Cell damage and reactive oxygen species production induced by fluorescence microscopy: effect on mitosis and guidelines for non-invasive fluorescence microscopy

Authors


For correspondence (fax +1 814 865 9131; e-mail rjc8@psu.edu).

Summary

The green fluorescent protein (GFP) and other intrinsically fluorescent proteins (IFPs) are popular reporters because they allow visualization of cellular constituents in living specimens. IFP technology makes it possible to view dynamic processes in living cells, but extended observation, using fluorescence microscopy (both wide-field and confocal), can result in significant light energy exposure. Therefore, it is possible that cells experience light-induced damage that alters cell physiology and confounds observations. To understand the impact that extended viewing has on cells, we obtained quantitative information about the effect of light energy dose and observation conditions on tobacco BY-2 cell physiology. Our results show a non-linear relationship between the excitation light intensity and mitotic arrest, and the frequency of mitotic arrest is dependent on the presence of an IFP that absorbs the excitation light. Moreover, fluorescence microscopy induces the production of reactive oxygen species (ROS), as assayed using BY-2 cells loaded with oxidation-sensitive dyes, and the level of ROS production increases if the cells express an IFP that absorbs the excitation light energy. The dye oxidation follows sigmoidal kinetics and is reversible if the cells are exposed to low irradiation levels. In addition, the dye oxidation rate shows a non-linear relationship to the excitation light intensity, and a good correlation exists between photobleaching, mitotic arrest, and dye oxidation. The data highlight the importance of ROS scavenging for normal mitotic progression, and provide a reference for judiciously choosing conditions that avoid photobleaching that can lead to ROS accumulation and physiological damage.

Introduction

The introduction of the green fluorescent protein (GFP) and other intrinsically fluorescent proteins (IFPs) as marker proteins has heralded a new era in cell biology. IFPs have been successfully used to study promoter activity and, in the form of fusion proteins, to visualize a variety of subcellular structures in living cells (Haseloff, 1999; Tsien, 1998). The use of IFPs has greatly facilitated the direct study of various dynamic processes that occur in living organisms, and the availability of the different spectral forms has added to the utility of IFP technology (Gadella et al., 1999; Haseloff, 1999; Reits and Neefjes, 2001).

IFP protocols are now widely used in plant biology because of the promise that they allow for non-invasive observation of cellular processes (Dixit and Cyr, 2002; Granger and Cyr, 2000; Hawes et al., 2001; Kost et al., 1998; Nebenführ et al., 2000). IFPs are typically visualized using wide-field or confocal microscopy, both of which can result in significant light exposure that can potentially damage cells. The potential for light-induced damage to cells increases during long-term observation or during the observation of multiple spectral forms of IFPs in the same cell. The lack of quantitative information about deleterious effects, resulting from extended fluorescence microscopy, prevents us from ensuring that the observation conditions are benign. Any unrecognized deleterious effects will confound observations and result in discrepancies between observations obtained using different observation conditions.

We were interested in obtaining quantitative information about light-induced cell damage, resulting from extended fluorescence microscopy, as well as gaining some understanding about cellular mechanisms that are affected by the incident radiant energy. Cells exposed to radiant energy may experience damage because of heating and/or the generation of reactive oxygen species (ROS). Fluorescence microscopy of unpigmented cells is unlikely to result in significant heating of cells because the peak absorption wavelengths for water lie in the infra-red range (Bernath, 2002; Matcher et al., 1994). However, fluorescence microscopy may lead to photooxidative damage through the well-known phenomenon of light-induced ROS production (Bartosz, 1997; Foyer et al., 1994; Wright et al., 2002).

ROS production may result in physiological changes in two possible ways: (i) ROS may directly damage key cellular molecules such as proteins, lipids, and nucleic acids (Halliwell and Gutteridge, 1989), resulting in physiological changes; and/or (ii) ROS may trigger changes through perturbation of the redox homeostasis (Schafer and Buettner, 2001). Free-radical-scavenging systems are important to protect cells from oxidative damage (Noctor and Foyer, 1998; Schafer and Buettner, 2001). In plants, the most abundant free-radical-buffering systems include ascorbate in the apoplast and vacuoles and glutathione in the cytoplasm (May et al., 1998; Noctor and Foyer, 1998). These systems themselves appear to be regulated because redox state changes accompany changes in the physiological state of the cell, both during the course of normal development (Schafer and Buettner, 2001) and as a result of environmental stress (Ren et al., 2002; Schafer and Buettner, 2001; Vranováet al., 2002). Importantly, changes in the cellular redox state affect cell viability and mitotic progression (Kato and Esaka, 1999; Kerk and Feldman, 1995; Kranner et al., 2002; Potters et al., 2000; Reichheld et al., 1999; Ren et al., 2002; Vernoux et al., 2000).

