(Received for publication, March 12, 1996, and in revised form, October 10, 1996)
From the Unit of Biochemistry, the Bruce Rappaport Faculty of Medicine, Technion, Israel Institute of Technology, P.O. Box 9649, Haifa 31096 Israel
Telomeres of vertebrate chromosomes terminate
with a short 5-d(TTAGGG)-3
single-stranded overhang that can form
in vitro tetrahelical structures. Here we describe a new
protein from rat hepatocyte nuclei designated quadruplex
telomere-binding protein 42 (qTBP42) that tightly binds
5
-d(TTAGGG)n-3
and 5
-d(CCCTAA)n-3
single-stranded and tetraplex forms of
5
d(TTAGGG)n-3
. The thermostable qTBP42 was isolated
from boiled nuclear extracts and purified to near homogeneity by
successive steps of column chromatography on DEAE-cellulose,
phosphocellulose, and phenyl-Sepharose. A subunit molecular size of
42.0 ± 2.0 kDa was determined for qTBP42 by Southwestern blotting
and SDS-polyacrylamide gel electrophoresis of the protein and its UV
cross-linked complex with labeled telomeric DNA. A native size of
53.5 ± 0.9 kDa, estimated by Superdex© 200 gel
filtration, suggests that qTBP42 is a monomeric protein. Sequences of
five tryptic peptides of qTBP42 contained motifs shared by a mammalian
CArG box-binding protein, hnRNP A/B, hnRNP C, and a human
single-stranded telomeric DNA-binding protein. Complexes of qTBP42 with
each complementary strand of telomeric DNA and with quadruplex forms of
the guanine-rich strand had 3.7-14.6 nM dissociation
constants, Kd, whereas complexes with double-stranded telomeric DNA had up to 100-fold higher
Kd values. By associating with tetraplex and
single-stranded telomeric DNA, qTBP42 increased their heat stability
and resistance to digestion by micrococcal nuclease.
Telomeres, the DNA-protein structures at the end of linear
eukaryotic chromosomes, form protective caps that shield chromosome termini against degradative processes or fusion with other chromosome ends (1-4). The evolutionarily conserved nucleotide sequence of
telomeric DNA consists of short nucleotide sequences repeated in
tandem. All vertebrates, slime molds, filamentous fungi, and Trypanosoma have a repeated 5-d(TTAGGG)-3
sequence of the
telomeric strand oriented 5
to 3
toward the chromosome terminus, the
"G-strand". The 3
-terminal G-strand stretch ends with an unpaired
12-16-nucleotide-long overhang (1-4). This single-stranded tract can
form in vitro under physiological conditions hairpin (5-7)
or unimolecular or bimolecular tetrahelical structures (5-13), which
may function in telomere transactions.
The telomere hypothesis of cellular senescence and tumorigenesis states that progressive shortening of telomeres in somatic cells leads to cessation of their division and to cellular aging (16). In contrast, maintenance of a stable telomere length in germ line and cancer cells is associated with their infinite division and immortality. This hypothesis is supported by evidence showing that whereas telomeric DNA becomes progressively shortened in the course of the aging of diverse somatic cells, its length remains stable in infinitely dividing stem and tumor cells (17-21). The G-strand of telomeric DNA is extended by telomerase, a ribonucleoprotein enzyme whose complementary cytosine-rich RNA component serves as a template for the synthesis of the G-strand (3, 4, 22). Whereas telomerase activity is undetectable in various somatic tissues and in dividing primary cells, many cancer cells retain an active telomerase (Refs. 18-21 and 23; reviewed in Ref. 17). More recent results suggest, however, that telomerase might not be the sole factor that determines telomere length. Thus, some tumor cells whose telomere length remains stable have no measurable telomerase activity (24), and conversely, an active telomerase was detected in normal human (25) and mouse cells (26). Further, the removal of 50-200-nucleotide-long segments of telomeric DNA with each round of replication in somatic cells (27, 28) suggests that, in addition to the loss of telomerase activity, termini of telomeric DNA may also be shortened by exonucleolytic attack.
Stabilization or destabilization of telomeric DNA may be affected by
telomeric DNA-binding proteins that have been identified in diverse
species. Proteins such as a monomeric 51-kDa polypeptide from
Euplotes crassus (29), a 34-kDa Chlamydomonas
protein (30), and a Xenopus protein (31) bind tightly to
single-stranded telomeric DNA. Another group of proteins bind to or
accelerate the formation of tetraplex structures of telomeric DNA. A
heterodimeric protein from Oxytricha nova consisting of a
56-kDa and a 41-kDa
subunit, dimerizes in the presence of
single-stranded telomeric DNA and binds to it cooperatively (32).
Without detectably binding to quadruplex DNA, the
subunit
facilitates tetraplex formation 105- to
106-fold (33). The yeast KEM1 protein binds to tetraplex
DNA and cuts a single-stranded sequence 5
to the quadruplex structure (34). Shortened telomeres and a senescent phenotype of KEM1 null
mutants suggest that the formation of G-strand tetrahelix and KEM1
nuclease activity are required for telomeric DNA extension (35). Two
additional yeast proteins, the 42-kDa G4p1 (36) and the 32-kDa G4p2
(37), bind with similar affinities to parallel and antiparallel
tetraplex DNA. The yeast protein RAP1 binds duplex telomeric DNA (38),
is required for telomere maintenance (39), and associates with and
promotes the formation of parallel stranded quadruplex telomeric DNA
(40, 41). A protein from Tetrahymena thermophila also binds
specifically to parallel-stranded tetraplex telomeric DNA (42).
