained from Sigma or sources as previously reported [28,29].
2.2. Molecular biology W409 mutations were made in the oxygenase domain (nNOSox) (a.k.a. heme domain) as described for full-length nNOS [28] using the same primers and the Quick Change PCR in vitro Mutagenesis Kit from Stratagene. Mutations were confirmed by DNA sequencing at the Cleveland Clinic core facility.
2.3. Protein expression and purification Wild-type and W409 mutant nNOS proteins (oxygenase domain and full length) were expressed in Escherichia coli and purified as described previously [28,29]. UV-visible spectra was recorded on a Hitachi U3110 Spectrophotome-
S. Adak, D. J. Stuehr / Journal of Inorganic Biochemistry 83 (2001) 301 – 308
ter in the absence or presence of 20 mM H4B and 1 mM Arg. The ferrous–CO adduct absorbing at 444 nm was used to quantitate the heme protein content using an extinction coefficient of 74 mM 21 cm 21 (A444–A500).
2.4. NO synthesis, NADPH oxidation and cytochrome c reduction
303
EDTA, 3 mM calmodulin (CaM), 200 mM DTT, 20 mM H 4 B, 80 mM NADPH, and 1 mM Arg, final volume 1 ml. Reactions were started by adding 1.2 mM Ca 21 to cause CaM binding and monitored by wavelength scanning at 158C. The concentration of ferrous–NO complex formed during NO synthesis was estimated from absorbance change at 436 nm using an extinction coefficient of 49 800 M 21 cm 21 [8].
The initial rate of NO synthesis, NADPH oxidation, or cytochrome c reduction by wild type and mutants were quantitated at 258C as described earlier [28,29].
2.8. Arg binding
2.5. H2 O2 -dependent NOHA oxidation
Arg binding affinity was studied at 308C by perturbation difference spectroscopy with imidazole according to methods described previously [30].
H 2 O 2 -dependent nNOS oxidation of NOHA to nitrite was assayed in 96-well microplates at 258C as described previously [28] with modification. The assay volume was 100 ml and contained 40 mM 4-(-2-hydroxyethyl)-1piperazinepropanesulfonic acid (EPPS) buffer pH-7.6, 250 nM nNOS or mutants, 1 mM N v hydroxy-L-arginine (NOHA), 1 mM DTT, 25 U / ml SOD, 0.5 mM EDTA and 4 mM (6R)-5,6,7,8-tetrahydro-L-biopterin (H 4 B). Reactions were initiated by adding 30 mM H 2 O 2 , and stopped after 10 min by adding 1300 U of catalase. Nitrite was detected at 550 nm using the Griess reagent (100 ml) and quantitated based on nitrite standards.
2.9. Characteristics and kinetics of NO binding Spectra were collected with the rapid scanning diode array detector (Hi-Tech, model SF-61). Anaerobic solutions containing wild type or mutant nNOS (3 mM), 100 mM EPPS buffer pH 7.6, 1 mM Arg, 10 mM H 4 B, 0.2 mM DTT were pre-reduced with excess dithionite (100 mM), and then rapid mixed with NO-saturated buffer at 108C. To study NO binding kinetics, the initial NO concentration was varied by diluting the saturated solution with anaerobic buffer, and absorbance change was recorded with a single wavelength detector.
2.6. Heme reduction 2.10. Reaction of ferrous–NO complexes with oxygen Optical spectra were recorded on a Hitachi 3010 UVvisible spectrophotometer at 158C. Anaerobic spectra were recorded using septum-sealed quartz cuvettes that could be attached through a quick-fit joint to a vacuum system. The samples were made anaerobic by alternating cycles of evacuation and refilling with nitrogen. The cuvette was maintained under positive nitrogen pressure. The ferric nNOS (3 mM) solution, containing 40 mM EPPS buffer pH 7.6, 6 mM CaM, 0.9 mM EDTA, 1 mM DTT, 1 mM Arg and 10 mM H4B were mixed with 20 mM of deoxygenated NADPH to cause flavin reduction. Heme reduction was initiated after adding 1.2 mM Ca 21 to trigger CaM binding and heme reduction. Spectra were recorded after reaching equilibrium. Kinetic measurements were carried out using a Hi-Tech stopped-flow apparatus (model SF-51) equipped for anaerobic work. Anaerobic buffer solutions containing 2 mM ferric enzyme, 100 mM EPPS buffer pH-7.6, 200 mM NO solution, 80 mM NADPH, 4 mM CaM, 0.5 mM EDTA, 10 mM H4B, and 5 mM Arg was rapid mixed with a solution containing 5 mM CaCl 2 at 108C. The heme reduction rate was monitored at 540 nm.
2.7. Heme–NO complex formation during steady-state Subsequently, 0.6 mM nNOS was diluted in an airsaturated 40 mM EPPS buffer, pH 7.6, containing 0.9 mM
Anaerobic solutions of full-length nNOS enzymes (4 mM) containing saturating Arg and H 4 B were reduced with a minimum amount of dithionite and then saturated NO solution was added to give an NO concentration of 0.1 mM. These ferrous–NO complexes were then rapid mixed with air-saturated buffer solutions at 108C. Absorbance change was monitored at 436 or 393 nm using a single wavelength detector. Signal to noise ratios were improved by averaging the 10 individual scans. The time course of absorbance change was best fit to a single exponential equation by use of a nonlinear least-squares method provided by the instrument manufacturer.
3. Results and discussion
3.1. Arginine binding To investigate if mutation of Trp 409 affected substrate binding on the distal side of the heme, we compared the ability of Arg to displace heme-bound imidazole by a spectral perturbation assay. Upon sequential addition of Arg, we obtained a concentration-dependent high-spin spectral shift that indicated a complete displacement of imidazole from the heme of wild type nNOS and the W409 mutants (Fig. 3). The apparent Arg K s values for wild
304
S. Adak, D. J. Stuehr / Journal of Inorganic Biochemistry 83 (2001) 301 – 308
Fig. 4. Steady-state NO synthesis and NADPH oxidation in wild type and Trp-409 mutants. Turnover number (k cat ) is expressed as mole of product formed (or NADPH oxidized) per mole of heme per min. Each value represents the mean6S.D. of three to six measurements. Rate measurements were taken over the first 3 min in reactions run at 258C as indicated under ‘Materials and methods’.
or electron transfer from the reductase domain to the oxygenase domain of nNOS.
3.3. H2 O2 -supported NOHA oxidation and cytochrome c reduction Fig. 3. Light absorbance spectra showing Arg displacement of imidazole from the heme. Proteins were diluted in cuvettes containing 40 mM EPPS buffer, pH 7.6, containing 5% glycerol (v / v) and 1 mM DTT, 20 mM H4B, and 10 mM imidazole. Light absorbance spectra were recorded at 258C prior to and 15 min after each sequential addition of concentrated Arg solution. The dotted lines indicate the initial imidazole-bound enzyme spectra and sequential solid lines are indicated for 1, 2, 10, 20, 30 and 50 mM Arg respectively.
type, W409F and W409Y in the presence of 10 mM imidazole and 20 mM H 4 B were derived by doublereciprocal analysis and were 55, 60 and 68 mM respectively. We conclude that Arg binding was not significantly altered by mutation of Trp 409 to Tyr or Phe.
Table 1 compares the catalytic turnover numbers of wild-type nNOS and the Trp-409 mutants with regard to H 2 O 2 supported-nitrite formation from NOHArg, and NADPH-dependent cytochrome c reduction in the presence of bound CaM. The turnover numbers for nitrite formation and cytochrome c reduction by the mutant proteins were equivalent to wild type nNOS [28]. The H 2 O 2 / NOHArg reaction occurs without electrons from the reductase domain and is not limited by heme-NO complex formation. Thus, our observing equivalent activities in this assay suggest that mutant hyperactivity during normal NO synthesis may involve increased electron transfer to the heme or a change in the dynamics of heme–NO complex formation.
