 |
INTRODUCTION |
Three genes encoding DNA ligases, LIG1,
LIG3, and LIG4, have been identified in the
mammalian genome (1). The products of these genes appear to be
responsible for the DNA-joining events that complete DNA replication,
genetic recombination, and DNA repair in mammalian cells. In the lower
eukaryote Saccharomyces cerevisiae, the DNA ligases encoded
by the CDC9 and DNL4 genes are functional
homologs of the LIG1 and LIG4 genes. However,
this organism apparently lacks a LIG3 gene homolog
(2-5).
Mammalian DNA ligases consist of a conserved catalytic domain flanked
by unrelated sequences (1). It has been suggested that protein-protein
interactions mediated by the unique regions flanking the catalytic
domain determine the cellular functions of these enzymes. In support of
this hypothesis, recent studies have identified specific protein
partners for each of the LIG gene products. The
non-catalytic N-terminal domain of DNA ligase I interacts with both
proliferating cell nuclear antigen and DNA polymerase 
, implicating
this species of DNA ligase in DNA replication and excision repair
pathways (6, 7). In contrast, the non-catalytic C-terminal extensions
of DNA ligases III
and IV contain BRCT motifs (8), a putative
protein-protein interaction domain first identified in the protein
encoded by the breast cancer susceptibility gene BRCA1 (9,
10). The single BRCT motif of DNA ligase III mediates complex
formation with the DNA repair protein XRCC1, linking this species of
DNA ligase with base excision repair and the repair of DNA
single-strand breaks (11, 12). Similarly, the region of DNA
ligase IV containing tandem BRCT motifs is involved in complex
formation with the DNA repair protein XRCC4,implicating DNA
ligase IV in non-homologous end joining
(NHEJ)1 and V(D)J
recombination (13, 14).
Human individuals with mutations in either the LIG1 (15) or
LIG4 gene (16) have been identified. The initial symptoms of
the individual with DNA ligase I deficiency were recurrent infections
caused by severe combined immunodeficiency (17). Cell lines established
from this patient exhibited defects in Okazaki fragment joining and
sensitivity to a wide range of DNA-damaging agents, a phenotype
consistent with the predicted role of DNA ligase I in DNA replication
and excision repair (18, 19). When the DNA ligase I-deficient cells
were activated for V(D)J recombination by ectopic expression of the
RAG proteins, there was no apparent defect in this type of
recombination (20, 21). However, studies with cell-free extracts
suggest that DNA ligase I contributes to the completion of V(D)J
recombination (22).
The individual with a mutated LIG4 gene presented with
leukemia. Unfortunately, attempts to treat this disease with
chemotherapy and then ionizing radiation caused a severe reaction (16,
23). The extreme radiosensitivity of cell lines established from this individual is consistent with the predicted role of DNA ligase IV in
the repair of DNA double-strand breaks by NHEJ. Surprisingly, the
DNA ligase IV-deficient individual did not appear to be
immunodeficient, and the DNA ligase IV-deficient cell line had no
apparent defect in V(D)J recombination (16, 23). In contrast, gene
targeting studies have provided compelling evidence linking DNA ligase
IV with V(D)J recombination. Inactivation of the LIG4 gene
either in human pre-B cells or in mouse embryonic fibroblasts abolishes V(D)J recombination (24, 25).
The processes of NHEJ and V(D)J recombination involve the joining of
DNA double-strand breaks (DSBs), whereas the ligation events that
complete DNA replication and excision repair pathways occur at
single-strand interruptions in duplex DNA. Thus, it is conceivable that
the DNA ligases may exhibit preferences for either double- or
single-strand breaks depending on the type of DNA transaction that they
participate in. In this regard, it was demonstrated that bovine DNA
ligases I and II could be discriminated by the ability of DNA ligase I
to join blunt-ended DNA duplexes (26). This observation provided the
rationale for studies examining the influence of DNA ligase I on the
spectrum of products produced by in vitro V(D)J
recombination reactions (22) and the functional interaction between DNA
ligase I and the DSB-binding factor Ku during the ligation of DSBs
in vitro (27). However, the initial biochemical studies on
blunt-end joining were performed prior to the identification of DNA
ligases III and IV (8). Here we have purified the human DNA ligase
IV-XRCC4 complex from baculovirus-infected insect cells and found that
the DNA ligase IV-XRCC4 complex binds specifically to the ends of
duplex DNA molecules, a property shared with another NHEJ factor, the
DNA-dependent protein kinase (DNA-PK) (28). Surprisingly,
these proteins do not compete for binding to DNA, but instead DNA
ligase IV-XRCC4 associates with both Ku and the catalytic subunit
(DNA-PKcs) of DNA-PK at the ends of DNA molecules. However, Ku and
DNA-PKcs have markedly different effects on DNA joining by DNA ligase
IV-XRCC4.
| |
EXPERIMENTAL PROCEDURES |
Recombinant Baculoviruses--
The 2.8-kb
BamHI/EcoRI human DNA ligase IV open reading
frame with an additional 3'-sequence encoding hemagglutinin (HA) and oligohistidine tags (a gift from Dr. Michael Lieber) was subcloned into
a pFastBac vector to generate a recombinant baculovirus expressing full-length tagged DNA ligase IV using the pFastBac baculovirus expression vector system (Life Technologies, Inc.). The full-length XRCC4 open reading frame was amplified from a human testis
cDNA library (CLONTECH) by the polymerase chain
reaction. After verifying the DNA sequence of the amplified product, it
was subcloned into a pFastBac vector to generate a recombinant
baculovirus expressing full-length XRCC4. To coexpress tagged DNA
ligase IV and untagged XRCC4, both cDNAs were subcloned into the
pFastBacDual vector.
Antibodies--
Yeast Dnl4, expressed as a glutathione
S-transferase fusion protein, was purified from
Escherichia coli. Rabbit antiserum raised against
glutathione S-transferase-Dnl4 cross-reacts with human DNA
ligase IV. Ku (p80) Ab-2 and DNA-PKcs Ab-3 were purchased from Neo
Markers (Fremont, CA). The monoclonal antibody 12CA5, which recognizes
the HA epitope, was a gift from Dr. Tom Boyer.
Purification of the DNA Ligase IV-XRCC4 Complex from
Baculovirus-infected Sf9 Cells--
At 60 h
post-infection, infected Sf9 insect cells (600 ml) were
harvested by centrifugation, flash-frozen, and stored at 
80 °C.
