Download Transcription Domain-Associated Repair in Human Cells

MOLECULAR AND CELLULAR BIOLOGY, Dec. 2006, p. 8722–8730
0270-7306/06/$08.00⫹0 doi:10.1128/MCB.01263-06
Copyright © 2006, American Society for Microbiology. All Rights Reserved.
Vol. 26, No. 23
Transcription Domain-Associated Repair in Human Cells䌤
Thierry P. Nouspikel,1,2* Nevila Hyka-Nouspikel,1,2 and Philip C. Hanawalt2
Institute for Cancer Studies, University of Sheffield Medical School, Beech Hill Road, Sheffield S10 2RX,
United Kingdom,1 and Department of Biological Sciences, Stanford University, Stanford, California2
Received 12 July 2006/Returned for modification 7 August 2006/Accepted 18 September 2006
A profound property of NER is that it can be coupled to
transcription, usually resulting in the preferential repair of the
transcribed strand (TS) over that of the nontranscribed strand
(NTS) in active genes, a subpathway termed transcriptioncoupled repair (TCR). The mechanistic details of TCR are still
unclear, although it is generally assumed that RNA polymerase
II (RNAPII) serves as a damage sensor that signals the NER
system when it encounters a blocking lesion in the TS (12).
Thus, RNAPII can substitute for XPC (and DDB) in lesion
detection, and XP group C (XP-C) patients, deficient in global
genomic repair (GGR), still retain TCR. Deficiencies in the
TCR pathway can result in several other genetic diseases, including Cockayne syndrome, in which the patients are not
cancer prone but suffer from developmental defects and numerous neurological problems, generally fatal at an early age.
Mutations in XPG, XPB, and XPD (the last two encoding
subunits of TFIIH) and in two other genes, CSA and CSB, can
result in Cockayne syndrome (17).
The strikingly different phenotypes of XP and Cockayne
syndrome can be understood in terms of the two principal
DNA transactions: replication and transcription. The presence
of a lesion in a transcribed gene is likely to result in a nonfunctional protein or no protein at all. It may also block translocation of RNAPII and make it unavailable for the transcription of other genes. Evidence has even been presented that a
stalled RNAPII can be ubiquitinated and targeted for degradation (7). In any case, the end result is cellular dysfunction
and possibly cell death. By contrast, some DNA polymerases
can bypass damage, in a process that often results in the introduction of an error into the newly synthesized strand opposite the lesion (29). Such an accumulation of mutations in
proliferating cells is a predominant mechanism leading to cancer (39).
However, the latter should not be a problem in postmitotic
cells, which no longer replicate their genomes. If they could
DNA is constantly under attack from numerous damaging
agents from the external environment and as a consequence of
intracellular metabolism. The resulting damage, adding to that
due to the spontaneous “decay” of DNA (18), would quickly
cripple a genome comprising several billion base pairs were it
not for continuous surveillance by multiple DNA repair systems. Among these systems, nucleotide excision repair (NER)
is the most versatile: it can recognize and repair a wide variety
of lesions, from UV-induced pyrimidine dimers to bulky chemicals or protein DNA adducts to intrastrand cross-links.
The mechanistic details of NER are well understood. Lesions are likely detected through the conformational change
they introduce in the double-helical DNA structure by the
heterotrimer XPC/HR23B/Centrin2 with, for some lesions, an
initial contribution by the DDB heterodimer. Then other recognition enzymes, XPA and RPA, come into play, in part to
verify the presence of a bona fide lesion and to identify the
damaged strand. A denaturation bubble is opened around the
lesion by the general transcription factor TFIIH, and the damaged strand is nicked by XPG on the 3⬘ side of the lesion and
by the heterodimer ERCC1/XPF on the 5⬘ side. Finally, an
oligonucleotide of roughly 30 nucleotides encompassing the
lesion is displaced, and the resulting gap is filled using the
intact strand as a template (19).
The importance of NER is dramatically illustrated by the
genetic disease xeroderma pigmentosum (XP), in which one of
the NER enzymes is absent or inactive. XP patients suffer from
multiple cancers in sun-exposed areas of their bodies, as well as
a modest increase in internal cancers (6).
* Corresponding author. Mailing address: Institute for Cancer Studies, University of Sheffield Medical School, Beech Hill Road, Sheffield
S10 2RX, United Kingdom. Phone: 44 114 271 2966. Fax: 44 114 271
3892. E-mail: [email protected].
䌤
Published ahead of print on 2 October 2006.
8722
Downloaded from mcb.asm.org at RUTGERS UNIVERSITY on April 1, 2007
Nucleotide excision repair (NER), which is arguably the most versatile DNA repair system, is strongly
attenuated in human cells of the monocytic lineage when they differentiate into macrophages. Within active
genes, however, both DNA strands continue to be proficiently repaired. The proficient repair of the nontranscribed strand cannot be explained by the dedicated subpathway of transcription-coupled repair (TCR), which
is targeted to the transcribed strand in expressed genes. We now report that the previously termed differentiation-associated repair (DAR) depends upon transcription, but not simply upon RNA polymerase II (RNAPII)
encountering a lesion: proficient repair of both DNA strands can occur in a part of a gene that the polymerase
never reaches, and even if the translocation of RNAPII is blocked with transcription inhibitors. This suggests
that DAR may be a subset of global NER, restricted to the subnuclear compartments or chromatin domains
within which transcription occurs. Downregulation of selected NER genes with small interfering RNA has
confirmed that DAR relies upon the same genes as global genome repair, rather than upon TCR-specific genes.
Our findings support the general view that the genomic domains within which transcription is active are more
accessible than the bulk of the genome to the recognition and repair of lesions through the global pathway and
that TCR is superimposed upon that pathway of NER.