We conducted a quantitative analysis of the effect of various excitation light intensities and observation regimes, during fluorescence microscopy, on mitotic progression and viability of tobacco BY-2 cells expressing a variety of IFPs. Our results show a non-linear relationship between the excitation light intensity and mitotic arrest and cell death. The presence of a fluorophore, which absorbs the excitation light energy, increases the frequency of mitotic arrest under conditions where photobleaching is observed. In addition, shorter wavelength excitation light incurs greater cell damage than longer wavelength excitation light of equivalent irradiance.

Fluorescence microscopy induces ROS production in cells, as assayed using BY-2 cells containing oxidation-sensitive dyes. The oxidation of the dyes follows sigmoidal kinetics and is reversible at low light intensities. Moreover, there is increased ROS production in BY-2 cells expressing an IFP that absorbs the excitation light energy, compared to wild-type cells. In addition, the dye oxidation rate has a non-linear relationship with the excitation light intensity, and this correlates with the non-linear relationship between mitotic arrest and the excitation light intensity. The data indicate that light-induced ROS production contributes to the physiological damage incurred by cells during extended fluorescence microscopy and supports the hypothesis that normal mitotic progression requires the regulation of ROS levels. More importantly, our results provide quantitative guidelines for avoiding physiological damage during extended fluorescence microscopic observation of plant cells expressing IFPs.

Results

The effect of light energy on transgenic tobacco cells

Mitotic arrest and cell death induced by fluorescence microscopy

A variety of transgenic cell lines were used to assay for the effect of light energy on mitotic progression: (i) BD2-5 cell line, expressing GFP-MBD (GFP fused to a microtubule-binding domain; Granger and Cyr, 2000); (ii) MBD-DsRed cell line, expressing the microtubule-binding domain fused to DsRed; (iii) A dual-marker, H2B-ECFP/GFP-MBD cell line that was generated by transforming the BD2-5 line with a second transgene encoding histone-2B fused to the enhanced cyan fluorescent protein (H2B-ECFP construct provided by Dr Jim Haseloff); and (iv) A dual-marker, H2B-mYFP/MBD-DsRed cell line, that was generated by transforming the MBD-DsRed cell line with a second transgene encoding histone-2B fused to the modified yellow fluorescent protein (H2B-mYFP construct provided by Dr Jim Haseloff; Boisnard-Lorig et al., 2001). The dual-marker cell lines allowed visualization of both microtubules and chromatin in living cells and therefore provided two sets of markers for mitotic progression.

To determine the relationship between light energy and mitotic arrest, wide-field, time-lapse microscopy was conducted, at various excitation light intensities, using the different cell lines. The progression of mitosis was used as an indicator for the effect of light on cell physiology. Cells were considered to be arrested in mitosis if they did not progress into the next mitotic stage after at least 2 h of being established in one stage, or if the mitotic apparatus disintegrated during observation.

At low light intensities, mitosis progressed normally (Figures 1a, S2a, and S3a), whereas high light intensities resulted in mitotic arrest. In most cases (>90%), the cells arrested in metaphase (Figures 1c, S2b, and S3b) and displayed aberrant spindles and loss of chromosome alignment at the metaphase plate (Figure S2b). Occasionally, the cells arrested in pre-prophase and the pre-prophase band either failed to mature or seemed to revert to its immature state (Figure 1b). Cells arrested in mitosis, at very high light intensities, subsequently plasmolysed (last panel in Figure S2b) and died.

Figure 1.

Fluorescence microscopy-induced mitotic arrest.

Progression of mitosis was visualized using the GFP-MBD microtubule marker. Cells with a pre-prophase band were chosen for observation, and images were captured at 30-sec intervals using 0.5-sec camera-exposure times for wide-field microscopy, or 16-sec scan times for confocal microscopy. The time of observation is indicated in hours:minutes (t = 0 indicates start of the time-lapse microscopy). Scale bars = 10 µm.

(a) Wide-field microscopy at 50% GFP light intensity. Mitosis progresses normally.

(b) Wide-field microscopy at 90% GFP light intensity. Mitosis arrested in pre-prophase.

(c) Wide-field microscopy at 75% GFP light intensity. Mitosis arrested in metaphase.

(d) Confocal laser scanning microscopy at idle laser power (3.2 × 10−5 J cm−2 sec−1). Mitosis arrested in metaphase.

Similar results were obtained when the cells were observed using laser scanning confocal microscopy. When time-lapse, confocal microscopy was conducted at the idle laser power (3.2 × 10−5 J cm−2 sec−1 or 32 µW cm−2), the mitotic spindle disintegrated over time and the cells failed to complete mitosis (Figure 1d). The progression of mitosis was unhindered if the laser power was attenuated at least threefold during observation (7.9 × 10−6 J cm−2 sec−1 or 7.9 µW cm−2).

Relationship between light energy dose and mitotic arrest

The wide-field, time-lapse microscopy experiments were conducted using the different cell lines, at various light intensities and observation regimes, to obtain quantitative information about the effect of excitation light energy on mitotic progression.

A non-linear relationship was observed between the excitation light energy dose and the frequency of mitotic arrest and cell death (Tables 1 and 2; Figure S4). The results indicate that, below a certain threshold of light energy dose, cell physiology is not significantly affected but above which there is persistent physiological damage.

