Direct Imaging of Dehydrogenase Activity within Living Cells Using
Enzyme-Dependent Fluorescence Recovery after
C. A. Combs and R. S. Balaban
Laboratory of Cardiac Energetics, National Heart, Lung, and Blood Institute, National Institutes of Health,
Bethesda, Maryland 20892-1061 USA
many redox reactions. We describe the use of confocal microscopy in conjunction with enzyme-dependent fluorescence
recovery after photobleaching (ED-FRAP) of NADH as a topological assay of NADH generation capacity within living cardiac
myocytes. Quantitative validation of this approach was performed using a dehydrogenase system, in vitro. In intact cells the
NADH ED-FRAP was sensitive to temperature (Q10of 2.5) and to dehydrogenase activation by dichloroacetate or cAMP
(twofold increase for each). In addition, NADH ED-FRAP was correlated with flavin adenine dinucleotide (FAD?) fluorescence.
These data, coupled with the cellular patterns of NADH ED-FRAP changes with dehydrogenase stimulation, suggest that
NADH ED-FRAP is localized to the mitochondria. These results suggest that ED-FRAP enables measurement of regional
dynamics of mitochondrial NADH production in intact cells, thus providing information regarding region-specific intracellular
redox reactions and energy metabolism.
Reduced nicotine adenine dinucleotide (NADH) is a key metabolite involved in cellular energy conversion and
The regulation of many cellular processes is thought to be
influenced by the regional distribution of enzymes and local
conditions (Srivastava and Bernhard, 1987; Saks et al.,
1991). The study of cellular compartmentation has gener-
ally been limited to evaluating the intracellular distribution
of enzymes/proteins/metabolites or the imaging of regional
ionic conditions. A recent study has shown direct visualiza-
tion of enzymatic processes in living cells using fluores-
cence resonance energy transfer (FRET) on GFP-modified
proteins (Nagai et al., 2000); however, quantification of
enzyme activity from regions within living cells has not
been reported. A method for measuring enzymatic activity
in vivo would assist in integrating regional cytosolic regu-
latory processes. We hypothesized that the recovery of
fluorescence of a fluorophore after photobleaching could be
used to assay regional intracellular enzymatic activity. In
this study we have focused on the regeneration of the
the cell due to its key role in cellular metabolism.
Fluorescence recovery from photobleaching (FRAP) has
been extensively used to measure translational mobility of
various fluorophores. This technique involves the photo-
bleaching of a fluorophore and observing the kinetics of the
fluorescence recovery dependent on passive diffusion and
active translational processes (protein transport) (Axelrod et
al., 1976; Icenogle and Elson, 1983; Smith et al., 1998). We
proposed that the motion effects in FRAP could be elimi-
nated by bleaching the entire cell, and be made dependent
on enzymatic activity through the resynthesis of the fluoro-
phore pool after photobleaching. This approach would re-
sult in an enzyme-dependent fluorescence recovery after
Conveniently, the NAD?/NADH ratio is large in most
cells (Erecinska et al., 1978), reducing the impact of the
NADH photobleaching on the total NADH-NAD?cellular
content. In addition, the intrinsic NADH fluorescence signal
from cells is a good candidate for ED-FRAP because this
molecule plays a critical role in energy metabolism and
numerous other redox coupled reactions in the cell (Chance
et al., 1972; Eng et al., 1989). The dehydrogenases that
generate NADH are highly regulated and compartmental-
ized enzymes involved in the effects of hormone action
(McCormack et al., 1990), as well as the orchestration of
workload with oxidative phosphorylation (Balaban, 1990).
To date, the cellular activity of dehydrogenases has been
extrapolated, indirectly, from NADH fluorescence ampli-
tude images. These studies cannot discriminate between a
change in NADH production and NADH utilization result-
ing in the net change in [NADH]. This is a general problem
for cellular fluorescence measurements of metabolic inter-
mediates. Here we show that ED-FRAP can provide a direct
measure of dehydrogenase activity, removing the ambiguity
of interpreting NADH fluorescent amplitude changes with
Preparation of cardiac myocytes and
mitochondria and experimental conditions
Cardiac myocytes were isolated from adult rabbits using standard proce-
dures (Chacon et al., 1994). Cells were resuspended after isolation in media
Received for publication 13 September 2000 and in final form 22 January
Address reprint requests to Dr. Christian A. Combs, Laboratory of Cardiac
Energetics, National Heart Lung and Blood Institute, National Institutes of
Health, Bldg. 10, Rm. B1D-416, Bethesda, MD 20892-1061. Tel.: 301-
496-0014; Fax: 301-402-2389; E-mail: firstname.lastname@example.org.
© 2001 by the Biophysical Society
2018 Biophysical JournalVolume 80April 2001 2018–2028
consisting of a 1:1 mixture of Joklik’s medium and medium 199 supple-
mented with 1 mM creatine, 1 mM carnitine, 1 mM taurine, 1 mM octanoic
acid, 10 mM Hepes, 5 mM hydroxybutyric acid, 0.05 units/ml insulin, 10
units/ml penicillin, and 10 ?g/ml streptomycin at pH 7.4. Cells were plated
onto coverslips coated with Matrigel (Becton, Dickinson, Franklin Lakes,
NJ) for attachment before each experiment. All experiments were con-
ducted within 8 h of isolation. Mitochondria were isolated as previously
described (Territo et al., 2000). Temperature was maintained at 37°C in
some experiments by blowing heated air from a commercial air-heating
system (World Precision Instruments Inc., Sarasota, FL) into a custom-
built box surrounding the microscope and by using a heated sample
chamber (Warner Instruments, Hamden, CT). Superfusion through the
sample chamber was performed by gravity-feed.
