| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
(Received for publication, July 5, 1995; and in revised form, August 30, 1995) From the
Expansion of a d(CGG)
Fragile X syndrome is an inherited, X-linked dominant mental
retardation disorder affecting about one male in 1500 and one female in
2500 (Webb, 1989; Richards and Sutherland, 1992). This syndrome is
frequently associated with a folate-sensitive fragile site, Xq27.3, on
chromosome X of cells of affected males (Sutherland, 1977). The fragile
X syndrome is characterized by a substantial expansion of a d(CGG)
trinucleotide repeat located in the 5`-untranslated region of a
housekeeping gene, FMR1, which was identified at a locus
coincident with the Xq27.3 breakpoint (Verkek et al., 1991).
Whereas normal individuals have 2-50 copies of the d(CGG)
sequence, the trinucleotide is amplified in affected subjects to
>200-2000 copies (Fu et al., 1991; Kremer et
al., 1991; Oberléet al., 1991;
Nakhahori et al., 1991; Verkek et al., 1991; Yu et al., 1991). Expansion of the d(CGG) repeat is accompanied
by methylation of the FMR1 promoter and of the amplified
trinucleotide tract (Bell et al., 1991; Vincent et
al., 1991; Pieretti et al., 1991; Luo et al.,
1993). Subsequent to d(CGG) expansion and hypermethylation, the FMR1 gene becomes transcriptionally silent (Pieretti et
al., 1991; Hansen et al., 1992; Sutcliffe et
al., 1992), and the replication of a chromosomal segment spanning
The
molecular mechanisms that govern d(CGG) The folding of exposed expanded single strand
runs of d(CGG)
Thymidine residues in oligomers that contain thymidine clusters were
probed by KMnO
Figure 3:
Formation by UV-irradiation of covalently
cross-linked forms of the H3T3 oligomer. End-labeled H3T3 DNA was
heat-denatured, annealed, and irradiated by UV light in the absence or
presence of 200 mM of the indicated salt as described under
``Experimental Procedures.'' Shown is an autoradiogram of 12%
polyacrylamide denaturing gel electrophoresis of control unirradiated
and UV irradiated H3T3 oligomer samples. Arrows mark the band
of rapidly migrating cross-linked DNA.
Figure 1:
Electrophoretic migration
of d(CGG)
The rapidly migrating forms of all of the examined
d(CGG) That
d(CGG)
Figure 2:
Absorbance thermal denaturation of
d(CGG)
To further demonstrate the formation of compact forms
of d(CGG)
Figure 4:
Modification by DEPC of
d(CGG)
Telomeric
DNA fragments contain clusters of thymidines (Guo et al.,
1993; Murchie and Lilley, 1994) or of thymidine and adenine residues
(Murchie and Lilley, 1994) interspersed between tracts of contiguous
guanines. Folding of these sequences into tetraplex structures involves
the looping-out of the thymidine or thymidine-adenine runs. Adenine
residues were also found to be interspersed within the d(CGG)
Figure 5:
Modification by DEPC and KMnO
Figure 6:
Modification by DEPC and KMnO
Figure 7:
Methylation-protection analysis of
d(CGG)
The substantial expansion of a d(CGG) trinucleotide repeat
within the 5` exon of the FMR1 gene in fragile X syndrome
cells (Fu et al., 1991; Kremer et al., 1991;
Oberléet al., 1991; Nakahori et
al., 1991; Verkerk et al., 1991; Yu et al.,
1991) and the hypermethylation of this tract and of the FMR1 gene promoter (Bell et al., 1991; Verkerk et
al., 1991; Yu et al., 1991), entail transcriptional
silencing of FMR1 (Pieretti et al., 1991; Hansen et al., 1992; Sutcliffe et al., 1992) and a delay in
the replication of a DNA region that encompasses FMR1 (Hansen et al., 1993). Illumination of the molecular basis for the
expansion of the d(CGG) In this communication, we
demonstrate that short stretches of d(CGG) That the compact structures of d(CGG) The gel mobility and chemical probing evidence provided
here, which indicates the facile formation of hairpin structures by
d(CGG) Based on the location of the looped-out sector in d(CGG)
Figure 8:
Models for the back-folded hairpin forms
of d(CGG)
As indicated in Fig. 8B, the H3T3 oligomer can assume either one of two
alternative folded structures. Both possible hairpin forms are
maintained by eight guanine-guanine pairs. However, whereas
conformation I is linear, the central loop in form II serves as a hinge
for a parallel arrangement of two duplex arms (Fig. 8B). Form II is perhaps less likely to exist than
conformation I since the former could easily convert into a tetraplex
that, as evidence indicates (Fig. 7), is not generated. The
propensity of the d(CGG) Folding of d(CGG)
Volume 270,
Number 48,
Issue of December 1, 1995 pp. 28970-28977
©1995 by The American Society for Biochemistry and Molecular Biology, Inc.
