Manganese superoxide dismutase can reduce cellular damage mediated by oxygen radicals in transgenic plants.

Abstract

In plants, environmental adversity often leads to the formation of highly reactive oxygen radicals. Since resistance to such conditions may be correlated with the activity of enzymes involved in oxygen detoxification, we have generated transgenic tobacco plants which express elevated levels of manganese superoxide dismutase (MnSOD) within their chloroplasts or mitochondria. Leaf discs of these plants have been analyzed in conditions in which oxidative stress was generated preferentially within one or the other organelle. It was found that high level overproduction of MnSOD in the corresponding subcellular location could significantly reduce the amount of cellular damage which would normally occur. In contrast, small increases in MnSOD activity were deleterious under some conditions. A generally applicable model correlating the consequences of SOD with the magnitude of its expression is presented.

Full text

Page 1

The EMBO Journal vol. 10 no.7 pp. 1 723 - 1732, 1991
Manganese superoxide dismutase can reduce cellular
damage mediated by oxygen radicals in transgenic
plants
Chris Bowler', Luit Slooten2,
Saskia Vandenbranden2, Riet De Rycke,
Johan Botterman3, Chris Sybesma2,
Marc Van Montagu and Dirk Inze
Laboratorium voor Genetica, Rijksuniversiteit Gent, B-9000 Gent,
2Laboratorium voor Biofysica, Vrije Universiteit Brussel, B-1050
Brussel and 3Plant Genetic Systems N.V., Plateaustraat 22, B-9000
Gent, Belgium
'Present address: Laboratory of Plant Molecular Biology, The
Rockefeller University, New York, NY 10021-6399, USA
Communicated by M.Van Montagu
In plants, environmental adversity often leads to the
formation of highly reactive oxygen radicals. Since
resistance to such conditions may be correlated with the
activity ofenzymes involved in oxygen detoxification, we
have generated transgenic tobacco plants which express
elevated levels of manganese superoxide dismutase
(MnSOD) within their chloroplasts or mitochondria. Leaf
discs of these plants have been analyzed in conditions in
which oxidative stress was generated preferentially within
one or the other organelle. It was found that high level
overproduction of MnSOD in the corresponding sub-
cellular location could significantly reduce the amount
of cellular damage which would normally occur. In
contrast, small increases in MnSOD activity were
deleterious under some conditions. A generally applicable
model correlating the consequences of SOD with the
magnitude of its expression is presented.
Key words: Nicotiana tabacum/oxidative stress/paraquat/
protein targeting/transgenic plants
Introduction
Reduced oxygen species such as superoxide radicals (02
hydrogen peroxide and hydroxyl radicals (OH * ) have been
the subject of much research in recent years [for review,
see Cadenas (1989)]. This interest has arisen because of the
potential harm they can cause to all organisms exposed to
an aerobic environment. Although neither O27 nor H202 at
physiological concentrations seem particularly harmful, their
toxicity in vivo arises by a metal ion-dependent conversion
into hydroxyl radicals (the Haber-Weiss reaction), one of
the most reactive species known to chemistry (Halliwell,
1987; Imlay and Linn, 1988). Of particular importance is
its ability to mutate DNA and to initiate chain reactions of
lipid peroxidation.
Oxy-radicals are by-products of many biological oxida-
tions. For example, the electron transport chain of mito-
chondria is a well documented source of reactive oxygen
species (Freeman and Crapo, 1982; Forman and Boveris,
1982; Ksenzenko et al., 1983) whilst the oxygen-evolving
functions within plant chloroplasts also make them particu-
)
Oxford University Press
larly susceptible to oxy-radical formation; electrons pass-
ing through the photosystems can react with oxygen to form
superoxide radicals and hydrogen peroxide by a mechanism
known as the Mehler reaction (Mehler, 1951; Asada and
Takahashi, 1987).
In addition to their generation from these indigenous
reactions, oxy-radical formation appears to be greatly
increased during stress conditions. In animals, the role played
by reduced oxygen species in ischemia/reperfusion injury,
in some aspects of the aging process, in the development
of inflammatory diseases such as rheumatoid arthritis and
in the cytotoxicity of tumor necrosis factor (TNF) has been
well documented (for reviews, see Bannister et al., 1987;
Jones, 1986; Halliwell, 1987). In plants, damage arising
from environmental stresses often appears to be caused by
oxy-radicals. For example, increased oxygen toxicity can
be caused by air pollutants such as ozone and sulfur dioxide
(Menzel,
1976; Chia et al.,
(Harbour and Bolton, 1975; Orr and Hogan, 1983), chilling
(Clare et al., 1984), and high light intensities (Krause, 1988).
Organisms have evolved a wide range ofenzymic and non-
enzymic mechanisms to contend with this problem. Since
hydroxyl radicals themselves are too reactive to be easily
controlled, these mechanisms are designed to ensure that
superoxide and hydrogen peroxide do not come into contact
with each other. Of importance in this defense are anti-
oxidants such as urate, glutathione, ascorbate and a-
tocopherol, catalase and peroxidase enzymes which remove
H202, and superoxide dismutase (SOD) which converts
superoxide to hydrogen peroxide. In many cases it appears
that SOD is a key enzyme for providing protection against
oxidative stress. Three forms of this enzyme exist, as
classified by their metal cofactor: copper/zinc, manganese,
and iron forms. The iron enzyme (FeSOD) is present in
prokaryotes and within the chloroplasts of some plants. The
manganese SOD (MnSOD) is widely distributed among
prokaryotic and eukaryotic organisms and in eukaryotes it
is most often found in the mitochondrial matrix. The
copper/zinc enzyme (Cu/ZnSOD), however, is found almost
exclusively in eukaryotic species where it is normally present
in the cytosol and in addition some plants contain a
chloroplastic isoform (for review, see Bannister et al., 1987).
The importance of SOD for aerobic growth has been
demonstrated in several ways. SOD-deficient mutants of
Escherichia coli (Carlioz and Touati, 1986) and yeast
(Bilin'ski et al., 1985; van Loon et al., 1986) are hyper-
sensitive to oxygen, and a null mutation of Cu/ZnSOD in
Drosophila results in infertility and a reduction of life span
(Phillips et al.,
superoxide dismutase in human cells partially protectscells
against the cytotoxicity of tumor necrosis factor (TNF)
(Wong et al., 1989). In plants, resistance to air pollutants
such as ozone and SO2 has been shown to correlate with
increased
levels
of enzymes
detoxification in poplar and in Phaseolus vulgaris (Tanaka
1984), certain herbicides
1989). Over-production of manganese
involved
in superoxide
1723

Page 2

C.Bowler et al.
and Sugahara, 1980; Lee and Bennett, 1982). Increased
resistance to photo-oxidative damage in ripening tomato fruits
is correlated with SOD (Rabinowitch et al., 1982) and the
presence of the enzyme in rhizomes of Iris under anaerobic
conditions has been found to be important for their ability
to recover from anoxic stress (Monk et al., 1987).
However, some data suggest that SOD activity is not the
limiting factor determining cell survival under adverse
conditions. Transgenic plants which contain high levels of
Cu/ZnSOD within their chloroplasts appear to be no more
resistant
to
superoxide
toxicity
(Tepperman and Dunsmuir, 1990). Moreover, in specific
cases SOD has also been shown to cause detrimental effects.
