Imaging of Conformational Changes of Proteins with a New

May 26, 2001 - Department of Chemistry, School of Science, The University of Tokyo, Hongo, Bunkyo-ku, Tokyo, 113-0013, Japan, Japan ..... Its pKa (∼...
0 downloads 0 Views 186KB Size
Download PDF
Anal. Chem. 2001, 73, 2920-2928

Imaging of Conformational Changes of Proteins with a New Environment-Sensitive Fluorescent Probe Designed for Site-Specific Labeling of Recombinant Proteins in Live Cells Jun Nakanishi,†,‡ Takahiro Nakajima,† Moritoshi Sato,†,‡ Takeaki Ozawa,†,‡ Kohji Tohda,†,‡ and Yoshio Umezawa*,†,‡

Department of Chemistry, School of Science, The University of Tokyo, Hongo, Bunkyo-ku, Tokyo, 113-0013, Japan, Japan Science and Technology Corporation (JST), Tokyo, Japan.

We demonstrate herein a new method for imaging conformational changes of proteins in live cells using a new synthetic environment-sensitive fluorescent probe, 9-amino6,8-bis(1,3,2-dithioarsolan-2-yl)-5H-benzo[a]phenoxazin5-one. This fluorescent probe can be attached to recombinant proteins containing four cysteine residues at the i, i + 1, i + 4, and i + 5 positions of an r-helix. The specific binding of the fluorescent probe to this 4Cys motif enables fluorescent labeling inside cells by its extracellular administration. The high sensitivity of the fluorophore to its environment enables monitoring of the conformational changes of the proteins in live cells as changes in its fluorescence intensity. The present method was applied to calmodulin (CaM), a Ca2+-binding protein that was wellknown to expose hydrophobic domains, depending on the Ca2+ concentration. A recombinant CaM fused at its C-terminal with a helical peptide containing a 4Cys motif was labeled with the fluorescent probe inside live cells. The fluorescence intensity changed reversibly depending on the intracellular Ca2+ concentration, which reflected the conformational change of the recombinant CaM in the live cells. Environment-sensitive fluorescent probes1-3 are fluorophores whose fluorescence wavelengths and quantum yields are quite sensitive to the change in hydrophobicity of their environment. Proteins labeled with these fluorescent probes have been thereby utilized for studying conformational changes of proteins.4-10 For * To whom correspondence should be addressed. Phone: +81-3-5841-4351. Fax: +81-3-5841-8349. E-mail: [email protected]. † The University of Tokyo. ‡ Japan Science and Technology Corporation. (1) Slavik, J. Biochim. Biophys. Acta 1982, 694, 1-25. (2) Weber, G.; Farris, F. J. Biochemistry 1979, 18, 3075-3078. (3) Diwu, Z.; Lu, Y.; Zhang, C.; Klaubert, D. H.; Haugland, R. P. Photochem. Photobiol. 1997, 66, 424-431. (4) Walkup, G. K.; Imperiali, B. J. Am. Chem. Soc. 1997, 119, 3443-3450. (5) Richieri, G. V.; Ogata, R. T.; Kleinfeld, A. M. J. Biol. Chem. 1992, 267, 23495-23501. (6) Hahn, K. M.; Waggoner, A. S.; Taylor, D. L. J. Biol. Chem. 1990, 265, 20335-20345. (7) Marvin, J. S.; Corcoran, E. E.; Hattangadi, N. A.; Zhang, J. V.; Gere, S. A.; Hellinga, H. W. Proc. Natl. Acad. Sci. U.S.A. 1997, 94, 4366-4371.