Nucleotide
Repeats Readily Fold Back to Form Unimolecular Hairpin Structures (*)
run within the
5`-untranslated region of the X-linked human gene FMR1 blocks FMR1 transcription, delays its replication, and precipitates
the fragile X syndrome. We showed previously that d(CGG)
tracts aggregate into interstrand tetrahelical complexes
(Fry, M., and Loeb, L. A.(1994) Proc. Natl. Acad. Sci. U. S. A. 91, 4950-4954). Here we show that these sequences also form
under physiological conditions in in vitro unimolecular
hairpin structures. Folding is demonstrated by temperature-dependent
mobility of d(CGG)
oligomers in a nondenaturing
polyacrylamide gel, by UV-hyperchromicity of thermally denaturing
oligomers, and by UV cross-linking of compact forms of d(CGG)
runs interspersed by thymidine clusters. That the compact
d(CGG)
structures are unimolecular is suggested by
their zero-order kinetics of formation. Diethyl pyrocarbonate
modification reveals a single, 4-5 residue-long central or
epicentral unpaired loop in folded d(CGG)
oligomers. The position of this loop remains unchanged by
insertion of thymidine clusters into 15- or 33-mer d(CGG) tracts as
indicated by KMnO
probing of unpaired thymidines. The
presence of a single loop in folded d(CGG) oligomers and the accessibility of every guanine to dimethyl
sulfate modification suggest that they are hairpin and not tetraplex
structures. Modeling indicates that different d(CGG)
hairpins are stabilized by guanine-guanine Hoogsteen hydrogen
bonds or by Hoogsteen and Watson-Crick bonds. If formed in
vivo, d(CGG)
hairpins could entail slippage
and trinucleotide expansion during replication and could obstruct FMR1 gene transcription and replication.
150 kb 5` and 34 kb 3` from the d(CGG) stretch is delayed (Hansen et al., 1993).
amplification and hypermethylation and that link d(CGG)
expansion to the suppressed transcription of FMR1 and to its
delayed replication are not known. In addressing these problems, we
showed recently that model d(CGG)
oligomers form
under physiological conditions a tetramolecular four-stranded structure
in a time-dependent and DNA concentration-dependent kinetics. We also
found that the multimolecular tetraplex structure of short
d(CGG)
tracts is stabilized by 5-methylation of
the cytosine residues (Fry and Loeb, 1994). The facile formation of
tetramolecular complexes by d(CGG)
raised the
hypothesis that single stranded stretches of this tract may also
generate unimolecular hairpin or tetraplex structures (Fry and Loeb,
1994). In this communication we use gel mobility analysis and chemical
probing to demonstrate that d(CGG)
oligodeoxynucleotides can fold back at a physiological range
of salt concentrations, temperatures, and pH values to form hairpin
structures. In contrast to the second-order kinetics of formation of
tetramolecular quadruplex d(CGG)
(Fry and Loeb,
1994), hairpin structures of this sequence are formed at a zero-order
kinetics. Our evidence for hairpin formation by d(CGG)
is sustained by reports by Gacy et al.(1995) and
Chen et al.(1995), which were published as this work was
completed and that provide NMR evidence to also show the formation of
hairpin structures by d(CGG)
and by other
trinucleotide tracts.
during the replication of fragile X
cell DNA could entail slippage and trinucleotide expansion.
Furthermore, guanine-rich tracts of RNA or DNA that fold back to
generate a hairpin (Christiansen et al., 1994) or tetraplex
structure (Christiansen et al., 1994; Woodford et
al., 1994) were shown to arrest DNA synthesis in vitro.
If formed in vivo along single strand stretches, back-folded
d(CGG)
structures may also obstruct DNA
transcription and replication. These structures might, therefore,
contribute to the observed blocking of FMR1 transcription and
to its delayed replication in fragile X cells.