Proteins of a third class such as a 10-kDa polypeptide from
Physarum polycephalum (43) and hTRF, a 50-kDa HeLa cell
protein (44-46) associate specifically with the duplex region of
telomeric DNA.
In searching for a mammalian cell protein that interacts with terminal telomeric DNA sequences, we identified in rat hepatocytes a protein that binds tightly both single strands of telomeric DNA and tetraplex forms of the guanine-rich strand. Here we describe the purification and characterization of this protein and report its effects on the bound DNA.
[32P]ATP (~3000
Ci/mmol) and molecular mass RainbowTM marker proteins were
from Amersham Corp. Synthetic DNA oligomers, listed in Table
I, were purchased from Operon Technologies. Midland Reagent supplied the HPLC-purified1 RNA
oligomer r(UUAGGG)4. The RNA oligomer rRAND was a gift of Dr. Ashwini Loeb (University of Washington). Sigma
provided boric acid, dithiothreitol (DTT), N-ethylmaleimide,
salmon sperm DNA, thymidine 3,5
-diphosphate, dimethyl sulfate,
leupeptin, aprotinin, benzamidine, phenylmethylsulfonyl fluoride;
Nonidet P-40, Sephadex G-50, phenyl-Sepharose, soybean trypsin
inhibitor (STI), trypsin, and micrococcal nuclease. Whatman supplied
DEAE-cellulose (DE-52); phosphocellulose (P-11), and DE-81 filter
paper. Bacteriophage T4 polynucleotide kinase was from Promega.
Acrylamide:bisacrylamide (19:1 or 30:1.2) was purchased from Amresco.
IBI provided Kodak autoradiographic film, urea, TEMED, bromphenol blue,
and xylene cyanol FF. Superdex© 200 gel filtration column
was the product of Pharmacia Biotech Inc. Gelman Sciences supplied
Biotrace polyvinylidene difluoride binding matrix membranes. Bio-Rad
provided reagents for SDS-polyacrylamide gel electrophoresis (SDS-PAGE)
and molecular weight protein standards.
|
Full-length DNA oligomers were purified by
electrophoresis through an 8 M urea, 15% polyacrylamide
denaturing gel (acrylamide:bisacrylamide, 19:1) as we described
previously (47). The purified DNA or RNA oligomers were 5-end-labeled
with 32P as described (48). Oligomers were maintained in
their single-stranded conformation when stored as a 0.25 µM solution in 1.0 mM EDTA, 10 mM
Tris-HCl buffer, pH 8.0 (TE buffer) and boiled immediately prior to
use. Double-stranded DNA was prepared by heating at 90 °C for 2 min
an equimolar mixture of complementary DNA oligomers (20 µM each in TE buffer) followed by slow cooling to room
temperature. Duplex DNA molecules were resolved from residual DNA
single strands by electrophoresis through a nondenaturing 15%
polyacrylamide gel (acrylamide:bisacrylamide, 30:1.2), and the
double-stranded DNA was eluted from an excised gel slice and isolated
as described (47). To prepare unimolecular tetraplex G
4 TeR DNA, a
solution of 0.25 µM TeR DNA in TE buffer was heated at
95 °C for 2 min and cooled rapidly to 4 °C, and NaCl was added to
a final concentration of 50 mM. A compact structure of TeR
DNA was formed in the presence of 50 mM NaCl and identified
by its rapid migration in a nondenaturing 15% polyacrylamide gel
relative to the mobility of bacteriophage M13 17-mer universal primer
DNA marker. Following UV cross-linking, this form also migrated rapidly
in a denaturing 8 M urea, 12% polyacrylamide gel (9). The
folded DNA structure was resistant to methylation with dimethyl
sulfate, indicating its stabilization by Hoogsteen hydrogen bonds (49).
To prepare G
2 TeR DNA, a bimolecular tetraplex form of oligomer TeR,
10 µM TeR2 DNA in TE buffer was heated at 95 °C for 2 min followed by incubation of the DNA at 37 °C for 20 h in the
presence of 1.0 M KCl. That a multimolecular DNA complex
was accumulated was shown by its retarded electrophoretic migration in
a nondenaturing 15% polyacrylamide gel relative to the mobility of
single-stranded oligomer TeR2. The bimolecular stoichiometry of the
electrophoretically retarded complex was demonstrated according to Sen
and Gilbert (50), using the short and long oligomers TeR2 and long
TeR2, respectively. A parallel G4 quadruplex form of oligomer Q was
prepared, and its tetramolecular stoichiometry was demonstrated, using
oligomers Q and long Q, as described (50).
