ABSTRACT
In mammalian cells, mRNA transcription is initiated with the aid of
transcription initiation factors. Of these, TFIIH has also been shown to play
an essential role in nucleotide excision repair (NER), which is a versatile
biochemical pathway that corrects a broad range of DNA damage. Since the dual
role of TFIIH is conserved among eukaryotes, including yeast and mammalian
cells, the sharing of TFIIH between NER and RNA transcription initiation might
provide some survival advantage. However, the functional relationship between
NER and RNA transcription initiation through TFIIH is not yet understood. We
have developed an optimized cell-free assay which allows us to analyze NER and RNA transcription under identical conditions. In this assay, NER
did not compete with RNA transcription, probably because the extracts contained sufficient amounts of TFIIH to support
both processes. Thus, NER can be considered functionally independent of RNA
transcription initiation despite the fact that both processes use the same
factor.
The basal transcription factor TFIIH has a dual role in eukaryotes, being
involved in both initiation of mRNA synthesis by RNA polymerase II (Pol II) and
in nucleotide excision repair (NER) (
1
-
4
). This factor from human cells contains multiple subunits, including XPB (
2
), XPD (
5
) and p62 (
6
). XPB and XPD are 3' -> 5' (
7
) and 5' -> 3' (
8
) helicases respectively. In studies with yeast mutants of TFIIH, the 3' -> 5' helicase has been shown to be essential for both NER and
initiation of mRNA synthesis (
9
), while the 5' -> 3' helicase is required only for NER (
10
,
11
).
Along with TFIIH, initiation of mRNA synthesis requires the basal transcription
factors TFIIB, TFIID, TFIIE, TFIIF and TFIIJ. These factors form a
preinitiation complex at the TATA box and Pol II is loaded onto DNA during the
process of complex formation (
12
). To facilitate RNA synthesis, the template DNA is probably melted by the TFIIH-associated helicase activities, resulting in activation of Pol II (
2
). In living organisms, a certain amount of these factors must be produced to
maintain competence in transcription, since mRNA synthesis is essential for growth and viability of
cells (
9
,
13
).
NER is a multienzymatic process that, in human cells, is initiated by binding of
the XPA-RPA complex to DNA damage (
14
-
17
). Then, an oligonucleotide of 27-30 bases containing the damage is excised from the DNA (
18
) by DNA nucleases, the XPF-ERCC1 complex (
19
) and the XPG protein (
20
). The helicase activities of TFIIH are also required for excision (
1
). The gapped duplex resulting from excision is filled by repair synthesis,
utilizing RPA (
21
), PCNA (
22
), DNA polymerases (
23
) and DNA ligase (
24
), which are also needed for DNA replication (
25
). This repair pathway removes a broad spectrum of DNA damage, including UV-induced cyclobutane pyrimidine dimers and 6-4 pyrimidine-pyrimidone photoproducts (
26
). In humans, NER deficiency results in the hereditary disorder xeroderma
pigmentosum (XP), characterized by hypersensitivity to sunlight exposure and
greatly increased skin tumor formation (
27
).
The dual functions of TFIIH are conserved among eukaryotes, including yeast (
4
) and mammals (
1
,
3
). Thus, eukaryotic cells may acquire a certain survival advantage by using
TFIIH in both mRNA transcription initiation and NER. However, the functional
relationship between NER and RNA transcription initiation through TFIIH is not
yet understood.
Therefore, we have developed an optimized cell-free DNA repair-RNA transcription system which allowed us to analyze NER and RNA
transcription under identical assay conditions and we have investigated the
relationship between NER and transcription initiation through TFIIH. Since NER
did not compete with RNA transcription in our assay system, we conclude that
NER is functionally independent of RNA transcription initiation.
RNase T1 was purchased from Gibco BRL. [[alpha]-
32
P]NTPs and [[alpha]-
32
P]dNTPs were obtained from Amersham.
Escherichia coli
endonuclease III and mouse antibody against RPA 70 kDa subunit (70C;
21
) were kindly provided by Dr R.D.Wood. Purified TFIIH from HeLa cells (
2
,
28
), mouse monoclonal antibody against the p62 subunit of TFIIH (3C9;
29
) and antibody to TFIIE (2A1) were kindly provided by Dr J.-M.Egly.
