Fluorescence lifetime imaging of metMyoglobin formation due to nitric oxide stress

Abstract

Myoglobin is a protein that is expressed quite unevenly among different cell types. Nevertheless, it has been widely acknowledged that the Fe³⁺ state of myoglobin, metmyoglobin (metMb) has a broad functional role in metabolism, oxidative/nitrative regulation and gene networks. Accordingly, real-time monitoring of oxygenated, deoxygenated and metMb proportions- or, more broadly, of the mechanisms by which metMb is formed, presents a promising line of research. We had previously introduced a Förster resonance energy transfer (FRET) method to read out the deoxygenation/oxygenation states of myoglobin, by creating the targetable oxygen (O₂) sensor Myoglobin-mCherry. In this sensor, changes in myoglobin absorbance features that occur with lost O₂ occupancy -or upon metMb production- control the FRET rate from the fluorescent protein to myoglobin. When O₂ is bound, mCherry fluorescence is only slightly quenched, but if either O₂ is released or met is produced, FRET will increase- and this rate competing with emission reduces both emission yield and lifetime. Nitric oxide (NO) is an important signal (but also a toxic molecule) that can oxidize myoglobin to metMb with absorbance increases in the red visible range. mCherry thus senses both met and deoxygenated myoglobin, which cannot be easily separated at hypoxia. In order to dissect this, we treat cells with NO and investigate how the Myoglobin-mCherry lifetime is affected by generating metMb. More discriminatory power is then achieved when the fluorescent protein EYFP is added to Myoglobin-mCherry, creating a sandwich probe whose lifetime can selectively respond to metMb while being indifferent to O₂ occupancy.

1. INTRODUCTION

We have exploited fluorescence lifetime imaging (FLIM) of a Förster resonance energy transfer (FRET) based probe, Myoglobin-mCherry, to monitor oxygen (O₂) concentration in the intracellular environment of living cells ^{1, 2}. With that sensor, changes in myoglobin absorbance features responsive to O₂ occupancy -or upon metmyoglobin (metMb; heme iron in the ferric state (Fe³⁺)) production- controlled the FRET rate from the fluorescent protein mCherry to the dark acceptor myoglobin. When O₂ is available in the intracellular environment and oxygenated Myoglobin (OxyMb) is dominant, mCherry fluorescence is only slightly quenched, but when the cells are deoxygenated or the met form is produced, FRET rates will increase- and this reduces both emission yield and lifetime ³.

Nitric oxide (NO) is an important cell signaling mediator with many diverse (and often opposing) biological activities that can be employed to oxidize myoglobin to the met state with absorbance increases in the red (>600nm) visible range ^{4, 5}. Our O₂ sensor, Myoglobin-mCherry, unfortunately senses both NO- induced met and deoxygenated myoglobin (DeoxyMb) species, states which cannot be easily separated at hypoxia. Fortunately, cells often have the ability to reduce metMb by enzymatic and/or non-enzymatic reductase mechanisms that apparently require cytochrome b₅ as an electron-transfer mediator ⁶. The mechanism of the enzymatic reduction establishes metMb reductase as a NADH-cytochrome b₅ oxidoreductase. Cytochrome b₅ is reduced at near diffusion-controlled rates by the enzyme with a turnover number of 1000 min^-1. Ferrous cytochrome b₅ then reduces metMb nonenzymatically with an apparent rate constant of 4.9 × 10⁴ M⁻¹ min⁻¹.

In this study, several cell lines: A549 human non-small-cell lung cancer, HeLa human cervical cancer, HCT116 human colorectal carcinoma cells and C2C12 immortalized mouse myoblast cells are treated with S-nitrosoglutathione (SNOG; a molecule that decomposes to provide a NO donor) at both normoxia and hypoxia to investigate how Myoglobin-mCherry lifetime is affected by metMb. We also use our more recently developed sandwich probe, EYFP-Myoglobin-mCherry, to observe (via reduced rather than increased FRET) the formation of metMb as a marker of nitrative stress in the intracellular environment ⁵. The EYFP-Myoglobin-mCherry sensor evaluates FRET from either EYFP or mCherry to myoglobin, and both the met- and O₂- dependent absorbance of that acceptor is thus revealed in their independent fluorescence. More detailed explanations of the sandwich probe and its mechanism of action can be found elsewhere ⁵.