The DNA binding activity of qTBP42 was monitored by
electrophoretic mobility shift assay as we described previously (47). Briefly, to assay for the binding of single-stranded TeR DNA, G4
quadruplex oligomer Q or double-stranded DNA, 0.2-2.0 ng of 32P-5-end-labeled DNA were incubated at 4 °C for 20 min
with 10-400 ng of purified or crude protein fraction in a 10-µl
final volume of buffer D (0.5 mM DTT, 1.0 mM
EDTA, 20% glycerol in 25 mM Tris-HCl buffer, pH 7.5). Gel
mobility shift electrophoresis was conducted as described previously
(47), and gels dried on DE-81 filter paper were exposed to x-ray film
or to a phosphor-imaging plate (Fuji). Amounts of free and qTBP42-bound
TeR DNA were deduced from phosphor-imaging quantification and the known
specific activity of the labeled DNA probe. One unit of qTBP42 DNA
binding activity was defined as the amount of qTBP42 that bound 0.05 ng
of single-stranded TeR DNA under the described standard conditions.
Binding of tetraplex G
2 TeR or G
4 TeR DNA was assayed as detailed
above except that 10 mM KCl or 50 mM NaCl,
respectively, was added to the DNA binding mixture to preserve the
quadruplex structure of the DNA, the 0.5 × TBE gel running buffer
contained 10 mM KCl or 50 mM NaCl, and electrophoresis was performed at 4 °C.
SDS-PAGE and silver staining of resolved protein bands were carried out as we described previously (47).
Southwestern analysis was conducted essentially according to Petracek
et al. (30) except that 1.6 µg of
32P-5-end-labeled TeR DNA were used and nonspecifically
adsorbed DNA was removed by washing the blots at 4 °C for 10 min,
first three times with 600 ml each of binding buffer that contained 0.25% Marvel nonfat dry milk and than three more times with 200 ml
each of binding buffer that contained 0.05% Nonidet P-40. The blots
were dried and exposed to autoradiographic film.
Protein qTBP42 was typically purified from 500-1000 g of liver tissue from adult rats. Salt extracts of non-histone nuclear proteins were prepared from isolated nuclei of hepatocytes as we described (52) except that the extraction buffer contained 0.4 M NaCl, 1 mM phenylmethylsulfonyl fluoride, 0.1 mM benzamidine, and 20 µg/ml each of STI, leupeptin, and aprotinin in buffer D. Preparation of protein extracts and every subsequent step of qTBP42 purification were conducted at 4 °C.
Electrophoretic mobility shift assays and Southwestern analysis showed that rat liver extracts contained a relatively heat-stable TeR DNA binding activity (see "Results"). Two alternative strategies were employed to purify this activity. In procedure 1, qTBP42 was purified to near homogeneity in a scheme that included boiling of the nuclear extract for 10 min as an initial step of purification. The binding activity that was recovered in the heat-resistant fraction was then further purified by successive steps of column chromatography. To ascertain that properties of the TeR DNA-binding protein were not modified by the initial heat treatment, it was also purified by an alternative method, procedure 2, in which the TeR DNA binding activity was resolved by steps of column chromatography without exposure of the protein to heat.
In procedure 1, the nuclear extract was heated at 100 °C for 10 min,
and denatured proteins were removed by ultracentrifugation at
80,000 × g for 35 min. TeR DNA binding activity that
was recovered in the supernatant was loaded onto a DE-52 column
equilibrated in buffer D at a ratio of 5.0 mg of protein/ml of packed
resin. The protein-loaded column was washed with a single packed column volume of the equilibration buffer, and bound proteins were eluted by a
10-column volume gradient of 0-500 mM NaCl in buffer D. Nonspecific adsorption of qTBP42 to glass and plastic was found to be
prevented by Nonidet P-40, which did not detectably decrease DNA
binding activity at detergent concentration of up to 5%. Thus, in this and in all of the following steps of chromatography, 40 fractions were
collected into Nonidet P-40 at a final concentration of 0.05%. Forty
fractions were collected and dialyzed overnight against ~200 volumes
of buffer D. To detect binding activity, parallel assays were conducted
using 32P-5-end-labeled single-stranded TeR DNA, G
4 TeR
DNA, or a G4 quadruplex form of oligomer Q as probes. In this and all
of the succeeding purification steps, electrophoretic mobility shift analysis revealed overlapping peaks of DNA binding activity with all
three probes. Binding activity was detected in fractions that were
eluted from the DE-52 column by 70-290 mM NaCl. These
fractions were pooled together, dialyzed overnight against ~200
volumes of buffer P (1.0 mM EDTA, 0.5 mM DTT,
20% glycerol in 80 mM KPO4 buffer, pH 8.5),
and loaded onto a P-11 column equilibrated in buffer P at a ratio of
2.0 mg of protein/ml of packed resin. The loaded column was washed with
a single packed column volume of buffer P, and resin-bound proteins
were eluted by a 10-column volume linear gradient of 80-350
mM KPO4 buffer, pH 8.5, that contained 1.0 mM EDTA, 0.5 mM DTT, 20% glycerol. Forty
fractions were collected and dialyzed overnight against ~200 volumes
of buffer D. Fractions eluted by 100-200 mM
KPO4 contained binding activity and were pooled together,
dialyzed overnight against ~200 volumes of buffer S (4.0 M NaCl, 1.0 mM EDTA, 0.5 mM DTT, in
25 mM Tris-HCl buffer, pH 7.5), and loaded at a ratio of
1.0 mg of protein/ml of packed resin onto a phenyl-Sepharose column equilibrated in buffer S. The loaded column was washed with a single
column volume of the equilibration buffer, and bound proteins were
eluted by a 10-column volume gradient of 4.0-0.0 M NaCl in 1.0 mM EDTA, 0.5 mM DTT, 25 mM
Tris-HCl buffer, pH 7.5. Forty fractions were collected and dialyzed
overnight against ~200 volumes of buffer D. DNA binding activity was
detected in fractions that were eluted from the phenyl-Sepharose column
by 0.5-0.0 M NaCl. These fractions contained a single
major 42-kDa protein band as revealed by SDS-PAGE (see "Results").
