Cellular Stress Responses and Cancer
Protein–protein interactions during mammalian
DNA single-strand break repair
K. W. Caldecott1
Genome Damage and Stability Centre, University of Sussex, Science Park Road, Falmer, Brighton BN1 9RH, U.K.
Abstract
The genetic stability of living cells is continually threatened by endogenous reactive oxygen species and
other genotoxic molecules. Of particular threat are the thousands of single-strand breaks that arise in
each cell every day. If left unrepaired, such breaks can give rise to potentially clastogenic or lethal
chromosomal double-strand breaks. This article summarizes our current understanding of how mammalian
cells detect and repair single strand breaks, and provides insights into novel polypeptide components of this
process.
Origin and structure of DNA single-strand
breaks (SSBs)
SSBs are discontinuities in the sugar–phosphate backbone of
one strand of a DNA duplex. Hundreds of thousands of
cellular SSBs arise from DNA damage each day, and if
these are not repaired, they can be converted into potentially clastogenic and/or lethal DNA double-strand breaks
(DSBs). The two major sources of endogenous SSBs are
sugar damage and DNA base damage arising primarily
from attack by reactive oxygen species (ROS) and other
electrophilic molecules, and from the intrinsic instability
of DNA [1,2]. SSBs can also be induced by exposure to
environmental genotoxins, UV light or ionizing radiation
(e.g. X-rays). While SSBs arise directly from sugar damage,
by disintegration of deoxyribose, those from base damage
arise indirectly, through enzymic removal of the damaged
base or nucleotide by DNA base excision repair (BER).
Another source of SSBs is topoisomerase I (Topo1), which
nicks and reseals DNA as part of its catalytic activity.
Under certain conditions, such SSBs can be ‘uncoupled’ from
Topo1 activity and converted into proper single-stranded or
double-stranded breaks during transcription or DNA replication. Most SSBs are accompanied by the loss of a single
nucleotide at the site of the break and are thus actually
single-stranded gaps. In the case of direct SSBs induced by
endogenous ROS or ionizing radiation, base loss arises when
the damaged sugar becomes fragmented. In the case of indirect
SSBs, base loss occurs when the damaged base is excised
during BER. SSBs typically possess ‘damaged’ termini that
lack the conventional 5 -phosphate or 3 -hydroxy moieties.
The 3 -termini of direct SSBs induced by endogenous ROS
Key words: aprataxin, base excision repair, poly(ADP-ribose) polymerase-1, polynucleotide
kinase, XRCC1.
Abbreviations used: AP, apurinic/apyrimidinic; APE1, AP endonuclease 1; BER, base excision
repair; DSB, double-strand break; Lig1, DNA ligase I; Lig3α, DNA ligase IIIα; PARP-1, poly(ADP–
ribose) polymerase-1; PCNA, proliferating cell nuclear antigen; Polβ, DNA polymerase-β; Polδ/ε,
DNA polymerase δ/DNA polymerase ε; ROS, reactive oxygen species; SSB, single-strand break;
SSBR, SSB repair; Topo1, topoisomerase I; XRCC1, X-ray repair cross complementing group 1.
1
e-mail [email protected]
or ionizing radiation possess primarily monophosphate
(approx. 70%) or phosphoglycolate (approx. 30%) endgroups [2]. Although most 5 -termini at such breaks are of
the conventional 5 -phosphate format, some may possess a
hydroxy end-group [3,4]. Topo1-induced breaks possess both
5 -phosphate and 3 -hydroxy damaged termini, i.e. the reverse
format to that present at conventional DNA ends [5]. Indirect
SSBs arising during BER occur in one of two formats,
depending on the mechanism of base excision: a normal 3 terminus and ‘damaged’ 5 -deoxyribose phosphate terminus,
or a normal 5 -terminus and damaged 3 -α,β-unsaturated
aldehyde (αβ) terminus [6,7].
Effect of SSBs on genetic stability and
cell death
It is generally accepted that SSBs are far less toxic to cells
than their double-stranded counterparts; however, this largely
reflects the ability of cells to rapidly repair large numbers of
SSBs, rather than an intrinsic lack of toxicity of these lesions.
