Kinetic basis for the activation of human cyclooxygenase-2 rather than cyclooxygenase-1 by nitric oxide

Jie Qiaoa, Lixin Maa, Justine Rothb, Yamin Li*c and Yi Liu*a
aHubei Collaborative Innovation Center for Green Transformation of Bio-resources, Hubei Key Laboratory of Industrial Biotechnology, College of Life Science, Hubei University, Wuhan, Hubei 430062, China. E-mail: yiliu0825@hubu.edu.cn
bDepartment of Chemistry, The Johns Hopkins University, Baltimore, MD 21210, USA
cSchool of Computer Science and Information Engineering, Hubei University, Wuhan, Hubei 430062, China. E-mail: yamin.li@hubu.edu.cn

Received 4th December 2017 , Accepted 21st December 2017

First published on 21st December 2017


Numerous studies have shown that nitric oxide (NO) interacts with human cyclooxygenase (COX); however, conflicting results exist with respect to their interactions. Herein, recombinant human COX-1 and COX-2 were prepared and treated with NO donors individually under anaerobic and aerobic conditions. The S-nitrosylation detection and subsequent kinetic investigations into the arachidonic acid (AA) oxidation of COX enzymes indicate that NO S-nitrosylates both COX-1 and COX-2 in an oxygen-dependent manner, but enhances only the dioxygenase activity of COX-2. The solution viscosity, deuterium kinetic isotope effect (KIE), and oxygen-18 KIE experiments further demonstrate that NO activates COX-2 by altering the protein conformation to stimulate substrate association/product release and by accelerating the rate of hydrogen abstraction from AA by catalytic tyrosine radicals. These novel findings provide useful information for designing new drugs with less cardiotoxic effects that can block the interaction between NO and COX.


Introduction

Cyclooxygenase (COX) is a bifunctional enzyme exhibiting coupled peroxidase and dioxygenase activities.1,2 Human COX-1 and COX-2 are two structurally homologous hemoproteins responsible for the biosynthesis of prostaglandin H2 from arachidonic acid (AA).1–3 They have been found to play great roles in certain diseases, including cancer,4 and nervous system5 and apoptosis-related diseases.6 COX-1 is constitutively expressed in a variety of cells,7 whereas COX-2 is induced by various inflammatory and proliferative stimuli. COX enzymes are important drug targets for the non-steroidal anti-inflammatory drugs. Over the past decades, COX-2 inhibitors have widely been used as anti-inflammatory medicine, although they usually elicit potential cardiotoxic side effects. Hence, many efforts have been made in recent years to eliminate or reduce such kinds of adverse effects.6,8,9 Crystal structure studies10,11 reveal that COX-2 has a larger substrate binding pocket than COX-1 (Fig. S1), providing the structural basis for developing new isoform-specific inhibitors (COXibs).

NO, a highly reactive small molecule produced within mammalian cells by the enzyme NO synthase (NOS) including iNOS (inducible NOS), eNOS (endothelial NOS) and nNOS (neuronal NOS), was initially reported by Needleman's group that can activate cyclooxygenase.12 Since then, numerous studies have shown that there is crosstalk between the NOS and COX pathways. Though a body of evidence indicated that NO elicits up-regulation of COX activity,13–16 a large number of reports support the idea that NO inactivates COX17–19 or has little effect on enzyme activity under certain conditions.20 It was suggested that the complex chemistry of NO species (NOx) and the crude enzyme preparations result in dichotomous effects with respect to COX activation by NO.21 Nevertheless, the mechanism of NO activation has not been understood yet and more in-depth kinetic investigations into the interaction between NO and COX are urgently needed.

