1 Novel Function of the Fanconi Anemia Group J or RECQ1 Helicase to Disrupt Protein-DNA ...

35 Pages·2014·1.23 MB·English

JBC Papers in Press. Published on June 3, 2014 as Manuscript M113.542456
The latest version is at http://www.jbc.org/cgi/doi/10.1074/jbc.M113.542456
Novel Function of the Fanconi Anemia Group J or RECQ1 Helicase to Disrupt
Protein-DNA Complexes in a Replication Protein A-Stimulated Manner*

Joshua A. Sommers1, Taraswi Banerjee1, Twila Hinds1, Marc S. Wold2,
Ming Lei3, and Robert M. Brosh, Jr.1*

1Laboratory of Molecular Gerontology, National Institute on Aging, National Institutes of Health,
NIH Biomedical Research Center, 251 Bayview Blvd, Baltimore, MD 21224 USA
2Department of Biochemistry, Carver College of Medicine,
University of Iowa, Iowa City, IA 52242 USA
3Department of Biological Chemistry, University of Michigan Medical School,
1150 W. Medical Center Drive, Ann Arbor, MI 48109 USA

*Running Title: FANCJ or RECQ1 disrupts protein-DNA complexes in an RPA-dependent manner
To whom correspondence should be addressed: Robert M. Brosh, Jr., Laboratory of Molecular
Gerontology, National Institute on Aging, National Institutes of Health, NIH Biomedical Research
D
Center, 251 Bayview Blvd, Baltimore, MD 21224 USA; Tel: [phone]; Fax: [phone]; Email: o
w
[email protected] n
lo
a
d
e
d
fro
Keywords: helicase, FANCJ, RECQ1, Replication Protein A, DNA metabolism, genomic instability, m
h
Fanconi anemia, DNA repair, DNA binding protein, genetic diseases ttp
://w
Background: DNA unwinding by helicases is protein-DNA interaction. The Fanconi Anemia w
w
blocked by proteins bound to duplex DNA. Group J (FANCJ) helicase partners with the .jb
c
single-stranded DNA binding protein .org
Results: The single-stranded DNA binding Replication Protein A (RPA) to displace b/
y
protein RPA stimulates FANCJ or RECQ1 BamHI-E111A bound to duplex DNA in a gu
e
helicase to disrupt protein-DNA complexes. specific manner. Protein displacement was st o
n
dependent on the ATPase-driven function of N
Conclusion: Helicases partner with RPA to o
the helicase and unique properties of RPA. ve
dislodge proteins bound to duplex DNA. m
Further biochemical studies demonstrated that b
e
the shelterin proteins TRF1 and TRF2, which r 2
Sprigonteiifnic dainspcela:c eRmegenutl aitsi ohnig ohfl yh erelilceavsaen-tc ainta clyezlleudl ar preferentially bind the telomeric repeat found 0, 201
at chromosome ends, effectively block FANCJ 8
nucleic acid metabolic processes that require
from unwinding the forked duplex telomeric
remodeling of chromatinized genomic DNA.
substrate. RPA, but not the E. coli single-
ABSTRACT stranded DNA binding protein or shelterin
factor Pot1, stimulated FANCJ ejection of
Understanding how cellular machinery TRF1 from the telomeric DNA substrate.
deals with chromosomal genome complexity is FANCJ was also able to displace TRF2 from
an important question because protein bound the telomeric substrate in a RPA-dependent
to DNA may affect various cellular processes of manner. The stimulation of helicase-catalyzed
nucleic acid metabolism. DNA helicases are at protein displacement is also observed with the
the forefront of such processes, yet there is only DNA helicase RECQ1, suggesting a conserved
limited knowledge how they remodel protein- functional interaction of RPA-interacting
DNA complexes and how these mechanisms are helicases. These findings suggest that
regulated. We have determined that partnerships between RPA and interacting
representative human RecQ and Fe-S cluster human DNA helicases may greatly enhance
DNA helicases are potently blocked by a their ability to dislodge proteins bound to
1

Copyright 2014 by The American Society for Biochemistry and Molecular Biology, Inc.
duplex DNA, an activity that is likely to be interstrand cross-link (ICL) repair (6). FANCJ is
highly relevant to their biological roles in DNA believed to play a role in homologous
metabolism. recombinational (HR) repair by virtue of its
catalytic DNA unwinding function and
Protein bound to single-stranded or interactions with protein partners including the
double-stranded DNA may impede the progress of tumor suppressor BRCA1 (7), the mismatch repair
replication or transcription, or affect processing protein MLH1 (8), and the single-stranded DNA
steps during DNA repair or recombination; binding protein Replication Protein A (RPA) (9).
