H2O2-Mediated Oxidation with Enhanced

Vivien Berthelot , Vincent Steinmetz , Luis A. Alvarez , Chantal Houée-Levin , Fabienne Merola , Filippo Rusconi , Marie Erard. Analytical and Bioana...
0 downloads 0 Views 240KB Size
Anal. Chem. 2005, 77, 2862-2867

Detection of Peroxidase/H2O2-Mediated Oxidation with Enhanced Yellow Fluorescent Protein Andrew Tsourkas, Gail Newton, J. Manuel Perez, James P. Basilion, and Ralph Weissleder*

Center for Molecular Imaging Research, Massachusetts General Hospital, Charlestown, Massachusetts 02129

The ability to sense oxidative stress in live cells and organisms would have far-reaching implications for biotechnology, drug discovery, and potentially medical imaging. We hypothesized that tyrosine-containing fluorescent proteins could be used as switches for sensing oxidative stress, based on their sensitivity to environmental and structural variations. We therefore tested purified EGFP, EYFP, ECFP, and DsRed proteins against the hemeperoxidase/H2O2 reaction. We found that peroxidasemediated oxidation resulted in up to 99.5% quenching of EYFP fluorescence (but not that of other fluorescent proteins) in a dose-dependent manner. Western blotting revealed inter- and intramolecular cross-linking. The observed detection limit for hydrogen peroxide was ∼100 nM, well below the extracellular levels previously reported to occur in mammalian tissue during signaling. Combined expression of EYFP (quenchable) and ECFP or EGFP (nonquenchable) is expected to allow sensitive monitoring of oxidative stress. Oxidative agents released by activated neutrophils play a key role in host defense by imposing a potent toxic effect on invading pathogens and tumor cells. Under normal conditions, these oxidants are subsequently eliminated from the body by enzymatic reduction, antioxidants, and metal chelators; however, in certain pathological states, normal tissue may also be a target for damage. Oxidative damage (stress) has been implicated with numerous diseases including atherosclerosis (and other inflammatory diseases),1-3 neurodegenerative disorders,4 carcinogenesis,5,6 ischemia, and reperfusion injury.7,8 Oxidative stress has also been associated with aging,9,10 exposure to ultraviolet light, and γ-irradiation. * To whom correspondence should be addressed. Tel.: 617-726-5788. Fax: 617-726-5708. E-mail: [email protected]. (1) de Nigris, F.; Lerman, A.; Ignarro, L. J.; Williams-Ignarro, S.; Sica, V.; Baker, A. H.; Lerman, L. O.; Geng, Y. J.; Napoli, C. Trends Mol. Med. 2003, 9, 351-359. (2) Cathcart, M. K. Arterioscler. Thromb. Vasc. Biol. 2004, 24, 23-28. (3) Salvemini, D.; Ischiropoulos, H.; Cuzzocrea, S. Methods Mol. Biol. 2003, 225, 291-303. (4) Barnham, K. J.; Masters, C. L.; Bush, A. I. Nat. Rev. Drug Discovery 2004, 3, 205-214. (5) Ohshima, H.; Tatemichi, M.; Sawa, T. Arch. Biochem. Biophys. 2003, 417, 3-11. (6) Olinski, R.; Gackowski, D.; Foksinski, M.; Rozalski, R.; Roszkowski, K.; Jaruga, P. Free Radical Biol. Med. 2002, 33, 192-200. (7) Touyz, R. M. Expert Rev. Cardiovasc. Ther. 2003, 1, 91-106. (8) Venditti, P.; De Rosa, R.; Cigliano, L.; Agnisola, C.; Di Meo, S. Cell Mol. Life Sci. 2004, 61, 2244-2252.