Frozen cells were thawed on ice and resuspended in 50 ml of lysis
buffer (50 mM Tris-HCl (pH 7.5), 300 mM NaCl,
10% glycerol, 10 mM 2-mercaptoethanol, 1 mM
phenylmethanesulfonyl fluoride, 1 mM benzamidine HCl, 1 µg/ml leupeptin, 2 µg/ml aprotinin, and 1 µg/ml pepstatin). After
a brief 30-s sonication to lyse the cells, the lysate was cleared by
centrifugation at 40,000 rpm for 30 min in Beckman Ti-45 rotor at
4 °C. The cleared lysate was then incubated with nickel beads (1 ml
of beads/10 mg of protein) in the presence of 10 mM
imidazole for 3 h at 4 °C. The beads were poured into a column
and washed extensively with lysis buffer containing 40 mM
imidazole. The DNA ligase IV-XRCC4 complex was eluted with lysis buffer
containing 200 mM imidazole and then further purified by
Resource Q fast protein liquid chromatography, by gel filtration
chromatography using a Superdex 200 column, and finally by Mono S fast
protein liquid chromatography (Amersham Pharmacia Biotech). Peak
fractions from the Mono S column were aliquoted, flash-frozen, and
stored at 80 °C. Approximately 0.2 mg of DNA ligase IV-XRCC4
complex was obtained from 109 infected insect cells.
Because the stoichiometry of the subunits within the DNA ligase
IV-XRCC4 complex is not known, the amount of complex was estimated by
comparing the Coomassie Blue staining of the DNA ligase IV polypeptide
with that of a known amount of -galactosidase after
SDS-polyacrylamide gel electrophoresis.
Purification of the Ku Heterodimer from Baculovirus-infected
Sf9 Cells--
Baculoviruses encoding human Ku86 and Ku70 were
obtained from Dr. Cathy Meek. Insect Sf9 cells (300 ml) were
co-infected with the Ku baculoviruses and then incubated for 65 h.
The purification of Ku was monitored by immunoblotting. After
collection and lysis of the cells as described above, proteins
precipitated by 30-70% ammonium sulfate were resuspended in and
dialyzed against 50 mM Tris-HCl (pH 7.5), 50 mM
NaCl, 10% glycerol, 0.5 mM dithiothreitol, 1 mM EDTA, 1 mM phenylmethanesulfonyl fluoride,
and 1 mM benzamidine HCl. The Ku heterodimer was then
purified by stepwise elution from a 10-ml phosphocellulose P-11 column
(Whatman; Ku eluted at 200 mM NaCl) and then from a 2-ml
Bio-Gel HTP hydroxylapatite column (Bio-Rad; Ku eluted at 200 mM potassium phosphate (pH 7.5)). Further purification of
the Ku heterodimer was achieved by gradient elution from a 1-ml
Resource Q column (Amersham Pharmacia Biotech; Ku eluted at 300 mM NaCl), by molecular sieving through a 120-ml Superdex
200 column (Amersham Pharmacia Biotech) at 150 mM NaCl, and
finally by gradient elution from a 1-ml Resource S column (Amersham
Pharmacia Biotech; Ku eluted at 220 mM NaCl). Peak
fractions from the Mono S column were aliquoted, flash-frozen, and
stored at 80 °C. Approximately 0.25 mg of Ku heterodimer was
obtained from 5 × 108 infected Sf9 cells.
Purification of DNA-PKcs from Nuclear Extracts of Raji
Cells--
Raji cells were purchased from the National Cell Culture
Center in Minneapolis. A clarified nuclear extract was prepared from a
100-liter culture as described (29, 30). The purification of DNA-PKcs
was monitored by immunoblotting. Ammonium sulfate (0.21 g/ml) was added
to the extract, and precipitated proteins were removed by
centrifugation. After the addition of ammonium sulfate (0.08 g/ml) to
the supernatant, precipitated DNA-PKcs was collected by centrifugation,
dialyzed, and then purified by conventional column chromatography using
Q-Sepharose and Sephacryl S300 (Amersham Pharmacia Biotech) and
hydroxylapatite (Bio-Rad) resins. Finally, DNA-PKcs was purified on
Mini Q and Mini S columns using the Smart System (Amersham Pharmacia
Biotech). Nearly homogenous DNA-PKcs was concentrated and stored at
80 °C in small aliquots.
Adenylation Assay--
Fractions (2 µl) were incubated for 15 min at 25 °C in 10-µl reaction mixtures containing 60 mM Tris-HCl (pH 8.0), 10 mM MgCl2, 5 mM dithiothreitol, 50 µg/ml bovine serum albumin, and
0.5 µCi of [-32P]ATP (3000 Ci/mmol; Amersham
Pharmacia Biotech). Reactions were stopped by the addition of SDS
sample buffer and analyzed by SDS-polyacrylamide gel electrophoresis.
Labeled DNA ligase-adenylate complexes were detected by either
autoradiography or Phosphor- Imager (Molecular Dynamics, Inc.,
Sunnyvale, CA) analysis.
DNA Substrates--
The oligo(dT)/poly(rA) substrate and the
oligonucleotide substrate containing a single defined nick were
prepared as described previously (31). Partial duplex oligonucleotide
substrates with either blunt ends or sticky ends were prepared
essentially as described by Grawunder et al. (14). End
labeling was achieved using T4 polynucleotide kinase (New England
Biolabs Inc.) and [
-32P]ATP (Amersham Pharmacia Biotech).
To detect joining of linear duplex DNA fragments, 2.9-kb and 400-bp
linear DNA duplexes with 5'-complementary single-strand ends 4 nucleotides in length were generated by digestion with EcoRI. Similarly, a 2.9-kb linear DNA duplex with
3'-complementary single-strand ends 4 nucleotides in length was
generated by PstI digestion. After separation by agarose gel
electrophoresis, the linear DNA fragments were purified using a
QIAquick gel extraction kit (QIAGEN Inc., Valencia, CA)
according to the manufacturer's instructions. DNA concentrations are
expressed as DNA molecules.
Ligation Assay--
DNA ligase and a labeled DNA substrate were
incubated in reaction mixtures (60 µl) containing 60 mM
Tris-HCl (pH 8.0), 10 mM MgCl2, 5 mM dithiothreitol, 1 mM ATP, and 50 µg/ml
bovine serum albumin at 25 °C. Reactions were terminated by the
addition of stop solution (40 mM EDTA and 80% formamide).
After separation by denaturing gel electrophoresis, labeled
oligonucleotides were quantitated in the dried gel by PhosphorImager
analysis. One unit of DNA ligase catalyzes the formation of 1 nmol of
phosphodiester bonds in 15 min at 25 °C.
Ligation reactions with DNA restriction fragments were deproteinized by
phenol/chloroform extraction prior to agarose gel electrophoresis.
