The development and application of a methodology for measurement of oxygen within single mammalian cells are presented, which employ novel macromolecular near infrared (NIR) oxygen probes based on new metalloporphyrin dyes. The probes, which display optimal spectral characteristics and sensitivity to oxygen, excellent photostability, and low cytotoxicity and phototoxicity, are loaded into cells by simple transfection procedures and subsequently analyzed by high-resolution fluorescence microscopy. The methodology is demonstrated by sensing intracellular oxygen in different mammalian cell lines, including A549, Jurkat, and HeLa, and monitoring rapid and transient changes in response to mitochondrial uncoupling by valinomycin and inhibition by antimycin A. Furthermore, the effect of ryanodine receptor-mediated Ca2+ influx on cellular oxygen uptake is shown by substantial changes in the level of intracellular oxygen. The results demonstrate the ability of this technique to measure small, rapid, and transient changes in intracellular oxygen in response to different biological effectors. Moreover, this technique has wide ranging applicability in cell biology and is particularly useful in the study of low oxygen environments (cellular hypoxia), mitochondrial and cellular (dys)function, and for therapeutic areas, such as cardiovascular and neurological research, metabolic diseases, and cancer.
- mitochondrial function
molecular oxygen is a key substrate in aerobic biological systems, providing the terminal electron acceptor of the electron transport chain (7). The rate of oxygen consumption is closely regulated by the ATP/ADP ratio (24), while also being highly dependent on the concentration of available oxygen (31) and the action of signaling molecules such as NO (9, 37) and Ca2+ (10, 23), which, in turn, are related to pathways of cell survival and death (22). Furthermore, the induction of cellular hypoxia results in the activation of a specific adaptive transcriptional response governed by the hypoxia inducible transcription factors (9). Oxygen consumption is, therefore, an informative marker of cellular metabolism, which is broadly applicable to various biological systems from mitochondria to cells to whole organisms. It may be used to elucidate biochemical pathways of the cell and alterations in metabolism caused by various stimuli or disease states. Accordingly, there is a requirement for methods that can effectively monitor cellular oxygen levels and oxygen consumption rates.
Optical schemes, based on the quenching by molecular oxygen of long-decay fluorescent and phosphorescent dyes, such as metalloporphyrins and ruthenium(II) complexes, have been under active development in recent years (32). The group of Wilson and coworkers (8, 47) have used extracellular probes based on phosphorescent dyes palladium(II) tetracarboxyphenyl porphyrin and palladium(II) tetrabenzo porphyrin (PdTBP) in low-resolution lifetime-based imaging of tissue and tumor oxygenation. A similar principle has been applied to develop simple formats for measuring biological oxygen consumption in microtiter plates on a fluorescent plate reader using phosphorescent solid-state oxygen sensors (18, 27) and water-soluble probes (11). These systems, which can measure oxygen consumption noninvasively, on a microscale and with high sample throughput, have been demonstrated with bacteria (18), isolated mitochondria (14), cell lines (12), and small organisms (26). All these, however, rely on measurement “extracellular” oxygen in relatively large samples.
The ability to measure “intracellular” oxygen provides a powerful tool for more detailed and information-rich studies of metabolism and cellular responses. Electrochemical microsensors have limited applicability due to their invasive and consumptive nature (45). Fluorescence-based schemes can potentially facilitate sensing of intracellular oxygen using different imaging formats, such as standard fluorescence microscopy, confocal microscopy (20), fluorescence lifetime imaging microscopy (46), and phase-fluorimetric imaging microscopy (38). Such systems have been explored to some degree mainly using particulate probes, such as polymeric nanoparticles impregnated with oxygen-sensitive dyes, such as ruthenium tris(diphenyl phenanthroline) (RuDPP) (43) and platinum(II) octaethyl porphyrin ketone (PtOEPK) (3, 20) and loaded into cells by microprojectile delivery. Other variations include “lipobeads” composed of a polymer particle impregnated with RuDPP and a phospholipid shell, which are incorporated into macrophages by phagocytosis (16, 17) and microspheres doped with platinum(II) tetrakis-(pentafluorophenyl) porphyrin introduced into large plant cells by microinjection (38).