Indeed, any SSB not repaired prior to DNA replication
or actively ‘tolerated’ in the appropriate way during DNA
replication will become a DSB. Given the frequency of
‘spontaneous’ SSBs, it is thus not surprising that cells have
evolved highly efficient mechanisms to minimize their impact.
Perhaps the most compelling experimental evidence for this
emerges from mammalian cells in which there is a defect in
SSB repair (SSBR). Rodent cells lacking the molecular scaffold
protein X-ray repair cross complementing group 1 (XRCC1)
[8] exhibit elevated cellular sensitivity to a range of agents
that induce the SSB directly or indirectly, including alkylating
agents, ionizing radiation, hydrogen peroxide, and camptothecin [9]. Moreover, XRCC1−/− embryos fail to develop
beyond embryonic day 7.5/8.5, suggesting that endogenous
levels of SSBs are incompatible with embryonic development in the absence of efficient SSBR [10]. Un-rejoined
SSBs also threaten genetic stability. XRCC1-mutant rodent
cells exhibit increased frequencies of ‘spontaneous’ and
induced chromosome aberrations, including translocations
C 2003
Biochemical Society
247
248
Biochemical Society Transactions (2003) Volume 31, part 1
and genetic deletion [9,11]. The simplest explanation for this
is that un-rejoined SSBs are converted into DSBs during
chromosomal replication and thereby increase the risk of
unwanted genetic rearrangement. In addition to gross genetic
changes, SSBs can also result in point mutations, due presumably to occasional DNA polymerase mis-incorporation
during chromosome replication. Finally, SSBs may also affect
transcription. In one study in which transcription by T7
polymerase was examined in vitro, a single nucleotide gap
in the template strand failed to stop elongation, resulting
rather in a transcript shortened by one nucleotide [12].
However, SSBs possessing ‘damaged’ 3 -phosphate termini
did block SSBR, due apparently to effects of charge repulsion
between the neighbouring 3 - and 5 -terminal phosphates.
Thus, whereas SSBs with normal termini may result in mutant
transcripts, SSBs with 3 -phosphate termini, such as those
caused by endogenous ROS or ionizing radiation, may block
transcription.
Detection and processing of SSBs
An early step in the repair of SSBs appears to be the
rapid binding by poly(ADP–ribose) polymerase-1 (PARP-1),
a molecular ‘nick sensor’ that is activated at SSBs and
that synthesizes negatively charged polymers of ADP-ribose
[7,13]. A model for the repair of mammalian SSBs is presented in Figure 1. The role of PARP-1 is unclear, but
may include inhibiting unwanted recombination at SSBs
by charge repulsion [14] or synthesizing ATP cofactor
for the DNA ligation step of SSBR [15]. PARP-1 may
help sequester other repair proteins to the sites of SSBs,
a notion supported by its ability to interact with XRCC1
[16–19]. As discussed above, the termini of many SSBs
are ‘damaged’ as they lack conventional 3 -hydroxy and/or
5 -phosphate termini. Such termini are often described as
‘blocked’, because they are unable to serve as primers
for DNA polymerases or support the activity of DNA
ligases. After the dissociation of PARP-1 from the SSB, the
damaged DNA ends must thus be restored to conventional
3 -hydroxy and 5 -phosphate end-groups to enable gap filling
and DNA ligation. The removal of 5 -deoxyribose phosphate
termini from indirect SSBs created during BER is achieved
primarily by the apurinic/apyrimidinic (AP)-lyase activity
of DNA polymerase-β (Polβ) [20]. This is an attractive
scenario because, as discussed below, Polβ is also the DNA
polymerase most commonly employed for gap filling at
SSBs. The 3 -phosphoglycolate or 3 -monophosphate termini
present at directly induced SSBs, and the 3 -α,β-unsaturated
aldehyde termini at indirectly induced SSBs, are substrates
for the diesterase activity of mammalian AP endonuclease 1
(APE1/HAP1). However, this activity is relatively weak and
it is possible that other enzymes can process these termini.
One possible candidate is the mammalian homologue of
T4 polynucleotide kinase. Polynucleotide kinase possesses
both 5 -hydroxyl and 3 -phosphatase activity, and could thus
remove both 5 -hydroxy and 3 -phosphate damaged termini
from SSBs [21,22]. XRCC1 interacts with polynucleotide
C 2003
Biochemical Society
kinase and stimulates both the DNA kinase and DNA phosphatase activity of this enzyme, and in so doing accelerates
SSBR reactions [8].