2920 Analytical Chemistry, Vol. 73, No. 13, July 1, 2001

this purpose, it is desirable to attach a single probe molecule at a specific site of a target protein; however most of the conventional environment-sensitive probes11 are reactive to the thiol group of generally existing cysteines, and accordingly, neither sites nor numbers of attached probe molecules are controllable. To rationally label a single probe molecule at a specific site of a particular protein, fluorescent labeling has been performed to the purified recombinant protein to which a single cysteine was introduced at the desired site, and other reactive cysteines were omitted by targeted mutagenesis.7-10 To utilize the labeled proteins for studies in live cells, the proteins have to be microinjected into the cells. Because of these complicated processes, only a few studies of their use in live cells have been reported.10,12 Recently, a potentially general approach for fluorescent labeling of proteins inside cells was developed by Tsien and co-workers.13,14 It is based on the specific binding of two appropriately spaced arsenics within a bisarsenical probe molecule to the genetically engineered four cysteines at the i, i + 1, i + 4, and i + 5 positions in an R-helix (4Cys motif) of a target protein. By taking advantage of this new technology, we intended to develop for the first time an environment-sensitive fluorescent probe that can be attached to a specific site (4Cys motif) of proteins inside cells (Figure 1a). By using this probe molecule, conformational changes of proteins in live cells were visualized (Figure 1b). We designed and synthesized a new environment-sensitive fluorescent probe, 9-amino-6,8-bis(1,3,2-dithioarsolan-2-yl)-5H-benzo[a]phenoxazin-5-one, which is an analogue of nile red with a bisarsenical moiety. Nile red is a well-known environment-sensitive fluorescent probe whose fluorescence is strongly influenced by changes in its molecular environment.15,16 The present probe molecule was applied to detect the conformational change of (8) Gether, U.; Lin, S.; Ghanouni, P.; Ballesteros, J. A.; Weinstein, H.; Kobilka, B. K. EMBO J. 1997, 16, 6737-6747. (9) Dunham, T. D.; Farrens, D. L. J. Biol. Chem. 1999, 274, 1683-1690. (10) Post, P. L.; Trybus, K. M.; Taylor, D. L. J. Biol. Chem. 1994, 269, 1288012887. (11) Haugland, R. P. Handbook of Fluorescent Probes and Research Chemicals, 6th ed.; Molecular Probes, Inc.: Eugene, OR, 1996. (12) Hahn, K.; DeBiasio, R.; Taylor, D. L. Nature 1992, 359, 736-738. (13) Griffin, B. A.; Adams, S. R.; Tsien, R. Y. Science 1998, 281, 269-272. (14) Griffin, B. A.; Adams, S. R.; Jones, J.; Tsien, R. Y. Methods Enzymol. 2000, 327, 565-578. (15) Sackett, D. L.; Wolff, J. Anal. Biochem. 1987, 167, 228-234. 10.1021/ac001528p CCC: $20.00

© 2001 American Chemical Society Published on Web 05/26/2001

Figure 1. Schematic representation of the present method: (a) The recombinant calmodulin (CaM) containing genetically encoded labeling site (4Cys motif: 4 cysteines at the i, i + 1, i + 4, and i + 5 positions of an R-helix) is specifically labeled with a new environment-sensitive probe (BArNile-EDT2) by its extracellular administration; (b) high environment sensitivity of the fluorescent probe enables the detection of conformational change of calmodulin as change in its fluorescence intensity and emission wavelength.

calmodulin (CaM), which is known to expose hydrophobic domains, depending on the Ca2+ concentration.15,17 Recombinant CaM fused at its C-terminal with a helical peptide containing a 4Cys motif could be labeled with the fluorescent probe in live cells, and Ca2+-dependent fluorescence intensity changes were imaged, which reflected the conformational change of CaM. This is the first demonstration of an environment-sensitive probe molecule that can be attached in vivo to a target protein and can be used for visualizing the conformational changes of proteins in live cells. EXPERIMENTAL SECTION Materials. All reagents used for the present organic synthesis were purchased from Tokyo Kasei Kogyo Co. (Tokyo, Japan) or Wako Pure Chemical Industries (Osaka, Japan). Fura 2, O,O′-bis(2-aminoethyl)ethyleneglycol-N,N,N′,N′-tetraacetic acid (EGTA), and 2-[4-(2-hydroxyethyl)-1-piperazynyl]ethanesulfonic acid (HEPES) were purchased from Dojindo Laboratories (Kumamoto, Japan). Ionomycin, ATP, pluoric F127, and 3-[(3-cholamidopropyl)dimethylammonio]-1-propane-sulfonate (CHAPS) were purchased from Sigma Chemical Co. (St. Louis, MO). pGEX-6P-2, PreScission Protease, and Glutathione Sepharose 4B were purchased from Amarsham Pharmacia Biotech. (Uppsala, Sweden). Mammalian expression vector pcDNA3.1(+) was from Invitrogen Co. (Carlsbad, CA), and restriction enzymes were from Takara (Kyoto, Japan). Hank’s balanced salt solution (HBSS), Dulbecco’s modified Eagle medium (DMEM), fetal calf serum (FCS), and lipofectAMINE 2000 reagent were from Life Technologies (Rockville, MD). Synthesis of the Fluorescent Probe. 1H NMR spectra were obtained on a JEOL R500 spectrometer (500 MHz; JEOL, Japan). Mass spectra were obtained on a JEOL SX-102 spectrometer (JEOL, Japan). (16) Greenspan, P.; Mayer, E. P.; Fowler, S. D. J. Cell Biol. 1985, 100, 965973. (17) Chabbert, M.; Piemont, E.; Prendergast, F. G.; Lami, H. Arch. Biochem. Biophys. 1995, 322, 429-436.