Table 1.  Relationship between GFP and DsRed excitation light and mitotic arrest
Light intensity
(HBO 103 W/2
short-arc lamp), %
Observation
conditions
Light energy per
observation period
(J cm−2)
% cells arrested
in mitosis
(n = sample size)
  • a

    Camera exposure time.

  • b

    Time-lapse interval.

  • c The light energy exposure over 1 h of time-lapse microscopy was calculated to allow comparison between the different observation conditions. See Experimental procedures for details about calculation.

  • d

    Comparison between 90 and 75% GFP light: P-value for the null hypothesis = 0.023.

GFP-MBD cells (460–500 nm)
 900.5 seca; 30 secb1.434c100 (10)
 750.5 sec; 30 sec1.19467 (12)d
 500.5 sec; 30 sec0.7900 (10)
 10–250.5 sec; 1 min0.081–0.1980 (>30)
MBD-DsRed (515–560 nm)
 750.5 sec; 30 sec4.031 (13)
 350.5 sec; 30 sec1.8660 (10)
Table 2.  Relationship between CFP and YFP excitation light and mitotic arrest
Light intensity
(HBO 103 W/2
short-arc lamp), %
Observation
conditions
Light energy per
observation period
(J cm−2)
% cells arrested
in mitosis
(n = sample size)
  • a

    Camera exposure time.

  • b

    Time-lapse interval.

  • c The light energy exposure over 1 h of time-lapse microscopy was calculated to allow comparison between the different observation conditions. See Experimental procedures for details about calculation.

  • d

    Comparison between 28 and 15% CFP light: P-value for the null hypothesis = 0.07.

  • e

    Comparison between 15% CFP and 76% YFP light: P-value for the null hypothesis = 0.013.

H2B-ECFP/GFP-MBD (426–446 nm)
 280.5 seca; 30 secb1.662c100 (10)
 150.5 sec; 30 sec0.9080 (15)d
 50.5 sec; 30 sec0.300 (10)
H2B-mYFP/MBD-DsRed (490–510 nm)
 760.5 sec; 30 sec0.8040 (15)e
 250.5 sec; 30 sec0.260 (11)

We also observed that shorter wavelength excitation light afflicted greater damage to the cells. In the case of the microtubule markers, the cells withstood much higher doses of DsRed excitation light (i.e. 515–560 nm), compared to GFP excitation light (i.e. 460–500 nm) doses, before they exhibited mitotic arrest (Table 1). Similarly, in the case of the chromatin markers, the YFP excitation light (i.e. 490–510 nm) was less damaging compared to the CFP excitation light (i.e. 426–446 nm), of equivalent irradiance (Table 2). This result suggests a wavelength-dependent effect on cell physiology that is probably the outcome of a combination of factors such as the energy per photon (shorter wavelength photons having more energy), spectral properties of biomolecules (biomolecules tend to absorb shorter wavelengths more efficiently) and photophysical characteristics of each IFP (such as quantum yield, relative bleaching time, etc.).

During the course of our experiments, we also observed that the detrimental effects of light energy on cell physiology were influenced by two variables of the light energy dose: the light energy dose received by cells during camera exposure (light energy per frame) and the light energy dose accrued by cells over a period of the time-lapse observation (light energy per period).

The light energy per frame is a key factor that determines the extent of damage that the cells will incur during a single observation. For example, cells exposed to 90% GFP light intensity (at 0.5 sec per exposure) suffer greater damage than cells exposed to 50% GFP light intensity (at 1.5 sec per exposure), even though observation conditions result in a energy dose of 1.434 J cm−2 in the former case, compared to an energy dose of 2.394 J cm−2 in the latter case, over 1 h of time-lapse microscopy (Table 3, upper panel). Therefore, it is advantageous to employ low excitation light intensity and a longer camera exposure time to capture high-resolution images than to resort to using high excitation light intensities and lower camera exposure times.

Table 3.  Contribution of light energy per frame and light energy per period to mitotic arrest
Light intensity
(HBO 103 W/2
short-arc lamp)
GFP-MBD (460–500 nm), %
Observation
conditions
Light energy per
image frame
(J cm−2 sec−1)
Light energy per
observation period
(J cm−2)
% cells arrested
in mitosis
(n = sample size)
  • a

    Camera exposure time.

  • b

    Time-lapse interval.

  • c The light energy exposure over 1 h of time-lapse microscopy was calculated to allow comparison between the different observation conditions. See Experimental procedures for details about calculation.

  • d

    Comparison between 90 and 50% light energy per frame conditions: P-value for the null hypothesis = 0.012.

