Treatment with iron(II) and ascorbate is essential for geNOps functionality
Standard cell culture settings under room air conditions (18 kPa O2) are unphysiological for most cell types; thus, we examined the effects of physiological O2 levels (5 kPa) in HEK293T and EA.hy926 cell lines stably expressing the orange variant of geNOps. Both cell lines adapted for five days to 18 kPa O2 levels showed robust geNOps expression, while cells adapted to 5 kPa O2 displayed significantly lower basal fluorescence intensity (Fig. 1a-c & e-g). Nevertheless, NOC-7, a potent NO donor23, evoked a robust geNOps signal in both cell lines adapted to physiological normoxia. In contrast, in cells adapted to standard culture under hyperoxia (18 kPa O2), geNOps displayed marginal changes in response to the potent NO donor. (Fig. 1d,h). Overall, these observations establish a critical role for physiological pericellular O2 in NO bioimaging, yet it remains unclear whether ambient O2 concentrations affect NO bioavailability39 or cellular iron(II) uptake and thereby geNOps activity.
Short-term (24 h) or long-term (21 days) culture of HEK293T cells under 18 kPa O2 in different commercially available media such as Dulbecco's Minimal Essential Media (DMEM), Advanced DMEM, F12, and F12K containing ferrous iron, ferric iron, or ascorbate, respectively, only led to marginal activation of geNOps (Supplementary Fig. 1). However, geNOps displayed full functionality upon treatment of cells with 1 mM FeSO4 and 1 mM ascorbate for 20 min prior to imaging experiments (Supplementary Fig. 1). Long-term (14 days) adaptation of cells to sub-toxic concentrations of ascorbate (ranging from 12–96 µM) or FeSO4 (ranging from 7–56 µM) in DMEM also failed to activate geNOps (Supplementary Fig. 2). Overall, these data demonstrate that geNOps functionality requires cell treatment with media containing freshly prepared iron(II) and ascorbate before an imaging experiment.
Imaging intracellular distribution of ferric and ferrous iron
Our results so far demonstrate the requirement for iron(II) supplementation to cells to achieve full functionality of geNOps, yet the spatial distribution and quantification of the labile iron pool remain unclear. We initially performed high-resolution confocal imaging experiments using the fluorescent indicator FeRhoNox-1, a specific probe to detect cytosolic labile iron29. Collecting multiple focal planes in the Z-direction in HEK293T cells adapted to 18 kPa O2 cells pretreated with FeSO4 and ascorbate revealed that a significant amount of the probe was also detectable on the surface of cells (Supplementary Fig. 3a). To further confirm whether iron(II) accumulates on the cell surface, we imaged cells stained with Perls and 3’-diaminobenzidine (DAB). Cells treated with ascorbate only were comparable to the control group. However, cells treated with FeSO4 in the absence of ascorbate displayed a remarkable accumulation of extracellular iron particles, indicating that most supplemented iron precipitates without entering the cells and aggregates on the surface of the cell membrane (Supplementary Fig. 3b).
In contrast, FeSO4 and ascorbate co-treatment led to negligible iron accumulation on the cell surface, whereas intracellular aggregates were detectable (Supplementary Fig. 3b). These results confirm that a strong reducing agent is essential to keep iron in solution reduced under hyperoxic conditions, allowing cells to internalize soluble iron from the culture media. FeRhoNox-1 imaging in cells adapted to either 5 or 18 kPa O2 and treated with FeSO4 only or in combination with ascorbate displayed higher levels of intracellular labile iron under physiological O2 conditions (Supplementary Fig. 3c,d), demonstrating the critical role of the redox state of iron in cellular uptake.
These results were further validated by electron microscopy (EM) (Fig. 2a). Correlative light and electron microscopy (CLEM) experiments confirmed the accumulation of extra- and intracellular iron. FeRhoNox-1 stained cells were less suitable for this protocol, as the chemical dye targets undefinable intracellular regions. Thus, we exploited Perls stained cells, intensified with diaminobenzidine (DAB), and then correlated high-resolution bright-field and EM images of the same cells (Fig. 2b,c). Our CLEM approach confirmed that cells treated with FeSO4 only displayed visible dark punctae, indicating the precipitation and accumulation of iron on the cell surface (Fig. 2b, left panel). These data correlate with EM images, confirming that ascorbate is required to internalize extracellular iron (Fig. 2b, right panel). Notably, there was an apparent accumulation around the endoplasmic reticulum (ER) without signs of ER stress (Fig. 2c, right panel).