For example increased activity of cytosolic Cu/ZnSOD
arising from a trisomy of chromosome 21 in Down's
syndrome (DS) patients may be a causative agent of some
of the disease symptoms. Low-level overproduction of
cytosolic Cu/ZnSOD in rat cells results in impaired neuro-
transmitter uptake (Elroy-Stein and Groner, 1988) and
transgenic mice which overproduce the enzyme develop
abnormalities within the tongue neuromuscular junctions
similar to those observed in tongue muscle of patients with
DS (Avraham et al., 1988).
If nothing else, these contradictory results demonstrate that
some cell types are affected by perturbations in SOD activity
levels, although the reasons why this is so have not been
elucidated. In an attempt to clarify some of the questions
raised by these results and also to evaluate the potential
capabilities of SOD for plant stress tolerance, we have
engineered plants which overproduce a MnSOD specifically
within their mitochondria or chloroplasts. Utilizing model
systems which generate an oxidative stress preferentially
within each of these compartments, we demonstrate here that
organellar MnSOD can provide protection against super-
oxide-mediated damage within these subcellular environ-
ments, although under some conditions
deleterious. We believe that the determining factor is the
ratio between superoxide and hydrogen peroxide within the
cell. In addition to being the SOD substrate and reaction
product, these are the two Haber-Weiss reaction substrates
from which hydroxyl radicals can be generated. The
implications of these results for further understanding the
general importance of SOD to the survival of multicellular
organisms and specifically for the engineering of plants more
tolerant to environmental adversity are discussed.
than normal
plants
it can also be
Results
Generation of plants which produce elevated levels
of MnSOD
In Nicotiana plumbaginifolia, MnSOD is a nuclear-encoded
protein which is targeted to the mitochondria. The isolation
of a full-length cDNA encoding this protein has been
previously described (Bowler et al., 1989a) and studies in
yeast have shown that the mitochondrial targeting is mediated
by an N-terminal leader sequence (Bowler et al., 1989b).
To make a construction which would overproduce MnSOD
in the mitochondria of plants, we inserted the entire coding
sequence downstream ofthe CaMV 35S promoter of binary
vector pGSJ780A (see Materials and methods), generating
plasmid pMitSOD (Figure la). Similarly, for targeting
MnSOD to chloroplasts, we first replaced the mitochondrial
leader sequence with a transit peptide for chloroplast
cDNA
a35Sp =z
b
35Sp :*E
mit. tp
mature protein
*
mit. tp
.F;;;
73FTMZM/IM
mature protein
Zzz/X//////////S
*
1
cp. tp
0.'.
at'-
mature protein
*
]
3'gene7 pMitSOD
3'gene7 pChISOD
AAG rGC ATG GAT CTG GGC TTG CAG
... Lys Cys Met Asp LeuGlyLuGln
...
Fig. 1. Schematic representation of constructions for mitochondrial and
chloroplastic targeting of MnSOD. The mitochondrial leader sequence-
encoding DNA (mit.tp) is denoted by a dotted box, the mature
protein-encoding sequence as a hatched box (with an asterisk
representing the stop codon) and the sequence encoding the chloroplast
transit peptide (cp.tp) as a striated box. Each cassette is flanked at the
5' end by the 35S promoter (35Sp) and at the 3' end by
polyadenylation signals of the T-DNA-encoded gene 7 (3' gene 7).
The sequence and corresponding amino acids at the fusion of the
chloroplast transit peptide and MnSOD mature protein in pChlSOD (b)
are shown. The predicted processing site is indicated by a vertical
arrow and the normal NH2-terminus of the MnSOD mature protein is
shown by a horizontal arrow. For details of the constructions, see
Materials and methods.
targeting and then placed this cassette downstream of the
35S promoter. This construct was designated pChiSOD
(Figure lb).
These constructions were subsequently introduced into
Nicotiana tabacum cv. PBD6 by conventional leaf disc
transformation (De Block et al.,
proteins were extracted from leaves of individual transgenic
plants (15 plants for each construction) and were first assayed
for MnSOD activity on native polyacrylamide gels. Most
of the putative transformants were found to contain a new
MnSOD activity (data not shown) and based on this analysis,
for each construction, two plants which contained the highest
levels were selected for further study. These were designated
MitSODl and MitSOD2 for the mitochondrial targeting
construct, and ChlSOD1 and ChlSOD2 for the chloroplast
targeting construct. It should be noted that we also designed
a construct to produce MnSOD in the cytosol of plants.
Although the protein product was active and transcript
abundance was equivalent to that in MitSOD and ChiSOD
plants, protein levels were extremely low, apparently a result
of protein instability (data not shown). Consequently these
plants were not studied further.
To determine whether the MnSOD was targeted to the
correct subcellular location, we prepared mitochondria and
chloroplasts from mature leaves ofMitSODI and ChlSOD1
transformants. Figure 2 details the results of this analysis.
Untransformed
N.tabacum
electrophoretically distinguishable SOD enzymes: three
Cu/ZnSODs, one MnSOD and one FeSOD (Figure 2a). The
MnSOD co-purifies with mitochondrial fractions and the
FeSOD is present within highly purified chloroplasts. A
similar situation exists in the related N.plumbaginifolia (Van
Camp et al., 1990) from which the MnSOD cDNA was
derived. However, the N.plumbaginifolia MnSOD migrates
faster than the N. tabacum MnSOD (e.g. Figure 2b), which
allows the engineered SOD to be distinguished from the
endogenous enzyme in the transgenic plants.
The activity of the new MnSOD in MitSODl and
ChlSOD1 plants adds significantly to the endogenous
activities (Figure 2b and c). The additional MnSOD activity
1987). Total soluble
appears
to
containfive
1724

Page 3

Reduction of cellular damage by MnSOD
a
C
-
C -j //Z ncS
Cs/Z
-[)
S CD
FeSOD [
Cu/Zn -SOU-D
b
C
A .4
..
-..-
Th
S
Fig. 2. Targeting of engineered MnSOD to mitochondria and chloroplasts of MitSODl and ChISODi plants, respectively, revealed by subcellular
fractionations (upper panel) and immunogold labelling of electron micrograph sections (lower panel). (a) SOD protein profile on native
polyacrylamide gels of untransformed tobacco (cv. PBD6). The three cytosolic Cu/ZnSODs, the chloroplastic FeSOD and the mitochondrial MnSOD
are indicated. The three types of SOD were identified using inhibitor studies (Van Camp et al., 1990). (b) SOD profile of MitSODl plants. (c) SOD
profile of ChISODI plants. The overproduced MnSOD is indicated in each case with an arrow. Abbreviations: total, 100 ytg total soluble proteins;
Chl., 50 ug protein extracted from Percoll-purified chloroplasts; Mit., 50lsgprotein from Percoll-purified mitochondria. The lower panel shows
representative immunogold labellings of sections derived from these plants. Abbreviations: C, cytosol; Chl, chlorop'ast; cw, cell wall; Mi,
mitochondria; Px, peroxisome; S, stroma; Th, thylakoid membrane.
in MitSOD1 co-purifies predominantly with mitochondrial
fractions, although some activity is present in the chloroplast
fractions (Figure 2b). We presume this to be a result of
mitochondrial contamination in the samples. In contrast to
MitSOD1, the new MnSOD activity in ChlSODl co-purifies
exclusively with the endogenous FeSOD in chloroplast
fractions (Figure 2c), demonstrating that the enzyme can be
targeted to, and retain its activity in, a foreign location.