9-Nitro-5-benzophenoxazone (1). Compound 1 was synthesized by the reported procedure18 with a slight modification. To 2-hydroxy-1,4-naphthoquinone (23.7 g, 0.136 mol) dissolved in 80% acetic acid (200 mL), 2-hydroxy-4-nitroaniline (20.7 g, 0.134 mol) was added and mixed for 12 h at 100 °C. After dilution of the reaction mixture with water, the solution was neutralized with Na2CO3. The precipitate was extracted several times with acetic acid, and the extracts were diluted with water. The formed precipitate was washed with diluted NaOH until the filtrate became colorless. The obtained yellow precipitate was purified using a silica gel chromatography eluted with dichloromethane (RF ) 0.45, 3.27 g, 8.33%). Anal. Calcd. for C16H8N2O4: C, 65.76; H, 2.76; N, 9.59. Found: C, 65.33; H, 3.04; N, 9.38. 9-Amino-5-benzophenoxazone (2). Compound 1 (2.69 g, 9.21 mmol) and Pd/C (0.275 g) were suspended in 54 mL of methanol under H2 and mixed until the spot corresponding to 1 [dichloromethane:ethyl acetate ) 6:1 (v:v), RF ) 0.83] disappeared on a silica gel TLC. The reaction mixture was dissolved in pyridine and filtered using Celite. The filtrate was diluted with water to form precipitate. The precipitate was filtered and washed with acetone and then dichloromethane until giving a single spot on a silica gel TLC [dichloromethane:ethyl acetate ) 9:1 (v:v), RF ) 0.35] (1.10 g, 45.7%). 1H NMR (DMSO-d6): δ 6.29 (s, 1H, ArH), 6.52 (d, 1H, ArH, J ) 2.0 Hz), 6.68 (d, 1H, ArH, J ) 9.0 Hz), 6.69 (s, 2H, NH2), 7.54 (d, 1H, ArH, J ) 9.0 Hz), 7.71 (dd, 1H, ArH, J ) 7.0 Hz), 7.79 (dd, 1H, ArH, J ) 7.3 Hz), 8.11 (d, 1H, ArH, J ) 7.5 Hz), 8.53 (d, 1H, ArH, J ) 8.0 Hz). Anal. Calcd for C16H10N2O2: C, 73.27; H, 3.84; N, 10.68. Found: C, 73.01; H, 4.05; N, 10.51. 9-Amino-6,8-bis(acetoxymercury)-5-benzophenoxazone (3). To the stirred suspension of mercury acetate (II) (1.65 g, 5.18 mmol) in 32 mL of acetic acid, 0.518 g (1.98 mmol) of 2 in 26 mL of acetic acid was slowly added and stirred at 50 °C overnight. The mixture was diluted with water and filtered. The obtained precipitate was washed with water, acetone, and then dichloromethane and dried in vacuo. This crude product was used without purification (1.41 g, 1.81 mmol). (18) Kehrmann, F.; Gauhe, E. Ber. Deut. Chem. Ges. 1897, 30, 2130-2138.

Analytical Chemistry, Vol. 73, No. 13, July 1, 2001

2921

Figure 2. Schematic representations of recombinant proteins containing a 4Cys motif used in this study: 4Cys-helix composed of 17 amino acids (AEAAAREACCRECCARA) contains a 4Cys motif and forms an R-helix, even in water. CaM, GST, and YFP represent Xenopus laevis calmodulin, gluthathione-(S)-transferase, and a yellow color mutant of the green fluorescent protein, respectively.

9-Amino-6,8-bis(1,3,2-dithioarsolan-2-yl)-5-benzophenoxazone (BArNile-EDT2). To 1.38 g of crude 3 suspended in 31.6 mL of N-methylpyrrolidone (NMP) was added 3 mL of arsenic trichloride (35.7 mmol) and 2.5 mL of N,N-diisopropylethylamine (DIEA) (14.4 mmol) with trace palladium acetate (II) as a catalyst. The reaction was performed at room temperature overnight and then was quenched using 200 mL of 80% acetone. The solution was neutralized with Na2CO3, 9.6 mL of ethanedithiol (EDT) was added, and the mixture was mixed overnight. The mixture was extracted with dichloromethane and washed with water, and the solvent was removed by evaporation. The residue was subjected to silca gel chromatography and eluted with dichloromethane in order to remove colorless byproduct. After eluting with dichloromethane:ethyl acetate [50:1 (v:v), RF ) 0.23], BArNile-EDT2 (68.5 mg, 6.5%) was obtained. 1HNMR (DMSO-d6): δ 6.75 (d, 1H, ArH, J ) 8.5 Hz), 7.24 (s, 2H, NH2), 7.54 (d, 1H, ArH, J ) 8.5 Hz), 7.70 (t, 1H, ArH, J ) 7.5 Hz), 7.79 (t, 1H, ArH, J ) 7.8 Hz), 8.05 (d, 1H, ArH, J ) 8.0 Hz), 8.50 (d, 1H, ArH, J ) 8.0 Hz); (C5D5N) δ 3.54 (m, 8H, -SCH2CH2S-). MS (FAB+), m/z (M + H)+ 595 (calcd for C20H16As2N2O2S4, 594.85). Plasmid Construction. The recombinant proteins containing a 4Cys motif used in this study are shown in Figure 2. The 4Cyshelix composed of 17 amino acids (AEAAAREACCRECCARA) contains a 4Cys motif and forms an R-helix, even in water.13,19 The cDNAs were generated by a standard polymerase chain reaction and inserted at BamHI and XhoI sites of either pGEX-6P-2 or pcDNA3.1(+) vectors for study outside or inside cells, respectively. Extraction of GST Fusion Proteins. GST fusion proteins were expressed in Escherichia coli and purified with a glutathione sepharose column. In the case of the GST-CaM-helix, the GST tag was removed using PreScission protease and again purified using a glutathione sepharose column. Protein concentrations were determined using a Biorad protein assay kit (Bio-rad Lab.; Hercules, CA) with bovine gamma globlin as a standard. Specific Labeling of the 4Cys Motif with BArNile-EDT2. The 4Cys motif rarely exists in endogenous proteins, which is (19) Marutka, G.; Shalongo, W.; Stellwagen, E. Biochemistry 1991, 30, 42454248.

