The Plant Journal (1998) 13(6), 803–814
Early stages of the apoptotic pathway in plant cells are
reversible
Iona E. W. O’Brien1,*, Bruce C. Baguley2, Brian G. Murray3,
Bret A. M. Morris1 and Ian B. Ferguson1
1Horticulture and Food Research Institute of New
Zealand, Private Bag 92169, Auckland, New Zealand,
2Cancer Research Laboratory, Faculty of Medicine and
Health Sciences, University of Auckland, Private Bag
92019, Auckland, New Zealand, and
3School of Biological Sciences, University of Auckland,
Private Bag 92019, Auckland, New Zealand
Summary
Chromatin condensation and nDNA fragmentation, indic-
ators of apoptosis in mammalian cells, occur in plant cells
during senescence and following induction by chemical
agents. In Nicotiana plumbaginifolia cells, camptothecin,
okadaic acid, salicylic acid, hydrogen peroxide, and the
calcium ionophore A23187 induced chromatin condensa-
tion and nDNA fragmentation. Exposure of cells to low
concentrations or removal of the chemical agent resulted
in an initial phase of chromatin condensation, followed by
its reversal. A further feature of apoptosis in mammalian
cells, annexin V binding, indicative of phosphotidylserine
exposure, was also confirmed in relation to the other
events in the apoptotic pathway. With respect to flow
cytometric characteristics, apoptosis triggered by a variety
of chemicals occurs in plant cells in a manner closely
related to that in mammalian cells. However, the extent
of chromatin condensation is substantially greater, and in
the early stages is reversible.
Introduction
Programmed cell death (PCD) is a genetically programmed
process involving a number of regulatory genes, stimula-
tory events and signalling pathways, and a range of mor-
phological features (Earnshaw, 1995; Hale et al., 1996;
Martin et al., 1994). It is distinct from necrosis, the latter
being cell death as a response to severe trauma, and not
genetically based. The distinctive characteristics of necrosis
centre around collapse of membrane function and cell
metabolism, manifest as cell swelling, loss of compart-
mentation and selective permeability of the plasma mem-
brane, and no obvious early changes in DNA (Darzynkiewicz
Received 11 June 1997; revised 4 November 1997; accepted 22 December
1997.
*For correspondence: (fax 1 64 98154201;
e-mail iweir@hort.cri.nz; iferguson@hort.cri.nz).
© 1998 Blackwell Science Ltd
et al., 1992; Dive et al., 1992). The general term apoptosis
has been used to describe a relatively ordered set of events
in cells undergoing PCD, the main morphological features
of which, in mammalian cells, are chromatin condensation,
cell shrinkage, systematic DNA cleavage with disintegra-
tion of the nucleus and fragmentation into discrete apop-
totic bodies, and ultimately cell death (Martin et al., 1994).
While input into the apoptotic pathway is likely to be
multichannelled, the features of chromatin condensation
and nDNA fragmentation in particular are central, and
relatively consistent in mammalian cells (Wertz and Hanley,
1996). These morphological features can be characterized
by various microscopic techniques, and particularly by
flow cytometry. Staining of permeabilized cells or nuclei
with propidium iodide (PI) results in a level of fluorescence
indicative of the extent of chromatin condensation.
Apoptosis in mammalian cells is characterized by a reduc-
tion in fluorescence as the tighter coiling of DNA results
in less PI intercalatory binding. DNA strand breakage can
be assessed by specific labelling with, for instance, fluor-
escein-dUTP catalyzed by terminal deoxynucleotidyl trans-
ferase (TdT). These cytometric techniques have proved
crucial in linking concomitant and sequential events in the
process (Darzynkiewicz et al., 1992).
Apart from the morphological and nuclear changes,
apoptosis can be characterized by the involvement of cell
signalling events such as an increase in free cytosolic
calcium concentration, an oxidative burst manifest as an
increase in free oxygen radical activity and consequent
increase in reduced glutathione, a collapse in mitochondrial
membrane potential, and increased exposure of the acidic
phospholipid phosphatidylserine (PS) on the outer surface
of the plasma membrane (Hale et al., 1996). Binding of
annexin V to exposed PS has proved a reliable and useful
indicator of apoptosis in both mammalian and plant cells
(O’Brien et al., 1997).
There is increasing evidence that programmed cell death
occurs in plants (Pennel and Lamb, 1997). This has been
demonstrated for the hypersensitive response associated
with disease resistance, with the description of Arabidopsis
mutants for ACD (accelerated cell death) genes (Greenberg
et al., 1994) and of mutants which developed necrotic
lesions in the absence of a pathogen, which were associ-
ated with characteristic markers of systemic acquired resist-
ance response (Dietrich et al., 1994). More recently, further
instances of programmed cell death have been described,
including developmental changes associated with root
growth and xylem differentiation (Groover et al., 1997;
Kossiak et al., 1997; Mittler and Lam, 1995; Wang et al.,
803

804 Iona E.W. O’Brien et al.
1996), and in carrot cells in response to low level exposure
to abiotic stresses (McCabe et al., 1997).
Some morphological features of apoptosis characteristic
of mammalian cells have also been observed in plant cells
(Pennell and Lamb, 1997). The evidence to date centres larg-
ely on DNA fragmentation detected by DNA laddering and
TdT-mediated dUTP in situ labelling in plant cells responding
to fungal infection or phytotoxin exposure (Ryerson and
Heath, 1996; Wang et al., 1996), in root cells (Kossiak et al.,
1997; Wang et al., 1996), and cells responding to abiotic
stress (Katsuhara, 1997; McCabe et al., 1997). However,
there is no system yet described which shows all the features
common to mammalian cell apoptosis. This is particularly
the case with chromatin condensation. Chromatin con-
densation has frequently been assayed by PI-fluoresence in
a number of plant tissues, and we have recently shown in
tobacco cells that it is a precursor to the onset of nucleosomal
DNA fragmentation as detected by both microscopy and
flow cytometry (O’Brien et al., 1998b). However, there is a
need to assess whether the various characteristics of mam-
malian apoptosis are present concomitantly in plants and, if
so, whether they may be linked in a pathway which has
similarities to that in mammalian cells.
We have used flow cytometry with suspension-cultured
tobacco cells to investigate whether chromatin condensa-
tion and DNA fragmentation can be associated with
apoptosis in plants, both in senescing cells, and in those
© Blackwell Science Ltd, The Plant Journal, (1998), 13, 803–814
1365313x, 1998, 6, Downloaded from https://onlinelibrary.wiley.com/doi/10.1046/j.1365-313X.1998.00087.x by Egyptian National Sti. Network (Enstinet), Wiley Online Library on [28/06/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Figure 1. Micrographs of normal (a) or
apoptotic (b) protoplasts as seen under light
microscopy show the normal protoplasts to
be cytoplasmically dense, in contrast to the
vacuolated apoptotic protoplasts in which
the nucleus has becomes more distinct as
the cytoplasm becomes more vacuolated.
Staining of normal (c) or apoptotic (d)
protoplasts with FDA shows plasma
membrane integrity in contrast to necrotic
(e) protoplasts which have lost membrane
integrity as shown by high background
fluorescence and altered shape. Apoptotic
protoplasts (70% chromatin condensation
as determined by PI fluorescence and flow
cytometry) when stained with ethidium
bromide (g), show nuclear marginalisation
onto the nuclear membrane with chromatin
loss resulting in the enlarged nucleolus
becoming prominent. Control protoplasts
show dense nuclear staining (f). Micro-
graphs (a,b,c,d,f,g) are at 10003 magnifica-
tion, with (e) at 4003 magnification. Cells
were viewed using a Leitz Fluovert
microscope.
exposed to chemical agents. In the latter case, we have
investigated whether the calcium and oxidative signalling
pathways may be associated with apoptosis in plants, using
the calcium ionophore A23187 and calcium indicator Indo-
1, and hydrogen peroxide and salicyclic acid as stimulants,
with reduced glutathione levels in the cell as a measure of
response. Some of these agents have been shown to induce
cell death with some of the physical features of programmed
cell death, such as cell shrinkage (Levine et al., 1996; McCabe
et al., 1997). In addition, as part of our approach to align
various indicators of apoptosis, we have used the marker
annexin V to test for exposure of PS at the outer surface of
the plasma membrane. The latter is a morphological feature
of apoptosis in mammalian cells. We have previously shown
that annexin V binding occurs in plant cells undergoing pro-
gammed cell death induced by camptothecin, at a stage prior
to DNA fragmentation (O’Brien et al., 1997). In the course of
our work, we have found in plant cells that the fluorescence
shifts obtained with PI labelling are markedly greater than
found in mammalian cells, and that chromatin condensation
may be reversible during the early stages of apoptosis.
Results
Chromatin condensation and DNA fragmentation
Under light microscopy (Figure 1a,b), the cytoplasm of
control (untreated) protoplasts was dense and evenly dis-

