2.1.

Reagent

Human recombinant epithelial growth factor (EGF) was purchased from Peprotech (Rocky Hill, NJ). 5(6)-Carboxy-2’,7’-dichlorofluorescein diacetate (carboxy-H2DCFDA), 4-Amino-5-(4-methylphenyl)-7-(t-butyl)pyrazolo[3,4-d]-pyrimidine (PP1), and sodium orthovanadate were purchased from Sigma-Aldrich (St. Louis, Missouri). Dulbecco’s modified Eagle’s medium (DMEM), Lipofectamine 2000 and ${\mathrm{CO}}_2$-independent medium were purchased from Invitrogen (Carlsbad, CA). Restriction enzymes were purchased from New England Biolabs (Ipswich, Massachusetts). Vanadate was prepared from sodium orthovanadate according to previously reported method.²⁸

2.2.

Cell Culture and Transfection

HeLa cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% FBS, $100\text{-}\mathrm{U}/\mathrm{ml}$ penicillin, and $100\text{-}\mathrm{mg}/\mathrm{ml}$ streptomycin at 37°C in 5% ${\mathrm{CO}}_2$.

Plasmids were transfected into HeLa cells by Lipofectamine 2000 according to manufacturer’s instructions (Invitrogen).

2.3.

Plasmid Construction

Ypet was generated by introducing S208F, V224L, H231E, and D234N into Venus. The architecture of our Src biosensor is identical to the previously reported Src biosensor.^26,27 Rac1 and Prx1 genes were amplified by PCR using cDNA prepared from HeLa cells. To generate the far-red fluorescent protein mLumin²⁹ tagged Rac1 (mLumin-Rac1), the mLumin fragment was amplified by PCR using a sense primer bearing a BamHI site and a reverse primer bearing an XhoI site. The Rac1 fragment was amplified by PCR using a sense primer bearing an XhoI site and a reverse primer bearing an EcoRI site. Finally, mLumin and Rac1 were ligated into pCDNA3.1 between BamHI and EcoRI sites. The PCR megaprimer method was used to generate mLumin-Rac1-N17.³⁰ For PrxI-Y194F, because Y194F is near the C-terminus, PrxI-Y194F was directly generated by normal PCR with an anti-sense primer bearing Y194F. To unambiguously identify cells expressing PrxI-Y194F, we used an internal ribosome entry site (IRES) sequence, which allowed expression of two genes from the same bi-cistronic mRNA transcript. The architecture of this sequence is as follows: BamHI-PrxI-(Y194F)-EcoRI-IRES-SacI-mLumin-NLS-XbaI. NLS represents nuclear localized sequence. First, IRES was amplified by PCR from a template pIRES vector (Invitrogen) with a sense primer bearing BamHI and EcoRI sites and an anti-sense primer bearing a SacI site. mLumin-NLS was obtained by PCR-addition of an NLS sequence (amino acid is PKKKRKVEDA) at the C-terminus of mLumin. This fragment was generated with SacI and XbaI sites at N- and C- terminus, respectively. Then, IRES and mLumin-NLS were fused into pCDNA 3.1 between the BamHI and XbaI sites. Finally, PrxI-Y194F was inserted before the IRES using BamHI and EcoRI restriction enzyme sites.

The mVenus gene is from Dr. Atsushi Miyawaki at Brain Science Institute, RIKEN, Japan.

2.4.

Cell Imaging

Living cell imaging was performed on an Olympus confocal laser-scanning microscope FV1000 (Olympus, Japan) equipped with a $60\times 1.35$ oil immersion objective. The FV1000 allows for dual-channel imaging. Thirty-six hours after transfection of exogenous plasmid, HeLa cells were starved for additional 6 to 12 h in serum-free DMEM before EGF ($100\mathrm{ng}/\mathrm{ml}$) stimulation. Immediately before imaging, DMEM was replaced with ${\mathrm{CO}}_2$-independent medium (Invitrogen). For the Src sensor, the following setting was used: 458-nm excitation, CFP emission: 470–500 nm, YFP emission: 520–560 nm. For mLumin, the 543-nm laser was used and emission wavelengths between 600 and 700 nm were collected. For the Grx1-roGFP2 biosensor, the following setting was used: 405- and 488-nm excitation, emission: 500–600 nm. The “time-control” function in the FV1000 software was used to quickly switch laser lines between 405 and 488 nm.