therefore, there has been considerable interest in Although FANCJ is implicated in downstream
the cellular mechanism to overcome protein events of double strand break (DSB) repair of
blockades. One of the first pieces of evidence that ICL-induced DNA damage, growing evidence
a helicase may facilitate displacement of protein suggests that FANCJ and other FA-linked gene
bound to DNA was provided by Bedinger et al. products have roles outside the classical FA
from the Alberts lab in which they showed that pathway. For example, FANCJ resolves G-
bacteriophage T4 Dda helicase can dislodge a quadruplex (G4) DNA structures to preserve
stationary RNA polymerase from the DNA chromosomal stability (10,11), and may have a
template (1). However, biochemical evidence unique role to suppress G4 accumulation or G4-
D
from the Matson lab demonstrated that the Lac induced DNA damage in vivo (12,13). G4 DNA o
w
repressor protein bound to DNA inhibited (14) and three-stranded T-loops (15) are believed nlo
a
unwinding by various bacterial and bacteriophage to be prominent structures at telomeres, and recent de
d
helicases to different extents (2), suggesting that evidence suggests that FANCJ is found at fro
cellular mechanisms may exist to promote the telomeres in telomerase negative human cells that hm
displacement of proteins bound to genomic DNA. preserve their chromosome ends by the alternative ttp
Mechanistic studies from the Raney lab showed lengthening of telomeres (ALT) pathway (16). ://ww
that Dda helicase displaces E. coli trp repressor Telomeres are bound by shelterin proteins (e.g., w.jb
from its operator sequence by a cooperative TRF1, TRF2, Pot1), and it is generally thought c.o
inchworm mechanism (3). Overall, only a limited that certain helicases (e.g., WRN, BLM, RTEL) brg/
number of studies have examined the ability of help to preserve telomere stability by unwinding y g
u
prokaryotic and eukaryotic DNA helicases to T-loops and G4 DNA; alternatively, certain es
displace proteins bound to DNA (4) (Table 1). helicases (e.g., PIF1) catalytically displace t on
N
For the limited number of helicases that have been telomerase from DNA to facilitate replication and o
v
shown to catalytically displace protein from DNA, DNA repair (17-19). Understanding how the em
b
the mechanism(s) are not well understood, catalytic functions of DNA helicases are regulated er 2
particularly in eukaryotes. Understanding how remains an important question in telomere 0, 2
helicase-enacted protein displacement is regulated biology. 018
should be informative in the context of cellular Previously, we reported that FANCJ
chromosomal DNA metabolism in mammalian associates with RPA in a DNA damage-inducible
cells, which is likely to be much more complex manner and that RPA physically binds to FANCJ
compared to biochemical reactions with naked and stimulates its helicase activity (9). RPA also
DNA substrates frequently used in in vitro studies. interacts with several other SF2 RecQ DNA
A number of superfamily 2 (SF2) DNA helicases (WRN, BLM (20), RECQ1 (RECQL,
repair helicases from the RecQ and Iron-Sulfur RECQL1) (21)) and stimulates their DNA
(Fe-S) cluster families are implicated in hereditary unwinding activities, suggesting a conservation of
disorders characterized by chromosomal instability the helicase-RPA protein interaction that is
(5), and it is unclear if and how efficiently these important for chromosomal stability. In the
helicases remodel proteins bound to DNA in vivo. current study, we have investigated a novel aspect
Among these helicase proteins, mutations in of the functional interaction between RPA and
FANCJ helicase are associated with breast cancer select SF2 DNA repair helicases that enables the
and genetically linked to Fanconi Anemia (FA), a helicase to displace protein tightly bound to
progressive bone marrow failure disorder duplex DNA. Our biochemical results
characterized by elevated cancer and a defect in demonstrate that FANCJ and other helicases are
2
strongly blocked from unwinding a forked duplex 22 or 34 base pairs flanked by 15 nucleotide 5’
DNA substrate bound by a catalytically inactive and 3’ overhangs.
restriction endonuclease or telomere-specific DNA
binding proteins. RPA enables FANCJ or RECQ1 Protein Displacement / Helicase Assays For
helicase to disrupt the protein-DNA interaction reactions with RECQ1 (23) and WRN (22), there
and unwind the double-stranded region to which was an initial 15 min incubation of the forked
the protein was bound. These findings suggest DNA substrate (0.5 nM) with BamHI-E111A
that partnerships between RPA and certain (indicated concentrations) at 24° C in reaction
helicases may greatly enhance their ability to buffer previously described, followed by the
dislodge proteins bound to duplex DNA, an addition of RECQ1 or WRN in the absence or
activity that is likely to be highly relevant to their presence of the indicated concentration of RPA or
biological roles in DNA metabolism. ESSB. The reactions were then incubated at 37 °C
for 15 min. For reactions with FANCJ (29), there
EXPERIMENTAL PROCEDURES was an initial incubation of the indicated forked
Recombinant Proteins Baculoviruses encoding DNA substrate with BamHI-E111A, TRF1 or
FANCJ with a C-terminal FLAG tag (12), His- TRF2 at 24° C for 15 min, followed by the
tagged WRN (22), or His-tagged RECQ1 (23) addition of FANCJ in the absence or presence of
D
were used to infect High Five insect cells, and the the indicated concentrations of RPA, Pot1, ESSB o
w
n
recombinant proteins were purified as previously homotetramer, or BLM proteins. The reactions lo
a
described. BamHI-E111A was provided by New were then incubated at 30 °C for 15 min. It is de
d
England Bio-Labs. Pot1 (Protection of Telomeres noted that the RPA heterotrimer concentration fro
m
1) (24), TRF1 (25), and TRF2 (25) were purified used for helicase experiments with the 22-bp h
as described previously. BLM was kindly telomeric forked duplex substrates was 5 nM ttp
provided by Dr. I. Hickson (Univ. Copenhagen) compared to 10 nM RPA for forked 26-bp or 34- ://ww
w
that was purified as described previously (26). bp duplex substrates because the shorter 22-bp .jb
Aro1-RPA and wtRPA were purified as described substrate was destabilized by greater RPA c.o
rg
previously (27). E. coli SSB (ESSB) was from concentrations. All reactions were terminated by b/
y
Promega. addition of 20 μl of Stop buffer containing final g
u
concentrations of 0.3 % SDS, 9 mM EDTA, es
t o
DNA Substrates and Oligonucleotides 0.02% bromophenol blue, 0.02% xylene cyanol, n
N
Polyacrylamide gel electrophoresis purified 12.5% glycerol, 5 nM unlabeled oligonucleotide, o
v
e
oligonucleotides used for the preparation of forked and 500 μg/ml Proteinase K. The reactions were m
b
e
DNA substrates were purchased from Loftstrand further incubated at 37 °C for 15 min, and resolved r 2
0
Labs (Gaithersburg, MD). All DNA substrates on non-denaturing 12% (19:1 , 2
0
were prepared by annealing 10 pmol of 32P-end- acrylamide:bisacrylamide) polyacrylamide gels in 18
labeled oligonucleotide to 25 pmol of unlabeled 1X TBE and quantitated as previously described
oligonucleotide as described previously (22). The (11).