Figure 2. Cells were sampled at progressive stages of the growth cycle,
protoplasts prepared, nuclei isolated and stained with PI.
Both cell division (G2 only) and chromatin condensation were measured
(see Experimental procedures). At day 10, cells were subcultured and
resampled at 14 days (f). A replicate flask was allowed to continue (e).
Values are the mean measurements from 6 replicate growth cycles, bars
represent SEs. Histograms of PI fluorescence (a–f) are shown at the various
sampling times during cell growth. These represent data from one of the
measured cycles. An increase in chromatin condensation is depicted as a
shift to the left (reduction in fluorescence) of the first (G0/G1) peak, and an
increase in cell division is displayed as an increase in the height of the
second peak (G2/M, channel number 926, b). An apoptotic peak (subdiploid)
indicating nDNA fragmentation was detected when cells were not
subcultured (channel number 72, e). Numbers above the peaks denote
fluorescence channel number. STD indicates measurement of fluorescence
of fluorospheres used as a standard for calibration.
tributed, with nuclei not especially obvious. Protoplasts at
an early stage of chromatin condensation (induced by the
addition of 130 µM H2O2) showed a more defined nucleus.
When protoplasts were stained with FDA, necrotic proto-
plasts (induced by the addition of 260 µM H2O2) showed a
loss of turgor and appeared to lose esterases resulting in
fluorescence of the media, whereas control and apoptotic
protoplasts showed retention of the dye in the cytoplasm
and nucleus, and functional plasma membrane structure
(Figure 1c,d,e). Fluorescence from ethidium bromide stain-
© Blackwell Science Ltd, The Plant Journal, (1998), 13, 803–814
1365313x, 1998, 6, Downloaded from https://onlinelibrary.wiley.com/doi/10.1046/j.1365-313X.1998.00087.x by Egyptian National Sti. Network (Enstinet), Wiley Online Library on [28/06/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Early stages of plant apoptosis are reversible 805
ing of tobacco nuclei from protoplasts with no chromatin
condensation (as determined by flow cytometry) was
evenly distributed across the nucleus (Figure 1f). However,
at a level of 70% chromatin condensation (see below),
nuclear marginalisation of the chromatin was evident,
together with the presence of the nucleolus as a central
core (Figure 1g).
To follow chromatin changes during the growth cycle,
cells were sampled at various stages, from time of subcul-
turing through to cessation of cell division and subsequent
subculturing (Figure 2). Over the first 3 days there was an
increase in the percentage of cells in the G2 phase, which
was accompanied by an increase in relative fluorescence,
suggesting a reduction in the percentage of chromatin
condensation (Figure 2a,b). This is shown by a shift in the
G0-G1 and G2 fluorescence peaks. A high mean channel
number (given above the peaks in Figure 2) indicates
greater intensity of fluorescence (see also Figure 3) and,
in this case, more PI binding and lower levels of chromatin
condensation. At the same time, there was an increase in
cell numbers in the G2 peak, and an increase in fluores-
cence, indicating again an opening up of the chromatin
with greater PI binding. By day 6, which was at the postlog
stage, cell division had declined and there was a marked
increase in chromatin condensation. On day 7, half the
cells were subcultured (Figure 2c), and the cycle from day
0 to day 6 was repeated (this was done six times for the
data in Figure 2). If the cells were left to grow on, by day
10 (Figure 2d) cell division dropped to less than 2% and
chromatin condensation reached 40% (shown by a further
decrease in channel number of the first peak). Subculturing
at this stage stimulated cell division and, correspondingly,
the percentage of condensation dropped. At day 14 in the
culture which was left to grown on, cell division ceased
and flow cytometric analysis showed a new fluorescence
peak (Figure 2e, channel number 72), recognised in mam-
malian cells as a sub-G1 or apoptotic peak (Hotz et al.,
1994), and as also previously identified in our earlier work
using both TUNEL assay and confocal microscopy (O’Brien
et al., 1997, 1998).
To relate DNA fragmentation with chromatin condensa-
tion, cells were sampled at increasing times from sub-
culture (days 1, 7 and 14), and single-strand DNA breaks
assayed by the TUNEL assay. With increasing levels of
chromatin condensation, an increase in fluorescein fluor-
escence occurred, shown by a shift (an increase) in fluores-
cence (Figure 3b,d,f). This shift is the result of increased
incorporation of dUTP due to DNA fragmentation.
Induction of apoptosis in plant cells
Agents chosen for this study were of three types: those
that either induced an increase in intracellular calcium
(A23187); resulted in oxidative stress (H2O2, salicylic acid);