Lymphoblastoid cells, GM01953C (normal) and GM01857 (Cockayne's syndrome group
A; CSA), were obtained from the NIGMS Human Mutant Cell Repository, NJ, and
cultured in RPMI 1640 supplemented with 20% fetal bovine serum and antibiotics.
HeLa S3 cells were maintained in suspension culture with MEM supplemented with
5% fetal bovine serum and antibiotics.
Actively growing cells were collected by centrifugation and dead cells, which
were often found even in optimized culture conditions, were removed with Ficoll-Paque (Pharmacia) as described (
30
). The cell-free extracts were prepared by the method of Manley
et al
. (
31
). In some experiments, TFIIH-depleted extracts were employed. To prepare the extracts, antibody against
the p62 subunit of TFIIH (3C9) was incubated with extract for 60 min at 0oC and the antibody-TFIIH complex was removed with protein G-Sepharose (GammaBind Plus Sepharose, Pharmacia Biothec).
About 95% of the TFIIH was removed from the extract (designated 95% TFIIH-depleted extract) by this procedure, as determined by Western blotting,
using antibody against the p62 subunit for the analysis.
pGf1 (Fig.
1
A) contained the adenovirus major late promoter (AdMLP) and a G-less cassette (
32
), which does not have any guanine residues in the non-transcribed strand. It was constructed from pAdHap (3.1 kb) (
33
). To replace an
Ssp
I site in pAdHap with a
Pst
I site (pAdHap-s), the plasmid was digested with
Ssp
I, ligated to a
Pst
I linker and the linker digested with
Pst
I. The plasmid was purified by 1% agarose gel electrophoresis containing 1 [mu]g/ml ethidium bromide (EtBr) and transfected into
E.coli
after ligation of the
Pst
I site. A
Mae
I site (labeled M in Fig.
1
A) was then replaced with a
Pac
I site (pAdHap-ps). Since there were three other
Mae
I sites (labeled m in Fig
1
A), pAdHap-s was digested with
Pst
I and
Bam
HI and the large fragment (2.2 kb) and the small fragment (0.8 kb) containing
the
Mae
I site (labeled M in Fig.
1
A) were purified by 1% agarose gel electrophoresis containing EtBr. The small
fragment was then digested with
Mae
I and the
Pst
I-
Mae
I fragment (0.6 kb) was purified by 2% agarose gel electrophoresis containing
EtBr. This
Pst
I-
Mae
I fragment was ligated with the large fragment (2.2 kb) at the
Pst
I site and the resulting
Bam
HI-
Mae
I fragments were purified by 1% agarose gel electrophoresis containing EtBr.
This fragment was then treated with mung bean nuclease, ligated with a
Pac
I linker, digested with
Pac
I, purified by 1% agarose gel electrophoresis containing EtBr and transfected
into
E.coli
after ligation of the
Pac
I site. The resulting pAdHap-ps plasmid was digested with
Pac
I and
Hin
dIII and the 2.7 kbp fragment was purified for ligation with the G-less cassette.
The typical reactions were carried out with cell-free extracts, 0.75 [mu]g pGf1 containing 430 fmol AdMLP and 0.25 [mu]g UV-irradiated pBluescript II KS
+
(pBS, 3.0 kb; Stratagene), which was prepared by irradiation with UV (450 J/m
2
at 254 nm), followed by treatment with
E.coli
endonuclease III as described by Wood
et al
. (
34
). To make the amount of pBS commensurate with that of pGf1, 0.5 [mu]g non-damaged pBS was also added to the reaction. Based upon the report of
Jones
et al
. (
35
), the level of cyclobutane pyrimidine dimers and 6-4 pyrimidine-pyrimidone photoproducts in 0.25 [mu]g UV-irradiated pBS was estimated to be 1300 and 430 fmol
respectively. It has been shown that the 6-4 pyrimidine-pyrimidone photoproduct is ~10 times better as a substrate than the cyclobutane pyrimidine dimer for NER in cell-free assays (
36
). In some reactions, pGf1 was replaced with p[Delta]Gf1 and UV-irradiated pBS
was replaced with non-damaged pBS.