Information obtained from the subcellular fractionations
was confirmed by immunogold labelling of sections derived
from untransformed and transgenic plants with polyclonal
antibodies
raised
against N.plumbaginifolia MnSOD
(Figure 2). Although expression of the endogenous MnSOD
was too low to be detected, the electron micrographs clearly
show labelling specifically within mitochondria of MitSOD 1
plants (Figure 2b) and chloroplasts of ChlSOD1 plants
(Figure 2c). We found no evidence for anomalous targeting
to other locations in any sections examined. Based on both
enzymatic activity and immuno-reactivity, the engineered
MnSOD in MitSOD2 and ChlSOD2 plants was identical to
that in MitSODl and ChlSOD1 plants, respectively (data
not shown).
Properties of chimeric MnSOD gene expression in
transgenic plants
Although transgenic plants which overproduce MnSOD in
either chloroplasts or mitochondria show no apparent
peculiarities during development, it was of great interest to
examine the effects of organellar SOD overproduction at the
cellular level, particularly during conditions which generate
oxidative stress. For these experiments, detailed in the
following sections, we chose to use assays which employed
leaf discs derived from leaves at different developmental
stages. Three types of leaves were studied: 'rosette leaves'
from young plants (6-9 weeks old), and immature and
mature (i.e. fully expanded) leaves from older bolting plants
(> 13 weeks old and with stems longer than 45 cm). These
are denoted, respectively, as 'young bolting' and 'old bolting'
leaves. The young bolting leaves have approximately half
the length and width of old bolting leaves.
Before assessing the consequences of increased SOD
activity, it was necessary to determine the level of aug-
mentation in these different leaf types. The results of this
analysis are shown in Figure 3. When SOD activity is
assayed in all the leaves down the stem of bolting plants,
1725

Page 4

C.Bowler et al.
*.f
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~.:..
....
*~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~..:.r-..
,.~~~ ~
~~~~~~~~~~~~~~~~~~~
........
0e
*.......
....
*........
;.S
:--I
:S.-.,w
S..
Fig. 3. MnSOD mRNA and SOD activity profiles in different leaves of MitSODl and ChiSODi plants. Total RNA and total solubleproteins were
isolated from leaves down the stem of bolting plants and from rosette plants (see diagrams) andanalyzed forsteady-state MnSOD mRNA and for
SOD activity. Panels a and c show the results from a MitSODl plant and panels b and d show the data obtained from aChlSODl plant. The
engineered (eng. MnSOD) and endogenous SOD (endog. SOD) activities are indicated. The asterisk shows the position of theendogenous
,senescence-related' MnSOD. The symbol A indicates the type of leaves referred to in the text as 'young bolting', and O denotes 'oldbolting'-type
leaves. Each sample loaded on the protein gels contains 80Agtotal soluble protein. MitSOD2 and ChlSOD2 plants showverysimilarprofiles to
those given here (data not shown). The abundance of MnSOD mRNA in different leaf types is also shown in eachpanel. Transcriptlevels are
always slightly higher in both ChlSOD plants than in the MitSOD plants. Since the number ofcopies of the introduced gene has not been assessed
this may, however, be a result of copy number rather than of a particular effect at the level oftranscription.
it can be seen that the endogenous SOD activity does not
change significantly (Figures 3a and b). However, activity
of the enzyme synthesized from the chimeric MnSOD genes
increases dramatically with age in both MitSOD and ChiSOD
plants. The endogenous MnSOD is too low to be detected
in the amounts ofprotein loaded on these gels but by making
a dilution series of extracts, we estimate that the increase
in MnSOD activity in mitochondria of old bolting leaves is
-30-fold in MitSOD plants. Young bolting leaves from
MitSOD plants only have a 3-fold increase in MnSOD
activity. Similarly, the levels of MnSOD activity in the
chloroplasts of old and young bolting leaves of ChlSOD
plants were estimated to be 50-fold and 10-fold higher,
respectively, when compared with the endogenous MnSOD
activity present in the mitochondria ofthese plants. However,
the total chloroplastic SOD activities (made up of the
combined
activities
of engineered MnSOD
endogenous FeSOD) are only augmented 3-fold and 0.6-fold
in old and young bolting leaves of ChlSOD plants, respec-
tively. The levels of elevated SOD activity in leaves at the
rosette stage from both MitSOD and ChlSOD plants are the
same as for young bolting leaves. The two MitSOD plants
did not differ significantly from each other with respect to
SOD activity levels, nor did the two ChlSOD plants (data
not shown). Also visible on such gels is an additional band
of SOD activity only present in older leaves (indicated by
1726
and
the
an asterisk in Figure 3a and b), migrating slightly faster than
the introduced MnSOD enzyme. This activity is also present
in untransformed plants (data not shown) and our data
suggests that it is a MnSOD. This 'senescence-related'
activity has hitherto been unreported.
In spite of this dramatic gradient of chimeric MnSOD
activity down the stem, the levels of steady-state mRNA are
equivalent in all leaves of MitSOD and ChlSOD plants
(Figure 3a-d), indicating that this phenomenon is a post-
transcriptional event.
Chloroplastic MnSOD provides resistance against
oxidative stress generated within the chloroplasts
To study the ceHular effects ofSOD over-production we have
developed model systems based on the use of methyl
viologen (MV), a herbicide also known as paraquat. It is
well known that during illumination, MV preferentially
accepts electrons from photosystem I and donates them to
oxygen, thus forming the superoxide radical within the
chloroplasts (reviewed in Halliwell, 1984). The development
of the methods to assess the amount of cellular damage
arising from such treatments will be documented in detail
in a separate article (Slooten,L., Bowler,C., Vanden-
branden,S., Van Montagu,M., Inze,D. and Sybesma,C., in
preparation). Essentially, rectangular leafdiscs ofequal size
cut from equivalent leaves of untransformed control,

Page 5

Reduction of cellular damage by MnSOD
w
0
z
c)
I-
0
0
0z0
w
a.
Go
z
cont. mit.
(16)
chi.
(12)
cont. mit.
(16)
chl.
(16)
cont. mit.
(27)
chl.
(27)
(12)
(17)
(27)
ROSETTE
a
YOUNG BOLTING
b
OLD BOLTING
C
Fig. 4. Light-dependent MV-induced increases in specific conductance
expressed as the values x 106 x f-1 cm-l per 35 mg fresh weight.
The values given show the mean + standard error of the mean for
control (cont.), MitSOD (mit.) and ChiSOD (chl.) material at the
rosette (a), young bolting (b) and old bolting (c) stages. The vertical
scale is slightly different for each panel, reflecting the differing
sensitivity of young and old material to MV. The MV concentrations
were 0.5AM(a) and 1.2AM(b and c), and the other experimental
conditions are as outlined in Materials and methods. The number of
leaf discs from which these values were calculated is shown
underneath each bar. NS indicates no significant difference in
comparison with control plants. P denotes the probability that the
results are not significantly different from those obtained with control
plants. These probabilities were calculated using Welch's modification
of Students t-test, applicable to populations with unknown variances
(Welch, 1947). The specific conductance of the floatin solutions of
leaf discs without MV was on the order of 16 x 10-0
35 mg fresh weight. The open boxes represent control plants; the
hatched boxes MitSOD plants, and the stippled boxes ChiSOD plants.
MitSOD and ChiSOD plants were floated on solutions of
MV in Petri dishes and illuminated as described in Materials
and methods. As mentioned in the previous section, we tested
rosette, old bolting and young bolting leaves.
Cellular injury within these leaf discs was assessed firstly
by measuring the conductance ofthe floating solution. This
measures the leakage of ionic solutes out of the cells and
hence gives an indication of membrane damage. Lipid
peroxidation resulting from oxidative stress is likely to initiate
this membrane deterioration. Light-dependent MV damage
as measured by conductance is shown for a typical series
of plants in Figure 4. This figure summarizes data from six
control plants and two plants each of MitSODl and
MitSOD2 (grouped
Ch1SOD2 (collectively ChiSOD). The different controlplants
did not differ significantly from each other in this assay,
likewise the four MitSOD plants and the four ChiSOD plants.
