A near-infrared fluorescent probe based on a hemicyanine dye with an oxazolidine switch for mitochondrial pH detection

Abstract

A near-infrared fluorescent probe (AH+) has been prepared by incorporating an oxazolidine switch into a near-infrared hemicyanine dye. The probe shows fast and sensitive responses to pH from an oxazolidine switch to the hemicyanine dye upon pH decreases from 10.0 to 5.0. The probe shows good photostability, low cytotoxicity, and reversible fluorescence responses to pH changes with a pK_a value of 7.6. It has been successfully used to determine pH changes in mitochondria.

Introduction

Mitochondria possess double-membranes composed of phospholipid bilayers and proteins and act as the energy-supplying organelles in almost all eukaryotic cells. Mitochondria have vital roles in cell metabolism such as regulation of the cellular redox state, production of reactive oxygen species (ROS), regulation of Ca²⁺ homeostasis, cell cycle and growth, cell death, and apoptosis.^1–5 Under ideal physiological conditions, mitochondria function in a slightly alkaline microenvironment at a pH value of ~8.0 compared to the other subcellular compartments of the cell.^6–8 These functions of the mitochondria depend on the pH level within the mitochondria. Deviations of the normal mitochondrial pH are associated with mitochondrial dysfunction that is present in many human disorders such as neurodegenerative and neuromuscular diseases, obesity and diabetes, cancer, and inherited mitochondrial diseases.^4,9,10 Therefore, accurate and sensitive detection of mitochondrial pH will provide a better understanding of mitochondrial biology.^11,12

Fluorescent probes based on small organic molecules are the essential tools for bioanalysis and real-time bioimaging technologies because of their superior features of excellent sensitivity and high spatial resolution.^3,11–16 Fluorescent probes with near-infrared absorption and emission wavelengths have the advantages of deep tissue penetration, low biological fluorescence background, and the least photodamage impact on live cells and tissues.^11,12,17 Near-infrared hemicyanine and rhodamine dyes bearing spirolactam switches have been developed to monitor pH changes in live cells based on open and close spirolactam ring configurations upon pH changes.^18–22 These neutral fluorescent probes with closed spirolactam ring structures often function as weak bases with low pK_a values related to spirolactam ring-opening and specifically target lysosomes instead of mitochondria.^18–22 Although the construction of fluorescent probes possessing higher pK_a values has been reported by introducing bulky residues into spirolactam switches, they are still used to visualize pH changes in lysosomes.²³ Ideal fluorescent probes for monitoring pH in mitochondria should possess the mitochondria-targeting capability and high pK_a values, which are difficult to achieve by using spirolactam switches in hemicyanine and rhodamine dyes. With this in mind, we developed a near-infrared fluorescent probe (AH+) for the detection of pH changes in the mitochondria by introducing an oxazolidine switch into a hemicyanine dye to overcome the pH insensitivity of the near-infrared hemicyanine dyes without spirolactam switches.²⁴ Under the basic conditions, the hemicyanine dye undergoes a ring-closing reaction to form an oxazolidine switch, resulting in fluorescence quenching as the free hydroxyl group engages in a nucleophilic attack on the indolenium moiety. Acidic pH effectively converts the oxazolidine switch into a hemicyanine moiety with significantly extended π-conjugation through proton-activated oxazolidine ring-opening and leads to a gradual fluorescence enhancement at 725 nm and absorbance increases at 713 nm upon pH decreases from 10.0 to 5.0. Probe AH+ shows good photostability, decent selectivity to pH, low cytotoxicity, and reversible fluorescence to pH changes with a high pK_a value of 7.6. The probe has been successfully used to determine pH changes in mitochondria.

Experimental

1. Materials

Unless specifically indicated, all reagents and solvents were obtained from commercial suppliers and used without further purification. 6-(Diethylamino)-2,3-dihydro-1H-xanthene-4-carbaldehyde (3) and 1-(2-hydroxyethyl)-2,3,3-trimethyl-3H-indol-1-ium salt (4) were prepared according to the literature.^25–27

Synthesis of fluorescent probe AH+.

