Imaging Dynamic Redox Changes in Mammalian Cells with Green Fluorescent Protein Indicators

Changes in the redox equilibrium of cells influence a host of cell functions. Alterations in the redox equilibrium are precipitated by changing either the glutathione/glutathione-disulfide ratio (GSH/GSSG) and/or the reduced/oxidized thioredoxin ratio. Redox-sensitive green fluorescent proteins (GFP) allow real time visualization of the oxidation state of the indicator. Ratios of fluorescence from excitation at 400 and 490 nm indicate the extent of oxidation and thus the redox potential while canceling out the amount of indicator and the absolute optical sensitivity. Because the indicator is genetically encoded, it can be targeted to specific proteins or organelles of interest and expressed in a wide variety of cells and organisms. We evaluated roGFP1 (GFP with mutations C48S, S147C, and Q204C) and roGFP2 (the same plus S65T) with physiologically or toxicologically relevant oxidants both in vitro and in living mammalian cells. Furthermore, we investigated the response of the redox probes under physiological redox changes during superoxide bursts in macrophage cells, hyperoxic and hypoxic conditions, and in responses to H2O2-stimulating agents, e.g. epidermal growth factor and lysophosphatidic acid. Changes in the redox equilibrium of cells influence a host of cell functions. Alterations in the redox equilibrium are precipitated by changing either the glutathione/glutathione-disulfide ratio (GSH/GSSG) and/or the reduced/oxidized thioredoxin ratio. Redox-sensitive green fluorescent proteins (GFP) allow real time visualization of the oxidation state of the indicator. Ratios of fluorescence from excitation at 400 and 490 nm indicate the extent of oxidation and thus the redox potential while canceling out the amount of indicator and the absolute optical sensitivity. Because the indicator is genetically encoded, it can be targeted to specific proteins or organelles of interest and expressed in a wide variety of cells and organisms. We evaluated roGFP1 (GFP with mutations C48S, S147C, and Q204C) and roGFP2 (the same plus S65T) with physiologically or toxicologically relevant oxidants both in vitro and in living mammalian cells. Furthermore, we investigated the response of the redox probes under physiological redox changes during superoxide bursts in macrophage cells, hyperoxic and hypoxic conditions, and in responses to H2O2-stimulating agents, e.g. epidermal growth factor and lysophosphatidic acid. Cells have elaborate homeostatic mechanisms to regulate the thiol-disulfide redox status of their internal compartments. Most thiol groups within the cytoplasm are normally reduced. Very few are present as disulfides. It has been speculated that the cytoplasm is reducing because many metabolic reactions evolved before oxygen became abundant in the atmosphere (1Krauth-Siegel R.L. Jockers-Scherubl M.C. Becker K. Schirmer R.H. Biochem. Soc. Trans. 1989; 17: 315-317Crossref PubMed Scopus (11) Google Scholar). Modest alterations in the thiol-disulfide equilibrium could have major consequences in the cell, including defective protein folding or enzyme activity (because many enzymes have a cysteine in their active site). When excess oxidation overwhelms the reductive capabilities of the cell, death results. Despite the dangers of excessive oxidation, cells sometimes use redox adjustments as signaling events, such as in the activation of transcription factors (NF-κB and AP-1), caspases, protein tyrosine phosphatases, or GTPases (Ras). Thus, changes in the redox equilibrium influence a host of cell functions, including but not limited to growth, stress responses, differentiation, metabolism, cell cycle, communication, migration, gene transcription, ion channels, and immune responses (for reviews see Refs. 2Forman H.J. Torres M. Am. J. Respir. Crit. Care Med. 2002; 166: S4-S8Crossref PubMed Scopus (723) Google Scholar, 3Forman H.J. Torres M. Mol. Aspects Med. 2001; 22: 189-216Crossref PubMed Scopus (441) Google Scholar, 4Droge W. Physiol. Rev. 2002; 82: 47-95Crossref PubMed Scopus (7532) Google Scholar, 5Yanping L. David D.G. Clin. Exp. Pharmacol. Physiol. 