Highly selective fluorescent probe with an ideal pH profile for the rapid and unambiguous determination of subcellular labile iron (III) pools in human cells

Abstract

A convenient tool for detecting iron ions in subcellular structures is desired for better understanding its roles in biological systems. In this work, a new Fe³⁺ sensor, 2-(2-((1-(6-acetylpyridin-2-yl)ethylidene)amino)ethyl)-3’,6’-bis(diethylamino)spiro[isoindoline-1,9’-xanthen]-3-one (RhPK), which operates across the entire cellular pH range and is capable of unambiguously detecting Fe³⁺ ion in live human cells at subcellular resolution, is reported. The sensor exhibits high selectivity and sensitivity toward Fe³⁺ with a rapid fluorescence response and a 12-fold increase in intensity upon the addition of 1 equivalent Fe³⁺ at pH 7.3. RhPK forms a 1:1 complex with Fe³⁺ with an apparent binding constant 1.54 × 10⁷ M⁻¹ and a detection limit of 50 nM. The sensor is stable between pH 4.2 and 9.0 and operates across the whole cellular pH range. Cell imaging demonstrates the ability of the sensor to unambiguously detect basal level Fe³⁺ as well as its dynamic changes in real-time in live cells at subcellular resolution, with one labile Fe³⁺ pool identified in mitochondria in human primary fibroblast (ws1) cells for the first time and two Fe³⁺ pools confirmed in mitochondria and endo/lysosomes in human SH-SY5Y neuroblastoma cells, suggesting different cell types have distinctive Fe³⁺ storage in subcellular compartments. The RhPK probe is powerful for rapid and sensitive bioimaging of Fe³⁺ at subcellular level, enabling the unambiguous detection of labile Fe³⁺ pools at the entire cellular pH range, which is of great significance to understand the biological chemistry of Fe³⁺ and its roles in physiological processes and diseases.

Introduction

Iron, as the most abundant transition metal in human body and being indispensable for almost all living systems, plays critical roles for a broad range of biological processes, such as oxygen and electron transport and cellular respiration, synthesis of DNA, RNA and enzymes, and redox regulation (Crichton 2001; Rouault 2006; Kell 2009). It is believed that a small fraction of intracellular labile iron, containing both ferric (Fe³⁺) and ferrous (Fe²⁺) ions, is redox-reactive and in exchange with multiple small ligands and intracellular enzymes (Hider 2013, Cabantchi 2014). Due to the reducing environments in the cytosol, ferrous ion is believed to be the major iron species (> 80%) in the labile iron pool (LIP) (Hider 2013, Cabantchi 2014). It has been estimated that low micromolar levels of Fe²⁺ are present in the cytosolic LIP and at slightly higher concentrations in certain organelles such as mitochondria (Hider 2013, Cabantchi 2014. However, precise quantification of iron levels in LIPs is quite challenging (Hider 2013, Cabantchi 2014, Camarena, 2021).

However, excess iron in living systems promotes reactive oxygen species (ROS) generation via Fenton reaction, leading to oxidative stress which damages nucleic acids, proteins, and lipids (Burdo and Connor 2003; Lv and Shang 2018). Moreover, iron ions have recently been discovered to trigger ferroptosis, a new type of cell death marked by iron-dependent phospholipid peroxidation, which has been implicated to the pathogenesis of numerous diseases including cancer, stroke and neurological disorders (Angeli et al. 2019; Jiang et al. 2021). Therefore, several analytical techniques have been developed for the detection of iron in living systems. Among the various techniques used for its detection, the method based on fluorescent probes has received considerable interest in recent years because of the ability to provide online monitoring of low concentrations with the advantages of spatial and temporal resolution (Andersen 2005; Domaille et al. 2008; Zhang et al. 2011; Pithadia and Lee 2012; Hirayama et al. 2013; Carter et al. 2014; Aron et al. 2016). As binding of Fe³⁺ with chemosensors induces intrinsic fluorescence quenching due to the paramagnetic nature of Fe³⁺, it is unsurprising that many early probes for Fe³⁺ detection in cells are “turn-off” sensors (Cabantchik 2014; Yang et al. 2012), with drawbacks such as poor spatial resolution, low sensitivity, or interferences from other metal ions (Espósito et al. 2002; Fakih et al. 2009).