2. MATERIALS AND METHODS

2.1. Cell transfection

A549, HeLa, HCT116 and C2C12 cells were kept in Modified Eagle’s Medium (DMEM, Gibco, Grand Island, NY, US) with 10% fetal bovine serum and 1% penicillin-streptomycin (Mediatech Inc. Manassas, VA, US). The cells were plated in 8- well chambers (Ibidi GmbH, Martinsried, Germany) with a density of 2 × 10⁴ cells/cm². To transfect the cells with Myoglobin-mCherry or EYFP-Myoglobin-mCherry, plasmid DNA diluted in 1000 µL of Opti-MEM® medium (Gibco) was combined with 3 µL of Lipofectamine® 2000 transfection reagent (Invitrogen, Carlsbad, CA, US) diluted in 1000 µL of Opti-MEM. After a 15-min, the DNA-Lipofectamine transfection complex was added to the cells with a final plasmid concentration of 4 ng/µL. After 24 h incubation at 37°C and 5% CO₂, the transfection media was removed and the cells were washed and covered with fresh cell medium.

2.2. Modulation of metMb using a NO donor

A NO donor, S-nitrosoglutathione (SNOG; Tocris Bioscience, Minneapolis, MN, USA) was used to release NO into cell culture media for modulation of metMb. On the day of imaging, SNOG solutions of 3, 50, 100, 250 or 350 µM were prepared by diluting 1 mM stock solution with additional DMEM (without serum). Then, the cell media was replaced with the SNOG solutions followed by imaging.

2.3. Imaging

Two-photon FLIM was performed using an Olympus IX81 confocal laser scanning microscope (Melville, NY, USA) equipped with a tunable Mai Tai BB DeepSee femtosecond laser (Spectra-Physics, Santa Clara, CA, USA) operating at 80 MHz repetition, with <150 fs pulses generated at 780 nm or 950 nm for the excitation of mCherry or EYFP, respectively. The laser light traversed a 690 nm dichroic mirror and was directed to an Olympus UPLANSAPO 60×, 1.2 NA water immersion objective. The epifluorescent emission was reflected and passed through a 675 nm short pass filter to reduce scattered laser light. The EYFP and mCherry signals were further filtered through 520/60 nm and 647/57 nm bandpass filters (Semrock BrightLine®, Rochester, NY, US), respectively. Then, the filtered emissions were focused on a PMC100 cooled detector (Becker & Hickl GmbH, Berlin, Germany) and the electrical pulse output from the detector was directed into a SPC-150 photon counting card (Becker & Hickl). The signals were synchronized with the (photodiode-monitored) pulses from the laser to allow for time-correlated single photon counting (TCSPC). Synchronization with the pixel, line, and frame clock from the raster scanning unit of the microscope was used for image construction in TCSPC mode. The cells were exposed to average laser powers ≤18 mW and imaged for ~30 to 50 s to accumulate an adequate number of photons per pixel and to avoid photocytotoxicity. Image size was set to 256 × 256 (pixels)², and TCSPC histograms were collected with 256 temporal channels in a 12.5 ns time window. To avoid any intermolecular FRET, low signal to noise ratio for the two lifetimes, or any effects of pH and refractive index upon the lifetime values, the infrequent cells showing unusually high or low signals were not imaged ⁷.

2.4. Controlling imposed O₂ during imaging

A miniature incubator was mounted on the microscope stage and connected to a gas mixing system (CO₂−O₂−MI, Bioscience Tools, San Diego, CA) to deliver mixtures of N₂, O₂, and CO₂ inside the chamber. During the imaging, the incubator kept the temperature at 37°C, maintained CO₂ at 5% and set O₂ at 20% (normoxia) or 0.5% (hypoxia); 0.5% is the lowest %O₂ reachable by our system. The cells in petri dishes with open lids and a ∼3 mm layer of medium above them were kept in hypoxia for at least 1 hour to reach a stable O₂ of 0.5%.

2.5. FLIM analyses

FLIM data were analyzed using the SPCImage software (Becker & Hickl). The decay curves at each pixel were fit using a least-squares method to follow a double-exponential decay model (described previously ^{1, 5}). Fitting was performed via iterative reconvolution with a default synthetically generated Instrument Response Function (IRF). The lifetime values were obtained by a optimizing the multi-exponential model in SPCImage for minimum goodness of fit (χ² R). Binning of adjacent pixels was used along with threshold rejection to avoid fitting decays with a peak photon count lower than 1000. Finally, color-mapped lifetime images of Myoglobin-mCherry or EYFP-Myoglobin-mCherry distribution were obtained for each cell.