2002; 29: 305-311Crossref PubMed Scopus (129) Google Scholar, 6Rojkind M. Dominguez-Rosales J.A. Nieto N. Greenwel P. Cell Mol. Life Sci. 2002; 59: 1872-1891Crossref PubMed Scopus (200) Google Scholar). Alterations in the redox equilibrium are reflected in changes of the glutathione/glutathione-disulfide ratio (GSH/GSSG) and the reduced/oxidized thioredoxin ratio. Glutathione is found in high concentrations in cells (5-10 mm) and is considered to be the major player in maintaining intracellular redox equilibrium. Ratios of GSH to GSSG are reported to range from 100 to 300:1 (7Hwang C. Sinskey A.J. Lodish H.F. Science. 1992; 257: 1496-1502Crossref PubMed Scopus (1598) Google Scholar, 8Liu S. Ansari N.H. Wang C. Wang L. Srivastava S.K. Curr. Eye Res. 1996; 15: 726-732Crossref PubMed Scopus (4) Google Scholar), but these measurements have been problematic because they require destruction of the tissue, during which great care must be taken not to allow further oxidation. The major source of error is the determination of GSSG concentration, because this species is at low abundance yet is measured only after complete removal of GSH to prevent oxidation. The spatial and temporal resolution of such destructive assays is very poor. Redox-sensitive green fluorescent proteins (GFP) 1The abbreviations used are: GFP, green fluorescent protein; AEBSF, 4-(2-aminoethyl)benzenesulfonyl fluoride; BMES, bis(2-mercaptoethyl)sulfone; BSO, buthionine sulfoximine; DMNQ, 2,3-dimethoxy-1,4-naphthoquinone; DTT, dithiothreitol; DTTox, trans-4,5-dihydroxy-1,2-dithiane; roGFP, redox-sensitive GFP; EGF, epidermal growth factor; IMDM, Iscove's modified Dulbecco's medium; PMA, phorbol 12-myristate 13-acetate; ROS, reactive oxygen species; LPA, lysophosphatidic acid. described recently (9Hanson G.T. Aggeler R. Oglesbee D. Capaldi R.A. Tsien R.Y. Remington S.J. J. Biol. Chem. 2004; 279: 13044-13053Abstract Full Text Full Text PDF PubMed Scopus (751) Google Scholar) allow real-time visualization of the oxidation state of the indicator. The indicators examined in this work are GFP mutants with two surface-exposed cysteine placed at positions 147 and 204 on adjacent β-strands close to the chromophore. Disulfide formation between the cysteine residues promotes protonation of the chromophore and increases the excitation spectrum peak near 400 nm at the expense of the peak near 490 nm. The ratios of fluorescence from excitation at 400 and 490 nm indicate the extent of oxidation and thus the redox potential while canceling out the amount of indicator and the absolute optical sensitivity. Because the indicator is genetically encoded, it can be targeted to specific proteins or organelles of interest and expressed in a wide variety of cells and organisms. Here we evaluate roGFP1 (GFP with mutations C48S, S147C, and Q204C) and roGFP2 (the same plus S65T) with physiologically or toxicologically relevant oxidants both in vitro and in living mammalian cells. The probes expressed in cell cytoplasm responded as expected to a variety of oxidants. Although lower concentrations of hydrogen peroxide were required to oxidize cytosolic roGFPs than oxidize the same proteins in vitro, attempts to detect growth factor-stimulated production of hydrogen peroxide were not successful. However, the probes could detect superoxide generated during the oxidative burst of HL60 cells. Aldrithiol, diamide, hydrogen peroxide, buthionine sulfoximine, 3-amino-1,2,4-triazole, apocynin (acetovanillone), and 4-(2-aminoethyl)-benzenesulfonyl fluoride were obtained from Aldrich. Oxidized and reduced lipoate, oxidized and reduced glutathione, menadione, diphenylene iodonium chloride, BCNU (1,3-bis-(2-chloroethyl)-1-nitrosourea) (carmustine), and apocynin were obtained from Sigma. 2,3-Dimethoxy-1,4-naphthoquinone (DMNQ) and DTT were obtained from Calbiochem. Bis(2-mercaptoethyl)sulfone (BMES) was from USB Corp. Its cyclic disulfide was prepared by oxidation with a stoichiometric quantity of iodine (10Goodrow M.H. Musker W.K. Synthesis. 