2922 Analytical Chemistry, Vol. 73, No. 13, July 1, 2001

essential for the site-specific labeling of bisarsenical compounds to the 4Cys motif inside cells. However, mere thiol residues also react with BArNile to a smaller extent. To decrease this nonspecific labeling, the fluorescent labeling should be performed in the presence of a low concentration of EDT,13 which dissociates BArNile from undesired sites. Although this trace EDT reduced nonspecific labeling, it also lowered the efficiency of labeling to the 4Cys motif. We determined that with 1 µM EDT, BArNile was able to label the 4Cys motif, and the nonspecific labeling could be reduced (data not shown). Therefore, in all of the experiments, fluorescent labeling was performed in the presence of 1 µM EDT. The fluorescent probe (at a final concentration 25 µM) was mixed with 30 µM GST-helix, in 10 mM HEPES buffer (pH 7.4) containing 1 mM mercaptoethanol, 1 µM EDT, 1% CHAPS, and 140 mM KCl, at 37 °C overnight. In the control experiment, GST was incubated with BArNile-EDT2 overnight under otherwise identical conditions. The reaction mixture was applied to a SMART system (Pharmacia Biotech.; Uppsala, Sweden) equipped with a C18 reverse-phase column. The eluting conditions were as follows: 100 µL/min; gradient of 0-100% acetonitrile in Milli-Q water, both containing 0.1% trifluoroacetic acid; and eluant absorbance monitored at 280 and 520 nm. Fluorescence Measurements with a Fluorescence Spectrometer. The fluorescence change of the probe upon its binding to the 4Cys motif was measured at 604 nm (excitation 520 nm) using a fluorescence spectrometer (JASCO Co.; Tokyo, Japan) in a quartz crystal cell with stirring at room temperature. The initial buffer composition was PBS (pH 7.4) containing 1 µM BArNileEDT2, 1 mM mercaptoethanol, 0.1 mM EGTA, and 1 µM EDT. To this solution, GST-helix (final concentration, 10 µM) was added. After 90 min, 1 mM EDT was added for dissociating the probe from the 4Cys motif. Ca2+-Dependent Fluorescence Changes of the CaM-Helix Labeled with the Fluorescent Probe. The fluorescent probe (1 µM) was mixed with 10 µM CaM-helix in the 10 mM HEPES buffer solution (pH 7.4) containing 140 mM KCl, 1 mM EGTA, 1 mM mercaptoethanol and 1 µM EDT, for 8 h at 25 °C. The observed fluorescence intensity was stable, and the labeling reaction reached its equilibrium. Although the fluorescence

Figure 3. Outline of the synthesis of BArNile-EDT2.

measurements were performed without removing unreacted BArNil-EDT2, the effect of the remaining BArNile-EDT2 on the fluorescence of BArNile attached to the CaM-helix was small, which had been checked by the control experiment using BArNile-EDT2 and CaM at differing Ca2+ concentrations (data not shown). Ca2+ concentration was changed by the addition of the above buffer solution containing 20 mM CaCl2, and changes in the sample volume were corrected for. The free Ca2+ concentration was calculated from the proton and Ca2+ association constants of EGTA at I ) 0.15 M, 22 °C (pK1-pK4 and pKCa ) 9.58, 8.97, 2.8, 2.12, and 10.955).20 The corresponding Hill constant and Hill coefficient for the binding of Ca2+ were obtained from changes in the fluorescence intensity normalized to the full Ca2+ saturation. Cell Culture and Transfection. HEK293 cells were cultured in DMEM supplemented with 10% FCS, 1 mM sodium pyruvate, and 0.1 mM nonessential amino acids at 37 °C in 5% CO2. The cells were transfected with LipofectAMINE 2000 reagent in 24well plates. During 8-16 h after transfection, the cells were transferred to glass-bottom dishes for imaging. Specific Labeling of the 4Cys Motif with BArNile-EDT2 inside Live Cells. During 3 to 5 days after transfection, the cells transfected with CaM-helix-YFP were washed with HBSS. The fluorescent labeling was performed in HBSS containing 0.01% pluoric F127, 1 µM BArNile-EDT2, and 1 µM EDT, and incubated for 3-5 h at 37 °C in 5% CO2. Before measurement, the cells were washed with HBSS three times to stop the progress of the labeling reaction. Images were obtained using a laser confocal microscope (LSM 510, Carl Zeiss; Planegg, Germany). Fluorescence images were obtained by illuminating an Ar laser at 514 nm and collecting at 535 to 590 nm and > 635 nm, which corresponds to YFP and BArNile, respectively. During the labeling procedure, the shapes and locations of the cells changed, which prevented the proper comparison of the fluorescence before and after labeling; hence, to evaluate the specificity of BArNile to the 4Cys motif, changes in the fluorescence intensity induced by the addition of 1 mM EDT were utilized. Imaging of the Conformational Change of CaM in Live Cells. Fluorescent labeling was performed under the same