Interestingly, a number of “turn-on” or ratio metric probes for Fe³⁺ have recently been reported (Aydin et al. 2012; Wei et al. 2012; Zhu et al. 2016; Zhou et al. 2017; Qiao et al. 2018; Wu et al. 2019; Shellaiah et al. 2020; Ghosh et al. 2020; Li et al. 2016; Liu et al. 2020; Chakraborty et al. 2020; Yang et al. 2020; Kaur et al. 2018; Zhang et al. 2014; Lohani et al. 2010; Heng et al. 2008; Ackerman et al. 2017; Lee et al. 2016; Wang et al. 2010; Wang et al. 2016; Mir et al. 2019; Qiu et al. 2014; Vishaka et al. 2019; Kim et al. 2019; Jothi et al. 2021; Gao et al. 2020; Zhang et al. 2021; Wei et al. 2010). Many iron probes, especially turn-on and ratiometric Fe³⁺ probes, have been discussed in detail in recent review papers (Sahoo and Crisponi 2019, Sun et al., 2021). These probes displayed varied selectivity and affinity to Fe³⁺ with detection limits from 3.18 nM to 10 μM and fluorescence enhancements from 8 to 800 fold. Many have been applied to bioimaging. Table S1 summarizes the basic properties and performance parameters of the recent Fe³⁺ sensors. However, their capabilities in unambiguously detecting native cellular Fe³⁺ pools across the entire biological pH range have yet to be established due to their various intrinsic drawbacks (Domaille et al. 2008; Zhu et al. 2016; Zhou et al. 2017; Qiao et al. 2018: Wu et al. 2019; Shellaiah et al. 2020; Ghosh et al. 2020; Liu et al. 2020; Chakraborty et al. 2020; Yang et al. 2020). Reaction-based Fe³⁺-sensors displayed nice turn-on fluorescent responses with cellular imaging capability but are incapable of monitoring the dynamic changes in Fe³⁺ due to the irreversible nature of the catalytic hydrolysis reactions (Kaur et al. 2018; Zhang et al. 2014; Lohani et al. 2010; Heng et al. 2008; Ackerman et al. 2017; Lee et al. 2010; Lee et al. 2016). Binding-based sensors have the advantages of reversible sensing and rapid on-off response. However, their fluorescent responses may also be triggered by acidic pH environment, limiting their application in Fe³⁺ detection in certain acidic cellular compartments such as endosomes and lysosomes (pH ~ 4.5 in active lysosomes (Mellman et al. 1986; Trombetta et al. 2003; Johnson et al. 2016). For example, a bisdiene macrocycle-based Fe³⁺ probe displayed “off-on” response to Fe³⁺ with intracellular imaging capability (Qiu et al. 2014) but does not operate at pH < 6.5 and it needs cell-damaging ultraviolet light for excitation. A recently reported rhodamine 6G-based Fe³⁺-selective sensor RG5NC works at pH > 6 (Vishaka et al. 2019). A Schiff-base Fe³⁺sensor FeP-1 (Kim et al. 2019) and a naphthalimide based Schiff-base Fe³⁺-selective probe NDSM (Jothi et al. 2021) both operate at pH > 5. A recently developed rhodamine linked dansylamide Fe³⁺-selective ratiometric probe DRhFe also operates at pH > 5 (Gao et al. 2020). A more recent rhodamine 6G-based Fe³⁺-selective sensor LXY works at pH between 6.8–8.0 (Zhang et al. 2021). All of these sensors are affected by the acidic pH in endosomes or lysosomes and thus do not have the capability to detect Fe³⁺ in these cellular compartments.

In our efforts to develop fluorescent sensors for detecting metal ions and other species in living systems (Wei et al. 2010; Aydin et al. 2012; Wei et al. 2012; Aydin et al. 2014; Maiti et al. 2015; Ozdemir et al. 2018; Aydin et al. 2020; Gupta et al. 2020), a highly sensitive “off-on” Fe³⁺ sensor RPE allowed the imagining of labile Fe³⁺ pools at subcellular resolution in living cells for the first time (Wei et al. 2012). However, a major limitation of this type of rhodamine-based sensors is the pH-dependence of the spiroclic lactone of rhodamine moiety. Their fluorescence can be turned-on at slightly acidic pH values in the absence of metal ions, limiting their applications in biological systems as certain common cellular organelles such as endosomes and lysosomes which have acidic environments (pH from 4.5 to 6.0) in their lumen (Mellman et al. 1986; Trombetta et al. 2003; Johnson et al. 2016). Our previously reported Fe³⁺ sensor RPE operates in the pH range from 5.5 to 9.0 (Wei et al. 2012) and an improved sensor RhHPA between 5.0 and 9.0 (Ozdemir et al. 2018). It is thus still necessary to develop Fe³⁺ sensors that work at more acidic pH, ideally, covering the whole cellular pH range (pH from 4.5 to 8.0) for detecting all Fe³⁺ pools in cells without ambiguity.

Inspired by the tactics in developing rhodamine-based pH sensors (Yuan et al. 2011; Lv et al. 2013), we have redesigned the Fe³⁺-binding moiety of the sensor to make it more acidic tolerant, and successfully overcome the pH issue with the creating of a new rhodamine-based Fe³⁺ sensor, RhPK, which operates from pH 4.2 to 9.0 that covers the whole cellular range. This highly selective sensor is capable of rapid detecting basal level Fe³⁺ in live-cells at subcellular resolution as well as the dynamic changes in cellular Fe³⁺ levels, with one labile Fe³⁺ pool identified in mitochondria in human primary fibroblast (ws1) cells for the first time and two Fe³⁺ pools confirmed in mitochondria and endo/lysosomes in human SH-SY5Y neuroblastoma cells.