VOL. 26, 2006
MECHANISM OF DAR
8723
maintain the integrity of their active genes, for instance via
TCR, such cells could largely dispense with NER at the global
genome level, provided that they do not alter their pattern of
gene expression. Cells that neither divide nor change phenotype are known as terminally differentiated and are common in
higher organisms. A classic example is neurons, and we have
reported that NER is greatly diminished at the global genomic
level in human hNT neurons when compared with their NT2
precursor cells (25). Similar results were observed in human
fetal neurons kept in culture for several months (24). The
repair of the principal UV-induced lesion, cyclobutane-pyrimidine dimers (CPDs), was virtually nil in either system, whereas
the more distorting, but less frequent, (6-4) pyrimidine-pyrimidone photoproducts [(6-4)PPs] were repaired much less rapidly in hNT neurons than in their NT2 precursor cells. Similarly, the repair of benzo[a]pyrene diol-epoxide adducts was
markedly reduced after differentiation (21). By contrast, TCR
was still very efficient in removing CPDs from the TS of active
genes. Surprisingly, the NTS was also proficiently repaired in
these active genes, a phenomenon that we originally termed
differentiation-associated repair (DAR) (25). We reasoned
that DAR may be needed to maintain the integrity of the NTS,
which must serve as a template for the patching step in TCR.
Attenuating GGR might be a valid strategy in cells that are not
expected to ever replicate their genomes, but it could be catastrophic if lesions were allowed to accumulate in the NTS of
active genes over a lifetime. Sooner or later, the repair DNA
polymerase would encounter a neglected lesion within the template for the 29-nucleotide single-strand DNA patch created by
TCR, which would likely result in the introduction of a mutation in the gene being repaired (26). The present study was
undertaken to further analyze the molecular mechanisms of
DAR, as well as to determine the generality of the DNA repair
phenotype that we have observed in neurons.
MATERIALS AND METHODS
Cell culture. HL60 (16) and THP1 (ATCC) cells were grown in RPMI 1640
medium (Gibco) with 10% fetal calf serum. HL60 cells were differentiated with
32 nM TPA (12-O-tetradecanoylphorbol-13-acetate; Calbiochem) for 16 or 48 h,
and nonadherent cells were washed away before harvest. THP1 cells were differentiated with 100 nM TPA for 48 h. Fresh monocytes were isolated as described previously (1) from buffy coats obtained from the Stanford Blood Center.
For experiments involving transcription inhibitors, cells were preincubated for
90 min with either 100 ␮M 5,6-dichlorobenzimidazole riboside (DRB; Calbiochem) or 10 ␮g/ml ␣-amanitin (Calbiochem). Cells were further incubated with
transcription inhibitors after UV irradiation.
GGR assay. Cells were suspended or rinsed with phosphate-buffered saline
(PBS) and irradiated with 254-nm UV light with a Westinghouse IL-782-30 15-W
lamp, at a fluency of ⬃111 mW/m2, for a total dose of 10 J/m2. Cells were
harvested by centrifugation or scraping, and DNA was prepared by phenol
extraction as described previously (32). Purified DNA was boiled for 5 min and
slot blotted in triplicate on nitrocellulose membranes (Bio-Rad): 1 ␮g per slot for
(6-4)PPs and 80 ng per slot for CPDs. Membranes were baked for 30 min at 80°C
and probed with monoclonal antibodies against CPDs and (6-4)PPs (20) as
described previously (25).
TCR assay. The same DNA samples prepared for the GGR assay were also
used to appraise strand-specific repair of CPDs, as described previously (32).
Briefly, 50 ␮g DNA was digested with an appropriate restriction enzyme (KpnI
for DHFR, XbaI for c-myc), ethanol precipitated, and dissolved in 10 mM Tris
(pH 8.0)–1 mM EDTA (TE). Each sample was precisely divided in two and
either digested with T4 endonuclease V or mock treated. DNA was resolved on
an alkaline agarose gel, transferred to a nitrocellulose membrane, and probed
with strand-specific RNA probes for DHFR intron IV (33) or for c-myc exon 3
(16). Results were quantified with a GS-363 PhosphorImager (Bio-Rad).
Chromatin immunoprecipitation-PCR assay. HL60 cells differentiated with
TPA for 16 h were incubated with 1% formaldehyde in PBS for 20 min at room
temperature to cross-link proteins to DNA. The reaction was quenched with 250
mM glycine (final), and the cells were washed with ice-cold PBS, resuspended in
FA buffer (10 mM Tris, pH 7.5, 1 mM EDTA, 150 mM NaCl, 1% Triton X-100,
Downloaded from mcb.asm.org at RUTGERS UNIVERSITY on April 1, 2007
FIG. 1. Global genome repair of UV-induced lesions. HL60 cells, naive (dotted lines) or differentiated with TPA for 16 h (dashed lines) or 48 h
(solid lines), THP1 cells, and fresh blood monocytes, naive or differentiated with TPA for 48 h, were irradiated with 10 J/m2 of 254-nm UV light.
DNA was isolated at various times after irradiation, blotted on a nitrocellulose membrane, and probed with monoclonal antibodies against CPDs
(top row) or (6-4)PPs (bottom row). Data are averages of two to four experiments, each performed in triplicate. Error bars are standard errors
of the means. Asterisks denote significant differences from naive cells: *, P ⬍ 0.1; **, P ⬍ 0.05; ***, P ⬍ 0.025; ****, P ⬍ 0.01.
8724
NOUSPIKEL ET AL.
MOL. CELL. BIOL.
