ABSTRACT
We examined DNA repair activities of a mitochondrial lysate derived from
Xenopus laevis
oocytes. Plasmid DNA, exposed to HCl, H
2
O
2
or UV light, was used as the substrate for the
in vitro
repair reaction. DNA synthesis in the lysate was stimulated 2-8-fold by such lesions, indicating the presence of excision repair
activities. This repair DNA synthesis was not affected by aphidicolin, but was
sensitive to
N
-ethylmaleimide. Thus the mitochondrial DNA polymerase, i.e., pol
[gamma]
is indeed involved in the reaction. Actual repair of the depurinated DNA was
demonstrated by using the polymerase chain reaction (PCR), where the amount of
the amplified DNA fragment increased significantly if the depurinated template
was incubated in the lysate prior to the PCR. UV-irradiated DNA, on the other hand, restored its ability as a PCR template
only if the repair reaction was carried out under the light. Therefore, in this
system, UV-induced damage is repaired mainly by photoreactivation. These results show
that mitochondria of
Xenopus
oocytes possess excision repair as well as photolyase activities, and that the
in vitro
repair system described here should be useful for further molecular
characterization of such DNA repair machinery.
The animal cell contains several hundreds of mitochondria, each having multiple
copies of circular DNA. This DNA carries 37 essential genes including those for
the components of the electron-transport chain, and thus its stable maintenance is important for survival
of the cell. However, unlike the nuclear DNA, the mitochondrial genome is not
protected by histones from various genotoxic agents. Furthermore, the oxidative
environment inherent to this organelle creates very unfavorable conditions for
the stability of DNA (
1
). Therefore it is reasonable to assume that mitochondria have some effective
means of repairing DNA damage frequently generated in their genome. Defects in
such mechanisms may be the cause of sequence alterations like those found in
patients afflicted with mitochondrial encephalomyopathy or in normal adults of
advanced age (
2
,
3
).
Because of their compartmentalized structure, mitochondria are presumed to have
their own repair machinery distinct from that of the nucleus. Although early
studies did not reveal the activity to repair pyrimidine dimers in mitochondria
(
4
,
5
), several repair-related enzymes such as AP endonuclease, UV endonuclease, uracil DNA
glycosylase and methyl transferase have been isolated and characterized (
6
-
8
). It has further been reported that mitochondria have the ability to remove DNA
damage generated by alkylation or oxidation (
8
-
10
). More recently, a homologue of the
Escherichia coli
MutS protein has been purified from yeast mitochondria, suggesting that a
mismatch repair pathway is also operative (
11
). Despite all these findings, systematic studies of each repair pathway as a
whole have not been fully explored. We decided to initiate such studies
employing an
in vitro
repair system derived from isolated mitochondria.
Oocytes of
Xenopus laevis
are well suited for the source of mitochondria, because they have
an extremely large number of this organelle that accumulates up to 10
7
per oocyte in the course of oogenesis (
12
-
14
), and are available in large quantities throughout the year. Furthermore it was
shown that a lysate of oocyte mitochondria contains the activity of replicating
mitochondrial DNA (
15
). Thus such an
in vitro
system may also be useful for biochemical studies of the DNA repair machinery.
In this paper, we provide evidence that a similarly prepared lysate of
mitochondria in fact possesses the ability to repair various types of DNA
lesions such as those caused by depurination, oxidation or UV-irradiation.
Ovaries were removed from five mature females of
X.laevis
. They were rinsed in OR-2 medium (5 mM HEPES pH 7.6, 1 mM Na
2
HPO
4
, 83 mM NaCl, 2.5 mM KCl, 1 mM MgCl
2
and 1 mM CaCl
2
), and gently homogenized in the Dounce homogenizer using five strokes of the
loose-fitting pestle. The homogenization buffer, which was used at 9 ml per 1 g
of ovaries, contained 10 mM Tris-HCl pH 7.5, 1 mM ethylenediaminetetraacetic acid (EDTA), 0.1 mM
phenylmethylsulfonyl fluoride (PMSF) and 15% (w/w) sucrose. Mitochondria were
obtained from the homogenate as described by Brun
et al.