After 6-(diethylamino)-2,3-dihydro-1H-xanthene-4-carbaldehyde (3) (283 mg, 1 mmol) and 1-(2-hydroxyethyl)-2,3,3-trimethyl-3H-indol-1-ium (4) (204 mg, 1 mmol) were added to 10 mL of acetic anhydride, the mixture was stirred for 5 hours at 80 °C. The reaction mixture was concentrated in vacuo and diluted with dichloromethane, washed by water and brine, and dried over anhydrous Na₂SO₄. The solution was filtered and the filtrate was concentrated. The residue was purified by using flash column chromatography through gradient elution with methanol to dichloromethane ratio ranging from 5% to 10%, affording the product as a green solid. ¹HNMR (400 MHz, chloroform-d) δ 8.48 (d, J = 14.1 Hz, 1H), 7.49 (s, 1H), 7.43–7.34 (m, 3H), 7.23 (d, J = 1.2 Hz, 1H), 6.79 (dd, J = 9.0, 2.5 Hz, 1H), 6.49 (d, J = 2.4 Hz, 1H), 6.23 (d, J = 14.2 Hz, 1H), 4.55 (d, J = 5.2 Hz, 2H), 4.50 (d, J = 5.3 Hz, 2H), 3.53 (q, J = 7.1 Hz, 4H), 2.75 (t, J = 6.1 Hz, 2H), 2.70 (t, J = 6.0 Hz, 2H), 2.03 (s, 2H), 1.87 (d, J = 0.7 Hz, 2H), 1.74 (s, 2H), 1.32–1.23 (m, 6H). ¹³CNMR (101 MHz, chloroform-d) δ 170.68, 170.59, 163.54, 156.42, 142.33, 140.72, 130.01, 128.88, 125.59, 122.43, 117.73, 116.01, 111.21, 77.50, 60.82, 60.62, 49.74, 45.47, 44.25, 29.98, 29.10, 28.84, 24.93, 20.99, 20.92, 16.72, 12.72. MS/Z = 469.4.

2. Optical measurements

A citrate–phosphate buffer (0.1 M) for the acidic pH from 5.0 to 7.0 and a carbonate–bicarbonate buffer (0.2 M) for the basic pH from 7.0 to 10.0 containing 5% ethanol were used to investigate the effect of pH on the absorption and fluorescence spectra of the fluorescent probe. The photostability and selectivity measurements of the fluorescent probes were conducted under similar conditions to those employed for the investigation of the pH dependency.

3. Live cell imaging

HeLa cells were seeded in 35 mm confocal glass-bottom dishes (MatTek) with 1 × 10⁵ cells per dish and cultured for 24 h before cellular imaging was conducted. For the co-localization experiment, HeLa cells were incubated with the 5 μM probe AH+, 1 μM Hoechst and either 1 μM Lysotracker Red or 1 μM Rhodamine 123 for 30 min, followed by washing the cells twice with PBS buffer before cellular imaging was performed. For visualization of intracellular pH changes, HeLa cells were rinsed with PBS buffer twice before they were incubated with nigericin (5 μg mL⁻¹) in different pH buffers 5.0, 6.0, 7.0, 8.0, 9.0, and 10.0 for 30 min to equilibrate the intracellular and extracellular pH.^20,28–32 The cells were further incubated with 5 μM probe AH+ for 30 min and followed by rinsing the cells with FBS buffer twice before fluorescence imaging was carried out. The cells were imaged using a 200X objective lens for the colocalization study and a 60X objective lens for the other imaging experiments. The fluorescence emission of the nucleus dye Hoechst (blue channel) under 405 nm excitation was collected from 425 nm to 475 nm; the fluorescence emission (red channel) of the probe under 635 nm excitation was collected from 725 nm to 775 nm. The fluorescence emissions of the commercial dyes Lysotracker Red and Rhodamine 123 (green channel) under the excitation of 559 nm and 488 nm were collected from 600 nm to 650 nm, and from 500 nm to 550 nm, respectively. The images were further processed using an Olympus FV10-ASW 3.1 viewer and Image Pro6.