2862 Analytical Chemistry, Vol. 77, No. 9, May 1, 2005

A prevalent pathway of oxidative stress involves the release of the heme protein, myeloperoxidase, and superoxide (O2•-) from activated neutrophils and monocytes. Superoxide subsequently dismutates to form hydrogen peroxide, which serves as a substrate for myeloperoxidase. The myeloperoxidase/H2O2 system is capable of reacting with various electron donors to produce oxidative agents including hypochlorous acid from Cl- and tyrosyl radicals from L-tyrosine. HOCl can react with tyrosine to form 3-chlorotyrosine.11 Alternatively, tyrosyl radicals can react with another tyrosine radical to form dityrosine or with nitric oxide to form 3-nitrotyrosine.12-16 Tyrosyl nitration can also occur in the presence of peroxynitrite, which is formed from the reaction between superoxide and nitric oxide.17 The occurrence of irreversibly altered tyrosine residues in the presence of oxidative reagents has been shown to result in impaired protein function and the subsequent loss of activity of cellular enzymes.18,19 Moreover, reports have shown that there is a strong correlation between various conditions of oxidative stress and the extent of tyrosine modifications in humans.20 This has aroused interest in using modified tyrosine as an indicator for oxidative stress. The prospect of utilizing tyrosine oxidation as an indicator for oxidative stress has recently led to the emergence of several molecular probes that specifically elicit a signal upon tyrosyl nitration or dityrosine formation. Examples include tyraminefluorescein conjugates (TyroFluor), which upon oxidation can cross-link and fluorescently label oxidized tyrosine residues on proteins;21-23 phenol-labeled magnetic agents (i.e., nanoparticles, chelates, etc.), which result in altered magnetic properties that (9) Sohal, R. S. Free Radical Biol. Med. 2002, 33, 37-44. (10) Stadtman, E. R. Science 1992, 257, 1220-1224. (11) Hazen, S. L.; Heinecke, J. W. J. Clin. Invest. 1997, 99, 2075-2081. (12) Giulivi, C.; Traaseth, N. J.; Davies, K. J. Amino Acids 2003, 25, 227-232. (13) Malencik, D. A.; Anderson, S. R. Amino Acids 2003, 25, 233-247. (14) Heinecke, J. W. Toxicology 2002, 177, 11-22. (15) Heinecke, J. W.; Li, W.; Daehnke, H. L., 3rd; Goldstein, J. A. J. Biol. Chem. 1993, 268, 4069-4077. (16) Jacob, J. S.; Cistola, D. P.; Hsu, F. F.; Muzaffar, S.; Mueller, D. M.; Hazen, S. L.; Heinecke, J. W. J. Biol. Chem. 1996, 271, 19950-19956. (17) Sampson, J. B.; Ye, Y.; Rosen, H.; Beckman, J. S. Arch. Biochem. Biophys. 1998, 356, 207-213. (18) Yamakura, F.; Taka, H.; Fujimura, T.; Murayama, K. J. Biol. Chem. 1998, 273, 14085-14089. (19) Di Stasi, A. M.; Mallozzi, C.; Macchia, G.; Petrucci, T. C.; Minetti, M. J. Neurochem. 1999, 73, 727-735. (20) Heinecke, J. W. Free Radical Biol. Med. 2002, 32, 1090-1101. (21) Sakharov, D. V.; Bunschoten, A.; van Weelden, H.; Wirtz, K. W. Eur. J. Biochem. 2003, 270, 4859-4865. (22) Czapski, G. A.; Avram, D.; Wirtz, K. W.; Pap, E. H.; Strosznajder, J. B. Med. Sci. Monit. 2001, 7, 606-609. (23) van der Vlies, D.; Wirtz, K. W.; Pap, E. H. Biochemistry 2001, 40, 77837788. 10.1021/ac0480747 CCC: $30.25