RESULTS
We made use of two human leukemia cell lines that can be
differentiated into macrophages, another canonical example of
terminally differentiated cells. HL60 is a promyelocytic leukemia cell line that can be differentiated into macrophage-like
cells with TPA (30) or into granulocyte-like cells with various
chemical agents (8). HL60 cells are unfortunately genetically
unstable, and they evolve in culture as additional mutations
and chromosomal abnormalities occur (10). The monocytic
leukemia cell line THP1 is farther down the monocyte/macrophage differentiation pathway, roughly corresponding to primary monocytes, and more stable than HL60. It too can be
differentiated into macrophage-like cells by addition of TPA to
the culture medium (37). To verify that our observations were
not due to aberrations specific to the leukemia cell lines, we
also employed primary monocytes freshly isolated from human
blood. These cells can also be induced to differentiate into
macrophages by treatment with TPA.
We assessed the NER system in these three cell types,
whether naive or differentiated, by measuring the repair of two
principal UV-induced lesions, CPDs and (6-4)PPs. Both involve noncanonical covalent bonds between two adjacent pyrimidines on the same DNA strand. CPDs constitute a strong
block for RNAPII and are therefore readily repaired by TCR
but are a weak substrate for GGR, and their detection requires
an additional component, DDB, at the global genome level. By
contrast, (6-4)PPs are among the most readily recognized and
rapidly repaired substrates for NER.
Global genome repair of UV-induced lesions. Freshly isolated human monocytes, HL60 cells, or THP1 cells were differentiated with TPA (or not) for 48 h and, in the case of HL60
cells, also for 16 h. To measure removal of UV-induced lesions
at the global genome level, cells were irradiated with 10 J/m2 of
FIG. 2. Strand-specific repair of CPDs in the DHFR gene. (A)
DNA prepared as for Fig. 1 was digested with KpnI and then with T4
endonuclease V and subjected to Southern analysis with RNA probes
specific for the transcribed (filled symbols) or nontranscribed (open
symbols) strand of the constitutively expressed DHFR gene. Dotted
lines, naive cells; solid lines, cells differentiated with TPA for 48 h.
Data are averages of two experiments, and error bars are standard
errors of the means. (B) Same data as those for panel A combined with
those in Fig. 1. Repair of the nontranscribed strand in differentiated
cells was replotted after subtraction of the amount of GGR observed
in Fig. 1. Upward triangles, monocytes; downward triangles, THP1;
leftward triangles, HL60. Error bars were omitted for clarity.
254-nm UV light and harvested after various periods of incubation at 37°C. DNA was purified and blotted on a nitrocellulose membrane, and the amounts of CPDs and (6-4)PPs were
measured with specific monoclonal antibodies (Fig. 1).
Naive HL60 cells repaired CPDs relatively efficiently,
whereas in cells differentiated with TPA for 16 h very little
repair of CPDs occurred within the first 8 h after irradiation.
Downloaded from mcb.asm.org at RUTGERS UNIVERSITY on April 1, 2007
0.1% sodium dodecyl sulfate, 0.1% sodium deoxycholate, 1 mM phenylmethylsulfonyl fluoride), and lysed by repeated aspiration through a 25-gauge needle.
The lysate was spun 30 min at 45,000 rpm, and the pellet was quickly rinsed with
TE, triturated into a solution containing 10 mM Tris, pH 8.0, 1 mM EDTA, and
150 mM NaCl, and sonicated for 5 min at 10% duty cycle with a Branson Sonifier
200 at 30% power. The DNA was then either digested with the restriction
enzyme AvaI for 4 h at 37°C or further sonicated for 50 min (both methods gave
identical results). The lysate was incubated in FA buffer with agarose beads
precoupled to an anti-RNA polymerase II antibody (H224; Santa Cruz Biotechnology) for 90 min at room temperature. The beads were washed extensively and
treated with 0.8 mg/ml pronase in TE for 1 h at 42°C and 5 h at 65°C. DNA was
phenol extracted and precipitated as described previously (32). PCR was performed as described previously (23) with primers specific for exon 1 (GTTTTC
GGGGCTTTATCTAACTCG and CGCTGCTATGGGCAAAGTTTCGTG)
or for exon 3 (GTCTCCACACATCAGCACAATAC and TTCCTTACTTTTC
CTTACGCACAA) of the c-myc gene. Amplification products were analyzed by
agarose gel electrophoresis and staining with ethidium bromide.
siRNA transfection. Proprietary transfection vectors were purchased from
GenScript to generate small interfering RNA (siRNA) against XPG, XPC, or
CSB, under the control of the U3 promoter. These vectors were stably transfected into THP1 cells, which were bulk selected with increasing doses of hygromycin (Gibco) up to 150 ␮g/ml for 5 weeks. The cultures were then serially
diluted in 96-well plates and kept in culture for several weeks with lower doses
of hygromycin (100 to 150 ␮g/ml). Sixteen clones corresponding to the most
dilute conditions were selected. RNA was prepared, reverse-transcribed with an
oligo(dT) primer, and amplified by PCR in the presence of [32P]dCTP with
primers specific for the genes of interest and located upstream of the siRNA
recognition site. Fragments of ␤-actin or GAPDH (glyceraldehyde-3-phosphate
dehydrogenase) were coamplified in the same reaction for normalization
purposes.
VOL. 26, 2006
MECHANISM OF DAR
8725
FIG. 3. Strand-specific repair in the c-myc gene in HL60 cells. (A)
The c-myc gene is expressed in naive HL60 cells, inactive in HL60 cells
differentiated for 48 h, and transcribed only to the end of the first exon
in HL60 cells differentiated for 16 h. X, XbaI restriction site; brackets,
location of the RNA probes. (B) Experiments similar to those in Fig.
2 but in which DNA was digested with XbaI and probed for a restriction fragment located at the end of the c-myc gene. Filled squares,
probe specific for the TS; open squares, probe specific for the NTS.
