ABSTRACT
Recent studies have shown that many proteins are involved in the early steps of
nucleotide excision repair and that there are some interactions between
nucleotide excision repair proteins, suggesting that these interactions are
important in the reaction mechanism. The xeroderma pigmentosum group A protein
(XPA) was shown to bind to the replication protein A (RPA) or the excision
repair cross complementing rodent repair deficiency group 1 protein (ERCC1),
and these interactions might be involved in the damage-recognition and/or incision steps of nucleotide excision repair. Here we
show that the XPA regions required for the binding to the 70 and 34 kDa
subunits of RPA are located in the middle and on N-terminal regions of XPA, respectively. These regions do not overlap with
the ERCC1-binding region of XPA, and a ternary protein complex of RPA, XPA and ERCC1
was detected
in vitro
. In addition, using the surface plasmon resonance biosensor, the binding of RPA
and ERCC1 to XPA was investigated. The dissociation constants (
K
D) of RPA and ERCC1 with XPA were 1.9*10-8 and 2.5*10-7 M, respectively. Moreover, our results suggest the sequential binding of RPA and ERCC1 to XPA.
Nucleotide excision repair (NER)
is the major pathway by which a broad spectrum of DNA damage, including
pyrimidine dimers induced by UV, are removed from the genome. The importance of
NER has been indicated by studies on an inherited human disease called
xeroderma pigmentosum (XP), which is characterized by the hypersensitivity of
skin to sunlight, a high incidence of sunlight-induced skin cancer, and neurological complications (
1
,
2
). Cells from XP patients are hypersensitive to killing by UV irradiation
because of a defect in NER. There are seven different genetic complementation
groups (A-G) and a variant form in XP. In addition, there are two genetic
complementation groups in Cockayne syndrome (CS-A and B), which is a NER-defective hereditary disorder with clinical symptoms of
neurodevelopmental delay and dwarfism, but with no increased incidence of skin
cancer, and 11 complementation groups (groups 1-11) in UV-sensitive, NER-defective, rodent mutant cell lines (
2
,
3
). In XP and some of these mutant cells, there are defects in the early stages
of NER, including the recognition of damaged DNA, generation of dual incisions
on either side of the damaged region, and excision of the oligonucleotides
containing the damaged site (
4
). In CS cells NER of total genomic DNA is normal but the increased rate of
repair of the transcribed strand of active genes (transcription-coupled repair) usually seen in repair-proficient cells is absent (
1
,
2
). So far,
XPA
,
XPB
(
ERCC3
),
XPC
,
XPD
(
ERCC2
),
XPF
(
ERCC4, ERCC11
),
XPG
(
ERCC5
),
CSA
(
ERCC8
),
CSB
(
ERCC6
) and
ERCC1
genes have been cloned (
5
-
15
). The
ERCC
genes are the human genes that can correct the repair deficiency in UV-sensitive rodent mutant cell lines.
In the early stages of NER, many proteins including the XP proteins are known to
be involved (
16
). The XPA protein (XPA) preferentially binds to UV- or chemical carcinogen-damaged DNA, suggesting that XPA is involved in the damage
recognition step of NER (
17
-
19
). XPB and XPD have helicase activity (
20
-
23
) and are components of the general transcription factor TFIIH (
23
-
25
). TFIIH was shown to participate in NER as well as in transcription. Possibly,
the bi-directional helicase activities of TFIIH serve to locally unwind DNA
around the damaged sites. XPG specifically nicks the DNA duplex at the 3' border of a single-stranded DNA region of a bubble structure (
26
) and cuts on the 3' side of the lesion (
26
,
27
). ERCC1 and XPF form a complex (
10
,
28
-
31
) which has nuclease activity (
31
) and cuts on the 5' side of the lesion (
10
,
27
,
28
). Using cell-free NER systems, RPA (replication protein A, also known as RFA and HSSB)
(
32
-
34
), a multisubunit single-stranded DNA binding protein, was shown to be involved in NER as well as
in DNA replication and recombination (
28
,
35
-
41
). XPC was purified as a tight complex with HHR23B (a human homologue of
Saccharomyces cerevisiae
Rad23) (
42
) which is required in the early stages of NER, although the function of the
complex is not clear yet. It was shown that a factor, designated IF7, is also
required for the
in vitro
reconstitution of the incision reaction (
38
).