Oligodeoxynucleotides
DNA oligomers listed under Table 1were prepared by Operon Technologies, Alameda, CA. To
purify these oligodeoxynucleotides, they were electrophoresed through a
15% polyacrylamide denaturing sequencing gel, and following detection
by UV absorption, DNA bands were cut out and the gel slices were
extracted by slow agitation overnight in 0.5 M ammonium
acetate, 1 mM magnesium acetate, 0.1 mM EDTA. To
remove gel residues, the extracts were passed through a SepPak
cartridge (Waters, Milford, MA) as instructed by the manufacturer. The
purified oligomers were dried, resuspended to a concentration of 5
µg/ml in 10 mM Tris-HCl buffer, pH 8.0, 1 mM EDTA
(TE buffer) and labeled at their 5`-end with
P using T4
bacteriophage polynucleotide kinase and
[
-
P]ATP (Sambrook et al., 1989).
Gel Electrophoresis
Intramolecular folding or
unfolding of 5` P-labeled oligodeoxynucleotides were
followed by electrophoresis of the DNA in a nondenaturing
polyacrylamide gel at 4 or at 75 °C, respectively (Williamson et al., 1989; Zahler et al., 1991; Murchie and
Lilley, 1994). Temperature-controlled electrophoresis was performed in
a Polar Bear isothermal electrophoresis apparatus (Owl Scientific,
Cambridge, MA) attached to a thermostatic circulator bath (Neslab
Instruments, Newington, NH). The 5` end-labeled oligodeoxynucleotides
were boiled for 3 min in TE buffer, pH 8.0, in the presence or absence
of salt as indicated. For electrophoresis at 75 °C, the boiled
oligomers were promptly loaded onto a preheated 12% polyacrylamide gel
(19:1, acrylamide/bisacrylamide), in 89 mM Tris borate buffer,
pH 8.3, 1 mM EDTA (TBE buffer) with or without salt as
indicated, and electrophoresis was conducted in the presence or absence
of salt, at 4.5 or 13-17 V/cm, respectively. For electrophoresis
at 4 °C, DNA oligomers were boiled, annealed on ice for 20 min,
loaded onto a gel precooled to 4 °C, and electrophoresed at this
temperature in the presence or absence of salt, at 6.7 or 20-31
V/cm, respectively. Electrophoresis was stopped when the bromphenol
blue dye marker reached a distance of 22.0 cm. The gels were dried, and
positions of the labeled oligomers were detected and recorded by
phosphor imaging (FujiX BAS 1000 bio-imaging analyzer). The
electrophoretic mobility of the d(CGG)
oligomers was
determined relative to the following marker oligodeoxynucleotides: Tet,
a Tetrahymena telomeric sequence
5`-d(T
G)-3`; bacteriophage M13
universal primer 5`-d(TCCCAGTCACGACGT)-3`; and cross-linked H3T2
oligomer that was generated by UV-light irradiation and purified by
sequencing gel electrophoresis as described below. Resolution of
oligodeoxynucleotides according to size was performed by
electrophoresis through a 12% polyacrylamide, 8 M urea
sequencing gel (Fry and Loeb, 1992).Absorbance Thermal Denaturation Analysis
Unlabeled
oligodeoxynucleotides in TE buffer, pH 8.0, 100 mM KCl were
boiled for 3 min, annealed at 4 °C for 30 min, and then subjected
to thermal denaturation. DNA denaturation was monitored by following
the change in absorption at 257 mµ as a function of temperature
using a Varian DMS 200S UV spectrophotometer attached to a thermostatic
circulator water bath.UV Cross-linking of DNA
Cross-linking of DNA by UV
irradiation was performed with oligomers H3T2 or H3T3, which contain
d(CGG) runs interrupted by three clusters of two or three
thymidines each (Table 1). These sequences were folded by first
denaturing for 2 min at 95 °C 0.5 µg of gel-purified 5` P-labeled oligomer in TE with or without added salt,
followed by annealing on ice for 20 min. The folded oligomers were
irradiated at 254 mµ for 5 min at 4 °C and at a distance of 5
cm by a UV light source (UVP, San Gabriel, CA; 580 microwatts/cm
at 6 inches). The irradiated DNA was denatured for 2 min at 90
°C in a sequencing gel loading buffer and resolved by 12%
polyacrylamide, 8.0 M urea denaturing gel electrophoresis. To
isolate the cross-linked form of H3T2, 0.1 µg of end-labeled
oligomer, mixed with 8 µg of unlabeled oligomer was UV irradiated
and then resolved by denaturing gel electrophoresis as detailed above.