These plants did not show any differences
'background' level of conductance (i.e. from leaf discs
floated on water alone) revealing that membrane integrity
was not altered by changes in SOD activity per se.
Rosette leaves and young bolting leaves, each having
comparatively low SOD overproduction, show only slight
differences with respect to each other (Figure 4a and b):
MitSOD material behaves very similarly to control leaves,
whereas ChiSOD leaves show slightly lessdamage (signific-
ant at 90% probability). However, in old bolting leaves,
where MnSOD activity is maximal, the differences become
much more pronounced (Figure 4c). Leaves from MitSOD
plants are slightly protected against light-dependent MV
damage (significant to 95%) and the resistance of ChiSOD
cm-' per
as MitSOD), and ChlSOD1
and
in their
Oil
0.81-
0.6 -
m~~~~
*
0
00
oo
0e
0.4F
0.2F
0
0
A
0AO
00
0
0
A
50
100
150
200
SPECIFIC CONDUCTANCE
250
Fig. 5. Scattergram of 411 and specific conductance data from leaf
discs for which both parameters were measured. Specific conductance
is expressed as in Figure 4. Symbols: 0,0 control (rosette and old
bolting, respectively); A, MitSOD (rosette); 5, *, ChlSOD (rosette
and old bolting, respectively). Small dots indicate the values obtained
from leaf discs of all types of plants not treated with MV. The 441
values were determined as described in Materials and methods.
material is now highly significant at 99.95% probability.
These leaves show
These data imply that the degree of damage is negatively
correlated with the level of MnSOD overproduction. The
observation that the effects of ChiSOD overproduction are
much more pronounced than equivalent levels of MitSOD
overproduction is consistent with the fact that the formation
of superoxide radicals is being biased to the chloroplasts in
light conditions.
To examine
chloroplast-localized
chlorophyll fluorescence techniques to determine the integrity
of various components of the photosynthetic apparatus (for
review, see Krause and Weis, 1984). Several parameters
were studied (Slooten,L., Bowler,C., Vandenbranden,S.,
Van Montagu,M., Inze,D. and Sybesma,C., in preparation)
and one example, 41H, the quantum yield ofexciton trapping
by the reaction center ofphotosystem II, will be documented
here. This parameter essentially gives a measure of the
integrity of the photosystem II reaction center. In some
experiments, after incubation in MV, the fluorescence
properties of the leaf discs were analyzed and -II values
determined. Hence, for these experiments we have a value
for conductance and for
plotted in a scatter diagram (Figure 5), it is clear that these
two parameters are correlated,
becomes progressively damaged (low
membrane damage also increases (high conductance).
Another important conclusion from this analysis is that
the data from all types of leaves from control, MitSOD and
ChISOD plants are scattered randomly along the line, except
for the values from old bolting ChiSOD leaves. These values
are uniformly grouped together at the upper left end of the
plot, the position corresponding to minimal cellular damage.
Hence high level MnSOD overproductionin thechloroplasts
-2.5-fold less damage than the controls.
effects, we
used
II. When the data from these are
i.e. as photosystem II
4II values), cell
1727

Page 6

5-
0°
gCont.
go
00
i:
S-/
ChtSOD2
5- 1S
0-
I5 HOURS
ILLUMINATION
0
5
l0 o3o0o0g
METHYL VIOLOGEN (uM)
Fig. 6. Light-dependent MV-induced increases in pheophytin in leaf
discs derived from old bolting control (0) and ChlSOD2 (@) leaves.
(a) Percentage of chlorophyll pheophytinized as a function of
illumination time in the presence of 5/AMMV. Experimental
conditions were exactly as described in Materials and methods, except
for these alterations in illumination time. (b) Pheophytin appearance
induced by increasing concentrations of MV with 5 h of illumination.
The dashed vertical line distinguishes between the two scales used on
the horizontal axis.
can protect both the photosynthetic apparatus and the cellular
membranes from MV-generated damage in this model
system.
Furthermore, we have examined the effects of oxygen
radical stress on the pigment content of leaves. One of the
consequences of the light incubations in the presence ofMV
is that in the subsequent dark period, chlorophyll
pheophytinized, i.e. it loses its central magnesium atom. It
has been shown recently that free fatty acids destabilize the
chloroplastic pigment-protein complexes, rendering the
chlorophyll susceptible to pheophytinization (Siefermann-
Harms, 1990). We assume that lipid peroxidation resulting
from oxidant damage to membranes is mediating this
destabilization and consequent pheophytin formation.
Figure 6 shows a representative example of pheophytin
formation in leaf discs derived from old bolting leaves of
control and ChiSOD plants as a function of illumination time
(Figure 6a) and MV concentration (Figure 6b). In both cases
less pheophytin is formed in the ChiSOD material than in
the untransformed leaf material.
In summary, we have shown that in this model system,
where MV damage is light-dependent, high level overpro-
duction of MnSOD in the chloroplasts can significantly
reduce the normal level ofoxidative damage, protecting both
the compartment where free radicals were generated and the
cellular membranes as well. The experiments described
above have been repeated with three sets of plants and in
all cases the results are comparable to those documented here
(data not shown).
is
Effects of MnSOD on the damage generated by
methyl viologen in the dark
Although the light-mediated MV toxicity is primarily a
consequence of electrons being donated to MV from
photosystem I, MV may also be accepting some electrons
from a mitochondrial source, most likely the respiratory
electron transport chain, and hence forming the superoxide
anion within this organelle. By incubating leaf discs with
MV in the dark, it may be possible to analyze the effects
of superoxide formation in the mitochondria since photo-
synthetic electron transport will be inoperative. Indeed, when
leaf discs are incubated with MV in the dark in the presence
of inhibitors ofmitochondrial electron transport, the resulting
180
140
,
120
100
z
0
80
o
80
tL
L)
40
20
0
cont. mit.
(11)
chl.
(11)
cont. mit.
(12)
chi.
(12)
cont. mit.
(28) (28) (28)
chi.
(10)
(12)
ROSETTE
a
YOUNG BOLTING
b
OLD BOLTING
C
Fig. 7. MV-induced increases in specific conductance in dark
experiments. All parameters are denoted as in Figure 4. The MV
concentrations were 0.5 /.tM (a) and 1.2 AM (b and c).
damage is reduced by -30%, whereas no reduction was
observed in the light (Slooten,L., Bowler,C., Vanden-
branden,S., Van Montagu,M., Inze,D., and Sybesma,C.,
in preparation).
Having developed a reproducible model system for
carrying out such dark incubations (see Materials and
methods) we then examined the damage generated in control,
MitSOD and ChiSOD material. Using the same plants as
were studied in the light experiments, rosette, young bolting
and old bolting leaves were again examined using con-
ductance as a measure for cellular damage. The results in
Figure 7 show clearly that different effects are observed in
the dark than were previously seen in the presence of light
(compare with Figure 4). Material in which MnSOD is only
moderately overproduced, either in chloroplasts or mito-
chondria (i.e. rosette and young bolting leaves) is actually
more sensitive to MV than control material (Figure 7a and
b), indicating that low levels ofMnSOD overproduction are
detrimental for the cells during these dark incubations.
However in old bolting leaves displaying maximal MnSOD
activity the effects are reversed: both mitochondrial and
chloroplastic MnSOD now provide high level protection
(Figure 7c) significant at the 99.95% probability levels.