forked DNA substrate with the BamHI consensus
sequence was prepared using oligonucleotides Electrophoretic Mobility Shift Assays (EMSA)
BAMHIFORKA Protein/DNA binding mixtures (20 µl) contained
(TTTTTTTTTTTTTTTTTTTTTTTTTTTTTTGA the indicated concentrations of BamHI-E111A and
ATTGCACCGGATCCCTAGGTCGAT) and 0.5 nM of the 32P-end-labeled forked DNA in
BAMHIFORKB binding buffer (10 mM Tris pH 7.4, 1 mM EDTA,
(ATCGACCTAGGGATCCGGTGCAATTCTTT 100 mM potassium acetate, 50 μM DTT, 100
TTTTTTTTTTTTTTTTTTTTTTTTTTT). This μg/ml BSA and 10% glycerol). For those
resulted in a substrate with a 26 base pair region experiments in which BamHI-digested calf thymus
and 30 nucleotide dT 5’ and 3’ overhangs. The DNA was included in the binding reaction
sequences of the oligonucleotides used for the mixture, the BamHI-digested calf thymus DNA
preparation of the forked duplex DNA substrates was prepared by incubating 5 μg of calf thymus
with two or four telomeric repeats are described DNA (Sigma) with 50 units of BamHI in
elsewhere (28). These substrates contained either NEBuffer 3 (New England Bio-Labs). The
3
binding mixtures were incubated at 37 °C for 30 annealed 26-bp forked duplex DNA molecule was
min. After the incubation, 4 µl of Loading dye used as a marker for the annealed product.
(74% glycerol, 0.01% xylene cyanol, 0.01%
bromophenol blue) was added to each mixture, RESULTS
and samples were loaded onto native 8% (19:1 Catalytically inactive BamHI inhibits DNA
acrylamide:bisacrylamide) polyacrylamide gels unwinding activity by human DNA repair
and electrophoresed at 200 V for 2.5 h at 4 °C helicases
using 0.5X TBE as the running buffer. The We began our studies by testing if a
resolved radiolabeled species were visualized catalytically inactive BamHI-E111A restriction
using a Phosphor-Imager and analyzed with endonuclease bound to a forked duplex DNA
ImageQuant software (GE Healthcare). substrate harboring a cognate palindromic BamHI
The reaction mixtures for EMSA recognition sequence (Fig. 1A) was able to inhibit
experiments with the telomeric 22-bp forked helicase-catalyzed unwinding of the DNA
duplex DNA substrate and indicated proteins were substrate. EMSA was used to demonstrate that
the same as those in the protein purified recombinant BamHI-E111A protein
displacement/helicase assays except that bound the radiolabeled forked duplex substrate in
Proteinase K digestion of the binding mixture a protein concentration dependent manner in the
D
products was not performed. Samples were presence or absence of excess BamHI-digested o
w
n
treated with Loading dye as described above and calf thymus DNA (Fig. 1B). These results are lo
a
resolved on native 5% consistent with previously published nitrocellulose de
d
(37.5:1 acrylamide:bisacrylamide) polyacrylamide filter binding studies (30) and unpublished EMSA fro
m
gels by electrophoresis at 200 V for 2 h at 4 °C data from Jurate Bitinaite (New England BioLabs h
using 0.5X TBE as the running buffer. (Ipswich, MA)). BamHI-E111A protein titration ttp
experiments revealed that inhibition of FANCJ ://ww
w
Restriction Enzyme Protection Experiments The helicase activity was observed to be dependent on .jb
26 bp forked duplex DNA substrate containing the the concentration of BamHI-E111A (Fig. 1C). A c.o
rg
BamHI restriction endonuclease recognition concentration of 38 nM BamHI-E111A inhibited b/
y
sequence was preincubated with BamHI-E111A FANCJ helicase activity to near completion (Fig. g
u
(38 nM) for 15 min at 30 °C followed by addition 1D). The greater concentration of BamHI-E111A es
t o
of RPA (10 nM) and further incubation at 30 °C required for nearly complete FANCJ helicase n
N
for 15 min. BamHI-WT (0.25 U; New England inhibition compared to the 9 nM BamHI-E111A o
v
e
Bio-Labs) was added to reaction mixtures and concentration required for the majority of substrate m
b
e
incubated for 15 min at 30 °C, followed by bound as measured by EMSA may reflect r 2
0
addition of Proteinase K (500 μg/ml) with Stop differences in reaction mixtures and conditions for , 2
0
buffer and further incubation at 37 °C for 10 min. the two assays and/or the effect of native 18
DNA products were resolved by electrophoresis polyacrylamide gel electrophoresis on the stability
on non-denaturing 12% polyacrylamide gels. of the protein-DNA complex.
Radiolabeled species were detected and Using this concentration of BamHI-
quantitated as described above. E111A, we observed strong inhibition of RECQ1
(Fig. 1E) or WRN helicase activity as well (Fig.