4. Theoretical calculations

Computer modeling of probes A and AH⁺ was accomplished using procedures published previously to establish base geometries.³³ The molecular data were initially refined using density functional theory (DFT) employed with the B3LYP functional³⁴ and electron basis sets initially at the 6–31*g(d) level to convergence in Gaussian 16³⁵ in a polarizable continuum model (PCM) of water.³⁶ The final models were calculated with the APFD functional³⁷ and 6–311+g(2d,p)^38,39 basis sets. Imaginary frequencies were not obtained in any frequency calculations. The excited states were assessed based on TD-DFT optimization⁴⁰ also in a PCM of water. The results were interpreted using GaussView 6⁴¹ for all data and figures. The results of the calculations (including drawings of LCAOs for all molecular orbitals discussed) are given in detail in the ESI.^†

Results and discussion

Probe design and synthesis

Near-infrared hemicyanine molecules possess unique advantageous photophysical properties such as high fluorescence quantum yield, large molar absorptivity, excellent chemical stability, and photostability with near-infrared emission at 725 nm.^24,25,27 However, hemicyanine dyes without spirolactam switches are insensitive to pH. In order to develop a near-infrared fluorescent probe for the sensitive detection of mitochondrial pH, we introduced an oxazolidine switch into a near-infrared hemicyanine dye through the Knoevenagel condensation of 6-(diethylamino)-2,3-dihydro-1H-xanthene-4-carbaldehyde (3) with 1-(2-hydroxyethyl)-2,3,3-trimethyl-3H-indol-1-ium salt (4) in acetic anhydride at 80 °C (Scheme 2). 6-(Diethylamino)-2,3-dihydro-1H-xanthene-4-carbaldehyde (3) was prepared by reacting β-bromoenal (2) with 4-(diethylamino)-2-hydroxybenzaldehyde (1) in the presence of Cs₂CO₃ in DMF solution at room temperature (Scheme 2).^25–27 The probe has been characterized via NMR and mass spectrometry.

Scheme 2.

The synthesis of fluorescent probe AH⁺.

Optical responses of the probe to pH changes

We investigated whether the probe can respond to pH changes by collecting absorption and fluorescence spectra in two different buffer solutions, 0.1 M citrate–phosphate (pH from 5.0 to 7.0) and phosphate–phosphate buffers (pH from 7.0 to 10.0) containing 5% ethanol (Fig. 1). Under basic pH 10.0 conditions, the dangling hydroxyl group reacts as a nucleophile with the indolenium moiety forming an oxazolidine switch through a ring-closing reaction (Scheme 1). As a result, probe AH+ shows low absorption at 713 nm, and a weak fluorescence peak at 727 nm (Fig. 1 and 2) at the basic pH 10.0. Gradual pH decreases from 10.0 to 5.0 result in significantly gradual absorbance increases at 713 nm and gradual fluorescence enhancement at 727 nm with a fluorescence quantum yield of 4.63% at pH 5.0 because acidic pH effectively converts the probe oxazolidine switch to a hemicyanine structure with significantly extended π-conjugation (Fig. 1 and 2). Probe AH+ possesses a pK_a value of 7.6 related to the opening of the oxazolidine switch to the hemicyanine structure (Fig. S4, ESI^†).

Fig. 1.

Absorption spectra of probe AH+ in different pH buffers containing 5% ethanol. Citrate–phosphate buffers were used for pH ranging from 5.0 to 7.0 while phosphate buffers were employed for pH ranging from 7.0 to 10.0.

Scheme 1.

Chemical structures of a fluorescent probe with an oxazolidine switch in response to pH changes.

Fig. 2.

Fluorescence spectra of probe AH+ in different pH buffers containing 5% ethanol under excitation at 670 nm. Citrate–phosphate buffers were used for pH ranging from 5.0 to 7.0 while phosphate buffers were employed for pH ranging from 7.0 to 10.0.