Glutamate dehydrogenase enzyme system
Equilibrium conditions were produced by mixing 5 mM ?-ketoglutarate, 5
mM NAD?, 1 mM glutamate, 50 mM Tris, and 300 ?M NH4
with varying amounts of glutamate dehydrogenase (Lowry and Passon-
neau, 1972). Using an equilibrium constant of 4.5 ? 10?14M2(Barman,
1969) the [NADH] was estimated at 1.5 ?M. Borosilicate pipettes drawn
to a tip of ?5 ?m were used to produce the micro-droplets. All experi-
ments were conducted under silicone oil to avoid evaporation.
?at pH 8.0
Image acquisition and data analysis
All fluorescence emission images of micro-droplets and cardiac myocytes
were obtained with a Zeiss LSM-510 confocal microscope system. Images
of the droplets were collected with a Fluar 10?, 0.5 NA lens. Images of the
isolated mitochondria and the cardiac myocytes were collected with a
C-Apochromat 63?, 1.2 NA, water lens. In all cases the center of the cell
or droplet was selected to image and the pinhole adjusted to produce a thin
slice (1.4–2.0 ?M in cells and mitochondria and 85 ?M in the droplets).
NADH was imaged with the 351-nm line of a UV laser and an LP 385-nm
emission filter (Eng et al., 1989). Flavin adenine dinucleotide (FAD?) was
imaged using the 488-nm line of an argon laser and a BP 505–550 nm
emission filter (Chance et al., 1972). All image processing was performed
using custom-written programs written in the IDL programming environ-
ment (RSI, Boulder, CO).
In vitro ED-FRAP validation
To validate the ED-FRAP approach, experiments were con-
ducted on small droplets (?350 ?m diameter) with gluta-
mate dehydrogenase (GDH) in an equilibrium mixture of its
substrates including NADH (?1.5 ?M). The photobleach-
ing was accomplished by bleaching the entire droplet (at
351 nm) to eliminate motion effects in the fluorescence
recovery. Images of an enzyme containing droplet before,
immediately after photobleaching, and after recovery are
shown in Fig. 1 A. An example of the time course of
fluorescent recovery is presented in Fig. 1 B. Droplets that
contained NADH in water alone without the enzyme reac-
tion components were found to bleach uniformly, but not
To quantify the NADH recovery rate versus enzyme
concentration we assumed that the exponential recovery of
NADH fluorescence is a reversible pseudo-first-order reac-
tion (NAD?7 NADH) that re-establishes steady state after
photobleaching according to the equation:
NADH ? NADHeq? ?NADHb? NADHeq??e?kNADH?).
Where kNADHis the sum of the forward (kf) and reverse (kr)
rate constants of the pseudo-first-order reaction, NADHbis
the initial fluorescence intensity after photobleaching, ? is
the time after photobleaching, NADH is the fluorescent
intensity, and NADHeqis the steady-state NADH fluores-
cence signal after bleaching. The assumption that the recov-
ery was exponential was supported by the observation that
the recovery rate could be fit to Eq. 1 (Chi-square, p ?
0.05). Because this reaction is enzyme-catalyzed, the ob-
served forward and reverse rate constants will be dependent
on the enzyme concentration (i.e., kNADH? [E]kf? [E]kr?
[E](kf? kr)). Thus, kNADHshould be proportional to the
enzyme concentration in a pseudo-first-order rate reaction.
The calculated kNADHvalue was found to be proportional to
[GDH] (R2? 0.99) in the droplets (Fig. 1 C). Varying the
degree of bleaching with light intensity or time did not
significantly change the kNADHvalue for a given enzyme
concentration. The initial rate under these conditions (1.5 ?
10?10mol/min/Unit Enzyme), estimated from the initial
linear region of the ED-FRAP recovery curve, indicated that
the enzyme was operating well below Vmaxdue to the
concentrations of metabolites used in this steady-state ex-
periment (Barman, 1969), as well as the fact that true initial
rate conditions might not have been obtained (i.e., incom-
plete photobleaching of NADH).
Effect of steady-state fluorescence imaging on
In previous studies using bright field microscopy, we found
that the NADH fluorescence signal is dominated by the
mitochondrial pool and that a slow decline in signal was
observed over time (Eng et al., 1989). In the current study
using confocal techniques, we observed a similar dynamic
phenomenon with imaging time. An initial decay in NADH
fluorescence was observed followed by an extended steady
state that was stable for hundreds of more images (Fig. 2).
This result suggested that a new steady state of NADH was
being created by the imaging procedure. This was likely due
to the competition between the optical destruction of NADH
with its metabolic production. This hypothesis was tested by
pausing the image acquisition and observing the effects on
the NADH signal. The results from one of these studies are
presented in Fig. 3. When imaging was paused, NADH
substantially recovered as metabolism regenerated NADH
without the competition of photobleaching. Over time, the
NADH signal returned to the original steady-state level.