3.2. NO synthesis and NADPH oxidation Fig. 4 compares catalytic turnover numbers of wild-type nNOS and the Trp-409 mutants with regard to NO synthesis form Arg and corresponding NADPH oxidation. The W409F and W409Y mutants had 3- and 1.8-fold faster rates of NO synthesis from Arg and NADPH oxidation compared with wild type, respectively [28]. The NADPH stoichiometry values are close to the theoretical minimum of 1.5 NADPH oxidized per NO formed [1,2], suggesting tight coupling between NADPH oxidation and NO synthesis in the mutants. These results suggested that Trp-409 mutations might affect catalysis by the oxygenase domain
Table 1 Analysis of oxygenase domain- and reductase domain-specific catalysis a Enzyme
H 2 O 2 supported NOHA oxidation (min 21 )
NADPH dependent cytochrome c reduction (min 21 )
nNOS W409F W409Y
17.662.0 16.061.5 13.061.6
70006220 50006200 55466300
a Turnover number is expressed as mole of product formation per mole of protein per min. Each value represents the mean and S.D. for two protein preparations each assayed in triplicate. Measurements were carried out at 258C. The details are described under ‘Materials and methods’.
S. Adak, D. J. Stuehr / Journal of Inorganic Biochemistry 83 (2001) 301 – 308
305
3.4. NADPH-dependent heme reduction We compared NADPH oxidation rates among wild type nNOS and W409 mutants under a number of different conditions [28]. The rates of NADPH oxidation by CaMbound nNOS in the absence of H 4 B and Arg, or only Arg, are higher than mutant proteins. However, in the presence of H 4 B and Arg the relationship is reversed, with NADPH oxidation rates being higher in the mutants compared to wild-type. Addition of the heme reduction inhibitor nitro-LArg methyl ester [31] to all of the H 4 B -bound proteins decreased their NADPH oxidation rates to a basal level seen for the nNOS reductase domain alone, indicating that any additional NADPH oxidation above this value was associated with heme reduction. Together, these results suggested that electron transfer from reductase domain to heme is actually slower in the two mutants than in wild type nNOS in the absence of NO synthesis. Abu-Soud and Stuehr [32] first reported that adding Arg to H 4 B-saturated nNOS lowered the rate of NADPH oxidation, which we subsequently proposed was due to heme-NO complex formation [8]. Here we utilized agmatine, a substrate analog that binds to nNOS without NO synthesis, to further compare electron flux through the mutants and wild-type in the absence of NO synthesis [28]. NADPH oxidation rates for the CaM-bound Trp-409 mutants were slower than wild type nNOS in the presence of agmatine and H4B. Together, these results indicated that electron flux through the heme is actually slower in the Trp-409 mutants under all conditions except when NO synthesis is taking place. To test if Trp-409 mutations affect the extent of heme reduction, we compared NADPH-dependent heme iron reduction under anaerobic conditions. Heme iron reduction in the presence of H 4 B and Arg was followed as a buildup of the ferrous enzyme, whose Soret peak absorbs maximally at 412 nm [33]. As shown in Fig. 5, heme reduction in wild type nNOS was complete under this condition and gave a peak at 412 nm, as previously reported [33]. Heme reduction in Trp-409 mutants was not complete under the same condition, and the equilibrium reached between ferric and ferrous heme was shifted toward ferric in both mutants compared with wild type. The percentage of NADPHdependent heme iron reduction was determined relative to complete reduction achieved by adding dithionite to the sample at the end of each experiment (not shown). Considering that heme reduction is proportional to the absorbance decrease at 400 nm, W409F and W409Y achieved 40% and 25% heme reduction, respectively. We next examined if the Trp-409 mutants would display different rates of heme reduction. Because these mutants form unstable CO complexes (data not shown), CO binding could not be used to follow heme reduction [25]. We instead followed reduction of their ferric heme–NO complexes at 540 nm in the stopped flow instrument (Fig. 6). Heme reduction was triggered at 108C by rapid mixing NADPH-reduced enzymes with excess Ca 21 . In wild-type
Fig. 5. NADPH-dependent heme iron reduction in wild type and Trp-409 mutant proteins. Experiments were done using excess NADPH under anaerobic conditions. The solid line indicates ferric nNOS diluted to 3 mM in 1 ml of 40 mM EPPS buffer pH 7.6, containing 6 mM CaM, 0.9 mM EDTA, 1 mM DTT, 1 mM Arg and 10 mM H4B. The dotted line was recorded after adding 20 mM of deoxygenated NADPH to cause flavin reduction. The dashed line was recorded after adding 1.2 mM Ca 21 to trigger CaM binding and heme reduction. Spectra were recorded after reaching equilibrium.
nNOS, reduction of the ferric heme–NO complex proceeded at 3 s 21 , similar to the rate observed (3.4 s 21 ) when using CO [25]. Heme reduction in the W409F and W409Y mutants was slower than in wild type (Table 2). Thus, both mutations decreased the extent and rate of heme iron reduction. This can explain why there is slower NADPH oxidation in the mutants when CaM is bound but NO synthesis is not taking place. The molecular basis of the effect remains to be determined. However, structural and resonance Raman data [34,35] predict the indole nitrogen of Trp-409 forms a hydrogen bond with the cysteine thiolate heme ligand (Fig. 1), and loss of this hydrogen bond by mutation to Phe or Tyr should increase the negative charge density on the cysteine thiolate and thus lower the midpoint potential of the nNOS heme iron.
3.5. Ferrous–NO complex formation under steady-state catalysis We compared heme–NO complex formation in the Trp-
306
S. Adak, D. J. Stuehr / Journal of Inorganic Biochemistry 83 (2001) 301 – 308 Table 2 Observed rate constants for NADPH-dependent heme iron reduction a Enzyme
Heme iron reduction k obs (s 21 )
nNOS W409F W409Y
3.0 1.8 0.7
a
Experiments were carried out under anaerobic conditions at 108C as described under ‘Materials and methods’. The NO bound ferric enzyme was mixed with 5 mM Ca 21 in presence of excess NADPH and calmodulin. Heme iron reduction was measured by conversion of ferric– NO to ferrous–NO at 540 nm. The data were best fit to single exponential equations in all cases to generate pseudo first-order rate constants. The values are the average obtained with two nNOS preparations.
reported earlier [8]. Under the same conditions, the Trp409 mutants had much less heme–NO complex formation, and their Soret difference peaks were blue-shifted compared to wild-type. We have also compared the kinetics of heme–NO complex formation and simultaneous NADPH oxidation during the initial phase of NO synthesis using a rapid-scanning stopped-flow spectrophotometer [39]. For wild type and W409F, heme–NO complex buildup was biphasic whereas buildup in W409Y was monophasic. The apparent k 1 values for W409F and W409Y were three and five times slower than wild-type, respectively, and the k 2 value of W409F was five times slower than wild-type. The rate of NADPH oxidation prior to NO complex buildup was about 2- and 4-fold slower in W409F and W409Y compared to wild-type nNOS [39], consistent with the mutations inhibiting heme reduction. However, subsequent heme–NO complex formation caused NADPH oxidation to decrease by 7-fold in wild-type, whereas in the W409 mutants almost no change was observed [39].
Fig. 6. Stopped flow analysis of heme reduction in nNOS and W409 mutants. The upper, middle and lower traces contain traces from wildtype, W409F, and W409Y nNOS, respectively. Anaerobic buffer solutions containing 2 mM enzyme (ferric–NO complex), 80 mM NADPH, 4 mM CaM, 0.5 mM EDTA, 10 mM H4B, and 5 mM Arg was rapid mixed with a solution containing 5 mM CaCl 2 at 108C. Absorbance change at 540 nm was recorded. Traces were fit to a single exponential equation (solid line) to obtain an observed rate constant.