Experimental

Materials and Methods

Rhodamine B base (Sigma-Aldrich, 95%) and 2, 6-diacetyl pyridine (Sigma-Aldrich, 99%) were purchased from Sigma–Aldrich. The other chemicals and solvents were purchased commercially as ACS regent grade or better. The solutions of metals ions were prepared from chloride salts of Fe²⁺, Fe³⁺ Cr³⁺, Hg²⁺, Mn²⁺, Ni²⁺, Cu²⁺, Zn²⁺, Ag⁺, Na⁺, Ca²⁺, and nitrate salts of Co²⁺, K⁺, Pb²⁺ and Mg²⁺. Stock solution of metals ions (10 mM) were prepared in deionized water while those for Fe²⁺, Fe³⁺ were prepared freshly in 0.01 M HCl solution. Tetrakis(acetonitrile)copper(I) (Sigma-Aldrich, 97%) was used to prepare for Cu⁺ solution by freshly dissolving it into deionized water.

A Bruker DRX-300 spectrometer or a Bruker AVANCE III HD 400 MHz digital spectrometer were used to acquire ¹H and ¹³C NMR spectra at ambient probe temperature, 298 K. Chemical shifts are reported in delta (δ) units per million (ppm) downfield from tetramethylsilane. Splitting patterns are abbreviated as follows: s, singlet; d, doublet; t, triplet; q, quartet; m, multiplet; br, broad. ESI-MS spectra were acquired on a PerkinElmer API 150EX mass spectrometer and data analyses was performed using the associated software. A Perkin-Elmer LS55 luminescence spectrometer was used to acquire fluorescence spectra at 298 K. Excitation and emission slits were 5.0 nm and emission spectra were collected from 530 to 700 nm with excitation at 510 nm. A Perkin-Elmer Lambda 25 spectrometer was used to record the absorption spectra at 298 K. A Corning pH meter equipped with a Sigma-Aldrich micro combination electrode was used for pH measurements. A Zeiss LSM 710 laser scanning confocal microscope was used to investigate the fluorescence responses of the probe in living cells.

Synthesis and characterization

Synthesis of RhPK.

Rhodamine B ethylenediamine was synthesized according to a published procedure (Ma et al. 2013). Briefly, 2, 6-diacetyl pyridine (200 mg, 1.23 mmol) was dissolved in ethanol (20 ml) and rhodamine B ethylenediamine (200 mg, 0, 48 mmol) dissolved in ethanol (40 ml) was added dropwise for 2 h. The mixture was refluxed and stirred for overnight. After the mixture was cooled to room temperature, the solvent was evaporated and resulting crude product was subjected to column chromatography with ethyl acetate/hexane (2:1) to obtain the sensor, 2-(2-((1-(6-acetylpyridin-2-yl)ethylidene)amino)ethyl)-3’,6’-bis(diethylamino)spiro[isoindoline-1,9’-xanthen]-3-one (RhPK), (91 mg, 35 % yield, purity≥95). ¹H NMR(DMSO-d₆, 300 MHz δ(ppm): 8.19 (s, 1H), 8.15 (m, 1H), 7.94 – 7.71 (m, 2H), 7.52 (m, 2H), 7.05 (s, 1H), 6.37–6.30 (m, 6H), 3.37–3.18 (m, 12H), 2.73 (s, 3H), 2.64 (s, 3H), 1.08 (t, 12H). ¹³C NMR (DMSO-d₆, 75 MHz δ(ppm): 200.1, 199.3, 168.2, 166.7, 156.5, 153.8, 153.3, 152.6, 151.9, 148.7, 137.9, 136.9, 132.3, 131.1, 128.02, 124.7, 124.3, 123.7, 122.7, 121.7, 108.04, 105.6, 97.7, 64.8, 58.1, 50.5, 44.3, 40.9,25.5, 18.4, 13.4, 12.5. ESI-MS: found: m/z = 630.2 [M+H]⁺, 652.3 [M+Na]⁺, calculated for C₃₉H₄₃N₅O₃ = 629.4 The NMR and ESI-MS spectra are in Supplementary Material Figure S3.

Limit of detection (LOD) determination

The limit of detection (LOD) for Fe³⁺ was determined by fluorescence measurements of a series of known RhPK/Fe³⁺ concentrations, following a method described in prior publications (Zheng et al., Anal. Sci., 2017; Aydin, et al., 2020) . Briefly, fluorescence intensities (λ_Ex/λ_Em 510/580 nm) of a series of similar experiments at various concentrations (0.1 to 1 μM) of RhPK in the absence and presence of Fe³⁺ were recorded, respectively, in acetonitrile(ACN)/water (3/1, pH 7.3). A calibration curve between the Fe³⁺ concentration and the induced fluorescence enhancement was obtained using linear regression. According to IUPAC recommendations, the limit of detection (LOD) was calculated based on 3σ/k, where σ is standard deviation of the blank signals of RhPK and k is the slope of the calibration curve.