Typically, the concentration of protein in this fraction was lower than
25 µg/ml, and the DNA binding activity was progressively lost in the
course of its storage at
80 °C. The DNA binding activity was
stabilized by collecting eluted fractions into Nonidet P-40 and STI
protein stabilizer at final concentrations of 0.05% and 200 µg/ml,
respectively. The STI and Nonidet P-40-stabilized protein retained full
TeR DNA binding activity for at least 6 months when stored at
80 °C.
In purification procedure 2, salt-extracted nuclear proteins were first precipitated by ammonium sulfate, and TeR DNA binding activity was recovered in the 50-70% (NH4)2SO4 precipitate and was dialyzed overnight against ~200 volumes of buffer D. The dialyzed fraction was incubated with denatured salmon sperm DNA at 4 °C for 30 min at a protein:DNA ratio of 30:1 (w/w). The TeR DNA binding activity bound to the denatured DNA, which (due to its polyanionic nature) bound strongly to DEAE cellulose. The protein-DNA mixture was loaded onto a DE-52 column equilibrated in buffer D at a ratio of 5.0 mg of protein/ml of packed resin. Weakly bound proteins were eluted from the column by successive stepwise washes with a single packed column volume each of 50, 100, and 150 mM NaCl in buffer D. The TeR DNA binding activity that tightly associated with the bound denatured DNA was eluted by a single packed column volume of 225 mM NaCl in buffer D. Following overnight dialysis against ~200 volumes of buffer P, the binding activity was chromatographed on a column of P-11 as described above. TeR DNA binding activity was eluted from the column by 140-160 mM KPO4, and after overnight dialysis against ~200 volumes of buffer S, it was chromatographed on a column of phenyl-Sepharose as detailed above. TeR DNA binding activity was detected in fractions that were eluted from phenyl-Sepharose by 0.5-0.0 M NaCl.
Determination of the Amino Acid Sequence of qTBP42 PeptidesTo determine a partial amino acid sequence of qTBP42, a Coomassie Blue-stained SDS-PAGE-resolved band of qTBP42 was excised, and the protein was eluted from the gel and digested with trypsin. Resulting tryptic peptides were separated by reverse-phase HPLC, and the amino acid sequences of selected peptides were determined by a standard automated procedure.
Determination of the Amount of ProteinThe Bio-Rad protein assay kit was used to determine the amount of protein in qTBP42 fractions that did not contain STI protein stabilizer.
Activity that bound tetraplex and single-stranded forms of
the vertebrate telomeric sequence TeR, 5-d(TTAGGG)4-3
was
detected by electrophoretic mobility shift analysis in extracts of
non-histone nuclear proteins from rat hepatocyte. Southwestern analysis
indicated that nuclear extracts exhibited two TeR DNA-binding protein
bands of ~40 and ~78 kDa, but only the ~40-kDa polypeptide, here
designated qTBP42, was heat-stable, partially resisting heating of the
extracts at 100 °C for 10 min (Fig. 1A).
To purify qTBP42, the nuclear extracts were initially boiled for 10 min, and the denatured proteins were removed by ultracentrifugation.
Heat-resistant TeR DNA binding activity that was recovered in the
supernatant was further purified by successive steps of chromatography
on columns of DE-52, P-11, and phenyl-Sepharose. Overlapping peaks of
binding activity were obtained when single-stranded TeR DNA, TeR G
4
DNA, or a G4 form of oligomer Q were used as probes in electrophoretic
mobility shift analysis (data not shown). SDS-PAGE of proteins resolved by the successive steps of extract boiling and chromatography on
columns of DE-52 and P-11 revealed a progressive depletion of proteins
(Fig. 1B). Note that a major 42-kDa protein band became prominent in the P-11-purified fraction of qTBP42 (Fig. 1B).
The level of TeR binding activity in the P-11-resolved fractions was found to be directly related to the intensity of this silver-stained 42-kDa protein (results not shown). Phenyl-Sepharose hydrophobic column
chromatography was ultimately used to obtain a nearly homogeneous fraction of qTBP42. As seen in Fig. 2A, TeR
DNA binding activity was detected by mobility shift electrophoresis in
phenyl-Sepharose fractions 34-38. Covalent UV cross-linking of labeled
TeR DNA to phenyl-Sepharose-resolved proteins revealed a 52.5-kDa
protein-DNA complex in fractions 34-38 (Fig. 2B) or a
51.0-kDa complex with the G
4 tetraplex form of TeR DNA and a 49.5-kDa
complex with the G4 form of oligomer Q (results not shown). Similar
average complex sizes of 51.5, 51.0, and 48.0 kDa were obtained when
labeled TeR DNA, G
4 TeR, and G4 oligomer Q, respectively, were
UV-cross-linked to qTBP42 that was purified by procedure 2. As seen in
Fig. 2C, silver staining of SDS-PAGE-resolved proteins
revealed in fractions 34-38 a single major protein band of 42 kDa.