409 mutants and wild type nNOS. Fig. 7 shows difference spectra of wild type and mutants after activating NO synthesis at 158C. The data indicate a majority (|70%) of wild type nNOS converted to the ferrous–NO complex as
Fig. 7. Spectral comparison of NO complex formation in wild-type and mutant nNOS during steady-state NO synthesis at 158C. The difference spectra were obtained by subtracting an initial spectrum of NADPHreduced, CaM-free nNOS in the presence of 1 mM Arg and 20 mM H 4 B from a spectrum collected after initiating steady-state NO synthesis with Ca 21 addition. The dashed, dotted, and solid lines indicate difference spectra for wild type, W409F, and W409Y, respectively. The absorbance decrease below 400 nm indicates NADPH depletion due to NO synthesis.
S. Adak, D. J. Stuehr / Journal of Inorganic Biochemistry 83 (2001) 301 – 308
This left the mutants with a faster rate of NADPH oxidation (and NO synthesis) than wild type, consistent with their being hyperactive in steady state measurements.
307
We have begun to investigate why less heme–NO complex builds up during NO synthesis in the Trp-409 mutants. Besides their slower k cat , possibilities include: (a) lower affinity for NO; (b) conversion from a 6- to 5coordinate ferrous–NO complex [36]; and (c) a faster decay of the heme–NO complex. We measured rates of NO binding to ferrous mutants and wild type nNOS and found that the relationships between binding rates and NO concentration were nearly identical in magnitude and concentration dependence [39]. This suggests mutant NO binding affinities are similar to wild-type nNOS. We studied the stability of the six-coordinate ferrous–NO complexes. Fig. 8 shows spectra obtained after mixing anaerobic ferrous enzymes with an equal volume of NO saturated buffer at 108C in the presence of Arg and H 4 B. In wild type nNOS a stable heme–NO complex formed with Soret absorbance at 436 nm and visible band at |560 nm, as reported previously [9]. In W409F an identical complex was formed initially but partially converted to another species within 3 min after mixing. In W409Y this new species was also formed at the outset and became the predominant species with Soret band at 417 nm and visible bands at 530 and 570 nm. The 417 nm NO complex clearly differs from the 5-coordinate ferrous–NO complex of nNOS which has a Soret absorbance at 395 nm [36], but is similar to ferrous–NO complexes of other heme proteins that contain weakened trans-axial bonds [37,38]. We
suspect an increase in cysteine thiolate electronegativity from the loss of the Trp-409 hydrogen bond may be responsible for these changes in heme–NO complex character among the mutants. Conversion to a 417 nm complex may also partially explain why we see less 436 nm heme–NO complex buildup in the mutants during steady state NO synthesis, and why their difference spectra are blue-shifted (see Fig. 7). We next compared ferrous heme–NO complex reaction with O 2 [39]. A faster reaction would speed the return of mutant enzyme molecules to the active cycle and help explain why they have less heme–NO complex buildup. Anaerobic ferrous–NO enzymes were rapidly mixed with an air-saturated solution. In all cases we saw a monophasic absorbance decrease at 436 nm (ferrous–NO complex decay) and a monophasic absorbance increase at 393 nm (ferric enzyme formation) with time. The observed rate constants are compared in Fig. 9. From these data it is apparent that decay of the heme-NO complex was concomitant with generation of ferric enzyme. Importantly, oxidation of the ferrous–NO complex was about seven times faster in the Trp-409 mutants compared to wild type nNOS. Based on all of our results, we propose that W409 mutants have less heme–NO complex buildup due to a combined effect of slower NO complex formation (due to slower heme reduction) and a faster reaction of their ferrous–NO complexes with O 2 . That both of these changes are mediated by mutation of a proximal Trp that hydrogen bonds with the thiolate heme ligand (the mutations weaken or prevent hydrogen bonding) is an intriguing possibility that deserves further study. At this point, our work implies that NOSs evolved a unique heme proximal environment in part to fulfill their special func-
Fig. 8. Spectrum of ferrous–NO complexes of Trp-409 mutants and wild-type nNOS. Spectra were recorded at 108C after mixing a buffered solution of dithionite-reduced nNOSox proteins (3 mM) containing H4B and Arg with an anaerobic NO-saturated buffer solution. The solid, dashed, and dotted lines indicate spectra obtained at equilibrium for ferrous–NO complexes of wild type, W409F and W409Y enzyme.
Fig. 9. Observed rate constants for reaction of ferrous–NO complexes with O 2 . NO-bound ferrous enzymes were mixed with air-saturated buffer at 108C. Ferrous–NO complex decay and buildup of ferric enzyme were followed at 436 and 393 nm, respectively, and fit to a single exponential function.
3.6. Characteristics of heme–NO complexes
S. Adak, D. J. Stuehr / Journal of Inorganic Biochemistry 83 (2001) 301 – 308
308
tion among heme-thiolate enzymes, namely, to synthesize NO.
4. Abbreviations Arg H4B nNOS nNOSox NO NOHArg NOS,
L-arginine
(6R)-5,6,7,8-tetrahydro-L-biopterin rat neuronal NOS oxygenase domain of nNOS nitric oxide N-hydroxy-L-arginine NO synthase
Acknowledgements We sincerely thank members of the Stuehr, Getzoff, and Tainer laboratories for their indispensable help on these projects, and we additionally thank Qian Wang for excellent technical assistance and generating Fig. 1. This work was supported by National Institutes of Health Grant GM51491 to D.J. Stuehr.
References [1] M.A. Marletta, A.R. Hurshman, K.M. Rusche, Curr. Opin. Chem. Biol. 2 (1998) 656–663. [2] D.J. Stuehr, Biochem. Biophys. Acta Bioenerg. 1411 (1999) 217– 230. [3] B.R. Crane, A.S. Arvai, D.K. Ghosh, C. Wu, E.D. Getzoff, D.J. Stuehr, J.A. Tainer, Science 279 (1998) 2121–2126. [4] C.S. Raman, H. Li, P. Martasek, V. Kral, B.S.S. Masters, T.M. Poulos, Cell 95 (1998) 939–950. [5] T.O. Fischmann, A. Hruza, X. DaNiu, J.D. Fossetta, C.A. Lunn, E. Dolphin, A.J. Prongay, R. Paul, D.J. Lundell, S.K. Narula, P.C. Weber, Nat. Struc. Biol 6 (1999) 233–242. [6] D.B. Goodin, M.G. Davidson, J.A. Roe, A.G. Mauk, M. Smith, Biochemistry 30 (1991) 4953–4962. [7] M. Sundaramoorthy, J. Terner, T.L. Poulos, Structure 3 (1995) 1367–1377. [8] H.M. Abu-Soud, J. Wang, D.L. Rousseau, J. Fukuto, L.J. Ignarro, D.J. Stuehr, J. Biol. Chem 270 (1995) 22997–23006. [9] J. Wang, D.L. Rousseau, H.M. Abu-Soud, D.J. Stuehr, Proc. Natl. Acad. Sci. USA 91 (1994) 10512–10516. [10] C.T. Migita, J.C. Salerno, B.S.S. Masters, P. Martasek, K. McMillan, M. Ikeda-Saito, Biochemistry 36 (1997) 10987–10992. [11] A.L. Tsai, V. Berka, P.F. Chen, G. Palmer, J. Biol. Chem. 271 (1996) 32563–32571. [12] A.R. Hurshman, M.A. Marletta, Biochemistry 34 (1995) 5627– 5634.
