All chemicals were purchased from Sigma-Aldrich or Acros and used without further purification.¹H-NMR and ¹³C-NMR spectra were recorded on a Bruker AMX-400 NMR spectrometer, using TMS as an internal standard. ESI-HRMS (high resolution mass spectrometry) spectra were obtained on a Thermo Electron LTQ Orbitrap hybrid mass spectrometer. UV-visible spectra were collected on a Cary 50 UV-Vis spectrophotometer; Fluorescence spectra were collected on a Cary Eclipse (Varian, Inc.) fluorescence spectrophotometer with slit widths set at 5 nm for both excitation and emission, respectively. The high voltage of the fluorescence spectrophotometer was set at 500 V for the resorufin-based probe 4. The pH measurements were carried out with an Orion 410A pH meter.

Probe 4 was synthesized in a one-step reaction of resorufin with acryloyl chloride (Scheme 2). To a solution of resorufin (220 mg) and Et₃N (1.5 equiv) in 15 mL of anhydrous CH₂Cl_2, acryloyl chloride (2.0 equiv in 5 mL of CH₂Cl₂) is added dropwise at 0 °C. After stirring at this temperature for 60 min, the resulting mixture is allowed to cool to room temperature and stirred overnight. The mixture is diluted with CH₂Cl₂ (20 mL), washed with H₂O (10 mL × 3) and dried over anhydrous Na₂SO₄. The solvent is removed by evaporation to afford an orange solid (172 mg, 70% yield). ¹H NMR (CDCl₃, 400 MHz), δ (ppm): 7.92 (d, 1H, J = 8.7 Hz), 7.59 (d, 1H, J = 9.8 Hz), 7.49 (s, 1H), 7.33 (dd, 1H, J₁ = 2.4 Hz, J₂ = 2.5 Hz), 6.83 (dd, 1H, J₁ = 2.0 Hz, J₂ = 2.0 Hz), 6.63 (d, 1H, J = 17.2 Hz), 6.49 (m, 1H), 6.24 (s, 1H), 6.22 (d, 1H, J = 1.0 Hz). ¹³C NMR (CDCl₃, 100 MHz), δ 186.48, 163.84, 153.24, 148.87, 147.78, 143.55, 134.85, 134.26, 134.15, 131.22, 131.16, 127.18, 118.82, 109.94, 105.86. ESI-MS m/z = 268.0833 [M+H]⁺, calc. 268.0810 for C₁₅H₁₀NO₄.

In buffer without a surfactant, as expected, 4 shows excellent selectivity towards Cys due to the reported kinetically favored 7-membered ring formation. Upon mixing Cys with 4, the conjugate addition product 4a is formed, which undergoes a rapid cyclization reaction to produce 3a while releasing the free resorufin (Scheme 3). A significant color change and fluorescence enhancement can be observed in response to Cys (Figure 1). In the case of Hcy, because it has an additional methylene group in its side chain, a kinetically less favored 8-membered ring would form. As for GSH, 1,4-addition of thiols to the α,β-unsaturated carbonyl moieties of 4 can occur readily; however, the ensuing intramolecular cyclization similar to that of Cys cannot proceed without the presence of surfactant. No significant color formation or fluorescence response is seen in the case of Hcy or GSH under these conditions.

The time course of the fluorescence assay is shown in Figure 1(b). It can be seen that the fluorescence upon reaction with Cys increases with time and reaches a plateau after about 90 min, whereas for Hcy and GSH the reactions are significantly slower. The sensitivity and linearity of the response of 4 towards Cys was also investigated. Linear response with submicromolar sensitivity was observed (See Figure S5 in the Electronic Supplementary Information). The reaction products between 4 and thiols were further studied using HRMS (See Figures S6 and S7 in the Electronic Supplementary Information).