These data are consistent with a balance between the de-
struction, via photobleaching, and generation, via metabo-
lism, resulting in the observed steady state of NADH during
FRAP Imaging of Enzyme Activity2019
Biophysical Journal 80(4) 2018–2028
imaging. This steady-state condition also suggests that the
photodamage to the NADH generation system is minimal
over the time of these experiments, because if the NADH
generating capacity was being reduced, the NADH levels
would be rapidly reduced by the background photobleach-
ing of the imaging processes itself. Further confirmation
that metabolic resynthesis of NADH was occurring is pre-
sented in the ED-FRAP experiments that follow.
NADH fluorescence photobleaching recovery
characteristics and distribution in living myocytes
Fig. 4 shows an NADH ED-FRAP experiment on a cardiac
myocyte. Again, the entire cell was photobleached to elim-
inate diffusional processes. The cell NADH fluorescence
recovered in a similar exponential function as seen in the in
vitro enzyme studies consistent with an enzymatic process
(Fig. 4, A and B) and the resynthesis of reduced NADH. The
photobleaching was found to pass through the entire cell
using the non-metabolized probe, calcein-AM, that demon-
strated out-of-plane bleaching using slower 3D imaging
procedures (data not shown). This result suggests that
through-plane diffusion of NADH was unlikely in these
studies. Photobleached cells did not exhibit signs of damage
such as blurred z-lines, spontaneous contractions, or bleb-
bing (Fig. 4 A). The photobleaching pulse had no direct
effect on mitochondrial membrane potential using tetra-
methylrhodamine methyl ester (TMRE) as a probe (data not
The effect of photobleaching magnitude on the rate of
NADH recovery was evaluated serially on individual myo-
cytes. This is presented in Fig. 5 A. As seen in Fig. 5, A and
B, the recovery rate of NADH increased with bleaching
level (ANOVA, p ? 0.05). There was an apparent nonlinear
(exponential) dependence of the rate of recovery of NADH
with the bleaching level as shown in Fig. 5 B. These data
suggest that the metabolic recovery rate is dependent on the
magnitude of the perturbation imposed and suggest that
standardized bleaching conditions will need to be used in
comparison studies. In addition, because these studies were
zyme-dependant in vitro. (A) NADH fluorescence im-
ages (85 ?m slice thickness) from a droplet containing
10 units/ml of GDH and substrates in equilibrium before
and immediately after the bleach pulse (75 ?W, ?2 ?s
dwell time/pixel), and after complete recovery. The
scale bar represents 50 ?m. (B) Representative plot of
NADH fluorescence intensity over time during an ED-
FRAP experiment from a droplet containing the GDH
enzyme system. (C) Plot of exponential rate constant
(kNADH) for recovery from photobleaching of NADH
from droplets containing variable concentrations of
GDH. n ? 5 for each [GDH].
NADH recovery is exponential and is en-
2020 Combs and Balaban
Biophysical Journal 80(4) 2018–2028
performed serially in the same myocytes, there is apparently
no significant damage to the NADH regeneration system
with the bleaching pulse. Indeed, the fact that the highest
bleaching level (i.e., highest power) resulted in the fastest
recovery rate is also inconsistent with a bleaching-related
lesion to the NADH production system.
A zoomed region of a myocyte NADH fluorescence
image is shown in Fig. 6 A. The characteristic fluorescent
patterns of mitochondria are observed localized to the re-
gions between the z-lines in the sarcomeres and densely
located in the perinuclear regions. Previous studies have
established that the mitochondrial NADH fluorescence sig-
nal dominates in this preparation (Eng et al., 1989). The
mitochondrial NADH signal is large due to concentration
and the large fluorescence enhancement associated with
protein binding in the matrix (Nuutinen, 1984; Jameson et
The kNADHvalue was calculated on a pixel-by-pixel basis
from an ED-FRAP experiment on this cell to create the
activity map in Fig. 6 B. This image demonstrates that
adequate signal-to-noise was attained to determine the dis-
tribution of this process within a single living cell. The
activity map shows the highest rates of NADH recovery
were localized to the mitochondria. NADH signal from the
cytosol (z-line regions) and nucleus were too low to quan-
To confirm that mitochondria have a significant NADH
recovery capacity, ED-FRAP experiments were conducted
on isolated porcine heart mitochondria. Fig. 7 shows an
NADH ED-FRAP experiment conducted on a dense field of
mitochondria at 25°C. The fluorescence recovery in this
isolated organelle is clearly observed, consistent with the
notion that the NADH ED-FRAP recovery is occurring in
the mitochondria. Quantitative comparisons of the mito-
chondrial kNADHand cellular rates are difficult due to the
differences in carbon substrates utilized as well as species
Because the NADH is bound in the mitochondria, intra-
cellular diffusion should not contribute to the NADH recov-
ery even if the whole cell was not irradiated. This was
confirmed by obtaining the same kNADHvalue in a small
bleached region (1⁄3 of the cell) as the whole-cell bleach
procedure (not shown). In addition, no evidence of an
increase in the NADH signal around a selectively bleached
region was observed, consistent with the notion that the
mitochondrial NADH is immobilized. One of the concerns
in interpreting the mitochondrial NADH ED-FRAP recov-
ery is the fact that NADH immobilization enhances the
fluorescence so markedly (Nuutinen, 1984; Jameson et al.,
1989). The kinetics of this binding process could contribute
to the ED-FRAP rate constant if binding is much slower that
the enzymatic rates. The contribution of the binding kinetics
will be discussed further below.