1981; : 457-459Crossref Scopus (38) Google Scholar) followed by recrystallization from hexane. HL60 cells were obtained from American Type Culture Collection. In Vitro Studies—roGFPs were subcloned into pRSETB (Invitrogen) using BamHI and EcoRI restriction sites. The plasmid encodes a fusion protein of the insert and an N-terminal extension containing a (His)6 tag, enabling purification by nickel affinity chromatography. The construct was expressed in the JM109 strain of Escherichia coli. Isolated protein was reduced daily by mixing concentrated protein (50-200 μm) with 10 mm DTT and diluting to the required concentration. When necessary, DTT was removed from the solution using Centri-spin 20 columns (Princeton Separations Inc.). Reactions were carried out in 125 mm KCl, 75 mm HEPES, and 1 mm EDTA, pH 7.3, which had been degassed by repetitive evacuation and nitrogen bubbling. Excitation scans (350-500 nm, 2.5-nm bandwidth) were run in 100-μl volumes in 96-well plates (sealed when required) on a Safire spectrofluorometer (Tecan), collecting emission at 530 nm, 7.5-nm bandwidth. Redox Titration Using Fluorescence Spectroscopy—Redox probes were titrated in degassed HEPES buffer containing 1 μm protein and 10 mm lipoate or BMES as redox buffers. Concentrations of oxidized and reduced forms were reciprocally varied from 0:10 to 10:0 mm in 1 mm increments. The eleven solutions comprising each titration series were at or 1 while under a nitrogen atmosphere before fluorescence excitation The redox potential of the redox probes using BMES or lipoate as redox were obtained by the to each of the concentrations in the ratio at a of redox ratio in 10 mm reduced redox and ratio in 10 mm oxidized redox a of redox redox the buffer ratio required to of the ratio can be from the was from The redox potential was from the in is the redox potential of the redox J. Chem. Scopus Google BMES is the is is the of and is Redox of BMES Using containing 10 mm and 10 mm oxidized BMES and solution containing 10 mm oxidized lipoate and 10 mm reduced BMES were prepared in buffer in were on a at K. The of solutions and taken a after were that equilibrium had been specific to each of the both oxidized and reduced were oxidized lipoate oxidized BMES and reduced BMES the equilibrium between and BMES, we used the under the the into the a of lipoate J. Chem. Scopus Google the redox potential of BMES in was to be by to superoxide was by 1 μm with of had been to and μm in of mm pH in were expressed in or HL60 cells using modified as and as of at in the cells were with solution Cells were on a with a by ratio used excitation and roGFP1 and and roGFP2 by a and an emission were used both Fluorescence were by of time was and were taken Cell HL60 cells were obtained on a with Cells were using at nm and nm were both excitation HL60 cells were in the of in plus The cell was in and in 1 of was to the solution and at 10 The cells were using a (Invitrogen) using a and Cells were in before oxidation Cells were and in and of cells were in μm or 100 μm cells were in either μm 4-(2-aminoethyl)benzenesulfonyl fluoride or μm mutations of and of roGFP1 and roGFP2 were using a with and with The in pRSETB (Invitrogen) were expressed using JM109 and as described in mammalian cells, the were subcloned into (Invitrogen) using BamHI and EcoRI restriction sites. a and was to the of roGFP1 and The was from a tyrosine protein to the The and were used to the into the was to the of the redox probes using to the construct in the redox probes were to using BamHI and restriction to the fluorescent protein and of Redox that roGFP1 and roGFP2 in the of mammalian cells were it was to the of reduced roGFP1 and roGFP2 in The proteins were expressed in and were reduced each were carried out in and (9Hanson G.T. Aggeler R. Oglesbee D. Capaldi R.A. Tsien R.Y. Remington S.J. J. Biol. Chem. 2004; 279: 13044-13053Abstract Full Text Full Text PDF PubMed Scopus (751) Google Scholar) measured the redox of roGFP1 