[13] H.M. Abu-Soud, C. Wu, D.K. Ghosh, D.J. Stuehr, Biochemistry 37 (1998) 3777–3786. [14] D.A. Wink, Y. Osawa, J.F. Darbyshire, C.R. Jones, S.C. Eshenaur, R.W. Nims, Arch. Biochem. Biophys. 300 (1993) 115–123. [15] Y. Shiro, M. Fujii, T. Iizuka, S. Adachi, K. Tsukamoto, K. Nakahara, H. Shoun, J. Biol. Chem. 270 (1995) 1617–1623. [16] D.H. O’Keeffe, R.E. Ebel, J.A. Peterson, J. Biol. Chem. 253 (1978) 3509–3516. [17] M. Tsubaki, A. Hiwatashi, Y. Ichikawa, H. Hori, Biochemistry 26 (1987) 4527–4534. [18] Y.-M. Kim, H.A. Bergonia, C. Muller, B.R. Pitt, W.D. Watkins, J.R. Lancaster, J. Biol. Chem. 270 (1995) 5710–5713. [19] O. Khatsenko, S.S. Gross, A. Rifkind, J.R. Vane, Proc. Natl. Acad. Sci. USA 90 (1993) 11147–11151. [20] M.L. Adams, B. Nock, R. Truong, T.J. Cicero, Life Sci. 50 (1992) PL35–PL40. [21] J. Stadler, J. Trockfeld, W.A. Schmalix, T. Brill, J.R. Siewert, H. Greim, J. Doehmer, Proc. Natl. Acad. Sci. USA 91 (1994) 3559– 3563. [22] H.M. Abu-Soud, D.L. Rousseau, D.J. Stuehr, J. Biol. Chem. 271 (1996) 32515–32518. [23] R.A. Dweik, D. Laskowski, H.M. Abu-soud, F.T. Kaneko, R. Hutte, D.J. Stuehr, S.C. Erzurum, J. Clin. Invest. 101 (1998) 660–666. [24] R.T. Miller, P. Martasek, T. Omura, B.S.S. Masters, Biochem. Biophys. Res. Commun. 265 (1999) 184–188. [25] R. Gachhui, H.M. Abu-Soud, D.K. Ghosh, A. Presta, M.A. Blazing, B. Mayer, S.E. George, D.J. Stuehr, J. Biol. Chem 273 (1998) 5451–5454. [26] C.R. Nishida, P.R. Ortiz de Montellano, J. Biol. Chem 273 (1998) 5566–5571. [27] T. Iwanaga, T. Yamazaki, S. Kominami, Biochemistry 38 (1999) 16629–16635. [28] S. Adak, C. Crooks, Q. Wang, B.R. Crane, J.A. Tainer, E.D. Getzoff, D.J. Stuehr, J. Biol. Chem. 274 (1999) 26907–26911. [29] S. Adak, S. Ghosh, H.M. Abu-Soud, D.J. Stuehr, J. Biol. Chem. 274 (1999) 22313–22320. [30] K. McMillan, B.S.S. Masters, Biochemistry 32 (1993) 9875–9880. [31] A. Presta, A.M. Weber-Main, M.T. Stankovich, D.J. Stuehr, J. Am. Chem. Soc. 120 (1998) 9460–9465. [32] H.M. Abu-Soud, D.J. Stuehr, Proc. Natl. Acad. Sci. USA 90 (1993) 10769–10772. [33] H.M. Abu-Soud, A. Presta, B. Mayer, D.J. Stuehr, Biochemistry 36 (1997) 10811–10816. [34] B.R. Crane, A.S. Arvai, R. Gachhui, C. Wu, D.K. Ghosh, E.D. Getzoff, D.J. Stuehr, J.A. Tainer, Science 278 (1997) 425–431. [35] J. Wang, D.J. Stuehr, D.L. Rousseau, Biochemistry 36 (1997) 4595– 4606. [36] L. Huang, H.M. Abu-Soud, R. Hille, D.J. Stuehr, Biochemistry 38 (1999) 1912–1920. [37] P. Ascenzi, M. Brunori, M. Coletta, A. Desideri, Biochem. J. 258 (1989) 473–478. [38] T. Yoshimura, T. Ozaki, Arch. Biochem. Biophys. 229 (1984) 126–135. [39] S. Adak et al., J. Biol. Chem. 275 (2000) 17434–17439.
A proximal tryptophan in NO synthase controls activity by a novel mechanism Subrata Adak, Dennis J. Stuehr* Department of Immunology, Lerner Research Institute, Cleveland Clinic Foundation, 9500 Euclid Avenue, Cleveland, OH 44195, USA Received 15 March 2000; received in revised form 15 June 2000; accepted 17 July 2000
Abstract The heme of neuronal nitric oxide synthase (nNOS) participates in O 2 activation but also binds self-generated NO, resulting in reversible feedback inhibition. We utilized mutagenesis to investigate if a conserved tryptophan residue (Trp409), which engages in p-stacking with the heme and hydrogen bonds to its axial cysteine ligand, helps control catalysis and regulation by NO. Mutants W409F and W409Y were hyperactive regarding NO synthesis without affecting cytochrome c reduction, reductase-independent N-hydroxyarginine oxidation, or Arg and tetrahydrobiopterin binding. In the absence of Arg electron flux through the heme was slower in the W409 mutants than in wild-type. However, less NO complex accumulated during NO synthesis by the mutants. To understand the mechanism, we compared the kinetics of heme–NO complex formation, rate of heme reduction, k cat prior to and after NO complex formation, NO binding affinity, NO complex stability, and its reaction with O 2 . During the initial phase of NO synthesis, heme–NO complex formation was three and five times slower in W409F and W409Y, which corresponded to a slower heme reduction. NO complex formation inhibited wild-type turnover 7-fold but reduced mutant turnover less than 2-fold, giving mutants higher steady-state activities. NO binding kinetics were similar among mutants and wild type, although mutants also formed a 417 nm ferrous-NO complex. Oxidation of ferrous-NO complex was seven times faster in mutants than in wild type. We conclude that mutant hyperactivity primarily derives from slower heme reduction and faster oxidation of the heme–NO complex by O 2 . In this way Trp409 mutations minimize NO feedback inhibition by limiting buildup of the ferrous–NO complex during the steady state. Conservation of W409 among NOS suggests that this proximal Trp may regulate NO feedback inhibition and is important for enzyme physiologic function. 2001 Elsevier Science B.V. All rights reserved. Keywords: Neuronal notric oxide synthase; heme–NO complex; ferrous–NO complex
1. Introduction Nitric oxide (NO) is synthesized from Arg by hemecontaining monooxygenases known as NO synthases (NOSs). The reaction requires NADPH and O 2 as cosubstrates, and occurs in two steps with generation of NhydroxyArg (NOHArg) as an intermediate [1,2]. Calmodulin binding activates NO synthesis in the NOSs by triggering heme reduction by the reductase domain. Although NOSs and cytochrome P-450s both utilize a cysteine thiolate-ligated heme to bind and activate O 2 , they have quite different heme site and tertiary structures [3–5]. In neuronal NOS (nNOS) (Fig. 1), the essential cofactor 6R-tetrahydrobiopterin (H4B) binds at the NOS heme edge perpendicular to its plane. Two conserved aromatic res*Corresponding author. Tel.: 11-216-445-6950; fax: 11-216-4449329. E-mail address: [email protected] (D.J. Stuehr).