DNA repair synthesis
Once the damaged 3 -termini at SSBs have been restored to
their conventional (hydroxy) configuration, gap filling can
occur. Since a single nucleotide is lost at most sites of SSB,
gap filling need only replace a single nucleotide; however,
several nucleotides can be replaced if strand displacement
occurs downstream of the gap followed by removal of
the resulting single-stranded flap, by flap endonuclease 1
(FEN1) [23]. These two distinct scenarios are defined as
‘short-patch’ or ‘single-nucleotide’ repair, and ‘long-patch’
repair respectively. Unfortunately, with the exception of
rodent cells lacking Polβ, cell lines that lack the ‘error-free’
polymerases that are candidates for gap filling are unavailable,
due presumably to essential roles in chromosome replication.
Consequently, most of our understanding of polymerase
utilization during DNA repair has emerged from DNA repair
reactions conducted in vitro using permeabilized cells, cell
extract, or purified proteins. Many studies have implicated
Polβ in gap filling at both directly induced SSBs and those
arising during BER [24–26]. Furthermore, both genetic and
biochemical analyses reveal that human Polβ can interact
with APE1 [27] and XRCC1 [16,28]. These interactions may
serve to recruit Polβ to SSBs for the process of gap filling.
Indeed, in the case of APE1, the interaction appears to occur
only in the presence of DNA substrate. An attractive model
of how the N-terminal domain of XRCC1might interact
with Polβ at a single-nucleotide gap has also been presented
[29]. Polβ induces a 90◦ kink in DNA at a SSB and binds
the outside bend of this kink, and the model of Marintchev
et al. [29] suggests that XRCC1 may simultaneously bind
both the inside of this DNA bend and the palm-thumb domains of Polβ. It is possible that by encompassing the SSB
in this way, the single-strand region is protected from nucleolytic attack, which might otherwise create a DSB.
In addition to Polβ, most in vitro studies also imply a
role for DNA polymerase δ and/or DNA polymerase ε
(Polδ/ε) in the repair of direct and indirect SSBs [23,30–35].
This has been suggested in part by reactions conducted
with cell-free extracts, in which some SSBR events are
dependent upon proliferating cell nuclear antigen (PCNA),
an accessory protein that is both bound by Polδ/ε and
required for their activity. However, recent experiments have
raised another possible role for PCNA during SSBR, in
stimulating FEN1 endonuclease activity during long-patch
repair. Nevertheless, there is much evidence, from inhibitor
studies employing compounds such as aphidicolin, that
Polδ/Polε can contribute to SSBR reactions. Polβ appears
to conduct most (50–80%) of the gap filling during SSBR,
and primarily inserts a single nucleotide. In addition to this
‘short-patch’ repair, a smaller fraction of BER events appear
to involve the insertion of multiple nucleotides (approx.