and roGFP2 to be and using DTT as the redox The between the roGFPs DTT that is with very ratios of to It be to the roGFPs using redox with to of We and BMES because they internal oxidation, to DTT and roGFPs with these with a Furthermore, both of these are which cell in of intracellular of the protein in redox of ratios of reduced to oxidized were from which the of the redox proteins were we found redox when using lipoate or BMES as the roGFP1 and roGFP2 by the Because BMES R. J. Chem. Scopus Google Scholar) had only been in we used to the BMES in a solution to the When the 1 was used to the we obtained close between the obtained in the two buffers. and roGFP1 and are than the redox reported by (9Hanson G.T. Aggeler R. Oglesbee D. Capaldi R.A. Tsien R.Y. Remington S.J. J. Biol. Chem. 2004; 279: 13044-13053Abstract Full Text Full Text PDF PubMed Scopus (751) Google Scholar) on DTT as and the of roGFP1 as reducing than roGFP2 is of Redox in roGFPs were reduced cells oxidants were we the in vitro of roGFPs to same oxidants. roGFPs are oxidized in vitro, they were reduced to and in degassed and the extent of oxidation by concentrations of oxidants after and of and were very reactive and oxidized roGFPs in at of the concentrations GSSG and the oxidized of BMES, required and concentrations to oxidize peroxide was not very in vitro at concentrations We two used to redox in cells. an in emission ratio with oxidation, but ratios and in fluorescence emission at of the excitation of the proteins after the of with the of of to the of by has been reported D. D. L. S. Biochem. PubMed Scopus Google Scholar, M. S. Biochem. Pharmacol. PubMed Scopus Google Scholar). GSH with to the as as hydrogen peroxide and superoxide D. S. P. Chem. Biol. PubMed Scopus Google Scholar). Chem. Biol. PubMed Scopus Google Scholar), reported to redox not fluorescence but only oxidized roGFPs after a and time oxidation of roGFP2 than of roGFP1 is a has a than of Redox in roGFPs were expressed in cells, was the and and ratios were of the of the such as the excitation ratio nm nm a few in a out the the cells reduced the DTT this these we to use of and oxidants to and ratios in However, we could not ratios as as from reduced or oxidized protein in of buffer with the same on the in vitro roGFP1 in cells was a redox potential of the of mm DTT or mm to the extent of roGFP1 oxidation to and cells roGFP2 after which DTT reduced the ratio to and oxidized the protein to a ratio than in of buffer on the Thus, roGFP2 a redox potential of with the by range of ratios in roGFP1 and roGFP2 were expressed in cells. of fluorescence obtained at or and ratio was protein under the same protein expressed in cells. Cells roGFPs were with mm of the oxidants in vitro with the of which not cell were both roGFP1 and roGFP2 The of menadione, oxidized BMES, and lipoate was to that in vitro, but mm hydrogen peroxide or the increases in ratio in cells within 10 1 mm concentrations were not of oxidation in Although oxidized BMES and lipoate were at at mm they oxidation. However, this was than that by mm we that redox from BMES and lipoate not intracellular redox to as in the of and in maintaining redox status under and conditions, cells were with buthionine an of GSH or a GSH by in a but in the excitation ratio of the and an extent of oxidation to μm) hydrogen peroxide with or results. with not the excitation ratio had on the oxidation but was as as at the response to 100 μm hydrogen of the reducing the roGFPs in the cell was but is because it is after removal of when is the of the are than We that of a with 100 μm J. Biol. Med. 2001; PubMed Scopus Google and μm N. D. H.J. L. Biol. Med. 22: PubMed Scopus Google both cells from reducing roGFPs after to BCNU with the roGFPs and could not be used of of The normally oxidation with 100 μm was a with μm or a with 100 μm and the in cells than in vitro is that in cells they can reactive oxygen species such as which with the The superoxide generated in vitro by of with was to oxidize roGFPs to a extent but was than oxidation in the