idues Phe-584 and Trp-409 stack with the heme on its distal and proximal sides, respectively. This aromatic stacking is absent in cytochrome P450s but is present in peroxidases and influences their electronic and catalytic properties [6]. The NOS crystal structures predict that the indole nitrogen of Trp-409 forms a strong hydrogen bond with the cysteine thiolate heme ligand in NOS. In cytochrome P450s, hydrogen bonding to the coordinating thiolate is often minimal, but stronger hydrogen bonding is present in the thiolate-ligated hemeprotein chloroperoxidase [7]. Thus, it is important to determine how these structural features impact NO synthesis in NOS. The NOS heme also functions under unique circumstances. Consider that during aerobic steady-state NO synthesis approximately 60–90% of nNOS is present as a ferrous–NO complex due to binding self-generated NO [8]. Ferric and ferrous NOSs both form stable six-coordinate NO complexes in the presence of H4B or substrate and regenerate fully active ferric NOS upon breakdown of
0162-0134 / 01 / $ – see front matter 2001 Elsevier Science B.V. All rights reserved. PII: S0162-0134( 00 )00176-8
302
S. Adak, D. J. Stuehr / Journal of Inorganic Biochemistry 83 (2001) 301 – 308
Fig. 1. Some unique aspects of the nNOS heme environment. The figure is based on the mouse iNOS oxygenase domain crystal structure (3). The view looks up from the proximal side of the heme to highlight aromatic stacking between the heme and F584 and W409 in nNOS. Thin lines indicate hydrogen bonds between a W409 indole nitrogen and the cysteine thiolate, and between H4B and a heme propionate.
the complex [9–13].This distinguishes NOSs from the cytochrome P450s which typically form unstable ferrous NO complexes that irreversibly inactivate the enzyme [14,15]. In cytochrome P450s, formation of a five-coordinate ferrous–NO complex typically occurs after NO binds and leads to formation of cytochrome P420 upon air oxidation, loss of heme in some cases, and irreversible loss of activity [14–18]. This inactivation links NO synthesis with down-regulation of tissue P450 activity during inflammation [19–21]. In NOSs, reversible NO complex formation allows NO to act as a feedback inhibitor of activity during the steady state and makes the rate of NO synthesis by the enzyme proportional to the O 2 concentration across the physiologic range [22,23]. We have mostly studied heme–NO complex formation in nNOS. After initiating NO synthesis, nNOS molecules quickly partition between a catalytically active form and an inactive ferrous–NO complex [8]. This creates two cycles that have their own rate-limiting step and together determine the observed rate of NO synthesis during the steady-state (Fig. 2). The rate-limiting step in the active cycle that generates NO has been suggested to be either heme reduction or product release [24–27]. If a nNOS molecule forms a ferrous–NO complex, it is placed in the inactive cycle and the rate-limiting step becomes the O 2 dependent decay of the heme–NO complex [22]. Because NO binding to the heme regulates NOS catalysis, it is important to identify the structural features that allow NOS to function in this capacity. To address this issue, we have used site-directed mutagenesis to investigate what role Trp-409 has in regulating nNOS activity. Recently we found that this proximal tryptophan controls its activity by regulating NO feedback inhibition [28]. Mutation of Trp-409 to Phe or Tyr in nNOS resulted in less heme–NO complex buildup and
Fig. 2. Proposed mechanism for autoinhibition of nNOS by NO under steady-state catalysis. nNOS cycles between an active, NO-generating cycle (Left) and an inactive cycle (Right) that involves formation and O 2 -dependent decay of a ferrous–NO complex. Product NO may bind to ferric or ferrous nNOS which partitions nNOS molecules to the inactive cycle.
hyperactivity with regard to NO synthesis. This report summarizes our past and more recent studies on W409 mutants that show hyperactivity results from changes in electron flux and heme–NO complex oxidation.
2. Materials and methods
2.1. Materials Oxygen and NO gas were purchased from Liquid Carbonic Company and Matheson Inc., respectively. All other reagents and materials were obtained from Sigma or sources as previously reported [28,29].
2.2. Molecular biology W409 mutations were made in the oxygenase domain (nNOSox) (a.k.a. heme domain) as described for full-length nNOS [28] using the same primers and the Quick Change PCR in vitro Mutagenesis Kit from Stratagene. Mutations were confirmed by DNA sequencing at the Cleveland Clinic core facility.
2.3. Protein expression and purification Wild-type and W409 mutant nNOS proteins (oxygenase domain and full length) were expressed in Escherichia coli and purified as described previously [28,29]. UV-visible spectra was recorded on a Hitachi U3110 Spectrophotome-
S. Adak, D. J. Stuehr / Journal of Inorganic Biochemistry 83 (2001) 301 – 308
ter in the absence or presence of 20 mM H4B and 1 mM Arg. The ferrous–CO adduct absorbing at 444 nm was used to quantitate the heme protein content using an extinction coefficient of 74 mM 21 cm 21 (A444–A500).
2.4. NO synthesis, NADPH oxidation and cytochrome c reduction
303
EDTA, 3 mM calmodulin (CaM), 200 mM DTT, 20 mM H 4 B, 80 mM NADPH, and 1 mM Arg, final volume 1 ml. Reactions were started by adding 1.2 mM Ca 21 to cause CaM binding and monitored by wavelength scanning at 158C. The concentration of ferrous–NO complex formed during NO synthesis was estimated from absorbance change at 436 nm using an extinction coefficient of 49 800 M 21 cm 21 [8].
The initial rate of NO synthesis, NADPH oxidation, or cytochrome c reduction by wild type and mutants were quantitated at 258C as described earlier [28,29].
2.8. Arg binding
2.5. H2 O2 -dependent NOHA oxidation
Arg binding affinity was studied at 308C by perturbation difference spectroscopy with imidazole according to methods described previously [30].
H 2 O 2 -dependent nNOS oxidation of NOHA to nitrite was assayed in 96-well microplates at 258C as described previously [28] with modification. The assay volume was 100 ml and contained 40 mM 4-(-2-hydroxyethyl)-1piperazinepropanesulfonic acid (EPPS) buffer pH-7.6, 250 nM nNOS or mutants, 1 mM N v hydroxy-L-arginine (NOHA), 1 mM DTT, 25 U / ml SOD, 0.5 mM EDTA and 4 mM (6R)-5,6,7,8-tetrahydro-L-biopterin (H 4 B). Reactions were initiated by adding 30 mM H 2 O 2 , and stopped after 10 min by adding 1300 U of catalase. Nitrite was detected at 550 nm using the Griess reagent (100 ml) and quantitated based on nitrite standards.
2.9. Characteristics and kinetics of NO binding Spectra were collected with the rapid scanning diode array detector (Hi-Tech, model SF-61). Anaerobic solutions containing wild type or mutant nNOS (3 mM), 100 mM EPPS buffer pH 7.6, 1 mM Arg, 10 mM H 4 B, 0.2 mM DTT were pre-reduced with excess dithionite (100 mM), and then rapid mixed with NO-saturated buffer at 108C. To study NO binding kinetics, the initial NO concentration was varied by diluting the saturated solution with anaerobic buffer, and absorbance change was recorded with a single wavelength detector.
2.6. Heme reduction 2.10. Reaction of ferrous–NO complexes with oxygen Optical spectra were recorded on a Hitachi 3010 UVvisible spectrophotometer at 158C. Anaerobic spectra were recorded using septum-sealed quartz cuvettes that could be attached through a quick-fit joint to a vacuum system. The samples were made anaerobic by alternating cycles of evacuation and refilling with nitrogen. The cuvette was maintained under positive nitrogen pressure. The ferric nNOS (3 mM) solution, containing 40 mM EPPS buffer pH 7.6, 6 mM CaM, 0.9 mM EDTA, 1 mM DTT, 1 mM Arg and 10 mM H4B were mixed with 20 mM of deoxygenated NADPH to cause flavin reduction. Heme reduction was initiated after adding 1.2 mM Ca 21 to trigger CaM binding and heme reduction. Spectra were recorded after reaching equilibrium. Kinetic measurements were carried out using a Hi-Tech stopped-flow apparatus (model SF-51) equipped for anaerobic work. Anaerobic buffer solutions containing 2 mM ferric enzyme, 100 mM EPPS buffer pH-7.6, 200 mM NO solution, 80 mM NADPH, 4 mM CaM, 0.5 mM EDTA, 10 mM H4B, and 5 mM Arg was rapid mixed with a solution containing 5 mM CaCl 2 at 108C. The heme reduction rate was monitored at 540 nm.