806 Iona E.W. O’Brien et al.
Figure 3. PI fluorescence (a,c,e) and the TUNEL assay (b,d,f) were used
to measure chromatin condensation and ssDNA breaks, respectively, in
senescing cells.
An increase in fluorescence in the TUNEL assay (denoted as FITC –see
Experimental procedures) corresponds to an increase in DNA breakage.
The fluorescence shift is depicted by overlaying the control population
(shaded peak, without TdT) with the TdT-labelled population (unshaded
peak) as a function of time. Channel numbers for the PI assay denote
reduction in fluorescence with increasing condensation. (a) and (b)
represent 0% condensation, (c) and (d) 40%, and (e) and (f) 70%. The
increase in DNA breakage coincides with the increase in chromatin
condensation, with maximum number of DNA breaks (f) coinciding with
the occurrence of an apoptotic peak (e). The control assay (g) represents
the greatest level of DNA breakage achieved from DNAse I treatment.
or altered chromatin structure, including camptothecin
(topoisomerase I inhibition) and okadaic acid (protein
phosphorylation inhibition). Tobacco protoplasts were
treated with a range of concentrations of each of these
agents. From the dose responses obtained, we were able to
identify concentrations of each agent which either induced
rapid necrosis or resulted in cell death by a process
showing characteristic features of apoptosis (dose–
response data not shown).
© Blackwell Science Ltd, The Plant Journal, (1998), 13, 803–814
1365313x, 1998, 6, Downloaded from https://onlinelibrary.wiley.com/doi/10.1046/j.1365-313X.1998.00087.x by Egyptian National Sti. Network (Enstinet), Wiley Online Library on [28/06/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Figure 4. Protoplasts were treated with 1 µM A23187 (c,d,e), 70 µM salicylic
acid (f,g,h) or 1 µM camptothecin (i,j,k). Chromatin condensation increased
as shown by a decrease in PI fluorescence (c,f,i), in contrast to the control
population (a), and ssDNA breaks increased as demonstrated by an increase
in TUNEL (FITC) fluorescence (d,g,j) with no increase in the control
population (b) over 3 h. An apoptotic peak developed after 4 h treatment
with A23187 (e), salicylic acid (h) and camptothecin (k).
Relatively high levels of the agents (e.g. 3 µM A23187)
caused cells to become necrotic. These protoplasts showed
membrane leakage (measured by FDA and PI staining), an
increase in forward scatter (therefore swelling prior to
rupture), and no change in DNA content (no quantitative
change in PI intensity). No nDNA fragmentation was
detected in necrotic cells. In direct comparison, protoplasts
treated with lower levels of the agents (e.g. 1 µM A23187)
retained membrane integrity and cell viability, had higher
side scatter (chromatin condensing) and lower forward
scatter (cell shrinkage; see Experimental procedures).
Treatments which stimulated apoptosis (lower concentra-
tions than those inducing immediate necrosis) resulted in
shifts in PI fluorescence (a decrease indicating chromatin
condensation, Figure 4c,f,i), and an increase in fluorescence
with TUNEL analysis, indicating an increase in DNA break-
age (Figure 4d,g,j). After 4 h with the same agents, an
apoptotic peak developed (Figure 4e,h,k), with increasing
DNA breakdown.
 1365313x, 1998, 6, Downloaded from https://onlinelibrary.wiley.com/doi/10.1046/j.1365-313X.1998.00087.x by Egyptian National Sti. Network (Enstinet), Wiley Online Library on [28/06/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Early stages of plant apoptosis are reversible 807

Treatment of protoplasts with 1 µM A23187, resulted in

an immediate increase in intracellular calcium (Figure 5a).
The increase was determined from the ratio of fluorescence
measurements at 405 and 525 nm. The former wavelength
is that for unbound Indo-1 and the latter for bound Indo-1
to calcium (Robinson, 1993). The mean of this ratio for
the protoplast population was found and concentration
calculated from a known calcium-Indo-1 standard. This
mean ratio is displayed in the density plots for Indo-I
fluorescence as shown in Figure 5. The increase was
maintained over 10 mins, after which time the intracellular
calcium concentration returned to base level. Following
this, after 3 h, there was a second increase in intracellular
calcium, which coincided with the time of appearance of
the maximum TUNEL response. The data allow an estimate
to be made of intracellular calcium concentration

Figure 5. Treatment of protoplasts with 1 µM A23187 (a), the same

concentration which induces chromatin condensation, resulted in an
immediate increase in intracellular calcium as detected by a rise in Indo-1
fluorescence (Y-axis).
An increase in intracellular calcium was also detected in plant protoplasts
taken from the normal growth cycle at the 35% stage of chromatin
condensation (b) in comparison to control protoplasts with no condensation
(c). The graph shows long-term changes in intracellular calcium after
treatment with 1 µM or 3 µM A23187. Protoplasts were treated with A23187
at time zero, Indo-1 added at the given time points, and fluorescence
measured after 10 min. Protoplasts treated similarly, but with 130 µM H2O2
did not increase in intracellular calcium over the 4-h period. Control
protoplasts with no A23187 added maintained a calcium concentration of
approximately 215 nM over the experimental time.

© Blackwell Science Ltd, The Plant Journal, (1998), 13, 803–814

 1365313x, 1998, 6, Downloaded from https://onlinelibrary.wiley.com/doi/10.1046/j.1365-313X.1998.00087.x by Egyptian National Sti. Network (Enstinet), Wiley Online Library on [28/06/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
808 Iona E.W. O’Brien et al.

(Robinson, 1993). The resting concentration is calculated

to be approximately 215 nM, rising to about 650 nM at the
peak for 1 µM A23187, and to approximately 850 nM at
the peak for 3 µM A23187. The former concentration of
ionophore induced chromatin condensation (Figure 4), with
eventual apoptotic cell death. With 3 µM A23187, there was
a greater increase, followed by a rapid collapse coinciding
with necrosis. However, when we compared control proto-
plasts with those at 35–40% chromatin condensation during
natural senescence of the cell population, we found higher
levels of intracellular calcium (Figure 5b,c). We were unable
to find any increase in intracellular calcium associated
with treatment of protoplasts with H2O2 at concentrations
(130 µM) which induced chromatin condensation (see
Figure 6 for H2O2 data). This is in contrast with data
from Levine et al. (1996) and is discussed further in the
Discussion section.
Protoplasts treated with H2O2 showed increases in
intracellular glutathione (GSH) correlated with specific
levels of chromatin condensation determined by PI fluor-
escence. This is visualised by a shift in the 2-D aggregation
of signal shown by plotting low fluorescence (405 nm) and
high fluorescence (450–500 nm) which measure non-bound
GSH and bound GSH, respectively (Figure 6). A vertical
shift in the intensity denotes an increase associated with
higher levels of reduced glutathione. The sequence of
histograms obtained from treating protoplasts with H2O2
shows an increase in reduced glutathione seen by a move-
ment of the fluorescence intensity up the Y-axis (Figure 6b).
After this increase, the reduced glutathione levels start to
fall (Figure 6c) until returning close to those at time zero
(Figure 6a) by the time chromatin condensation has
reached about 35% (Figure 6e). A similar increase and
decrease in internal reduced glutathione was found in cells
from a growing cell population (Figure 6f–h).
Figure 6. Treatment of protoplasts with 130 µM H2O2 resulted in an increase
in intracellular reduced glutathione as depicted by an increase up the Y-
Reversibility of chromatin condensation axis (high fluorescence indicates monochlorobimane binding).
Protoplasts were exposed to 130 µm H2O2 at time zero (a), and mono-
During our experiments, we noticed on several occasions chlorobimane added at 30 min intervals (a–e) and fluorescence measured.
that fluorescence shifts associated with increased chro- At 30 min (b), reduced glutathione increased, followed by a gradual
reduction until at 2 h (e), there was little evidence for reduced glutathione
matin condensation showed reversibility. The first instance
present. Beyond this point, protoplasts did not respond to further addition
of this was when cells were subcultured after reaching a of peroxide. The population in the lower left corner has decreased in
level of 40% chromatin condensation. Four days after forward scatter, and therefore is in the late stages of apoptosis. When
protoplasts taken from different stages of the growth cycle were stained
subculturing, fluorescence levels had returned to those
with monochlorobimane, higher levels of reduced glutathione were found
associated with 0% chromatin condensation (Figure 2). The in protoplasts with up to 20% chromatin condensation (0%, f; 20%, g). At
first interpretation of this would be that in the population 35% condensation (h), there was a lower level of intracellular reduced
glutathione.
of cells at the time of subculturing, there were uncondensed
cells which were candidates for new cell division. However,
these would have been obvious in the flow cytometric The apparent reversibility was investigated further. When
analysis as two peaks representing condensed and uncon- protoplasts were treated with 1 µM A23187 for 2 h, the
densed chromatin. Furthermore, there was no evidence of agent was washed out and the protoplasts held for a
a large population of dead cells in the subculture, which further 18 h, there was an initial increase in chromatin
would have indicated that cells at the 40% level had condensation (Figure 7a,b) followed by a reduction back
proceeded to die. to the initial levels (Figure 7c). A similar shift and recovery