Light energy per frame
 900.5 seca; 30 secb0.02391.434c100 (10)
 501.5 sec; 30 sec0.01332.39460 (10)d
Light energy per period
 750.5 sec; 30 sec0.01991.19467 (12)
 750.5 sec; 1 min0.01990.5970 (10)
 501.5 sec; 30 sec0.01332.39460 (10)
 500.5 sec; 30 sec0.01330.7980 (10)

The light energy per period, at a given light intensity, is determined by the duration of the camera exposure time for each observation and the frequency at which images are captured. As expected, an increase in the light energy per period to cells, at a constant excitation light intensity, resulted in greater damage to the cells (Table 3, lower panel). In addition, our results reveal that the damage resulting from the light energy per period can be prevented by sacrificing either the temporal-information aspect (the time-lapse interval) of time-lapse microscopy or the spatial-information aspect (camera exposure time) of time-lapse microscopy (Table 3, lower panel).

Contribution of IFP expression towards physiological damage during fluorescence microscopy

It is possible that IFP expression increases the chances of physiological damage, suffered by cells, during fluorescence microscopy. To test this possibility, the frequency of mitotic arrest was determined for three cell lines at selected excitation wavelengths: H2B-ECFP/GFP-MBD cells, GFP-MBD cells, and MBD-DsRed cells, observed using either CFP optics or DsRed optics. The results indicate that there is a significant increase in the probability of mitotic arrest because of fluorescence microscopy if the cells are expressing an IFP form that is excited by the incident radiation (Table 4). In other words, the presence of a fluorophore that absorbs the excitation light energy increases the sensitivity of cells to damage by fluorescence microscopy. Note that the CFP excitation wavelengths (426–446 nm) overlap slightly with the excitation spectra for GFP and DsRed, and this is correlated to mitotic arrest observed in the GFP-MBD and MBD-DsRed cell lines at 28% CFP light intensity (Table 4). On the other hand, the DsRed excitation wavelengths (515–560 nm) do not significantly overlap with the excitation spectra for GFP and CFP, and this is correlated to no mitotic arrest observed in the GFP-MBD and H2B-ECFP/GFP-MBD cell lines at 75% DsRed light intensity (Table 4).

Table 4.  Contribution of IFP expression towards mitotic arrest during fluorescence microscopy
Light intensity
(HBO 103 W/2
short-arc lamp), %
% cells arrested in mitosis (n = sample size)
H2B-ECFP/GFP-MBD cellsGFP-MBD cellsMBD-DsRed cells
  • The same observation conditions (images captured at 30-sec intervals using 0.5-sec camera exposure times) were used for all experiments.

  • a

    Comparison between the H2B-ECFP/GFP-MBD and GFP-MBD cells at 28% CFP light: P-value for the null hypothesis = 0.023.

  • b

    Comparison between the GFP-MBD and MBD-DsRed cells at 28% CFP light: P-value for the null hypothesis = 0.086.

  • c Comparison between the H2B-ECFP/GFP-MBD and GFP-MBD cells at 15% CFP light: P-value for the null hypothesis = 10−5.

  • d

    Comparison between the MBD-DsRed and GFP-MBD cells at 75% DsRed light: P-value for the null hypothesis = 0.028.

426–446 nm
 28100 (10)67 (12)a38 (13)b
 1580 (15)10 (10)c0 (10)
515–560 nm
 750 (13)0 (10)d31 (13)
 350 (10)0 (10)

The effect of light energy on wild-type tobacco cells

Fluorescence microscopy and the production of reactive oxygen species

We were interested in determining the basis for the physiological changes or damage incurred by BY-2 cells during extended fluorescence microscopy. In particular, we wanted to determine whether fluorescence microscopy was accompanied by light-induced formation of ROS in BY-2 cells, which could lead to oxidative damage to the cells. To determine this, oxidation-sensitive dyes were used to test whether ROS production occurred in BY-2 cells during fluorescence microscopy. These dyes are non-fluorescent in the reduced state and become fluorescent upon oxidation by ROS. Therefore, an increase in fluorescence of the dyes indicates ROS production.

Wild-type BY-2 cells were loaded with either the 6-carboxy-2′,7′-dichlorodihydrofluorescein diacetate, di(acetoxymethyl ester) dye or the 5- and 6-chloromethyl-2′,7′-dichlorodihydrofluorescein diacetate, di(acetoxymethyl ester) dye. Similar results were obtained for both dyes; the results for the chloromethyl version of the dye are presented in this paper. The redox dye partitions to the cytoplasm and the nucleoplasm of BY-2 cells, so an area encompassing the nucleus and the perinuclear cytoplasm of each cell was chosen to conduct fluorescence intensity measurements (Figure 2).

Figure 2.

Distribution of the ROS dye in wild-type BY-2 cells.

Wild-type BY-2 cells were exposed to 20 µm 5- and 6-chloromethyl-2′,7′-dichlorodihydrofluorescein diacetate, di(acetoxymethyl ester) dye for 2 min in the dark. The cells were then washed three times to remove excess dye and visualized using GFP optics. Note that the dye localizes to the cytoplasm and nucleoplasm of BY-2 cells. The boxes represent the areas chosen for measurement of dye intensity. Scale bar = 10 µm.