Of the kinetic models proposed for COX catalysis, the “branch-chain” model is generally favored because it reveals the independent peroxidase and dioxygenase cycles (Scheme 1).22 Recently, we found that the first irreversible hydrogen transfer step in COX-2 catalysis altered along with oxygen variations,23 suggesting that the oxygen concentration has a huge impact on enzyme activity. In addition, we demonstrated that the dioxygenase activity of COX-2 can be analyzed independently of peroxidase activity by adding sufficient phenol, which acts as a sacrificed reductant,22,23 at a concentration of up to 2.5 mM. In contrast, the acceleration rate of COX-1 is followed by inhibition upon the addition of phenol above 0.3 mM,24 indicating that the kinetics of AA oxidation might be much more complicated for COX-1. Therefore, in spite of the sequence and structural similarities of COX-1 and COX-2, the different kinetics on substrate oxidation could be the main factor to determine their distinct functions under physiological and pathological conditions. Furthermore, we hypothesize that the interaction of NO with COX-1 and COX-2 might lead to different effects on enzyme activity under various oxygen conditions.


image file: c7ob02992f-s1.tif
Scheme 1 A proposed “branch-chain” mechanism of cyclooxygenase. The COX enzyme is a bifunctional enzyme having coupled peroxidase and dioxygenase activities.

In this study, we prepared recombinant COX-1 and COX-2 in the presence or absence of NO and measured their kinetics on AA oxidation. The results showed that NO can S-nitrosylate both COX enzymes in an oxygen-dependent manner, but activates only COX-2 under aerobic conditions. The UV-vis and CD measurements indicate that NO does not interact with the heme portion of COX enzymes, but largely alters the protein conformation of COX-2 rather than COX-1 by S-nitrosylation. In addition, the kinetic isotope effects on hydrogen extraction, and solvent and oxygen incorporation, combined with solvent viscosity experiments, reveal that NO activates COX-2 by stimulating the substrate association/product release, as well as by accelerating the rate of hydrogen abstraction from AA by catalytic tyrosine radicals.

Results

NO does not interact with the heme of COX enzymes

We firstly purified the apo-recombinant human COX-1 and COX-2, and then reconstituted them with hematin under anaerobic conditions (<1 ppm O2) in a glove box, affording Fe(III) protoporphyrin IX (Por) binding tightly in the enzyme active center. By using this method, the O2-free holo-COX enzymes were obtained. In the meantime, the holo-COX enzymes were prepared under aerobic conditions following the same protocol. Next, 1 mM NO donor sodium nitroprusside (SNP) was added and incubated with each type of the holo-COX enzyme for one hour under anaerobic and aerobic conditions, respectively. After this, the NO-treated COX enzymes were passed through spin columns to remove excess NO. Alternatively, other NO donors including nitroso-S-glutathione (GSNO) and (Z)-1-[N-(2-aminoethyl)-N-(2-ammonioethyl)amino]diazen-1-ium-1,2-diolate (DETA NONOate) were used and showed identical results.

A previous report showed that NO did not significantly alter the Soret band absorbance of COX-1,15 and suggested that NO has a relatively weak affinity for heme in ferric COX-1.15,20 Our spectral studies (Table S1) are in accordance with these results, and further prove that NO does not interact with the heme portion of COX-2 either.

Kinetics of O2 consumption by COX and NO-treated COX enzymes

Recently, the Snyder group revealed that NO can elicit the S-nitrosylation of the selected cysteines of COX-2 to increase its activity.25 They measured the dioxygenase activity of COX-2 by the rate constant of AA consumption under a saturated O2 condition. The results indicate that NO can activate COX-2 by increasing its apparent Vmax without changing KM(AA). In this study, we detected the kinetics of O2 consumption by COX and NO-treated COX enzymes at a saturated AA concentration.

The kinetic data of AA oxidation (Fig. 1) obtained over a range of oxygen concentrations were fitted by the Michaelis–Menten equation, giving the values as shown in Table 1. The data shown in Fig. 1A indicate that NO-treated COX-2 which was prepared under aerobic conditions has a significantly enhanced dioxygenase activity. In this case, NO accelerates COX-2 activity by increasing apparent Vmax from 16.9 s−1 to 27.5 s−1, as well as by decreasing KM(O2) from 15.0 μM to 10.3 μM. By contrast, no activation effect was observed for NO-treated COX-2 which was prepared under anaerobic conditions (Fig. 1B). On the other hand, no activation occurred for NO-treated COX-1 prepared under aerobic (Fig. 1C) and anaerobic conditions (Fig. 1D). In conclusion, the kinetic studies indicate that a certain amount of O2, in general over six-fold of KM(O2),26,27 is essential for the NO activation of COX-2.