© Blackwell Science Ltd, The Plant Journal, (1998), 13, 803–814

 1365313x, 1998, 6, Downloaded from https://onlinelibrary.wiley.com/doi/10.1046/j.1365-313X.1998.00087.x by Egyptian National Sti. Network (Enstinet), Wiley Online Library on [28/06/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Early stages of plant apoptosis are reversible 809

Figure 7. Protoplasts were treated with 1 µM A23187 (a–c) and underwent chromatin condensation, which at 2 h is shown by a substantial shift in fluorescence (b).
The protoplasts were then washed and within 18 h, fluorescence returned to the original channel (c). Protoplasts treated with 7.5 µM okadaic acid (d–f) 70 µM
H2O2 (g–i) developed chromatin condensation (as determined by decreased PI fluorescence) over 3 h (e and h, respectively). Under continued exposure to
the agents for 5 h, peak fluorescence returned to the original channel (f and i, respectively). In all three treatments some cells died as depicted by the debris
in the lower channel. Histograms a, d and g are controls before addition of the inducing chemicals. (a) and (g) are the same since, for these treatments, the
same batch of protoplasts was used and split.

was found in the continued presence over 5 h of sublethal
levels of agents such as 0.5 µM A23187 (data not shown)
or 7.5 µM okadaic acid (Figure 7d–f) and 70 µM H2O2
(Figure 7g–i).
Where treatment replicates from the same batch of
protoplasts were analysed by flow cytometry, standard
deviations (n 5 3) were very low. For example, with PI
binding repeated analysis of control (time 0) samples in
experiments gave standard deviations typically of less than
5%. Repeated sampling of protoplasts treated with various
agents gave SDs of up to 10%. Thus, for example, fluores-
cence changes shown in Figure 7 are beyond this variability
and therefore significant. Effects such as those shown in
Figure 7 were confirmed by repeating experiments.
ation of apoptotic protoplasts binding annexin V (upper
left quadrant), a normal subpopulation with neither annexin
V nor PI binding (lower left quadrant), and a necrotic
subpopulation strongly positive for PI indicative of mem-
brane leakage (right hand population). After 1 h (Figure 8b),
a subpopulation of apoptotic cells can be seen, with most
cells in a relatively uncondensed condition. After 2 and 3 h
(Figure 8c,d, respectively), both the apoptotic and necrotic
subpopulation increased and became more distinct.
Necrotic cells can develop both as an end-point of the
apoptotic pathway and by non-apoptotic death. Thus
annexin V binding, combined with PI, was used to detect
levels of apoptosis and to differentiate between subpopula-
tions at different stages.

Annexin V binding Discussion

Annexin V binding was first detectable at 10–20% chromatin

condensation and was present throughout the process of
apoptosis in plant cells during senescence, or treated with
A23187, salicylic acid or camptothecin. To detect annexin
V binding, whole protoplasts were used as we were looking
for binding on the plasma membrane. Since PI (when used
as vital stain) is taken up by necrotic cells but excluded
from apoptotic ones, we could use PI as an indicator of
membrane damage (it is a commonly used vital stain) at
the same time as measuring annexin V binding. We treated
control protoplasts (no detectable chromatin condensation,
Figure 8a) with 1 µM A23187 and after 1, 2, and 3 h labelled
them with annexin V and PI (Figure 8b,c,d, respectively).
In each of the 2-D histograms in Figure 8 representing
treated protoplasts (b,c and d), there is a distinct subpopul-
Plant apoptosis
The flow cytometric and microscopic characteristics of
chromatin condensation and nucleosomal DNA fragmenta-
tion that we have found in tobacco cells are almost identical
to those described for apoptosis in mammalian cells. The
only mammalian characteristic which we did not observe
was plasma membrane blebbing, although we did not
look closely for this. A wide range of examples of DNA
fragmentation associated with whole cell shrinkage and
condensation have been published (see Pennell and Lamb,
1997 for review). However, a particular association between
chromatin condensation and the other features of pro-
grammed cell death such as DNA fragmentation, other
indicators such as annexin V binding, with induction by