The cross-linked forms of the oligomers, which migrated ahead of the
main oligomer band, were excised from the gel, extracted, and purified
according to the standard procedure employed for gel purification of
DNA fragments (see above).DEPC
DEPC detects
unpaired purine residues in DNA (Herr, 1985; Furlong and Lilley, 1986;
Huertas et al., 1993; Murchie and Lilley, 1994). To probe
unpaired guanines in folded d(CGG)
and KMnOProbing of d(CGG) Oligomers
, 0.2-0.25
µg/µl 5` P-labeled oligomer in 40 µl of TBE
buffer, pH 8.3, 100 mM KCl, was denatured at 90 °C for 2
min and reannealed on ice for 30 min. Following annealing, 0.1 volume
of DEPC (Sigma) was added, and the DNA samples were incubated for 3 h
at 4 °C. To probe unpaired guanine residues in fully unfolded DNA,
labeled oligomers were denatured, incubated at 80 °C for additional
5 min and then exposed to 0.1 volume of DEPC for 4.5 min at 80 °C.
Reaction with the longer oligomers (33-mer, H3T1, H3T2, and H3T3) was
terminated by the addition of an equal volume of 3.0 M sodium
acetate, pH 7.0, 2.0 M
-mercaptoethanol, and 200 µg
of yeast tRNA carrier, and DEPC was removed by twice precipitating the
DNA by 2 volumes of ethanol. With all of the remaining oligomers, the
reaction was terminated by the addition of water to a final volume of
100 µl, and DEPC was removed by centrifugation through a Sephadex
G-25 minicolumn. The isolated DNA was dried, resuspended in 40 µl
of freshly diluted 1.0 M piperidine, and incubated for 5 min
at 90 °C. The piperidine-treated DNA samples were dried,
resuspended in sequencing gel loading buffer and electrophoresed
through a 20% polyacrylamide, 8.0 M urea sequencing gel.
Patterns of oligomer fragmentation were detected by autoradiography. , which selectively modifies unpaired
thymidine residues in folded nucleotide sequences (Burton and Reilly,
1966; Hayatsu and Ukita, 1967; Friedman and Brown, 1978). DNA
oligomers, folded or denatured as described above, were exposed to
KMnO at a final concentration of 0.5 mM for either
30 min at 4 °C (folding conditions) or for 5 min at 80 °C
(denaturation conditions). Termination of the KMnO reaction, isolation of the DNA, its treatment by piperidine and
electrophoresis in a 20% polyacrylamide sequencing gel were as
described above for DEPC.Methylation-protection Analysis of d(CGG)
The accessibility to methylation of guanine N7
groups in folded or denatured d(CGG) Oligomers tracts was assessed by
reacting oligomers with dimethyl sulfate (DMS). A folded DNA structure
was formed by heating 5` P-labeled oligomer for 2 min at
90 °C in 40 µl of TBE buffer, pH 8.3, 100 mM KCl
followed by 30 min annealing at 4 °C. After adding 5 µl of 1
mg/ml dG
carrier, the folded oligomer was
alkylated by the addition of 5 µl of 0.5% DMS freshly diluted in
TBE buffer, pH 8.3, 100 mM KCl. The reaction mixture was
incubated for 15 min at 4 °C, and methylation was terminated by the
addition of 25 µl of 0.9 M sodium acetate pH 7.0, 0.3 M
-mercaptoethanol, and 300 µg/ml yeast tRNA carrier.
The DNA was twice precipitated at 4 °C by 2.0 volumes of ethanol,
dried, and hydrolyzed by 1.0 M piperidine as described for
DEPC above. Denatured oligomers were similarly methylated except that
carrier dG and DMS were added immediately after
the labeled d(CGG)
oligomer was denatured for 2 min at 90
°C, and modification by DMS was conducted for 45 s at 90 °C.
Termination of the methylation reaction, cleavage by piperidine, and
sequencing gel electrophoresis were performed as described for DEPC
above.
Folding of d(CGG)
To
examine whether or not d(CGG) Oligomers oligomers are capable of
forming folded structures, their electrophoretic migration in a
nondenaturing polyacrylamide gel was compared at 4 and 75 °C.