Cell culture and confocal studies

Human primary fibroblast ws1 cells and SH-SY5Y neuroblastoma cells were purchased from American Type Culture Collection (ATCC). SH-SY5Y cells were maintained similarly as described previously (Wei et al, 2012) in a 1:1 mixture of EMEM (ATCC) and Ham’s F12 medium (ATCC) supplemented with 10% fetal bovine serum (FBS) (Atlanta Biologicals) while ws1 cells were maintained in EMEM (ATCC) supplemented with 10% FBS. Both cells were cultured without antibiotics and incubated at 37 °C in a humidified 5% CO₂ atmosphere. Sub-culturing of the cells was routinely performed using 0.05% trypsin-EDTA solution (ATCC). Seeding of the cells onto 35 mm confocal culture chambers (MatTek) for 48 h at 5 × 10⁴ cells/chamber to allow them to grow until each chamber was 30 to 40% confluent. The medium in the dishes was replaced by fresh medium without FBS supplement before the imaging experiment and a stock solution of RhPK (10 mM in DMSO) was added to provide a final concentration of 10 μM. The cells were scanned under a Zeiss LSM 710 laser scanning confocal microscope to monitor the fluorescence of the sensor in living cells. The excitation wavelength of the laser was 543 nm and the emission signals were integrated between 547 and 703 nm for images with the RhPK sensor. For imaging the subcellular organelles mitochondria and lysosomes, MitoTracker Green FM and LysoTracker BlueDND-22 were used employing the excitation wavelengths recommended by the manufacturer (488 nm for MitoTracker and 405 nm for LysoTracker). Emission signals between 492 and 548 nm and 409 and 484 nm were integrated for MitoTracker and LysoTracker, respectively. The REUSE function (Zeiss software) was used to ensure that all images were acquired under the same instrumental conditions.

Results and Discussion

Sensor design and synthesis

Harnessing the structure-activity relationships learned from the development of rhodamine-based pH sensors (Yuan et al. 2011; Lv et al. 2013), we redesigned the Fe³⁺-receptor moiety to make it more acidic tolerant, creating a new sensor RhPK with a mixed O/N/N/O binding motif for Fe³⁺ (Schemes 1 and 2). The RhPK sensor was synthesized via a two-step process with overall yield of 35% (Scheme 1).

Scheme 1.

Synthesis of the Fe³⁺ sensor RhPK.

Scheme 2.

Proposed reversible 1:1 binding mode of RhPK with Fe³⁺ in acetonitrile/Tris-HCl buffer (10 mM, pH 7.3, v/v 3:1). X and Y represent possible co-ligands from the solvent.

pH stabilities of RhPK and RhPK + Fe³⁺

One of the disadvantages of rhodamine-based sensors is that the spirolactone gives response to hydrogen ions (Yuan et al. 2011; Lv et al. 2013). We thus firstly evaluated the pH response of RhPK in Acetinitrile(ACN)/water solution (3/1) and monitored by absorbance and fluorescence spectroscopy. As shown in Fig. 1, as expected, RhPK does not show any significant changes in absorption or fluorescence (excitation at 510 nm) over the pH range from 4.2 to 9.0, suggesting its good stability across this range. In the presence of 1 equivalent Fe³⁺, significantly elevated absorption (555 nm) and emission (580 nm) were observed, and the absorption is stable from pH 4.2 to 9.0 while the emission dropped slightly at pH 9.0. These interesting results indicate that RhPK is stable from pH 4.2 to 9.0 and the fluorescence of RhPK + Fe³⁺ system is stable from 4.2 to 8.0, suggesting that RhPK operates in the whole physiological pH range (including lysosomal pH) with a low background fluorescence.

Fig. 1.

Variation of (a) absorption and (b) fluorescence of RhPK and RhPK and 25 μM Fe³⁺. The absorption was measured at 555 nm and fluorescence at 580 nm with 510 nm excitation at various pH values in acetonitrile/H₂O (3/1, v/v) solution. The pH was adjusted using HCl and NaOH.