Theoretical modeling

The structures of probes A and AH⁺ were assessed to understand the nature of the p-delocalization that occurs in basic media and to speculate as to any other conformational changes that may be occurring. We find that for probe A as illustrated in Fig. 3 and Fig. S11 (ESI^†), the rhodamine plane and that comprising the oxazolidine switch are at an acute angle of approximately 72°. Excited-state calculations employing the TD-DFT calculation for six excited states reveal two possible transitions at 290.68 nm (i.e., excited state 6 in Table S2, ESI^†) and 417 nm (i.e., excited state 1 in Table S2, ESI^†). The nature of the lower energy transition was depicted in the form of a current density plot, Fig. 3 (left) which shows that the plane encompassing the oxazolidine ring is not involved in this transition. The individual LCAOs from which the image for probe A in Fig. 3 was derived are presented as Fig. S14 and S15 (ESI^†).

Fig. 3.

Current density illustrations as iso-surfaces of probes A (left) and AH⁺ (right). Red to blue areas indicate −ve to +ve values for the different densities of 1.428 × 10⁻⁴ for A and 1.2023 × 0⁻² for AH⁺.

For probe AH⁺ illustrated in Fig. 3 and Fig. S16 (ESI^†), there is a major conformation change when the oxazolidine ring switch is activated. Specifically, the planes formed by the rhodamine moiety and that for the hemicyanine are roughly coplanar (angle of 1.72°) as is evident in the overall delocalization in the current density image for probe AH⁺depicted in Fig. 3. The LCAOs from which this image was obtained are given as Fig. S19 and S20 (ESI^†). The excited-state calculations for six excited states revealed two transitions as evident in Fig. S18 (ESI^†) at 333.75 nm for excited state 4 and 593.26 nm for excited state 1 as listed in Table S4 (ESI^†). The transition at 593.26 nm (expt. 725 nm) results from the extended p-delocalization as a consequence of the activation of the oxazolidine switch.

Probe selectivity, photostability, and pH response reversibility

The possible interference of other analytes was tested by recording the fluorescence spectrum of the 5 μM probe AH+ in the absence and presence of a high concentration (200 μM) of various essential metal ions such as Al³⁺, Fe³⁺, Fe²⁺, Cr³⁺, Ca²⁺, Co²⁺, Hg²⁺, Mg²⁺, Mn²⁺, Ni²⁺, Zn²⁺ K⁺, and Na⁺ ions, or I⁻, Br⁻, Cl⁻, ${\text{SO}}_3^{2-}$, NO₂⁻, NO₃⁻, S²⁻, ${\text{CO}}_3^{2-}$ and HCO₃⁻. The results indicate that the probe maintains high selective fluorescence responses to pH without interference from these cations and/or anions (Fig. 4).

Fig. 4.

Fluorescence responses of the 5 μM fluorescent probe AH+ to pH at 5.0 and 10.5 in the absence and presence of different metal ions (200 μM), respectively.

The fluorescent probe was excited continuously (670 nm) for 5 min intervals and the fluorescence intensity was measured every 5 min at pH 5.0 and pH 7.4, respectively. The result indicates that probe A displayed a moderate photostability with 6.9% and 4.7% decrease only in fluorescence intensity under two-hour excitation at pH 5.0 and pH 7.4 (Fig. 5A). Probe A exhibits a reversible response to pH changes between 5.0 and 9.2 (Fig. 5B).

Fig. 5.

Photostability of 5 μM fluorescent probe AH+ at pH 5.0 and pH 7.4 in 10% ethanol solution (Fig. 5A). The sample was exposed under the respective optimal excitation wavelength (670 nm) and fluorescence intensities were measured at 5 min intervals (Fig. 5B). The reversible response of the probe to pH changes between pH 5.0 and 9.2 under excitation at 670 nm with three repeated experiments.

Probe cytotoxicity

We evaluated the cell cytotoxicity of the probe for its biocompatibility via MTT assay. The cytotoxicity of the probe increases slightly with the probe concentration with lower cell viability. A high concentration (50 μM) of the probe did not cause any considerable cytotoxicity because the cell viability was still higher than 86.5%, indicating that the probe showed excellent biocompatibility and low toxicity (Fig. 6).