2.5.

Image Processing

ImageJ software ( http://rsbweb.nih.gov/ij/) was used to process confocal images and to generate the CFP/YFP emission ratio images of Src biosensor and the $405/488\mathrm{nm}$ excitation ratio images of Grx1-roGFP2 biosensor. The key procedures are identical with previous reports.^31,32 For Src biosensor, raw CFP and YFP images were firstly background subtracted at each time point. Gaussian blur function (sigma value 1) was used for image smoothing. For each time point, because the YFP image had the largest signal-to-noise ratio, the YFP image was used to generate a binary mask image with a value of zero outside the cell and a value of one inside the cell. The CFP and YFP images were then respectively multiplied by this mask image. Finally, the updated CFP image was divided by the updated YFP image to yield the CFP/YFP emission ratio image of Src biosensor. For better visualization of ratiometric images, brightness/contrast was adjusted and a linear pseudo-color lookup table (color bar) was applied. For generating the kinetics of the normalized CFP/YFP emission ratio of Src biosensor, a region of interest in the cytoplasm of ratio images was selected to extract ratio values. And, the ratio values were normalized to the average ratio value before EGF stimulation by Excel (Microsoft).

For generating the $405/488\text{-}\mathrm{nm}$ excitation ratio images of Grx1-roGFP2, the image excited at 488 nm at each time point was used to generate a binary mask image. Then the following processing steps were identical with above procedures for generating the CFP/YFP ratio images of Src biosensor. For generating the kinetics of the normalized $405/488\text{-}\mathrm{nm}$ excitation ratio of Grx1-roGFP2, a region of interest in the cytoplasm of ratio images was selected to yield ratio values. Then the ratio values were normalized to the average ratio value before EGF or ${\mathrm{H}}_2{\mathrm{O}}_2$ stimulation by Excel (Microsoft).

The curves of the normalized CFP/YFP emission ratio of Src biosensor were plotted using OriginPro 8 software (Northampton, MA). The mathematical parameters, PAS and DS, were obtained from the curves of the normalized CFP/YFP emission ratio of Src biosensor. And the ISS was calculated using integration function in OriginPro 8 software.

2.6.

Statistical Analysis

The student’s t-test function in GraphPad Prism software (San Diego, CA) was used to evaluate statistical differences between experimental groups. P values below 0.05 were considered as significantly different.

3.1.

Three Parameters to Quantitatively Characterize the Kinetics of Src Signaling upon EGF Stimulation

The CFP/YFP emission ratio of the Src biosensor provides an assay for measuring the magnitude and kinetics of Src signaling in living cells. To obtain the kinetics of Src signaling under EGF stimulation, HeLa cells were transfected with a plasmid encoding the Src biosensor. The concentration of EGF used throughout this study was $100\mathrm{ng}/\mathrm{ml}$. As shown in Fig. 1(a), activated Src mainly exists in the cytoplasm throughout the imaging period. Thus, a region of interest in the cytoplasm was selected for analyzing the kinetics of Src signaling. The kinetic profile of Src signaling upon EGF stimulation displayed biphasic characteristics, consisting of a rapidly increasing phase and a slowly declining phase [black curve in Fig. 1(b)]. Strikingly, the kinetic profile of EGF-induced Src signaling resembles kinetic profile of EGF-induced ERK-mediated phosphorylation signaling as reported before.^8–11 This similarity suggests that growth factor-induced and kinase-mediated phosphorylation signaling possesses analogous dynamic mechanisms. Therefore, we reasoned that the models concerning ERK signaling might be used to characterize EGF-induced Src signaling.^8,9,11 Based on previous models which described ERK phosphorylation signaling cascades,^8,9 we introduced the following three simplified mathematical parameters to more precisely define the kinetics of EGF-induced Src signaling: the peak amplitude of signaling (PAS), the duration of signaling (DS), and integral signaling strength (ISS). These parameters are calculated from the normalized CFP/YFP emission ratio curves (refer to Sec. 2.5 Image Processing), in which PAS is the maximum value of the normalized CFP/YFP emission ratio change within 5 min, DS is the full width of the curve at half of the maximum, and ISS is the area under the curve within 60 min.