After staining with ethidium bromide, substrates and ligated products
were visualized using an IS 1000 imaging system (Alpha Innotech
Corp.).
Electrophoretic Mobility Shift Assay (EMSA)--
A 400-bp
restriction fragment with 5'-complementary 4-nucleotide overhangs was
end-labeled with T4 polynucleotide kinase and [-32P]ATP (3000 Ci/mmol). Labeled DNA (50 fmol) and
protein were incubated in EMSA buffer (20 mM Tris-HCl (pH
7.5), 1 mM dithiothreitol, 0.1% Triton X-100, 150 mM KCl, 60 µg/ml bovine serum albumin, and 5% glycerol)
for 20 min at 25 °C. Samples were loaded directly onto a 0.7%
agarose gel, and electrophoresis was carried out at 120 V for 3 h
at 4 °C in 20 mM Tris acetate (pH 7.5) and 1 mM EDTA. Labeled DNA fragments were detected in the dried
gel by autoradiography.
Atomic Force Microscopy--
Atomic force microscopy (AFM) was
performed using a Nanoscope Scanning Probe system (Digital Instruments
Inc., Santa Barbara, CA) with an E-type vertical engage piezoelectric
scanner operating in tapping mode. Nanoprobe SPM probes (length = 125 µm; resonance frequency between 278 and 359 kHz) were used
throughout the study. Images were captured with scan frequencies
between 1.02 and 1.97 Hz.
The DNA substrate was the 400-bp fragment described above. DNA binding
was carried out in EMSA buffer without bovine serum albumin and
glycerol at 25 °C for 5 min. Samples were then diluted 1:50 in
buffer containing 20 mM HEPES (pH 7.9) and 10 mM MgCl2 before depositing on freshly cleaved
mica. After ~30 s, samples were washed off with filter-sterilized,
distilled water and blow-dried with a stream of filtered air.
In preliminary experiments with purified DNA ligase IV-XRCC4, DNA-PKcs,
and Ku, distinct particulate species of uniform size were detected.
When each of the proteins was incubated with DNA, DNA-protein complexes
of uniform size were also observed. Initially, the height of the
complexes formed by each of the proteins was measured. The average
height of 10 DNA-protein complexes formed by DNA ligase IV-XRCC4 shown
in Fig. 3A was 1.25 ± 0.21 nm. In similar experiments,
the average heights of DNA-protein complexes formed by Ku and DNA-PKcs
were 2.02 ± 0.16 and 3.34 ± 0.33 nm, respectively. When DNA
ligase IV-XRCC4 was co-incubated with either Ku or DNA-PKcs,
significantly larger DNA-protein complexes were observed. The average
height of the larger complexes formed by DNA ligase IV-XRCC4 and Ku
shown in Fig. 6B was 2.9 ± 0.36 nm. In three
independent experiments (111 DNA-protein complexes), the larger
DNA-protein complexes constituted 15% of the total DNA-protein
complexes. The average height of the larger complexes formed by DNA
ligase IV-XRCC4 and DNA-PKcs shown in Fig. 7B was 5.62 ± 1.1 nm. In three independent experiments (67 DNA-protein complexes),
the larger DNA-protein complexes constituted 8% of the total
DNA-protein complexes. Viewed from above, all the DNA-protein complexes
had an oval shape. The average lengths and widths of the ovals were as
follows; DNA ligase IV-XRCC4, 19.6 × 16.6 nm; Ku, 23.3 × 19.0 nm; DNA-PKcs, 31.8 × 28.6 nm; DNA ligase IV-XRCC4 and Ku,
25.3 × 23.6 nm; and DNA ligase IV-XRCC4 and DNA-PKcs, 40.9 × 32.2 nm.
| |
RESULTS |
Purification of the Recombinant Human DNA Ligase IV-XRCC4
Complex--
Previous studies have shown that DNA ligase IV and XRCC4
copurify as a stable complex and that coexpression of XRCC4 stimulates DNA joining by DNA ligase IV (13, 14). Moreover, the levels of DNA
ligase IV protein are severely reduced in the XRCC4
mutant cell line, XR-1 (32). Together, these results indicate that at
least one important function of XRCC4 is to stabilize DNA ligase IV. In
agreement with this idea, human DNA ligase IV protein was undetectable
after infection of insect Sf9 cells with a recombinant baculovirus encoding DNA ligase IV, whereas coexpression of DNA ligase
IV and XRCC4 in insect cells resulted in readily detectable levels of
DNA ligase IV protein (data not shown). The DNA ligase IV-XRCC4 complex
was purified from insect cells infected with a baculovirus encoding
oligo-His-tagged DNA ligase IV and untagged XRCC4. After an initial
metal-chelating affinity chromatography step, the DNA ligase IV-XRCC4
complex was purified to >95% homogeneity by gel filtration and
ion-exchange chromatography (Fig.
1A).
View larger version (55K):
[in this window]
[in a new window]
|
Fig. 1.
Purification of the human DNA ligase IV-XRCC4
complex. The human DNA ligase IV-XRCC4 complex was purified from
baculovirus-infected insect cells as described under "Experimental
Procedures." A, proteins in the peak fractions eluting
from the final Mono S fast protein liquid chromatography column were
separated by SDS-polyacrylamide gel electrophoresis and then stained
with Coomassie Blue. B, the DNA ligase IV-XRCC4 complex
(fraction 8; 0.5 pmol) was incubated with [-32P]ATP as
described under "Experimental Procedures." After separation by
SDS-polyacrylamide gel electrophoresis, labeled polypeptides were
detected by autoradiography. Molecular mass standards (in kilodaltons)
are shown on the left. DNA ligase IV and XRCC4 are indicated on the
right. C, DNA ligase III and the DNA ligase IV-XRCC4
complex (0.2 pmol of each) were incubated with the oligo(dT)/poly(rA)
substrate as described under "Experimental Procedures." Aliquots
were taken at the indicated times and separated by denaturing gel
electrophoresis. The positions of the oligo(dT) substrate (25-mer) and
the ligated product (50-mer) are indicated on the left.
|
|
The purification of the DNA ligase IV-XRCC4 complex was monitored by
assaying for formation of the 110-kDa labeled enzyme-adenylate intermediate (Fig. 1B). In the initial purification of DNA
ligase IV from HeLa nuclei, it was noted that a significant proportion of DNA ligase IV, unlike the other DNA ligases, remained adenylated during purification (33). By measuring DNA joining in the absence of
ATP, we estimate that ~50% of the DNA ligase IV molecules in the
most highly purified fraction are adenylated. As expected, the DNA
ligase IV-XRCC4 complex joined oligo(dT) molecules that were annealed
to a poly(rA) template (Fig. 1C). In assays with this
substrate, the rate of phosphodiester bond formation catalyzed by the
DNA ligase IV-XRCC4 complex (0.04 ligations/enzyme/min) was
indistinguishable from that catalyzed by DNA ligase III (Fig. 1C), with both enzymes having a specific activity of 5 units/mg. In addition, the recombinant DNA ligase IV-XRCC4 complex was
active with the oligo(dT)/poly(dA) substrate, but not with the
oligo(rA)/poly(dT) substrate (data not shown). Thus, the reactivity of
the recombinant DNA ligase IV-XRCC4 complex with the homopolymer
substrates is identical to that of DNA ligase IV purified from HeLa
cell nuclei (33).