Notwithstanding these advances, the goal of high-resolution imaging of oxygen within the cell by relatively simple and convenient means has still to be achieved. The main requirement is an oxygen probe that should combine the properties of high photostability, simple and efficient means of delivery into the cell, minimal cytotoxicity and interference with cell function, optimal photophysical properties, and operational performance in conditions of a typical live cell imaging experiment. Particulate probes can contain photostable dyes; however, loading by projectile delivery or endocytosis is rather inefficient, may cause irreparable damage to the cell, while random distribution of the small number of particulate sensors within the cell may give a poor representation of the intracellular oxygen distribution.
The use of hydrophilic molecular oxygen probes can circumvent the problems of delivery into the cell, but such probes known so far display inadequate photostability for fluorescence microscopy applications, where high-intensity illumination and prolonged exposures are normally used. In this paper, we describe a new class of phosphorescent probes and simple experimental procedures for sensing intracellular oxygen. The performance of new oxygen probes and the utility of measurement methodology are demonstrated by monitoring oxygen in individual mammalian cells and cellular responses to various effectors of metabolism by high-resolution fluorescent microscopy imaging.
MATERIALS AND METHODS
Phosphorescent oxygen probes comprising covalent conjugates of platinum(II) coproporphyrin (PtCP), platinum(II) coproporphyrin ketone (PtCPK), and palladium(II) coproporphyrin ketone (PdCPK) with BSA were synthesized using procedures described elsewhere (11, 21) and are now commercially available from Luxcel Biosciences (Cork, Ireland). Oxyphor G2 probe (PdTBP dendrimer) was obtained from Harvard Apparatus (Kent, UK). 4,6-Diamidino-2-phenylindole (DAPI) and fura-2 AM ester were obtained from Invitrogen. HeLa human cervical carcinoma, A549 human lung carcinoma, and Jurkat T-cell cell lines were obtained from The American Type Culture Collection. Fugene transfection agent was from Roche, 4-chloro-m-cresol (CMC) was from Merck Biosciences. DMEM, RPMI, BSA, Escort III transfection reagent, FBS, penicillin streptomycin solution, trypsin, valinomycin, antimycin A, methanol, Mowiol, and all analytical reagents were obtained from Sigma-Aldrich.
Absorbance spectra of the probes were measured in a 1-cm quartz cell on a HP8453 diode array spectrophotometer (Agilent). Excitation and emission spectra, lifetimes, quantum yields, and oxygen calibrations of the probes were measured on LS-50B (Perkin Elmer) and Cary Eclipse (Varian) spectrofluorometers. Phosphorescence decay curves were measured using phosphorescence decay measurement mode; lifetimes were calculated from single and double exponential decay fitting using Origin software. Quantum yields of each probe were determined in PBS, with respect to PtCP standard having a yield of 28% in 3 mM CTAB-sulfite at 22°C (29), following spectral correction for PMT response. Oxygen calibrations were performed at 1 μM of probe in PBS. Standard gas mixtures (BOC Gases, Cork, Ireland) containing 20.5%, 10%, 5%, 2%, 0.5%, and 0% of O2 balanced with N2 (100%, 48.7%, 24.4%, 9.7%, 2.4%, and 0% of air saturation, respectively) were bubbled through the cuvette until gas equilibrium was reached, at which point the phosphorescent signal was recorded.
Cell culture and probe loading.
A549 and HeLa cell lines were cultured in 75-cm2 adherent cell flasks in DMEM supplemented with 10% FBS, 2 mM l-glutamine, 100 U/ml penicillin, and 100 μg/ml streptomycin. Twenty-four hours before microscopy or transfection, A549 cells were removed from the flask surface using PBS containing 2 mM EDTA and 1× trypsin, and aliquotted in 1-ml volumes into 35-mm glass bottom dishes (Mattek) or 10-mm round glass coverslips (Scientific Laboratory Supplies). Jurkat T-cells were grown in RPMI medium supplemented with 10% FBS, 2 mM l-glutamine, 100 U/ml penicillin and 100 μg/ml streptomycin.