Data are averages of two experiments, and error bars are standard
errors of the means. (C) PCR amplification of the first or third
exon of c-myc, using as template either total DNA or DNA that was
immunoprecipitated (IP) with antibodies against RNA polymerase II,
after formaldehyde cross-linking and DNA shearing by ultrasounds
and restriction digestion. A successful amplification indicates that
RNA polymerase II was present in this portion of the gene at the time
of cross-linking.
Downloaded from mcb.asm.org at RUTGERS UNIVERSITY on April 1, 2007
Significant repair was apparent 24 h after irradiation, but much
less than that in naive cells. CPD repair was completely abolished when HL60 cells were differentiated for 48 h with TPA.
This effect of differentiation was less clear with THP1 cells and
primary monocytes, mainly because these cells displayed little
repair of CPDs even in their naive state. Nevertheless, it is
obvious from Fig. 1 that neither macrophage-like cells differentiated from THP1 cells nor macrophages differentiated from
primary monocytes significantly repaired CPDs.
As is generally the case, (6-4)PPs were repaired much more
efficiently than CPDs in all three cell types, although, interestingly, repair was somewhat slower in primary monocytes. Repair of (6-4)PPs was decreased upon differentiation but not
completely abolished. Rather, the kinetics of repair were significantly slower in all three cell types after differentiation by
TPA for 48 h. This effect was not apparent in HL60 cells after
only 16 h of differentiation with TPA.
Strand-specific repair in the DHFR gene. To measure repair
of CPDs in a transcribed gene, we performed strand-specific
Southern blot analyses on the same DNA samples used for
GGR appraisal, after digestion with T4 endonuclease V. This
enzyme generates a single-strand break at the site of a CPD,
which results in the disappearance of the restriction fragment
of interest from the Southern blot, allowing us to appraise the
number of lesions in a given gene. By using denaturing gels and
strand-specific RNA probes, this method can also reveal any
difference in repair rates from one strand to the other (32).
Figure 2 demonstrates that CPDs were efficiently removed
from both strands of the constitutively active DHFR gene in
naive cells (dotted lines), although significantly faster from the
transcribed strand than from the nontranscribed strand (filled
symbols versus open symbols). This strand bias is the hallmark
of TCR and probably reflects the fact that RNAPII constitutes
an efficient, processive detection system for lesions that arrest
transcription.
In primary macrophages or macrophage-like cells (solid
lines) differentiated from either HL60 or THP1 cells, the TS of
the DHFR gene was proficiently repaired, as expected for an
actively transcribed gene. Interestingly, the NTS was also proficiently repaired, in spite of the lack of GGR for CPDs in
differentiated cells (Fig. 1). Figure 2B combines data from Fig.
1 and 2A and shows repair of the NTS in differentiated cells
after subtraction of the amount of global repair observed in
these cells. In other words, it displays a supplementary repair
component, approximately equal to 25%, which cannot be
accounted for by GGR. Since RNAPII does not detect CPDs
in the NTS (9), this additional repair also cannot be attributed
to TCR. We thus concluded that the NTS had to be repaired
by DAR, confirming the phenotype that we have observed in
neurons (25).
8726
NOUSPIKEL ET AL.
Downloaded from mcb.asm.org at RUTGERS UNIVERSITY on April 1, 2007
Strand-specific repair in the c-myc gene. In HL60 cells, the
c-myc gene is subject to a very peculiar regulation (Fig. 3A): it
is active in naive cells and inactive in differentiated cells 48 h
after treatment with TPA. However, 16 h after treatment with
TPA, the gene is still partially transcribed but RNAPII never
proceeds beyond the first exon (3). We verified this fact in our
laboratory, both with “run-on” experiments (16) and with a
technique involving chromatin immunoprecipitation and PCR
(Fig. 3C). Both approaches confirmed that the distal part of
the c-myc gene was not transcribed in HL60 cells differentiated
for 16 h.
We then performed strand-specific Southern blotting on the
c-myc gene, selecting a restriction site and a probe located in
the distal part of the gene, thereby to measure repair of CPDs
in the region that translocating RNAPII does not reach in
TPA-treated cells. Figure 3B illustrates the results of these
experiments. As expected, in naive cells (top panel) both
strands of the c-myc gene are proficiently repaired, with a
noticeable bias in favor of the TS, indicating that both GGR
and TCR are proficient. After 48 h of treatment with TPA
(bottom panel), very little repair is observed in either strand of
the now-silent c-myc gene. This was expected as GGR is downregulated (Fig. 1), and TCR occurs only in active genes. The
surprising results were those obtained with HL60 cells differentiated with TPA for only 16 h (middle panel): both strands
were equally well repaired, in spite of the fact that GGR of
CPDs is essentially inactive in these cells (Fig. 1). Because
RNAPII never reaches the part of the gene that we were
assaying, TCR cannot account for repair in that domain, as
demonstrated by the absence of strand bias. We thus concluded that both strands were repaired by DAR.
Effect of transcription inhibitors. The above experiment
could not be performed in THP1 cells, because the c-myc gene
is inactive in this cell line, even prior to differentiation. We
therefore decided to use two transcription inhibitors, DRB and
␣-amanitin. DRB inhibits the phosphorylation of the C terminus of the large subunit of RNAPII, thereby preventing it from
entering the elongation mode and leaving the promoter. By
contrast, ␣-amanitin interacts with an elongating RNAPII and
prevents incorporation of ribonucleotides into the nascent
RNA chain, thereby stalling the polymerase. Thus, using these
inhibitors we could theoretically distinguish between two situations: a polymerase bound to the promoter that never initiates
transcription and a polymerase stalled within the gene during
RNA synthesis.
As shown in Fig. 4, both treatments resulted in the same
phenotype, similar to that observed for the c-myc gene in HL60
cells differentiated for 16 h. There was no strand bias in favor
of the TS, as expected, since transcription had been shut off.