This protein also remained the only detectable major stained band when
no STI protein stabilizer was added to these phenyl-Sepharose-resolved
fractions (results not shown). The heat-stable ~40-kDa binding
protein detected by Southwestern analysis of nuclear extracts (Fig.
1A) as well as the 42-kDa size of the highly purified active
protein (Fig. 2, A and C) and the 46-53-kDa size
of its complex with DNA (Fig. 2B) strongly suggested that
the 42-kDa protein represented qTBP42. The yield of qTBP42 purified by
procedure 1 was low, but whereas 96% of the extract proteins became
denatured by boiling, 8-15% of the initial TeR DNA binding activity
was recovered in the supernatant.
Chemical-physical Properties of qTBP42
Some properties of qTBP42 are shown in Table II. The proteinaceous nature of qTBP42 was demonstrated by its inactivation by SDS or trypsin. Protein qTBP42 appeared to be relatively heat-stable such that purified qTBP42, which resisted 10 min of boiling in the nuclear extract, retained 60% of its activity after an additional 2 min at 100 °C and became nearly fully inactivated only after 60 min of boiling (Table II). The resistance of qTBP42 to digestion with micrococcal nuclease (Table II) suggested that it did not contain an essential nucleic acid component. An only slight decrease of qTBP42 activity in the presence of 8.5 mM N-ethylmaleimide (Table II) indicated that reduced protein sulfhydryl groups were not directly involved in its interaction with DNA.
|
Protein qTBP42 migrated on SDS-PAGE as a 42.0 ± 2.0-kDa polypeptide (n = 6). Protein qTBP42 that was purified by procedure 1 or procedure 2 had apparent native molecular size of 53.5 ± 0.9 kDa (n = 3) or 55.0 kDa, respectively, as determined by Superdex 200© gel filtration. The different sizes of denatured and native qTBP42, ~42 versus ~54 kDa, respectively, could be due to the contribution of the shape of the protein molecule to its migration through the Superdex 200© gel. The ~40-kDa apparent size of qTBP42 in a Southwestern blot (Fig. 1A) and the 48-52.5-kDa mass of its UV cross-linked complexes with different DNA probes (see above) indicated that in its native form qTBP42 was most probably a 42-kDa monomeric protein.
Specificity of DNA Binding by qTBP42DNA sequence and
structure specificity of binding to qTBP42 were first assessed by
binding competition between 32P-5-end-labeled TeR DNA and
excess of different unlabeled DNA sequences. A 50-fold molar excess of
unlabeled d(A)16, d(G)16, or the oligomer
5
-d(A2TC2GTCGTCGAGCAGAGT2)-3
failed to detectably compete with the binding of either single-stranded
or quadruplex G4 oligomer Q (results not shown). Essentially identical
results were obtained for qTBP42 that was purified by either procedure 1 or procedure 2 (data not shown). Several DNA sequences and structures did compete, however, with TeR DNA at variable efficiencies. To accurately determine the relative affinity of qTBP42 for the competing DNA species, the dissociation constants, Kd, of
their complex with the protein were determined. Fig. 3
shows a typical equilibrium binding measurement of the binding of
qTBP42 to G
4 TeR DNA. A constant amount of the P-11 fraction of qTBP42
purified by procedure 1, was incubated at 4 °C for 20 min with
increasing amounts of 32P-5
-end-labeled G
4 TeR DNA, and
the protein-DNA complex was then separated from unbound DNA by mobility
shift electrophoresis (Fig. 3A). Amounts of bound and free
DNA were determined by phosphor-imaging measurements of the respective
bands. The dissociation constant, Kd, was inferred
from the negative reciprocal of the slope of a Scatchard plot of the
results (Fig. 3B). Table III compiles Kd values determined for complexes of qTBP42 with
DNA sequences that measurably competed with the TeR DNA on its binding to the protein. The presented results indicate that qTBP42 bound the
guanine-rich telomeric single strand with nanomolar binding affinity.
Interestingly, the cytosine-rich DNA strands anti-TeR or TeR C DNA,
which contained four and three clusters, respectively, of three
cytosine residues each, formed complexes with qTBP42 that had
Kd values of 3.7-4.8 × 10
9
mol/liter, values similar to qTBP42 complexes with TeR DNA (Table III).