Fig. 1

Visualization of kinetics of epidermal growth factor (EGF)-induced Src signaling by Src biosensor; (a) Cyan/yellow fluorescent protein (CFP/YFP) emission ratio images of the Src biosensor in response to $100\text{-}\mathrm{ng}/\mathrm{mL}$ EGF; (b) Kinetics of normalized CFP/YFP emission ratios of the Src biosensor in response to EGF in a control HeLa cell as demonstrated in (a), or in a HeLa cell pre-treated with 200-nM PP1 (Src specific inhibitor) for 1 h, or in a HeLa cell pre-treated with 500-μM vanadate (PTP specific inhibitor, denoted as Van) for 30 min (c, d, and e). Bar graphs represent the $\text{mean}\pm \mathrm{SEM}$ of peak amplitude, (c) duration, (d) and integral signal strength, (e) of Src signaling from at least three independent experiments with the same conditions, as in (b). “n” denotes the sample number in each group. “*”, “**” and “***” indicate $p<0.05$, $p<0.01$ and $p<0.001$ respectively. “ns” indicates insignificant differences between groups ($p>0.05$). Scale bars, 10 μm.

To correlate the three kinetic parameters with the activity of Src kinase and PTP in living cells upon EGF treatment, PP1 and vanadate, which are specific inhibitors of Src kinase and PTPs, respectively, were used. As demonstrated in Fig. 1(a) and 1(b), exposure of HeLa cells to EGF led to rapid phosphorylation of the Src biosensor, and this phosphorylation was maintained at relatively high levels throughout the imaging session, demonstrating sustained activation of Src signaling [Fig. 1(c)–1(e); PAS: $1.14\pm 0.06$; DS: $57.9\pm 4.2\min$; ISS: $53.1\pm 3.8$, $n=7$]. The initially increasing phase typically reached a maximum value within 5 min, consistent with the experimentally tested model asserting that kinases are rapidly activated whereas phosphatases are transiently inhibited upon growth factor treatment, permitting rapidly signal amplification.^8,9 Hence, we conceived that Src activity is dominant to PTP activity in this phase. And PAS was used to measure the relative level of activated Src. We used this experiment as control group. Figure 1(b) demonstrates that inhibition of Src activity by PP1 resulted in marked decrease of PAS [Fig. 1(c); PAS: $0.74\pm 0.05$, $n=8$], whereas DS was not significantly influenced [Fig. 1(d); DS: $56.0\pm 3.3\min$, $n=8$]. Correspondingly, ISS was reduced due to decreases in the CFP/YFP ratio amplitude [Fig. 1(e); ISS: $35.0\pm 3.4$, $n=8$]. In contrast to effect of PP1, suppression of PTPs activity by vanadate led to a broader peak and substantially longer duration of CFP/YFP emission ratio [Fig. 1(b) and 1(d); DS: $110.8\pm 8.4\min$, $n=6$]. Additionally, there was a modest increase in PAS [Fig. 1(c); PAS: $1.33\pm 0.06$, $n=6$]. ISS was correspondingly increased with vanadate treatment [Fig. 1(e); ISS: $78.7\pm 2.8$, $n=6$].

Actually, vanadate alone is able to gradually induce weak activation of Src signaling (weakly activated pattern) [Fig. 2(b)], but with quite different kinetics as compared to that of EGF-induced Src signaling (strongly activated pattern), suggesting those two stimulants utilize distinct mechanisms to initiate Src signaling. In some vanadate-pretreated HeLa cells upon EGF stimulation, those two types of kinetics co-exist and can be easily distinguished. As shown in Fig. 2(a), there is a slowly increasing phase after the rapidly increasing phase. Under this experimental condition, PAS values refer to the maximum value of the CFP/YFP emission ratio change within first 5 min following EGF treatment.