However, even though these cells displayed almost no repair of
CPDs at the global genome level (data not shown; similar to
Fig. 1), both strands of the DHFR gene were repaired by DAR.
Depleting NER enzymes with siRNA. The above results are
compatible with a model in which DAR is a subset of GGR,
which operates only in transcriptionally active domains of the
genome. If this hypothesis is correct, DAR should require the
same NER enzymes as those needed for GGR. In particular, it
should rely on XPC, which is required for GGR but not for
TCR. Conversely, if DAR were a variant of TCR, it should
require CSA and CSB, both specific for TCR.
MOL. CELL. BIOL.
FIG. 4. Strand-specific repair in THP1 cells treated with transcription inhibitors. (A) Experiment similar to that in Fig. 2, except that
THP1 cells were treated with transcription inhibitors DRB or ␣-amanitin. Only the results in differentiated cells are presented. (B) Quantification of the results from panel A. Filled triangles, probes specific
for the TS; open triangles, probe specific for the NTS.
We transfected THP1 cells with stable siRNA constructs
designed to destroy mRNA for XPC, CSB, or XPG (as a
control). Clones that achieved at least 70% decrease in the
respective mRNA levels were selected (data not shown). As a
control for the efficiency of the siRNA approach, we also
measured the effect of siRNA transfection on GGR. Given
that CPDs are poorly repaired in THP1 cells, only the repair of
(6-4)PPs in differentiated cells is shown in Fig. 5A. It can be
VOL. 26, 2006
MECHANISM OF DAR
8727
seen that transfection of an anti-XPC siRNA construct resulted
in a very significant decrease in the removal of (6-4)PPs. By
contrast, transfection of an anti-XPG siRNA construct had a
much less marked effect. This was not completely unexpected,
as it has been shown that even small amounts of XPG mRNA
are enough to greatly alleviate the phenotype of XP-G patients
(36). It has even been observed that healthy individuals display
an abnormal splicing event in XPG, sometimes resulting in a
90% decrease in mRNA levels, with no obvious phenotypic
consequences (22), although the possibility that the amount of
abnormally spliced mRNA varies in a tissue-specific manner
and was higher in the samples provided by the donors than in
more-vital organs cannot be excluded. Nevertheless, we were
not surprised that a 70% decrease in XPG mRNA did not have
a marked effect on GGR. As for CSB, since this enzyme only
affects TCR, it would not be expected to have an effect greater
than a few percent at the global genome level.
We then measured strand-specific repair of CPDs in the
various transfectants before (not shown) and after differentiation. Figure 5B demonstrates a strict dependence on XPG for
DAR; neither strand is repaired in cells transfected with antiXPG siRNA. This is the result we were expecting since XPG is
absolutely required for NER, both at the global genome level
and when coupled to transcription. It is interesting that, in
spite of the relatively proficient repair of (6-4)PPs at the global
genome level, TCR of CPDs is almost completely abolished in
the transfected cells after differentiation. We do not have a
definitive explanation for this difference in behavior between
the two lesions. It may be that, when XPG becomes limiting,
the vast majority of (6-4)PPs (generally repaired by GGR more
rapidly than CPDs are repaired by TCR) result in titrating of
XPG away from the few molecules of RNAPII stalled by a
CPD in a transcribed gene. Another possible explanation is
that a 70% decrease in XPG may have indirect effects on NER,
for instance, through the transcription machinery or chromatin
structure. In this respect, it is worth noting that an analysis of
XPG mutations suggested that XPG must have an additional
function besides the incision role it plays in NER (27).
Cells transfected with anti-XPC siRNA displayed the typical
phenotype of XP-C cells: proficient repair of the TS (by TCR
Downloaded from mcb.asm.org at RUTGERS UNIVERSITY on April 1, 2007
FIG. 5. Effect of siRNA directed against CSB, XPC, or XPG. (A) Global genomic repair in siRNA-transfected cells, measured as for Fig. 1.
Only (6-4)PPs are shown. (B) Strand-specific repair in the DHFR gene, measured as for Fig. 2. Only differentiated cells are shown. Filled symbols,
probe specific for the TS; open symbols, probe specific for the NTS. Error bars are standard errors of the means of two experiments.
8728
NOUSPIKEL ET AL.
only, hence lower than in naive cells in which it benefits from
GGR and TCR) and complete lack of repair of the NTS. Thus,
XPC is required for proficient DAR. By contrast, cells transfected with anti-CSB siRNA proficiently repaired CPDs in
both strands of the active DHFR gene. There was no strand
bias in favor of the TS, confirming that the absence of CSB
inhibited TCR but that DAR was still observed on both
strands.
DISCUSSION
(25). In this respect, it is interesting that DAR is present in
white blood cells, which have a limited life span. The probability is fairly low of a lesion occurring in the TS close enough
(i.e., within 30 nucleotides at most) to a lesion in the NTS that
a mutation would result upon TCR. It would probably take
years before the NTS is crippled to such an extent that it
eventually compromises TCR in any given gene. Clearly, this
might be a problem for neurons, which are supposed to last for
a lifetime, but not for short-lived white blood cells. Thus, we
would not have been surprised if DAR had not been present in
macrophages. Yet we did observe it, suggesting that DAR may
be a more common mode of repair than we originally thought.