To ascertain that these cytosine-rich oligomers remained single-stranded rather than tetraplex structures, they were
electrophoresed through a nondenaturing 10% polyacrylamide gel at pH
8.0 with or without added salt. Control TeR DNA formed under the same
experimental conditions a rapidly migrating unimolecular tetraplex in
an alkali ion-dependent fashion, and a UV-cross linked TeR
oligomer remained in a compact form with or without salt (see
"Experimental Procedures"). By contrast, neither anti-TeR nor TeR C
DNA exhibited any change in its electrophoretic mobility, and both
migrated as unfolded structures with or without the presence of ions
(results not shown). A possible mechanism for the observed association
between qTBP42 and the cytosine-rich sequences might have been the
pairing of these sequences with residual complementary guanine-rich DNA
that could have remained bound to the protein. To examine this
possibility, phosphocellulose-purified qTBP42 was incubated at 37 °C
for 10 min with 16.6 ng/µl of micrococcal nuclease in the presence of 3.3 mM CaCl2. The nucleolytic digestion was
terminated by the addition of EGTA and thymidine 3
,5
-diphosphate to a
final concentration of 8.0 and 0.45 mM, respectively, and
the efficiency of anti-TeR and TeR C binding to the protein was
assessed by electrophoretic mobility shift analysis. Although control
unbound TeR DNA was completely digested under the described conditions,
the binding of both anti-TeR and TeR C DNA by qTBP42 remained unaltered
after exposure of the protein to the nuclease (data not shown). Hence, the binding of the cytosine-rich telomeric DNA to qTBP42 appeared not
to be due to interaction between this DNA strand and a
protein-associated complementary sequence. That qTBP42 displayed a
preference for telomeric sequences was demonstrated by the ~85-fold
higher Kd value of its complex with the
single-stranded guanine-rich oligomer Q (Table III). Complexes of
qTBP42 with the G
4 unimolecular and the G
2 bimolecular quadruplex
structures of TeR DNA, as well as complexes with the tetramolecular
quadruplex form of oligomer Q, also had nanomolar Kd
values similar to those of single-stranded TeR, anti-TeR, and TeR C
DNA. In contrast to its tight binding of tetraplex telomeric DNA,
qTBP42 bound poorly to double-stranded telomeric DNA with or without a
single-stranded d(TTAGGG)2 overhang. Complexes that qTBP42
formed with these molecules had Kd values
40-73-fold higher than the rate constants of complexes of qTBP42 with
single-stranded TeR DNA or G4 oligomer Q (Table III). It is noteworthy
that qTBP42 bound with a similar nanomolar affinity to the
single-stranded telomeric sequence RNA oligomer rTeR as to the
divergent RNA sequence rRAND (Table III). The sequence-independent binding of RNA by qTBP42 and its similar binding affinities for RNA and
DNA sequences are in contrast to the highly specific binding of
telomeric RNA sequence by previously described mammalian hnRNPs (see
Ref. 59 and "Discussion"). Last, we note that qTBP42 did not bind
exclusively to the described DNA sequences and tetraplex structures.
Competition and equilibrium binding measurements revealed that qTBP42
also bound to d(T)16 and d(C)17, forming
complexes that had Kd values of 8.0 and 28.0 × 10
9 mol/liter, respectively. To test whether the observed
DNA sequence and structure binding preferences of qTBP42 were not
artificially modified by the heat treatment of proteins during
purification by procedure 1, Kd values were measured
for complexes that were formed between DNA ligands and qTBP42 that was
purified by procedure 2. It was found that the rate constants were very similar for heat-treated and untreated binding protein. For instance, Kd values for complexes of unheated qTBP42 with TeR
DNA, anti-TeR, and the G4 form of oligomer Q were 1.75 × 10
9, 30.0 × 10
9, and 7.0 × 10
9 mol/liter, respectively, values close to those
obtained with complexes of heat-treated qTBP42 and the same DNA ligands
(Table III).
|
To investigate the minimal sequence requirements for the binding of
guanine-rich telomeric DNA strand by qTBP42, protein purified by
procedure 1 was incubated at 4 °C for 20 min with
32P-5-end-labeled single-stranded TeR DNA in the presence
of an increasing molar excess of either unlabeled TeR1 or long TeR1 DNA
that contained a single d(TTAGGG) repeat unit or with an excess of
either unlabeled TeR2 or long TeR2 DNA, which contained two d(TTAGGG)
repeats each (Table I). Whereas 32P-5
-end-labeled TeR DNA
was dislodged from its complex with qTBP42 in direct proportion to the
molar excess of the unlabeled TeR DNA homologous competitor, both TeR1
and long TeR1 competitor oligomers were ineffective in displacing TeR
DNA from its complex with qTBP42 (Fig. 4). Competing
TeR2 DNA that contained two telomeric DNA repeat units dislodged TeR
DNA from its complex with qTBP42 at an efficiency lower than the
homolog TeR sequence (Fig. 4). Long TeR2 DNA competitor, which also
contained two d(TTAGGG) clusters, was more efficient in displacing TeR
DNA, probably because of its greater length (Fig. 4). These results
suggest, therefore, that whereas a single d(TTAGGG) cluster was not
sufficient for qTBP42 binding, two repeats of this sequence allowed
binding at an efficiency approaching that of TeR DNA, which contained
four repeats.
Association with qTBP42 Stabilizes Tetraplex and Single-stranded TeR DNA
To inquire whether the stability of quadruplex TeR DNA
was affected by its association with qTBP42, we compared the rate of heat denaturation of naked and protein-bound bimolecular tetraplex G2
TeR DNA. In a procedure originally applied to the thermostable quadruplex DNA-binding protein QUAD (53), free or qTBP42-bound 32P-5
-end-labeled G
2 TeR DNA was incubated at 50 °C
for different periods of time, incubation was terminated, the protein
residue was removed by the addition of SDS to a final concentration of 0.25%, and the tetraplex and single-stranded forms of TeR DNA were
resolved by electrophoresis through a nondenaturing polyacrylamide gel.