Obviously protein-protein interactions might play important roles in NER as well as in DNA
replication, recombination and transcription. Actually, the interactions
between proteins related to NER have been detected in the yeast two-hybrid system and in
in vitro
binding experiments; XPA and ERCC1 (
43
-
46
), XPA and RPA (
47
-
49
), XPA and TFIIH (
50
) and XPG and RPA (
48
). These interactions may function to co-ordinate the early stages of NER. Here, we further analyze the
interactions between XPA, RPA and ERCC1. This is the first study to provide
quantitative measurements of interactions between NER proteins by biophysical
techniques. The results suggest that there is a sequential order of binding of
the proteins XPA, RPA and ERCC1 during assembly of the DNA repair complex.
Furthermore, we present a more precise mapping than previously available of the
sites of interaction between XPA and the subunits of RPA.
XPA with N-terminal T7.Tag (XPA) was expressed in
Escherichia coli.
Construction of the expression vector and purification of XPA has been
described elsewhere (
19
,
51
). To prepare GST-XPA fusion proteins, a series of
XPA
cDNA fragments (
52
) were ligated to a pGEX vector.
Escherichia coli
cells transformed with each vector were grown to
A
600
of 0.5, and fusion proteins were induced with 0.2 mM IPTG. After 8-10 h of additional growth at 25oC, the
E.coli
cells were pelleted, washed once with phosphate buffered saline (PBS), and
stored at -80oC. Cells were resuspended in 1/20-1/40 culture volume of a buffer containing 50 mM Tris-HCl, pH 7.8, 500 mM NaCl, 1 mM EDTA, 1% NP-40, 1 mM DTT and 0.5 mM PMSF and lysed by
sonication on ice. Lysates were centrifuged at 35 000
g
for 15 min, and the cleared lysates were incubated with glutathione-Sepharose (Pharmacia Biotech) for 30 min at 4oC. The beads were washed three times with NETN (50 mM Tris-HCl, pH 7.8, 150 mM NaCl, 1 mM EDTA, 1% NP-40, 1 mM DTT and 0.5 mM PMSF). Elution of the GST
fusion proteins was performed with a buffer containing 50 mM Tris-HCl, pH 8.0, 100 mM NaCl and 10 mM glutathione. The XPA and GST-XPA used
in the experiments had damaged DNA binding activity (
19
,
51
) and were active in an
in vitro
repair assay (data not shown). GST-ERCC1 was induced at 37oC with 1 mM IPTG for 4 h and purified as described above. Human RPA was
purified from HeLa cells (
47
), which was used for the BIAcore measurements. Recombinant human RPA was
purified as described (
53
), and used in the
in vitro
binding assay. Recombinant p34 (the 34 kDa subunit of RPA) was partially
purified from a fraction eluted with a buffer containing 0.5 M NaSCN from an
Affi-Gel blue column during the purification of recombinant RPA, as described above. RPA was active, as shown by an
in vitro
NER assay (
37
,
38
,
54
,
55
).
In vitro
transcription and translation was performed using a TNT coupled reticulocyte
lysate system (Promega) according to the manufacturer's instructions.
GST fusion proteins (500 ng-1 [mu]g) adsorbed onto glutathione-Sepharose (10 [mu]l) were mixed with purified proteins or
in vitro
translation products in 1 ml of NETN containing 10 mg/ml skim milk at 4oC for 2 h and then washed five times with NETN. The bound proteins were
extracted by boiling in SDS sample buffer and separated by SDS-PAGE and analyzed by immunoblotting using the ECL Western blotting
detection system (Amersham) or by the BAS 2000 image analyzer (FUJIX).