2–10) in a long-patch process that can involve either Polβ
Cellular Stress Responses and Cancer
Figure 1 Model for mammalian SSBR
SSBs arise either directly (e.g. sugar damage or ‘spontaneous’ disintegration of abasic sites), or from Topo1 cleavage, or
indirectly (e.g. during BER of base damage). (a) PARP-1 binds to SSBs and is activated, thereby synthesizing chains of
poly(ADP–ribose). Poly(ADP–ribose) recruits XRCC1–Lig3α (L3) complex to the SSB via the affinity of these polypeptides for
this polymer. PARP-1 dissociates from the SSB and is replaced by XRCC1 to establish a molecular scaffold. (b) AP sites created
during BER are bound by APE1. Note that while APE1 can couple the creation of the SSB by its own endonuclease activity
to the following steps of SSBR by molecular ‘hand-off’ to subsequent enzymes, it is possible that some cleaved abasic sites
are not ‘coupled’ in this way. The resulting ‘exposed’ breaks may be chanelled into the PARP-1-dependent process (broken
arrow). (c) Damaged 5 - and 3 -termini (open unlabelled circles) are converted back into 5 -phosphate and 3 -hydroxy termini
by recruitment of the enzyme appropriate for the type of damaged terminus, e.g. APE1, tyrosylphosphodiesterase 1 (TDP1),
polynucleotide kinase (PNK), Polβ. Note that the affinity of Polβ, PNK and APE1 for the SSB, and the enzymic activity of
the latter two, may be increased by binding to XRCC1.(d) At AP sites during BER, APE1 recruits Polβ and possibly XRCC1 via
physical interaction, and cleaves the AP site, creating an SSB. This reaction may be stimulated by XRCC1. Polβ then removes
the deoxyribose phosphate moiety from the 5 -terminus of the break in a reaction that may be stimulated by APE1. (e) In the
case of most (approx. 80%) of direct and indirect SSBs, Polβ binds the gapped substrate and XRCC1, and gap filling inserts
one nucleotide (short-patch/single-nucleotide repair). (f) Under some circumstances (e.g. during BER if the 5 -terminus is an
oxidized deoxyribose phosphate moiety that Polβ cannot remove) gap filling extends beyond one nucleotide in the long-patch
mode (2–10 nucleotides). Polβ or Pol δ/ε can conduct this process. After long-patch synthesis, the displaced single-stranded
flap is removed by FEN1 in a reaction that can be stimulated by PCNA. DNA ligation of short-repair patches is conducted by
Lig3α (g) and long-repair patches by Lig1 (h).
or Polδ/ε. Recent studies with extracts from wild-type and
Polβ −/− cells suggest that most long-patch repair events
are conducted by Polβ −/− , with an aphidicolin-sensitive
polymerase (presumably Polδ/ε) fulfilling a ‘back-up’ role.
In addition to long patch repair, Polδ/ε may also be
able to fulfil some ‘back-up’ repair in a single-nucleotide,
short-patch process. A major challenge for the future is to
determine how polymerase choice is made in living cells; for
example, it is possible that Polβ operates globally throughout
the genome, whereas the replicative polymerases, Polδ/Polε,
have a specific role at SSBs encountered by a replication fork.
DNA ligation
In addition to multiple DNA polymerases, mammalian
cells also possess multiple DNA ligases. So far, three
C 2003
Biochemical Society
249
250
Biochemical Society Transactions (2003) Volume 31, part 1
Table 1 Yeast two-hybrid screen for XRCC1-interacting human
proteins
cDNA recovered
Number of clones
Lig3α
Polβ
DNA polynucleotide kinase phosphatase
Aprataxin
2
29
2
2
human DNA ligase genes have been identified, which encode
at least five different polypeptides [36]. A number of studies
have implicated DNA ligase IIIα (Lig3α) in the repair of
both direct SSBs and indirect SSB during BER. Moreover,
these studies suggest that Lig3α is involved primarily in shortpatch repair events involving Polβ; for example, during BER
of abasic sites in cell extracts, Lig3α appears to conduct
ligation specifically following short-patch gap filling [37],
which is mediated primarily by Polβ, and a preference
for Lig3α during Polβ-mediated repair has been observed
at direct SSBs [38]. Moreover, cellular Lig3α is bound
and stabilized by the molecular scaffold protein, XRCC1
[39,40], providing a physical link between this DNA ligase
and Polβ. In a similar situation that was observed during
gap filling, DNA ligation during SSBR does not appear
to be restricted to a single enzyme. In addition to Lig3α,
DNA ligase I (Lig1) is implicated in the repair of direct
SSBs and of indirect SSBs during BER, as determined by
biochemical experiments and by the cellular phenotype of
46BR, a Lig1-defective cell line [23,35,38]. Whereas Lig3α
appears to operate in SSBR mediated by Polβ, Lig1 appears
to operate primarily during long-patch SSBR mediated by
Polδ/ε [23,35,37,42]. Indeed, while Lig3α is linked to Polβ
via interaction with XRCC1, Lig1 is linked to Polδ/ε via
interaction with PCNA [42]. If Lig3α is primarily involved
in short-patch repair and Lig1 in long-patch repair, how
can Polβ, which interacts with Lig3α, be the major longpatch polymerase? One possible explanation for this arises
from the observation that Polβ can also interact with Lig1
[43]. Perhaps, via this interaction, Lig1 is sequestered into
the subset of SSBR events mediated by Polβ that are long
patch.