As seen in Fig. 5, unbound bimolecular tetraplex G
2 TeR DNA became 50% denatured into single strands after 4.5 min at 50 °C. By contrast, the tetrahelical structure of the qTBP42-bound G
2 TeR DNA was almost fully retained after 8 min at 50 °C (Fig. 5),
and only 25% of the tetrahelix became denatured after 15 min at
50 °C (results not shown). Parallel measurements disclosed that the
amount of protein-DNA complex in samples not exposed to SDS was only
slightly diminished during a similar heat treatment (results not
shown). Additional experiments showed that G
2 TeR DNA and the G4
tetramolecular quadruplex form of oligomer Q were also stabilized by
qTBP42 at 37 and 60 °C, respectively (data not shown).
To further inquire whether qTBP42 affected the stability of the bound
telomeric DNA, unbound single-stranded TeR DNA or G4 TeR DNA or their
complexes with qTBP42 were exposed to 0.75 ng/µl micrococcal nuclease
at 20 °C for different lengths of time. Following termination of the
nucleolytic digestion by the addition of SDS to a final concentration
of 0.25%, the DNA samples were electrophoresed through a 6%
nondenaturing polyacrylamide gel to separate the intact DNA oligomer
from its digestion products, which migrated at the front of the gel.
Fig. 6A shows that under the experimental conditions employed, unbound single-stranded TeR DNA became completely digested within 30 s. By contrast, 50 or 25% of the qTBP42-bound TeR DNA remained unbroken after exposure to the nuclease for 1 or 2 min, respectively (Fig. 6A). Similarly, unbound G
4 TeR DNA became fully digested 15 s after being exposed to micrococcal nuclease, whereas 50 or 25% of the protein-associated tetraplex DNA
remained intact after 30 or 60 s of digestion, respectively (Fig.
6B). To ascertain that the observed protection of TeR G
4 and TeR DNA by qTBP42 was a direct result of the interaction between the DNA ligands and their binding protein, the kinetics of nuclease digestion was examined for the oligomer d(CCAGA16G), which
was not detectably bound by qTBP42. Under the described experimental conditions, d(CCAGA16G) was digested at the same kinetics
with or without added qTBP42 (results not shown). Hence, qTBP42 by itself appeared not to have interfered with the nuclease action, and
the observed protection of TeR G
4 and TeR DNA was due to their binding
to qTBP42.
Protein qTBP42 Possesses Sequence Motifs Shared by Other Mammalian DNA- and RNA-binding Proteins
Amino acid sequences of five
tryptic peptides of qTBP42 were determined (see "Experimental
Procedures"). A computerized search of GenBankTM revealed
close homology between the amino acid sequence of every tested qTBP42
peptide and sequences of several known nucleic acid-binding proteins.
The amino acid sequences of qTBP42 peptides III and V were closely
homologous to two consensus RNA binding motifs (RNP motifs) shared by
many RNA-binding proteins. Peptide V of qTBP42, GFGFILF(K), displays
similarity to the consensus sequence of RNP1,
(K)VX
; and qTBP42 peptide III,
(K)SFVGGL(S), corresponds to the consensus sequence RNP2,
L
N
(54, 55). More
specifically, the sequences of peptide V and peptide III of qTBP42,
respectively, were found to be fully or nearly homologous to the RNP1
or RNP2 motifs, respectively, of the human RNA-binding protein hnRNP
A/B (Table IV and Ref. 56). Additional amino acid
sequences of hnRNP A/B shared extensive similarity with qTBP42 peptides
I, II, and IV (Table IV). Although the sequence of qTBP42 peptides I,
II, and IV was highly similar to a partial sequence deduced from a cDNA fragment of rat hnRNP C (Table IV and Ref. 57) they were not
identical to each other (Table IV). Hence, rat qTBP42 cannot be equated
with the known hnRNP C from this species. The highest degree of
sequence homology was noted for peptides of qTBP42 and of the Mus
musculus DNA-binding protein CBF-A, which specifically binds to
CArG box motifs in myogenic cells (58). As shown in Table IV, qTBP42
peptides II, IV, and V displayed 100% sequence homology with CBF-A,
whereas qTBP42 peptides I and III differed from CBF-A by only a single
residue each out of total peptide lengths of 14 and 8 amino acids,
respectively (Table IV). Last, qTBP42 peptide sequences I, II, and V
were found to closely resemble, but not to be identical with, the
sequences of tryptic peptides derived from a group of HeLa cell nuclear
proteins, B39, B41, and B37, which bind both the pre-mRNA 3
-splice
site r(UUAG/G) and single-stranded telomeric DNA d(TTAGGG)n
(Table IV and Ref. 59).
|
The new rat telomeric DNA-binding protein qTBP42 appears to differ from every other protein that was found to bind telomeric DNA in vitro. Telomeric DNA-binding proteins from a wide variety of organisms fall into two classes, based on their binding preference for double stranded or single-stranded DNA. Some of these proteins were also found to associate with tetrahelical telomeric DNA or to promote its formation.
The relatively low affinity of qTBP42 for double-stranded telomeric DNA, with or without a guanine-rich single-stranded overhang (Table III) clearly distinguishes it from double-stranded telomeric DNA-binding proteins such as a P. polycephalum protein (43) or the HeLa cell hTRF protein (44, 45).