Ligation of Single-strand Nicks in Duplex DNA by Mammalian DNA
Ligases--
Although the homopolymer substrates were invaluable in
the identification of the multiple species of DNA ligases in mammalian cells (26, 33, 34), it is unlikely that the differential activities of
DNA ligases I, III, and IV with these substrates have any biological
significance. In contrast, the relative ability to join single-strand
nicks in duplex DNA and the double-strand ends of duplex DNA molecules
may reflect the type of cellular DNA transaction that the DNA ligase
participates in. Therefore, we have compared the rates of joining of a
single defined nick in a linear duplex by DNA ligases I and III and
DNA ligase IV-XRCC4 (Fig. 2A),
each of which was purified from baculovirus-infected insect cells (35,
36). In agreement with a previous study (35), the rates of nick
ligation by DNA ligases I and III were not significantly different.
Under these reaction conditions, DNA ligases I and III catalyzed
0.54 and 0.44 ligations/enzyme/min, respectively. In contrast, the rate
of nick ligation by the DNA ligase IV-XRCC4 complex was significantly
slower (Fig. 2A). Under these reaction conditions, the
turnover number of the DNA ligase IV-XRCC4 complex was 0.25 ligations/enzyme/min. To determine whether the lower joining activity
of DNA ligase IV-XRCC4 is due to its expression and purification in a
heterologous system, similar joining assays were carried out with the
functionally homologous yeast enzyme Dnl4-Lif1 that was purified after
overexpression in S. cerevisiae. We again observed that the
rate of ligation by the Dnl4-Lif1 complex was also ~2-fold lower than
the rate of ligation by Cdc9 DNA ligase, the functional homolog of
human DNA ligase I (data not shown). This indicates that the lower rate of nick joining by DNA ligase IV-XRCC4 is intrinsic to this enzyme.
View larger version (31K):
[in this window]
[in a new window]
|
Fig. 2.
Ligation of nicks and DNA double-strand
breaks by mammalian DNA ligases. A, ligation of a
labeled DNA substrate containing a single defined nick measured as
described under "Experimental Procedures." B, ligation
of partial duplex oligonucleotide substrates with complementary ends
was measured as described under "Experimental Procedures."
Reactions contained 4 pmol of substrate and 0.2 pmol of each DNA
ligase.  , DNA ligase I;  , DNA ligase III;  , DNA ligase
IV-XRCC4. C, ligation of DNA molecules with complementary
ends. A 2.9-kb DNA restriction fragment (0.1 µg) with either 5'- or
3'-complementary single-strand ends 4 nucleotides in length and 0.2 pmol of the indicated DNA ligase were incubated at 25 °C for 2 h. D, ligation of DNA molecules with blunt ends. A 2.9-kb
DNA restriction fragment with blunt ends (0.1 µg) and 0.2 pmol of the
indicated DNA ligase were incubated at 16 °C overnight.
Con. (control), substrate only; T4, T4 DNA ligase
(1 unit; Roche Molecular Biochemicals); I, DNA ligase I;
III, DNA ligase III; IV/4, DNA ligase
IV-XRCC4 complex. After agarose gel electrophoresis, DNA was stained
with ethidium bromide and then visualized using the IS 1000 imaging
system.
|
|
Ligation of Double-strand Breaks in Duplex DNA by Mammalian DNA
Ligases--
Genetic studies in yeast and mammals have indicated that
DNA ligase IV functions in the repair of DSBs by completing NHEJ reactions (3-5, 24, 25). Although there appear to be multiple mechanisms involved in end-to-end joining, many of these events involve
short homologous sequences (microhomologies, usually 2-6 nucleotides
in length) at the ligation site, suggesting that the DNA ligase joins
DNA molecules held together by short complementary terminal sequences
(37). To mimic this type of structure, we constructed oligonucleotide
substrates in which 2 oligonucleotides are annealed to the opposite
ends of an 80-mer, generating partial duplex regions with complementary
single-strand ends 3 nucleotides in length (14). In assays with this
substrate, the rates of DNA joining catalyzed by DNA ligases I and
III and DNA ligase IV-XRCC4 were not significantly different (Fig.
2B). We estimate that the turnover number for this reaction
was ~0.04 ligations/enzyme/min. Thus, the ligation of complementary
duplex ends is markedly slower than the ligation of nicks in duplex DNA.
Using the same type of DNA substrate, we examined the influence of
complementary overhang length on DSB joining. Under the same reaction
conditions as in Fig. 2B, no ligation of DNA molecules with
blunt ends was detected in assays with the human DNA ligases, but DNA
ligase III did generate significant amounts of ligated product after
overnight incubation. In assays with the DNA ligase IV-XRCC4 complex,
we observed an increase in the amount of ligated product as the length
of overhang increased from 1 to 6 nucleotides, whereas a 9-nucleotide
overhang resulted in a sharp decrease in ligation (data not shown).
Next we examined the joining of a 2.9-kb linear duplex with either
complementary or blunt ends since this structure is likely to more
closely resemble the in vivo substrate compared with the short partial duplex oligonucleotide. After a 2-h incubation at 25 °C, DNA ligase III and DNA ligase IV-XRCC4 produced similar amounts of ligated product in assays with sticky-ended restriction fragments (Fig. 2C). In contrast, DNA ligase I did not
generate detectable ligated product under the same reaction conditions. Since the concentration of complementary DNA ends was significantly lower in the assays with the restriction fragment compared with the
oligonucleotide substrate, it is possible that the relatively low
activity of DNA ligase I reflects the DNA end binding properties of
this enzyme.