Inclusion of surfactants (see Figure 2) dramatically alters the intrinsic selectivity of 4. For example, solutions of GSH, Cys and Hcy (2 equiv) with the resorufin-acrylate probe 4 in pH 7.4 phosphate buffer generate the characteristic strong pink color immediately upon addition of CTAB only in the presence of GSH (Figure 3(a)). Other amino acids, such as Hcy and Cys, did not exhibit significant changes under the same conditions. This interesting feature indicates that 4-CTAB system can serve as a selective visual inspection dosimeter for GSH. Corresponding fluorescence increases were also observed (Figure 3(b)). Fluorescence upon reaction with GSH reaches a plateau in less than 2 min.

To study the effects of different surfactants, SDS, BC and Triton X-100 were evaluated. It was found that 4 serves as an outstanding indicator for GSH only in the presence of cationic surfactants such as BC and CTAB (Figure 4). Negatively charged SDS suppressed the response of 4 towards all three thiols; while the non-ionic surfactant Triton X-100 enhanced the selectivity towards Cys resulting in zero response for Hcy or GSH (See Figure S8 in the Electronic Supplementary Information). The result indicates that selectivity of 4 is tunable by simply changing the reaction media rather than the chromogen.

To further evaluate the selectivity of the 4-CTAB system towards GSH, control experiments using a series of other amino acids were performed measuring the fluorescence responses. Other amino acids produced no significant fluorescence enhancement as compared to GSH, further demonstrating the excellent selectivity for GSH (Figure 5(a)). Moreover, significant fluorescence enhancement can still be observed for GSH even in the presence of excess Cys (see Figure S9 in the Electronic Supplementary Information). Additionally, the fluorescence enhancement of the 4-CTAB system displayed submicromolar sensitivity and a linear relationship with the concentration of GSH over a wide range (Figure 5(b)). The system functions well within the range of typical GSH concentrations in human plasma [21].

Based on the above results, we propose a plausible mechanism for the GSH selectivity of the 4-CTAB system that involves the conjugate addition of GSH to 4 to generate 6, which in turn undergoes an intramolecular cyclization/elimination reaction sequence catalyzed by the CTAB micelle releasing the free resorufin dye (Scheme 4), resulting in the recovery of the fluorescence. Formation of the 12-membered ring (7) and the free resorufin dye were further investigated via HRMS (See Figures S10, S11 in the Electronic Supplementary Information).

While the 4-CTAB-GSH signal is high at pH 7.4 (see above), it is known that acrylate esters are susceptible to hydrolysis at high pH and that GSH is also relatively unstable under the same conditions (half-life of 9 hours at pH 7.4 compared to 16 hours at pH 6.5) [22]. In this regard, spectral behavior of the 4-CTAB system was investigated over a wide range of pH values (pH 5.5 to pH 8) and its response to the various thiols was studied at lower pH values. As expected, the acrylate-based probe 4 showed the greatest background signal at the higher pH values investigated. The 4-CTAB system was most stable at lower pH values (See Figure S12 in the Electronic Supplementary Information). The reaction with thiols at pH 6 is slightly less rapid as compared to pH 7.4; however, the selectivity towards GSH was further enhanced through the complete elimination of any residual Cys and Hcy response (Figure 6). Thus pH 6 was chosen for further application studies in plasma.

The detection of GSH was performed in 10% deproteinized human plasma. Plasma proteins were precipitated using MeCN (two thirds of the reconstitution volume) and removed by centrifugation at 4,000 rpm for 30 min. The supernatant liquid was diluted [23] and added to a solution of 4 (1 μM) in phosphate buffer (pH 6, 50 mM) in the presence of 2.0 mM CTAB. The fluorescence response of replicate (n = 3) samples and GSH spiked samples were monitored as above and the GSH content in the plasma sample was determined from the regression equation of a standard calibration curve (Figure 7). The GSH content of the plasma sample was determined to be 3.24 ± 0.14 μM, which is well within the reported GSH concentration range for human plasma samples from healthy individuals [21]. Recoveries of the known spiked amounts of GSH were between 99.2% and 102.3% with a generally satisfactory precision (Table 1). These results revealed the potential applicability and reliability of using 4 in quantitative detection of GSH in human plasma. In addition, Cys exhibited significant interference with the GSH determination (See Figure S13 in the Electronic Supplementary Information).