Time-lapse microscopy was conducted on dye-loaded, wild-type BY-2 cells at various excitation light intensities using GFP optics. Our results show that the oxidation of the dye follows sigmoidal kinetics (Figure 3a). The lag phase of the curve represents a measure of the ROS-scavenging capacity of the cell, after which there is a linear increase in the level of oxidation of the dye pool that ultimately reaches maximal oxidation. If the dye-loaded cells are exposed to low light for a short period (10 sec), and then allowed to recover in the dark for 5 min, the fluorescence intensity of the dye drops to the initial levels (data not shown), indicating that the cells can recover from the oxidation induced by fluorescence microscopy. However, if the cells are exposed to high light intensities, for longer time periods, the fluorescence intensity of the dye does not recover to the initial level after 5 min of incubation in the dark (data not shown).

Figure 3.

Fluorescence microscopy-induced ROS production in BY-2 cells.

(a) Wild-type BY-2 cells containing 20 µm 5- and 6-chloromethyl-2′,7′-dichlorodihydrofluorescein diacetate, di(acetoxymethyl ester) redox dye were visualized at various light intensities, and digital images were captured at 1-sec intervals for 50 sec. The fluorescence intensity of the dye, in an area encompassing the nucleus and perinuclear cytoplasm, was determined at each time point. The traces represent sample curves obtained for the dye fluorescence intensity at various light intensities.

(b) Graph showing the relationship between oxidation rate of the chloromethyl redox dye and excitation light intensity. The inverse time for half-maximal oxidation of the dye, at the light intensities that saturated dye oxidation (50, 75, and 90% light intensities), was determined as a measure of the dye oxidation rate. Each data point represents the average of at least 30 separate measurements, and the error bars denote standard deviation.

We observed that the dye oxidation rate was dependent on the excitation light (460–500 nm wavelength) intensity: at 90% light intensity, there was a sharp rise in the oxidation of the dye that reached saturation by 20–30 sec, whereas at 25% light intensity, the dye exhibited gradual oxidation that did not reach saturation during the observation period (Figure 3a). The time for half-maximal oxidation of the dye was used as a measure of the dye oxidation rate at the different light intensities. A comparison of the time of half-maximal oxidation of the dye to the excitation light intensity revealed a non-linear relationship between the two parameters (Figure 3b). This non-linear relationship between the rate of ROS production of BY-2 cells and the excitation light intensity correlates to the non-linear relationship observed between mitotic arrest and excitation light intensity.

Excitation of the dye, during the course of these experiments, could also contribute to ROS production and influence the measurements. However, we did not detect any significant photobleaching of the dye during the experiments, which would predictably occur if the dye generated significant levels of free radicals. In addition, we observed that wavelengths that do not excite the dye (515–560 nm) also result in oxidation of the dye that shows a non-linear relationship with the excitation light intensity: the net oxidation of the dye at 35% DsRed excitation light intensity was (2.06 ± 1.15) × 106 (arbitrary pixel grayscale value; 40 cells), whereas the net oxidation of the dye at 75% DsRed excitation light intensity was (7.51 ± 3.19) × 106 (29 cells; P-value < 10−10 for the null hypothesis). Therefore, the dye itself does not significantly interfere with measurements of the ROS production in the cells. Moreover, experiments conducted using DsRed exciting wavelengths, with MBD-DsRed expressing cells loaded with the oxidation-sensitive dye, show an increased level of ROS production in these cells compared to the wild-type cells under the same conditions ((3.04 ± 1.93) × 106 at 35% light intensity, and (9.36 ± 3.01) × 106 at 75% light intensity; >27 cells in each case; P-value is 0.02 for the null hypothesis compared to the wild-type cells for both light intensities). This result indicates that the presence of an IFP that absorbs the excitation light energy enhances ROS production during extended fluorescence microscopy.

Discussion

The advantage of using an IFP as a marker protein is that it can be visualized in living cells to observe various dynamic cellular processes. However, the benefits of IFP technology can only be exploited through a careful consideration of the microscopic conditions used to visualize the IFP. Both wide-field and confocal microscopy can result in significant light dosage to cells and, under harsh observation conditions, prolonged visualization of IFP-expressing cells will result in physiological changes and eventually cell death.

We used mitotic progression as an indicator of the physiological status of cells in our experiments because it consists of events that can be easily scored and because mitotic progression is known to be sensitive to the redox state of cells (Kato and Esaka, 1999; Kerk and Feldman, 1995; Potters et al., 2000; Reichheld et al., 1999; Vernoux et al., 2000). The correlation between mitotic arrest, ROS production rate and excitation light intensity suggests the presence of redox-sensitive checkpoints that could be responsible for arresting mitosis in our experiments. The nature of these checkpoints is not known, but possibly includes changes in the expression and activity of cell cycle regulatory proteins (Reichheld et al., 1999; Russo et al., 1995; Schafer and Buettner, 2001). The adaptive significance of mitotic arrest may be to limit the propagation of heritable changes in cells exposed to oxidative stress (May et al., 1998; Reichheld et al., 1999).