conditions as described in the previous section. The cells that emitted bright fluorescence were used. Fluorescence images were recorded at room temperature on a Carl Zeiss Axiovert S100 microscope equipped with a cooled CCD camera. Cells were illuminated at 550 nm using a Polychrome II monochromatizing device (Carl Zeiss; Planegg, Germany), and fluorescence was collected using a 590-nm-long path filter. The images obtained were analyzed using TILLvisION (T.I.L.L. Photonics GmbH; Martinsried, Germany), a software program for data analysis. RESULTS AND DISCUSSION Design and Synthesis of an Environment-Sensitive Fluorescent Probe which Binds to the 4Cys Motif. The bisarsenical compound used in the study of Griffin et al. (FlAsH)13 was a derivative of fluorescein, to which two arsenic atoms were introduced at the 4 and 5 positions of xanthene. On the basis of this structure, we chose nile red (9-diethylamino-5H-benzo[a]phenoxazin-5-one), a well-known environment-sensitive probe,15,21 as a good candidate for the present purpose; however, on the basis of its CPK model, a bulky diethylamino group at the 9 position of nile red appeared to hinder the entrance of ethanedithiol. Therefore, we used a 9-amino analogue of nile red [Figure 3(2)] as a fluorophore. This compound was also environment-sensitive, because its fluorescence spectra were highly dependent on the solvent (Figure 4). Its pKa (∼1) is thought to be well below the physiological pH. The final product, 9-amino-6,8-bis(1,3,2-dithioarsolan-2-yl)-5H-benzo[a]phenoxazin-5-one, will be hereafter called BArNile-EDT2 (bisarsenical nile red analogue, bis-EDT adduct) for convenience. The synthesis of BArNile-EDT2 is outlined in Figure 3. To introduce arsenic atoms at the 6 and 8 positions, 2 was mercurated, followed by transmetalation with arsenic trichloride.13 The final bis-EDT adduct was obtained by reacting with excess EDT and purified with silica gel chromatography. Specific Binding of BArNile to the 4Cys Motif Studied by HPLC. Direct evidence for covalent binding of BArNile to the (20) Harrison, S. M.; Bers, D. M. Am. J. Physiol. 1989, 256, C1250-C1256. (21) Deye, J. F.; Berger, T. A.; Anderson, A. G. Anal. Chem. 1990, 62, 615622.

Analytical Chemistry, Vol. 73, No. 13, July 1, 2001

2923

Figure 4. Emission spectra of 9-amino-5-benzophenoxazone (the fluorophore of BArNile) in different solvents (excited at 550 nm).

Figure 5. Specific binding of BArNile to the 4Cys motif studied by HPLC. 1 µM BArNile-EDT2 overnight incubated with GST-helix (b, c, e), GST (d) or no protein (a) at 4 °C were applied to a SMART system equipped with a C18 reverse-phase column. In c, 1 mM EDT was added just before the application. The eluant absorbance was monitored at 520 nm (a-d) or 280 nm (e).

4Cys motif was demonstrated by a shift in the HPLC retention time. The reaction mixture of BArNile-EDT2 and GST-helix was applied to a C18 reverse-phase column and the eluant absorbance was monitored at 520 nm, which corresponded to the absorption of BArNile (Figure 5). BArNile-EDT2 eluted at 27 min (trace a). When BArNileEDT2 was incubated overnight with a small excess of GST-helix, a new peak appeared at 21 min (trace b), together with the original peak at 27 min. This new peak was ascribed to BArNile attached to the GST-helix, because trace e, in which eluant absorbance was monitored at 280 nm, indicated that the labeled and unlabeled proteins eluted at 21 min. Addition of 1 mM EDT to the same reaction mixture reduced the peak at 21 min (trace c), which showed that the labeling proceeded following the reversible reaction

BArNile-EDT2 + 4Cys motif a BArNile-4Cys motif + 2EDT (1)

For a control experiment, GST without 4Cys motif was mixed with 2924 Analytical Chemistry, Vol. 73, No. 13, July 1, 2001

Figure 6. Fluorescence time profile of BArNile upon binding to and dissociating from the 4Cys motif: In a quartz crystal cell, 10 µM GSThelix was added to PBS buffer (pH 7.4) containing 1 µM BArNileEDT2, 1 mM mercaptoethanol, 0.1 mM EGTA, and 1 µM EDT at 12 min. The labeling was reversed by the addition of 1 mM EDT (excitation ,550 nm; emission, 604 nm).