© Blackwell Science Ltd, The Plant Journal, (1998), 13, 803–814


810 Iona E.W. O’Brien et al.
Figure 8. Control (uncondensed) protoplasts were treated with 1 µM
A23187.
(a) represents an annexin V control in the absence of A23187 (showing no
binding of annexin V or PI). After 1 (b), 2 (c) and 3 h (d) treatment with
A23187, the protoplasts were treated with annexin V and PI. Fluorescence
from annexin V and PI was measured. Three populations of protoplasts
are visible. The population in the upper left quadrant has high annexin V
binding and no PI fluorescence, indicative of apoptotic protoplasts. The
population on the right has relatively high PI fluorescence indicative of
necrotic protoplasts. The lower left population is typical of normal healthy
protoplasts showing no annexin V binding or PI uptake. Note that
fluorescence in these histograms is displayed on a log scale.
various agents, has been made possible by flow cytometric
analysis and has not yet been part of published plant
studies.
A distinctive feature of our results is the magnitude of
the change in stainability of DNA by PI. Mammalian studies
usually show much smaller channel shifts with PI fluores-
cence on the settings we have used, followed by a subdi-
ploid peak known as the apoptotic peak. This peak is
indicative of the start of DNA loss from the nucleus and is
a specific characteristic of apoptosis. However, in tobacco
cells we have found shifts of up to 480 channels (70%),
followed by the apoptotic peak (e.g. Figure 2). The greater
shift in plant cells suggests that chromatin in plants is
more dynamic, i.e. there is a greater degree of coiling of
DNA around the histones over a longer period, suggesting
that this may be a protective mechanism valuable for an
organism vulnerable to static exposure to stress such as
high UV radiation.
Flow cytometry also allows for analysis of cell size
changes. We found a decrease in cell size (forward scatter)
with an increase in granularity (side scatter), both conven-
tionally characteristic of mammalian apoptosis. By compar-
ison, necrotic cells swelled (increased forward scatter).
Analysis of nuclear DNA fragmentation by TUNEL, and our
© Blackwell Science Ltd, The Plant Journal, (1998), 13, 803–814
1365313x, 1998, 6, Downloaded from https://onlinelibrary.wiley.com/doi/10.1046/j.1365-313X.1998.00087.x by Egyptian National Sti. Network (Enstinet), Wiley Online Library on [28/06/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
microscopic evidence for DNA marginalization (Figure 1),
appear to be no different from that found in mammalian
cells (O’Brien et al., 1997).
Signalling pathways
We have shown that the above features of apoptosis
are present in plant cells which are undergoing natural
senescence, and in the presence of chemical agents. As
with mammalian cells, chromatin condensation and nDNA
fragmentation appear to be characteristic of cells at a later
stage in their growth curve. This was sufficiently consistent
for us to use, for other experiments, cells at various stages
of the growth cycle identified by specific percentages of
condensation. The sequence of changes found in cultured
cells leads us to suggest that apoptosis could be reversible
at certain stages, as is discussed below.
The chemicals and concentrations used were selected on
the basis of their ability to induce chromatin condensation
followed by nDNA fragmentation in mammalian cells.
Mammalian apoptosis is an active cell process which
includes a number of signalling stages (Hedley and McCul-
loch, 1996; Martin et al., 1994). This sequence includes an
oxidative burst, followed by an increase in cytosolic calcium
concentration and then protease action inducing changes
in chromatin structure (Hedley and McCulloch, 1996; Martin
et al., 1994). The chemicals we chose were of three types:
those resulting in oxidative stress (hydrogen peroxide,
salicylic acid), those inducing an increase in intracellular
calcium (A23187), and those altering nuclear morphology
(camptothecin, okadaic acid). Plant cells treated with rela-
tively high levels of the agents (e.g. 3 µM A23187) became
necrotic, showing similar flow cytometric characteristics to
necrotic mammalian cells (Dive et al., 1992; Darzynkiewicz
et al., 1992), with membrane leakage identified by dual
FDA and PI staining, an increase in forward scatter (there-
fore swelling prior to rupture), and no change in DNA
content as measured by PI binding. No nDNA fragmenta-
tion was detected in necrotic cells with the TUNEL
technique.
In contrast, plant cells treated with moderate levels of
an inducing agent (e.g. 1 µM A23187) retained membrane
integrity and cell viability, increased in side scatter while
decreasing in forward scatter (therefore cell shrinkage),
with a progressive increase in chromatin condensation
followed by nDNA fragmentation indicative of apoptosis.
The third, low concentration level identified resulted in
chromatin condensation in conjunction with a small num-
ber of single-stranded DNA breaks as determined by
TUNEL. These cells recovered, indicative of repair (see
below).
The role of calcium in the induction of apoptosis has
been described for mammalian cells (Bertrand et al., 1993;
Martin et al., 1994), and our results with A23187 showing

chromatin condensation followed by nDNA fragmentation
approximately 3 h following treatment confirm published
plant results (Jabs et al., 1997; Levine et al., 1996; McCabe
et al., 1997), suggesting that calcium plays a critical role in
plant programmed cell death. Necrotic cells showed a
rapid increase in intracellular calcium, followed by an even
more rapid drop as the cells died (Figure 5), results similar
to those for necrosis in mammalian cells (Darzynkiewicz
et al., 1992). The A23187 concentrations effective in our
study (0.5–3 µM) are substantially lower than those used
by Levine et al. (1996) (15 mM) and McCabe et al. (1997)
(0.1 mM). In both of these published studies, only cell death
and whole cell morphological changes were recorded after
relatively long treatment periods; neither followed earlier
nuclear changes. The differences between these results
and ours may therefore lie in the greater sensitivity of
protoplasts and in our measurement of early apoptotic
features. Indo-1 has not been used previously for flow
cytometric analysis in plant cells, although it was used to
follow intracellular calcium changes in soybean suspension
cells (Levine et al., 1996). We preferred this to other fluor-
escent agents such as Fura-red or Fluo-3, since with Indo-
1 we can more precisely determine ratios of bound to
unbound agent, and there appears to be a slower transfer
of agent into the vacuole (data not shown). There are
significant differences in our methodology (short-term
loading and analysis) and that of Levine et al. (1996) who
loaded over several hours. However, the same conclusions
are reached: raised internal calcium concentrations are
part of the signalling pathway, and perturbation of the
cellular calcium economy can induce major features of
apoptosis.
To identify whether oxidative stress triggers apoptosis
in plants we tested hydrogen peroxide and salicylic acid.
Both these treatments resulted in chromatin condensation
leading to apoptosis and, again, we were able to distinguish
between treatments which directly induced necrosis and
those which induced apoptotic changes in chromatin and
DNA. An oxidative burst has been associated with the
hypersensitive response (Lamb and Dixon, 1997), and there
are a number of instances where treatment with H2O2 and
increased intracellular levels of reduced oxygen species
have been shown to lead to programmed cell death (Jabs
et al., 1996; Levine et al., 1994, 1996). Whilst the effects of
H2O2 are explainable in terms of the published evidence,
there is still uncertainty as to how salicylic acid may
work since it does not necessarily induce increased H2O2
concentrations in cells (Bi et al., 1995). As with our results
using A23187, there are large differences between our
effective concentrations of H2O2 (70–260 µM) and those
quoted in the literature. Levine et al. (1994) have pointed
out that H2O2 has a relatively short life in a cultured cell
suspension, but also that a short pulse may be all that is
necessary to induce cell death. These authors also showed
© Blackwell Science Ltd, The Plant Journal, (1998), 13, 803–814
1365313x, 1998, 6, Downloaded from https://onlinelibrary.wiley.com/doi/10.1046/j.1365-313X.1998.00087.x by Egyptian National Sti. Network (Enstinet), Wiley Online Library on [28/06/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Early stages of plant apoptosis are reversible 811
that lower concentrations of H2O2 (2 mM) than required for
hypersensitive cell death (8 mM) were sufficient to induce
cellular protectant genes such as glutathione S-transferase
and glutathione peroxidase. Protoplasts will be consider-
ably more sensitive that whole cells to agents such as
peroxide and again, as with our calcium results, we were
detecting early stages in progammed cell death – chromatin
condensation, DNA fragmentation and changes in
intracellular reduced glutathione, as well as necrosis with
the higher concentrations.
Dose responses with inducing agents (lower concentra-
tions causing apoptosis and higher necrosis) are found
with mammalian cells (Wertz and Hanley, 1996). We also
found the same effect, which has also been noted by
McCabe et al. (1997). Effective concentrations of an agent
will be a function of duration of treatment and the initial
condition of the cells. Therefore, although we consistently
found that increasing concentrations of stimulating agents
such as A23187 and H2O2 delivered over the same time
periods resulted in either apoptosis or, at higher concentra-
tions, necrosis without passing through an apoptotic stage,
the absolute concentrations having these effects could
change according to cell stage.
The juxtaposition of calcium and oxidative signalling
events is still unresolved in mammalian cells, although
there is a belief that the oxidative burst may be a relatively
early event (Hedley and McCulloch, 1996). The glutathione
changes that we found in tobacco cells undergoing natural
senescence (occurring at about the 20% chromatin con-
densation level, where we also found the first TUNEL
response, Figure 3) occurred before any increase in
intracellular calcium was detectable (at about the 35%
level), supporting the same situation occurring in plants
(Figure 6). Direct measurements of H2O2-induced increases
in intracellular calcium suggest that, in some plant systems,
a calcium rise may follow oxidative activity (Levine et al.,
1994), although other data suggest that the sequence is
much less clear (e.g. Jabs et al., 1997). We were not able
to detect any increase in intracellular calcium after treating
cells with H2O2. As mentioned above, our experiments
involved much lower peroxide concentrations than those
used by Levine et al. (1996), and much shorter term
intracellular calcium measurements (whereby we tried to
exclude intravacuolar calcium measurements). The
sequence of events may well differ among cell types
and under different induction, suggesting multisignalling
pathways (Martin et al., 1994). A further consideration
is that most methods of detecting intracellular calcium
increases associated with programmed cell death in plant
cells are not designed to detect release from intracellular
pools. The conventional mechanism for calcium signalling,
e.g. the Indo-1 loading methods used by Levine et al.
(1996), would have led to the presence of the indicator in
the vacuole as well as the cytosol. Often, the applied