The results of DNA joining assays with a blunt-ended restriction
fragment were qualitatively the same as those with the blunt-ended oligonucleotide substrate. No ligation was observed for any of the
three human enzymes after 2 h of incubation at 25 °C,
confirming that blunt-end ligation is less efficient than sticky-end
ligation. After overnight incubation at 16 °C, ligated product was
observed only with DNA ligase III (Fig. 2D), showing
that, under these in vitro conditions, DNA ligase III is
the most active enzyme for blunt-end ligation. Previous studies have
shown that DNA ligase I was capable of blunt-end joining in the
presence of polyethylene glycol (26). We reiterated this result and
found that the addition of polyethylene glycol also stimulated the
blunt-end substrate joining by DNA ligase III and DNA ligase
IV-XRCC4 to a similar extent (data not shown). This suggests that
molecular crowding by polyethylene glycol nonspecifically stimulates
the joining of blunt ends by human DNA ligases.
Visualization of Interactions between DNA and the DNA Ligase
IV-XRCC4 Complex by AFM--
The DNA substrate specificity of the
mammalian DNA ligases is likely to be a consequence of their DNA
binding properties. Based on the results shown above, we suspected that
DNA ligase IV-XRCC4 would bind preferentially to the ends of linear DNA
molecules with short single-strand overhangs. To examine this directly, we visualized DNA-protein complexes by AFM. A 400-bp DNA restriction fragment with complementary ends 4 nucleotides in length was absorbed onto mica. As expected, the majority of the molecules (>96%) behaved as unit length monomers (Fig.
3A, left panel)
with the remaining molecules appearing as dimers. When DNA ligase
IV-XRCC4 was preincubated with the DNA fragment (Fig. 3A,
right panel), a significant fraction of the DNA molecules
(~50%) behaved as oligomers consisting of two or more unit length
DNA molecules. Approximately 70% of the oligomers had internally bound
DNA ligase IV-XRCC4, whereas in DNA-protein complexes formed with unit
length DNA molecules, DNA ligase IV-XRCC4 bound exclusively to the ends
of the DNA molecules. Examples of unit length DNA molecules with a DNA
ligase IV-XRCC4 protein complex at either one (Fig. 3B,
upper left panel) or both (upper right panel)
ends are shown. In addition, circular DNA molecules, which appear to be
formed as a result of the DNA ligase IV-XRCC4 complex binding to both
ends of the same linear unit length DNA molecule, were observed (Fig.
3B, lower right panel). In similar experiments
with DNA ligase I, the majority of protein molecules did not associate
with the DNA molecules, and DNA oligomerization was not affected. When
DNA ligase IV-XRCC4 was preincubated with a blunt-ended DNA fragment,
the frequency of both DNA oligomers and DNA-protein complexes was
significantly reduced (data not shown).
View larger version (68K):
[in this window]
[in a new window]
|
Fig. 3.
Tapping mode AFM topographs of DNA-protein
complexes formed by DNA ligase IV-XRCC4. A: left
panel, DNA only; right panel, DNA incubated with the
DNA ligase IV-XRCC4 complex (1 × 1-µm scans). B:
upper left panel, DNA ligase IV-XRCC4 complex
bound to one end of unit length DNA; upper right panel, DNA
ligase IV-XRCC4 complexes bound to both ends of unit length DNA;
lower left panel, linking of three unit length DNA molecules
by DNA ligase IV-XRCC4 (350 × 350-nm images, zoomed in
from 2 × 2-µm scans); lower right panel, a circular
unit length DNA molecule with a bound DNA ligase IV-XRCC4 complex
(300 × 300-nm image, zoomed in from 2 × 2-µm scans). The
initial concentrations of DNA ligase IV-XRCC4 and DNA were 200 and 300 nM, respectively.
|
|
As mentioned above, longer DNA molecules, presumably formed by the
linking together of unit length molecules, were also detected (Fig.
3A, right panel). The frequency of these
oligomers increased as the amount of protein in the reactions was
increased. In the oligomers with internally bound DNA ligase IV-XRCC4
molecules, the positions of the protein complexes appeared to
correspond to the junction between unit length DNA molecules. An
example of such a structure is shown in Fig. 3B (lower
left panel). Because a significant proportion of the DNA ligase IV
molecules are adenylated, it is possible that at least some of the unit
length DNA molecules have been covalently linked by ligation.
Consistent with this idea, preincubation of the DNA substrate with calf
intestinal phosphatase significantly reduced the number of oligomeric
DNA molecules without internally bound protein. In contrast, oligomeric DNA molecules with protein bound at defined intervals, corresponding to
the length of the DNA monomer, were still readily detectable with
phosphatase-treated DNA (Fig. 3B, lower left
panel). Since these DNA molecules cannot be ligated, it seems
likely that DNA ligase IV-XRCC4 acts as a bridging factor, bringing
together the ends of two DNA molecules. The sizes of the
internal DNA-protein complexes on both the circular and oligomeric DNA
molecules do not differ from those at the ends of unit length DNA
molecules (Fig. 3B), suggesting that the two DNA ends are
held together by a single DNA ligase IV-XRCC4 complex rather than each
end having a DNA ligase IV-XRCC4 complex bound.
Effect of Ku on the Catalytic and DNA Binding Activities of the DNA
Ligase IV-XRCC4 Complex--
Since DNA ligase IV-XRCC4 and the Ku
heterodimer participate in the same NHEJ pathway in vivo
(3-5) and both protein complexes bind to DNA ends in vitro,
the influence of Ku on DNA joining catalyzed by the DNA ligase IV-XRCC4
complex was examined. The addition of Ku inhibited intramolecular
ligation of an oligonucleotide substrate with cohesive ends by DNA
ligase IV-XRCC4 under various reaction conditions, including different
salt concentrations (Fig. 4), different
temperatures (25 or 37 °C), and the presence or absence of
polyethylene glycol (data not shown). At both 50 and 100 mM
KCl and a 1:1 molar ratio of Ku to DNA ends, DNA joining was ~70%
inhibited. The inhibitory effect of the Ku fraction was abolished by
heat denaturation, indicating that it is mediated by Ku protein rather
than a buffer component (data not shown). It has been shown that the
DNA-dependent interaction of Ku and DNA-PKcs occurs only if
the linear DNA molecule is long enough to allow both proteins to bind
at the same time (38, 39). To determine whether the Ku-mediated
inhibition of DNA ligase IV-XRCC4 is due to this type of effect, we
carried out similar ligation assays using a 400-bp labeled linear DNA
fragment. Ku also inhibited joining of this longer substrate (see Fig.
8).
View larger version (20K):
[in this window]
[in a new window]
|
Fig. 4.