Strand Annealing Experiments Partially 1F). In the WRN reactions, the unwound product
complementary single-stranded oligonucleotides migrated slightly faster than the heat-denatured
used to prepare 26-bp forked duplex DNA DNA substrate control (Fig. 1F), consistent with
substrate containing the BamHI restriction earlier observations that WRN helicase product
endonuclease recognition sequence were was slightly degraded by intrinsic WRN nuclease
incubated with the indicated concentration of RPA (31). The residual unwound product observed in
or ESSB for 15 min at 30 °C in the presence or WRN reactions containing BamHI-E111A
absence of BamHI-E111A (38 nM) and migrated even slightly farther than the products
subsequent digestion with Proteinase K (500 from reactions lacking BamHI-E111A, suggesting
g/ml) in Stop buffer. Products were resolved on further degradation by WRN nuclease when the
non-denaturing 12% polyacrylamide gels. The helicase activity is blocked.
4
Recent studies provided evidence for the duplex substrate. We also tested a FANCJ-A349P
coordination of E. coli DNA helicases protein, derived from a clinically relevant FANCJ
translocating on opposite strands to displace mutation in the conserved iron-sulfur domain that
protein bound to duplex DNA (see Discussion and is genetically linked to FA (34). The FANCJ-
(32)). Based on this paradigm, we tested if the A349P mutant protein was previously
purified recombinant human FANCJ and BLM demonstrated to bind DNA, hydrolyze ATP, and
helicases, which are known to unwind DNA with translocate on single-stranded DNA, but fail to
opposite polarities yet physically and functionally unwind duplex or G-quadruplex DNA substrates
interact (29), might collaborate to displace (35). In the current study, RPA failed to stimulate
BamHI-E111A from the forked duplex DNA FANCJ-A349P to unwind the BamHI-E111A-
substrate. The results from this experiment proved bound fork duplex substrate, indicating that
to be negative. BamHI-E111A failed to be FANCJ-catalyzed DNA unwinding activity is
displaced from the DNA substrate in the presence responsible for RPA-dependent displacement of
of FANCJ, BLM, or BLM + FANCJ under BamHI-E111A from the forked duplex DNA
conditions that each DNA helicase could substrate. Collectively, the studies using catalytic
effectively unwind the naked protein-free forked mutants of FANCJ suggest that FANCJ and RPA
duplex (Fig. 1G), attesting to the strong blockade collaboratively displace BamHI-E111A from the
D
imposed by the catalytically inactive restriction forked duplex as the helicase unwinds the DNA o
w
n
endonuclease bound to the double-stranded region substrate. lo
a
of the partial duplex DNA molecule. Kinetic assays were performed to measure de
d
DNA unwinding of the BamHI-E111A-bound fro
m
RPA stimulates FANCJ displacement of forked duplex DNA substrate by FANCJ in the h
eBnaambHlinI-gE t1h1e1 hAe lbicoausned t oto u dnowuibnlde -tshter asnudbesdtr aDtNe A pshreoswenedce t hoart aRbPseAn cceo uolfd R sPtiAm.u lTahtei sF aAnNalCysJi sto ttp://ww
w
Previously, it was reported that the single- displace BamHI-E111A and unwind the forked .jb
stranded DNA binding protein (SSB) Replication duplex in a kinetic manner throughout a 16 min c.o
rg
Protein A (RPA) physically interacts with FANCJ time course (Fig. 2C). By the completion of the b/
y
and stimulates its DNA unwinding reaction (9). time course, FANCJ was able to unwind nearly g
u
This led us to ask if RPA was able to stimulate 60% of the BamHI-E111A-bound forked duplex es
t o
FANCJ to dislodge BamHI-E111A from the when RPA was present; however, in the absence n
N
forked duplex DNA substrate and unwind the of RPA, FANCJ unwound only approximately o
v
e
underlying duplex. As shown in Fig. 2A, the 15% of the forked duplex bound by BamHI- m
b
e
presence of RPA (10 nM) enabled FANCJ to E111A. From a kinetic analysis of the initial rates r 2
0
unwind the BamHI-E111A-bound forked duplex of DNA unwinding, the percent of BamHI-E111A- , 2
0
to nearly the level of FANCJ helicase activity on bound forked duplex unwound by FANCJ in the 18
the naked forked duplex. In control reactions, presence of RPA was 2-fold greater than that
RPA alone had only a marginal effect on the observed for the BamHI-E111A-bound substrate
stability of the naked forked duplex DNA unwound by FANCJ acting alone. For the naked
substrate or the BamHI-E111A-bound forked DNA substrate, RPA increased the extent of DNA
duplex (Fig. 2A). To ascertain if the observed substrate unwound by FANCJ in the latter
RPA-dependent unwinding of the forked duplex timepoints (8, 16 min). These results demonstrate
DNA substrate bound by BamHI-E111A was that RPA significantly enhanced the rate of
dependent on intrinsic FANCJ ATPase/helicase FANCJ-catalyzed BamHI-E111A displacement
activity, we tested a catalytically inactive FANCJ- from the forked duplex DNA substrate.