812 Iona E.W. O’Brien et al.
treatments are sufficiently traumatic to have effects at the
plasma membrane which may not mimic those occurring
naturally.
Both okadaic acid, which in plant cells inhibits phosphat-
ases I and II resulting in the hyperphosphorylation of
histone 1 leading to chromatin condensation (Zhang et al.,
1992), and camptothecin, which inhibits topoisomerase I
thereby inducing apoptosis in HL-60 cells (Bertrand et al.,
1993; O’Brien et al., 1996), induced chromatin condensation
in tobacco cells. Together, these data show that nuclear
changes characteristic of animal apoptosis may occur in
plants, and these compounds may be useful for direct
targeting of histones and DNA in determining character-
istics of plant apoptosis.
Reversibility and control
Perhaps our most novel result is that which suggests that
the apoptotic pathway in plant cells is reversible at specific
stages. In our earlier work, we found that early stages of
chromatin condensation, in response to mild HCl treatment
of nuclei, could be reversed (O’Brien et al., 1998). We
assume in this case, with acid treatment, reversibility is a
result of disassociation of histones, allowing increased PI
binding. In the present work, we have confirmed reversibil-
ity of chromatin condensation by using the chemicals at
sublethal levels (e.g. 0.5 µM A23187) or by washing the
cells after varying treatment times to remove the agent
(Figure 7). In both instances, chromatin condensation which
had been induced in the presence of the agent (e.g. a shift
from channel number 356–67 in 2 h) was fully reversed
within 20 h (with a return to channel number 362; Figure 7).
It appears likely that early repair mechanisms were able to
prevent cell death. The only way in which such a reversal
could occur was by chromatin relaxation.
When we followed chromatin condensation during nat-
ural senescence (Figure 2), we were able to show that
renewal of medium was followed by a reversible PI shift.
We do not believe that this represents a new subpopulation
of cells, since we have no cytometric evidence for G2 cells
in the culture before subculturing, nor for two peaks
representing condensed and uncondensed chromatin at
the early stages of the new cycle. Our conclusion is that
cells in which chromatin had condensed to some degree
were able to recover and be stimulated to divide. Our data
also suggest that early DNA breaks may be reversible
(Figure 3).
Our results with annexin V binding also confirm our
earlier results (O’Brien et al., 1997), that this indicator of
the progress of apoptosis in animal cells can also be used
to detect stages of apoptosis in plant cells. Annexin V is
an indicator of exposure of PS at the outer surface of animal
cell plasma membranes, a process linked to apoptosis. It
appears that annexin V binding may occur at an earlier
© Blackwell Science Ltd, The Plant Journal, (1998), 13, 803–814
1365313x, 1998, 6, Downloaded from https://onlinelibrary.wiley.com/doi/10.1046/j.1365-313X.1998.00087.x by Egyptian National Sti. Network (Enstinet), Wiley Online Library on [28/06/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
stage than in mammalian cells, perhaps before the loss of
reversibility.
In summary, we may be able to identify decision points
at early stages of chromatin condensation in the apoptotic
pathway in plants. The greater degree of condensation of
plant chromatin may allow us to dissect the pathway to
an extent not appreciated in mammalian studies. This has
led us to identify reversibility in the early stages, and
show the potential for distinguishing progressive stages
of the pathway.
Experimental procedures
Cells and culture conditions
Suspension-cultured tobacco (Nicotiana plumbaginifolia) cells
were originally obtained from Dr P. Larkin (CSIRO, Canberra,
Australia). Cells were maintained in liquid CSV media in conical
flasks on an orbital shaker (100 rpm), grown at 25°C with 16 h
photoperiod, and subcultured at 7 day intervals.
Natural cell death
Senescence of plant cell cultures was determined by measuring
the percentage of cell division in the culture over 14 days. Cells
were stained with propidium iodide (PI) and the cell cycle analysed
by flow cytometry, using the Multicycle AV Cell Cycle Analysis
Program (Phoenix Flow Systems, San Diego, CA, USA). This cell
cycle programme determines the percentage of cells in any stage;
we are also following the protocol of Galbraith et al. (1983) on
cell cycle analysis in plants. This technique provided data on the
percentage of cells in each part of the cell division cycle. Cultures
were considered to be senescent when the percentage of cells in
G2 phase dropped below 2%.
Induced cell death
A number of chemical agents were used to induce chromatin
condensation and DNA fragmentation. A23187 (Molecular Probes,
Inc., Eugene, Orm USA), camptothecin, and okadaic acid were
dissolved in DMSO at concentrations shown in the results. The
A23187 used was the non-fluorescent form. Salicylic acid was
dissolved in 1 M sodium hydroxide, and hydrogen peroxide was
used in an aqueous solution. Where DMSO was used, the possible
effects of the solvent alone on the plant cells were also tested.
The effects of these agents were tested as follows. Protoplasts
from Nicotiana cells were isolated and washed according to
O’Brien and Lindsay (1993). Sub-samples (™ 106 ml–1) were incub-
ated for varying times with the different agents. To examine the
effect of removing the agents, protoplast suspensions were divided
into two batches. Both batches were treated with chemical agents
as above. One batch was then analysed for chromatin condensa-
tion immediately after appropriate incubation. In order to follow
recovery, the other batch was washed twice after varying times,
put back into culture medium, and analysed 24 h later.
Chromatin condensation
Protoplasts were isolated and washed according to O’Brien and
Lindsay (1993), except that cellulase (Worthington Biochemical