The PtCPK or PdCPK probe (10 nmol/vial) was reconstituted in 100 μl of PBS to produce 100 μM stock. Similarly, the PtCP and Oxyphor G2 probes were made up to 100 μM in PBS. These stock solutions were stored in the dark at +4°C until further use. Probe loading using the Escort III reagent was carried out as per manufacturer's instructions. Briefly, cells in 35-mm dishes were washed three times and incubated in serum-free medium. Five microliters of transfection reagent were added to 200 μl of serum-free medium followed by the addition of an aliquot of probe to give a concentration of 5 μM. This solution was incubated at 37°C for 15 min followed by the addition to the cell culture dish containing 800 μl of serum-free medium. The cells were incubated for 5 h in a 5% CO2 incubator at 37°C and then were washed three times with 1 ml of serum-free medium and then used for optical measurements.
The effect of probe loading on the viability of cells was measured using a ViaCount assay kit (Guava Technologies), carried out as per manufacturer's instructions. Briefly, after a 5-h incubation period, either loaded cells or unloaded (control) cell samples (both in serum-free medium) were mixed with the ViaCount reagent and ran on a PCA-96 flow cytometer (Guava Technologies) with results analyzed using Cytosoft 2.5.5 software.
For confocal fluorescence imaging, additional cell fixing was carried out by submerging HeLa cells cultured on cover slips and loaded in 100% methanol. Costaining of the nucleus with DAPI was carried out by permeabilization of the fixed cells using PBS containing 0.2% Triton X-100 and incubation with 5 μM dye for 20 min. After three washes with PBS, the coverslips were mounted onto slides with Mowiol.
For imaging intracellular calcium and its fluctuations, A549 cells were loaded with fura-2 AM ester by incubation with 2 μM of the dye in PBS for 45 min at 37°C followed by washing and incubation for a further 30 min at 37°C before measurement.
Live-cell imaging experiments with the oxygen probes were carried out on an Olympus IX51 inverted fluorescence microscope equipped with 75 W Xenon Arc Lamp (Cairn), an Optoscan Monochromator (Cairn), and an Orca-ER charge-coupled device (CCD) camera (Hamamatsu), using an Olympus UplanFl 1.3 NA 100 × oil-immersion objective. The images were analyzed using Kinetic Imaging AQM Advanced software (ver. 6). The PtCPK and PdCPK probes were imaged by excitation at 395 nm and 400 nm, respectively, while the Oxyphor G2 probe was excited at 430 nm and the PtCP probe at 380 nm. The phosphorescence emission of all probes was collected using a 600-nm cut-off filter. Measurement of changes in intracellular Ca2+ concentration was carried out using the same instrument with ratiometric imaging of fura-2 following excitation at 340 and 380 nm and measurement of emission at 510 nm using a 400-nm cut-off filter. The ratiometric image was transformed and analyzed using the software described above.
Confocal fluorescence microscopy of fixed A549 cells was carried out using an Olympus FV1000 confocal laser scanning biological microscope. The PtCPK dye was imaged using an oil immersion 60× objective with excitation by a 488-nm laser and measurement of emission using a 520-nm cut-off filter. The DAPI nuclear stain was imaged by excitation at 405 nm and measurement of emission using a 420- to 480-nm bandpass filter. Images were analyzed using FV1000 Viewer software (Olympus) and Adobe Photoshop.
Design of intracellular oxygen probes and their photophysical and sensing properties.
The oxygen probes based on the PtCP dye conjugated to a polypeptide carrier developed for extracellular use in oxygen consumption assays (11, 12, 33) were seen to incorporate a number of features important for intracellular oxygen imaging applications. Their hydrophilic nature and relatively small molecular size facilitate more simple loading procedures than those used for the particulate probes. In particular, the hydrophilic polycarboxylic dye and anionic carrier protein (acidic pI for albumin) are compatible with cationic transfection reagents (liposomal and nonliposomal) developed for transfection of cells with nucleic acids and proteins. However, insufficient photostability of PtCP-based probes makes them inappropriate for microscopy imaging.