image file: c7ob02992f-f1.tif
Fig. 1 Initial rates of O2 consumption in the presence of 50 μM AA and 2 mM phenol at pH 8.0 under 30 °C. Squares represent COX and circles represent NO-treated COX. (A) COX-2 and NO-treated COX-2 prepared under aerobic conditions; (B) COX-2 and NO-treated COX-2 prepared under anaerobic conditions; (C) COX-1 and NO-treated COX-1 prepared under aerobic conditions; (D) COX-1 and NO-treated COX-1 prepared under anaerobic conditions.
Table 1 Rate constants for the oxidation of AA by different COX enzymes prepared under aerobic and anaerobic conditionsa
Enzyme Aerobic Vmax (S−1) KM(O2) (μM) Anaerobic Vmax (S−1) KM(O2) (μM)
a Limiting rate constants are quoted as ±1 S. E. in parameters.
COX-2 16.9 (1.1) 15.0 (1.8) 16.7 (0.6) 13.2 (1.2)
NO-treated COX-2 27.5 (0.8) 10.3 (0.6) 15.2 (0.8) 12.8 (1.8)
COX-1 150.5 (7.2) 8.9 (0.8) 130.7 (7.4) 8.9 (1.1)
NO-treated COX-1 136.6 (4.5) 10.4 (0.7) 123.4 (4.2) 9.2 (0.8)


S-Nitrosylation detection

COX has been identified as a target of S-nitrosylation modification by various laboratories;13,28–30 however, the clear S-nitrosylation mechanism has not been understood. Herein, we performed a modified biotin switch assay31 and compared the fluorescence intensity of COX enzymes with and without SNP. The data (Fig. 2) indicate that both COX-1 and COX-2 have much higher S-nitrosylation ratios under aerobic conditions. Besides, COX enzymes showed median fluorescence intensity values at 50 μM O2, which is a common physiological oxygen concentration within mammalian cells.23 These findings support that NO can elicit the S-nitrosylation of both COX-1 and COX-2 enzymes in an oxygen-dependent manner; however, it can enhance only the dioxygenase activity of COX-2.
image file: c7ob02992f-f2.tif
Fig. 2 A biotin switch assay is specific for S-nitrosylation detection and the relative S-nitrosylation ratios of each of the COX enzymes are shown as (A) NO-treated COX-2 under aerobic and (B) anaerobic conditions; (C) NO-treated COX-1 under aerobic and (D) anaerobic conditions; (E) NO-treated COX-2 at 50 μM O2; (F) NO-treated COX-1 at 50 μM O2.

Solvent viscosity effects on kcat/KM(O2)

To ascertain the kinetic basis for the NO activation of COX-2 rather than COX-1, the studies on the O2 consumption rate constant (kcat/KM(O2)) as a function of solvent viscosity were carried out to expose possible contributions from diffusion controlled steps (Fig. 3). For COX-2, the ratio of enzyme activity to the activity in more viscous solutions increased a bit due primarily to a decrease in kcat. This increased trend was diminished in the S-nitrosylated COX-2, indicating that the interaction of NO with COX-2 can stimulate substrate association and/or product release from the enzyme, possibly by altering the structure or stability of the enzyme. In contrast, the S-nitrosylated COX-1 showed an unaltered viscosity dependence (Fig. 3B).
image file: c7ob02992f-f3.tif
Fig. 3 Effect of solvent viscosity upon kcat/KM(O2) for COX and S-nitrosylated COX. The {kcat/KM(O2)}0 represents the value at the initial solvent viscosity, whereas the kcat/KM(O2) represents the value under certain solvent viscosity by adding sucrose.

The analysis of the circular dichroism (CD) spectral data (Table S2) indicates that NO induced the significant secondary structural changes of both COX-1 and COX-2 by a dramatic increase in the percentage of β-sheet conformation. However, in contrast to COX-1, more turns and random coil conformation in COX-2 was transformed to β-sheets, therefore making the S-nitrosylated COX-2 more structurally ordered and less dynamic. A less dynamic structure usually corresponds to a lower solvent-viscosity effect, which is consistent with the observation (Fig. 3A). As shown in Fig. S2, Cys526 is located at the end of a region which orients perpendicular to the AA binding channel. Thus, it is possible that the S-nitrosylation of Cys526 may alter the geometry of the AA binding channel to facilitate easier substrate entry and product release, thereby, increasing the catalytic efficiency. However, this kind of activation effect was absent in COX-1, possibly because it has a more compact substrate binding pocket as shown in Fig. S2,[thin space (1/6-em)]10,11 and also because a solvent accessible pocket existing in COX-2 is inaccessible in COX-1.