However, the placed cells a and superoxide have to the to the the of superoxide to be cells. We considered the that generated from by be the reactive but the of to and roGFPs in vitro the protein fluorescence than the in emission ratio. be that or enzymes either with roGFPs or GSSG to oxidize the roGFPs by thiol-disulfide However, to in vitro with roGFPs with or the the response in cells Redox during roGFPs detect redox changes under physiological and not oxidants and as The was to the of from to 1 to from to with or were or cells in a but in the excitation ratios with when the cells had been with GSH or Thus, cytosolic roGFPs the intracellular redox to be very changes in in to their to and of the oxidative in mammalian cells is the oxidative burst in immune cells, which a major in W. Physiol. Rev. 2002; 82: 47-95Crossref PubMed Scopus (7532) Google Scholar). we expressed roGFPs in HL60 cells to into by in S.J. PubMed Google Scholar) and with to protein C. of time by was because of by by and of the cells. However, we were to the ratios by of 20 μm PMA, the roGFPs oxidation within could not be used to Biol. Med. PubMed Scopus Google Scholar), because with this oxidation of the roGFPs by was not in of the oxidation was with P. J. Biol. Chem. Full Text Full Text PDF PubMed Scopus Google Scholar) and apocynin A.J. J. Pharmacol. 2001; PubMed Scopus Google Scholar), two of of roGFP2 to production of has been in growth factor responses such as in cells with epidermal growth factor J. Biol. Chem. Full Text Full Text PDF PubMed Scopus Google Scholar) or cells with lysophosphatidic M. J. Biochem. J. PubMed Scopus Google Scholar). The concentrations of and during such signaling are but than used to during oxidative We were to detect oxidation of the the of to cells or to cells. a with not the reported in oxidation of this to the J. Biol. Chem. Full Text Full Text PDF PubMed Scopus Google Scholar). the is than the proteins because oxidation of the is that it and production of roGFPs are indicators that the redox equilibrium of the cell, which be the responses of the to growth factors was that the was increases in the redox potential of the cell, using we redox responses were near the growth factor on the We targeted roGFP2 to the using a N-terminal and from a tyrosine protein to the into and cells, the redox was to the to and to DTT were than of However, the protein to by or the roGFPs to the and the not in in activity to expressed in the the of roGFPs to of the roGFPs or which be to detect signaling Redox of are limited by the to the thiol before it can near cysteine residues increases to oxidation by Biochem. PubMed Scopus Google Scholar), by we a series of with residues placed in close to the cysteine and were with mutants were prepared in which both positions and or positions and were with proteins were and in to oxidation by of the proteins were into cells, but only and a response to than of the roGFP2 two mutants were responses to but oxidation of the roGFPs was not probes of redox potential many redox that the probes can be into cell or that can that the proteins can be targeted to specific or and that are of redox potential with a indicator is and in spatial and temporal resolution than using destructive assays of thiol and disulfide redox on fluorescent protein has been reported J. 2001; PubMed Scopus Google Scholar), but this only an not a In each disulfide formation protonation of the which but excitation of to nm as is not only of the of to such as cell and but to redox changes from such as of the as with and the of roGFPs changes in ratio from reduced to to in vitro and to in of mammalian cells. of redox reported by roGFPs We that roGFPs be in redox potential was a the proteins than that of we with and the of the had to be in a The of and roGFP1 and are only than the measured of (9Hanson G.T. Aggeler R. Oglesbee D. Capaldi R.A. Tsien R.Y. Remington S.J. J. Biol. Chem. 2004; 279: 13044-13053Abstract Full Text Full Text PDF PubMed Scopus (751) Google Scholar) on DTT, and (9Hanson G.T. Aggeler R. Oglesbee