To address the problem of photostability, new probes based on hydrophilic porphyrin-ketone dyes, namely, PtCPK and PdCPK, have been deployed. As opposed to the hydrophobic porphyrin-ketones (mainly PtOEPK), which have been in use for some time and known to have high photostability and red-shifted near infrared (NIR) spectral characteristics (34), water-soluble coproporphyrin ketones have so far not been exploited for oxygen sensing. The use of unconjugated dyes for intracellular oxygen imaging is associated with partitioning, high cytotoxicity due to the production of singlet oxygen (4), and dye leakage. To offset this effect, PtCPK and PdCPK were conjugated to serum albumin to produce a new generation of macromolecular oxygen probes. Previous studies with these probes used in the extracellular environment have shown that their oxygen-sensing properties remain stable and unaltered by parameters such as pH, common biological ions, and molecules, including proteins, DNA, and metabolites. So far, any interferences that have been reported (e.g., by heavy metals, some extrinsic compounds) occurred at concentrations far outside those present in a standard biological assay and were compensated for (25, 42).
The photophysical properties and oxygen-sensing characteristics of the PtCP-, PtCPK- and PdCPK-based probes are summarized in table 1 and their absorption and emission spectra are shown in Fig. 1, A and B. The new probes display two intense absorbance bands and can be effectively excited at 370–410 nm (Soret band) or at 580–600 nm (Q band). Compared with the PtCP probe, the PtCPK and PdCPK probes have their Q bands enhanced and red-shifted by ∼60 nm. They emit in the near-infrared with maxima at 767 and 796 nm, respectively. The quantum yields are smaller than for the PtCP probe, while the phosphorescence lifetimes are threefold shorter for PtCPK and twofold longer for PdCPK (2, 28, 35). The probes are therefore compatible with conventional fluorescent microscopy systems.
The sensitivity to oxygen for the PtCPK and PdCPK probes was within the required range and close to that of the PtCP probe (11). For the PtCPK probe, the changes in phosphorescence intensity over the entire physiological oxygen range (0–100% of air saturation or 0–21 kPa) are shown in Fig. 1C. This probe displays an overall change in intensity of 2.14 with a measurable change between even 2.5% and 0% of air saturation. Phosphorescence lifetime measurements showed a slightly less 1.7-fold quenching at 100% air saturation, calibration function (measured on a spectrometer), and Stern-Volmer plots are presented in Fig. 1D. The PdCPK probe, as expected, displayed a greater 4.44-fold signal change between 100% and 0% of air saturation; its intensity and lifetime changes are shown in Fig. 1, E and F. The PdCPK probe is more suitable for low oxygen range.
Probe loading and live cell imaging.
Loading of all three oxygen probes was carried out in a number of cell lines, including A549, Jurkat, and HeLa cells, using liposomal transfection reagents Escort (Sigma) and Fugene (Roche). Loading was initially assessed by measurement on a fluorescent lifetime spectrometer of the PtCP probe, both in the extracellular medium and in the intracellular environment, that is, after loading and washing of cells. A sample of Jurkat cells with extracellular PtCP probe displayed a lifetime of 24 μs, equivalent to air-saturated oxygen concentration. In loaded cells, the probe lifetime was 31 μs, indicating an average level of intracellular oxygenation of 71%. These experiments showed both efficient and reproducible loading of the PtCP probe into transfected cells (phosphorescent signal corresponded to ∼10 nM concentration), maintained oxygen gradients in viable cells, and no probe leakage from the cells. Similar behavior was expected for the PtCPK and PdCPK probes, which are of the same nature. The effect of probe loading on viability of the cells was analyzed by flow cytometry and showed that cells loaded in serum-free medium for 5 h retained 93.7% viability with respect to control cells incubated for the same period in the same medium.
To demonstrate loading efficiency and intracellular probe localization, A549 cells loaded with the PtCPK probe, fixed and costained with the nuclear stain DAPI were imaged by confocal fluorescence microscopy (Fig. 2A). This illustrates that, even using off-peak excitation at 488 nm, the intracellular probes and loading technique provide sufficiently high, even, and reproducible loading of the cells, to facilitate a detailed examination of the cells by fluorescence microscopy and for measuring oxygen concentration and its dynamic changes within the cells.