Corp, NJ, USA) was used instead of Cellulysin (Calbiochem, La
Jolla, CA, USA) due to the high levels of DNAse activity in the
Cellulysin. The protoplasts were lysed in 10 ml of ice-cold modified
(45 mM sodium citrate) Galbraith’s buffer (Galbraith et al., 1983)
for 5 min. The nuclei were centrifuged at 300 g for 4 min at 4°C,
the supernatant removed, and the pellet resuspended in 50 µg
ml–1 PI and 1000 units ml–1 deoxyribonuclease-free ribonuclease
A (Sigma) in modified Galbraith’s buffer. Final dye saturation was
adjusted to 1.5 µg 1000–1 nuclei.
The nuclei were analysed using an EPICS Profile II flow cytometer
(Coulter Electronics, Hialeah, FL, USA). The dye PI was excited
with 15 mW of the 488 nm line of an argon ion laser. At least
10 000 nuclei were analysed with clumped nuclei eliminated by
gating on peak versus integral fluorescence of the PI signal. Data
were analysed using the Multicycle AV Cell Cycle Analysis Program
(Phoenix Flow Systems, San Diego, CA, USA).
Linearity of the flow cytometer was ensured firstly by using
linearity beads (Coulter Electronics, Hialeah, FL, USA), and sec-
ondly by using plant standards (cotyledons of germinated seed-
lings were excised and nuclei prepared according to Galbraith
et al., 1983) of varying DNA contents. On a daily basis, chicken
red blood cells in combination with DNA check beads (Coulter
Electronics, Hialeah, FL, USA) were used to ensure that there was
no variation in the instrument by standardising the fluorescence
signal. A 100 µg ml–1 PI solution was run through the instrument
for 5 min prior to analysis in order to saturate the tubing.
DNA fragmentation
Protoplasts were fixed in 2% paraformaldehyde for 30 min at
room temperature with gentle rotation. The protoplasts were
washed twice, and the final pellet slowly resuspended in ice-cold
Piersolve (ethylene glycol monomethyl ether, Pierce Chemical Co,
Rockford, IL, USA) by adding the solution dropwise during 15 min
to avoid clumping. This dehydrated the protoplasts, which were
then held overnight at – 20°C. The dehydrated protoplasts were
then washed twice in the protoplast wash solution and lysed in
0.1% triton X-100 and 45 mM sodium citrate with gentle swirling
for 6 min. The nuclei were washed and incubated for 1 h at 37°C
with terminal deoxynucleotidyl transferase (TdT) and fluorescein-
dUTP according to the in situ Cell Death Detection Kit (Boehringer
Mannheim). Nuclei were then washed and stained with 20 µg
ml–1 of PI, and analysed by flow cytometry for both nuclear DNA
fragmentation by detection of fluorescein-labelled dUTP (TUNEL),
and chromatin condensation (PI). A potential source of error with
in situ DNA strand break labelling is that necrotic cells can give a
false positive signal with the nick translation assay. We took care
to use both a necrotic cell control, and to use the TUNEL assay
which has been shown to preferentially label cells undergoing
apoptosis (Hotz et al., 1994), to ensure that apoptotic and not
necrotic DNA breaks were being measured.
Intracellular calcium and glutathione
For calcium measurements, protoplasts from different stages of
senescence and/or treated with different chemicals as described
above were loaded with 3 µM Indo-1 (Molecular Probes Inc) at
28°C for 10 min at pH 5.5 and then returned to pH 7.0. Microscopic
observation showed that beyond 30 min Indo-1 accumulated in
the vacuole and, accordingly, protoplasts were analysed for no
longer than 30 min. Indo-1 fluorescence was collected on a BD
Facstar Plus flow sorter running dual lasers with 488 nm primary
beam (200 mW) and UV secondary beam (100 mW). Forward/side
© Blackwell Science Ltd, The Plant Journal, (1998), 13, 803–814
1365313x, 1998, 6, Downloaded from https://onlinelibrary.wiley.com/doi/10.1046/j.1365-313X.1998.00087.x by Egyptian National Sti. Network (Enstinet), Wiley Online Library on [28/06/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Early stages of plant apoptosis are reversible 813
scatter signals were collected from the primary beam using 488/
10BP filters. The Indo-1 signal was split such that low fluorescence
(bound calcium) was collected on the FL3–2 detector through a
dichroic 405/30BP, and high fluorescence (free calcium) on the FL4
detector using a D485/25BP, with the signal split through a D440LP
filter. Binding of calcium to Indo-I results in a fluorescence wave-
length emission shift from violet (calcium-free) to blue-green
(calcium-bound).
To measure intracellular glutathione, protoplasts were stained
with 40 µM monochlorobimane (Molecular Probes Inc) at 25°C for
10 min and analysed immediately. Identical results were achieved
with monobromobimane. The glutathione signal is for reduced
glutathione and the same filter set up was used except that the
FL4 detector was fitted with a 520LP filter, in order to cope with
high levels of autofluorescence. A background subtraction was
made by exposing protoplasts to diamide (diazine-dicarboxylic
acid bis[N,N-dimethylamide], Sigma), which inactivates gluta-
thione, and measuring fluorescence. An increase in reduced gluta-
thione is seen as an increase in fluorescence with the FL4 detector.
This is due to GSH forming a fluorescent adduct when bound to
monochlorobimane.
Annexin V binding
To detect annexin V binding to phosphatidylserine, protoplasts
were incubated with fluorescein-labelled annexin V protein
(APOPTEST©-FITC, NeXins Research BV, The Netherlands),
according to the manufacturer’s recommendations, for 5 min in
the presence of PI, washed twice with protoplast wash and
analysed by flow cytometry. In this case, PI is used as a vital stain,
being excluded by live cells, and taken up by necrotic ones.
Flow cytometry
For all flow cytometric analyses in this study, unstained protoplasts
at the same stage of cell division or under the same stress
conditions, were used as autofluorescence controls. In our results,
we use the terms ‘forward scatter’ and ‘side scatter’. These
refer to flow cytometric responses. Forward scatter indicates the
amount of laser light which passes through the cell to be detected
behind it. The level of forward scatter indicates cell size, i.e. the
greater the forward scatter, the larger the cell. Side scatter indicates
the granularity of the cellular contents, the greater the side scatter,
the more dense or granular the cell.
Channels with channel numbers refer to quantitative measures
of fluorescence. Fluorescence response is covered by a linear set
of 1024 channels, with the maximum levels of fluorescence at
1023 and the minimum at 1. The numbers associated with peaks
given in the figures refer to this 0–1024 range, e.g. high numbers
indicate high amounts of fluorescence. Peak heights and areas in
histograms represent the number of protoplasts or nuclei
recorded.
Acknowledgements
We wish to acknowledge the assistance and advice of Karen
Holdaway (Auckland Medical School), Geoff Osborne and Sabine
Gruninger (ANU, Canberra, Australia), and David Hedley (Princess
Margaret Hospital, Toronto, Canada).
References
Bertrand, R., Solary, E., Jenkins, J. and Pommier, Y. (1993)
Apoptosis and its modulation in human promyelocytic HL-60