We observed a non-linear relationship between the excitation light intensity and the frequency of mitotic arrest in BY-2 cells. The non-linear nature of this relationship indicates that the cells can tolerate light dose to a certain threshold beyond which the cells rapidly suffer physiological change/damage, probably through a ROS-mediated mechanism. This observation is consistent with the hypothesis that cells possess a light-induced damage repair mechanism that is subject to saturation. We hypothesize that the processes responsible for scavenging and/or downregulating ROS production constitute at least one aspect of the repair mechanism.

The experiments using oxidation-sensitive dyes show that fluorescence microscopy induces ROS production in cells. ROS generation under these conditions probably occurs as a result of energy transfer from excited state IFPs and cellular biomolecules to molecular oxygen (Bensasson et al., 1993). Although ROS probably represent the predominant oxidant species causing oxidative stress during fluorescence microscopy, other oxidant species (e.g. reactive nitrogen species) could also be generated and contribute to oxidative stress during fluorescence microscopy.

ROS production during fluorescence microscopy follows sigmoidal kinetics. Studies of photobleaching of fluorescent dyes have also revealed non-linear fluorescence decay curves with respect to excitation light intensities (Benson et al., 1985). However, where the photobleaching process corresponds to first-order kinetics (Benson et al., 1985), our data shows that ROS production corresponds to more complex reaction kinetics. The sigmoidal pattern indicates that cells can scavenge/downregulate a certain degree of ROS production resulting from the excitation light; however, high light doses saturate the scavenging/downregulating mechanisms, at which point the cells will suffer oxidative damage. Therefore, we suggest that all observations be conducted under conditions where the ROS-scavenging/downregulating system is well below saturation.

The light energy dose received by cells provides a quantitative criterion for performing non-invasive fluorescence microscopy (Table 5). This criterion incorporates several parameters: (i) excitation light intensity; (ii) wavelength of the excitation light; (iii) camera exposure time; and (iv) frequency of observation.

Table 5.  Guidelines for conducting non-invasive fluorescence microscopy
ParameterSuggested setting
Light intensityUse of low excitation light intensity is a key factor. Increase
camera exposure time, instead, for greater spatial resolution.
Spatial versus temporal resolutionInherent trade-off between spatial and temporal resolution.
Sacrifice one or the other according to the experimental goal.
IFP expression levelMay need to select for cell lines with lower IFP expression to
facilitate long-term, non-invasive, fluorescence microscopy.
Excitation wavelengthLonger excitation wavelengths are less invasive.
Use IFPs requiring shorter excitation wavelengths (GFP, CFP,
and BFP) to label structures expected to result in high contrast.

The excitation light intensity (light energy per frame) is a key factor that determines the extent of cell damage during fluorescence microscopy. High light intensities inflict greater damage to cells, even after brief exposure, compared to longer exposure of cells to low light intensities. Therefore, the use of low excitation light intensities is the foremost condition for non-invasive fluorescence microscopy.

The wavelength of the excitation light is determined by the spectral form of the IFP used for observation. Shorter wavelength excitation light is more damaging to cells than longer wavelength excitation light because of increased efficiency of ROS production at shorter wavelengths (Rozanowska et al., 1998). In practical terms, if multiple spectral forms of IFP are to be used in the same cell, we recommend using the shorter excitation wavelength-demanding spectral forms (such as CFP or the blue fluorescent protein) for labeling structures that are expected to result in a high density of fluorophore labeling (such as organelles, chromatin, etc.). High-density fluorophore labeling will result in high contrast (i.e. high signal-to-noise ratio) and will therefore tolerate lower excitation light intensity for optimal observation.

An added complication arises from the fact that the expression of IFP enhances the probability of physiological damage suffered by cells during extended fluorescence microscopic observation. This is probably because of the fact that IFP expression enhances the level of light-induced ROS production during fluorescence microscopy. Therefore, we recommend selection of cells that express low levels of IFP for extended fluorescence microscopy.

Fluorescence microscopy of IFP-labeled cells is usually conducted with the aim of obtaining both high spatial and temporal resolution of cellular structures and processes. Confocal microscopy is the most popular method to obtain high spatial resolution of fluorescent structures because of the enhanced contrast afforded by low, out-of-focus fluorescence (Hepler and Gunning, 1998), and the commonly held belief that this method is less invasive. However, image acquisition can be a slow process during confocal microscopy and cells are subject to damage if the laser power is not suitably attenuated. We observed that high laser power (3.2 × 10−5 J cm−2 sec−1 or 32 µW cm−2, in our case) during confocal microscopy results in damage to BY-2 cells as well as to Arabidopsis root cells (unpublished observations).