(
16
) with minor modifications. Thus the oocyte homogenate was cleared by
centrifugation, first at 1000
g
for 5 min and then at 2500
g
for 10 min. The cleared homogenate was centrifuged further at 10 000
g
for 10 min. The pellet containing mitochondria was suspended to a concentration
of 5 mg protein/ml in the homogenization buffer supplemented with 15 % (w/w)
sucrose. The suspension was subjected to centrifugation (21 000
g
, 45 min) through a two-step sucrose gradient [25 and 42.5% (w/w) sucrose in the homogenization
buffer]. Material recovered from the boundary of the two sucrose solutions was pelleted at 12 000
g
for 20 min, and resuspended to a concentration of 10 mg protein/ml in the
homogenization buffer supplemented with 7.5% (w/w) sucrose and 3 mg/ml
digitonin. It was incubated for 15 min at 4oC to remove the mitochodrial outer membrane, and centrifuged at 12 000
g
for 20 min. The pellet was resuspended to a concentration of 5 mg protein/ml in
a lysis buffer (20 mM Tris-HCl pH 7.5, 0.5 M KCl, 1 mM dithiothreitol, 0.5% Triton X-100 and 0.1 mM PMSF). It was incubated for 30 min at 4oC to break the inner membrane of mitochondria, and centrifuged
at 250 000
g
for 3 h. The final supernatant was dialysed against a buffer containing 20 mM
potassium phosphate pH 7.5, 14% (w/w) glycerol and 2 mM [beta]-mercaptoethanol, and concentrated by filtration to 10 mg protein/ml.
It was aliquoted and stored at -85oC. The lysate thus prepared was stable for at least 6 months.
An oocyte extract, which contains nuclear and cytosolic material but is depleted
of mitochondrial proteins, was prepared as follows. The details of the
procedure will be published elsewhere (Ryoji
et al.
, manuscript in preparation). Briefly, ovaries were treated with 300 [mu]g/ml of collagenase for 16 h at 20oC to separate oocytes from follicle cells. Oocytes were washed,
transferred to the extraction buffer (25 mM HEPES pH 7.5, 50 mM KCl, 2.5 mM
MgCl
2
, 1 mM dithiothreitol and 250 mM sucrose). They were packed in a centrifuge
tube, and squashed by centrifugation at 15 000
g
for 15 min. The resulting supernatant (38 mg protein/ml) was aliquoted and
stored at -85oC.
Partially depurinated DNA was obtained by treatment of supercoiled plasmid pUC9
with 2-10 mM HCl for 30 min at 20oC. The solution was neutralized with 1 M Tris-HCl pH 7.6, and DNA was recovered by ethanol precipitation.
Nicked, open circular DNA was prepared in a 110 [mu]l mixture containing 10 [mu]g of supercoiled pUC9 DNA, indicated quantities of DNase I, 50 mM Tris-HCl pH 7.6, 85 mM KCl, 3 mM MgCl
2
, 50 [mu]M CaCl
2
, 5.5 [mu]g bovine serum albumin and 5% (w/w) glycerol. The mixture was incubated for
2 min at 24oC. DNA was purified by phenol-chloroform extraction. Oxygen free radical-induced lesions were generated in supercoiled pUC9 DNA by
incubation with 3.2 mM H
2
O
2
and 50 [mu]M CuSO
4
for 2 min at 37oC as described (
18
). UV-induced damage was formed by irradiating supercoiled pUC9 DNA (94 ng/[mu]l) under the short-wavelength UV lamp to a fluence of 900 J/m
2
.
The reaction mixture (11.7 [mu]l) for excision repair consisted of 10 ng pUC9 DNA, 40 [mu]g mitochondrial lysate proteins, 2-10 [mu]Ci [[alpha]-
32
P]dCTP, 10 [mu]M dNTP but dCTP, 50 mM HEPES pH 7.5, 5 mM MgCl
2
, 3 mM ATP, 30 mM creatin phosphate, 0.12 [mu]g creatin kinase and 1.5 mM [beta]-mercaptoethanol. Unless stated otherwise, it was incubated for 2
h at 21oC. The reaction was terminated by addition of 59 [mu]l of a proteinase K solution (20 [mu]g proteinase K, 30 mM Tris-HCl pH 7.4, 20 mM EDTA, 1% sodium dodecyl sulfate). Plasmid
DNA was phenol-extracted from the mixture, and separated by 1% agaraose gel electrophoresis in 0.5* TBE buffer (45 mM Tris, 45 mM boric acid, 1.3 mM EDTA).