Cellular NADH fluorescence photobleaching
recovery rate is sensitive to temperature and
To further evaluate whether the NADH ED-FRAP rates in
the myocytes were dependent on enzymatic processes, ex-
periments were conducted to examine the effects of tem-
myocytes. Representative plot of NADH fluorescence intensity over time
at a rate of imaging of 3 Hz. Intensity values are the mean of all pixel
values greater than the background fluorescence in the myocyte. Note that
there is an initial steep decrease followed by a slow constant decline in
fluorescence intensity where NADH bleaching rate is balanced by meta-
bolic generation of NADH.
NADH fluorescence over time in isolated perfused cardiac
of NADH over time in an isolated perfused cardiac myocyte during image
acquisition as evidenced by regeneration of NADH during pauses in the
imaging experiment. Plot of NADH fluorescence intensity over time at a
rate of imaging of 3 Hz with two pauses. Intensity values are the mean of
all pixel values greater than the background fluorescence in the myocyte.
Note that NADH substantially recovers during the two pauses and follows
an exponential decay similar to the initial decay upon resumption of
Balance between photobleaching and metabolic generation
FRAP Imaging of Enzyme Activity2021
Biophysical Journal 80(4) 2018–2028
perature and dehydrogenase activation with dichloroacetic
Stacpoole, 1997). Fig. 8 shows the average kNADHvalue in
cardiac myocytes as a function of these treatments. Raising
the temperature by 12°C increased the rate by more than a
factor of 2.5. This is consistent with the effect of tempera-
ture on the activity of most enzymes (Hochachka and Som-
ero, 1984). However, as discussed above, the temperature
dependence of NADH binding to matrix proteins could have
a similar temperature dependence.
DCA increased average kfby factors of 3 and 1.5 at 25
and 37°C, respectively. Theophylline was used to simulate
hormone activity by increasing cellular cyclic adenosine
monophosphate (cAMP) levels. Theophylline increased kf
by 2.5-fold (Fig. 8). These increases in kNADHare similar to
the enhancement of pyruvate dehydrogenase activity, deter-
mined in vitro, caused by these agents in intact hearts
(Bersin and Stacpoole, 1997; Depre et al., 1998). These later
data are consistent with kNADHbeing proportional to the
enzyme activity in the mitochondria and not other physical
processes, such as NADH protein binding, as it is unlikely
that both of these agents would affect NADH binding rates.
The distribution of the effect of DCA is presented in Fig.
6, C and D, where a reference NADH intensity image (Fig.
6 C) and a pixmap of the difference in kNADHvalues before
and after DCA (Fig. 6 D) are presented. Fig. 6 D shows a
clearly punctate pattern of activation by DCA. These
changes correlate with the distribution of mitochondria
based on NADH fluorescence correlation with mitochon-
dria-specific dyes and metabolic perturbations (Eng et al.,
1989). This image demonstrates that adequate signal-to-
noise is attained in this approach to image the difference in
dehydrogenase activity within a single cell during experi-
Effects of NADH photobleaching on
Another internal verification of the enzymatic requirement
for NADH recovery after photobleaching is the FAD?
images (2 ?M slice thickness) from a myocyte before photobleaching (75 ?W, ?2 ?s dwell time/pixel), immediately after, and following recovery.
Obvious structures that can be seen in the cardiac myocyte are the nucleus (large dark spot in the left center) and the z-lines of the sarcomeres (repeating
dark lines). The scale bar represents 10 ?m. (B) Representative plot of NADH fluorescence intensity recovery after a bleach pulse. Intensity values are the
mean of all pixel values greater than the background fluorescence in the myocyte. Note that the recovery is exponential and similar to the recovery curve
in Fig. 1. .
NADH fluorescence recovery and cellular viability after photobleaching in isolated perfused cardiac myocytes. (A) NADH fluorescence
2022 Combs and Balaban
Biophysical Journal 80(4) 2018–2028
fluorescence response. FAD?and NAD?redox states are
linked via the NADH ubiquinone reductase, pyruvate dehy-
drogenase, and other dehydrogenase reactions. Predictably,
with NADH depletion, the FADH pool should also be
reduced, as it is used to regenerate NADH through these
reversible reactions. In contrast to NADH, the oxidized
form of FAD?is fluorescent, not the reduced form (i.e.,
FADH), resulting in a decrease in fluorescence intensity
with increasing level of FADH (Chance et al., 1972; Kunz
and Kunz, 1985; Romashko et al., 1998). Thus, the fluores-
cence intensities from these two metabolite pools change
inversely with increasing reduction level. To test whether
the NADH and FAD?pools were enzymatically linked, we
monitored the NADH and FAD?fluorescence signal recov-
ery after a NADH photobleaching pulse in a single cell. An
example of these studies is presented in Fig. 9. A transient
increase in FAD?fluorescence was observed that mirrored
the decrease in NADH fluorescence in response to the
NADH-directed photobleaching pulse. These data are con-
sistent with the enzymatic coupling of the FAD?and
NADH metabolite pools as probed using the ED-FRAP
approach. Finally, the similar kinetics of FAD?enhance-
ment and NADH recovery after NADH photobleaching also
suggest that the binding kinetics of NADH to proteins,
enhancing the NADH fluorescence signal, are not dominat-
ing the NADH fluorescence recovery kinetics.