A convenient target gene to study the mechanism of DAR is
the c-myc gene in HL60 cells. This gene is active in naive HL60
cells and eventually becomes silent once the cells have been
differentiated for 48 h. After only 16 h of differentiation, however, the gene still appears to be transcribed, but only to the
end of the first exon. There are two schools of thought about
this phenomenon: some believe that the polymerase actually
travels to the end of the first exon and is arrested there (3), and
other groups believe that the RNAPII never actually leaves the
promoter in undisturbed cells and that the observed stretch of
transcription occurs upon preparation of nuclei (34). We attempted to mimic either situation in the DHFR gene in THP1
cells by using different transcription inhibitors: DRB to prevent
RNAPII from leaving the promoter, or ␣-amanitin to stall it
while actively engaged in transcription. The same NER phenotype resulted from both inhibitors: despite the lack in global
genome repair, the DHFR gene was still repaired on both
strands, without any bias in favor of the TS. This was also
observed in the c-myc gene in HL60 cells differentiated for
16 h. These experiments lead to important conclusions about
the mechanism of DAR. They demonstrate, first, that DAR is
active on both DNA strands and, second and most importantly,
that DAR depends upon active transcription but does not rely
upon translocating RNAPII to actually detect lesions. The
simplest model to account for these observations is that DAR
is merely a subset of GGR, restricted to the nuclear subcompartments within which transcription occurs. Bringing a gene
into such “transcription factories” would facilitate proficient
repair on both strands, regardless of whether the gene is transcribed or not.
This is reminiscent of previous observations in the laboratory of George Kantor, in which large DNA domains (over 50
kb) were shown to be preferentially repaired in domains containing active transcription units. However, this “proximity effect” was apparent only on the TS and occurred in XP-C cells
(2) but not in Cockayne syndrome group B cells (31), implying
that it was somehow an extension of TCR.
By contrast, we observed DAR on both strands, and our
model predicts that DAR should rely upon the same NER
enzymes as GGR. In particular, it should require the early
detection enzyme XPC. XP-C cells lack GGR but have proficient TCR (13), a phenotype very similar to what we observed
in differentiated cells, except that the NTS of active genes is
not repaired in XP-C cells. This may be because XPC is required for DAR, but it could also be due to the fact that the
only cells available from XP-C patients are actively growing
fibroblasts and lymphoblasts. Conceivably, if we had access to
differentiated cells from XP-C patients, we might observe
Downloaded from mcb.asm.org at RUTGERS UNIVERSITY on April 1, 2007
We have examined modulations of NER in human cells of
the myelomonocytic lineage when they are induced to differentiate into macrophages. We observed a general drop in efficiency of NER at the global genome level upon differentiation, whereas the repair of both strands of transcribed genes
remained proficient. Proficient repair of the TS can be attributed to TCR, whereas repair of the NTS in GGR-deficient
cells constitutes what we have previously named DAR. This
phenotype is similar to what we have previously observed in
human neurons, either primary (24) or differentiated from a
precursor neuroteratoma cell line (25). It is thus possible that
all differentiated cells may experience an attenuation of global
genome repair, possibly an energy-saving strategy, while maintaining proficient repair of active genes via TCR and DAR.
We did not observe any marked changes in the expression of
the various NER enzymes upon macrophage differentiation,
with the exception of a twofold reduction in the level of the
p48/DDB2 subunit of DDB (14). However, this reduction was
not observed in neurons (25) or in macrophages originating
from a third acute myeloid leukemia cell line, KG-1, which
displayed a GGR phenotype similar to that of HL60 and THP1
(14). This makes it unlikely that the decrease in p48/DDB2
could be the cause of the GGR defect we observed.
In addition, it is worth noting that GGR of CPDs was already relatively slow in naive cells from the monocytic lineage,
when compared to that generally observed in actively dividing
cells (25). In the case of THP1, this could be attributed to the
deficient p53 status of this cell line (28); it has been shown that
p53 regulates basal levels of the p48/DDB2 regulatory subunit
of the DDB enzyme (15), as well as regulating its induction
after DNA damage (11). Since DDB greatly enhances GGR of
CPDs (35), a p53-deficient cell line is not likely to repair CPDs
efficiently at the global genome level. The situation is less clear
for HL60 cells, due to the unstable nature of this cell line.
There have been reports of p53-proficient as well as of p53-null
HL60 sublines, and even sublines in which p53 is present but
inactive (10). We have tested two different sublines of HL60
(14, 16), and only one of these was reasonably proficient in the
repair of CPDs. That was, of course, the one used in the
present study.
TCR of CPDs does not require DDB, presumably because
these lesions constitute a strong block for RNAPII, which then
serves as a sensor (in place of DDB and XPC) and leads to
proficient repair of CPDs in the TSs of active genes. Of course,
while repairing the TS, TCR must use the NTS as a template
to fill the ⬃30-nucleotide gap left by the excision of an oligonucleotide containing the lesion. A teleological explanation for
DAR is that it serves to maintain the integrity of the NTS of
active genes, so as to preserve a flawless template for TCR
MOL. CELL. BIOL.
VOL. 26, 2006
MECHANISM OF DAR
ACKNOWLEDGMENTS
We thank Toshio Mori for the gift of anti-CPD and anti-(6-4)PP
antibodies, Stephen Lloyd for the gift of T4 endonuclease V, David
Mellert for assistance with RT-PCR, and Ann Ganesan for critical
reading of the manuscript and helpful suggestions.
This work was supported by a Senior Scholar Award from the Ellison
Medical Foundation and grant CA77712 from the National Cancer
Institute, U.S. Department of Health and Human Services.
REFERENCES
1. Armant, M., M. Rubio, G. Delespesse, and M. Sarfati. 1995. Soluble CD23
directly activates monocytes to contribute to the antigen-independent stimulation of resting T cells. J. Immunol. 155:4868–4875.
2. Barsalou, L. S., G. J. Kantor, D. M. Deiss, and C. E. Hall. 1994. DNA repair
in the genomic region containing the beta-actin gene in xeroderma pigmentosum complementation group C and normal human cells. Mutat. Res.
315:43–54.
3. Bentley, D. L., and M. Groudine. 1986. A block to elongation is largely
responsible for decreased transcription of c-myc in differentiated HL60 cells.
Nature 321:702–706.