2.7. Heme–NO complex formation during steady-state Subsequently, 0.6 mM nNOS was diluted in an airsaturated 40 mM EPPS buffer, pH 7.6, containing 0.9 mM
Anaerobic solutions of full-length nNOS enzymes (4 mM) containing saturating Arg and H 4 B were reduced with a minimum amount of dithionite and then saturated NO solution was added to give an NO concentration of 0.1 mM. These ferrous–NO complexes were then rapid mixed with air-saturated buffer solutions at 108C. Absorbance change was monitored at 436 or 393 nm using a single wavelength detector. Signal to noise ratios were improved by averaging the 10 individual scans. The time course of absorbance change was best fit to a single exponential equation by use of a nonlinear least-squares method provided by the instrument manufacturer.
3. Results and discussion
3.1. Arginine binding To investigate if mutation of Trp 409 affected substrate binding on the distal side of the heme, we compared the ability of Arg to displace heme-bound imidazole by a spectral perturbation assay. Upon sequential addition of Arg, we obtained a concentration-dependent high-spin spectral shift that indicated a complete displacement of imidazole from the heme of wild type nNOS and the W409 mutants (Fig. 3). The apparent Arg K s values for wild
304
S. Adak, D. J. Stuehr / Journal of Inorganic Biochemistry 83 (2001) 301 – 308
Fig. 4. Steady-state NO synthesis and NADPH oxidation in wild type and Trp-409 mutants. Turnover number (k cat ) is expressed as mole of product formed (or NADPH oxidized) per mole of heme per min. Each value represents the mean6S.D. of three to six measurements. Rate measurements were taken over the first 3 min in reactions run at 258C as indicated under ‘Materials and methods’.
or electron transfer from the reductase domain to the oxygenase domain of nNOS.
3.3. H2 O2 -supported NOHA oxidation and cytochrome c reduction Fig. 3. Light absorbance spectra showing Arg displacement of imidazole from the heme. Proteins were diluted in cuvettes containing 40 mM EPPS buffer, pH 7.6, containing 5% glycerol (v / v) and 1 mM DTT, 20 mM H4B, and 10 mM imidazole. Light absorbance spectra were recorded at 258C prior to and 15 min after each sequential addition of concentrated Arg solution. The dotted lines indicate the initial imidazole-bound enzyme spectra and sequential solid lines are indicated for 1, 2, 10, 20, 30 and 50 mM Arg respectively.
type, W409F and W409Y in the presence of 10 mM imidazole and 20 mM H 4 B were derived by doublereciprocal analysis and were 55, 60 and 68 mM respectively. We conclude that Arg binding was not significantly altered by mutation of Trp 409 to Tyr or Phe.
Table 1 compares the catalytic turnover numbers of wild-type nNOS and the Trp-409 mutants with regard to H 2 O 2 supported-nitrite formation from NOHArg, and NADPH-dependent cytochrome c reduction in the presence of bound CaM. The turnover numbers for nitrite formation and cytochrome c reduction by the mutant proteins were equivalent to wild type nNOS [28]. The H 2 O 2 / NOHArg reaction occurs without electrons from the reductase domain and is not limited by heme-NO complex formation. Thus, our observing equivalent activities in this assay suggest that mutant hyperactivity during normal NO synthesis may involve increased electron transfer to the heme or a change in the dynamics of heme–NO complex formation.
3.2. NO synthesis and NADPH oxidation Fig. 4 compares catalytic turnover numbers of wild-type nNOS and the Trp-409 mutants with regard to NO synthesis form Arg and corresponding NADPH oxidation. The W409F and W409Y mutants had 3- and 1.8-fold faster rates of NO synthesis from Arg and NADPH oxidation compared with wild type, respectively [28]. The NADPH stoichiometry values are close to the theoretical minimum of 1.5 NADPH oxidized per NO formed [1,2], suggesting tight coupling between NADPH oxidation and NO synthesis in the mutants. These results suggested that Trp-409 mutations might affect catalysis by the oxygenase domain
Table 1 Analysis of oxygenase domain- and reductase domain-specific catalysis a Enzyme
H 2 O 2 supported NOHA oxidation (min 21 )
NADPH dependent cytochrome c reduction (min 21 )
nNOS W409F W409Y
17.662.0 16.061.5 13.061.6
70006220 50006200 55466300
a Turnover number is expressed as mole of product formation per mole of protein per min. Each value represents the mean and S.D. for two protein preparations each assayed in triplicate. Measurements were carried out at 258C. The details are described under ‘Materials and methods’.
S. Adak, D. J. Stuehr / Journal of Inorganic Biochemistry 83 (2001) 301 – 308
305
3.4. NADPH-dependent heme reduction We compared NADPH oxidation rates among wild type nNOS and W409 mutants under a number of different conditions [28]. The rates of NADPH oxidation by CaMbound nNOS in the absence of H 4 B and Arg, or only Arg, are higher than mutant proteins. However, in the presence of H 4 B and Arg the relationship is reversed, with NADPH oxidation rates being higher in the mutants compared to wild-type. Addition of the heme reduction inhibitor nitro-LArg methyl ester [31] to all of the H 4 B -bound proteins decreased their NADPH oxidation rates to a basal level seen for the nNOS reductase domain alone, indicating that any additional NADPH oxidation above this value was associated with heme reduction. Together, these results suggested that electron transfer from reductase domain to heme is actually slower in the two mutants than in wild type nNOS in the absence of NO synthesis. Abu-Soud and Stuehr [32] first reported that adding Arg to H 4 B-saturated nNOS lowered the rate of NADPH oxidation, which we subsequently proposed was due to heme-NO complex formation [8]. Here we utilized agmatine, a substrate analog that binds to nNOS without NO synthesis, to further compare electron flux through the mutants and wild-type in the absence of NO synthesis [28]. NADPH oxidation rates for the CaM-bound Trp-409 mutants were slower than wild type nNOS in the presence of agmatine and H4B. Together, these results indicated that electron flux through the heme is actually slower in the Trp-409 mutants under all conditions except when NO synthesis is taking place. To test if Trp-409 mutations affect the extent of heme reduction, we compared NADPH-dependent heme iron reduction under anaerobic conditions. Heme iron reduction in the presence of H 4 B and Arg was followed as a buildup of the ferrous enzyme, whose Soret peak absorbs maximally at 412 nm [33]. As shown in Fig. 5, heme reduction in wild type nNOS was complete under this condition and gave a peak at 412 nm, as previously reported [33]. Heme reduction in Trp-409 mutants was not complete under the same condition, and the equilibrium reached between ferric and ferrous heme was shifted toward ferric in both mutants compared with wild type. The percentage of NADPHdependent heme iron reduction was determined relative to complete reduction achieved by adding dithionite to the sample at the end of each experiment (not shown). Considering that heme reduction is proportional to the absorbance decrease at 400 nm, W409F and W409Y achieved 40% and 25% heme reduction, respectively. We next examined if the Trp-409 mutants would display different rates of heme reduction. Because these mutants form unstable CO complexes (data not shown), CO binding could not be used to follow heme reduction [25]. We instead followed reduction of their ferric heme–NO complexes at 540 nm in the stopped flow instrument (Fig. 6). Heme reduction was triggered at 108C by rapid mixing NADPH-reduced enzymes with excess Ca 21 . In wild-type
Fig. 5. NADPH-dependent heme iron reduction in wild type and Trp-409 mutant proteins. Experiments were done using excess NADPH under anaerobic conditions. The solid line indicates ferric nNOS diluted to 3 mM in 1 ml of 40 mM EPPS buffer pH 7.6, containing 6 mM CaM, 0.9 mM EDTA, 1 mM DTT, 1 mM Arg and 10 mM H4B. The dotted line was recorded after adding 20 mM of deoxygenated NADPH to cause flavin reduction. The dashed line was recorded after adding 1.2 mM Ca 21 to trigger CaM binding and heme reduction. Spectra were recorded after reaching equilibrium.