© 2005 American Chemical Society Published on Web 03/26/2005

can be detected by magnetic resonance upon the peroxidase induced cross-linking;24,25 and enhanced green fluorescent proteins (EGFP), which is effectively quenched upon nitration of tyrosyl residues (but apparently not by peroxidase/H2O2).26 EGFP is a particularly versatile probe because it not only provides information on tyrosine nitration under homogeneous conditions but also is continuously produced within the cells of interest providing a true real-time analysis of protein modification. Here we examined whether fluorescent proteins could also be used to detect peroxidase/H2O2-mediated oxidation. Specifically, we monitored the fluorescence of EGFP, EYFP (yellow), ECFP (cyan), and DsRed in the presence and absence of peroxidase/H2O2. We show that EYFP can act as a molecular reporter for peroxidase/H2O2mediated oxidation and is much more sensitive than other fluorescent proteins tested. The ability to use endogenous molecular reporters as biosensors for oxidative stress could provide (a) unique insight into the occurrence of protein modifications at the site of oxidative stress and (b) allow real-time measurement of such stress. EXPERIMENTAL SECTION Gene Construction and Protein Expression. The vectors pECFP, pEGFP, pEYFP-C1, pDsRed-C1, and pHAT10/11/12 were purchased from BD Biosciences. The ECFP, EGFP, and EYFP genes were amplified by PCR using the following primers: 5′CGC GGA TCC ATG GTG AGC AAG GGC GAG G-3′ and 5′-CCG GAA TTC TTA CTT GTA CAG CTC GTC CAT GCC-3′, which introduce BamHI and EcoRI restriction sites into the 5′- and 3′end of each amplicon, respectively. The DsRed gene was amplified using the primers 5′-CGC GGA TCC ATG GCC TCC TCC GAG AAC G-3′ and CCG GAA TTC TTA CAG GAA CAG GTG GTG GCG-3′, which also introduced BamHI and EcoRI restriction sites onto the 5′- and 3′-ends, respectively. Each fluorescent protein gene was then ligated into the BamHI and EcoRI sites of pHAT10, which contains an N-terminal histidine affinity tag. The fluorescent proteins were expressed in Escherichia coli strain XL1-Blue (Stratagene). For the growth and induction of the strains carrying the desired constructs, 1 L of LB medium was inoculated with a 50-mL overnight culture and allowed to grow at 37 °C for 1.5 h. Expression was induced by adding IPTG to a concentration of 0.5 mM and incubating for a further 5 h. The cells were then harvested and resuspended in 35 mL of extraction/wash buffer (50 mM sodium phosphate, 300 mM NaCl, pH 7.0) per 500 mL of bacterial culture. Lysozyme was then added to a final concentration of 0.75 mg/mL, and PMSF was added to a final concentration of 1 mM. The cells were left standing for 15-20 min and then frozen at -80 °C overnight or until needed. Subsequently, the cells were thawed at 37 °C and lysed by sonicating 3 times for ∼10 s with ∼60-s intervals, chilling on ice between each round of sonication. The cell extract was then purified on TALON resin (BDClontech), a cobalt-charged immobilized metal affinity chromatography resin, and eluted with extraction/wash buffer containing 0.15 M imidazole (Sigma). Samples were concentrated using Microcon cen(24) Perez, J. M.; Simeone, F. J.; Tsourkas, A.; Josephson, L.; Weissleder, R. Nano Lett. 2004, 4, 119-122. (25) Chen, J. W.; Pham, W.; Weissleder, R.; Bogdanov, A., Jr. Magn. Reson. Med. 2004, 52, 1021-1028. (26) Espey, M. G.; Xavier, S.; Thomas, D. D.; Miranda, K. M.; Wink, D. A. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 3481-3486.