BArNile-EDT2 overnight. Trace d was almost the same as that of BArNile-EDT2 (trace a). These results confirmed that the binding of BArNile to the 4Cys motif actually occurred with the reaction described above, and the labeling was specific to the 4Cys motif. As shown in trace b, there still remained unreacted BArNileEDT2 after the incubation overnight. This was because the labeling was performed in the presence of 1 µM EDT in order to reduce nonspecific labeling (see Experimental Section), and EDT shifted the equilibrium to the left of eq 1. The Fluorescence Change of BArNile upon Its Binding to the 4Cys Motif. Next, the fluorescence change of BArNile upon binding to the 4Cys motif was evaluated (Figure 6). When 10 µM GST-helix was added to the solution containing 1 µM of BArNile-EDT2, a time-dependent fluorescence increase was observed. This fluorescence increase was reversed by the addition of a high concentration (1 mM) of EDT. This fluorescence profile is explained as follows. The rotation of two C-As bonds in BArNile-EDT2 quenches the excited state by vibrational deactivation or intramolecular charge transfer. Upon binding of BArNile to the 4Cys motif, the rigid conformation of a BArNile-helix complex hinders conjugation of the arsenic lone-pair electrons with the π orbitals of the fluorophore, which causes its increased fluorescence. A high concentration of EDT promotes the reverse reaction of the above equilibrium, and dissociation of BArNile from the 4Cys motif results in a decrease in fluorescence. The fluorescence increase of BArNile upon binding to the 4Cys motif was smaller than that of FlAsH.13 A possible reason for this is that aminobenzophenoxazone, the fluorophore of BArNile, fluoresces weakly in water and strongly in nonpolar solvents, whereas fluorescein (fluorophore of FlAsH) is highly fluorescent in water. Specific Labeling of BArNile to the 4Cys Motif inside Live Cells. To confirm the specific binding of BArNile to the 4Cys motif inside live cells, we measured fluorescence energy transfer (FRET) between BArNile and a yellow color mutant (YFP) of the green fluorescent protein (Figure 7a). As shown in Figure 7b, the fluorescence of BArNile (∼600 nm) increased and that of YFP

Figure 7. Confirmation with FRET of specific labeling of BArNile to the 4Cys motif: (a) Schematic representation of this experiment; (b) emission spectra of GST-CaM-helix-YFP and GST-CaM-helix-YFP incubated with BArNile-EDT2 for 100 min and after the addition of 1 mM EDT (excitated at 500 nm).

Figure 8. Observation of dissociation of BArNile from a 4Cys motif of CaM-helix-YFP in the cells upon loading with BArNile-EDT2: HEK293 cells transiently expressing CaM-helix-YFP were first loaded with 1 µM BArNile-EDT2 for 3 h, followed by treatment with 1 mM EDT. Fluorescence images corresponding to (a) YFP (530 to 590 nm); (b) BArNile (> 635 nm) before (left) and after (right) addition of 1 mM EDT to the cells (excitation 514 nm); (c) phase-contrast image; scale bar, 50 µm; (d) fluorescence changes of YFP and BArNile of two selected cells shown in (c) before and after the addition of 1 mM EDT.

(∼530 nm) decreased when BArNile bound to GST-CaM-helixYFP. These fluorescence changes were reversed upon the addition of 1 mM EDT to dissociate BArNile from the 4Cys motif (Figure 7b). HEK293 cells transiently expressing CaM-helix-YFP were incubated with 1 µM BArNile-EDT2 for 3 h at 37 °C. The fluorescence images corresponding to the fluorescence of YFP and BArNile after incubation for 3 h are shown in Figure 8. Only the cells expressing YFP showed bright red fluorescence of BArNile. Upon addition of 1 mM EDT, the fluorescence of YFP increased and that of BArNile decreased (Figure 8d). On the other hand, there was no fluorescence change with addition of 1 mM EDT when the cells expressing YFP but without 4Cys motif were

used (Figure 9). The change in the FRET efficiency in the cells expressing CaM-helix-YFP was not due to mere changes in distance and orientation between two fluorophores; rather, it was caused by the dissociation of BArNile from the 4Cys motif, because the fluorescence of BArNile also decreased when images were obtained by direct excitation of BArNile (data not shown). From these results, it is concluded that BArNile specifically labels the 4Cys motif, even in live cells. Detection of Ca2+-Dependent Conformational Change of Recombinant Calmodulin Labeled with BArNile. Calmodulin (CaM) is a Ca2+-binding protein that controls various physiological responses by binding to its target proteins in response to extracellular stimuli.22 CaM changes its conformation, depending Analytical Chemistry, Vol. 73, No. 13, July 1, 2001