Deuterium KIE and solvent KIE experiments

Recently, we revealed that hydrogen transfer from the reduced catalytic tyrosine (Y371) in COX-2 to a terminal peroxyl radical is the first irreversible step that controls regio- and stereo-specific production formation at physiologically relevant oxygen concentrations.23 Herein, the deuterium KIE on the AA oxidation of the S-nitrosylated COX-2 was monitored by comparing the oxidation kinetics of AA with that of 13,13-d2-AA. The apparent Dkcat and Dkcat/KM(O2) values were determined and are shown in Table 2. According to the dioxygenase model shown in Scheme 2, the primary deuterium KIEs (Dkcat) reflect k1, representing the C13 hydrogen homolysis of AA by tyrosine radicals (Y371˙). Compared to COX-2, the S-nitrosylated COX-2 showed an inflated Dkcat value from 3.1 to 5.8, suggesting that nitrosylation makes this hydrogen atom abstraction step more rate-limiting. One possible explanation is that the nitrosylation causes a conformational change in COX-2, which puts the tyrosine radical in a better position for hydrogen atom abstraction. In addition, the S-nitrosylated COX-2 showed a large Dkcat/KM(O2) value, which is similar to what we observed for COX-2 before,23 revealing that the secondary hydrogen transfer from the O–H homolysis of the reduced catalytic tyrosine radical to a terminal peroxyl radical (k5 in Scheme 2) becomes a rate-determining step at limiting O2 concentrations.
image file: c7ob02992f-s2.tif
Scheme 2 A proposed dioxygenase mechanism of human cyclooxygenase-2. Two hydrogen abstraction steps could be sensitive to AA deuteration: the abstraction of C13 pro-(S) hydrogen by the Tyr371 radical (k1) and abstraction of hydrogen from Tyr371 by the PGG2 radical (k5).
Table 2 Deuterium KIEs and 18O KIEs upon AA oxidation by COX-2 and S-nitrosylated COX-2
Enzyme COX-2a S-Nitrosylated COX-2b
a Data derived from ref. 23.b Prepared by treating COX-2 with 1 mM SNP under aerobic conditions.c Apparent solvent kinetic isotope effects are determined as described in ref. 24.
Dkcat 3.1 (0.4) 5.4 (0.6)
Dkcat/KM(O2) 18.6 (3.1) 15.1 (2.7)
SKIEc 1.2 (0.2) 1.1 (0.3)
18O KIEs 1.0194 (0.0018) 1.0211 (0.0023)


Next, the solvent KIE experiments were performed by measuring the oxidation of AA in deuterium water. No solvent KIE (Table 2) was observed for COX-2 and the S-nitrosylated COX-2, indicating that there is no proton transfer or hydrogen atom exchange with the solvent in rate-determining steps. Therefore, the observed inflated Dkcat of the S-nitrosylated COX-2 arises directly from the acceleration of the hydrogen atom abstraction of AA.

Competitive oxygen-18 KIE experiments

The competitive oxygen-18 KIE experiments were carried out to monitor any possible kinetic changes with respect to the oxygen involved steps in S-nitrosylated COX-2 catalysis. Previously, the 18O KIE upon kcat/KM(O2), reflecting steps that begin with O2 entering the catalysis cycle and leading up to the first irreversible step, was measured to be 1.0194 ± 0.0018 for AA oxidation by COX-2.23 Herein, the 18O KIE studies of the S-nitrosylated COX-2 resulted in a comparable value of 1.0211 ± 0.0023 for AA oxidation (Table 2), suggesting that there is no kinetic alternation after O2 encounters. That is to say, the NO activation step should exist before O2 enters the catalysis cycle (k2 in Scheme 2). This result reinforces the above conclusion that NO activates COX-2 by the acceleration of the hydrogen atom abstraction of AA.