Spectroscopic studies on metal ion binding

The absorption and fluorescent properties of RhPK were evaluated in acetonitrile (ACN)/Tris-HCl buffer (10 mM, pH 7.3, v/v, 3:1) and the interactions between RhPK and various metal ions were investigated in the same buffer after 30 min incubation. The colorless solution of RhPK displays almost no absorption peak in the visible wavelength range (> 400 nm). The changes in color and absorption spectra after the addition of various metal ions are shown in Fig. S1. Only the addition of Fe³⁺ to the solution of RhPK showed an obvious pinkish red color with a strong absorption peak at 555 nm (ε = 3.2 ×10³ M⁻¹ cm⁻¹) in acetonitrile/Tris-HCl buffer (10 mM, pH 7.3, v/v 3:1). Compared with Fe³⁺, other metal ions, Zn²⁺, Cr³⁺, Ni²⁺, Hg²⁺, Mn²⁺, Ag⁺, Pb²⁺, Fe²⁺, Cu⁺, Cu²⁺, Co²⁺, Na⁺, K⁺, Mg²⁺ and Ca²⁺ did not induce any changes in color or significant absorption. However, Hg²⁺ and Cr³⁺ at 100 μM did induce minor absorbance of up to 12% of the Fe³⁺ signal intensity at 555 nm. As Hg²⁺ and Cr³⁺ are generally not considered to be essential biometals and are present in cells at negligible concentrations under normal conditions (Williams and Fraústo da Silva, 2000), they are unlikely to significantly interfere with the detection of Fe³⁺ in cells.

The fluorescence responses of RhPK to different metal ions are shown in Fig. 2a. The RhPK solution shows weak fluorescence at 580 nm (excitation at 510 nm) in the absence of metal ions. When Fe³⁺ was added, a significant enhancement of fluorescence (>12-fold with 1.0 equivalent of Fe³⁺) was observed, demonstrating that RhPK is a highly specific turn-on fluorescent sensor for Fe³⁺. Kinetics studies revealed a rapid reaction with Fe³⁺, with turn-on fluorescent response in seconds (Fig. S2), a property necessary for real-time imaging applications.

Fig. 2.

(a) Fluorescence responses for excitation at 510 nm of 100 μM RhPK sensor to various metal ions (100 μM for Ni²⁺, Cu⁺, Cu²⁺, Zn²⁺, Pb²⁺, Fe²⁺, Fe³⁺, Cr³⁺, Hg²⁺, Mn²⁺, Ag⁺, and Co²⁺; 200 μM for K⁺, Na⁺, Ca²⁺ and Mg²⁺) in acetonitrile/Tris buffer (10 mM, pH 7.3, v/v 3:1). (b) Fluorescence of 100 μM sensor in the presence of various metal ions (grey bars) and the subsequent addition of Fe³⁺ (black bars) in acetonitrile/Tris-HCl buffer (10 mM, pH 7.3, v/v 3:1). The bars represent the fluorescence intensity at 580 nm.

We also examined the interferences from the other metal ions with RhPK for its response to Fe³⁺. Although we have shown that the 12 metal ions tested do not significantly turn-on the color nor the fluorescence of RhPK, quenching is also a possibility. Each of the 12 metal ions was pre-incubated with RhPK before 1 equivalent of Fe³⁺ was added and the fluorescence response was measured. As shown in Fig. 2b, the fluorescent intensity of RhPK with Fe³⁺ is unaffected significantly by the presence of any of the other metal ions tested, except Hg²⁺ and Cr³⁺ (at 100 μM) which showed ~13% increase and ~29% decrease in fluorescence intensity at 580 nm, respectively. The reasons for the interference from Hg²⁺ and Cr³⁺ are not known but it is possible for Cr³⁺ to partially bind to the ONNO donors at the metal receptor moiety of RhPK, producing non-fluorogenic species. A detailed study on the mechanisms of interference is beyond the scope of this study. This interference may affect Fe³⁺ detection in test tubes. However, as Hg²⁺ and Cr³⁺ are present in cells at negligible concentrations under normal conditions (Williams and Fraústo da Silva, 2000), they are unlikely to significantly interfere with the qualitative detection of Fe³⁺ in cells.

Stoichiometry, binding affinity and detection limit

Job’s method and an absorption titration were applied to study the binding stoichiometry between RhPK and Fe³⁺. As shown in Fig. 3, the absorption band at 555 nm was enhanced upon the addition of increasing concentration of Fe³⁺ into RhPK solution, and saturated after 1 equivalent of Fe³⁺ was added. The titration curve (a plot of RhPK versus Fe³⁺ concentration) increased linearly and plateaued at 1:1 ratio of the sensor and Fe³⁺ (Fig. 3 inset). The saturation of absorption suggests the structural changes on RhPK induced by Fe³⁺ is stoichiometric, most likely owing to the formation of a 1:1 Fe³⁺–RhPK complex in solution (Scheme 2). This stoichiometry was also confirmed by a Job’s plot (Fig. S3). The binding constant between Fe³⁺ and RhPK (1:1 complex) was determined using a previously reported method (Wei et al.2012; Ozdemir et al.2018; Aydin et al. 2020) with absorption values at 555 nm, and determined to be 1.54 × 10⁷ M⁻¹ (S4).

Fig. 3.

Absorption spectra for the titration of 100 μM RhPK sensor with the concentration of FeCl₃ (10, 15, 25, 35, 50, 70, 90, 100, 150 μM) in acetonitrile/Tris-HCl buffer (10 mM, pH 7.3, v/v 3:1). The absorbance was measured at 555 nm. The inset shows the titration curve.