Fig. 2

Kinetics of normalized cyan/yellow fluorescent protein (CFP/YFP) emission ratios of the Src biosensor in response to EGF in a vanadate-pretreated HeLa cell; (a) Kinetics of normalized CFP/YFP emission ratios of the Src biosensor in response to epidermal growth factor (EGF) in HeLa cells pretreated with 500-μM vanadate for 30 min; (b) Kinetics of normalized CFP/YFP emission ratios of the Src biosensor in response to 500-μM vanadate in HeLa cells.

Taken together, when EGF induces Src signaling, Src kinase controls PAS [Fig. 1(c)], whereas PTPs have more effect on DS than PAS [Fig. 1(c) and 1(d)]. And, ISS is controlled by both Src kinase and PTPs [Fig. 1(e)]. Since the CFP/YFP emission ratio of Src biosensor is only determined by Src kinase and PTPs, the three kinetic parameters can be used to quantitatively assess how cellular regulators exert their effect on kinetics of Src signaling.

3.2.

${\mathrm{H}}_2{\mathrm{O}}_2$ Tunes the Duration of EGF-Induced Src Signaling Through PTP Inhibition

To investigate how ${\mathrm{H}}_2{\mathrm{O}}_2$ modulates the kinetics of Src signaling upon EGF stimulation, HeLa cells were transfected with two plasmids encoding two ${\mathrm{H}}_2{\mathrm{O}}_2$ regulating proteins: Rac1-N17 (dominant negative form of Rac1) and PrxI-Y194F (a variant of the ${\mathrm{H}}_2{\mathrm{O}}_2$ degrading enzyme PrxI).

A dominant negative form of Rac1 (Rac1-N17) is reported to prevent production of ${\mathrm{H}}_2{\mathrm{O}}_2$ through its inhibition of NADPH oxidase upon growth factor stimulation.^33–36 Actually, measurement of ${\mathrm{H}}_2{\mathrm{O}}_2$ by DCF dye showed that Rac1-N17 dramatically decreased intracellular EGF-induced ${\mathrm{H}}_2{\mathrm{O}}_2$ level (Fig. 3). To elucidate the role of ${\mathrm{H}}_2{\mathrm{O}}_2$ in regulating Src signaling, HeLa cells were co-transfected with two plasmids, encoding the Src biosensor and the far-red fluorescent protein mLumin²⁹ tagged Rac1-N17. We found that overexpression of Rac1-N17 remarkably prevented propagation of EGF-induced Src signaling [Fig. 4(a), 4(c), and 4(i); ISS: $26.4\pm 2.1$, $n=9$] and had more effects on DS than PAS [Fig. 4(g) and 4(h)]. In these cells, PAS was reduced to $0.92\pm 0.05$ ($n=9$; 19% lower than control) and DS was shortened to $23.2\pm 2.2\min$ ($n=9$; 60% shorter than control). This result suggested that ${\mathrm{H}}_2{\mathrm{O}}_2$ production is critical to propagation of EGF-induced Src signaling.

Fig. 3

$\text{Carboxy}\text{-}{\mathrm{H}}_2\mathrm{DCFDA}$, demonstrated Rac1-N17 prevented production of ${\mathrm{H}}_2{\mathrm{O}}_2$ in HeLa cells in response to epidermal growth factor (EGF). (Scale bars, 10 μm)