Discussion

It has been proposed that local hydrophobic compartments might promote specific S-nitrosylation, especially for S-nitrosylation that is based on NO auto-oxidation or the direct reaction of NO with the thiolate of cysteines.32 In these cases, NO itself or other NO oxides, including in particular NO2/N2O3/N2O4, which is generated from NO/O2 reaction products, can S-nitrosylate specific cysteines that usually reside in a juxtamembrane zone. In support of this view, herein, our results show that NO S-nitrosylates both COX-1 and COX-2 in an oxygen-dependent manner, pointing out that a certain amount of O2 is essential for the high ratios of S-nitrosylation. Snyder et al. reported that multiple cysteines in COX-2 can be S-nitrosylated but only C526 is responsible for COX-2 activation, since the C526S mutant prevented the activation of COX-2 by SNP.25 This cysteine residue is located in the high hydrophobic region within COX-2 (Fig. S2), surprisingly as far as 30 Å away from the heme prosthetic group. The analysis of the sequence alignment of human COX-1 and COX-2 (Fig. S3) demonstrates that C526 in COX-2 corresponds to C540 in COX-1. Thus, the activation of COX-2 rather than COX-1 by NO cannot be simply explained by the S-nitrosylation of specific cysteine(s). Our solution viscosity effects and CD spectroscopy studies reveal that NO can induce the conformation changes of COX-2 and COX-1 a bit, but accelerates only substrate association and /or product release for COX-2.

Conclusion

In summary, our studies demonstrate that NO binds to COX independently of heme, S-nitrosylates it, and enhances the dioxygenase activity of COX-2 rather than COX-1 under aerobic conditions. The S-nitrosylation modification is in an oxygen-dependent manner. The solvent viscosity in combination with the CD experiment indicates that NO can activate COX-2 by altering the protein conformation to stimulate substrate association/product release. The kinetics, deuterium KIE, solvent KIE, and 18O KIE results all suggest that the enhancement of the activity of the S-nitrosylated COX-2 is most likely attributed to the acceleration of the hydrogen abstraction from AA. Currently, researchers are reflecting on the growing prospect that S-nitrosylation is a post-translational protein modification that is regulated precisely in time and space.33 Herein, our data reinforce this idea and prove that sufficient oxygen is essential for the high-level S-nitrosylation of COX enzymes, especially for the activation of COX-2. The findings supply meaningful information for designing new drugs that block the interaction between NO and COX.

Experimental

General methods

The chemicals were procured in the highest purity available and used as received unless noted. Sodium phosphate (monobasic), sodium pyrophosphate, sodium chloride, sucrose, polyethylene glycol, and H2O2 were obtained from National Medicine Company of China. Hematin, phenylmethanesulfonyl fluoride (PMSF), 2,2-azinobis(3-ethylbenzothiazoline-6-sulfonic acid)diammonium salt, trimethyl phosphite, glycerol, Tween 20, 2-methoxyphenol, imidazole, phenol, 5(Z),8(Z),11(Z),14(Z)-eicosatetraenoic acid (AA), and 15(S)-hydroperoxyeicosatetraenoic acid (HPETE) were obtained from Sigma Aldrich. Selectively labeled 13,13-d2-AA was a generous gift from Prof. Wilfred van der Donk (University of Illinois at Urbana-Champaign). AA and COX enzymes when necessary were stored at −30 °C inside an argon-filled glove box (MBRAUN LABmaster) containing <1 ppm O2. Ethanolic stock solutions of AA were prepared prior to use in kinetic experiments.

Protein purification, reconstitution and characterization

The recombinant holo-COX enzymes were prepared under aerobic conditions according to the procedures as described previously.23 Alternatively, the holo-COX enzymes were prepared under anaerobic conditions in a glove box filled with argon gas (<1 ppM O2) following the same protocol. The protein concentrations were determined by UV/vis spectroscopy according to the Soret band absorbance molar extinction coefficient.31 The active enzyme concentration in reaction mixtures was determined by performing the standard enzyme activity assays. The solutions of holo-COX enzymes were concentrated to <5 ml, apportioned into 50–100 μM and stored at −80 °C.