The reactions were carried out in one of four different reaction mixtures.
Repair assay.
The reaction mixture was that used in the cell-free DNA repair assay reported by Wood
et al
. (
34
) and contained 20 mM HEPES-KOH, pH 7.9, 2 mM ATP, 8 [mu]M dATP, 25 [mu]M each of dCTP, dTTP and dGTP, 40 mM phosphocreatine, 13 U/ml
creatine phosphokinase (CPK), 70 mM KCl, 5 mM MgCl
2
, 3.4% glycerol, 300 [mu]M dithiothreitol, 300 [mu]g/ml bovine serum albumin (BSA) and 0.75 [mu]Ci [[alpha]-
32
P]dATP (1 [mu]Ci/150 pmol).
Transcription assay.
This was the slightly modified reaction mixture used in cell-free RNA transcription assays reported by Dignam
et al
. (
37
) and contained 20 mM HEPES-KOH, pH 7.9, 600 [mu]M each of ATP, UTP and GTP, 25 [mu]M CTP, 40 mM phosphocreatine, 13 U/ml CPK, 0.3 U RNase T1, 60 mM
KCl, 10 mM MgCl
2
, 3.4% glycerol, 300 [mu]M dithiothreitol, 300 [mu]g/ml BSA and 2.1 [mu]Ci [[alpha]-
32
P]CTP (1 [mu]Ci/300 pmol).
Repair-transcription assay (unoptimized).
The reaction mixture for a DNA repair-RNA transcription assay was the same as for the transcription assay, but
the assay also contained dNTPs at the same concentrations as in the repair
assay and the reactions were carried out with either 0.75 [mu]Ci [[alpha]-
32
P]dATP (1 [mu]Ci/150 pmol) for repair or 2.1 [mu]Ci [[alpha]-
32
P]CTP (1 [mu]Ci/300 pmol) for RNA transcription.
Repair-transcription assay (optimized).
The reaction mixture for an optimized DNA repair-RNA transcription assay contained 20 mM HEPES-KOH, pH 7.9, 1.2 mM ATP, 360 [mu]M GTP, 180 [mu]M UTP, 270 [mu]M CTP, 600 [mu]M dATP, 180 [mu]M dGTP, 120 [mu]M dTTP, 480 [mu]M dCTP, 10 mM phosphocreatine, 40 U/ml
CPK, 0.3 U RNase T1, 60 mM KCl, 10 mM MgCl
2
, 3.4% glycerol, 300 [mu]M dithiothreitol, 300 [mu]g/ml BSA and either 30 [mu]Ci [[alpha]-
32
P]dGTP (1 [mu]Ci/150 pmol) for repair or 30 [mu]Ci [[alpha]-
32
P]UTP (1 [mu]Ci/150 pmol) for transcription. In the repair-transcription assay (optimized), the specific activity of [[alpha]-
32
P]UTP was twice that of [[alpha]-
32
P]CTP in the transcription assay and repair-transcription assay (unoptimized), since the number of thymidine residues
in the transcripts from the G-less cassette was half the number of cytosine residues.
These cell-free reactions were carried out in 25 [mu]l volumes at 30oC for 60 min and were terminated by addition of 25 [mu]l 1.2% SDS, 40 mM EDTA and 2.4 mg/ml proteinase K followed by
incubation for 30 min at 37oC. Plasmid DNA and RNase T1-resistant transcripts were purified by phenol/chloroform extraction
and precipitated with ethanol with ammonium acetate and tRNA. For analysis of
DNA repair, plasmid DNA was incubated with 50 U
Hin
dIII, 50 U
Pst
I and 75 [mu]g RNase A in a 60 [mu]l reaction mixture for 30 min at 37oC and 20 [mu]l was fractionated by electrophoresis on a 1% agarose gel
containing 1 [mu]g/ml EtBr. A photographic negative of the gel was taken to analyze DNA
recovery and phosphorimaging (GS-363 Molecular Image System; BioRad) and autoradiography of the dried
agarose gel were used to determine repair activity. For RNA transcription,
ethanol-precipitated RNase T1-resistant transcripts were dissolved in gel loading buffer
containing 7 M urea, 10 mM EDTA, 5% glycerol, 0.05% bromophenol blue and 0.05%
xylenene cyanol, denatured at 65oC for 10 min and fractionated on a 10% acrylamide-4 M urea gel. Phosphorimaging and autoradiography of the dried urea
gel were used to quantify the RNA.