Wide-field microscopy generally provides greater temporal resolution, compared to confocal microscopy, because of the short exposure times required to capture images, at least when using high-sensitivity, cooled, charge-coupled-device cameras. In addition, high signal-to-noise ratios can also be obtained using wide-field fluorescence microscopy by increasing camera exposure time or through the use of a variety of image restoration techniques (McNally et al., 1999). However, the light energy dose received by cells during wide-field microscopy presents a limitation as to the observation conditions that can be used to conduct non-invasive fluorescence microscopy. Time-lapse observations designed to obtain three-dimensional (two-dimensional images collected over time) or four-dimensional (stacks of Z-section images collected over time) information about cellular processes can lead to high light energy doses to cells and alter cell physiology. Our results indicate that non-invasive, time-lapse microscopy can be conducted by lowering either the temporal resolution (using longer time-lapse intervals) or the spatial resolution (lowering camera exposure time). Because of this trade-off, observation conditions may need to sacrifice either high temporal or high spatial resolution to maintain specimen health.

Based on our results, we recommend that fluorescence microscopy be conducted using observation conditions that avoid fluorescence bleaching. Under non-photobleaching conditions, the ROS scavenging/downregulating capacity of cells should be below saturation and observations can be conducted in a physiologically, non-invasive manner.

Experimental procedures

Transgenic constructs and cell cultures

All cell cultures were maintained at 26°C on a rotary platform shaker, at 100 r.p.m., and subcultured weekly. Wild-type tobacco BY-2 cells and BD2-5 cells, expressing the GFP-MBD marker protein, have been described by Granger and Cyr (2000).

The MBD-DsRed fusion construct was generated using recombinant PCR (Dixit and Cyr, 2002) with the following primers: 5′ MBD, AGTCGACTCCCGGCAAGAAGAAGC; 3′ MBD, TCCGGTTGTTGCGGCACCTCCTGCAGGAAAGTG; 5′ DsRed, GCCGCAACAACCGGAGCCATGAGGTCTTCCAAG; 3′ DsRed, TGAGCTCCTAAAGGAACAGATGGTG. The construct was introduced into the pBI101 vector between a 35S CaMV promoter and a nos terminator and introduced into BY-2 cells using Agrobacterium-mediated transformation. Transformed cells were selected using 100 mg l−1 kanamycin, and individual calli were used to generate cell cultures.

The H2B-ECFP-encoding transgene and the H2B-mYFP-encoding transgene were generously provided by Jim Haseloff. The transgenes were inserted into the pCAMBIA 1300 vector (CAMBIA, Canberra, Australia) between a 35S CaMV promoter and a nos terminator and introduced into BD2-5 cells, or MBD-DsRed cells, using Agrobacterium-mediated transformation. Transformed cells were selected using 30 mg l−1 hygromycin and 100 mg l−1 kanamycin on plates. Individual, putatively transformed, calli were transferred to liquid medium, containing the selective antibiotics, to initiate cell cultures.

ROS dye experiments

Experiments were conducted using both 6-carboxy-2′,7′-dichlorodihydrofluorescein diacetate, di(acetoxymethyl ester) and 5- and 6-chloromethyl-2′,7′-dichlorodihydrofluorescein diacetate, di(acetoxymethyl ester) dyes obtained from Molecular Probes (Eugene, OR, USA). 10 mm stock solutions were freshly prepared using anhydrous DMSO. BY-2 cells, 2–3 days after subculture, were exposed to 20 µm dye for 2 min at 26°C in the dark. The cells were then washed three times with fresh culture medium lacking the dye and immediately used for observations. The dyes localize to the cytoplasm and nucleoplasm for about 15 min, after which they gradually partition into the vacuoles. Therefore, all observations were conducted within 10–15 min of washing of the dye-loaded cells.

Determination of the excitation light energy dosage

A mercury HBO 103 W/2 short-arc lamp (OSRAM), under the control of the variable-intensity Atto-Arc system (Zeiss, Thornwood, NY, USA), was used as a light source for wide-field microscopy. The excitation light irradiance, at various light intensities, was determined using a LI-1000 light meter, equipped with a LI-190SA Quantum sensor (LI-COR, Inc., Lincoln, NE, USA). The sensor was placed at the specimen focal plane, in the presence of oil, for the measurements, and the data for the different filter sets is shown in Figure S1.

The mercury arc-lamp spectrum is discontinuous and consists of peaks of varying intensities. The peak output wavelengths corresponding to the DsRed, YFP, GFP, and CFP filter sets are 546 nm (energy = 3.64 × 10−19 J), 497 nm (energy = 4 × 10−19 J), 497 nm (energy = 4 × 10−19 J), and 436 nm (energy = 4.6 × 10−19 J), respectively. The excitation light energy, at the different light intensities, was estimated as follows:

image(1)

As a comparison of the light energy dose at the different observation regimes, the light energy exposure over 1 h of time-lapse microscopy was calculated according to the camera exposure time and the time-lapse interval for each observation regiment:

image(2)

For confocal microscopy, the LI-190SA sensor was placed at the specimen focal plane, in the presence of oil, for the measurements. The light energy exposure was determined using Equation 1, with the energy of the 488-nm laser photons as 4.1 × 10−19 J.

Microscopy

All observations were carried out on living cells, immobilized on poly-l-lysine-coated coverslips, in a humid chamber. Images were collected using a Plan-Neofluar ×40 (numerical aperture, 1.3) oil-immersion objective (Zeiss, Thornwood, NY, USA).