The limit of detection (LOD) for Fe³⁺ was determined by fluorescence measurements of a series of known RhPK/Fe³⁺ concentrations as described in the experimental section. The calibration curve (Fig.S4) presented a linear relationship between Fe³⁺ concentration and the induced fluorescence enhancements, with a correlation coefficient of 0.9786. According to IUPAC recommendations, the limit of detection (LOD) was calculated to be approximately 0.05 μM (or 50 nM) based on 3σ/k.

Reversibility in Fe³⁺-binding

Reversibility experiments were carried out by adding EDTA to a solution of RhPK and Fe³⁺ complex in acetonitrile/Tris-HCl buffer (10 mM, pH 7.3, v/v 3:1). In the absence of EDTA, the complex was colored and fluorescent. After adding EDTA, the absorption of the complex decreased in intensity and finally, disappeared (Fig. S5), suggesting reversible binding between RhPK and Fe³⁺. The spectroscopy, affinity constant, and binding properties of RhPK and Fe³⁺ are similar to those of previously developed binding-based rhodamine Fe³⁺ probes (Chakraborty et al. 2020; Shellaieh et al. 2020; Wei et al.2012; Aydin et al. 2014; Ozdemir et al.2018; Aydin et al. 2020), suggesting a similar mechanism involving a coordination-inducted fluorescent activation (Wei et al. 2012; Aydin et al. 2014; Ozdemir et al. 2018; Aydin et al. 2020). Another mechanism without the involvement of direct Fe³⁺-coordination, instead, a Fe³⁺ -induced ring-opening of the rhodamine spiroclic lactone was proposed recently in a new rhodamine-based Fe³⁺probe RXY (Zhang et al. 2021). However, RXY has a much lower Fe³⁺-binding constant (3.0 × 10⁴ M⁻¹) (Zhang et al. 2021) than for RhPK (1.54 × 10⁷ M⁻¹). Thus it is unlikely that RhPK has a similar sensing mechanism as RXY. A possible mechanism for the reaction of RhPK and Fe³⁺ and the binding mode is shown in Scheme 2.

Cell imaging

Encouraged by the above promising results, we used a laser scanning confocal microscope to examine the ability of RhPK to detect Fe³⁺ in living cells. The labile Fe³⁺ pools in human primary fibroblast cells (ws1) were tested first. After ws1 cells were incubated with RhPK (10 μM) for 30 min, the cells were scanned by the confocal microscope. As shown in Fig. 4, we observed weak “turn-on” fluorescence (Fig. 4b) which is likely being triggered by endogenous labile Fe³⁺. Another possibility may be cleavage of the sensor intracellularly or due to the acidic environment of certain intracellular organelles (e.g., lysosomes/endosomes), as low pH may also trigger the turn-on response of spicrolactone based sensors (Chartres et al. 2011).

Fig. 4.

Laser confocal microscopy images with differential interference contrast (DIC) of (a) ws1 cells only; (b) ws1 cells with 10 μM RhPK probe after 30 min incubation; (c) ws1 cells incubated with 50 μM Fe³ for 30 min; and (d) ws1 cells incubated with a cell permeable Fe³⁺-chelator salicylaldehydeisonicotinoyl hydrazone (SIH) (100 μM) overnight and the RhPK probe was added. Fluorescence was collected from 547 to 703 nm for RhPK. (e) Bar chart of the mean intensity of each of the above conditions.

To more carefully examine these possibilities, two approaches were applied using a similar previously reported strategy (Wei et al., 2012; Ozdemir et al. 2018). First, the cells were loaded with Fe³⁺ by supplementation of the cell culture and second by Fe³⁺ depletion by a cell-permeable Fe³⁺-specific chelator. We carefully examined the fluorescence changes under both conditions. The confocal fluorescence images (Fig. 4) of RhPK in fibroblast (ws1) cells before (Fig. 4b) and after (Fig. 4c) loading with 50 μM Fe³⁺ were compared. Brighter and more widely distributed fluorescence signals in the cytosol were observed in the Fe³⁺-loaded fibroblast (ws1) cells (Fig. 4c), demonstrating a positive response of RhPK to elevated exchangeable Fe³⁺ levels in the cells. To deplete exchangeable Fe³⁺ from the cells, the fibroblast (ws1) cells were incubated overnight with a cell permeable Fe³⁺-chelator salicylaldehydeisonicotinoyl hydrazone (SIH) (Buss et al. 2003; Sheftel et al. 2007; Richardson et al. 2010; Wei and Guo 2007). When being treated with RhPK at the same concentration, the iron-depleted cells displayed lower intensity fluorescence (Fig. 4d). The bar chart in Fig. 4e compared the fluorescent intensities under the various conditions. These results demonstrate that the fluorescence responses of RhPK in fibroblast (ws1) cells are triggered by cellular exchangeable Fe³⁺, not by cleavage or acidic pH, thus confirming that RhPK unambiguously detects endogenous exchangeable basal level Fe³⁺ in the fibroblast (ws1) cells and the dynamic changes of Fe³⁺ in the fibroblast (ws1) cells.