K52R mutant protein characterized by an amino To address if RPA acting alone was able
acid replacement of the highly conserved lysine-52 to displace BamHI-E111A from the forked duplex
in the Walker A (motif I) box with an arginine. DNA substrate harboring its cognate recognition
The FANCJ-K52R protein was previously shown sequence element, restriction enzyme protection
to be defective as an ATPase or helicase (33). As experiments were performed. BamHI-E111A was
shown in Fig. 2B, RPA failed to stimulate FANCJ- preincubated with the DNA substrate followed by
K52R to unwind the BamHI-E111A-bound forked addition of RPA for 10 min. The reaction mixture
5
was then incubated with catalytically active determined that RPA stimulated FANCJ
BamHI-WT enzyme, and the DNA products were displacement of BamHI-E111A from the DNA
resolved by electrophoresis on native substrate 3.5-fold, whereas ESSB actually showed
polyacrylamide gels. As shown by a inhibition of FANCJ-catalyzed BamHI-E111A
representative gel in Fig. 2D, a similar level of displacement (Fig. 4B), presumably due to ESSB
DNA substrate was protected from BamHI-WT competing with FANCJ for loading on to the DNA
digestion by the inactive BamHI-E111A in the substrate. These results suggest that RPA
presence or absence of RPA (10 nM). These stimulates FANCJ to displace BamHI-E111A from
results suggest that RPA alone does not displace the forked duplex and unwind the DNA substrate
BamHI-E111A from the DNA substrate, enabling in a specific manner because a heterologous SSB
FANCJ to unwind the duplex DNA. failed to stimulate FANCJ protein displacement.
To further address the mechanism of
Specificity and DNA binding requirement of stimulation, we evaluated the effect of BamHI-
RPA to stimulate FANCJ protein displacement E111A on rewinding the complementary strands
To assess if the ability of RPA to avidly of the DNA substrate and the effect of single-
bind single-stranded DNA was required for it to stranded DNA binding proteins RPA or ESSB on
stimulate FANCJ displacement of BamHI-E111A rewinding. As shown in Fig. 4C, D, the presence
D
from the forked duplex substrate, we tested a of BamHI-E111A with the unwound o
w
n
previously studied RPA mutant heterotrimer complementary strands effectively promoted their lo
a
(Aro1-RPA), characterized by two amino acid annealing. However, the presence of RPA (Fig. de
d
substitutions within the DNA binding domain of 4C) or ESSB (Fig. 4D) inhibited BamHI-E111A- fro
m
the RPA70 subunit (RPA1) (Fig. 3A), that is promoted strand annealing at concentrations as h
significantly compromised in single-stranded low as 1.3 nM single-stranded DNA binding ttp
DNA binding (27). As shown in Fig. 3B, the protein. Based on these results, we conclude that ://ww
w
Aro1-RPA mutant heterotrimer poorly stimulated although both RPA and ESSB can prevent .jb
FANCJ to displace BamHI-E111A from the forked rewinding of the complementary strands, only c.o
rg
duplex under reaction conditions that the wild-type RPA could stimulate FANCJ-catalyzed protein b/
y
RPA heterotrimer readily did so. Based on these displacement, indicating a specific RPA-FANCJ g
u
results, we conclude that the single-stranded DNA interaction to promote protein displacement during es
t o
binding activity of RPA is necessary for it to the DNA unwinding reaction by the helicase. n
N
stimulate FANCJ-catalyzed displacement of o
v
e
BamHI-E111A from the DNA substrate during the RPA stimulates RECQ1 to disrupt BamHI- m
b
e
DNA unwinding reaction. E111A-DNA substrate interaction r 2
0
To address if stimulation of FANCJ to Previously, it was demonstrated that RPA , 2
0
displace BamHI-E111A and unwind the forked physically and functionally interacts with several 18
duplex was specific, we substituted the E. coli SF2 RecQ DNA repair helicases to stimulate their
Single-stranded DNA Binding Protein (ESSB) for DNA unwinding reactions (38), suggesting that
RPA in the FANCJ reaction mixtures with the the ability of RPA to stimulate the Fe-S cluster
BamHI-E111A-bound forked duplex DNA DNA helicase FANCJ to disrupt protein-DNA
substrate. For these experiments, we used a complexes might also be observed for a human
concentration of ESSB homotetramer the same as RecQ helicase. To address this, the human
RPA heterotrimer (10 nM) because one ESSB RECQ1 helicase, which is known to interact with
homotetramer binds 35 nucleotides (36), and one RPA (21), was tested for its ability to dislodge
RPA heterotrimer binds 30 nucleotides (37). BamHI-E111A from the forked duplex DNA
ESSB completely failed to stimulate FANCJ to substrate in the presence of RPA. As shown in
displace BamHI-E111A from the forked duplex Fig. 4E, RPA stimulated RECQ1-catalyzed
and unwind the DNA substrate under conditions BamHI-E111A displacement. Quantitative
that FANCJ was able to efficiently unwind the analysis of the products from RECQ1-RPA
naked forked duplex or RPA effectively stimulated reaction mixtures with the BamHI-E111A-bound
FANCJ to displace BamHI-E111A from the DNA forked duplex revealed a 2-fold stimulation
substrate (Fig. 4A). By quantitative analysis, we compared to the additive effect of RECQ1 or RPA
6
acting separately on the BamHI-E111A-bound modestly destabilized the 22-bp telomeric forked
substrate (Fig. 4F). The stimulation of RECQ1- duplex, presumably by its ability to bind the
catalyzed protein displacement was specific as single-stranded tails and capture fraying of the
evidenced by the inability of ESSB to substitute short duplex. Taken together, these results
for RPA in the reaction. strongly suggest that RPA alone poorly displaces
TRF1 from the DNA substrate and the substrate
RPA stimulates FANCJ- or RECQ1-catalyzed bound by TRF1 is effectively unwound by the
displacement of TRF1 bound to a telomeric concerted action of FANCJ and RPA.
duplex substrate We next wanted to ascertain if TRF1 was
The shelterin protein TRF1 is known to actually removed from the DNA substrate by
preferentially bind duplex DNA harboring the FANCJ and RPA acting together. To address this,
telomeric sequence repeat TTAGGG-AATCCC we examined the DNA-bound protein species from
and contribute to the protection of the reaction mixtures containing FANCJ + RPA with
chromosome end (39,40). However, telomere the TRF1-bound forked telomeric duplex
structural dynamics may require helicase-mediated substrate. In this case, Proteinase K was not
unwinding of protein-bound t-loops and other included during the quench with STOP buffer.