Effect of Ku on the joining of DNA
double-strand breaks by DNA ligase IV-XRCC4. A, the
intramolecular joining of partially duplex DNA molecules with
complementary single-strand ends 3 nucleotides in length was measured
as described under "Experimental Procedures." Ligation reactions,
which contained 4 pmol of labeled DNA substrate and 0.2 pmol of DNA
ligase IV-XRCC4 as indicated (+), were incubated at 25 °C for 30 min
at 0, 50, and 100 mM KCl. Ku (0.2 and 8 pmol) was added as
indicated. B, the DNA joining reactions shown in
A were quantitated by PhosphorImager analysis, and the
results are shown as a bar graph. At 0 and 50 mM
KCl, 0.5 pmol of the substrate was ligated, whereas at 100 mM KCl, 0.3 pmol of the substrate was ligated in the
absence of Ku. The effect of Ku on DNA joining is expressed as a
percentage of the ligation product produced in the absence of Ku.
Black bars, no added Ku; gray bars, 0.2 pmol of
Ku; white bars, 8 pmol of Ku.
|
|
Since the DNA end binding properties of Ku have been extensively
studied, we considered the possibility that Ku may inhibit DNA ligase
IV-XRCC4 by preventing access to the DNA ends. In an EMSA, a specific
complex formed between a 400-bp labeled restriction fragment with
complementary ends and DNA ligase IV-XRCC4 (Fig. 5, A, lane 3; and
B, lane 4). This complex was supershifted by an
anti-DNA ligase IV antibody (Fig. 5A, lane 2) and
by an antibody specific for the HA tag of DNA ligase IV (Fig.
5B, lane 3), but not by the anti-Ku antibody
(Fig. 5B, lane 2). Formation of the DNA-protein
complex was more effectively inhibited by the addition of increasing
amounts of unlabeled linear DNA compared with supercoiled DNA (data not
shown). As expected, Ku also formed a specific DNA-protein complex with
the labeled linear DNA substrate (Fig. 5, A, lane 5; and B, lane 6). This complex was
supershifted by the anti-Ku antibody (Fig. 5B, lane
7), but not by the antibodies that recognize DNA ligase IV (Fig.
5, A, lane 4; and B, lane
8). When Ku and DNA ligase IV-XRCC4 were co-incubated with the
labeled DNA substrate, a DNA-protein complex with an electrophoretic
mobility distinct from that of the complexes formed by either protein
alone was observed (Fig. 5, compare A, lanes
3-5; and B, lanes 4-6). The addition of
increasing amounts of Ku protein did not abolish the appearance of the
novel DNA-protein complex, strongly suggesting that Ku does not inhibit
the binding of DNA ligase IV-XRCC4 to DNA (data not shown). To
demonstrate that this novel DNA-protein complex contained both Ku and
DNA ligase IV, we performed supershift assays. As expected, the complex
was quantitatively supershifted by antibodies specific for either Ku
(Fig. 5, A, lane 7; and B, lane
10) or DNA ligase IV (Fig. 5, A, lane 8; and
B, lane 9).
View larger version (22K):
[in this window]
[in a new window]
|
Fig. 5.
DNA-protein complexes formed by Ku and DNA
ligase IV-XRCC4 in electrophoretic mobility shift assays. The DNA
ligase IV-XRCC4 complex (600 fmol) and Ku (350 fmol) were incubated as
indicated with a 400-bp end-labeled restriction fragment with
complementary single-strand ends 4 nucleotides in length (50 fmol) as
described under "Experimental Procedures." Antibodies to either DNA
ligase IV or Ku were added as indicated. After native agarose gel
electrophoresis, labeled oligonucleotides were detected by
autoradiography. A, EMSA using anti-DNA ligase IV and
anti-Ku86 antibodies; B, EMSA using anti-HA and anti-Ku86
antibodies as indicated.
|
|
Visualization of DNA-Protein Complexes Formed by Ku and DNA Ligase
IV-XRCC4--
Our EMSA studies demonstrate that Ku and DNA ligase
IV-XRCC4 can bind to the same linear DNA molecule, but do not provide insights into the relationship between the protein complexes on the DNA
molecule. To address this issue, we have used AFM to directly visualize
the DNA-protein complexes. In agreement with published studies (28), Ku
bound to the ends of the 400-bp DNA restriction fragment and also
translocated internally. An example of a unit length DNA molecule with
two internally translocated Ku complexes is shown in Fig.
6A. The height of the
DNA-bound Ku complexes was 2.02 ± 0.16 nm, whereas the DNA-bound
DNA ligase IV-XRCC4 complexes had a height of 1.25 ± 0.21 nm
(Fig. 3A, right panel). When Ku and DNA ligase
IV-XRCC4 were co-incubated with the 400-bp DNA fragment, ~15% of the
DNA-protein complexes were clearly larger than the complexes formed by
either protein alone (Fig. 6B). These complexes, an example
of which is indicated by arrowheads c in Fig. 6B,
had an average height of 2.91 ± 0.36 nm. DNA-protein complexes
similar in size to those observed in reactions with either DNA ligase
IV-XRCC4 (Fig. 3A, right panel) or Ku (Fig. 6A) alone are indicated by arrowheads a and
b, respectively, in Fig. 6B. These results
demonstrate that Ku and DNA ligase IV-XRCC4 associate in a DNA
end-dependent manner.
View larger version (112K):
[in this window]
[in a new window]
|
Fig. 6.
Tapping mode AFM topographs of DNA-protein
complexes formed by DNA ligase IV-XRCC4 and Ku. A, DNA
incubated with Ku. A unit length DNA molecule with two internally
translocated Ku complexes is shown. B, DNA ligase IV-XRCC4
and Ku co-incubated with DNA (500 × 500-nm images, zoomed in from
2 × 2-µm scans). Arrowhead a, DNA ligase
IV-XRCC4 bound to DNA (average height of 1.26 nm); arrowhead
b, Ku bound to DNA (average height of 2.02 nm); arrowheads
c, complex of Ku and DNA ligase IV-XRCC4 bound to DNA (average
height of 2.91 nm). The initial concentrations of DNA ligase IV-XRCC4
and DNA were the same as in Fig. 3. The initial concentration of Ku was
300 nM.
|
|
Visualization of DNA-Protein Complexes Formed by DNA-PKcs and DNA
Ligase IV-XRCC4--
Previous studies have shown that the 470-kDa
catalytic subunit of the DNA-dependent protein kinase
interacts with DNA double-strand ends independently of Ku and has
binding sites for both single- and double-strand DNAs (28, 40). In
agreement with these studies, we observed large DNA-protein complexes
with a height of 3.34 ± 0.33 nm at the end of DNA molecules when
DNA-PKcs was incubated with the 400-bp DNA restriction fragment (Fig.