These data demonstrate that NADH ED-FRAP can be used
to image the distribution of enzyme-dependent reaction
rates in vitro and in intact cells. The correlation between
NADH ED-FRAP kNADHand enzyme content (Fig. 1 C)
confirms that quantitative enzymatic activity can be deter-
mined using this approach in vitro. The temperature, DCA,
and theophylline sensitivity of the kNADHand the correlation
with FAD?fluorescence levels found in cardiac myocytes
cence of NADH is proportional to bleach level in
isolated perfused cardiac myocytes. (A) Plots of
NADH fluorescence intensity recovery after
bleach pulses designed to produce different de-
grees of photobleaching in repeated measurements
on a single myocyte. Intensity values are the mean
of all pixel values greater than the background
fluorescence in the myocyte. (B) Mean and SEM
values for the exponential rate constant (averaged
over the entire cell volume) of recovery after
photobleaching versus treatments designed to pro-
duce three different levels of photobleaching (five
iterations of ?38 (curve 1) or 75 ?W (curve 2), or
10 iterations of 75 ?W (curve 3) at a wavelength
of 351 nM and a dwell time of 2 ?s/pixel) for
superfused cardiac myocytes. All bleach levels
and recovery rates were significantly different (re-
peated measures ANOVA, SNK, p ? 0.05). The
line represents an exponential fit of recovery rate
versus bleach level (R2? 0.99).
Rate of recovery of the autofluores-
FRAP Imaging of Enzyme Activity 2023
Biophysical Journal 80(4) 2018–2028
support the notion that ED-FRAP is determining the en-
zyme-dependent NADH generating capacity of regions
within the cell.
In this initial study we have not attempted to determine
the initial rate of NADH formation (i.e., [E]*kf) because we
have not been able to attain complete removal of the NADH
pool with photobleaching. Attempts to perform complete
photobleaching with higher power or longer times of irra-
diation resulted in irreversible cell damage based on gross
structural changes. This damage could be due to direct
photodamage of the cell or depletion of the total NAD?/
NADH pool as the NADH is rapidly cycled through the
reaction and photobleaching process. This latter effect was
observed in the enzyme system in vitro with long irradiation
times. However, no evidence of direct GDH destruction was
observed. We selected a photobleaching level of ?30% in
these studies based on empirical observations that this pro-
vided a minimal perturbation but adequate signal-to-noise to
A surprising result was the dependence of kNADHon the
bleaching level in the cells, getting faster with larger de-
creases in NADH (Fig. 5). This was not observed in the
isolated enzyme system. The increasing recovery rate with
NADH bleaching level suggests that bleaching is not sig-
nificantly destroying the dehydrogenases, because just the
opposite would occur. The fact that the rate of recovery is
dependent on the magnitude of the displacement from the
steady state could be due to thermodynamic effects on the
metabolic driving forces (i.e., NADH/NAD (Wilson, 1994))
or direct negative-allosteric effects of NADH on the dehy-
drogenases (Patel and Roche, 1990)). These possibilities
need to be explored further with regard to the quantitation of
ED-FRAP rates and the regulation of cellular NADH pro-
duction. Without a complete model it is clear that the
images (1.5 ?M slice thickness). (B) False color pixmap of exponential recovery rate ? 10?3s?1. The NADH intensity image was smoothed over two pixels
before exponential fitting (?1 ?m in plane and z-resolution). Only pixels that passed a chi-square distribution function were fit. Temporal resolution was
0.49 s for exponential fits. (D) Pixmap of the difference in exponential recovery rate between superfusion with nutrient medium and 5 mM DCA of the
cell depicted in (C). Imaging and photobleaching parameters are the same as in Fig. 1.
Pixmaps of NADH recovery rate reveal regional dehydrogenase activity in isolated cardiac myocytes. (A and C) NADH fluorescence intensity
2024 Combs and Balaban
Biophysical Journal 80(4) 2018–2028
perturbation, or bleaching level, must be consistent between
experimental conditions to access differences in NADH
One other interesting consequence of the NADH ED-
FRAP is what it means in the evaluation of NADH using
standard fluorescence microscopy techniques. Many in-
vestigators have assumed that the bleaching of NADH is
minimized in studies where the NADH fluorescence sig-
nal is nearly constant or slowly drifts down during an
experimental observation period (Eng et al., 1989). How-
ever, this apparent constant NADH signal is reflecting
the balance between the photobleaching destruction and
enzymatic production of NADH described in the current
study (Fig. 3). Similar concerns may exist for steady-
state fluorescence studies on FAD?and other metaboli-
cally active markers. Potentially, with a more complete
model of NADH bleaching and production in these types
of experiments, the approach to steady state alone in an
imaging experiment may provide additional information
on NADH production.