DNA structures in order for telomeres to be Here we observed that TRF1 inhibited FANCJ
D
properly replicated or repaired (17). Previously, it helicase activity as before, and that the additional o
w
n
was reported that FANCJ can be found at presence of RPA with FANCJ and the TRF1- lo
a
telomeres of living cells that preserve their bound DNA substrate during the incubation period de
d
chromosome ends by the ALT pathway (16), resulted in the appearance of RPA-bound single- fro
m
suggesting a potential role of FANCJ in telomeric stranded DNA (Fig. 5D). These experimental data h
DNA metabolism. The ability of RPA to stimulate provide evidence that TRF1 was removed from the ttp
FANCJ to displace BamHI-E111A and effectively DNA substrate by the combined presence of ://ww
w
unwind the protein-bound forked duplex raised the FANCJ and RPA in the reaction mixture. .jb
experimental question if RPA would exert a To evaluate if the effect of RPA on TRF1 c.o
rg
similar effect and enable FANCJ to displace TRF1 displacement from the telomeric forked duplex by b/
y
from the telomeric forked duplex substrate and FANCJ was specific, we assessed the ability of g
u
unwind the underlying duplex. To perform this ESSB to substitute for RPA in this capacity. es
t o
experiment, we used a 22-bp forked duplex ESSB failed to stimulate FANCJ displacement of n
N
harboring two direct telomeric repeats (Fig. 5A). TRF1 under conditions that RPA effectively did so o
v
e
TRF1 binds the telomere repeat as a homo-dimer (Fig. 5E), attesting to the specificity of the m
b
e
in which each monomer binds a minimal FANCJ-RPA interaction. To further address the r 2
0
consensus sequence of 5’-YTAGGGTTR-3’ (41); specificity of RPA-FANCJ functional interaction, , 2
0
therefore, the DNA substrate with two direct we tested if the shelterin protein POT1 was able to 18
telomeric repeats would be expected to bind one stimulate FANCJ to unwind the forked duplex
TRF1 dimer. TRF1 binding to the forked duplex with TRF1 bound to the 22-bp telomeric duplex.
caused the DNA substrate to be strongly retarded Pot1 shares a conserved
upon electrophoresis on a native 5% oligonucleotide/oligosaccharide binding (OB) fold
polyacrylamide gel (Fig. 5B), a result that is also found in RPA and interacts with strong
consistent with a previous report (41). RPA (5 specificity to the telomeric ssDNA sequence 5’-
nM) alone had only a very minor effect on TRF1 (T)TAGGGTTAG-3’ (42). The experimental
displacement (Fig. 5B). Analysis of helicase results demonstrated that POT1 did not stimulate
reaction products that had been Proteinase K FANCJ under conditions that RPA did (Fig. 5F),
treated revealed that the presence of TRF1 in the further attesting to the specificity of the FANCJ-
reaction mixture effectively inhibited FANCJ RPA interaction. Taking into account the effect of
unwinding of the telomeric forked duplex RPA alone on destabilization of the forked duplex,
substrate (Fig. 5C). However, in the presence of a quantitative analysis demonstrated that RPA
RPA (5 nM), FANCJ unwound the telomeric stimulated FANCJ-displacement of TRF1
forked duplex to near completion in the 15-min approximately 10-fold (Fig. 5G).
reaction. In control reactions, RPA alone only
7
Recent evidence suggests a potential role We next tested if FANCJ could displace
of RECQ1 in telomere metabolism (43). The TRF2, another shelterin protein that binds
ability of RPA to stimulate RECQ1 removal of the telomeric repeats (44), from the forked duplex
catalytically inactive BamHI mutant protein bound substrate harboring the 34-bp region with four
to a DNA substrate suggested RPA may also direct TTAGGG-AATCCC repeats. Like TRF1,
stimulate RECQ1 displacement of the shelterin TRF2 binds as a homo-dimer in which each
proteins TRF1 and TRF2 bound to telomeric monomer binds a minimal consensus sequence of
duplex sequence. Using the 22-bp forked duplex 5’-YTAGGGTTR-3’ (42,45). Therefore, the 34-
with the two direct telomeric repeats described bp duplex with four telomeric repeats would be
above, we found that TRF1 inhibited RECQ1 expected to maximally bind two homo-dimers of
unwinding of the telomeric substrate (Fig. 5H, I). TRF2. TRF2 prebound to the 34-bp telomeric
In the presence of RPA (5 nM), RECQ1 forked duplex prevented FANCJ from unwinding
unwinding of the TRF1-bound substrate was the substrate (Fig. 6D). However, in the presence
stimulated ~2-fold, taking into account the fraction of RPA, FANCJ was able to unwind the TRF2-
of TRF1-bound substrate destabilized by RPA. bound substrate, displacing TRF2 bound to the
four consecutive telomeric repeats.