Real-time interactions of ERCC1 and RPA with XPA were monitored using the
surface plasmon resonance biosensor BIAcoretm system (Pharmacia Biotech). Measurements are based on changes in surface
plasmon resonance signals from chips which are covalently cross-linked with ligands (
56
). Changes in the signal were measured in real time, reflecting changes in the
refractive index due to increased protein concentrations on the surface of the
chip (
57
). XPA was diluted to a concentration of 1 mg/ml in HBS buffer (10 mM HEPES, pH
7.5, 150 mM NaCl, 3.4 mM EDTA and 0.05% Tween 20) and coupled to the
carboxymethylated dextran-modified gold surface of a CM5 sensor chip as described in the BIAcoretm system's manual. Briefly, the carboxyl groups of the chip were activated with 50 mM
N
-hydroxysuccinimide and 200 mM
N
-ethyl-
N
'-(3-diethylaminopropyl)-carbodiimide for 1 min, and then the XPA solution was injected onto the activated chip for 3 min at a flow
rate of 5 [mu]l/min. The remaining activated groups were blocked with 1 M ethanolamine (pH 8.5)
for 6 min, and any uncoupled protein was washed with HBS buffer. Under these conditions, surfaces containing densities of 9000 resonance units (RU; an arbitrary unit for the BIAcoretm system: ~9 ng/mm
2
) of XPA were generated. To collect sensorgrams, GST-ERCC1 and RPA were dialyzed against, and diluted with HBS buffer, and each concentration of the proteins was passed over the sensor surface at a flow rate of 10 [mu]l/min. Sensorgrams were recorded, normalized to a base line of 0 RU, and
analyzed using BIA Evaluation software. The equilibrium dissociation constants
(
K
D
) were calculated either from the kinetic parameters or from Scatchard analysis
of steady-state binding.
RPA is a heterotrimer consisting of 70, 34 and 11 kDa subunits (
32
-
34
) and we previously showed that XPA binds the 34 kDa subunit (p34) of RPA (
47
). We also examined whether the other subunits of RPA bind XPA. The GST-XPA
beads were mixed with either
in vitro
translated p70, p34 or p11 (Fig.
1
). p70 bound to XPA as well as p34, but p11 did not. GST-XPA used in this study
was active in an
in vitro
repair assay.
To determine the RPA-binding regions of XPA, a series of truncated XPA proteins fused with GST
were produced and used for binding to p34, p70 or RPA (Fig.
2
). In the case of p34, the truncated XPA, which had deletions in C-terminal regions and lacked the three N-terminal amino acids, was able to bind the
in vitro
translated p34, while the truncated XPA, which has >29 N-terminal amino acids deleted, was unable to bind. However, truncated XPA, which
has >86 C-terminal amino acids or 137 N-terminal amino acids deleted, was unable to bind the
in vitro
translated p70. These results indicate that amino acids 4-29 and 98-187 of XPA are required for binding to p34 and p70, respectively.
When recombinant RPA, which was active in an
in vitro
repair assay, was bound to XPA, it showed a combination of the p34 and p70
patterns. However, deletions of the p34-binding region in XPA (N[Delta]29, N[Delta]52 and N[Delta]97) had little effect on RPA binding, whereas
deletion of the p70-binding region (C[Delta]136 and N[Delta]137) caused a drastic decrease in RPA binding (Fig.
2
). These results indicate that p70 more predominantly contributed to the XPA-binding than p34. The binding of p34 to XPA seems to be rather weak and
the significance of p34 binding is unknown. However, it may costitute a binding
domain in conjunction with p70.
As the RPA-binding regions of XPA are different from the ERCC1-binding region (Fig.
6
, 44-46), it may be possible that XPA binds to both RPA and ERCC1
simultaneously. To verify the formation of a XPA-RPA-ERCC1 ternary complex
in vitro
, GST-ERCC1 beads were mixed with XPA (N-terminal T7.Tag XPA which was active in an
in vitro
repair assay) and RPA. RPA did not bind to GST-ERCC1 beads in the absence of XPA (Fig.
3
, lane 4), indicating that RPA does not bind to ERCC1 directly. But RPA bound to GST-ERCC1 beads in the presence of XPA (Fig.
3
, lane 3). These results indicate the formation of a ternary complex
in vitro
. Under these conditions, the interactions between XPA and RPA (Fig.
3
, lane 1) or XPA and GST-ERCC1 (Fig.
3
, lane 6) were also observed.
Direct interactions between ERCC1 or RPA and XPA were monitored using the
BIAcore system. GST-ERCC1 and RPA solutions of various concentrations were
passed over the XPA-immobilized sensor chip. Figure
4
A displays an example of overlaid sensorgrams using five different concentrations of ERCC1 (25-800 nM). The association rate constant (
k
ass
) and the dissociation rate constant (
k
diss
) for ERCC1 binding to the surface of XPA were 1.4 * 10
3
/M/S
and 3.5 * 10
-4
/S, respectively. The equilibrium dissociation constant (
K
D
), determined from the ratio of these two kinetic constants (
k
diss
/
k
ass
), was 2.5 * 10
-7
M. When RPA solutions (25-400 nM) were examined (Fig.
4
B), the binding was too fast to determine the reliability of the rate constants.