To quantify the incorporation of [[alpha]-
32
P]dCTP, the gel was dried and subjected to autoradiography employing the
Molecular Imager GS250 (Bio Rad). The image intensity was proportional to the
signal over the range described, and the data were processed without any
enhancement of the image contrast. DNA in each band in the gel was quantified
by Southern analysis using
32
P-labeled pUC9 DNA as the hybridization probe.
Conditions for the photoreactivation reaction were basically the same as above
except that [[alpha]-
32
P]dCTP was replaced with unlabeled dCTP. Where indicated, the reaction tubes
were illuminated from outside at ~2500 lux with a 30 W fluorescent lamp.
Plasmid pSO1 used for the PCR assay was a pUC9 derivative that carries a 792 bp
Xenopus
cDNA inserted into the
Eco
RI site of the vector. The primers were both 17mers, i.e., 5'-CCCAGTCACGACGTTGT-3' and 5'-CAGGAAACAGCTATGAC-3' that are complementary to
the vector regions on either side of the multiple cloning sites. PCRs were
carried out for 25 cycles using
Taq
DNA polymerase according to the standard protocol (
17
). The size of the amplified fragment was expected to be 889 bp.
We first studied whether DNA synthesis for excision repair can be seen in the
lysate of mitochondria. Figure
1
A shows a time course of [[alpha]-
32
P]dCTP incorporation into depurinated pUC9 DNA (lanes 7-9). Undamaged DNA was used as a control (lanes 1-3). Although DNA synthesis continued for at least 4 h in both
cases, depurinated DNA displayed a much higher incorporation. A Southern
analysis performed in parallel confirmed that the template DNA was present in
about equal quantities in all the cases (Fig.
1
B). Note in this figure that the radioactivity contributed by the DNA synthesis
was negligibly small in comparison with the hybridization signals of the
Southern analysis. Therefore, the densities of DNA bands can be taken directly
as the quantities of the template DNA.
As seen in Figure
1
A, DNA molecules which incorporated [[alpha]-
32
P]dCTP were mostly open circles with only a minor population of supercoiled
ones. It should be noted that the open circular molecules include both nicked
circles (form II) and covalently closed, relaxed ones (form Ir). Therefore it
is not clear whether the
excision repair reaction described here proceeded efficiently to the final
ligation step. To know the extent of the repair reaction in this system, we
electrophoresed the radioactive DNA molecules in a buffer containing 15 [mu]g/ml of chloroquine phosphate. It is known that, in the presence of this
concentration of chloroquine phosphate, covalently closed but relaxed circles
(form Ir) migrate ahead of the other types of molecules, whereas nicked circles
(form II) move as the slowest species (
19
: see arrows in Fig.
2
). It is evident that the majority of the radioactive DNA molecules carried
nicks or gaps, and thus the repair reaction was not completed in most DNA
molecules. Nevertheless, once the repair patches were sealed, DNA molecules
became supercoiled rapidly as evidenced by the faint bands corresponding to the
fully supercoiled form, i.e
.
, form I.
Eukaryotic cells contain at least five species of DNA polymerase, i.e., pol [alpha], [beta], [gamma], [delta] and [epsilon] (see ref.
20
for a review). Among them, pol [gamma] is localized exclusively within the mitochondrion, and participates in
replication of the mitochondrial genome (
16
). The other polymerases are present largely in the nucleus, and are responsible
for replication as well as repair of the chromosome. To establish that DNA
synthesis described above was carried out by the mitochondrial polymerase but
not by nuclear or cytosolic contaminants, we looked at the sensitivity of the
DNA synthesis to aphidicolin. This inhibitor blocks pol [alpha], [delta] and [epsilon] but does not affect pol [beta] and [gamma] activities (
21
-
24
). For comparison, the excision repair reaction was also performed in a crude
extract of oocytes depleted of the mitochondrial fraction. It was found that
DNA synthesis in the mitochondrial lysate was refractory to aphidicolin up to a
concentration of at least 100 [mu]g/ml (Fig.