To explore the issue of photostability, the metallocoproporphyrin ketones (MeCPK)-based probes were tested compared with the PtCP probe and another hydrophilic oxygen probe Oxyphor G2 (8). Live HeLa cells were loaded with different probes, then fixed and analyzed on a fluorescent microscope. Fig. 3A shows that under continuous illumination at the Soret maximum of each probe over a 20–25 min time frame, the PtCPK probe is stable, whereas both PtCP and Oxyphor G2 probes degrade considerably. The photostability of the PtCPK probe was also conserved in live cells (see Fig. 3B), whereas the instability of the PtCP and Oxyphor G2 probes was amplified, possibly due to the increased effect of singlet oxygen photosensitization. Intracellular imaging of the MeCPK probes using confocal fluorescence microscopy further verified their stability under continuous excitation with a 405-nm laser (see Fig. 2A).
Monitoring cellular responses by fluorescent oxygen imaging.
The intracellular imaging methodology has been applied to monitoring of dynamic changes in cellular oxygen. Adherent A549 and HeLa cells were grown under normal conditions, loaded with either probe, and subjected to various conditions or treated with various effector compounds with known mechanisms of action on cell metabolism, while observing them under the fluorescent microscope. Figure 4 represents processed fluorescent imaging data showing cellular responses to various treatments, which induce changes in intracellular oxygen concentration.
Valinomycin is an ionophore that funnels K+ ions across the inner mitochondrial membrane, dissipating the negative potential in the mitochondrial matrix and thus uncoupling oxidative phosphorylation causing increased O2 consumption (36). The action of valinomycin on A549 cells and HeLa cells (Fig. 4, A and B) showed the response to be rapid and cell specific. In the A549 cells, the PtCPK and PdCPK probes displayed a similar response to valinomycin with a greater range displayed by the PtCPK probe. This indicates that the average base concentration of O2 in respiring A549 cells was greater than 50% of air saturation. In HeLa cells, greater sensitivity to the effects of valinomycin was displayed by the PdCPK probe, indicating a lower base level of intracellular oxygen than in A549 cells. Images of the cells with PdCPK probe before and after the addition of valinomycin are shown in Fig. 2, B and C.
Antimycin A inhibits oxidative phosphorylation by binding cytochrome b of the electron transport chain and blocking electron transport between cytochrome b and cytochrome c1 (39). On the addition of 2 μM antimycin A to A549 cells (Fig. 4C), an approximate 40% drop in PtCPK phosphorescence was observed, as the base rate of respiration was inhibited, resulting in intracellular oxygen reaching extracellular levels (100% of air-saturation). Control experiments with pure DMSO (solvent for valinomycin and antimycin A) showed no change in intracellular PtCPK phosphorescence (Fig. 4D).
CMC is an agonist of ryanodine receptor type 1 (RYR1) located in the sarcoplasmic recticulum, causing release of Ca2+ into the cytosol (41). The A549 cell line possess RYR1 (44), while recent studies have demonstrated specific use of CMC for stored calcium release in this cell line (30). This increased cytosolic calcium results in an increase in mitochondrial calcium by Ca2+ influx through the mitochondrial Ca2+ uniporter (19), present in the inner mitochondrial membrane. Increased mitochondrial calcium is known to increase O2 consumption by a combination of mitochondrial membrane depolarization (10) and stimulation of the Ca2+-dependent dehydrogenase enzymes of the tricarboxylic acid cycle (5, 6). This double effect may result in the double peak in probe phosphorescence with intensities dropping to below base levels, demonstrating both the toxicity of high levels of released calcium and the reversibility of the sensor. Figs. 2, D and E, and 4E display the images and graphical representations, respectively, of the effect of 500 μM CMC on a population of A549 cells loaded with the PtCPK probe, which is in agreement with expectations. The specificity of the CMC was confirmed by incubating A549 cells with the RYR1 blocker procaine followed by CMC treatment. In this case, phosphorescence intensity increased by less than 10% (Fig. 4E) after the introduction of CMC. Using probe calibration data (Fig. 1, C and D) and assuming that complete inhibition of respiration was achieved at the end of the experiment, we could convert the fluorescence intensity profile into an oxygen scale (Fig. 4E, blue trace). To confirm that the effect of CMC on intracellular oxygen was a function of Ca2+ release, analysis of intracellular calcium was carried out in parallel, using the calcium-sensitive dye fura-2 and ratiometric imaging. Results in Fig. 4F show that the same concentration of CMC causes a large and immediate release of Ca2+, while comparison of Figs. 4, E and F shows that the changes in intracellular oxygen and calcium are on the same time scale as the change in oxygen uptake rate and that these changes occur slightly subsequent to the release of Ca2+, as expected.