Perspectives
A number of critical questions remain to be answered. For
example, how is the choice of polymerase and choice of
long-patch versus short-patch repair made in living cells? Is
there any cell cycle specificity for the different mechanisms of
SSBR? What is the relationship of these processes to human
disease? Intriguingly, with respect to the latter question, a
recent two-hybrid screen for novel candidate SSBR proteins
that interact with XRCC1 recovered two independent cDNA
clones encoding aprataxin, the protein mutated in the ataxia
telangiectasia-like neurodegenerative disorder ataxia ocular
apraxia (Table 1). The validity of this screen was demonstrated
C 2003
Biochemical Society
by the recovery from the same screen of multiple cDNA
clones of Lig3α, Polβ, and polynucleotide kinase. The impact
of SSBR processes on human disease may thus be more
significant than previously thought.
References
1 Lindahl, T. (1993) Nature (London) 362, 709–715
2 Ward, J.E. (1998) in DNA Damage and Repair Vol. 2: DNA Repair in
Higher Eukaryotes (Nickoloff, J.A. and Hoekstra, M.F., eds), pp. 65–84,
Human Press, Totowa, NJ
3 Coquerelle, T., Bopp, A., Kessler, B. and Hagen, U. (1973) Int. J. Radiat.
Biol. Relat. Stud. Phys. Chem. Med. 24, 397–404
4 Lennartz, M., Coquerelle, T., Bopp, A. and Hagen, U. (1975) Int. J. Radiat.
Biol. Relat. Stud. Phys. Chem. Med. 27, 577–587
5 Nitiss, J.L. (1998) in DNA Damage and Repair (Nickoloff, J.A. and
Hoekstra, M.F., eds), pp. 517–537, Human Press, Totowa, NJ
6 Piersen, C.E., McCullough, A.K. and Lloyd, R.S. (2000) Mutat. Res. 459,
43–53
7 Wilson, III, D.M., Engelward, B.P. and Samson, L. (1998) in DNA Damage
and Repair Vol. 1: DNA Repair in Prokaryotes and Eukaryotes
(Nickoloff, J.A. and Hoekstra, M.F., eds), pp. 29–64, Human Press,
Totowa, NJ
8 Whitehouse, C.J., Taylor, R.M., Thistlethwaite, A., Zhang, H.,
Karimi-Busheri, F., Lasko, D.D., Weinfeld, M. and Caldecott, K.W. (2001)
Cell 104, 1–11
9 Thompson, L.H. and West, M.G. (2000) Mutat. Res. 459, 1–18
10 Tebbs, R.S., Flannery, M.L., Meneses, J.J., Hartmann, A., Tucker, J.D.,
Thompson, L.H., Cleaver, J.E. and Pedersen, R.A. (1999) Dev. Biol. 208,
513–529
11 Dominguez, I., Daza, P., Natarajan, A.T. and Cortes, F. (1998) Mutat. Res.
398, 67–73
12 Zhou, W. and Doetsch, P.W. (1994) Biochemistry 33, 14926–14934
13 de Murcia, G. and Menissier de Murcia, J. (1994) Trends Biochem. Sci. 19,
172–176
14 Lindahl, T., Satoh, M.S., Poirier, G.G. and Klungland, A. (1995) Trends
Biochem. Sci. 20, 405–411
15 Oei, S.L. and Ziegler, M. (2000) J. Biol. Chem. 275, 23234–23239
16 Caldecott, K.W., Aoufouchi, S., Johnson, P. and Shall, S. (1996)
Nucleic Acids Res. 24, 4387–4394
17 Dantzer, F., de La, R.G., Menissier-de Murcia, J., Hostomsky, Z.,
de Murcia, G. and Schreiber, V. (2000) Biochemistry 39, 7559–7569
18 Masson, M., Niedergang, C., Schreiber, V., Muller, S.,
Menissier-de Murcia, J. and de Murcia, G. (1998) Mol. Cell Biol. 18,
3563–3571
19 Pleschke, J.M., Kleczkowska, H.E., Strohm, M. and Althaus, F.R. (2000)
J. Biol. Chem. 275, 40974–40980
20 Matsumoto, Y. and Kim, K. (1995) Science 269, 699–702
21 Jilani, A., Ramotar, D., Slack, C., Ong, C., Yang, X.M., Scherer, S.W. and
Lasko, D.D. (1999) J. Biol. Chem. 274, 24176–24186
22 Karimi-Busheri, F., Daly, G., Robins, P., Canas, B., Pappin, D.J., Sgouros, J.,
Miller, G.G., Fakhrai, H., Davis, E.M., Le Beau, M.M. and Weinfeld, M.