One of the most important facets of ED-FRAP is that it
can provide significantly more information than steady-state
metabolite measurements usually obtained in conventional
studies. Steady-state measurements of metabolite fluores-
cence are complicated by the inability to correlate mecha-
nism to the observed concentration changes. For instance,
an increase in cellular NADH fluorescence could be a
function of decreased NADH oxidation or increased NAD?
reduction by metabolism. In contrast, an increase in ED-
FRAP rates of recovery must be ascribed to either an
increase in dehydrogenase activity and/or increases in rate
limiting substrate concentrations in the region of the en-
zyme. Thus, ED-FRAP provides much more information
than the determination of concentration alone.
mitochondria before photobleaching (75 ?W, ?2 ?s dwell time/pixel), immediately after, and following recovery. The box represents the bleaching area.
The scale bar represents 10 ?m. (B) Plot of NADH fluorescence intensity recovery after a bleach pulse. Intensity values are the mean of all pixel values
greater than the background fluorescence. Note that the recovery is exponential and similar to the recovery curve in Figs. 1 and 2 .
NADH fluorescence recovery is exponential in isolated perfused mitochondria. (A) NADH fluorescence images (1.4 ?M slice thickness) from
FRAP Imaging of Enzyme Activity 2025
Biophysical Journal 80(4) 2018–2028
As discussed earlier, these data are consistent with the
notion that kNADHfrom ED-FRAP of NADH in cardiac
myocytes is correlated with mitochondrial dehydrogenase
activity based on its localization in the mitochondria. How-
ever, it is unclear which mitochondrial dehydrogenase sys-
tem(s) contributes to the NADH recovery rate. Based on in
vitro activity measurements (Table 1) many dehydrogenases
could contribute to the ED-FRAP rate. Based on these
activities determined in vitro, malate dehydrogenase has the
largest activity and can enhance the fluorescence of NADH
(Jameson et al., 1989). However, DCA and theophylline
were found to increase kNADH?3-fold at 25°C in this study.
DCA and cAMP have been shown to activate pyruvate
dehydrogenase (PDH) (Sugden and Holness, 1994; Bersin
and Stacpoole, 1997). These data suggest that PDH may
also play an important role in NADH ED-FRAP despite its
low in vitro activity (Table 1). These hormone effects could
also be affecting the net flux through several upstream and
downstream metabolic steps kinetically via substrate and
product concentrations, as well as a variety of allosteric
effects. Thus, localizing these effects to one dehydrogenase
system is difficult. The reciprocal effects of NADH photo-
bleaching on FAD?(Fig. 7) suggest that ED-FRAP is
interrogating an NADH pool enzymatically coupled to the
FAD?/FADH system, implying that the pool is in general
metabolic communication within the matrix and not an
isolated pool. Though more work will be required to iden-
tify the specific enzymes and bound pools of NADH par-
ticipating in the ED-FRAP reaction, it is apparent that
ED-FRAP is monitoring one aspect of the delivery of the
primary energy source, NADH reducing equivalents, to the
electron transport system.
Given these considerations it is clear that ED-FRAP
could contribute to the understanding of the cellular regu-
latory processes involved in energy metabolism in living
SEM values for the exponential rate constant (averaged over the entire cell volume) of recovery after photobleaching versus treatment for superfused cardiac
myocytes. Control at the two temperatures stands for superfusion rates of ?3 ml/min with nutrient medium (n ? 11 and 12 at 25 and 37°C, respectively).
DCA at the two temperatures stands for superfusion with 5 mM DCA and 5 mM pyruvate added to the nutrient medium (n ? 5 at each temperature).
Theo 25°C stands for superfusion with theophylline (17 ?g/ml). The combination of DCA and pyruvate is reported as a potent stimulator of
dehydrogenase activity. The lines at the top of the panel indicate significant differences in mean values between treatments (ANOVA, Tukey, p ? 0.05).
The NADH recovery exponential rate constant is temperature- and dehydrogenase activity level-sensitive in cardiac myocytes. Mean and
veals inverse metabolic coupling with NADH bleaching. Plot of NADH
(open circles) and FAD?(closed circles) fluorescence intensity over time
from a single cardiac myocyte. The drop in NADH fluorescence around
29 s is due to the photobleaching pulse (at 351 nm) that is selective for
NADH. The FAD?(flavin adenine nucleotide) signal (excitation 488 nm,
emission 520 nm (Chance et al., 1972) increases immediately after the
NADH photobleaching. Intensity values are the mean of all pixel values
greater than the signal from the nucleus.
Dual acquisition of NADH and FAD?autofluorescence re-
2026Combs and Balaban
Biophysical Journal 80(4) 2018–2028
cells in ways that were previously not possible. Reducing
equivalent delivery is likely a critical factor in the regulation
of oxidative phosphorylation in the heart via the modifica-
tion of the maximum ATP production rate (Hansford, 1980;
Balaban, 1990). Additionally, ED-FRAP may play an im-
portant role in examining the relationship between Ca2?and
dehydrogenase activity within living cells, which has been
shown to have a large impact on substrate utilization and
ATP production in isolated mitochondria (Hansford, 1980).