These results imply that the level ofMnSOD overproduction
plays a critical role in determining the effects which follow
in this model system. Additionally, since chloroplastic
MnSOD has a similar influence to the mitochondrial
MnSOD, it would appear that superoxide is still present
within the chloroplasts in dark conditions. For each leaf disc,
fluorescence measurements were also made, again revealing
a direct correlation between 411 values and conductance
values (data not shown). Once more the conclusion is that
whilst 02' generation may be confined to specific cell
locations under different physiological conditions, the
damage that is produced pervades the entire cell.
Discussion
The potential use ofSOD for the genetic engineering ofplant
stress resistance has been often suggested but never
thoroughly investigated. In this study we have chosen to
evaluate the properties of MnSOD overproduction in plant
mitochondria and chloroplasts. MnSOD was chosen rather
C.Bowler et al.
a
a
0 ;-
T
tL
-i
T
b 3!
Z 2
Ii
15
30
45
60
ILLUMINATION TIMElmin.)
1728

Page 7

Reduction of cellular damage by MnSOD
than Cu/ZnSOD or FeSOD because it is not inactivated by
the reaction product H202, as are the other two enzymes.
Consequently any effects observed can be thought of simply
as a result ofa change in the ratio between the SOD substrate
and product (i.e. 02
and H202), this change being directly
proportional to the level of SOD augmentation and being
unaffected by H202 concentration.
The effects of MnSOD overproduction were analyzed on
leaf discs using MV as a superoxide radical-generating agent.
This reproducible model system to study the effects of
oxygen radical formation at the cellular level has several
advantages over the use of whole plants. First, it is easier
to quantify the amount of MV applied to a leaf disc than
it is to quantify that applied to an intact leaf or an intact plant.
Experiments on whole plants show much variation and
produce large statistical errors which can only be averaged
out by using a much larger number of plants than can be
handled in laboratory conditions. A second problem ofMV
application on whole plants is that the degree of damage
caused by the compound is light dependent. Since all leaves
will not receive the same amount of light, the observed MV
damage will greatly vary.
We were able to test the effects of different levels ofSOD
overproduction because of the unexpected gradient of
MnSOD activity which exists from young to old leaves of
transgenic plants. Although its cause is not known, its
experimental exploitation has revealed that the level ofSOD
overproduction can be instrumental in determining the
amount ofdamage generated by the redox-active compound
methyl viologen (MV). MV acts by accepting electrons and
donating them to oxygen, thus forming the superoxide anion.
Its effects have been studied in leaf discs incubated in light
and dark conditions. The most surprising observation we
made was that both positive and negative effects can arise,
apparently dependent upon the level ofSOD overproduction.
To what extent do these results conform to the theories
currently held of oxygen radical toxicity? Central to these
ideas is a metal (M) ion-catalyzed Haber-Weiss reaction,
involving the formation of highly reactive hydroxyl radicals
(OH *) from the comparatively unreactive superoxide anion
and hydrogen peroxide (Halliwell, 1987; Imlay and Linn,
1988):
027 + H202Mn
M(n)
OH. + OH- + 02
Hence, SOD would be likely to play an important role in
cellular defense against oxidative stress since it has the unique
property of controlling the ratio of 02
Haber-Weiss reaction substrates. The remainder of this
discussion will examine the support which our results give
to these ideas and will attempt to explain other contradictions
previously reported from both prokaryotic and eukaryotic
systems.
In our light experiments, the superoxide anion is likely
to be formed predominantly within the chloroplasts due to
the high rate of electron transfer through the photosystems.
In the absence of sufficient SOD, the presence of both O2
and H202 would ensure the formation of OH
then initiate destructive processes. The consequence of
MnSOD over-production in the chloroplasts would be to
reduce the levels of02
The H202 can then be controlled by the H202-scavenging
pathway ofchloroplasts, which has been shown to be active
in the light but not in the dark (Foyer and Halliwell, 1976;
and H202, the two
which can
and to increase the levels ofH202.
Nakano and Asada, 1980). SOD can thus act synergistically
with this pathway to remove the Haber-Weiss reaction
substrates, thereby reducing the damage generated by MV
by amounts correlating with its level of overproduction.
In dark experiments this simple relationship between SOD
activity levels and the extent of damage does not exist. It
appears that low levels of chloroplastic MnSOD are
detrimental, but above a certain level this effect is completely
reversed.
Since chloroplastic H202-scavenging
operational in the dark, it is conceivable that low level
MnSOD overproduction merely serves to upset the normally
optimized balance between 02
the formation of OH-
more likely and consequently
increasing the damage. High level overproduction would
shift the 02 :H202 equilibrium almost completely over to
H202, thus ensuring that little 02
Haber-Weiss reaction and resulting in a reduction in cellular
damage. These effects were clearly observed (Figure 7).
Since the chloroplastic MnSOD shows such striking effects
in our dark experiments, a significant proportion of super-
oxide must still be present in the chloroplast. It is possible
that this is generated elsewhere and diffuses in, following
its protonation (pKa = 4.7), or that it is generated within
the chloroplast. Isolated chloroplasts have been shown to
reduce MV in the dark (Law et al., 1983) suggesting that
02
could in fact be generated in situ. A chlororespiratory
electron transport pathway is known to be stimulated in the
dark in photosynthetic prokaryotes and green algae and
recent evidence suggests that it is also present in plant
chloroplasts (Garab et al., 1989). This may then provide a
chloroplastic source of electrons for MV in the dark.
The effects of increased mitochondrial MnSOD activity
are not particularly pronounced in the light experiments
presumably because the majority of superoxide is being
generated within the chloroplasts. However when activity
levels are high, even this spatially restricted enzyme appear
to provide low levels of protection, perhaps destroying the
superoxide formed within mitochondria or acting as a sink
for superoxide diffusing in from the cytosol. In the dark,
however, the effects are more pronounced and are essentially
the same as observed for the chloroplastic MnSOD-low
levels being detrimental whilst high levels are beneficial.
Although almost nothing is known about H202-scavenging
within plant mitochondria, the results can again be explained
by the perturbation of the normal02 :H202 ratio, i.e. low
MnSOD activity levels changing the equilibrium closer to
1:1, favoring OH
formation, and high MnSOD activity
levels ensuring the presence of only low amounts of 027
and hence reduced OH
formation.
This model is fully consistent with the hypothesis that the
Haber-Weiss reaction is the prime culprit for oxygen
toxicity in vivo. It also incorporates the phenomenon oflight-
dependent
H202
scavenging
acknowledges the possible existence ofchlororespiration in
plants. Additionally both positive and negative effects of
Cu/ZnSOD overproduction have been observed in human
and mouse cells (Elroy-Stein et al., 1986)and with MnSOD
in E. coli (Bloch and Ausubel, 1986; Gruber et al., 1990),
in all cases the likely cause beingthe differences in the levels
of SOD overproduction, i.e. the particular consequence is
dependent on how SOD affects the ratio between 02
H202,
either
making
OH-
Haber-Weiss reaction more likely or less likely.
is not
and H202, thereby making
is available for the
in
chloroplasts
and
and
theformationfrom
1729

Page 8

C.Bowler et al.
Nevertheless, other factors should be considered. Obvious
differences between the above mentioned experiments are
the type of SOD studied, either Cu/ZnSOD or MnSOD, and
the subcellular location.
Cu/ZnSOD is inactivated by H202 whereas MnSOD is not.