The cell-free reactions were carried out with pGf1, UV-irradiated pBS and non-damaged pBS as described above in the presence of either 0.5 [mu]Ci [[alpha]-
32
P]dATP, [[alpha]-
32
P]dGTP, [[alpha]-
32
P]dCTP, [[alpha]-
32
P]dTTP, [[alpha]-
32
P]ATP, [[alpha]-
32
P]GTP, [[alpha]-
32
P]CTP or [[alpha]-
32
P]UTP. After incubation for 60 min at 30oC, the reactions were terminated by addition of 25 [mu]l loading buffer. Samples were fractionated by electrophoresis on 10%
polyacrylamide-4 M urea gels. Phosphorimaging and autoradiography of dried gels were
used to visualize the radioactivity.
Cell-free extracts (25 [mu]g) were denatured in gel loading buffer for 5 min at 100oC, fractionated on an SDS-12.5% polyacrylamide gel and transferred to a
nitrocellulose membrane. The membrane was then incubated with mouse antibody
against the p62 subunit of TFIIH in buffer containing 0.1% Tween-100, 150 mM NaCl and 10 mM Tris-HCl, pH 7.5, (TBST) with 5% skimmed milk overnight at 4oC. After washing the membrane four times for 5 min with TBST,
it was incubated with anti-mouse secondary antibody conjugated with horse radish peroxidase (HPR-Goat anti-mouse IgG; Zymed Laboratories Inc., San Francisco, CA) in TBST
containing 5% skimmed milk for 1 h at 25oC. The membrane was then washed four times for 5 min with TBST and
incubated with chemiluminescence reagents (Renaissance; Du Pont). Protein was
visualized by exposure of the membrane to X-ray film. Over six different exposure times were taken and the films were
scanned (Scan Jet IIP, Hewlett Packard) for quantitation.
To establish a DNA repair-RNA transcription assay which allows analysis of NER and transcription
through TFIIH under identical assay conditions, the cell-free assay for NER and RNA transcription was carried out with two
different kinds of closed circular plasmid DNA (Fig.
1
B): pGf1 for RNA synthesis by Pol II and UV-irradiated pBS for NER.
pGf1 contained AdMLP and a G-less cassette (
32
), which is a stretch of DNA without guanine residues in the non-transcribed strand. Thus the transcripts did not contain guanine residues
and they were resistant to RNase T1 in the reaction mixture. After
fractionation by urea gel electrophoresis, the RNA transcription activity of
Pol II was determined by quantitation of
32
P in the transcripts.
UV-irradiated pBS contained pyrimidine dimers and 6-4 pyrimidine-pyrimidone photoproducts. After digestion of plasmid DNA with
Pst
I and
Hin
dIII and separation of UV-irradiated pBS from fragments of pGf1 by EtBr-agarose gel electrophoresis, repair activity was determined by
comparing the amount of
32
P in UV-irradiated pBS
with that in fragments of pGf1.
The originally reported DNA repair assay found the optimal KCl concentration to
be between 40 and 100 mM and was carried out at pH 7.9 with 2 mM ATP, 8 [mu]M dATP, 25 [mu]M each of dCTP, dTTP, dGTP and an ATP-regenerating system with 13 U/ml CPK (
34
). For transcription, the optimal KCl concentration and pH reported by Dignam
et al
. (
37
) are 60 mM and between 7.5 and 8.5 respectively and the reaction contained 600 [mu]M each of ATP, UTP and CTP and 25 [mu]M GTP. To establish a DNA repair-RNA transcription assay, we have tested an assay with 60 mM KCl,
HEPES-KOH, pH 7.9, 8 [mu]M dATP, 25 [mu]M each of dCTP, dTTP and dGTP, 2 mM ATP, 600 [mu]M each of UTP and GTP, 25 [mu]M CTP and an ATP-regenerating system with 13 U/ml CPK (see Materials
and Methods for detailed reaction conditions).