Optimization of the iron-supplementation procedure
We next sought to optimize acute iron(II) supplementation by lowering the reducing agent and iron compound concentrations. Employing Taguchi guidelines40, we optimized all three parameters: FeSO4 and ascorbate concentrations and the incubation time (Supplementary Table 1). This approach established that a minimum concentration of iron(II) compound, reducing agent, and incubation time of 300 µM, 500 µM, and 15 min, respectively, is required to activate geNOps biosensors (Supplementary Fig. 4). As shown in Fig. 3a, the optimized iron supplementation protocol resulted in FeRhoNox-1 signals similar to those achieved by our initial iron loading protocol (1 mM FeSO4 and 1 mM ascorbate for 20 min). These results show that lower levels of FeSO4 and ascorbate provision under room air conditions yield similar levels of intracellular labile iron documented by geNOps functionality and FeRhoNox-1 (Fig. 3b).
Importantly, scanning electron microscopy with energy-dispersive X-ray spectroscopy (SEM/EDX) analysis for elemental identification and quantitative compositional information confirmed that optimized iron(II) concentrations did not lead to accumulation of intracellular iron(II) in undefinable structures (Fig. 3c, and Supplementary Fig. 5). Overall, our findings suggest that lower FeSO4 and ascorbate concentrations are sufficient and necessary to activate metalloprotein geNOps under standard cell culture conditions (18 kPa O2).
Cell viability and mitochondrial reactive oxygen species (ROS) generation were examined following acute iron(II) supplementation in HEK293T cells under 18 kPa O2. Cell viability and mitochondrial ROS levels remained unaffected by iron(II) supplementation (Fig. 4). However, treatment of HEK293T and EA.hy926 cells with 1 mM FeSO4 and 1 mM ascorbate led to robust increases in mitochondrial H2O2 levels measured using the ultrasensitive H2O2 biosensor HyPer730,41 (Fig. 4c,d). Significantly, optimized iron(II) concentrations did not increase mitochondrial H2O2 levels 24 h after iron(II) supplementation (Fig. 4c,d). Basal intracellular GSH levels were lower in EA.hy926, but not in HEK293T cells adapted to 5 kPa than cells adapted to 18 kPa O2, consistent with diminished oxidative stress in cells cultured under physiological normoxia36 (Supplementary Fig. 6a). Moreover, basal mitochondrial H2O2 levels in EA.hy926 cells were comparable under 18 kPa and 5 kPa, suggesting that mitochondrial ROS levels were affected negligibly by physiological normoxia (Supplementary Fig. 6c).
The role of ambient oxygen levels on geNOps functionality in endothelial cells
We next sought to examine the effects of optimized iron(II) supplementation in EA.hy926 endothelial cells capable of generating intracellular NO in response to the GPCR agonist adenosine triphosphate (ATP), which robustly triggers intracellular calcium mobilization to activate eNOS42 (Fig. 5a). Treatment of endothelial cells adapted to 18 kPa O2 with ATP caused a robust intracellular geNOps signal (inhibitable by nitro-L-arginine methyl ester) in cells pretreated with the optimized iron(II) and ascorbate concentration. In contrast, the geNOps signal in non-iron treated cells was negligible (Fig. 5b). Our results indicate that treating endothelial cells with iron(II) in combination with ascorbate is necessary for activating geNOps under standard hyperoxic culture conditions. We hypothesized that the iron(II) concentration could be lowered further by adapting endothelial cells to physiological normoxia (5 kPa O2) to mimic O2 levels in vivo22,38. EA.hy926 cells stably expressing O-geNOp-NES were adapted to either 18 or 5 kPa O2 for at least five days. We initially treated cells adapted to 5 kPa O2 with even lower iron(II) (150 µM FeSO4 and 300 µM ascorbate). ATP stimulated NO production in these cells induced a robust geNOps signal that was diminished upon subsequent addition of the NO synthase inhibitor L-NAME (Fig. 5c). These results confirm our hypothesis that culturing endothelial cells under physiological normoxia requires treatment with significantly lower iron(II) and ascorbate concentrations. However, when EA.hy926 cells were adapted to standard cell culture hyperoxia (18 kPa O2) and pretreated with the same concentrations of iron(II) and ascorbate (150 µM FeSO4 and 300 µM ascorbate), the geNOps signal in response to ATP was significantly decreased (Fig. 5d).