Fig. 4

Overexpression of either mLumin-Rac1-N17 or PrxI-Y194F attenuated epidermal growth factor (EGF)-induced Src signaling in HeLa cells; (a, b) cyan/yellow fluorescent protein (CFP/YFP) emission ratio images of the Src biosensor in a HeLa cell expressing either mLumin-Rac1-N17 or PrxI-Y194F upon EGF treatment. PrxI-Y194F and nuclear-targeted mLumin (mLumin-NLS) were expressed from the same mRNA by using an internal ribosome entry site (IRES) sequences that allowed expression of two genes from the same bi-cistronic mRNA transcript; (c, d) Kinetics of normalized CFP/YFP emission ratios in (a) and (b). (e, f) Kinetics of normalized CFP/YFP emission ratios of the Src biosensor in response to EGF in either Rac1-N17 (e) or PrxI-Y194F-expressing (f) HeLa cells pretreated with 500-μM vanadate (denoted as Van) for 30 min (g, h, i, j, k and l). Bar graphs represent the $\text{mean}\pm \text{standard}$ error (SEM) of peak amplitude (g, j), duration (h, k) and integral signal strength (i, l) of Src signaling from at least three independent experiments with the same conditions as in c, d, e and f. “n” denotes the sample number in each groups. “*”, “**” and “***” indicate $p<0.05$, $p<0.01$ and, $p<0.001$, respectively. Scale bars, 10 μm.

To provide additional evidence, we decreased the amount of endogenous ${\mathrm{H}}_2{\mathrm{O}}_2$ by expressing a variant of the ${\mathrm{H}}_2{\mathrm{O}}_2$ degrading enzyme PrxI (Prx1-Y194F), which is an abundant cytosolic enzyme.³⁷ Previous evidence demonstrated that PrxI is phosphorylated by Src kinase upon EGF stimulation, leading to its inactivation. This inactivation allows localized accumulation of ${\mathrm{H}}_2{\mathrm{O}}_2$.³⁸ The PrxI-Y194F protein was demonstrated to be insulated from phosphorylation-mediated inactivation while retaining catalytic activity.³⁸ Time-lapse imaging of the Src biosensor in PrxI-Y194F-expressing HeLa cells demonstrated that sustained phosphorylation of Src biosensor was attenuated [Fig. 4(b), 4(d), and 4(i); ISS: $29.2\pm 1.3$, $n=11$], strongly implying that ${\mathrm{H}}_2{\mathrm{O}}_2$ was implicated in maintaining Src signaling. Like in Rac1-N17 transfected cells, the amplitude and duration of Src signaling were both compromised in PrxI-Y194F-expressing HeLa cells [Fig. 4(g) and 4(h)]. PAS was reduced to $0.82\pm 0.03$ ($n=11$; 28% lower than control) and DS was shortened to $36.0\pm 1.9\min$ ($n=11$; 38% shorter than control). This result again suggested that ${\mathrm{H}}_2{\mathrm{O}}_2$ plays a crucial role in maintaining propagation of EGF-induced Src signaling.

Because duration of Src signaling is controlled by PTPs’ activity [Fig. 1(d)], we first examined whether the shortened duration of Src signaling resulted from the relatively high PTPs’ activity in Rac1-N17 or PrxI-Y194F- expressing cells as compared with the control group. Previous reports demonstrated that growth factor-induced ${\mathrm{H}}_2{\mathrm{O}}_2$ can inhibit activity of PTPs.^6,7 It is highly likely that overexpressing PrxI-Y194F or Rac1-N17 reduced the amount of ${\mathrm{H}}_2{\mathrm{O}}_2$ available to inhibit PTPs, leading to the shortened duration of Src signaling. To test this hypothesis, the activity of PTPs in HeLa cells expressing Rac1-N17 or PrxI-Y194F was inhibited with vanadate. The results showed that addition of vanadate increased the duration of the Src signaling in cells expressing either Rac1-N17 [Fig. 4(e) and 4(k); DS: $57.5\pm 6.7\min$, $n=8$] or PrxI-Y194F [Fig. 4(f) and 4(k); DS: $63.9\pm 3.3\min$, $n=9$], verifying our above hypothesis that the activity of PTPs was insufficiently inhibited by ${\mathrm{H}}_2{\mathrm{O}}_2$ in those cells. Actually, for HeLa cells expressing Rac1-N17 or PrxI-Y194F, vanadate treatment increased the duration of Src signaling to the level of the control group [Fig. 4(k)], indicating that the effect of ${\mathrm{H}}_2{\mathrm{O}}_2$ on duration of Src signaling is mediated by its inhibition of PTPs. Taken together, we concluded that ${\mathrm{H}}_2{\mathrm{O}}_2$ modulates the duration of Src signaling through its inhibition of PTPs’ activity.