RPA-dependent FANCJ or RECQ1 We next evaluated the ability of RECQ1
D
displacement of TRF1 or TRF2 bound to a with RPA to displace TRF2 bound to the 34-bp o
w
n
DNA substrate harboring four telomeric telomeric forked duplex substrate. When RPA (10 lo
a
repeats nM) was present in the reaction mixture de
d
To further examine the effect of RPA on containing RECQ1 and the TRF2-bound telomeric fro
m
FANCJ displacement of TRF1 bound to telomeric substrate, 13% of the DNA substrate was unwound h
duplex, reactions were performed with a longer whereas only ~0.5% of the TRF2-bound DNA ttp
34-bp forked duplex substrate characterized by the substrate was unwound by RECQ1 alone (Fig. ://ww
w
presence of four TTAGGG-AATCCC repeats in 6E). Based on these results, we conclude that .jb
the duplex region (Fig. 6A). Thus, the 34-bp RPA stimulates RECQ1 displacement of TRF1 or c.o
rg
duplex with four telomeric repeats would be TRF2 bound to telomeric substrates. b/
y
expected to maximally bind two homo-dimers of g
u
TRF1. Due to their low processivity, RECQ1, like DISCUSSION es
t o
FANCJ, alone very poorly unwound the naked Recent biological data has suggested that n
N
DNA substrate; however, in the presence of RPA protein-DNA complexes are the primary source of o
v
e
RECQ1 and FANCJ unwound over 50% of the replication fork pausing in E. coli (46). Emerging m
b
e
telomeric forked duplex in an RPA concentration- evidence supports a model in which replication r 2
0
dependent manner (Fig. 6B). Preincubation of the collisions are overcome by accessory helicases , 2
0
helicase substrate with RPA followed by the that assist the replisome to progress through 18
initiation of the helicase reaction with the addition protein-DNA blockades (32). At present, the
of FANCJ showed a comparable level of DNA strongest biochemical and genetic evidence for the
unwinding as observed for reactions in which role of accessory replicative motors to displace
FANCJ and RPA were simultaneously added to proteins bound to double-stranded DNA ahead of
reaction mixtures (data not shown), suggesting the fork has been garnered from studies in
that RPA does not recruit FANCJ to the helicase prokaryotic systems. Guy et al. provided evidence
substrate. Further studies are warranted to study that Rep and UvrD are able to promote replisome
the mechanism of RPA stimulation of FANCJ and progression along protein-bound DNA both in
RecQ DNA helicases. vitro and in vivo (47). Studies from the McGlynn
TRF1 prebound to the 34-bp telomeric (47-50) and Michel (51-53) labs have provided
forked duplex prevented FANCJ from unwinding both molecular and cellular evidence delineating
the substrate (Fig. 6C). However, in the presence the roles of accessory helicases in bacteria to
of RPA, FANCJ displaced TRF1 bound to the four disrupt protein-DNA complexes that would
consecutive telomeric repeats and unwind the impede the replication fork and pose a source of
telomeric duplex. chromosomal instability. While considerable
evidence suggests that eukaryotic DNA helicases
8
(e.g., Srs2) serve to dislodge proteins (Rad51 replication and transcription (48). It is
recombinase) bound to single-stranded DNA to conceivable that a specific interaction exists in
regulate HR (54-57), it has also been proposed that mammalian cells between the replicative MCM
helicases or helicase-like proteins (e.g., Rad54) helicase and an auxiliary helicase yet to be
dislodge proteins bound to double-stranded DNA determined.
leading to the disassembly of toxic recombination FANCJ, which interacts with RPA (9),
intermediates or static dead-end complexes (58) was able to efficiently displace BamHI-E111A
(for review, see (59)). The human helicase-like from the forked duplex and unwind the underlying
transcription factor (HLTF) coordinates duplex. Protein displacement from the DNA
remodeling of stalled replication forks by substrate was dependent on the intrinsic ATPase
displacing the clamp PCNA or clamp loader RFC activity of FANCJ, and required a helicase-
(60). Although these studies suggests a potential proficient version of FANCJ because the
role of helicase-like proteins to dislodge proteins translocase-proficient but helicase-inactive
bound to DNA, very little is known about the FANCJ-A349P mutant protein failed to do so.
mechanism, regulation, and biological significance Furthermore, the ability of RPA to promote
of helicase-catalyzed protein displacement, FANCJ disruption of the BamHI-E119A-DNA
especially in eukaryotes. interaction required the high affinity DNA binding
D
In the current study, we have carefully activity of RPA. The dysfunctional DNA o
w
n
examined the effect of protein bound to duplex replication and repair phenotypes of human cells lo
a
DNA on the catalytic DNA unwinding function of expressing the Aro1-RPA mutant (27) may be de
d
representative human SF2 DNA repair helicases attributed to at least in part the disruption of fro
m
that are important for helping cells deal with functional interactions between RPA and DNA h
replication stress. BamHI-E111A served as a helicases responsible for displacing proteins bound ttp
useful tool for this analysis because the restriction to genomic DNA and unwinding the underlying ://ww
w
endonuclease with the active site mutation DNA duplexes. .jb
displays a high affinity for its cognate recognition In addition to the high affinity binding of c.o
rg
sequence (Kd = 2.95E-11 M) (Jurate Bitinaite, single-stranded DNA by RPA, the protein b/
y
unpublished data), approximately 300-fold greater interaction between RPA and FANCJ (as well as g
u
than normal BamHI (30). We determined that the interaction between RPA and RECQ1) is likely es
t o
DNA helicase activity catalyzed by FANCJ, to be important for protein ejection from DNA n
N
RECQ1, and WRN is potently blocked by a based on our observation that a heterologous SSB o
v
e
catalytically inactive BamHI-E111A restriction was unable to stimulate FANCJ- or RECQ1- m
b
e
endonuclease bound to a forked duplex DNA catalyzed protein displacement. This result r 2
0
substrate with a centrally positioned cognate parallels our previous findings that the physical , 2
0
palindromic hexanucleotide double-stranded DNA interaction between RPA and the WRN helicase is 18
sequence recognized by the endonuclease. We required for stimulation of helicase-catalyzed
next asked if FANCJ might function with one of DNA unwinding of protein-free partial duplex
its known helicase protein partners to efficiently DNA substrates (20). The current results robustly
displace protein bound to a DNA substrate. demonstrate that a protein bound to duplex DNA
However, FANCJ failed to collaborate with the 3’ potently blocks unwinding by representative SF2
to 5’ DNA repair helicase BLM, with which it is Fe-S and RecQ helicases and this inhibition can be
known to interact (29), to dislodge BamHI-E111A overcome in a specific manner by the
bound to the forked duplex. Further studies will collaborative action of RPA and its interacting
elucidate if certain opposite polarity DNA helicase.