UV/vis spectroscopy

Electronic absorption measurements were performed using a UV/vis spectrophotometer (Agilent 8453) according to the published work.15 The maximum Soret band of holo-COX-1 is at 410 nm, whereas it is at 407 nm for holo-COX-2. The Soret band absorbance of holo-COX enzymes did not significantly change after the addition of NO, as shown in Table S1.

Steady-state kinetics

All experiments were performed as described previously.23 The trace hydroperoxide impurities present in AA are sufficient to initiate the cyclooxygenase catalytic cycles. Alternatively, 3 μM 15S-HTEPE was added as the initiator, and no kinetic changes were observed with respect to AA oxidation. Kinetic parameters including Vmax and KM(O2) were obtained by fitting data to the hyperbolic Michaelis–Menten equation by using the software Oringin Pro 9.0.

S-Nitrosylation assay (botin switch assay)

Cayman's S-nitrosylated Protein Detection kit (Biotin Switch), based on a modification of the Jaffrey et al. ‘Biotin-switch’ method,31 was employed to visualize the S-nitrosylated COX enzymes prepared at various oxygen concentrations. By using this method, free SH groups were first blocked and any S–NO bonds present in the sample were then cleaved. The biotinylation of the newly formed SH groups provides the basis for the visualization of separated proteins in gel by using fluorescence detection.

Solvent viscosity assay

The experiments were carried out to probe diffusion-limited contributions to kcat/KM(O2) in accord with the Stokes–Einstein relation,34 following the protocol by adding sucrose.

Circular dichroism spectroscopy

The CD spectra of COX enzymes (5 μM in 100 mM Tris-HCl, pH 8.0) including COX-1, S-nitrosylated COX-1, COX-2 as well as S-nitrosylated COX-2 were recorded at 22 °C in 0.1 mm cuvettes by using a Jasco 710 spectrophotometer. The data were analyzed by CAPITO, which is a newly developed web server-based analysis and plotting tool for CD data.35 The quantitative analysis of data is shown in Table S2.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was supported by the National Foundation of Hubei Province (no. 201700963 to Y. Liu) and the Shanghai Pujiang Grant (no. 16PJ1403400 to Y. Liu).