The mean lifetime of mCherry in the FRET probe of Myoglobin-mCherry and the mean lifetime of EYFP in the sandwich FRET probe of EYFP-Myoglobin-mCherry were calculated by taking τ_mean (amplitude weighted) from each single image and averaging them across multiple cells (n > 20). These values were correlated with the SNOG (NO donor) concentration by the Curve Fitting Toolbox in MATLAB R2020a (The MathWorks Inc., Natick, MA, USA). The relationship between lifetime (in both FRET sensors) and SNOG concentration was found to be approximately sigmoid:

$$f\left(x\right)=\tau min+\frac{\tau max-\tau min}{1+exp\left(ax-b\right)}$$ (1)

where τmin and τmax are the shortest and longest lifetime values of mCherry or EYFP for each dataset obtained at hypoxia and normoxia, respectively. a and b are independent fitting parameters.

2.6. Statistical analyses

At least two independent experiments were performed for each imaging; the mean values are presented with standard deviation. A Mann-Whitney U test was used to evaluate whether the values in the independent groups are significantly different from each other. Analyses were carried out using SPSS 14.0 software (IBM, Chicago, Illinois, USA) and statistical significance was defined at p < 0.05 (95% confidence level).

3. RESULTS AND DISCUSSION

3.1. The effect of metMb on Myoglobin-mCherry lifetime

In the original Myoglobin-mCherry O₂ sensor, the amount of energy transferred from mCherry to myoglobin is controlled by changes in myoglobin spectral features linked with either O₂ occupancy or with metMb production. When O₂ is bound, mCherry fluorescence is only slightly quenched, and when either O₂ is released or met is produced, FRET will increase- and this reduces both yield and the excited population lifetime ^{1, 2}.

The intracellular OxyMb, DeoxyMb, and metMb proportions are regulated by the oxygen concentration (conventional myoglobin binding curve) and reduced-nicotinamide adenine dinucleotide (phosphate)[H] (NAD(P)H)-dependent met/free radical reducing activity ^{8, 9}. In cells like myocytes, where a significant amount of myoglobin naturally exists, rapid oxidation of NO by OxyMb occurs under normoxia with the formation of nitrate (NO₃^-) and metMb ^{10, 11}. In the presence of strong oxidizing agents such as ROS, the met state is exhibited while myoglobin clears NO/ROS for cell protection ^{10, 11}. The interaction of DeoxyMb with nitrite (NO₂ ^-) might also generate NO under hypoxia ^12–14. The metMb state could also form via nitrosylation of DeoxyMb ^{10, 15}. However, the concentration of metMb in most cells is usually low due to the presence of the NAD(P)H-dependent cytochrome b₅ reductase clearance of met/free radicals -- as well as the low expression of myoglobin in most cells ⁸.

In order to investigate the effect of metMb formation on Myoglobin-mCherry lifetime, A549, HeLa, HCT116 and C2C12 cells were transfected with this O₂ probe and treated with SNOG (a calibrated NO donor) at normoxia and hypoxia. Then, the lifetime values were compared to those in control cells (those without NO treatment). Pseudocolor mapping of mCherry fluorescence lifetime (in the FRET sensor Myoglobin-mCherry) in the intracellular environment of the NO-treated and control cells are shown in Fig. 1; red indicates lower lifetime values, whereas blue shows higher values.

Figure 1.

Pseudocolor mapping of mCherry fluorescence lifetime in the FRET sensor (Myoglobin-mCherry) in the intracellular environment of (A) A549, (B) C2C12, (C) HeLa and (D) HCT116 cells treated with 350 μM SNOG (a nitric oxide (NO) donor) as compared to the control cells. The cells were at imposed O₂ of 20 and 0%. In the color bars, red indicates lower lifetime values, whereas blue indicates higher values.

Based on these results, NO application resulted in a decreased mCherry lifetime (in both normoxia and hypoxia) due to the conversion of both OxyMb and DeoxyMb to metMb. As shown in Fig. 2, SNOG produced a concentration-dependent decrease of mCherry lifetime, showing approximate sigmoid behavior (see Eq. (1)); the fitting parameters are shown in Table 1. Under moderate NO challenges (< 3 μM), FLIM didn’t expose any metMb formation in the cells. This might be due to the background effects of metMb reductase (e.g. NAD(P)H-dependent cytochrome b₅ reductase) that protect the cells with clearance of met/free radicals ¹⁶. At higher NO doses (those that appear to saturate and likely were sufficient to completely convert myoglobin to metMb), any putative clearance mechanism was clearly overwhelmed. The maximum metMb effect was observed following 350 μM SNOG treatment at imposed O₂ of 20%, when the probe lifetime decreased by 19% (from 1.22 to 0.98 ns) in A549 cells, 16% (from 1.19 to 1.00 ns) in C2C12 cells, 22% (from 1.23 to 0.96 ns) in HCT116 cells and 20% (from 1.23 to 0.98 ns) in HeLa cells. At hypoxia, the mean lifetime values decreased by 11% in A549 and 10% in C2C12, HCT116 and HeLa cells treated with 350 μM SNOG. In all but the SNOG concentration of 3 μM, Mann-Whitney U tests showed a significant decrease of the lifetime values upon treatment at each imposed O₂% (the maximum p-value = 0.02).