D. Capaldi R.A. Tsien R.Y. Remington S.J. J. Biol. Chem. 2004; 279: 13044-13053Abstract Full Text Full Text PDF PubMed Scopus (751) Google Scholar). we of and roGFP1 and (9Hanson G.T. Aggeler R. Oglesbee D. Capaldi R.A. Tsien R.Y. Remington S.J. J. Biol. Chem. 2004; 279: 13044-13053Abstract Full Text Full Text PDF PubMed Scopus (751) Google Scholar). Thus, the roGFPs require reducing than was reported to be J. 2001; PubMed Scopus Google Scholar). J. 2001; PubMed Scopus Google Scholar) reported that was oxidized when expressed in of the probes by a of oxidants was in vitro to oxidation before in oxidants such as and oxidized the probes in peroxide was found to be a 1 mm concentrations not oxidize the protein after a to a in fluorescence when at either and it was found to the cysteine of the redox probes as it has been found to with proteins D. D. L. S. Biochem. PubMed Scopus Google Scholar). However, the only oxidation after in When expressed in the cytoplasm of cells, roGFP1 and roGFP2 were either or a redox potential of and at pH The redox potential of the is (9Hanson G.T. Aggeler R. Oglesbee D. Capaldi R.A. Tsien R.Y. Remington S.J. J. Biol. Chem. 2004; 279: 13044-13053Abstract Full Text Full Text PDF PubMed Scopus (751) Google Scholar), because the pH of that is measured with roGFPs that the and are reducing than from measurements of intracellular GSH and GSSG to C. Lodish H.F. Sinskey A.J. PubMed Scopus Google Scholar, Biol. Med. 2001; PubMed Scopus Google Scholar). The during which has been reported to cells we that the redox in the cell are not in equilibrium with each the and thioredoxin redox have to be than the redox in to of to Although reduced roGFPs with GSSG in vitro, oxidized roGFPs are or to with GSH in is because the are that the of GSH with GSSG is to the at concentrations of roGFPs and with reducing such as they the potential of the reducing redox in the of oxidants to cells a in ratio of the Although the extent of the ratio and the time required to a the the followed that in in The major was hydrogen peroxide, which was found to at 100 μm concentrations in the cells, concentrations were required in Thus, we to the redox probes be to oxidation by such as superoxide and we found that an source superoxide J. Biol. Chem. Full Text PDF PubMed Google oxidation of the roGFPs both and Although reduced roGFPs in vitro, from to to the ratio of roGFPs expressed in a of cell including and not be to of the because has been found to be to hyperoxic J.A. Biochem. PubMed Scopus Google Scholar). of cells to not the ratio. The ratio under or to cells to oxidative It that the cells a oxidation in the which was not in the of such as superoxide as a response to cells high of a of superoxide as a of oxidation of to superoxide production in the activation of protein C. We used a cell HL60 to to with cells S.J. J. Exp. Med. PubMed Scopus Google Scholar). of the cell HL60 to to with roGFPs with an in the fluorescence ratio with oxidation of the The roGFP2 the response to in with to oxidation. When cells were with an diphenylene the oxidation by was We were that diphenylene iodonium the because the of oxidize the after with diphenylene iodonium not oxidation. However, of such as apocynin and AEBSF, which by the oxidation. have that as a in signaling by growth factors such as growth EGF, and and that R. P. J. PubMed Scopus Google Scholar) and M. 2001; PubMed Scopus Google Scholar). It has been that high of and that GSH between as an oxidative or a in growth factor K. J. Biol. Chem. Full Text Full Text PDF PubMed Scopus Google Scholar). Using we were to response of the redox probes to two growth factor reported to in cells or in cells. the to the or the roGFPs to to to to in cells or by in cells, the modified probes to oxidants. we in that the in cells responded to EGF, has been reported to be to oxidation or C. J. Biol. Chem. Full Text Full Text PDF PubMed Scopus Google Scholar, Biol. Med. PubMed Scopus Google Scholar). we that the growth factors in redox is a in thiol-disulfide redox it have to be We L. and P. and C.