Subcellular imaging

The fluorescent signals detectable by RhPK in live fibroblast ws1 cells are scattered in the cytosol, suggesting that the exchangeable Fe³⁺ ions may be located in certain subcellular organelles in ws1 cells. To further probe the capability of RhPK in subcellular Fe³⁺ imaging, ws1 cells with Fe³⁺-loaded for stronger signals were treated with RhPK, MitoTracker Green FM to reveal possible exchangeable Fe³⁺ pools in mitochondria, and LysoTracker blue DND-22 to reveal possible exchangeable Fe³⁺ pools in endo/lysosomes in ws1 cells. Fig. 5 shows the subcellular distribution of exchangeable Fe³⁺ pools in ws1 cells revealed by the RhPK sensor. The fluorescent signals of RhPK-Fe(III) complex in cells are shown in red (Fig. 5b). Fig. 5c shows the MitoTracker dyed mitochondria (green) and Fig. 5d shows the LysoTracker blue DND-22 dyed endo/lysosomes (blue) in ws1 cells. Merging Figs. 5b and 5c gives rise to Fig. 5e which reveals the co-localization signals in yellow. As shown in Fig. 5e, MitoTracker signals are completely co-localized with the RhPK-Fe(III) signals, suggesting the labile Fe³⁺ ions are located in mitochondria in ws1 cells. In contrast, as shown in Fig. 5f which is from the merging of Figs. 5b and 5d, no co-localization was observed between RhPK-Fe(III) signals and the LysoTracker blue signals, suggesting little or no labile Fe³⁺ ions in endo/lysosomes in ws1 cells. Merging Figs. 5b, 5c and 5d gives rise to Fig. 5g which displays only yellow and blue but no red, green or purple, suggesting that all and only the mitochondria in ws1 cells contain detectable labile Fe³⁺ ions while endo/lysosomes contain little or no labile Fe³⁺ ions under these conditions. This conclusion is confirmed by independent co-localization experiments shown in Figs. S6 and S7. Interestingly, our ferrous ion (Fe²⁺) sensor RhT also detected labile Fe²⁺ ions only in mitochondria in ws1 cells (Maiti et al. 2015), corroborating the mitochondrial location of iron pools in ws1 cells.

Fig. 5.

Laser confocal microscopy images of intracellular colocalization experiments of 20 μM RhPK incubated with Fe³⁺-loaded fibroblast cells (ws1) (pre-incubated with 20 μM Fe³⁺ for 30 min) co-dyed with MitoTracker Green FM (100 nM, incubated for 30 min) and LysoTracker Blue DND-22 (50 nM, incubated for 120 min). (a) Differential interference contrast (DIC) image of the ws1 cells with 20 μm scale bar. (b) RhPK fluorescence collected from 580 to 703 nm (red). (c) Fluorescence of MitoTracker from 492 to 548 nm (green). (d) Fluorescence of LysoTracker from 409 to 484 nm (blue). (e) Merged images of the DIC image (a) and the fluorescence images of (b) and (c). Colocalization regions are in yellow and non-overlapping regions are in red or green. (f) DIC image of (a) and fluorescence images of (b) and (d) merged together. Overlapping regions would be in purple (not observed) and non-overlapping regions in red or blue. (g) Images of (a), (b), (c) and (d) merged together showing that the RhPK-Fe³⁺ images are almost completely colocalized with those of MitoTracker but not colocalized with LysoTracker. (h) Images of (a), (c) and (d) merged showing no overlapping region between lysosomes and mitochondria.

In our previous work, the RPE sensor revealed chelatable Fe³⁺ pools in mitochondria and endo/lysosomes in human SH-SY5Y neuroblastoma cells (Wei et al. 2012) and our RhHPA sensor revealed chelatable Fe³⁺ pools in mitochondria and endo/lysosomes in bovine aortic endothelial cells (BAEC) (Ozdemir et al. 2018). However, the RhPK sensor in this work detected chelatable Fe³⁺ in mitochondria only in human primary fibroblast ws1 cells but not in endo/lysosomes. To verify whether RhPK has the capability to detect endo/lysosomal Fe³⁺ ions, RhPK was investigated in human SH-SY5Y neuroblastoma cells with co-localization experiments. SH-SY5Y cells were first pre-incubated with Fe³⁺ ions and incubated with RhPK, MitoTracker and LysoTracker. The fluorescent signals in the cells were monitored by confocal microscopy. As shown in Fig. 6, a portion of the RhPK-Fe(III) signals are co-localized with all MitoTracker signals (Fig. 6e) and a fraction of the RhPK-Fe(III) signals are co-localized with all LysoTracker blue DND-22 signals (Fig. 6f). A complete co-localization was observed among the signals from RhPK-Fe(III), MitoTracker and LysoTracker (Fig. 6g), suggesting chelatable Fe³⁺ ions detected by RhPK are indeed in both mitochondria and endo/lysosomes in human SH-SY5Y neuroblastoma cells. These are the same results observed previously using the RPE probe (Wei et al. 2012) and demonstrate that the RhPK sensor is capable of detecting Fe³⁺ ions in subcellular structures such as mitochondria and the acidic organelles such as endo/lysosomes. However, human primary fibroblast ws1 cells appears to have little chelatable Fe³⁺ in endo/lysosomes, suggesting different cell types may handle iron differently and have different Fe³⁺ storage at subcellular compartments.