This may be particularly relevant in regard to determining
the importance of the association of mitochondria with other
intracellular organelles, like the sarcoplasmic reticulum,
that histological studies have shown varies regionally
(Scales, 1983; Shimada et al., 1984). Recent studies have
shown that these associations may play a critical role in
calcium buffering and sequestration (Rutter and Rizzuto,
2000). Studies have also suggested a metabolic difference
between different cellular pools of heart mitochondria, such
as the subsarcolemma and intrafibrillar mitochondria
(Palmer et al., 1985). Again, ED-FRAP may provide a
direct evaluation of these proposed metabolic differences
within an intact cell.
ED-FRAP measurements may not be limited to NADH
linked reactions alone. We have already shown that even
indirect responses to the photobleaching can be detected in
FAD?-linked reactions (see Fig. 8). In addition, specific
fluorescence markers of enzymatic modifications of metab-
olites could be used to monitor other reaction mechanisms
inside cells. In these applications, the substrates must be
converted between a fluorescent species and a molecule
with different absorption and fluorescence spectral proper-
ties to selectively eliminate a product or a substrate using
the photobleaching approach. If exogenous tracer probes are
used, the effect of bleaching on probe concentration will not
be as critical as naturally occurring metabolites like NADH
In summary, we have demonstrated that NADH ED-
FRAP provides a new tool for unraveling the role of NADH
generation in the regulation of mitochondria ATP produc-
tion, as well as potential new insights into complexities of
cellular compartmentation in this process.
We thank Drs. Anthony Aletras and Han Wen for many useful discussions
and assistance in IDL programming. We also thank Stephanie French and
Raimundo Correa for preparation of porcine mitochondria and rabbit
ventricular myocytes, respectively.
Adisa, A. O., and A. A. Odutuga. 1998. Changes in the activities of three
diagnostic enzymes in the heart of rats following the consumption of
diets deficient in zinc and essential fatty acids. Biochem. Mol. Biol. Int.
Axelrod, D., D. E. Koppel, J. Schlessinger, E. Elson, and W. W. Webb.
1976. Mobility measurement by analysis of fluorescence photobleaching
recovery kinetics. Biophys. J. 16:1055–1069.
Balaban, R. S. 1990. Regulation of oxidative phosphorylation in the
mammalian cell. Am. J. Physiol. Cell Physiol. 258:C377–C389.
Balaban, R. S., V. K. Mootha, and A. A. Arai. 1996. Spectroscopic
determination of cytochrome c oxidase content in tissues containing
myoglobin or hemoglobin. Anal. Biochem. 237:274–278.
Barman, T. E. 1969. Enzyme Handbook. Springer-Verlag, New York.
Bersin, R. M., and P. W. Stacpoole. 1997. Dichloroacetate as metabolic
therapy for myocardial ischemia and failure. Am. Heart J. 134:841–855.
Chacon, E., J. M. Reece, A. L. Neiminen, G. Zahrebelski, B. Herman, and
J. L. Lemasters. 1994. Distribution of electrical potential, pH, free Ca2?,
and volume inside cultured adult rabbit cardiac myocytes during chem-
ical hypoxia: a multiparameter digitized confocal microscopic study.
Biophys. J. 66:942–952.
Chance, B., I. A. Salkovitz, and A. G. Kovach. 1972. Kinetics of mito-
chondrial flavoprotein and pyridine nucleotide in perfused heart.
Am. J. Physiol. 223:207–218.
Depre, C., S. Ponchaut, J. Deprez, L. Maisin, and L. Hue. 1998. Cyclic
AMP suppresses the inhibition of glycolysis by alternative oxidizable
substrates in the heart. J. Clin. Invest. 101:390–397.
Eng, J., R. M. Lynch, and R. S. Balaban. 1989. Nicotinamide adenine
dinucleotide fluorescence spectroscopy and imaging of isolated cardiac
myocytes. Biophys. J. 55:621–630.
Erecinska, M., D. F. Wilson, and K. Nishiki. 1978. Homeostatic regulation
of cellular energy metabolism: experimental characterization in vivo and
fit to a model. Am. J. Physiol. Cell Physiol. 234:C82–C89.
Genova, M. L., C. Castelluccio, R. Fato, C. G. Parenti, P. M. Merlo, G.
Formiggini, C. Bovina, M. Marchetti, and G. Lenaz. 1995. Major
changes in complex I activity in mitochondria from aged rats may not be
detected by direct assay of NADH:coenzyme Q reductase. Biochem. J.
Hansford, R. G. 1980. Control of mitochondrial substrate oxidation. Curr.
Top. Bioenerg. 10:217–278.
Hochachka, P. W., and G. N. Somero. 1984. Biochemical Adaptation.
Princeton University Press, Princeton, NJ.
literature for rat cardiac muscle
Mitochondrial enzyme activities obtained from the
(U/g wet weight)Reference
Genova et al.,
256* Link et al.,
*Converted from U/mg protein in homogenized to U/g wet weight by
multiplication by 140 mg protein/g wet weight (Balaban et al., 1996).
†Converted from U/mg mitochondrial protein by multiplication by 36 mg
mitochondrial protein/g wet weight (Balaban et al., 1996).
FRAP Imaging of Enzyme Activity2027
Biophysical Journal 80(4) 2018–2028
Icenogle, R. D., and E. L. Elson. 1983. Fluorescence correlation spectros-
copy and photobleaching recovery of multiple binding reactions. I.
Theory and FCS measurements. Biopolymers. 22:1919–1948.