This may be the possible reason why small increases of
Cu/ZnSOD can provide resistance against MV in human and
mouse cells whereas higher levels cannot (Elroy-Stein et al.,
1986). As such, in high producing lines the burst of H202
generated by the combined action of MV and SOD may
simply inactivate the Cu/ZnSOD, leading to the existence
of both O2
and H202 within the cell, and hence hydroxyl
radicals. This can also explain the discrepancy in the results
obtained by Tepperman and Dunsmuir (1990) in comparison
with ours. They analyzed the effects of high level over-
production of a Cu/ZnSOD in tobacco chloroplasts in
systems analogous to ours and found neither positive nor
negative effects with regard to light-dependent MV damage
to plant cells. In the leaves examined, the level of over-
production appeared comparable to the high levels which
we observed in older leaves, so again the lack of any effect
may simply be due to the inactivation of the Cu/ZnSOD by
H202.
Unfortunately, Cu/ZnSOD activity levels in the
material before and after stress treatment was not examined
in either case.
The subcellular location is also likely to play an important
role in determining whether positive or negative effects arise
from SOD overproduction. Furthermore, not only the
amount of SOD present within the different
compartments, but also their endogenous H202-scavenging
abilities will influence the turnover and accumulation of
reactive products as well as the continued activity of SOD
under stress conditions. Therefore, it remains possible that
the dependence of the observed effects on leaf age reflects
differences not only in SOD activity levels but also in the
physiological states of the different leaf types.
In summary, it is clear that SOD can be manipulated to
give resistance to oxidative stress although there is clearly
a fine line between benefit and injury, dependent upon the
level of overproduction, the type of SOD used and the
endogenous scavenging systems of the organism, with
everything being ultimately dependent upon the O2 :H202
ratio. Our results support the use of MnSOD expressed at
high levels as the most promising route of SOD for
developing stress tolerance in plants. Also, since our model
systems have revealed both mitochondrial and chloroplastic
MnSOD to be beneficial,
contains elevated MnSOD in both organellar compartments
will be even more resistant to MV. This awaits analysis,
as does the effects of cytosolic MnSOD overproduction.
Finally it remains to be seen whether visible phenotypes
will be observed at the whole plant level with regard to
economically relevant stress conditions such as chilling, high
light intensities and pollutants such as ozone. Plant varieties
which are tolerant to MV (in the form of the herbicide
paraquat) often show cross-tolerance to adverse environ-
mental conditions, thus implicating the involvement of
oxidative damage in many of these processes (e.g. Tanaka
et al., 1988; Shaaltiel et al., 1988), and suggesting that the
transgenic plants which we have generated may also show
similar phenotypes. However,
protection observed at the cellular level will be sufficient
toprovidesuch a resistant phenotype at the whole plant level
It is prudent to remember that
cellular
it is likely that a plant which
whether the degree of
remains to be tested and it may prove necessary to supple-
ment an engineered SOD enzyme with other oxygen-
detoxifying enzymes such as catalase or peroxidase to
produce a desirable phenotype.
Materials and methods
Construction of MnSOD expression cassettes
Expression cassettes were constructed in binary vectors pGSJ780A (derived
from pGSFR780A; De Block et al., 1989) and pGSC1702 (derived from
pGSC1700; Cornelissen and Vandewiele, 1989). Between the T-DNA
borders, each of these vectors contains the neomycin phosphotransferase
II (nptll) gene under control of the nopaline synthase (nos) promoter
(Depicker et al., 1982) for kanamycin selection of transformed cells. In
addition, pGSJ780Acontains the cauliflower mosaic virus (CaMV) 35S
promoter (from pGSJ280;Deblaere et al., 1987) immediately upstream of
unique ClaI and BamHI sites which can be used, respectively, to create
transcriptionalor translational fusions to foreign protein-coding sequences.
pMitSOD was constructed by inserting the entire coding sequence of the
MnSOD cDNA as a 960bp HpaI-SmaI fragment from pSODI (Bowler
etal., 1989a)into the blunt-ended, dephosphorylated ClaI site of pGSJ780A
(Figure la).Thisgeneratesa transcriptional fusion of the MnSOD to the
35S promoter.
To constructpChlSODit was necessary to create a fusion between the
MnSOD mature protein and a transit peptide for chloroplast targeting. This
was doneby cutting pSODIwith Sad, which recognizes a unique site at
the junction between the mitochondrial leader sequence-encoding DNA and
the mature MnSOD-encoding sequence, flushing with T4 DNA polymerase
and adding an octameric Bgll linker to create plasmid pSODB. The mature
protein-encoding sequencewas then inserted as a BgllI-BamHI fragment
into thedephosphorylatedBamHI site of pKAHI(J.Botterman, unpublished
data).This BamHI site is immediately downstream of a sequence encoding
the35S promoter coupledto the sequence encoding the transit peptide of
a small subunitgeneofribulose-1,5-bisphosphate carboxylase (SStp) from
pea (Cashmore, 1983).Thisgeneratesan in-frame fusion of the chloroplast
transit peptidewith the MnSODmature protein. Once processed the MnSOD
will contain four additional N-terminal amino acids (Met-Asp-Leu-Gly)
generatedas aconsequenceof the cloning procedure (Figurelb). The entire
35S-SStp-MnSODcassette was then isolated as a BglII-BamHI fragment
and inserted into the dephosphorylated BamHI site of binary vector
pGSC1702, yielding pChlSOD.
Downstream from the MnSODsequences within pMitSOD and pCh1SOD
are thepolyadenylation signalsfrom the T-DNA-encoded gene 7 (Dhaese
etal., 1983)which are in addition to the endogenous MnSOD 3' sequences
also present in each construct.
Transformation and propagation of plant material
Constructions were mobilized toAgrobacterium tumefaciens by 'menage
atrois' and theresultingstrains were used to transform N.tabacum cv. PBD6
byleaf disc transformation (De Block et al., 1987). Cuttings taken from
regenerated plants were transferred to soil and grown under standard
greenhouse conditions. Primary transformants were maintained and
multiplied vegetatively.
Proteinanalysis, subcellular fractionations and immunogold
labelling
Solubleproteinswere extracted from plant material and assayed for SOD
activityon native 10%polyacrylamide gels using the in situ staining technique
ofBeauchampand Fridovich (1971) as previously described (Van Camp
et al., 1990). For subcellular fractionation, leaf material (20 g) was
homogenizedin 80 ml ice-cold homogenization buffer [50 mM Tris-HCl,
pH 8.0, 0.33 Msucrose, 0.05% 13-mercaptoethanol, 0.1% bovine serum
albumin(BSA)], usingaWaringblender (three low-speed pulses, 3 s each).
After filtrationthrough two layers of Miracloth®, the homogenate was
centrifugedin anSS34 rotor at 6000 r.p.m. for 30 s; chloroplasts were
thenpreparedfrom thepelleton two-step Percoll gradients essentially as
described (Mullet and Chua, 1983). A crude mitochondrial pellet was
obtained from thesupernatantof the initial low-speed centrifugation by an
additionalcentrifugationat 15 000 r.p.m. for 12 min in an SS34 rotor. The
pellet was resuspended in 2mlsuspension buffer (50 mMHEPES-KOH,
pH 8.0, 0.33 M sorbitol, 0.1% BSA) and loaded onto two-step Percoll
gradients aspreviously described (Bowler et a!., 1989a). The miitochondrial
fraction was removed from the interface and was diluted 10-foldin suspension
buffer. Both chloroplasts and mitochondria were pelleted by centrifugation
1730

Page 9

Reduction of cellular damage by MnSOD
as before and were washed once more in suspension buffer. They were
then resuspended in lysis buffer (25 mM Tris-HCI, pH 7.5, 0.5% fi-
mercaptoethanol) and insoluble material was removed by a furthercentrifuga-
tion step (13 000 r.p.m. for 12 min). All manipulations were carried out
at 40C.