3.3.

${\mathrm{H}}_2{\mathrm{O}}_2$ Regulates Peak Amplitude of EGF-Induced Src Signaling Through Inhibiting PTP

As indicated in Fig. 4(g), reduction of endogenous ${\mathrm{H}}_2{\mathrm{O}}_2$ levels by overexpression of Rac1-N17 or PrxI-Y194F led to decreased PAS of Src signaling, suggesting that ${\mathrm{H}}_2{\mathrm{O}}_2$ is involved in activating Src kinase upon EGF treatment. In cells overexpressing Rac1-N17 or PrxI-Y194F, suppression of PTPs’ activity by vanadate not only prolonged the duration of Src signaling, but also increased the peak amplitude of the Src signal [Fig. 4(j); PAS in Rac1-N17-expressed cells: $1.14\pm 0.04$, $n=8$; PAS in PrxI-Y194F-expressed cells: $0.99\pm 0.06$, $n=9$]. Notedly, vanadate treatment recovered the peak amplitude [Fig. 4(j)] of Src signaling in cells expressing Rac1-N17 or PrxI-Y194F to the level of the control group, implying that ${\mathrm{H}}_2{\mathrm{O}}_2$ enhances Src activation through suppressing PTPs’ activity.

To further clarify the role of ${\mathrm{H}}_2{\mathrm{O}}_2$ on Src activity, we utilized exogenous ${\mathrm{H}}_2{\mathrm{O}}_2$ to examine its effect on the peak amplitude of EGF-induced Src signaling.³⁹ Three concentration gradients of ${\mathrm{H}}_2{\mathrm{O}}_2$(100-μM, 500-μM, and 1-mM ${\mathrm{H}}_2{\mathrm{O}}_2$) were used. When compared to the control group, both 500-μM and 1-mM ${\mathrm{H}}_2{\mathrm{O}}_2$ were able to enhance the PAS of EGF-induced Src signaling [Fig. 5(a) and 5(b)]. Notably, 100-μM ${\mathrm{H}}_2{\mathrm{O}}_2$ impaired EGF-induced Src signaling [Fig. 5(a) and 5(b); PAS: $0.60\pm 0.04$, $n=9$]. This phenomenon is explained in the next section.

Fig. 5

The impact of exogenous ${\mathrm{H}}_2{\mathrm{O}}_2$ on the kinetics of epidermal growth factor (EGF)-induced Src signaling in HeLa cells; (a) Kinetics of normalized cyan/yellow fluorescence protein (CFP/YFP) emission ratios of the Src biosensor in HeLa cells in response to EGF and ${\mathrm{H}}_2{\mathrm{O}}_2$. Three concentration gradients of ${\mathrm{H}}_2{\mathrm{O}}_2$ (100 μM, 500 μM, and 1 mM ${\mathrm{H}}_2{\mathrm{O}}_2$) were used. (b) Bar graphs represent the mean ± standard error (SEM) of peak amplitude of Src signaling in (a). The statistical results were from at least five independent experiments (c). Bar graphs represent the $\text{mean}\pm \mathrm{SEM}$ of peak amplitude of Src signaling from at least five independent experiments in d and e; (d and e) Kinetics of normalized CFP/YFP emission ratios of the Src biosensor in response to EGF and 1-mM ${\mathrm{H}}_2{\mathrm{O}}_2$ in either Rac1-N17 (d) or PrxI-Y194F-expressing (e) HeLa cells. (f) Kinetics of normalized CFP/YFP emission ratios of the Src biosensor in response to EGF and 1-mM ${\mathrm{H}}_2{\mathrm{O}}_2$ in either Rac1-N17- expressing HeLa cells pretreated with 500-μM vanadate for 30 min. “n” denotes the sample number in each group. “*”, “**” and “***” indicate $p<0.05$, $p<0.01$ and $p<0.001$, respectively.