2925

Figure 9. Control experiment of Figure 8: HEK293 cells transiently expressing YFP were loaded with BArNile-EDT2 for 4.5 h, followed by treatment with 1 mM EDT. Fluorescence images of HEK293 cells expressing YFP corresponding to (a) YFP (530 to 590 nm) and (b) BArNile (> 635 nm) before (left) and after (right) the addition of 1 mM EDT to the cells (excitation, 514 nm); (c) phase-contrast image; scale bar, 50 µm; (d) fluorescence changes of YFP and BArNile of two selected cells shown in (c) before and after the addition of 1 mM EDT.

on the intracellular Ca2+ concentration,23-25 and exposes its hydrophobic domains for interacting with its target proteins.17,26 We utilized our fluorescent probe in order to detect the conformational change of CaM. At first, we tried to mutate an intrinsic R-helix in CaM to generate a 4Cys motif. However, on the basis of the threedimensional structure of CaM of Ca2+-bound23,27 and free form,28 there was no appropriate R-helix to be labeled with BArNile where the probe was expected to change its molecular environment through the conformational change of CaM. In fact, when one of the intrinsic R-helices of calmodulin was changed to the 4Cys motif and labeled with BArNile, the labeled protein did not show any fluorescence change through the conformational change of the calmodulin (not shown). Therefore, we extended the C-terminal of CaM and conjugated a 4Cys-helix and used this recombinant calmodulin (CaM-helix). BArNile attached to this 4Cys motif was expected to interact with hydrophobic domains of CaM (Figure 1b), because a CaM-binding peptide M13 conjugated to the C-terminal of CaM was known earlier to interact with hydrophobic domains of CaM.29,30 (22) Berridge, M. J.; Lipp, P.; Bootman, M. D. Nat. Rev. Mol. Cell. Biol. 2000, 1, 11-21. (23) Ikura, M.; Clore, G. M.; Gronenborn, A. M.; Zhu, G.; Klee, C. B.; Bax, A. Science 1992, 256, 632-638. (24) Crivici, A.; Ikura, M. Annu. Rev. Biophys. Biomol. Struct. 1995, 24, 85116. (25) Meador, W. E.; Means, A. R.; Quiocho, F. A. Science 1992, 257, 12511255. (26) O’Neil, K. T.; Henry R. Wolfe, J.; Erickson-Viitanen, S.; DeGrado, W. F. Science 1987, 236, 1454-1456. (27) Rao, S. T.; Wu, S.; Satyshur, K. A.; Ling, K. Y.; Kung, C.; Sundaralingam, M. Protein Sci. 1993, 2, 436-47. (28) Zhang, M.; Tanaka, T.; Ikura, M. Nat. Struct. Biol. 1995, 2, 758-767.

2926 Analytical Chemistry, Vol. 73, No. 13, July 1, 2001

Figure 10. Detection of Ca2+-dependent conformational change of CaM-helix using BArNile-EDT2: (a) Emission spectra of CaMhelix labeled with BArNile in the presence of 1 mM EGTA (- -) and after the addition of excess CaCl2 to the solution (s); (b) Ca2+ titrations of CaM-helix labeled with BArNile (emission, 604 nm, excited at 520 nm, n ) 3). The error bars represent a standard deviation of three independent measurements. The changes in fluorescence intensity were normalized to the effects of full Ca2+ saturation. The fitted curve was constructed from the Hill coefficient and Hill constant described in the text.

The fluorescence of BArNile labeled to CaM-helix increased with elevating Ca2+ concentrations (Figure 10a). The change in intensity was about 10% without change in emission wavelength. This fluorescence increase was reversed by an addition of excess EGTA for chelating Ca2+ (not shown). The observed 10% change in the fluorescence intensity is not a physical constant that is specific to the present probe molecule, but rather is dependent on the kind and site of target proteins where the probe is attached. The fluorescence spectrum of the fluorophore of BArNile was actually found to be strongly dependent on its solvent polarity, as shown in Figure 4, which suggests larger responses to the conformational changes of proteins by introducing 4Cys motifs to appropriate sites on the basis of the structural information from X-ray crystallography or NMR studies. However, in the case of calmodulin in the present study, there were no appropriate intrinsic R-helices to be labeled with BArNile for the detection of its conformational change (vide supra), and therefore, the response would be improved alternatively by inserting a flexible peptide linker between calmodulin and the 4Cys motif for facile interaction of attached BArNile and the hydrophobic domains of calmodulin. (29) Porumb, T.; Yau, P.; Harvey, T. S.; Ikura, M. Protein Eng. 1994, 7, 109115. (30) Miyawaki, A.; Llopis, J.; Heim, R.; Caffery, J. M. M.; Adams, J. A.; Ikura, M.; Tsien, R. Y. Nature 1997, 388, 882-887.