4. Berneburg, M., and A. R. Lehmann. 2001. Xeroderma pigmentosum and
related disorders: defects in DNA repair and transcription. Adv. Genet.
43:71–102.
5. Bielas, J. H. 2006. Non-transcribed strand repair revealed in quiescent cells.
Mutagenesis 21:49–53.
6. Bootsma, D., and J. H. Hoeijmakers. 1991. The genetic basis of xeroderma
pigmentosum. Ann. Genet. 34:143–150.
7. Bregman, D. B., R. Halaban, A. J. van Gool, K. A. Henning, E. C. Friedberg,
and S. L. Warren. 1996. UV-induced ubiquitination of RNA polymerase II:
a novel modification deficient in Cockayne syndrome cells. Proc. Natl. Acad.
Sci. USA 93:11586–11590.
8. Collins, S. J., F. W. Ruscetti, R. E. Gallagher, and R. C. Gallo. 1978.
Terminal differentiation of human promyelocytic leukemia cells induced by
dimethyl sulfoxide and other polar compounds. Proc. Natl. Acad. Sci. USA
75:2458–2462.
9. Donahue, B. A., S. Yin, J. S. Taylor, D. Reines, and P. C. Hanawalt. 1994.
Transcript cleavage by RNA polymerase II arrested by a cyclobutane pyrimidine dimer in the DNA template. Proc. Natl. Acad. Sci. USA 91:8502–8506.
10. Fisher, A. M., K. Danenberg, D. Banerjee, J. R. Bertino, P. Danenberg, and
C. J. Gomer. 1997. Increased photosensitivity in HL60 cells expressing wildtype p53. Photochem. Photobiol. 66:265–270.
11. Ford, J., and P. C. Hanawalt. 1997. Expression of wild-type p53 is required
for efficient global genomic nucleotide excision repair in UV-irradiated human fibroblasts. J. Biol. Chem. 272:28073–28080.
12. Hanawalt, P. C. 2002. Subpathways of nucleotide excision repair and their
regulation. Oncogene 21:8949–8956.
13. Hanawalt, P. C., J. M. Ford, and D. R. Lloyd. 2003. Functional characterization of global genomic DNA repair and its implications for cancer. Mutat.
Res. 544:107–114.
14. Hsu, P. S., P. C. Hanawalt, and T. Nouspikel. 4 August 2006, posting date.
Nucleotide excision repair phenotype of human acute myeloid leukemia cell
lines at various stages of differentiation. Mutat. Res. [Online.] doi: 10.1016/
j.mrfmmm.2006.06.008.
15. Hwang, B. J., J. M. Ford, P. C. Hanawalt, and G. Chu. 1999. Expression of
the p48 xeroderma pigmentosum gene is p53-dependent and is involved in
global genomic repair. Proc. Natl. Acad. Sci. USA 96:424–428.
16. Islas, A., and P. C. Hanawalt. 1995. DNA repair in the MYC and FMS
proto-oncogenes in ultraviolet light-irradiated human HL60 promyelocytic
cells during differentiation. Cancer Res. 55:336–341.
17. Lehmann, A. R. 2003. DNA repair-deficient diseases, xeroderma pigmentosum, Cockayne syndrome and trichothiodystrophy. Biochimie 85:1101–1111.
18. Lindahl, T. 1993. Instability and decay of the primary structure of DNA.
Nature 362:709–715.
19. Lindahl, T., and R. D. Wood. 1999. Quality control by DNA repair. Science
286:1897–1905.
20. Mori, T., M. Nakane, T. Hattori, T. Matsunaga, M. Ihara, and O. Nikaido.
1991. Simultaneous establishment of monoclonal antibodies specific for
either cyclobutane pyrimidine dimer or (6-4) photoproduct from the same
mouse immunized with ultraviolet-irradiated DNA. Photochem. Photobiol.
54:225–232.
21. Nouspikel, T.22 August 2006, posting date. DNA repair in differentiated
cells: some new answers to old questions. Neuroscience [Online.] doi:
10.1016/j.neuroscience.2006.07.006.
22. Nouspikel, T. 1996. Study of the role of XPG in xeroderma pigmentosum
and Cockayne’s syndrome. Ph.D. dissertation. University of Geneva, Geneva, Switzerland.
23. Nouspikel, T., and S. G. Clarkson. 1994. Mutations that disable the DNA
repair gene XPG in a xeroderma pigmentosum group G patient. Hum. Mol.
Genet. 3:963–967.
24. Nouspikel, T., and P. C. Hanawalt. 2001. DNA repair in terminally differentiated cells. DNA Repair 1:59–75.
25. Nouspikel, T., and P. C. Hanawalt. 2000. Terminally differentiated human
neurons repair transcribed genes but display attenuated global DNA repair
and modulation of repair gene expression. Mol. Cell. Biol. 20:1562–1570.
Downloaded from mcb.asm.org at RUTGERS UNIVERSITY on April 1, 2007
DAR. Since such cells are not available, we resorted to stably
transfecting THP1 cells with siRNA designed to inhibit production of XPC and subsequently differentiating transfected
cells into macrophages. These XPC-deficient macrophages displayed the same phenotype as XP-C fibroblasts or XP-C lymphoblasts, with little or no repair of the NTS, thereby confirming that DAR does require XPC. Conversely, when we used
siRNA to knock down the mRNA level of CSB, an enzyme
required for TCR but dispensable for GGR, we did not observe any effect on DAR but did observe the expected disappearance of the TCR-related strand bias. Therefore, DAR
cannot be considered a subset of TCR, but rather it is a subset
of GGR.