trifugation filters (Millipore) and dialyzed against buffer A (50 mM Na2HPO4, 100 µM DTPA, pH adjusted to 7.5 with acetic acid). The concentration of each fluorescent protein was determined by spectroscopy using the following information provided by BD Biosciences (ECFP: λabs ) 434,  ) 26 000. EGFP: λabs ) 489,  ) 55 000. EYFP: λabs ) 514,  ) 84 000. DsRed: λabs ) 558,  ) 22 5000). Fluorescent Protein-Peroxidase End-Point Assay. To determine whether peroxidase/H2O2-mediated oxidation could be detected by an attenuation of fluorescent signal, the following five samples were prepared with ECFP, EGFP, EYFP, and DsRed: (1) 1 µM fluorescent protein in buffer A, (2) 1 µM fluorescent protein and 50 µM H2O2 in buffer A, (3) 1 µM fluorescent protein and 40 units/mL horseradish peroxidase (HRP) in buffer A, (4) 1 µM fluorescent protein, 50 µM H2O2, and 40 units/mL HRP in buffer A, and (5) 1 µM fluorescent protein, 40 units/mL HRP and 150 mM NaCl in buffer A. Each sample was made in triplicate and allowed to react overnight at 4 °C to ensure that the reaction reached equilibrium. Fluorescence measurements were taken of each sample on a Hitachi F-4500 fluorescence spectrophotometer, using the following excitation and emission wavelengths (ECFP: λexc ) 434 nm, λemm ) 477 nm. EGFP: λexc ) 489 nm, λemm ) 508 nm. EYFP: λexc ) 514 nm, λemm ) 527 nm. DsRed: λexc ) 558 nm, λemm ) 583 nm). The fluorescence of each sample was normalized by the average fluorescence of sample 1. The fluorescence spectrum of EYFP was obtained using an excitation wavelength of 514 nm and emissions ranging from 525 to 600 nm. Samples containing EYFP were also excited at 325 nm, and the emissions ranging from 375 to 475 nm were recorded to determine whether the inherent fluorescence of dityrosine could be detected. As an alternative method to determine whether intermolecular protein cross-linking had occurred, 16 µL of each sample was subjected to SDS-PAGE and western blotting. Western blotting was performed using an ECL Western Blotting Kit (Amersham) according to the manufacturer’s directions. A rabbit anti-GFP/ YFP/CFP (Sigma) was used as the primary antibody at a 1:1500 dilution, and a goat anti-rabbit antibody with HRP (Sigma) was used as the secondary antibody at a 1:5000 dilution. Kinetics of Peroxidase-Mediated Oxidative Stress. The real-time kinetics of peroxidase-induced EYFP quenching was determined for samples containing 1 µM EYFP, 50 µM H2O2, and 40 units/mL HRP in buffer A by recording the fluorescence signal (λexc ) 514, λemm ) 527) every second for a period of 1 h. A fluorescence emission of a control sample containing 1 µM EYFP in buffer A was also monitored to determine the extent of photobleaching. The fluorescence of each sample was normalized by the fluorescence of the control sample at time 0. Additional samples were run in the presence of 150 mM NaCl to determine whether NaCl interfered with the quenching of EYFP that resulted from peroxidase-induced oxidation. Further, a sample containing 1 µM EYFP, 40 units/mL HRP, and 150 mM NaCl in buffer A was also tested to determine whether chlorination occurred or had an impact on the fluorescent signal. Following the 4-h kinetic experiment, each sample was incubated at 4 °C overnight and the end-point fluorescence was recorded the following day. Each of the samples tested was run in triplicate. Peroxidase Sensitivity Assay. To determine the dependence of EYFP quenching on the concentration of HRP, varying amounts Analytical Chemistry, Vol. 77, No. 9, May 1, 2005