helicases in eukaryotic cells act together to We next evaluated the effect of the
synergistically disrupt proteins bound to duplex shelterin proteins TRF1 or TRF2 bound to a
DNA. Up to this point, it is unknown if eukaryotic forked duplex substrate harboring direct telomeric
cells follow the precedent set by the Rep 3’ to 5’ repeats on DNA unwinding by FANCJ, a protein
auxiliary helicase which functionally cooperates that was previously shown to be found at
with the E. coli DnaB replicative 5’ to 3’ telomeres in ALT cells (16) and implicated in
hexameric helicase to facilitate conflicts between helping cells deal with replicative stress (29).
9
TRF1 and TRF2 are known to bind exclusively to replication of the telomere repeats. It is
double-stranded telomeric repeat DNA as homo- conceivable that certain DNA helicases, acting
dimers, resulting in highly stable protein-DNA with RPA, dislodge proteins of the shelterin
interactions (39,44,61). Here, we also observed complex once they have fulfilled their functions to
potent inhibition of FANCJ helicase activity by repress telomere fragility, providing the
the biologically relevant interaction of TRF1 or opportunity for the telomere sequence to be copied
TRF2 with the human telomeric sequence. These faithfully to preserve chromosome ends. The
findings may be in accord with biochemical data helicase-RPA interaction would provide a higher
showing that TRF2 protects Holliday Junction level of coordination between events associated
DNA from unwinding by the human RecQ with telomere capping by shelterin proteins and
helicase WRN in part by its ability to bind the telomeric DNA synthesis by the replication
telomeric sequence duplex arms (62). In the machinery.
current study, we demonstrated that the presence FANCJ or another DNA repair helicase
of RPA in the reaction mixture stimulated FANCJ (e.g., member of the RecQ family) implicated in
to dislodge TRF1 or TRF2 from the telomeric HR might utilize its motor ATPase function to
duplex and efficiently unwind the DNA substrate. displace Rad51 from single-stranded or double-
The ability of RPA to stimulate FANCJ in this stranded DNA. Indeed, the DNA helicases Srs2
D
capacity was specific based on our observations (54-57), RECQ5 (67), FBH1 (68) and FANCJ (69) o
w
n
that ESSB or the telomeric SSB Pot1 failed to have all been previously shown to strip Rad51 lo
a
enhance FANCJ activity on the TRF1-bound DNA from single-stranded DNA. However, it is also de
d
substrate. RPA also stimulated FANCJ to displace possible that a DNA helicase may disrupt Rad51- fro
m
TRF1 or TRF2 from a forked duplex harboring double-stranded DNA complexes which would be h
four direct telomeric repeats, a substrate that inhibitory to DNA strand exchange (for review, ttp
would be expected to bind two homo-dimers under see (59)). For example, the yeast Srs2 helicase ://ww
w
TRF1- or TRF2-saturating conditions. The ability was shown to unwind double-stranded DNA .jb
of RPA to stimulate FANCJ displacement of covered with Rad51 (58). Based on our finding c.o
rg
BamHI-E111A, TRF1, or TRF2 from the forked that RPA modulates helicase-catalyzed protein b/
y
duplex substrates is compelling in light of an displacement in vitro, future studies should g
u
earlier observation that RPA does not promote address if RPA promotes DNA helicases to disrupt es
t o
BLM helicase to unwind a forked duplex toxic dead-end complexes such as Rad51 bound to n
N
mononucleosome substrate (63). In the future, it double-stranded DNA. This mechanism of o
v
e
will be of interest to test if RPA can stimulate regulation would be distinct from one in which C. m
b
e
FANCJ, or another DNA helicase with which it elegans DNA helicase HELQ-1 was reported to r 2
0
interacts, to unwind nucleosomal DNA in an promote the disassembly of RAD-51 from double- , 2
0
efficient manner to regulate chromatin state. stranded DNA in an ATP-independent manner 18
Both FANCJ and BLM were found to be (70).
specifically bound to telomeric DNA in telomerase Cells from FA-J patients display
negative cells engaged in the ALT maintenance chromosomal instability and sensitivity to agents
pathway (16), raising the possibility that FANCJ that impose replication stress (29). FANCJ was
and BLM, two helicases which are known to observed to show an enhanced association with
physically and functionally interact (29), may chromatin as cells progress through S phase (71),
cooperate to unwind telomeric DNA structures or and co-localize with RPA after DNA damage (9).
displace proteins bound to telomeres. Mammalian Our current work suggests that FANCJ performs
telomeres display a fragile phenotype that is its catalytic functions in HR repair or replication
suppressed by TRF1, which is required for restart in vivo on chromatinized DNA substrates
efficient replication of TTAGGG repeats to which with its protein partner RPA, which promote
it binds (64,65). However, it was also observed FANCJ’s ability to displace proteins bound to
that DNA helicases (BLM (65), RTEL1 (66)) are duplex DNA. Although the emphasis of this work
required to repress the fragile-telomere phenotype, has been on the RPA-dependent ejection of protein
suggesting that specialized DNA unwinding from DNA by FANCJ, it is reasonable that
functions are necessary for timely and efficient FANCJ may help to remodel proteins on DNA to
10