Fig. 6.

Laser confocal microscopic images of intracellular colocalization experiments of 10 μM RhPK incubated with Fe³⁺-loaded human SH-SY5Y cells (pre-implemented with 10 μM Fe³⁺) co-dyed with MitoTracker Green FM (100 nM, incubated for 30 min) and LysoTracker Blue DND-22 (50 nM, incubated for 120 min). (a) Differential interference contrast (DIC) image of cells with 10 μm scale bar. (b) RhPK-Fe³⁺ fluorescence from 580 to 703 nm (red). Stronger fluorescence suggests elevated Fe³⁺ levels in Fe³⁺-loaded human SH-SY5Y cells. (c) Fluorescence of MitoTracker fromt 492 to 548 nm (green). (d) Fluorescence of LysoTracker from 409 to 484 nm (blue). (e) Merged images of the DIC image (a) and the fluorescence images of (b) and (c). The yellow regions represent the co-localization signals and non-overlapping regions remain in red, showing that RhPK-Fe³⁺ is partially colocalized with mitochondria. (f) Merged images of the DIC image (a) and the fluorescence images of (b) and (d). The purple regions represent the colocalization signals and non-overlapping regions are in red, showing that RhPK-Fe³⁺ is partially co-localized with endo/lysosomes. (g) Merged images of (a), (b), (c) and (d). Co-localization regions are in yellow or purple exclusively; no non-overlapping regions were observed, suggesting that chelatable Fe³⁺ is still localized in mitochondria and endo/lysosomes under the conditions. (h) Merged images of (a), (c) and (d), showing no overlapping region between mitochondria and endo/lysosomes.

The cell morphology, growth rate and cell density in RhPK treated cells (both the SH-SY5Y and ws1 cells) are normal and similar as those of the control cells without RhPK treatment, even after 24 h incubation, suggesting that the cells are healthy in the presence of RhPK which is not toxic to the cells under the experimental conditions. RhPK exhibits excellent selectivity for Fe³⁺ with an ideal working pH range (pH 4.2 to 8.0) that covers the whole cellular pH range. It binds Fe³⁺ tightly yet reversible with fast fluorescent responses and excellent sensitivity (detection limit of 50 nM in solution) and offers excellent cell permeability and detects basal level chelatable Fe³⁺ and its dynamic changes in living cells at subcellular resolution. The excellent biochemical and spectroscopic properties of the RhPK sensor provide a valuable tool to study the cell biology of Fe³⁺ in a broad range of cell types.

The synthesis of RhPK involves only two steps under mild conditions with a good yield. The total chemical and material cost for the synthesis of 1 g of RhPK is ~$141 USD and may be completed in 3 days including purification. The low cost of synthesizing RhPK probe may provide its broad applications in research and the biomedical industry.

Conclusion

By carefully tuning the metal-binding moiety, we have designed and developed a new fluorescent probe RhPK which displays highly selective and sensitive “off-on” fluorescence responses towards Fe³⁺ in solution and in living cells. The major advantage of the RhPK sensor is its acidic resistance and operation in an ideal pH range (4.2 to 8.0) that covers all cellular pH values. The probe binds strongly and rapidly to Fe³⁺ at 1:1 (RhPK/Fe³⁺) ratio with fluorescence response in seconds and a detection limit of 50 nM for Fe³⁺ in solution. Confocal imaging has demonstrated that the RhPK probe is capable of monitoring basal level labile Fe³⁺ and its dynamic changes in living cells at subcellular resolution. It revealed one Fe³⁺ pool in the mitochondria of human primary fibroblast cells for the first time and confirmed two Fe³⁺ pools in the mitochondria and endo/lysosomes in human SH-SY5Y neuroblastoma cells. Our results also reveal that different cell types have distinct Fe³⁺ storage in subcellular compartments and suggest they may handle iron distinctly. The RhPK probe is powerful for rapid and sensitive bioimaging of Fe³⁺ at subcellular level, enabling the unambiguous detection of labile Fe³⁺ pools at the entire cellular pH range, which is of great significance to understand the biological chemistry of Fe³⁺ and its roles in physiological processes and diseases.