We have recently demonstrated that the GGR deficiency in
THP1 cells can be complemented by the ubiquitin-activating
enzyme E1 (27a). This suggested to us that an NER enzyme
might be activated by ubiquitination, and we have evidence
that it is the TFIIH complex. In particular, purified rat TFIIH
was able to complement extracts from differentiated THP1
cells in an in vitro NER assay. Our current model is that
ubiquitination of TFIIH is required for GGR but dispensable
for TCR and for transcription. Such a differential regulation
would not be surprising, as there are mutations in the XPB or
XPD subunits of TFIIH known to affect only TCR, with no
effect on GGR or transcription (4). Clearly, the various functions of TFIIH can be regulated independently from each
other.
Since TFIIH is a general transcription factor, required for
RNAPII to enter the elongation mode, it would not be surprising if it were concentrated inside “transcription factories”
within the nucleus. Thus, if even a small fraction of TFIIH
remained ubiquitinated in terminally differentiated cells, which
we did observe (data not shown), it would be available preferentially to genes entering transcription factories, thereby resulting in the preferential repair of both strands in active genes,
which we have called DAR.
Whether this phenomenon occurs in all differentiated cells,
and only in differentiated cells, is still an open question. Although DAR was not observed in terminally differentiated
cardiac myocytes from rats (38), it is quite possible that DAR
is present in many cell types but can be visualized only in cells
that have little or no GGR. Indeed, we (14) and others (5) have
reported evidence for DAR in cells that were not terminally
differentiated but in which global repair of CPDs was low
enough that DAR became obvious in active genes. It has been
proposed that DAR should thus be renamed “nontranscribed
strand repair,” given that it is not unique to differentiated cells
(5). But this alternative nomenclature would not be appropriate, since we demonstrate here that DAR operates on both
strands. We suggest instead that it be called “domain-associated repair,” which has the advantage of maintaining the same
acronym, while also more appropriately reflecting the mechanistic details of DAR: preferential repair within transcribed
DNA domains that have privileged access to NER enzymes
within transcription factories.
8729
8730
NOUSPIKEL ET AL.
33. Spivak, G., T. Itoh, T. Matsunaga, O. Nikaido, P. Hanawalt, and M.
Yamaizumi. 2002. Ultraviolet-sensitive syndrome cells are defective in transcription-coupled repair of cyclobutane pyrimidine dimers. DNA Repair
1:629–643.
34. Strobl, L. J., and D. Eick. 1992. Hold back of RNA polymerase II at the
transcription start site mediates down-regulation of c-myc in vivo. EMBO J.
11:3307–3314.
35. Tang, J. Y., B. J. Hwang, J. M. Ford, P. C. Hanawalt, and G. Chu. 2000.
Xeroderma pigmentosum p48 gene enhances global genomic repair and
suppresses UV-induced mutagenesis. Mol. Cell 5:737–744.
36. Thorel, F., A. Constantinou, I. Dunand-Sauthier, T. Nouspikel, P. Lalle, A.
Raams, N. G. Jaspers, W. Vermeulen, M. K. Shivji, R. D. Wood, and S. G.
Clarkson. 2004. Definition of a short region of XPG necessary for TFIIH
interaction and stable recruitment to sites of UV damage. Mol. Cell. Biol.
24:10670–10680.
37. Tsuchiya, S., Y. Kobayashi, Y. Goto, H. Okumura, S. Nakae, T. Konno, and
K. Tada. 1982. Induction of maturation in cultured human monocytic leukemia cells by a phorbol diester. Cancer Res. 42:1530–1536.
38. van der Wees, C. G., M. P. Vreeswijk, M. Persoon, A. van der Laarse, A. A.
van Zeeland, and L. H. Mullenders. 2003. Deficient global genome repair of
UV-induced cyclobutane pyrimidine dimers in terminally differentiated myocytes and proliferating fibroblasts from the rat heart. DNA Repair 2:1297–
1308.
39. Venitt, S. 1996. Mechanisms of spontaneous human cancers. Environ. Health
Perspect. 104(Suppl. 3):633–637.
Downloaded from mcb.asm.org at RUTGERS UNIVERSITY on April 1, 2007
26. Nouspikel, T., and P. C. Hanawalt. 2003. When parsimony backfires: neglecting DNA repair may doom neurons in Alzheimer’s disease. Bioessays
25:168–173.
27. Nouspikel, T., P. Lalle, S. A. Leadon, P. K. Cooper, and S. G. Clarkson. 1997.
A common mutational pattern in Cockayne syndrome patients from xeroderma pigmentosum group G: implications for a second XPG function. Proc.
Natl. Acad. Sci. USA 94:3116–3121.
27a.Nouspikel, T., and P. C. Hanawalt. Impaired nucleotide excision repair upon
macrophage differentiation is corrected by E1 ubiquitin-activating enzyme.
Proc. Natl. Acad. Sci. USA, in press.
28. Olivier, M., R. Eeles, M. Hollstein, M. A. Khan, C. C. Harris, and P.
Hainaut. 2002. The IARC TP53 database: new online mutation analysis and
recommendations to users. Hum. Mutat. 19:607–614.
29. Prakash, S., R. E. Johnson, and L. Prakash. 2005. Eukaryotic translesion
synthesis DNA polymerases: specificity of structure and function. Annu. Rev.
Biochem. 74:317–353.
30. Rovera, G., D. Santoli, and C. Damsky. 1979. Human promyelocytic leukemia cells in culture differentiate into macrophage-like cells when treated with
a phorbol diester. Proc. Natl. Acad. Sci. USA 76:2779–2783.
31. Shanower, G. A., and G. J. Kantor. 1997. A difference in the pattern of repair
in a large genomic region in UV-irradiated normal human and Cockayne
syndrome cells. Mutat. Res. 385:127–137.
32. Spivak, G., and P. Hanawalt. 1995. Determination of damage and repair in
specific DNA sequences. Methods Companion Methods Enzymol. 7:147–
161.
MOL. CELL. BIOL.