However, in this assay [designated repair-transcription assay (unoptimized)] NER activity was significantly reduced
compared with that in the assay carried out under the previously reported
conditions (designated repair assay; see Materials and Methods) (Fig.
2
A, lanes 1 and 3). Such a reduction may be due to increased degradation of dCTP.
As shown in Figure
3
(dCTP, lane 3), the recovered dCTP after incubation with 100 [mu]g cell-free extract for 60 min incubation in this assay was significantly
reduced and a degradation product, which migrated more slowly than dCTP on the
urea gel, had appeared. However, reactions with a higher initial dCTP
concentration (480 [mu]M) did not exhibit this reduction of dCTP (Fig.
3
, dCTP, lane 4) and resulted in increased NER activity (Fig.
2
A, lane 4).
Consistent with the report by Coverley
et al
. (
23
), the addition of 20 [mu]g/ml aphidicolin suppresses >95% of damage-specific
32
P incorporation in the assay and an antibody against RPA (70C) reduced
incorporation by 50%. In addition, the 95% TFIIH-depleted extract exhibited only 40% of the control NER activity due to
limited availability of TFIIH for NER; the addition of 0.83 [mu]g (~1800 fmol) purified TFIIH restored NER activity to 70% of the non-depleted extract. These results confirmed that the
32
P incorporation detected in our assay was mainly due to NER.
For transcription, an antibody against TFIIE (2A1) or against TFIIH (3C9)
reduced transcription activity by 70 and 40% respectively and negligible
amounts of RNase T1-resistant transcripts were generated from p[Delta]Gf1, which lacked AdMLP but contained the G-less cassette. Furthermore, transcription activity was
reduced to 30% when the 95% TFIIH-depleted extract was employed. However, addition of 1800 fmol purified
TFIIH, restored the activity to 50% of the non-depleted extract, although addition of >1800 fmol TFIIH did not further
increase RNA transcription activity, possibly due to removal of other transcription factors,
which interact with TFIIH, during depletion. Taken together, these results suggest that our assay measured the occurrence of
TFIIH-dependent transcription initiation from AdMLP by Pol II.
NER as well as RNA transcription activities increased linearly as a function of
amount of extract used, at least up to 120 [mu]g protein (4.8 mg protein/ml; Fig.
5
A and B), and both reactions continued up to 4 h (data not shown). NER and
transcription activities increased linearly, at least up to 0.75 [mu]g UV-irradiated DNA and 0.75 [mu]g pGf1 with 100 [mu]g HeLa extract, for 60 min incubation (data not shown).
Using the repair-transcription assay (optimized), relationships between NER and RNA
transcription were analyzed with 100 [mu]g extracts for 60 min incubation. To investigate the effect of RNA
transcription on NER, DNA repair activity was compared between the assay with
pGf1 and p[Delta]Gf1. As shown in Figure
6
A, RNA transcription did not have a major effect on NER. Similarly, the effect
of NER on RNA transcription was analyzed by performing reactions with pGf1 and
either UV-irradiated pBS or non-damaged pBS. RNA transcription was not affected by NER, as shown in
Figure
6
B, although a reproducible but slight suppression of RNA transcription activity
was found when GM01953C extracts were employed. These results suggest that NER
is functionally independent of RNA transcription. Such independence may be a
consequence of an excess of TFIIH activity over that required for NER and RNA
transcription. In fact, in 100 [mu]g HeLa and GM01953C extracts, ~2300 and 2400 fmol TFIIH respectively were found, as determined by
Western blotting (data not shown). On the other hand, with 100 [mu]g HeLa and GM01953C extracts 60 min incubation removed only 134 and 34 fmol
UV-induced DNA lesions [calculated from the amount of incorporated dGMP into
repair patches by assuming that the repair patch size is 29 nt (
18
) and that the patch contained an equal number of each dNMP] and produced 0.32
and 0.6 fmol RNase T1-resistant transcripts (calculated from the amount of UMP incorporation
into RNase T1-resistant transcripts by dividing by 30, which is the number of uracil
residues in the transcript).