nNOS, reduction of the ferric heme–NO complex proceeded at 3 s 21 , similar to the rate observed (3.4 s 21 ) when using CO [25]. Heme reduction in the W409F and W409Y mutants was slower than in wild type (Table 2). Thus, both mutations decreased the extent and rate of heme iron reduction. This can explain why there is slower NADPH oxidation in the mutants when CaM is bound but NO synthesis is not taking place. The molecular basis of the effect remains to be determined. However, structural and resonance Raman data [34,35] predict the indole nitrogen of Trp-409 forms a hydrogen bond with the cysteine thiolate heme ligand (Fig. 1), and loss of this hydrogen bond by mutation to Phe or Tyr should increase the negative charge density on the cysteine thiolate and thus lower the midpoint potential of the nNOS heme iron.
3.5. Ferrous–NO complex formation under steady-state catalysis We compared heme–NO complex formation in the Trp-
306
S. Adak, D. J. Stuehr / Journal of Inorganic Biochemistry 83 (2001) 301 – 308 Table 2 Observed rate constants for NADPH-dependent heme iron reduction a Enzyme
Heme iron reduction k obs (s 21 )
nNOS W409F W409Y
3.0 1.8 0.7
a
Experiments were carried out under anaerobic conditions at 108C as described under ‘Materials and methods’. The NO bound ferric enzyme was mixed with 5 mM Ca 21 in presence of excess NADPH and calmodulin. Heme iron reduction was measured by conversion of ferric– NO to ferrous–NO at 540 nm. The data were best fit to single exponential equations in all cases to generate pseudo first-order rate constants. The values are the average obtained with two nNOS preparations.
reported earlier [8]. Under the same conditions, the Trp409 mutants had much less heme–NO complex formation, and their Soret difference peaks were blue-shifted compared to wild-type. We have also compared the kinetics of heme–NO complex formation and simultaneous NADPH oxidation during the initial phase of NO synthesis using a rapid-scanning stopped-flow spectrophotometer [39]. For wild type and W409F, heme–NO complex buildup was biphasic whereas buildup in W409Y was monophasic. The apparent k 1 values for W409F and W409Y were three and five times slower than wild-type, respectively, and the k 2 value of W409F was five times slower than wild-type. The rate of NADPH oxidation prior to NO complex buildup was about 2- and 4-fold slower in W409F and W409Y compared to wild-type nNOS [39], consistent with the mutations inhibiting heme reduction. However, subsequent heme–NO complex formation caused NADPH oxidation to decrease by 7-fold in wild-type, whereas in the W409 mutants almost no change was observed [39].
Fig. 6. Stopped flow analysis of heme reduction in nNOS and W409 mutants. The upper, middle and lower traces contain traces from wildtype, W409F, and W409Y nNOS, respectively. Anaerobic buffer solutions containing 2 mM enzyme (ferric–NO complex), 80 mM NADPH, 4 mM CaM, 0.5 mM EDTA, 10 mM H4B, and 5 mM Arg was rapid mixed with a solution containing 5 mM CaCl 2 at 108C. Absorbance change at 540 nm was recorded. Traces were fit to a single exponential equation (solid line) to obtain an observed rate constant.
409 mutants and wild type nNOS. Fig. 7 shows difference spectra of wild type and mutants after activating NO synthesis at 158C. The data indicate a majority (|70%) of wild type nNOS converted to the ferrous–NO complex as
Fig. 7. Spectral comparison of NO complex formation in wild-type and mutant nNOS during steady-state NO synthesis at 158C. The difference spectra were obtained by subtracting an initial spectrum of NADPHreduced, CaM-free nNOS in the presence of 1 mM Arg and 20 mM H 4 B from a spectrum collected after initiating steady-state NO synthesis with Ca 21 addition. The dashed, dotted, and solid lines indicate difference spectra for wild type, W409F, and W409Y, respectively. The absorbance decrease below 400 nm indicates NADPH depletion due to NO synthesis.
S. Adak, D. J. Stuehr / Journal of Inorganic Biochemistry 83 (2001) 301 – 308
This left the mutants with a faster rate of NADPH oxidation (and NO synthesis) than wild type, consistent with their being hyperactive in steady state measurements.
307
We have begun to investigate why less heme–NO complex builds up during NO synthesis in the Trp-409 mutants. Besides their slower k cat , possibilities include: (a) lower affinity for NO; (b) conversion from a 6- to 5coordinate ferrous–NO complex [36]; and (c) a faster decay of the heme–NO complex. We measured rates of NO binding to ferrous mutants and wild type nNOS and found that the relationships between binding rates and NO concentration were nearly identical in magnitude and concentration dependence [39]. This suggests mutant NO binding affinities are similar to wild-type nNOS. We studied the stability of the six-coordinate ferrous–NO complexes. Fig. 8 shows spectra obtained after mixing anaerobic ferrous enzymes with an equal volume of NO saturated buffer at 108C in the presence of Arg and H 4 B. In wild type nNOS a stable heme–NO complex formed with Soret absorbance at 436 nm and visible band at |560 nm, as reported previously [9]. In W409F an identical complex was formed initially but partially converted to another species within 3 min after mixing. In W409Y this new species was also formed at the outset and became the predominant species with Soret band at 417 nm and visible bands at 530 and 570 nm. The 417 nm NO complex clearly differs from the 5-coordinate ferrous–NO complex of nNOS which has a Soret absorbance at 395 nm [36], but is similar to ferrous–NO complexes of other heme proteins that contain weakened trans-axial bonds [37,38]. We
suspect an increase in cysteine thiolate electronegativity from the loss of the Trp-409 hydrogen bond may be responsible for these changes in heme–NO complex character among the mutants. Conversion to a 417 nm complex may also partially explain why we see less 436 nm heme–NO complex buildup in the mutants during steady state NO synthesis, and why their difference spectra are blue-shifted (see Fig. 7). We next compared ferrous heme–NO complex reaction with O 2 [39]. A faster reaction would speed the return of mutant enzyme molecules to the active cycle and help explain why they have less heme–NO complex buildup. Anaerobic ferrous–NO enzymes were rapidly mixed with an air-saturated solution. In all cases we saw a monophasic absorbance decrease at 436 nm (ferrous–NO complex decay) and a monophasic absorbance increase at 393 nm (ferric enzyme formation) with time. The observed rate constants are compared in Fig. 9. From these data it is apparent that decay of the heme-NO complex was concomitant with generation of ferric enzyme. Importantly, oxidation of the ferrous–NO complex was about seven times faster in the Trp-409 mutants compared to wild type nNOS. Based on all of our results, we propose that W409 mutants have less heme–NO complex buildup due to a combined effect of slower NO complex formation (due to slower heme reduction) and a faster reaction of their ferrous–NO complexes with O 2 . That both of these changes are mediated by mutation of a proximal Trp that hydrogen bonds with the thiolate heme ligand (the mutations weaken or prevent hydrogen bonding) is an intriguing possibility that deserves further study. At this point, our work implies that NOSs evolved a unique heme proximal environment in part to fulfill their special func-
Fig. 8. Spectrum of ferrous–NO complexes of Trp-409 mutants and wild-type nNOS. Spectra were recorded at 108C after mixing a buffered solution of dithionite-reduced nNOSox proteins (3 mM) containing H4B and Arg with an anaerobic NO-saturated buffer solution. The solid, dashed, and dotted lines indicate spectra obtained at equilibrium for ferrous–NO complexes of wild type, W409F and W409Y enzyme.