The
K
D
determined by Scatchard analysis of steady state binding levels (
R
eq
, relative response at equilibrium) at each concentration was 1.9 * 10
-8
M. Binding of RPA and GST-ERCC1 was not observed by the BIAcore system (data not shown).
Next we examined whether RPA and ERCC1 bound in any specific order to XPA.
First, 150 nM of RPA was passed over the XPA-sensor chip. The resonance signal increased to 50 RU 300 s after the injection. Then 150 nM of ERCC1 was injected. The
resonance signal increased to 206 RU, and the total resonance signal summed up
to 256 RU (Fig.
5
A). These results indicate that the ternary XPA-RPA-ERCC1 complex was formed in the order of injection. Conversely,
when ERCC1 was passed over the XPA-sensor chip first, the resonance signal increased to 193 RU 300 s after
the injection (Fig.
5
B). The amount and binding pattern was almost identical to those obtained by
ERCC1 injection in the above experiment, indicating that ERCC1 could bind to
XPA irrespective of the binding of RPA to XPA. RPA was then injected onto the
sensor chip which was pre-bound with ERCC1. Only a little signal increase was observed (7 RU) after
the RPA injection (Fig.
5
B), indicating that the pre-binding of ERCC1 to XPA prevents the additional binding of RPA to XPA.
When RPA and ERCC1 were injected simultaneously, the resonance signal increased
to 268 RU, which was almost equal to that obtained by the sequential bindings
of RPA and ERCC1 to XPA (Fig.
5
A and C). Injection of RPA or ERCC1 alone could not increase this score. All
these results suggest that RPA bound XPA first and then ERCC1 bound to XPA.
Previously we showed that p34 bound to XPA using the yeast two-hybrid system, and the direct binding of XPA and RPA was confirmed by
in vitro
experiments. We have also shown that RPA was co-immunoprecipitated with XPA from whole cell lysates, indicating the
existence of a complex of XPA and RPA
in vivo
(
47
). In this study, we showed that p70 bound to XPA
in vitro
as well as p34, which is consistent with the results of other groups (
48
,
49
). We confirmed that amino acids 4-29 and 98-187 of XPA were required for the binding of p34 and p70,
respectively. Recently, Li
et al.
showed that the 58 N-terminal amino acids and amino acids 153-176 of XPA are required for p34 and p70 binding, respectively (
49
). The region for the binding of p70 which we determined was broader. Since Li
et al.
used various C-terminal truncated forms of XPA it is not excluded that other regions of
XPA are required for p70 binding. Therefore, we think that these results are
consistent. We recently identified a region of XPA which is responsible for its
preferential binding to DNA damaged by UV or cisplatin, which consists of amino
acids 98-219 (MF122) and contains a C
4
type zinc finger motif (
51
,
58
; Fig.
6
). This region overlaps with the p70-binding region. It was shown that the damaged-DNA binding activity of XPA increased when XPA bound to RPA (
48
,
49
; our unpublished results), suggesting that the enhancement of damaged DNA-binding activity of XPA was caused by the direct binding of p70 to the DNA
binding domain of XPA and that the XPA-RPA complex is involved in the damage-recognition step of NER. It was recently shown that RPA itself
preferentially binds to damaged DNA (
48
,
59
,
60
), but complex formation between RPA and UV-damaged DNA was not affected by pre-treatment with dimer photolyase, which removes the cyclobutane
pyrimidine dimers from DNA without affecting the content of the (6-4) photoproducts (
60
). In fact, RPA bound to substrates with bubble or loop structures (
55
). RPA, and probably XPA also, may recognize the single-strandedness resulting from UV lesions.
We thank Dr M. S. Wold for human RPA expression vector (p11d-tRPA) and Dr J. Hurwitz for anti-p70 and p34 monoclonal antibodies. This work was supported by a
grant-in-aid for scientific research on priority areas from the Ministry of
Education, Science and Culture of Japan and Human Frontier Science Program.
*To whom correspondence should be addressed. Tel: +81 6 879 7971; Fax: +81 6 877
9136; Email: ktanaka@imcb.osaka-u.ac.jp
REFERENCES
Return