814 Iona E.W. O’Brien et al.
cells treated with DNA topoisomerase I and II inhibitors. Exptl.
Cell Res. 207, 388–397.
Bi, Y.-M., Kenton, P., Mur, L., Darby, R. and Draper, J. (1995)
Hydrogen peroxide does not function downstream of salicyclic
acid in the induction of PR protein expression. Plant J. 8,
235–245.
Darzynkiewicz, Z., Bruno, S., Del Bino, G., Gorczyca, W., Hotz,
M.A., Lassota, P. and Traganos, F. (1992) Features of apoptotic
cells measured by flow cytometry. Cytometry, 13, 795–808.
Dietrich, R.A., Delaney, T.P., Uknes, S.K., Ward, E.R., Ryals, J.A.
and Dangl, J.L. (1994) Arabidopsis mutants simulating disease
resistance response. Cell, 77, 565–577.
Dive, C., Gregory, C.D., Phipps, D.J., Evans, D.L., Milner, A.E. and
Wyllie, A.H. (1992) Analysis and discrimination of necrosis and
apoptosis (programmed cell death) by multiparameter flow
cytometry. Biochim. Biophys. Acta, 1133, 275–285.
Earnshaw, W.C. (1995) Apoptosis: lessons from in vitro systems.
Trends Cell Biol. 5, 217–220.
Galbraith, D.W., Harkins, K.R., Maddox, J.M., Ayres, N.M., Sharma,
D.P. and Firoozabady, E. (1983) Rapid flow cytometric analysis
of the cell cycle in intact plant tissues. Science, 220, 1049–1051.
Greenberg, J.T., Guo, A., Klessig, D.F. and Ausubel, F.M. (1994)
Programmed cell death: a pathogen-triggered response
activated coordinately with multiple defense functions. Cell, 77,
551–564.
Groover, A., De Witt, N., Heidel, A. and Jones, A. (1997)
Programmed cell death of plant tracheary elements
differentiating in vitro. Protoplasma, 196, 197–211.
Hale, A.J., Smith, C.A., Sutherland, L.C., Stoneman, V.E.A.,
Longthorne, V.L., Culhane. A.C. and Williams, G.T. (1996)
Apoptosis: molecular regulation of cell death. Eur. J. Biochem.
236, 1–26.
Hedley, D.W. and McCulloch, E.A. (1996) The generation of reactive
oxygen intermediates after treatment of the blasts of acute
myeloblastic leukemia with cytosine arabinoside: The role of
BCL-2. Leukemia, 10, 1143–1149.
Hotz, M.A., Gong, J., Traganos, F. and Darzynkiewicz, Z. (1994)
Flow cytometric detection of apoptosis: Comparison of the
assays of in situ DNA degradation and chromatin changes.
Cytometry, 15, 237–244.
Jabs, T., Dietrich, R.A. and Dangl, J.L. (1996) Initiation of runaway
cell death in an Arabidopsis mutant by extracellular superoxide.
Science, 273, 1853–1856.
Jabs, T., Tschope, M., Colling, C., Hahlbrock, K. and Scheel, D.
(1997) Elicitor-stimulated ion fluxes and O2- from the oxidative
burst are essential components in triggering defense gene
activation and phytoalexin synthesis in parsley. Proc. Natl Acad.
Sci. USA, 94, 4800–4805.
© Blackwell Science Ltd, The Plant Journal, (1998), 13, 803–814
1365313x, 1998, 6, Downloaded from https://onlinelibrary.wiley.com/doi/10.1046/j.1365-313X.1998.00087.x by Egyptian National Sti. Network (Enstinet), Wiley Online Library on [28/06/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Katsuhara, M. (1997) Apoptosis-like cell death in barley roots
under salt stress. Plant Cell Physiol. 38, 1091–1093.
Kossiak, R.M., Chamberlin, M.A., Palmer, R.G. and Bowen, B.A.
(1997) Programmed cell death in the root cortex of soybean
root necrosis mutants. Plant J. 11, 729–745.
Lamb, C. and Dixon, R.A. (1997) The oxidative burst in disease
resistance. Annu. Rev. Plant Phyiol. Plant Mol. Biol. 48, 251–275.
Levine, A., Pennell, R.I., Alvarez, M.E., Palmer, R. and Lamb, C.
(1996) Calcium-mediated apoptosis in a plant hypersensitive
disease resistance response. Curr. Biol. 6, 427–437.
Levine, A., Tenhaken, R., Dixon, R. and Lamb, C. (1994) H2O2 from
the oxidative burst orchestrates the plant hypersensitive disease
resistance response. Cell, 79, 583–593.
Martin, S.J., Green, D.R. and Cotter, T.G. (1994) Dicing with death:
dissecting the components of the apoptosis machinery. Trend
Biochem. Sci. 19, 26–30.
McCabe, P.F., Levine, A., Meijer, P.-J., Tapon, N.A. and Pennell, R.I.
(1997) A programmed cell death pathway activated in carrot
cells cultured at low cell density. Plant J. 12, 267–280.
Mittler, R. and Lam, E. (1995) In situ detection of nDNA
fragmentation during the differentiation of tracheary elements
in higher plants. Plant Physiol. 108, 489–493.
O’Brien, I.E.W. and Lindsay, G.C. (1993) Protoplasts to plants in
Gentianaceae. Regeneration of lisianthus (Eustoma grandi-
florum) is affected by calcium ion preconditioning osmolality
and pH of the culture media. Plant Cell Tiss. Org. Cult. 33, 31–37.
O’Brien, I.E.W., Murray, B.G., Baguley, B.C., Morris, B.A.M. and
Ferguson, I.B. (1998) Major changes in chromatin condensation
are indicative of apoptosis in plant cells. Exptl. Cell Res. in press.
O’Brien, I.E.W., Reutelingsperger, C.P.M. and Holdaway, K.M.
(1997) The use of annexin-V and TUNEL to monitor the
progression of apoptosis in plants. Cytometry, 29, 28–33.
Pennell, R.I. and Lamb, C. (1997) Programmed cell death in plants.
Plant Cell, 9, 1157–1168.
Robinson, J.P. (1993) Handbook of Flow Cytometry Methods. New
York: Wiley-Liss. pp. 154–195.
Ryerson, D.E. and Heath, M.C. (1996) Cleavage of nuclear DNA
into oligonucelosomal fragments during cell death induced by
fungal infection or by abiotic treatments. Plant Cell, 8, 393–402.
Wang, H., Li, J., Bostock, R.M. and Gilchrist, D.G. (1996) Apoptosis:
a functional paradigm for programmed cell death induced by a
host-specific phytotoxin and invoked during development. Plant
Cell, 8, 375–391.
Wertz, I.E. and Hanley, M.R. (1996) Diverse molecular provocation
of programmed cell death. Trends Biochem. Sci. 21, 359–364.
Zhang, K., Tsukitani, Y. and John, P.C.L. (1992) Mitotic arrest in
tobacco caused by the phosphoprotein phosphatase inhibitor
okadaic acid. Plant Cell Physiol. 33, 677–688.