7A). When DNA-PKcs and DNA
ligase IV-XRCC4 were co-incubated with the linear DNA, ~8% of the
DNA-protein complexes (Fig. 7B, arrowhead c) were
significantly larger than those formed by either DNA-PKcs
(arrowhead a) or DNA ligase IV-XRCC4 (arrowhead
b). The height of the larger DNA-protein complex was
5.62 ± 1.1 nm. Thus, it appears that DNA ligase IV-XRCC4 also
associates with the catalytic subunit of DNA-PK in a DNA end-dependent manner.
View larger version (44K):
[in this window]
[in a new window]
|
Fig. 7.
Tapping mode AFM topographs of DNA-protein
complexes formed by DNA ligase IV-XRCC4 and DNA-PKcs.
A, DNA incubated with DNA-PKcs. A unit length DNA molecule
with a DNA-PKcs molecule bound at each end is shown. B, DNA
ligase IV-XRCC4 and DNA-PKcs co-incubated with DNA (500 × 500-nm
images, zoomed in from 2 × 2-µm scans). arrowhead a,
DNA-PKcs bound to DNA (average height of 3.34 nm); arrowhead
b, DNA ligase IV-XRCC4 bound to DNA (average height of 1.26 nm);
arrowhead c, complex of DNA-PKcs and DNA ligase IV-XRCC4
bound to DNA (average height of 5.62 nm). The initial concentrations of
DNA ligase IV-XRCC4 and DNA were the same as in Fig. 3. The
initial concentration of DNA-PKcs was 200 nM.
|
|
Effect of DNA-PKcs and DNA-PK on DNA Joining by the DNA Ligase
IV-XRCC4 Complex--
Since DNA ligase IV-XRCC4 appears to associate
with Ku and DNA-PKcs at DNA ends, we compared the effects of the DNA-PK
subunits on the ability of DNA ligase IV-XRCC4 to ligate a 400-bp
labeled DNA restriction fragment with complementary ends. In agreement with the results of assays using the oligonucleotide substrate (Fig.
4), Ku inhibited ligation by DNA ligase IV-XRCC4 (Fig.
8). By contrast, the inclusion of
DNA-PKcs did not significantly inhibit DNA joining, but it did alter
the type of ligation product generated. In reactions with the DNA
substrate alone, the predominant product generated by both DNA ligase
IV-XRCC4 and T4 DNA ligase was a circular monomer generated by
intramolecular ligation. Under these conditions, DNA ligase IV-XRCC4
converted 10 fmol of the linear substrate into circular molecules.
Preincubation of the DNA substrate with DNA-PKcs did not significantly
alter the amount of DNA substrate (12 fmol) ligated by DNA ligase
IV-XRCC4. The inclusion of DNA-PKcs did, however, have a dramatic
effect on the type of ligation product generated (Fig. 8). Under these
reaction conditions, ~50% of the ligation events (6 fmol) were now
intermolecular. Since Ku and DNA-PKcs function together in
vivo as the DNA-PK, we preincubated these proteins with the DNA
substrate prior to the addition of DNA ligase IV-XRCC4. Under these
reaction conditions, products generated by intramolecular ligation were
not detectable (Fig. 8). Since Ku did not have the same inhibitory
effect on intermolecular ligation, the inclusion of both subunits of
DNA-PK resulted in DNA ligase IV-XRCC4 joining 5 fmol of the linear DNA
substrate by intermolecular ligation events (Fig. 8). The DNA-PK
inhibitor wortmannin had no effect on either the amount or the
distribution of ligation products (data not shown), indicating that the
kinase activity of DNA-PKcs is not involved in the modulation of the catalytic activity of DNA ligase IV-XRCC4 in these reactions.
View larger version (43K):
[in this window]
[in a new window]
|
Fig. 8.
Effects of DNA-PK and its subunits on the
joining of DNA double-strand breaks by DNA ligase IV-XRCC4. The
joining of a 400-bp labeled restriction fragment with complementary
single-strand ends 4 nucleotides in length was measured as described
under "Experimental Procedures." Where indicated, the DNA substrate
(20 fmol) was preincubated with Ku (200 fmol) and DNA-PKcs (200 fmol)
for 5 min at 25 °C. After the addition of DNA ligase IV-XRCC4 (500 fmol), reactions were incubated at 25 °C for 1 h.
|
|
| |
DISCUSSION |
Genetic studies in both yeast and mammals indicate that DNA ligase
IV participates in NHEJ (3-5, 24, 25). In addition, DNA ligase IV also
plays a critical role during V(D)J recombination in immune cells (24,
25). Both of these DNA transactions involve the end-to-end joining of
duplex DNA molecules. In contrast, the enzymes encoded by the
LIG1 and LIG3 genes complete DNA metabolic pathways, in which the final step is the sealing of a single-strand nick in duplex DNA (1). Based on these observations, one might expect
DNA ligase IV to preferentially join linear duplex DNA molecules
in vitro. However, efforts to characterize the biochemical properties of DNA ligase IV have been hindered by the low levels of
this enzyme activity in mammalian cells and the instability of DNA
ligase IV in the absence of its partner protein XRCC4 (32, 33). In this
study, we have coexpressed human DNA ligase IV and XRCC4 in insect
cells and purified an active complex of these proteins to near homogeneity.
In agreement with published studies (27, 33), DNA ligase IV-XRCC4 was
less active than either DNA ligase I or III with nicked DNA substrates.
Previously, it has been shown that, in the presence of polyethylene
glycol, DNA ligase I ligates blunt-ended DNA molecules, whereas DNA
ligase II does not (26). From this study, it has been inferred that DNA
ligase I is the only mammalian DNA ligase capable of joining blunt ends
and so is likely to be the enzyme completing NHEJ events involving
blunt ends and signal-joint formation during V(D)J recombination (27).
In fact, purified recombinant DNA ligases I and III and DNA ligase
IV-XRCC4 all joined blunt-ended DNA molecules to a similar extent in
the presence of polyethylene glycol, whereas only DNA ligase III
generated detectable amounts of ligated product in the absence of
polyethylene glycol. Analysis of DNA molecules repaired by NHEJ
revealed that a significant proportion of these events appear to be
mediated by short complementary sequences or microhomologies (37). This suggests that the ends of DNA molecules may be aligned by base pairing
between short complementary single-strand sequences, presumably generated by limited nucleolytic degradation, and then ligated. Therefore, we compared the ability of the mammalian DNA ligases to join
linear DNA molecules with cohesive ends. DNA ligase I was significantly
less active than either DNA ligase III or DNA ligase IV-XRCC4 with
this type of substrate.