Notes and references

  1. C. A. Rouzer and L. J. Marnett, Chem. Rev., 2003, 103, 2239–2304 CrossRef CAS PubMed.
  2. C. A. Rouzer and L. J. Marnett, Chem. Rev., 2011, 111, 5899–5921 CrossRef CAS PubMed.
  3. W. L. Smith, Y. Urade and P. J. Jakobsson, Chem. Rev., 2011, 111, 5821–5865 CrossRef CAS PubMed.
  4. W. L. Smith, D. L. DeWitt and R. M. Garavito, Annu. Rev. Biochem., 2000, 69, 145–182 CrossRef CAS PubMed.
  5. B. Halliwell and J. M. C. Gutteridge, Free Radicals in Biology and Medicine, Oxford University Press, New York, 3rd edn, 2005 Search PubMed.
  6. L. J. Marnett, Annu. Rev. Pharmacol. Toxicol., 2009, 49, 265–290 CrossRef CAS PubMed.
  7. J. Tian, S. F. Kim, L. Hester and S. H. Snyder, Proc. Natl. Acad. Sci. U. S. A., 2008, 105, 10537–10540 CrossRef CAS PubMed.
  8. D. G. Menter, R. L. Schilsky and R. N. DuBois, Clin. Cancer Res., 2010, 16, 1384–1390 CrossRef CAS PubMed.
  9. D. Mukherjee, S. E. Nissen and E. J. Topol, J. Am. Med. Assoc., 2001, 286, 954–959 CrossRef CAS PubMed.
  10. A. J. Vecchio, D. M. Simmons and M. G. Malkowski, J. Biol. Chem., 2010, 285, 22152–22163 CrossRef CAS PubMed.
  11. A. J. Vecchio, B. J. Orlando, R. Nandagiri and M. G. Malkowski, J. Biol. Chem., 2012, 287, 24619–24630 CrossRef CAS PubMed.
  12. A. R. Morrison, H. Moritz and P. Needleman, J. Biol. Chem., 1978, 253, 8210–8212 CAS.
  13. D. Salvemini, T. P. Misko, L. J. Masferrer, K. Seibert, M. G. Currie and P. Needleman, Proc. Natl. Acad. Sci. U. S. A., 1993, 90, 7240–7244 CrossRef CAS.
  14. A. H. Liu, G. L. Borchert and J. M. Phang, J. Biol. Chem., 2004, 279, 18694–18700 CrossRef PubMed.
  15. D. P. Hajjar, H. M. Lander, S. F. Pearce, A. R. K. Upmacis and K. B. Pomerantz, J. Am. Chem. Soc., 1995, 117, 3340–3346 CrossRef CAS.
  16. S. W. Park, M. W. Sung, D. S. Heo, H. Inoue, S. H. Shim and K. H. Kim, Oncogene, 2005, 24, 6689–6698 CrossRef CAS PubMed.
  17. J. Stadler, B. G. Harbrecht, S. M. Di, R. D. Curran, M. L. Jordan, R. L. Simmons and T. R. Billiar, J. Leukocyte Biol., 2005, 53, 165–172 Search PubMed.
  18. L. Minghetti, E. Plazzi, A. Nicolini, C. Creminon and G. Levi, J. Neurochem., 1996, 66, 1963–1970 CrossRef CAS PubMed.
  19. T. A. Swierkosz, J. A. Mitchell, T. D. Warner, R. M. Botting and J. R. Br. Vane, J. Pharmacol., 1995, 114, 1335–1342 CAS.
  20. A. L. Tsai, C. Wei and R. J. Kulmacz, Arch. Biochem. Biophys., 1994, 313, 367–372 CrossRef CAS PubMed.
  21. S. F. Kim, Nitric Oxide, 2011, 25, 255–264 CrossRef CAS PubMed.
  22. H. Danish, I. S. Doncheva and J. P. Roth, J. Am. Chem. Soc., 2012, 133, 15846–15849 CrossRef PubMed.
  23. Y. Liu and J. P. Roth, J. Biol. Chem., 2016, 291, 948–958 CrossRef CAS PubMed.
  24. A. Mukherjee, D. W. Brinkley, K. M. Chang and J. P. Roth, Biochemistry, 2007, 46, 3975–3989 CrossRef CAS PubMed.
  25. S. F. Kim, D. A. Huri and S. H. Snyder, Science, 2005, 310, 1966–1970 CrossRef CAS PubMed.
  26. J. P. Roth, Acc. Chem. Res., 2009, 42, 399–408 CrossRef CAS PubMed.
  27. J. P. Roth, Curr. Opin. Chem. Biol., 2007, 11, 142–150 CrossRef CAS PubMed.
  28. A. V. D. Vliet, P. A. Hoen, P. S. Wong, A. Bast and C. E. Cross, J. Biol. Chem., 1998, 273, 30255–30262 CrossRef PubMed.
  29. T. A. Kennedy, C. J. Smith and L. J. Marnett, J. Biol. Chem., 1994, 269, 27357–27364 CAS.
  30. L. Xu, C. Han, K. Lim and T. Wu, J. Biol. Chem., 2008, 283, 3077–3087 CrossRef CAS PubMed.
  31. S. R. Jaffrey, H. C. D. Ferris, P. Tempst and S. H. Snyder, Nat. Cell Biol., 2001, 3, 193–197 CrossRef CAS PubMed.
  32. Isotope Effects On Enzyme-catalyzed Reactions, ed. A. J. Kresge, W. W. Cleland, M. HD. B. O'Leary, Northrop, University Park Press, Baltimore, 1977, pp. 37–63 Search PubMed.
  33. D. Hess, T. A. Matsumoto, S. O. Kim, H. E. Marshall and J. S. Syamler, Nat. Rev. Mol. Cell Biol., 2005, 2, 150–166 CrossRef PubMed.
  34. A. C. Brouwer and J. F. Kirsch, Biochemistry, 1982, 21, 1302–1307 CrossRef CAS PubMed.
  35. C. Wiedemann, P. Bellstedt and M. Görlach, Bioinformatics, 2013, 29, 1750–1757 CrossRef CAS PubMed.

Footnote

Electronic supplementary information (ESI) available. See DOI: 10.1039/c7ob02992f

This journal is © The Royal Society of Chemistry 2018