Immunogold labelling of sections derived from leaves of control and
transgenic plants were performed using polyclonal antibodies raised in rabbits
against recombinant N.plumbaginifolia MnSOD purified from yeast (Bowler
et al., 1989b) and following the procedures described in De Clercq et al.
(1990).
RNA analysis
Total RNA was prepared from fresh tissue as described by Jones et al. (1985)
and quantified spectrophotometrically. For Northern analysis, 12
denatured in formaldehyde, electrophoresed and transferred to nylon
membranes according to the manufacturer's recommendations (Amersham,
Bucks, UK). To obtain a highly specific probe, the internal HpaI-HindIll
fragment from pSODI (Bowler et al.,
SmaI-HindIII sites of pGem2 and 32P-labeled single-stranded riboprobes
were
synthesized using T7 RNA polymerase (Promega).
hybridizations were performed as described (Bowler et al., 1989a).
tg was
1989a) was recloned in the
Northern
Methyl viologen treatment
Rectangular leaf discs of equal size (0.7 x 2.2 cm) were cut with a new
surgical blade from leaves of soil-grown plants and floated on 0.4 M sorbitol.
Each was then weighed and floated individually, top side up, on 3 ml water
and methyl viologen (MV) solutions in 35 mm Petri dishes.
For light experiments leaf discs were first incubated in the dark for 16 h
at room temperature ( 4 230C) to allow diffusion of the MV into the leaf,
after which they were placed under a white light source for 2 hours (at
40 cm distance from two 20W cool-white fluorescent tubes; light intensity
2.7 W/m2). Thereafter 20AM3-(3,4-dichlorophenyl)-1,1'-dimethyl urea
(DCMU) was added (final ethanol concentration 0.1 S% by volume) and the
leaf discs were returned to the dark for 16 hours at 30°C. Ion leakage
occurred in the course of this second dark incubation and was accelerated
by the elevated temperature. DCMU was added in order to inhibit the
reoxidation ofQA, the first stable electron acceptor within the reaction center
of photosystem H (Velthuys, 1981; Pfister et al., 1981; Oettmeier and Soll,
1983). This was necessary for the fluorescence measurements and had no
effect on ion leakage. In dark controls containing MV (which were given
the same treatment except that they were not illuminated) no damage was
observed by any of the criteria used, showing that all damage in these
experiments was light mediated.
For dark experiments, leaf discs were incubated in complete darkness
for 68 h at 30°C. This extended period and elevated temperature was
necessary to generate sufficient quantifiable damage. DCMU was added
after 54 h of dark incubation.
As mentioned in the text, three sets of control, MitSOD, and ChISOD
plants were tested during the course of these experiments. Although the
results given are for only one of these sets, they are highly representative
of the results obtained for each series. Nonetheless, the sensitivity of the
leaf discs varied with age (the older the material the more resistant) and
season (the higher the light intensity the more resistant) so it was necessary
each time to determine the optimal non-saturating MV concentrations to
use in order to reveal the differences. This lay between 0.5 and 3JIMfor
our experimental conditions.
Conductance and fluorescence measurements
The solution on which the leaf discs had been floating was collected, made
up to 3 ml (to correct for evaporation) and conductance was measuredusing
KCI as a standard. The contribution of MV alone to the conductance was
corrected for (this amounted to <5% of the reported values) and
conductances were calculated per 35 mg fresh weight.
For fluorescence measurements, leaf discs were transferred to Petri dishes
containing water plus 5 AMDCMU and allowed to preadapt at room
temperature in the dark for at least 1 h. Fluorescence of the leaf discs was
excited by a broad band (400-650 nm) of blue light, and measured at
wavelengths above 690 nm. The excitation light was switched on and off
by means of an electronic shutter which opened within 2.5 ms. Theoutput
of the photomultiplier was fed into a storage oscilloscope (Nicolet Explorer
IILA) and then plotted on an X-Y plotter. Two measurements were made
for each disc. For the first, the excitation light intensitywas reduced 22-fold
by means of a perforated aluminium disc, and thephotomultiplier output
was amplified by the same factor (by adjustment of the photomultiplier
voltage). This allowed us to measure F0, the 'dark' fluorescence level. For
the second measurement, we used the full light intensity (40 W/m2) with
no amplification of the signal output. This allowed us to measure Fmax,
the maximum fluorescence level. MII was calculated as (Fmax-Fo)/Fmax
(Genty et al., 1989).
In some experiments we used a pulse-amplitude modulated fluorimeter
(PAM101, Walz, Effeltrich, FRG) as described (Schreiber et al., 1986),
using leaf discs without DCMU. The results obtained with this apparatus
were the same as those obtained with the above described method.
Pheophytin measurements
Pigment extracts were prepared in chloroform:methanol essentially according
to Bligh and Dyer (1959) and the absorbance spectra of the extracts were
measured using a Cary spectrophotometer (Model 2300). The percentage
of pheophytinized chlorophyll was calculated from the ratioAE538/E665,
where AE538 is the absorbance at 538 nm relative to a straight line
connecting the two adjacent minima (at -525 and 553 nm) andE665is
the adsorbance at 665 nm (for chlorophyll, AE538 is zero). This calculation
was based upon a calibration experiment in which all the chlorophyll in
a pigment extract from untreated leaf discs (prepared as described above)
was converted to pheophytin, and absorbance spectra were taken before
and after pheophytinization (Slooten,L., Bowler,C., Vandenbranden,S., Van
Montagu,M., Inze,D. and Sybesma.C., in preparation).
Acknowledgements
We
Pamela Dunsmuir and Barry Halliwell for their constructive advice on our
experiments, ideas and final manuscript. We thank Martine De Cock,
Vera Vermaercke, Stefaan Van Gijsegem and Karel Spruyt for their
invaluable help for the preparation of this manuscript, Paul Holvoet for the
MnSOD antiserum, and Hilde Van de Wiele and Chris Genetello for plant
transformation. This work was supported by grants from the International
Atomic Energy Agency (no. 5285), the Services of the Prime Minister (UIAP
120CO187), and the 'ASLK-Kankerfonds' to M.V.M., and 'Fonds voor
Geneeskundig Wetenschappelijk Onderzoek' (FGWO 33.0104.90)
M.V.M. and C.S. C.B. is supported by an SERC (NATO) Predoctoral
Overseas Fellowship;
L.S.
29.0221.0530) and D.I. is a Research Director of the Institut National de
la Recherche Scientifique (France).
are
greatly
indebted
to
Allan
Caplan,
Arlette
Reynaerts,
to
is supported by the FGWO (grant
no.
References
Asada,K. and Takahashi,M. (1987) In Kyle,D.J., Osmond,C.B. and
Arntzen,C.J. (eds), Photoinhibition. Elsevier Science Publishers B.V.,
Amsterdam, pp. 227-287.
Avraham,K.B., Schickler,M., Sapoznikov,D., Yarom,R. and Groner,Y.
(1988) Cell, 54, 823-829.
Bannister,J.V., Bannister,W.H. and Rotilio,G. (1987) CRC Crit. Rev.
Biochem., 22, 111-180.
Beauchamp,C.O. and Fridovich,I. (1971) Anal. Biochem., 44, 276-287.
Biliiski,T., Krawiec,Z., Liczmaniski,A. and Litwinska,J. (1985) Biochem.
Biophys. Res. Commun., 130, 533-539.