Wide-field microscopy was conducted with a shutter-equipped Zeiss Axiovert S100 TV microscope, and digital images were captured with a Hamamatsu Orca charge-coupled-device camera (Hamamatsu Corp., Bridgewater, NJ, USA) controlled by esee software (Inovision Corp., Durham, NC, USA). DsRed (515–560 nm excitation, LP 590 nm emission), YFP (490–510 nm excitation, 520–550 nm emission), GFP (460–500 nm excitation, 510–560 nm emission), and CFP (426–446 nm excitation, 480–500 nm emission) filter sets were used to visualize and discriminate between the fluorophores. In all experiments, a computer-controlled shutter ensured that cells were exposed to light only during image acquisition.

For the time-lapse microscopy experiments of the transgenic BY-2 cells, cells were exposed to varying light intensities and digital images were captured under different observation regimes as mentioned in the text. In all cases, 2 × 2 pixel binning was used to capture digital images. The time, t = 0, indicates the start of time-lapse microscopy.

For the ROS dye experiments, 50 images were captured at 1-sec intervals using the GFP filter set, and the fluorescence intensity of the redox dye in each cell was determined in an area encompassing the nucleus and the peri-nuclear cytoplasm at each time point. For some experiments, 52 images were captured at 1-sec intervals using the following scheme: first image at 1% GFP light intensity, followed by 50 images at different DsRed light intensities, followed by a final image at 1% GFP light intensity to record the change in dye fluorescence. In all cases, a 0.1-sec camera exposure time was used to record images. Fluorescence intensity measurements were corrected for background fluorescence for each frame.

Laser scanning confocal microscopy was conducted using an LSM model 410 (Zeiss Corp.) microscope, and cells were imaged using the 488-nm line of an argon laser. Optical sections were collected at 30-sec intervals using a 488/543 nm dual dichroic excitation mirror, a 510–540 nm emission filter and 16-sec total scan time (4-sec scan and four-line averaging).

Acknowledgements

We thank Jim Haseloff and Sarah Hodge for kindly providing us with the H2B-ECFP and H2B-mYFP transgenes, Melissa Ho and Neil Barto for providing us with the LI-1000 light meter, and Deb Fisher for critical reading of the manuscript. We also thank the two anonymous reviewers for their comments and suggestions. This work was supported by USDA grant # 98-35304-6668 and DOE grant # DE-FG02-91ER20050.

Supplementary Material

The following material is available from http://www.blackwellpublishing.com/products/journals/suppmat/TPJ/TPJ1868/TPJ1868sm.htm

Figure S1. Light meter measurements for the different filter sets.

The LI-190SA sensor of a LI-1000 Datalogger light meter was placed at the specimen plane of the ×40 Plan-Neofluar objective, in the presence of immersion oil, for the measurements. The light meter measurements were conducted in a dark room, with aluminum foil shielding stray light from the arc-lamp. Irradiance for the different filter sets was expressed in µmol m−2 sec−1. The error bars on the graphs indicate standard deviation.

(a) YFP filter set (490–510 nm excitation wavelengths).

(b) GFP filter set (460–500 nm excitation wavelengths).

(c) DsRed filter set (515–560 nm excitation wavelengths).

(d) CFP filter set (426–446 nm excitation wavelengths).

Figure S2. Wide-field microscopy-induced mitotic arrest and plasmolysis of H2B-CFP GFP-MBD cells.

Progression of mitosis was recorded using the H2B-CFP chromatin marker along with occasional snapshots in the GFP channel to record the microtubule configuration. Cells with a pre-prophase band were chosen for observation, and digital images were captured at 30-sec intervals using 0.5-sec camera exposure times. The time of observation is indicated in hours:minutes (t = 0 indicates the start of time-lapse microscopy). Scale bars = 10 µm.

(a) 5% CFP light intensity. Mitosis progresses normally.

(b) 15% CFP light intensity. Mitosis arrested in metaphase. The last frame is a bright-field image of the cell at the onset of plasmolysis.

Figure S3. Wide-field microscopy-induced mitotic arrest in MBD-DsRed cells.

The MBD-DsRed microtubule marker was used to record mitotic progression. Cells with a pre-prophase band were chosen for observation, and digital images were captured at 30-sec intervals using 0.5-sec camera exposure times. The time of observation is indicated in hours:minutes (t = 0 indicates the start of time-lapse microscopy). Scale bars = 10 µm.

(a) 35% DsRed light intensity. Mitosis progresses normally.

(b) 75% DsRed light intensity. Mitosis arrested in metaphase.

Figure S4. Non-linear relationship between excitation light intensity and mitotic arrest.

Graphs plotting the extent of mitotic arrest observed at different excitation light intensities. Note the presence of a threshold of excitation light intensity, below which the cells complete mitosis normally, and above which mitotic arrest is observed.

(a) GFP-MBD cells observed using GFP optics (460–500 nm excitation).

(b) H2B-CFP/GFP-MBD cells observed using CFP optics (426–446 nm excitation).

Ancillary