Figure 2.

Changes of Myoglobin-mCherry fluorescence lifetime versus SNOG concentration in (A) A549, (B) C2C12, (C) HeLa and (D) HCT116 cells at imposed O₂ of 20 and 0%. Error bars show the standard deviation of at least 20 cells.

Table 1.

Parameters of the sigmoid fits. Fitting parameters a and b were obtained from fitting the data presented in Fig. 2 to Eq. (1). τmin and τmax are the shortest and longest average lifetime for Myo-mCherry in each cell type measured at hypoxia (O₂ = 0.5%) and normoxia (O₂ = 20%), respectively. Each parameter is shown with its standard deviation.

Cells	Imposed O2 = 20%					Imposed O2 = 0.5%
	τmin (ns)	τmax (ns)	a	b	R2	τmin (ns)	τmax (ns)	a	b	R2
A549	0.98 ± 0.01	1.22 ± 0.03	0.05 ± 0.03	2.89 ± 1.61	0.99	0.87 ± 0.04	0.98 ± 0.05	0.05 ± 0.04	3.76 ± 3.35	0.96
C2C12	1.00 ± 0.03	1.19 ± 0.03	0.03 ± 0.02	2.08 ± 1.33	0.96	0.90 ± 0.04	1.02 ± 0.05	0.03 ± 0.01	1.50 ± 0.53	0.99
HeLa	0.98 ± 0.04	1.23 ± 0.05	0.03 ± 0.02	2.24 ± 1.62	0.96	0.91 ± 0.06	1.00 ± 0.04	0.06 ± 0.03	2.30 ± 1.88	0.99
HCT116	0.96 ± 0.06	1.23 ± 0.03	0.04 ± 0.02	2.31 ± 1.59	0.97	0.89 ± 0.03	0.99 ± 0.02	0.013 ± 0.01	2.10 ± 1.45	0.93

Note that the apparent met state reduces lifetime another ~10% below the hypoxic floor in all cases.

3.2. Lifetime imaging of EYFP in EYFP-Myoglobin-mCherry probe for metMb detection

As mentioned previously, metMb can form due to a rapid oxidation of NO by OxyMb or by nitrosylation of DeoxyMb at hypoxia, yielding MbNO as an intermediate. The presence of the enzyme metMb/free radical reductase (mainly NAD(P)H-dependent coenzyme cytochrome b₅) usually means the metMb level in living cells is likely to be small ^{6, 8, 15}. The changes in mCherry lifetime with even saturating NO are modest and entangled with oxygenation. To monitor metMb formation, we therefore expressed EYFP-Myoglobin-mCherry in the cytoplasm of HCT116 and C2C12 cells. Then, we performed lifetime imaging on cells expressing the probe at imposed O₂ values of 20% and 0.5% and used the lifetime of EYFP as a more sensitive and direct reporter of metMb. Figs. 3A and B show the pseudocolor mapping of EYFP fluorescence lifetime at O₂ of 20% (in the sandwich FRET sensor EYFP-Myoglobin-mCherry) in the intracellular environment of the C2C12 and HCT116 cells treated with 250 μM SNOG as compared to the controls (cells at the same imposed O₂% and no SNOG treatment); red indicates lower lifetime values, whereas blue shows higher values.

Figure 3.

Pseudocolor mapping of EYFP (in EYFP-Myoglobin-mCherry) fluorescence lifetime in A) HCT116 and B) C2C12 cells treated with 250 μM SNOG at O₂ of 20% as compared to the control cells (no SNOG treatment). Changes of the average lifetime of EYFP in the FRET sensor vs. SNOG concentration in C) HCT116 and D) C2C12 cells at normoxia (O₂= 20%) and hypoxia (O₂= 0.5%). Error bars show the standard deviation of at least 20 cells.