Bligh,E.G. and Dyer,W.J. (1959) Can. J. Biochem. Physiol., 37, 911-917.
Bloch,C.A. and Ausubel,F.M. (1986) J. Bacteriol., 168, 795-798.
Bowler,C., Alliotte,T., De Loose,M., Van Montagu,M. and Inze,D. (1989a)
EMBOJ., 8, 31-38.
Bowler,C., Alliotte,T., Van den Bulcke,M., Bauw,G., Vandekerckhove,J.,
Van Montagu,M. and Inze,D. (1989b) Proc. Natl. Acad. Sci. USA, 86,
3237-3241.
Cadenas,E. (1989) Annu. Rev. Biochem., 58, 79-110.
Carlioz,A. and Touati,D. (1986) EMBO J., 5, 623-630.
Cashmore,A.R. (1983) In Kosuge,T., Meredith,C.P. and Hollaender,A.
(eds) Genetic Engineering ofPlants. AnAgricultural Perspective. Plenum
Press, New York, pp. 29-38.
Chia,L.S., Mayfield,C.I. and Thompson,J.E. (1984)Plant Cell Environ.,
7, 333-338.
Clare,D.A., Rabinowitch,H.D. and Fridovich,I. (1984) Arch. Biochem.
Biophys., 231, 158-163.
Cornelissen,M. and Vandewiele,M. (1989)Nucleic Acids Res., 17, 19-29.
Deblaere,R., Reynaerts,A., Hofte,H., Hernalsteens,J.-P., Leemans,J. and
Van Montagu,M. (1987) Methods Enzymol., 153, 277-292.
De Block,M.,
Botterman,J.,
Vandewiele,M.,
Gossele,V., Movva,R., Thompson,C.,VanMontagu,M.and Leemans,J.
(1987) EMBO J., 6, 2513-2518.
De Block,M., De Brouwer,D. andTenning,P. (1989)PlantPhysiol., 91,
694-701.
De
Clercq,A.,
VanMontagu,M., Krebbers,E. and Vandekerckhove,J. (1990) Plant
Physiol., 92, 899-907.
Dockx,J.,
Thoen,C.,
Vandewiele,M.,
De
Rycke,R.,
Van
Damme,J.,
1731

Page 10

C.Bowler et al.
Depicker,A., Stachel,S., Dhaese,P., Zambryski,P. and Goodman,H.M.
(1982) J. Mol. Appl. Genet., 1, 561 -573.
Dhaese,P., De Greve,H., Gielen,J., Seurinck,J., Van Montagu,M. and
Schell,J. (1983) EMBO J., 2, 419-426.
Elroy-Stein,O. and Groner,Y. (1988) Cell, 52, 259-267.
Elroy-Stein,O., Bernstein,Y. and Groner,Y. (1986) EMBOJ., 5, 615 -622.
Forman,H.J. and Boveris,A. (1982) In Pryor,W.A. (ed.), Free Radicals
in Biology. Academic Press, New York, pp. 65-90.
Foyer,C.H. and Halliwell,B. (1976) Planta, 133, 21-25.
Freeman,B.A. and Crapo,J.D. (1982) Lab. Invest., 47, 412-426.
Garab,G., Lajk6,F., Mustardy,L. and Marton,L. (1989) Planta, 179,
349-358.
Genty,B., Briantais,J.-M. and Baker,N.R. (1989) Biochim. Biophys. Acta,
990, 87-92.
Gruber,M.Y., Glick,B.R. and Thompson,J.E. (1990) Proc. Natl. Acad.
Sci. USA, 87, 2608-2612.
Halliwell,B. (1984) Chloroplast Metabolism-The Structure and Function
of Chloroplasts in Green Leaf Cells. Clarendon Press, Oxford.
Halliwell,B. (1987) FASEB J., 1, 358-364.
Harbour,J.R. and Bolton,J.R. (1975) Biochem. Biophys. Res. Commun.,
64, 803-807.
Imlay,J.A. and Linn,S. (1988) Science, 240, 1302-1309.
Jones,G.R.N. (1986) Med. Hypothesis, 21, 267-271.
Jones,J.D.G., Dunsmuir,P.
and Bedbrook,J.
2411-2418.
Krause,G.H. (1988) Physiol. Plant., 74, 566-574.
Krause,G.H. and Weis,E. (1984) Photosynthesis Res., 5, 139-157.
Ksenzenko,M., Konstantinov,A.A., Khomutov,G.B., Tikhonov,A.N. and
Ruuge,E.K. (1983) FEBS Lett., 155, 19-24.
Law,M.Y., Charles,S.A. and Halliwell,B. (1983) Biochem. J., 210,
899-903.
Lee,E.H. and Bennett,J.H. (1982) Plant Physiol., 69, 1444-1449.
Mehler,A.H. (1951) Arch. Biochem. Biophys., 33, 65-77.
Menzel,D.B. (1976) In Pryor,W.A. (ed.), Free Radicals in Biology.
Academic Press, New York, Vol II, pp. 181-202.
Monk,L.S., Fagerstedt,K.V. and Crawford,R.M.M. (1987) Plant Physiol.,
85, 1016-1020.
Mullet,J.E. and Chua,N.-H. (1983) Methods Enzymol., 97, 502-509.
Nakano,Y. and Asada,K. (1980) Plant Cell Physiol., 21, 1295-1307.
Oettmeier,W. and Soll,H.-J. (1983) Biochem. Biophys. Acta, 724, 287-290.
Orr,G.L. and Hogan,M.E. (1983) Pestic. Biochem. Physiol., 20, 311-319.
Pfister,K., Steinback,K.E., Gardner,G. and Arntzen,C.J. (1981) Proc. Natl.
Acad. Sci. USA, 78, 981-985.
Phillips,J.P.,
Campbell,S.D.,
Michaud,D.,
Hilliker,A.J. (1989) Proc. Natl. Acad. Sci. USA, 86, 2761 -2765.
Rabinowitch,H.D., Sklan,D. and Budowski,P. (1982) Physiol. Plant, 54,
369-374.
Schreiber,U., Schliwa,U. and Bilger,W. (1986) Photosynth. Res., 10,
51-62.
Shaaltiel,Y., Glazer,A., Bocion,P.F. and Gressel,J. (1988) Pestic. Biochem.
Physiol., 31, 13-23.
Siefermann-Harms,D. (1990) In Baltscheffsky,M. (ed.), Current Research
in Photosynthesis. Kluwer, Dordrecht, Vol. II, pp. 245-248.
Tanaka,K. and Sugahara,K. (1980) Plant Cell Physiol., 21, 601-611.
Tanaka,K., Furasawa,I., Kondo,N. and Tanaka,K. (1988) Plant Cell
Physiol., 29, 743-746.
Tepperman,J.M. and Dunsmuir,P. (1990) Plant Mol. Biol., 14, 501-511.
Van Camp,W., Bowler,C., Villarroel,R., Tsang,E.W.T., Van Montagu,M.
and Inze,D. (1990) Proc. Natl. Acad. Sci. USA, 87, 9903-9907.
van Loon,A.P.G.M., Pesold-Hurt,B. and Schatz,G. (1986) Proc. Natl.
Acad. Sci. USA, 83, 3820-3824.
Velthuys,B.R. (1981) FEBS Lett., 126, 277-281.
Welch,B.L. (1947) Biometrika, 34, 28-35.
Wong,G.H.W., Elwell,J.H., Oberley,L.W. and Goeddel,D.V. (1989) Cell,
58, 923-931.
(1985) EMBO J.,
4,
Charbonneau,M.
and
Received on February 1, 1991; revised on April 2, 1991
1732