In the following, we used 1-mM ${\mathrm{H}}_2{\mathrm{O}}_2$. In HeLa cells expressing Rac1-N17 or Prx1-Y194F, addition of 1-mM ${\mathrm{H}}_2{\mathrm{O}}_2$ effectively rescued the peak amplitude of EGF-induced Src signaling [Fig. 5(c), 5(d), and 5(e)]. Given that ${\mathrm{H}}_2{\mathrm{O}}_2$ is a robust inhibitor of PTP,^5,6 and that vanadate could rescue the peak amplitude of Src signaling in cells expressing Rac1-N17 or Prx1-Y194F [Fig. 4(j)], the ability of exogenous ${\mathrm{H}}_2{\mathrm{O}}_2$ to enhance Src activation could be attributed to its inhibition of PTPs’ activity. In fact, it has been reported that Src activity is regulated by ${\mathrm{H}}_2{\mathrm{O}}_2$-mediated oxidation at cysteine residues.²⁰ Nevertheless, the consequences (e.g., activation versus inactivation) of ${\mathrm{H}}_2{\mathrm{O}}_2$-dependent Src oxidation are in conflict.^21–25 Our results suggested that, in the context of an EGF-induced signal transduction pathway, ${\mathrm{H}}_2{\mathrm{O}}_2$-dependent Src oxidation per se does not exert an essential effect on Src activity. Collectively, we concluded that ${\mathrm{H}}_2{\mathrm{O}}_2$ regulates the peak amplitude of the EGF-induced Src signaling by inhibiting PTPs’ activity.

Notably, the kinetic profile of Src signaling displayed a striking difference in Rac1-N17-expressing HeLa cells upon treatment of EGF and 1-mM ${\mathrm{H}}_2{\mathrm{O}}_2$ [Fig. 5(d)]. The Src signaling rapidly declined to a low stable level at approximately 30 min, indicating that most Src biosensors were recovered in the dephosphorylation state. Since dephosphorylation of the Src biosensor is specifically mediated by action of PTPs, we reasoned that the rapidly decrease of the CFP/YFP emission ratio (Src signaling) was due to recovery of PTPs’ activity after temporary inhibition. To test this hypothesis, we pre-treated HeLa cells with vanadate. As demonstrated by Fig. 5(f), the CFP/YFP emission ratio remained stable during the imaging session, supporting above hypothesis. Thus, we verified again that ${\mathrm{H}}_2{\mathrm{O}}_2$ tunes the duration of EGF-induced Src signaling through its inhibition of PTPs’ activity. The values of three kinetic parameters under different experimental conditions are summarized in Table 1.

Table 1

Quantification of EGF-induced Src signaling with three parameters (Mean±SEM). “N.A” indicates that the data is not given because the groups displayed distinct temporal profiles of EGF-induced Src signaling.

3.4.

Global Elevation of Intracellular ${\mathrm{H}}_2{\mathrm{O}}_2$ Level Suppresses Propagation of EGF-Induced Src Signaling