(1999) J. Biol. Chem. 274, 24187–24194
23 Matsumoto, Y., Kim, K., Hurwitz, J., Gary, R., Levin, D.S., Tomkinson, A.E.
and Park, M.S. (1999) J. Biol. Chem. 274, 33703–33708
24 Dianov, G., Price, A. and Lindahl, T. (1992) Mol. Cell Biol. 12,
1605–1612
25 Singhal, R.K., Prasad, R. and Wilson, S.H. (1995) J. Biol. Chem. 270,
949–957
26 Sobol, R.W., Horton, J.K., Kuhn, R., Gu, H., Singhal, R.K., Prasad, R.,
Rajewsky, K. and Wilson, S.H. (1996) Nature (London) 379, 183–186
27 Bennett, R.A., Wilson, III, D.M., Wong, D. and Demple, B. (1997)
Proc. Natl. Acad. Sci. U.S.A 94, 7166–7169
28 Kubota, Y., Nash, R.A., Klungland, A., Schar, P., Barnes, D.E. and Lindahl, T.
(1996) EMBO J. 15, 6662–6670
29 Marintchev, A., Mullen, M.A., Maciejewski, M.W., Pan, B., Gryk, M.R. and
Mullen, G.P. (1999) Nat. Struct. Biol. 6, 884–893
30 DiGiuseppe, J.A. and Dresler, S.L. (1989) Biochemistry 28, 9515–9520
31 Fortini, P., Pascucci, B., Parlanti, E., Sobol, R.W., Wilson, S.H. and
Dogliotti, E. (1998) Biochemistry 37, 3575–3580
32 Frosina, G., Fortini, P., Rossi, O., Carrozzino, F., Raspaglio, G., Cox, L.S.,
Lane, D.P., Abbondandolo, A. and Dogliotti, E. (1996) J. Biol. Chem. 271,
9573–9578
Cellular Stress Responses and Cancer
33 Miller, M.R. and Chinault, D.N. (1982) J. Biol. Chem. 257, 46–49
34 Miller, M.R. and Chinault, D.N. (1982) J. Biol. Chem. 257,
10204–10209
35 Pascucci, B., Stucki, M., Jonsson, Z.O., Dogliotti, E. and Hubscher, U.
(1999) J. Biol. Chem. 274, 33696–33702
36 Timson, D.J., Singleton, M.R. and Wigley, D.B. (2000) Mutat. Res. 460,
301–318
37 Cappelli, E., Taylor, R., Cevasco, M., Abbondandolo, A., Caldecott, K. and
Frosina, G. (1997) J. Biol. Chem. 272, 23970–23975
38 Winters, T.A., Russell, P.S., Kohli, M., Dar, M.E., Neumann, R.D. and
Jorgensen, T.J. (1999) Nucleic Acids Res. 27, 2423–2433
39 Caldecott, K.W., Mckeown, C.K., Tucker, J.D., Ljungquist, S. and Thompson,
L.H. (1994) Mol. Cell Biol. 14, 68–76
40 Caldecott, K.W., Tucker, J.D., Stanker, L.H. and Thompson, L.H. (1995)
Nucleic Acids Res. 23, 4836–4843
41 Reference deleted
42 Levin, D.S., McKenna, A.E., Motycka, T.A., Matsumoto, Y. and
Tomkinson, A.E. (2000) Curr. Biol. 10, 919–922
43 Prasad, R., Singhal, R.K., Srivastava, D.K., Molina, J.T., Tomkinson, A.E. and
Wilson, S.H. (1996) J. Biol. Chem. 271, 16000–16007
Received 10 October 2002
C 2003
Biochemical Society
251