Since the mammalian DNA ligases catalyze phosphodiester bond formation
by the same reaction mechanism (1), the differences in DNA joining
activity presumably reflect differences in DNA binding properties. In
support of this idea, DNA ligase I does not bind to DNA at
physiological salt concentrations, whereas the zinc finger of DNA
ligase III allows this enzyme to bind to single- and double-strand
breaks in DNA (7, 35). Furthermore, it has been shown recently that
XRCC4 binds preferentially to DNA with nicks and ends, suggesting that
this protein may mediate the binding of DNA ligase IV to DNA (41). Here
we have shown by atomic force microscopy that the DNA ligase IV-XRCC4
complex binds specifically to the ends of duplex DNA molecules and
apparently prefers ends with single-strand overhangs. Furthermore, a
single DNA ligase IV-XRCC4 complex was able to bind simultaneously to the complementary but non-ligatable ends of two DNA molecules. Although
the DNA ligase IV-XRCC4 complex could stabilize DNA molecules interacting via their complementary ends, it seems more likely that DNA
ligase IV-XRCC4 is acting as a bridging factor, binding to the end of
one molecule and then forming a stable complex with the complementary
end of another molecule.
Genetic studies in eukaryotes suggest that Ku and DNA ligase IV
function in the same pathway (3-5, 16, 24, 25). Interestingly, Ku also
binds to DNA ends and can act as a bridging factor between two DNA
molecules (27, 42). Moreover, it was reported that Ku stimulates the
intermolecular joining of both blunt- and sticky-ended DNA molecules by
DNA ligases I, III, and IV (27). The nonspecific stimulation of
ligation by all the mammalian DNA ligases tested suggests that this
effect is due to the structure formed by Ku and the DNA molecules
rather than protein-protein interactions between Ku and the mammalian
DNA ligases. However, the genetic studies predict that there should be
a specific relationship between Ku and DNA ligase IV-XRCC4 (3-5). In
our studies, Ku inhibited intramolecular ligation of DNA molecules with
cohesive ends. This inhibition could not be explained by Ku preventing
DNA ligase IV-XRCC4 from interacting with DNA. Instead, DNA ligase
IV-XRCC4 and Ku were able to bind to the same DNA end. Recent studies
on the initial binding of Ku to DNA by cross-linking suggest that the
DNA end is exposed (43), so the DNA ligase IV-XRCC4 complex may still
be interacting with the terminal nucleotides. This association between
Ku and DNA ligase IV-XRCC4 at the ends of DNA molecules, which is
consistent with the previous observation that XRCC4 stimulates DNA
binding by Ku (44), presumably prevents DNA ligase IV-XRCC4 from
interacting productively with two DNA ends.
Analysis of the DNA-PK complex assembled at a DNA end by cross-linking
revealed that Ku is shifted away from the DNA end by about one helical
turn, with DNA-PKcs binding closest to the DNA end (45). Based on these
results, it appears likely that, in vivo, DNA ligase
IV-XRCC4 interacts with DNA ends bound by DNA-PKcs. Here we have shown
that DNA ligase IV-XRCC4 and DNA-PKcs were able to bind simultaneously
to the same DNA end. Unlike Ku, DNA-PKcs did not markedly inhibit
ligation by DNA ligase IV-XRCC4. Intriguingly, the addition of DNA-PKcs
resulted in a significant fraction of the ligation products being
generated by intermolecular ligation events, whereas reactions with DNA
ligase IV-XRCC4 alone produced predominantly intramolecular ligation
products. To explain intramolecular ligation, we suggest that the DNA
ligase IV-XRCC4 complex first binds to a DNA end and then finds the
second end by random diffusion. For relatively short molecules, this
random process favors interactions with the other end of the same DNA
molecule. In support of this model, decreasing the length of the linear
DNA substrate from 3 kb to 400 bp increased the proportion of ligated
product generated by intramolecular ligation (compare Figs.
2B and 8). To explain the stimulation of intermolecular
ligation by DNA-PKcs, we suggest that DNA-PKcs-containing complexes
formed at DNA ends have affinity for each other. The specific
association of DNA ends via protein-protein interactions would favor
intermolecular interactions and therefore promote intermolecular ligation.
Recently, it has been reported that extracts of human lymphoblastoid
cells catalyze DNA end joining in a reaction that is dependent upon DNA
ligase IV-XRCC4, Ku, and DNA-PKcs (46). Interestingly, the majority of
the end-joining events catalyzed by these extracts were intermolecular.
When we incubated both subunits of DNA-PK with DNA ligase IV-XRCC4,
intramolecular ligation was inhibited, as it was in reactions with Ku,
but intermolecular joining still occurred. Since the distribution of
reaction products generated by purified Ku, DNA-PKcs, and DNA ligase
IV-XRCC4 is similar to that observed in assays with the cell extract
(46), we conclude that intermolecular joining of DNA molecules with
cohesive ends can be carried out in vitro by DNA-PK and DNA
ligase IV-XRCC4. However, there is compelling evidence that the DNA-PK
end-joining reaction mediated by DNA-PK in vivo is more
complex. First, the reaction catalyzed by the cell extract was
sensitive to wortmannin (46), whereas the reconstituted reaction was
not. Second, fractionation of the cell extract revealed the
participation of factors in addition to DNA-PK and DNA ligase IV-XRCC4
in the end-joining reaction (46). Finally, genetic studies in yeast
have implicated other proteins such as the Rad50-Mre11-Xrs2 complex in
Ku-dependent DSB repair (47).
In summary, we have shown that the DNA ligase IV-XRCC4 complex binds
specifically to DNA ends and acts as an alignment factor, holding
together the short complementary single-strand ends of linear duplex
DNA molecules. Both of these properties are consistent with the
predicted role of this enzyme in NHEJ and V(D)J recombination. Furthermore, DNA ligase IV-XRCC4 associates with the Ku and DNA-PKcs subunits of DNA-PK to form specific complexes at DNA ends. However, Ku
and DNA-PKcs have markedly different effects on DNA end joining by DNA
ligase IV-XRCC4. Ku inhibits intramolecular DNA joining, whereas
DNA-PKcs stimulates intermolecular DNA joining even in the presence of
Ku. Further studies will provide insights into the multiprotein
complexes involving DNA ligase IV-XRCC4, Ku, and DNA-PKcs that are
assembled on DNA ends.
TOP
ABSTRACT
INTRODUCTION
To whom correspondence should be addressed: Dept. of Molecular
Medicine, Inst. of Biotechnology, University of Texas Health Science
Center, 15355 Lambda Dr., San Antonio, TX 78245. Tel.: 210-567-7327;
Fax: 210-567-7324; E-mail: Tomkinson@uthscsa.edu.
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