As shown in Figures 3A and B, NO application resulted in an increased EYFP lifetime. This is due to the conversion of either (higher FRET equivalent states) OxyMb and DeoxyMb to (lower FRET state, higher lifetime) metMb ⁵. When the full image averaged values of EYFP lifetime (in the sandwich FRET probe), taken for multiple cells, were correlated to the SNOG concentration, the NO donor produced a concentration-dependent increase of the lifetime. As shown in Figures 3C and D, the relationship between lifetime and SNOG concentration was found to be approximately sigmoid; the fitting parameters are shown in Table 2. The maximum lifetime increase was observed following 250 μM SNOG treatment, when the probe lifetime increased by ~16% in HCT116 cells and by ~18–19% in C2C12 cells at the imposed O₂ of 20% and 0.5%; cells incubated without SNOG were used as a control. Note the indifference to oxygenation.

Table 2.

Parameters of the sigmoid fits. Fitting parameters a and b were obtained from fitting the data presented in Fig. 3 to Eq. (1). τmin and τmax are the shortest and longest average lifetime for EYFP (in the sandwich probe EYFP-Myoglobin-mCherry) in each cell type measured at hypoxia (O₂ = 0.5%) and normoxia (O₂ = 20%), respectively. Each parameter is shown with its standard deviation.

Cells	Imposed O2 = 20%					Imposed O2 = 0.5%
	τmin (ns)	τmax (ns)	−a	−b	R2	τmin (ns)	τmax (ns)	−a	−b	R2
C2C12	1.68 ± 0.07	2.06 ± 0.22	0.04 ± 0.02	2.22 ± 1.35	0.98	1.72 ± 0.08	2.12 ± 0.21	0.03 ± 0.02	2.70 ± 1.29	0.99
HCT116	1.70 ± 0.10	2.03 ± 0.20	0.04 ± 0.03	1.73 ± 1.21	0.94	1.69 ± 0.09	2.02 ± 0.19	0.04 ± 0.03	1.55 ± 1.13	0.91

4. CONCLUSIONS

In the last few years, the importance of NO levels has emerged as a key determinant of its biological function ⁴. It is clear now that NO is not only endogenously generated but that it is an essential part of the immune response and many other physiological signal transduction pathways ⁴. Although NO chemical reactivity can lead to toxicity, the biological properties of these same reactive species can be beneficial and explain their apparently dichotomous actions. Processes ranging from apoptosis, senescence, angiogenesis, inflammation, immunological responses, vascular tone control, cardiac contractility and relaxation, to neuronal death all show distinct and at times, contradictory behavior in response to NO.

We had previously introduced Myoglobin-mCherry for the detection of O₂ in the intracellular environment ^{1, 2}. It is well-documented that myoglobin scavenges free radicals through an intermediate state of metMb ¹⁷. Therefore, the presence of NO in living cells can oxidize myoglobin to the met state with absorbance increases in the red visible range similar to DeoxyMb ^{4, 5}. The original O₂ probe thus senses both met and DeoxyMb, an entanglement not easily separated at hypoxia. In this study, A549, HeLa, HCT116 and C2C12 cells transfected with Myoglobin-mCherry were treated with SNOG (a NO donor) at normoxia and hypoxia to investigate just how much Myoglobin-mCherry lifetime is affected by NO-induced metMb. The results showed that met can decrease the mCherry lifetime by maximum 22% in normoxia and 11% in hypoxia. In addition, we used fluorescence lifetime imaging of EYFP in the FRET probe EYFP-Myoglobin-mCherry to quantify the intracellular accumulation of metMb as an indicator of NO. This is an increased- rather than decreased- lifetime indicator. We correlated metMb generation with NO levels in our experimental cell lines subjected to hypoxia OR normoxia. The increased lifetime of the EYFP portion of the probe showed metMb was produced at equivalent levels from either OxyMb or DeoxyMb (using treatment with SNOG).

As an aside, we have previously shown that this probe can also serve to report O₂ concentration via the intensity ratio of EYFP/mCherry ⁵. This employs the indifference of EYFP to O₂ to create an internal reference. The latter approach is of use for those lacking FLIM access, but FLIM has the feature of being nearly independent of concentration. This feature also makes the FLIM analysis of both O₂ and met by sandwich probe lifetime components more robust to varying cellular levels of probe.

In all, the combined probe (or, if preferred, its internal paired parts) provide us the opportunity to glean valuable redox and free radical information within living cells, just as the relevant events occur.