Fig. 9. Observed rate constants for reaction of ferrous–NO complexes with O 2 . NO-bound ferrous enzymes were mixed with air-saturated buffer at 108C. Ferrous–NO complex decay and buildup of ferric enzyme were followed at 436 and 393 nm, respectively, and fit to a single exponential function.
3.6. Characteristics of heme–NO complexes
S. Adak, D. J. Stuehr / Journal of Inorganic Biochemistry 83 (2001) 301 – 308
308
tion among heme-thiolate enzymes, namely, to synthesize NO.
4. Abbreviations Arg H4B nNOS nNOSox NO NOHArg NOS,
L-arginine
(6R)-5,6,7,8-tetrahydro-L-biopterin rat neuronal NOS oxygenase domain of nNOS nitric oxide N-hydroxy-L-arginine NO synthase
Acknowledgements We sincerely thank members of the Stuehr, Getzoff, and Tainer laboratories for their indispensable help on these projects, and we additionally thank Qian Wang for excellent technical assistance and generating Fig. 1. This work was supported by National Institutes of Health Grant GM51491 to D.J. Stuehr.
References [1] M.A. Marletta, A.R. Hurshman, K.M. Rusche, Curr. Opin. Chem. Biol. 2 (1998) 656–663. [2] D.J. Stuehr, Biochem. Biophys. Acta Bioenerg. 1411 (1999) 217– 230. [3] B.R. Crane, A.S. Arvai, D.K. Ghosh, C. Wu, E.D. Getzoff, D.J. Stuehr, J.A. Tainer, Science 279 (1998) 2121–2126. [4] C.S. Raman, H. Li, P. Martasek, V. Kral, B.S.S. Masters, T.M. Poulos, Cell 95 (1998) 939–950. [5] T.O. Fischmann, A. Hruza, X. DaNiu, J.D. Fossetta, C.A. Lunn, E. Dolphin, A.J. Prongay, R. Paul, D.J. Lundell, S.K. Narula, P.C. Weber, Nat. Struc. Biol 6 (1999) 233–242. [6] D.B. Goodin, M.G. Davidson, J.A. Roe, A.G. Mauk, M. Smith, Biochemistry 30 (1991) 4953–4962. [7] M. Sundaramoorthy, J. Terner, T.L. Poulos, Structure 3 (1995) 1367–1377. [8] H.M. Abu-Soud, J. Wang, D.L. Rousseau, J. Fukuto, L.J. Ignarro, D.J. Stuehr, J. Biol. Chem 270 (1995) 22997–23006. [9] J. Wang, D.L. Rousseau, H.M. Abu-Soud, D.J. Stuehr, Proc. Natl. Acad. Sci. USA 91 (1994) 10512–10516. [10] C.T. Migita, J.C. Salerno, B.S.S. Masters, P. Martasek, K. McMillan, M. Ikeda-Saito, Biochemistry 36 (1997) 10987–10992. [11] A.L. Tsai, V. Berka, P.F. Chen, G. Palmer, J. Biol. Chem. 271 (1996) 32563–32571. [12] A.R. Hurshman, M.A. Marletta, Biochemistry 34 (1995) 5627– 5634.
[13] H.M. Abu-Soud, C. Wu, D.K. Ghosh, D.J. Stuehr, Biochemistry 37 (1998) 3777–3786. [14] D.A. Wink, Y. Osawa, J.F. Darbyshire, C.R. Jones, S.C. Eshenaur, R.W. Nims, Arch. Biochem. Biophys. 300 (1993) 115–123. [15] Y. Shiro, M. Fujii, T. Iizuka, S. Adachi, K. Tsukamoto, K. Nakahara, H. Shoun, J. Biol. Chem. 270 (1995) 1617–1623. [16] D.H. O’Keeffe, R.E. Ebel, J.A. Peterson, J. Biol. Chem. 253 (1978) 3509–3516. [17] M. Tsubaki, A. Hiwatashi, Y. Ichikawa, H. Hori, Biochemistry 26 (1987) 4527–4534. [18] Y.-M. Kim, H.A. Bergonia, C. Muller, B.R. Pitt, W.D. Watkins, J.R. Lancaster, J. Biol. Chem. 270 (1995) 5710–5713. [19] O. Khatsenko, S.S. Gross, A. Rifkind, J.R. Vane, Proc. Natl. Acad. Sci. USA 90 (1993) 11147–11151. [20] M.L. Adams, B. Nock, R. Truong, T.J. Cicero, Life Sci. 50 (1992) PL35–PL40. [21] J. Stadler, J. Trockfeld, W.A. Schmalix, T. Brill, J.R. Siewert, H. Greim, J. Doehmer, Proc. Natl. Acad. Sci. USA 91 (1994) 3559– 3563. [22] H.M. Abu-Soud, D.L. Rousseau, D.J. Stuehr, J. Biol. Chem. 271 (1996) 32515–32518. [23] R.A. Dweik, D. Laskowski, H.M. Abu-soud, F.T. Kaneko, R. Hutte, D.J. Stuehr, S.C. Erzurum, J. Clin. Invest. 101 (1998) 660–666. [24] R.T. Miller, P. Martasek, T. Omura, B.S.S. Masters, Biochem. Biophys. Res. Commun. 265 (1999) 184–188. [25] R. Gachhui, H.M. Abu-Soud, D.K. Ghosh, A. Presta, M.A. Blazing, B. Mayer, S.E. George, D.J. Stuehr, J. Biol. Chem 273 (1998) 5451–5454. [26] C.R. Nishida, P.R. Ortiz de Montellano, J. Biol. Chem 273 (1998) 5566–5571. [27] T. Iwanaga, T. Yamazaki, S. Kominami, Biochemistry 38 (1999) 16629–16635. [28] S. Adak, C. Crooks, Q. Wang, B.R. Crane, J.A. Tainer, E.D. Getzoff, D.J. Stuehr, J. Biol. Chem. 274 (1999) 26907–26911. [29] S. Adak, S. Ghosh, H.M. Abu-Soud, D.J. Stuehr, J. Biol. Chem. 274 (1999) 22313–22320. [30] K. McMillan, B.S.S. Masters, Biochemistry 32 (1993) 9875–9880. [31] A. Presta, A.M. Weber-Main, M.T. Stankovich, D.J. Stuehr, J. Am. Chem. Soc. 120 (1998) 9460–9465. [32] H.M. Abu-Soud, D.J. Stuehr, Proc. Natl. Acad. Sci. USA 90 (1993) 10769–10772. [33] H.M. Abu-Soud, A. Presta, B. Mayer, D.J. Stuehr, Biochemistry 36 (1997) 10811–10816. [34] B.R. Crane, A.S. Arvai, R. Gachhui, C. Wu, D.K. Ghosh, E.D. Getzoff, D.J. Stuehr, J.A. Tainer, Science 278 (1997) 425–431. [35] J. Wang, D.J. Stuehr, D.L. Rousseau, Biochemistry 36 (1997) 4595– 4606. [36] L. Huang, H.M. Abu-Soud, R. Hille, D.J. Stuehr, Biochemistry 38 (1999) 1912–1920. [37] P. Ascenzi, M. Brunori, M. Coletta, A. Desideri, Biochem. J. 258 (1989) 473–478. [38] T. Yoshimura, T. Ozaki, Arch. Biochem. Biophys. 229 (1984) 126–135. [39] S. Adak et al., J. Biol. Chem. 275 (2000) 17434–17439.