Jameson, D. M., V. Thomas, and D. M. Zhou. 1989. Time-resolved
fluorescence studies on NADH bound to mitochondrial malate dehydro-
genase. Biochim. Biophys. Acta. 994:187–190.
Kunz, W. S., and W. Kunz. 1985. Contribution of different enzymes to
flavoprotein fluorescence of isolated rat liver mitochondria. Biochim.
Biophys. Acta. 841:237–246.
Link, G., A. Saada, A. Pinson, A. M. Konijn, and C. Hershko. 1998.
Mitochondrial respiratory enzymes are a major target of iron toxicity in
rat heart cells. J. Lab. Clin. Med. 131:466–474.
Lowry, O. H., and J. V. Passonneau. 1972. A flexible system of enzymatic
analysis. Academic Press, New York.
Lucas, D. T., and L. I. Szweda. 1999. Declines in mitochondrial respiration
during cardiac reperfusion: age-dependent inactivation of alpha-
ketoglutarate dehydrogenase. Proc. Natl. Acad. Sci. USA. 96:
McCormack, J. G., A. P. Halestrap, and R. M. Denton. 1990. Role of
calcium ions in regulation of mammalian intramitochondrial metabo-
lism. Physiol. Rev. 70:391–425.
Nagai, Y., M. Miyazaki, R. Aoki, T. Zama, S. Inouye, K. Hirose, M. Iino,
and M. Hagiwara. 2000. A fluorescent indicator for visualizing cAMP-
induced phosphorylation in vivo. Nat. Biotechnol. 18:313–316.
Nuutinen, E. M. 1984. Subcellular origin of the surface fluorescence of
reduced nicotinamide nucleotides in the isolated perfused rat heart. Bas.
Res. Card. 79:49–58.
Palmer, J. W., B. Tandler, and C. L. Hoppel. 1985. Biochemical differ-
ences between subsarcolemmal and interfibrillar mitochondria from rat
cardiac muscle: effects of procedural manipulations. Arch. Biochem.
Patel, M. S., and T. E. Roche. 1990. Molecular biology and biochemistry
of pyruvate dehydrogenase complexes. FASEB J. 4:3224–3233.
Poston, J. M., and G. L. Parenteau. 1992. Biochemical effects of ischemia
on isolated, perfused rat heart tissues. Arch. Biochem. Biophys. 295:
Romashko, D. N., E. Marban, and B. O’Rourke. 1998. Subcellular meta-
bolic transients and mitochondrial redox waves in heart cells. Proc. Natl.
Acad. Sci. USA. 95:1618–1623.
Rutter, G. A., and R. Rizzuto. 2000. Regulation of mitochondrial metab-
olism by ER Ca2?release: an intimate connection. Trends. Biochem. Sci.
Saks, V. A., Y. O. Belikova, and A. V. Kuznetsov. 1991. In vivo regulation
of mitochondrial respiration in cardiomyocytes: specific restrictions for
intracellular diffusion of ADP. Biochim. Biophys. Acta. 1074:302–311.
Scales, D. J. III. 1983. Three-dimensional electron microscopy of mam-
malian cardiac sarcoplasmic reticulum at 80 kV. J. Ultrastruct. Res.
Seymour, A. M., and J. C. Chatham. 1997. The effects of hypertrophy and
diabetes on cardiac pyruvate dehydrogenase activity. J. Mol. Cell Car-
Shimada, T., K. Horita, M. Murakami, and R. Ogura. 1984. Morphological
studies of different mitochondrial populations in monkey myocardial
cells. Cell Tissue Res. 238:577–582.
Shonk, C. E., and G. E. Boxer. 1964. Enzyme patterns in human tissue. 1.
Methods for the determination of glycolytic enzymes. Can. Res. 24:
Smith, C. L., K. J. Zaal, and J. Lippincott-Schwartz. 1998. Looking at the
bright side of photobleaching: analysis of intracellular protein dynamics
by confocal microscopy. Scanning. 20:147–148.
Srivastava, D. K., and S. A. Bernhard. 1987. Biophysical chemistry of
metabolic reaction sequences in concentrated enzyme solution and in the
cell. Annu. Rev. Biophys. Biophys. Chem. 16:175–204.
Sugden, M. C., and M. J. Holness. 1994. Interactive regulation of the
pyruvate dehydrogenase complex and the carnitine palmitoyltransferase
system. FASEB J. 8:54–61.
Territo, P. R., V. K. Mootha, S. A. French, and R. S. Balaban. 2000. Ca2?
activation of heart mitochondrial oxidative phosphorylation: role of the
F0/F1-ATPase. Am. J. Physiol. Cell Physiol. 278:C423–C435.
Wilson, D. F. 1994. Factors affecting the rate and energetics of mitochon-
drial oxidative phosphorylation. Med. Sci. Sports Exerc. 26:37–43.
Zonderland, M. L., P. R. Bar, J. C. Reijneveld, B. M. Spruijt, H. A. Keizer,
and J. F. Glatz. 1999. Different metabolic adaptation of heart and
skeletal muscles to moderate-intensity treadmill training in the rat. Eur.
J. Appl. Physiol. 79:391–396.
2028 Combs and Balaban
Biophysical Journal 80(4) 2018–2028