Figure 11. Conformational change of CaM-helix in live cells visualized with BArNile attached to its 4Cys motif: (a) Fluorescence images of CaM-helix labeled with BArNile-EDT2 in a HEK293 cell during the profile shown in b; (b) open squares, left-hand ordinate axis, time profile of the fluorescence of CaM-helix labeled with BArNile in a HEK293 cell; filled squares, right-hand ordinate axis, fluorescent time profile of a mocktransfected cell loaded with BArNile-EDT2 (excitation, 550 nm; emission, > 590 nm); (c) time profile of the fluorescence of a HEK293 cell loaded with Fura 2 (excitation, 340 nm; emission, > 500 nm).

To compare the Ca2+ dependence of the CaM-helix labeled with BArNile with native CaM, fluorescence intensities at 604 nm were replotted against the logarithm of Ca2+ concentration (Figure 10b). The data were fitted as a Hill plot, and the corresponding dissociation constant (Kd) and Hill coefficient (n) for the binding of Ca2+ were obtained (Kd ) 110 nM; n ) 1.7). The Ca2+ dissociation constant was lower than that of native Xenopus CaM determined by NMR (7.2 µM).29 The lower Ca2+ dissociation constant is probably due to the stabilization of the Ca2+-bound form of CaM by the interaction of hydrophobic domains of CaM with the helix or probe, as was the case in earlier reports.6,29 Imaging of Conformational Change of CaM in Live Cells. Using the developed probe molecule, the conformational change of CaM was imaged in live cells. There are endogenous P2y2 purinergic receptors in HEK293 cells.31 Stimulating HEK293 cells with ATP results in Ca2+ release from endoplasmic reticulum (ER) by the following signal transduction: activation of phospholipase C (PLC), production of inositol 1,4,5-trisphosphate (IP3), and binding of the IP3 to the IP3 receptor existing in the ER membranes. HEK293 cells transiently expressing CaM-helix were incubated with BArNile-EDT2 for 3 h, and after washing the cells, fluorescence images were obtained. Figure 11a shows the fluorescence images at different time points through a measurement. (31) Mundell, S. J.; Benovic, J. L. J. Biol. Chem. 2000, 275, 12900-12908.

CaM-helix distributed in cytosol and nuclei of HEK293 cells and both cytosolic and nuclear CaM-helix were labeled. Upon stimulation with 1 mM ATP, the fluorescence sharply increased and was followed by a gradual decrease (Figure 11b, open squares). When a Ca2+ ionophore, ionomycin, was added for increased permeability of Ca2+ through plasma membranes, the fluorescence increased further. This fluorescence increase was reversed upon chelating extracellular Ca2+ with EGTA. Similar fluorescence profiles as in Figure 11b were also observed in other cells at differing rates and extents of fluorescence changes; however, the change in fluorescence intensity induced by ionomycin was as much as 10% in every cell, which was the same extent as observed in vitro (vide supra). When mock-transfected HEK293 cells were loaded with BArNile-EDT2, the cells showed weak fluorescence as a result of the background staining, and the fluorescence intensity did not change when the intracellular Ca2+ concentration increased (Figure 11b, filled squares). This result confirmed that the observed fluorescence changes of the cells transfected with CaM-helix were not simple noises but, rather, reflected the changes in the environment of BArNile attached to the CaM-helix. When a cell was loaded with Fura 2-AM, the membranepermeable form of Ca2+ indicatior Fura 2, a fluorescence time profile (Figure 11c) that was nearly synchronized with that of the BArNile-loaded cell was obtained (Figure 11b, open squares). From this, we conclude that the fluorescence change obtained Analytical Chemistry, Vol. 73, No. 13, July 1, 2001

2927

with BArNile labeled to CaM-helix in fact represents the Ca2+induced conformational change of CaM. CONCLUSION In the present study, we designed and synthesized a new environment-sensitive fluorescent probe that is capable of being attached at its genetically engineered 4Cys motif of recombinant proteins in live cells. Using this fluorescent probe, BArNile-EDT2, the recombinant calmodulin containing the 4Cys motif was labeled inside live cells, and its conformational change upon Ca2+ increase was imaged in live cells. Although this probe molecule showed a moderate fluorescence change through the conformational change of calmodulin in in vitro and in vivo experiments in the present study, the fluorescence spectra of the fluorophore of BArNile in various solvents imply its high potentiality for detecting a change in its microenvironment. Larger responses to the conformational changes of

2928

Analytical Chemistry, Vol. 73, No. 13, July 1, 2001

other proteins are expected by rationally designing the locations of 4Cys motifs in the proteins on the basis of the structural information from X-ray crystallography and NMR. The present probe molecule will thus be accepted as a useful probe for studying signal transduction in live cells. ACKNOWLEDGMENT This work has been supported by CREST (Core Research for Evolutional Science and Technology) of JST (Japan Science and Technology) and grants to Y.U. from the Ministry of Education, Science, and Culture, Japan. J.N. expresses thanks for a fellowship from the Japan Society for the Promotion of Science (JSPS).

Received for review December 28, 2000. Accepted April 4, 2001. AC001528P