Evidence suggests ${\mathrm{H}}_2{\mathrm{O}}_2$ is locally produced to specifically organize ${\mathrm{H}}_2{\mathrm{O}}_2$ signaling events upon growth factor stimulation.^15,40 However, the consequences for growth factor-induced signaling under global elevation of intracellular ${\mathrm{H}}_2{\mathrm{O}}_2$ level are unclear. In the above section, we determined that low concentration of exogenous ${\mathrm{H}}_2{\mathrm{O}}_2$ (100-μM) impaired EGF-induced Src signaling [Fig. 5(a) and 5(b)]. We reasoned that exogenous ${\mathrm{H}}_2{\mathrm{O}}_2$ would cause a global change in cellular redox state, leading to oxidative stress and inevitably triggering cellular antioxidant defenses, which might be implicated in regulating Src signaling. Because the glutathione (GSH) based redox buffer system is the major mechanism for resisting oxidative stress,^41,42 we attempted to determine the kinetics of the GSH-GSSG equilibrium by employing a Grx1-roGFP2 biosensor.⁴³ The $405/488\text{-}\mathrm{nm}$ excitation ratio of Grx1-roGFP2 in cells reflects the ratio of reduced form of glutathione (GSH) and oxidized form of glutathione (GSSG). Under oxidative stress, GSH is converted to GSSG, leading to reduction of GSH-GSSG ratio, which can be reflected by the increase of $405/488\text{-}\mathrm{nm}$ excitation ratio of Grx1-roGFP2. As indicated in Fig. 6(a) and 6(b), the GSH-GSSG ratio exhibited a transient decrease and rapid recovery, returning to the basal level within 10 min upon treatment with 100-μM ${\mathrm{H}}_2{\mathrm{O}}_2$ alone or together with EGF. The GSH-GSSG ratio typically reached a minimum at approximately 2 min. On the contrary, it takes more time for cells to remove higher concentrations of ${\mathrm{H}}_2{\mathrm{O}}_2$ (500 μM and 1 mM) as indicated in Fig. 6(a) (typically 15–20 min). This finding suggested that cells possess a powerful ${\mathrm{H}}_2{\mathrm{O}}_2$ elimination system, which might explain the results in Fig. 5(a). Because EGF-induced Src signal typically reach peak within 5 min, we hypothesized that the ${\mathrm{H}}_2{\mathrm{O}}_2$ degradation system triggered by low exogenous ${\mathrm{H}}_2{\mathrm{O}}_2$ would also act to remove the endogenous ${\mathrm{H}}_2{\mathrm{O}}_2$ that was induced by EGF for inhibition of PTPs’ activity. This removal would thus result in impairment of Src signaling. For higher concentration of ${\mathrm{H}}_2{\mathrm{O}}_2$ (500 μM and 1 mM), the exogenous ${\mathrm{H}}_2{\mathrm{O}}_2$ would instead enhance the peak amplitude of Src signal because cells are unable to eliminate such high concentration of ${\mathrm{H}}_2{\mathrm{O}}_2$ within 5 min.

Fig. 6

Imaging of GSH-GSSG ratio kinetics by the Grx1-roGFP2 biosensor in HeLa cells in response to treatment with epidermal growth factor (EGF) and ${\mathrm{H}}_2{\mathrm{O}}_2$; (a) The $405/488\text{-}\mathrm{nm}$ excitation ratio images of Grx1-roGFP2 in a HeLa cell in response to EGF, and (b) time course of normalized $405/488\text{-}\mathrm{nm}$ excitation ratios of Grx1-roGFP2 in response to different concentrations of ${\mathrm{H}}_2{\mathrm{O}}_2$(100 μM, 500 μM, and 1 mM) alone or EGF and 100-μM ${\mathrm{H}}_2{\mathrm{O}}_2$ as in (a); (c) time course of normalized $405/488\text{-}\mathrm{nm}$ excitation ratios of Grx1-roGFP2 in Rac1-N17-expressing HeLa cells in response to EGF and 1-mM ${\mathrm{H}}_2{\mathrm{O}}_2$. Scale bars, 10 μm.

The GSH-GSSG redox buffer system could also explain the distinct Src signaling profile in Fig. 5(d) (EGF and 1-mM ${\mathrm{H}}_2{\mathrm{O}}_2$). For cells co-expressing Rac1-N17 and Grx1-roGFP2, the $405/488\text{-}\mathrm{nm}$ excitation ratio rapidly increased following addition of EGF and 1-mM ${\mathrm{H}}_2{\mathrm{O}}_2$, and eventually declined to basal levels at approximately 20 min, indicating that the balance of GSH-GSSG was recovered [Fig. 6(c)]. Notably, the recovery of GSH-GSSG balance was a few minutes ahead of the low stable level of Src signaling in Fig. 5(d), strongly suggesting that GSH-dependent removal of ${\mathrm{H}}_2{\mathrm{O}}_2$ accounts for recovery of PTPs’ activity,⁴⁴ leading to decreased Src signaling.

In summary, we concluded that global elevation of intracellular ${\mathrm{H}}_2{\mathrm{O}}_2$ level can attenuate EGF-induced Src signaling. This evidence also suggests that EGF-induced ${\mathrm{H}}_2{\mathrm{O}}_2$ has to be locally generated. Otherwise, it triggers antioxidant defenses, leading to attenuation of EGF signal transduction.