In this report we have described a combined DNA repair-RNA transcription assay, which has required optimization of the relative
concentrations of dNTP and NTP as well as CPK activity. During the cell-free reaction, nucleoside triphosphates were turned over (data not shown)
and probably converted to nucleoside diphosphates. To regenerate dATP and ATP
from corresponding nucleoside diphosphates, an ATP-regenerating system was added to the cell-free reactions. As described in this report, the assay with
lymphoblastoid cell extract requires at least 40 U/ml CPK. Other nucleoside diphosphates may be regenerated by an endogenous
enzyme, presumably nucleoside diphosphokinase (
38
), in cell-free extracts. This enzyme has a high
K
m
for dCDP and CDP relative to that for the other nucleoside diphosphates (
38
). Thus, dCDP and CDP tended to accumulate during the cell-free assay if the relative concentrations of dNTPs and NTPs are not
optimized. Thus, optimization of the cell-free DNA repair-RNA transcription assay required adequate and balanced amounts of
nucleoside triphosphates. We have determined the optimal concentrations for
dNTPs and NTPs, based mainly upon the
K
m
of nucleoside diphosphokinase for nucleoside diphosphates (
38
).
In the repair-transcription assay (optimized), NER and RNA transcription activities
were increased about five to six times when compared with the activity obtained
in the repair assay and transcription assay. In this optimized assay, no
apparent competition for TFIIH between NER and RNA transcription was found.
Furthermore, addition of purified TFIIH to the assay did not significantly
increase NER or RNA transcription activities. These results indicate that the
cell-free extracts contain a sufficient amount of TFIIH to support both NER and
RNA transcription. Thus, it can be considered that NER is functionally
independent of RNA transcription initiation despite the fact that both
processes are using the same factor. However, the sharing may provide certain
survival advantages, since the dual role of TFIIH is conserved among
eukaryotes, including yeast (
4
) and mammalian cells (
1
,
3
).
In
E.coli
, the NER enzymes UvrA and UvrB are induced to higher levels following the
introduction of DNA damage, leading to modestly enhanced repair capacity (
39
), while mammalian cells seem to maintain a high NER capacity constitutively. In
fact, we have confirmed significant amounts of NER activity in the extracts. As
demonstrated by Donahue
et al
. (
33
), DNA damage on the transcribed strand of an actively transcribed gene can
cause termination of RNA synthesis by stalling of Pol II. Thus, a single DNA
lesion can abolish mRNA synthesis and this could be lethal if the damage in an
essential gene is not repaired. The risk of termination of transcription may be
reduced by maintaining high NER activity. On the other hand, there are
sometimes very low levels of unscheduled DNA synthesis in cultured cells (see
for example
27
), so most of the time, NER may be considered to be an infrequent event relative
to RNA transcription initiation. By using TFIIH, which is a factor with a
housekeeping function, in NER, cells would be able to maintain high NER
capacity regardless of the immediate requirement for NER activity and thus
facilitate an efficient response to elevated levels of DNA damage. In this
regard, the possible presence of two distinct forms of TFIIH, such as for
transcription initiation and for NER, has been suggested (
40
). However,
in vitro
reconstitution experiments with purified factors clearly demonstrate that NER
and RNA transcription initiation require only one form of TFIIH, since TFIIH
which is active for RNA transcription initiation (
29
) is also active for NER (
24
). NER also shares RPA, PCNA, DNA polymerases and DNA ligase with the process of
DNA replication. Such sharing of factors between NER and replication may have a
similar physiological relevance to that of TFIIH.
We thank Dr R.D.Wood for providing the
E.coli
endonuclease III and RPA antibody. We also thank Dr J.-M.Egly for providing various antibodies and purified TFIIH and for his
thoughtful comments on the manuscript. We also thank Dr A.Ganesan for
discussion and critical reading of the manuscript. This work was supported by
an Outstanding Investigator Grant from The National Cancer Institute (CA44349)
and a Program Project Grant from The National Institute on Aging (AG2908).
REFERENCES
Return