3
, upper panel). In contrast, DNA synthesis in the crude oocyte extract was
inhibited by 50% at 5 [mu]g/ml. Thus aphidicolin-resistant pol [beta] or [gamma] is responsible for the repair DNA synthesis in the
lysate of mitochondria.
Plasmid pUC9 DNA exposed to H
2
O
2
was incubated in the lysate of mitochondria. The oxidative damage presumably
includes strand breaks and altered bases like 8-hydroxyguanine and thymine glycol (
25
). Since we noticed that the open circular form of DNA increased in quantity
from 12 to 36% upon oxidation of DNA, we prepared a control DNA sample which was intentionally nicked with DNase I to the extent where a similar percentage of the open
circular DNA was present. This DNA was compared with the oxidized one in regard
to their abilities to induce DNA synthesis. As shown in Figure
4
, DNase I-induced nicks alone did not significantly stimulate DNA synthesis, whereas
the template carrying oxidative damage caused a 4-fold increase in incorporation. This indicates that oxidative damage can
also be repaired through an excision repair pathway in
Xenopus
mitochondria.
The
damage-dependent DNA synthesis described above does not necessarily infer that
the lesion is actually repaired. We wanted to see an actual reduction of the
lesions upon incubation in the lysate. It was shown that the movement of
Taq
DNA polymerase on the template is arrested when it encounters damaged bases.
Thus several investigators reported that the PCR product increases in quantity
if the damaged template is subjected to a repair reaction prior to the PCR
(see, e.g., ref.
26
). We carried out such an assay to examine whether depurinated or UV-irradiated DNA is actually restored from the damage when incubated with
the mitochondrial lysate. To determine how much of the template should be used
for the PCR analysis, various amounts of undamaged plasmid DNA were subjected
to PCR. The 889 bp PCR product increased linearly if no more than 1 pg of the
template was used (Fig.
6
, upper panel). Thus we took an aliquot containing 1 pg of DNA from the repair
reaction mixture, and then used it for PCR (Fig.
6
, lower panel). When the template that had been depurinated with 3 mM HCl was
subjected directly to PCR, the amount of the amplified fragment was only 20-30% of the control. On the other hand, prior incubation of the same
template in the lysate resulted in an ~70% increase in the amount of the 889 bp fragment. Therefore some of the
abasic sites are repaired in the mitochondrial lysate. If the template was
exposed to 10 mM HCl, the repair was not effective enough to be detected by
this method.
The damage-dependent DNA synthesis we observed is an indication of excision repair
activities of
Xenopus
oocyte mitochondria. We tested three types of DNA damaging agents, i.e., HCl, H
2
O
2
and UV-light, and all of them generated lesions that caused considerable
increases of DNA synthesis. In the case of HCl-mediated, depurinated lesions, the PCR analysis revealed that the
restoration of DNA bases was only partial when the damaged DNA was incubated in
the lysate for 4 h. This result is due in part to the fact that the excision
repair reaction did not proceed efficiently to the completion of gap filling
and subsequent ligation of the nick (Fig.
2
). Although the damage itself might be excised, the remaining gap, if not sealed
completely, would impede the movement of
Taq
DNA polymerase. It should be mentioned that the same PCR analysis cannot be
applied to the oxidative damage, since lesions like 8-hydroxyguanine allow improper base pairings, and do not effectively block
Taq
DNA polymerase (
2
7
). These data provide the first direct evidence of the excision repair activity
of mitochondria. To further characterize the nature of such reactions in
detail, it would be necessary to introduce more defined damage at specific
sites in the template DNA.
According to the present studies, DNA synthesis in response to UV-damage was relatively low compared with that observed in the crude extract
of oocytes. This is consistent with the fact that the PCR analysis did not
reveal any restoration of the UV-induced damage if the reaction was performed in the dark. However, the
damage was repaired by photoreactivation, though the complete repair required
an extended incubation time (6 h) and a larger quantity of the lysate (80 [mu]g protein). Incubation under the standard conditions (4 h, 40 [mu]g protein) resulted in only ~15% of restoration (data not shown). It remains to be seen whether
the mitochondrial photolyase described here differs from that observed in the
nucleus.
We hope that further studies in this system will uncover a number of features of
the mitochondrial repair machinery distinct from that of the nucleus.
REFERENCES
Return