The mechanism of oxygen sensing by phosphorescence quenching by dioxygen is a fully reversible dynamic process by which physical collision of the probe and quencher cause a decrease in the phosphorescence intensity and lifetime. The results presented demonstrate the efficiency of this approach, and the utility of the new family of intracellular oxygen probes based on PtCPK and PdCPK dyes, for sensing intracellular oxygen. Their hydrophilic nature and peripheral functional groups facilitate conjugation of the dyes to polypeptide carriers such as serum albumin, to produce highly photostable and biocompatible macromolecular oxygen probes. Such NIR probes display an optimal sensitivity to oxygen and compatibility with existing imaging equipment. They can be effectively excited with continuous wave or pulsed LEDs and lasers (e.g., 590, 405, or 390 nm) and detected by CCD cameras. The hydrophilic and anionic nature of the probes and relatively small molecular size facilitate passive loading into cells by simple transfection procedures. Macromolecular carrier reduces probe partitioning in the intracellular environment, leaching, cytotoxic and phototoxic effects. These probes can be used in conjunction with standard fluorescence imaging systems, including basic wide field, confocal microscopes and phosphorescence lifetime-based imaging systems to allow simple analysis and imaging of intracellular oxygen over long periods of time. These features are particularly advantageous for biological imaging and cannot be achieved with the other oxygen probes.
The measurement of intracellular oxygen has a number of intrinsic advantages over techniques that measure extracellular oxygen gradients. High-resolution fluorescent imaging facilitates measuring changes in oxygen within individual cells, whereas the microplate-based systems are only applicable to large populations of cells (103 cells/well or greater in a sealed compartment) with end-point parameter readout (14). Furthermore, the kinetics of local changes in intracellular oxygen are much faster than the formation of global oxygen gradients in bulk sample, which eliminates the need for exclusion of ambient oxygen and sealing test samples (1). Thus, subtle and transient metabolic responses, such as those observed following CMC treatment of A549 cells, may be clearly seen by measuring intracellular oxygen but are outside the sensitivity of extracellular sensing methodologies. Finally, as intracellular oxygen is a biomarker of numerous processes from overall cellular function to specific signaling pathways, this results in a technique with wide-ranging applicability, which can complement other important fluorescent probes for intracellular parameters, such as calcium indicators, fluorescent protein tags, probes for reactive oxygen species, and mitochondrial membrane potential. It is amenable to high-content screening applications.
A number of avenues are seen for further increasing the impact of this technique. The long-decay emission of these probes allows time-resolved fluorescent imaging in the microsecond time domain, which can increase the sensitivity and contrast of the sensing system (11) and facilitate multiplexing (40). Phosphorescence lifetime-based oxygen imaging can be applied, allowing more straightforward determination of the absolute oxygen concentrations and eliminating the need of frequent calibrations (46). These advanced techniques, which are relatively easy to implement, will augment what is already an extremely useful tool for cell biologists to study mitochondrial and cellular (dys)function.
Authors declare no competing interests, except for D. B. Papkovsky, who is an equity shareholder in Luxcel Biosciences.
Financial support of this work by the Irish Higher Educational Authority and National Development Plan (2000-2006), the European Commission FP6 project LSHM-CT-2005-018725, and Science Foundation Ireland is gratefully acknowledged.
Authors wish to thank Berenice Riedewald, Dr. Kieran McDermott [Anatomy Department, University College Cork (UCC)] and Dr. Alexander Zhdanov (Biochemistry Department, UCC) for assistance with confocal imaging experiments.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2007 the American Physiological Society