2863

of HRP (0, 0.001, 0.0025, 0.005, 0.01, 0.025, 0.05, 0.1, 0.25, 0.5, 1, 2.5, 5, 10 units/mL) were incubated with 1 µM EYFP, 150 mM NaCl, and 50 µM H2O2 in buffer A overnight at 4 °C. The maximum fluorescent signal (λexc ) 514, λemm ) 527) was then recorded for each sample and normalized by the fluorescent signal of EYFP in the presence of 50 µM H2O2 but no HRP. An SDS-PAGE was run with 16 µL of the samples containing 0, 0.01, 0.1, 1, 10, and 10 units/mL HRP, and a western blot was performed as described above. Experiments were performed in duplicate. H2O2 Sensitivity Assay. Samples containing 1 µM EYFP, 40 units/mL HRP, and 150 mM NaCl in buffer A were incubated with various concentrations of H2O2 (0, 0.01, 0.1, 1.0, 10, 100, 1000 µM) overnight at 4 °C. The maximum fluorescent signal was then recorded for each sample and normalized by the fluorescence signal of EYFP in the presence of 40 units/mL HRP but no H2O2. Experiments were performed in duplicate. EYFP Sensitivity Assay. Samples containing 0.2, 0.5, 1, 2, or 3 µM EYFP were incubated overnight at 4 °C with 50 µM H2O2 and 150 mM NaCl in buffer A in the presence and absence of 40 units/mL HRP. The maximum fluorescent signal was then recorded for each sample and normalized by the fluorescence signal of EYFP at the same concentration in the presence of 50 µM H2O2 but no HRP. Site-Directed Mutagenesis. Specific mutations were introduced into the pHAT10-EYFP plasmid using the Quick-Change site-directed mutagenesis kit (Stratagene). The oligonucleotides that were used to generate the desired mutations (Y201S or Y152S), and their complement were as follows: Y201S, 5′-GCC CGA CAA CCA CAG CCT GAG CTA CCA GTC C-3′; 5′-GGA CTG GTA GCT CAG GCT GTG GTT GTC GGG C-3′; Y152S, 5′-TAC AAC AGC CAC AAC GTC AGC ATC ATG GCC GAC AAG CAG3′; 5′-CTG CTT GTC GGC CAT GAT GCT GAC GTT GTG GCT GTT GTA-3′. Each mutant EYFP protein was expressed and purified as described above. Further, each mutant (at a concentration of 1 µM) was incubated at 4 °C overnight with 50 µM H2O2 and 150 mM NaCl in buffer A in the presence and absence of 40 units/mL HRP. The maximum fluorescent signal was then recorded for each sample and normalized by the fluorescence signal of EYFP at the same concentration in the absence of HRP. Myeloperoxidase Sensitivity Assay. To determine whether mammalian myeloperoxidase (MPO) also induces EYFP quenching, cross-linking (intra- or intermolecular), or both , 1 µM fluorescent protein in buffer A was incubated with 0, 1, or 8 units/ mL MPO, 0 or 50 µM H2O2, 0, 0.05, or 0.2 mM free L-tyrosine, and 0 or 150 mM NaCl. Each sample was made in triplicate and allowed to react overnight at 4 °C to ensure that the reaction reached equilibrium. Fluorescence measurements were taken of each sample using excitation and emission wavelengths of 514 and 527 nm, respectively. The fluorescence of each sample was normalized by the average fluorescence of 1 µM EYFP in buffer A. The fluorescence spectrum of EYFP was obtained using an excitation wavelength of 514 nm and emissions ranging from 525 to 600 nm. Samples containing EYFP were also excited at 325 nm, and the emissions ranging from 375 to 475 nm were recorded to determine whether the fluorescence of dityrosine could be detected. 2864

Analytical Chemistry, Vol. 77, No. 9, May 1, 2005

RESULTS AND DISCUSSION Fluorescent Proteins as Biosensors for Oxidative Stress. A series of experiments were performed to determine whether the fluorescent proteins EGFP, EYFP, ECFP, or DsRed could be used as a biosensor for detecting peroxidase/H2O2-mediated oxidative stress. It has previously been established that a HRP/ H2O2 system can effectively induce the oxidative dimerization of tyrosol residues on proteins in vitro, and it was hypothesized that modification of tyrosol residues on a fluorescent protein would induce a measurable change in its fluorescent properties.27,28 To explore this possibility, each purified fluorescent protein (1 µM) was suspended in buffer A and exposed to 40 units/mL heme protein, HRP, and/or 50 µM H2O2. The fluorescent signal of each sample was measured using the peak excitation and emission wavelength of the respective fluorescent protein. It was found that the presence of HRP or H2O2 alone does not induce a significant change in fluorescence (