A distinct amino acid sequence and characteristic physical properties
and DNA binding preferences distinguish qTBP42 from known
single-stranded or tetraplex telomeric DNA-binding proteins from a
variety of sources. Despite their similar subunit size and their
ability to bind tetraplex DNA, qTBP42 and the yeast 42-kDa telomeric
DNA-binding protein G4p1 differ by their native size, their lack of
shared amino acid sequence (Table IV and Ref. 36), and their different
relative affinities for single- or double-stranded guanine-rich DNA and
for tetraplex G4 DNA (36). The ~32-kDa yeast tetraplex DNA-binding
protein, G4p1, does not share an amino acid sequence homology with
qTBP42 and fails to bind single- or double-stranded guanine-rich DNA
(37). The major telomeric DNA binding protein in yeast RAP1 (60) is
distinguished from qTBP42 by its 92.5-kDa molecular mass, lack of amino
acid sequence homology, and binding preference for duplex telomeric DNA. Protein qTBP42 also differs from an E. crassus 51-kDa
protein that binds the guanine-rich telomeric DNA strand but fails to bind the complementary cytosine-rich sequence (61). The O. nova heterodimer of and
subunits binds specifically to the
single-stranded overhang
d(T4G4T4G4) of each
macronuclear terminus and protects it from nucleolytic attack (62).
However, while promoting the formation of quadruplex DNA, the
subunit, unlike qTBP42, does not measurably bind to it (34). A 34-kDa
GBP protein from Chlamydomonas binds a d(TTTTAGGG)n
single strand, but in contrast to qTBP42, it fails to associate with
the complementary cytosine-rich DNA sequence or with
r(UUUUAGGG)n (30). The T. thermophila TGP protein
binds a parallel stranded G4 tetraplex form of the sequence
d(T2G4)4, but in contrast to
qTBP42, it does not recognize single-stranded
(T2G4)4, duplex
d(T2G4)4/(C4A2)4,
or r(U2G4)4 (42). Unlike qTBP42
(Table III) the Xenopus laevis 50-kDa XTEF protein binds
preferentially a single-stranded d(TTAGGG) overhang at the 3
-end of
duplex telomeric DNA but does not bind tetraplex telomeric DNA
(31).
In addition to its ability to bind tightly single-stranded and tetraplex telomeric DNA, qTBP42 associates at similar nanomolar affinity with the RNA sequences rTeR and rRAND (Table III). In addition, qTBP42 shares an extensive amino acid sequence homology with mammalian hnRNP A/B and hnRNP C (Table IV), and it contains amino acid tracts similar to the consensus motifs RNP1 and RNP2 of RNA-binding proteins (54, 55). Further, qTBP42 displays amino acid sequences that are closely similar but not identical to sequences identified in human proteins that bind both the pre-mRNA splice junction sequence r(UUAG/G) and the telomeric sequence d(TTAGGG)n (Ref. 59 and Table IV). These human cell proteins are distinguished from qTBP42 by small differences in their amino acid sequence as well as by their complete heat inactivation after 5 min at 70 °C, their strict sequence specificity of binding r(UUAGGG)n and d(TTAGGG)n, and their clear preference for the binding of the RNA over the DNA oligomer (59). Another hnRNP species that does not share sequence homology with qTBP42 was also shown to bind RNA more tightly than DNA (63, 64). It is notable, however, that qTBP42 displays its closest sequence homology with the mouse CBF-A protein that binds the muscle-specific CArG box motif as well as single-stranded DNA (Table IV and Ref. 58).
The ability of qTBP42 to efficiently form complexes with both RNA and DNA (Table III) and the amino acid sequence homology of this protein with both DNA- and RNA-binding proteins (Table IV) raise questions about its primary in vivo nucleic acid binding target. Numerous nucleic acid-binding proteins interact with both RNA and DNA or fulfill more than a single function. For instance, the yeast RAP1 protein is essential for telomere maintenance but it also functions as either an activator or repressor of transcription (14). A multifunctional protein is the yeast G4p2 protein that binds tetraplex DNA but also participates in protein kinase-mediated signal transduction pathways and in cell cycle progression (37). Further, the G4p1 (36) and G4p2 proteins (37) bind tetraplex DNA and RNA structures at a similar efficiency. Likewise, some RNA-binding hnRNPs also bind telomeric DNA (59, 63, 64). Hence, as suggested for these proteins, qTBP42 might also function in transactions of both RNA and telomeric DNA. A potential role for qTBP42 in telomere metabolism might be the binding and stabilization of the telomeric DNA strands as they separate to allow their extension by telomerase and DNA polymerase. Another potential function of qTBP42 could be the stabilization of a tetraplex structure of the guanine-rich strand overhang. It was proposed that unimolecular quadruplex telomeric DNA structures prevent overextension of telomeric DNA by telomerase (15) and that multimolecular tetraplex telomeric synapses participate in the pairing of meiotic homolog chromosomes (51). The stabilization by qTBP42 of tetraplex conformations of telomeric DNA could thus contribute to one or more of these functions.
We thank the Technion Protein Research Center (Professor Arie Admon) for skillfully conducting peptide microsequencing.