Fig. 6.

Cytotoxicity and cell proliferation effects of the probe were tested via MTT assay. The HeLa cells were incubated with different concentrations of the probe for 24 hours for mitochondria staining.

Selective staining of mitochondria

We elevated the probe performance by testing its cell permeability. We found out that the probe displayed excellent cell permeability. Probe AH+ carrying a positive charge could be used to specifically target mitochondria with a negatively charged matrix. We carried out colocalization experiments by using Lysotracker Red and rhodamine 123 to confirm our hypothesis that the probe possessed mitochondrial specificity.⁴² The probe fluorescence is impeccably overlapped with fluorescence mitochondria-targeting rhodamine 123 with a Pearson correlation coefficient of 0.923 (Fig. 7) while the probe fluorescence shows a poor Pearson correlation coefficient of 0.631 with lysosome-targeting Lysotracker Red in HeLa cells (Fig. 8). The probe also displays lower Pearson correlation coefficients of 0.194 (Fig. S6, ESI^†) and 0.138 with Golgi-specific Golgi-GFP (Fig. S9, ESI^†), and ER-specific ER-Tracker^™ Green (Fig. S10, ESI^†). These results confirm that the probe exhibits excellent mitochondria selectivity.⁴²

Fig. 7.

Confocal microscopic cellular images and merged images of the probe colocalized with Hoechst, and Lysotracker Red DND-99 in HeLa cells. Colocalization scatterplot of the probe with Lysotracker Red. Scale bar: 20 μm. Pearson correlation coefficient: 0.631.

Fig. 8.

Confocal microscopic cellular images and merged images of the probe colocalized with Hoechst, and rhodamine 123 in HeLa cells. Colocalization scatterplot of the probe with rhodamine 123.⁴² Scale bar: 20 μm. Pearson correlation coefficient: 0.923. Scale bar: 20 μm.

In order to further demonstrate that the probe specifically stains mitochondria, we employed carbonyl cyanide 4-((trifluoromethoxy)phenylhydrazone) (FCCP), an uncoupler of oxidative phosphorylation in mitochondria, to further treat HeLa cells after the cells were incubated with the probe (Fig. 9). Our results show that FCCP treatment increases the fluorescence intensity of the probe because FCCP can disrupt the mitochondrial H+ gradient and lead to acidification of mitochondria (Fig. 9).^17,43,44

Fig. 9.

Confocal microscopic cellular images of HeLa cells with a 10 μM probe AH+ with and without 200 nM carbonyl cyanide 4-((trifluoromethoxy)phenylhydrazone) (FCCP) treatment.

Visualization of mitochondrial pH changes in live cells

Since we demonstrated that the probe can selectively accumulate in mitochondria, we further studied whether the probe could be used to determine mitochondrial pH changes in live cells. We incubated HeLa cells with the 10 μM probe AH+ in different pH buffers containing 5 μM nigericin, which was used to adjust the intracellular pH to the external buffer pH. Our results show that gradual decreases of intracellular pH values from pH 10.0 to 5.0 significantly enhance the cellular fluorescence of the probe (Fig. 10), which is consistent with the fluorescence responses of the probe to buffer pH changes (Fig. 2), as acidic pH effectively converts the oxazolidine switch into hemicyanine with significant π-conjugation through the proton-activated oxazolidine ring-opening. The cells may have a short lifetime under pH 10.0 conditions.

Fig. 10.

Confocal microscopic cellular images of HeLa cells with 10 μM probe AH+ incubated in different pH buffers containing 5 μM nigericin. Scale bar: 50 μm.

Conclusion

We have successfully developed a near-infrared fluorescent probe by incorporating the oxazolidine switch into hemicyanine for specific targeting of mitochondria. Effective monitoring of mitochondrial pH changes in live cells is achieved through π-conjugation modulation between oxazolidine switch and hemicyanine structure upon pH changes.