Mobilities of d(CGG) oligomers of different length were
assessed relative to each other and to the migration of the following
marker DNA fragments: Tet, which forms a unimolecular folded quadruplex
structure at 4 °C and unfolds at 75 °C (Williamson et
al., 1989); bacteriophage M13 universal primer, which lacks an
intrastrand complementarity necessary for folding; and UV cross-linked
H3T2 oligomer, which maintains a folded form under both native or
denaturing conditions (see Fig. 3below). As shown in Fig. 1, d(CGG) oligomers display different
mobilities at 4 and 75 °C. The oligomers migrate at 75 °C at
rates inversely proportional to their length, both in the presence of
salt (Fig. 1D) or in its absence (Fig. 1B). This mobility pattern is similar to the
pattern of migration in a denaturing sequencing gel (Fig. 1E). By contrast, at 4 °C and with (Fig. 1C) or without salt (Fig. 1A),
every tested d(CGG) oligomer displays an increased
mobility. In contrast to the progressively slowed mobility of oligomers
of increasing length at 75 °C (Fig. 1, B and D), 8-mer, 11-mer, and 15-mer d(CGG) sequences display
comparable mobilities at 4 °C, indicating an accelerated migration
of the 15- and perhaps the 11-mer oligonucleotides (Fig. 1, A and D). The temperature dependence of the mobility
of d(CGG) oligomers is also underscored by the similar
electrophoretic behavior of 24-mer d(CGG) and of the telomeric Tet
sequence, which forms a compact, rapidly migrating unimolecular
tetraplex in the presence of a monovalent alkali cation (Williamson et al., 1989). As seen in Fig. 1, A and B, both the mobilities of the 24-mer and of Tet are strongly
affected by temperature. At 4 °C, the two DNA oligomers migrate at
a rate similar to that of the M13 universal primer (Fig. 1A), but at 75 °C, both sequences migrate at
a significantly slower rate than the M13 primer (Fig. 1B). Note, however, that whereas the migration at
75 °C of the 24-mer is retarded in the presence of potassium, Tet
retains a high mobility despite the elevation in temperature (Fig. 1D). It appears, therefore, that unlike
quadruplex Tet, which requires potassium for stabilization (Sen and
Gilbert, 1991; Williamson, 1992), the compact form of the 24-mer is
less stabilized by this cation (see below). The strong temperature
dependence of the electrophoretic mobility of the d(CGG) oligomers is also apparent for the 33-mer and for H3T2. These
oligomers migrate at 4 °C ahead of the folded UV cross-linked H3T2
marker (Fig. 1, A and C) but display a slower
relative migration at 75 °C (Fig. 1, B and D).
oligomers in a nondenaturing
polyacrylamide gel at 4 or 75 °C. A, electrophoregram of
d(CGG)
oligomers run under nondenaturing
conditions at 4 °C and in the absence of salt. End-labeled oligomer
or DNA marker,
500 pg of DNA in 2.5 µl of TE buffer pH 8.0,
was heat-denatured for 3 min at 90 °C, annealed for 30 min at 4
°C, and electrophoresed at 4 °C through a nondenaturing 12%
polyacrylamide gel (see ``Experimental Procedures''). B, electrophoregram of d(CGG)
oligomers
run under nondenaturing conditions at 75 °C and in the absence of
salt. End-labeled oligomer or DNA marker,
500 pg of DNA in 2.5
µl TE buffer pH 8.0, was heat-denatured for 3 min at 90 °C,
incubated for additional 30 min at 75 °C, and electrophoresed at 75
°C through a nondenaturing 12% polyacrylamide gel (see
``Experimental Procedures''). C, electrophoregram of
d(CGG)
oligomers run under nondenaturing
conditions at 4 °C and in the presence of salt. DNA sample
preparation and electrophoresis were performed as described for A above, except that DNA annealing and electrophoresis were
conducted in buffers contained 100 mM KCl. D,
electrophoregram of d(CGG)
oligomers run under
nondenaturing conditions at 75 °C and in the presence of salt. DNA
sample preparation and electrophoresis were performed as in B above except that DNA annealing and electrophoresis were conducted
in buffers contained 100 mM KCl. E, electrophoregram
of d(CGG)
oligomers run through a 12%
polyacrylamide, 8.0 M urea denaturing gel. DNA sample
preparation and electrophoresis were conducted as described under
``Experimental Procedures.'' Marker DNA designations were as
follows: Univ., bacteriophage M13 universal primer; Tet, Tetrahymena telomeric sequence
5`-d(T
G)-3`; H3T2 CL,
UV-light cross-linked oligomer H3T2.
oligomers are found to form within 5-7 min,
which is the shortest testable annealing time (data not presented).
Furthermore, these structures are generated at an identical efficiency
over a range of DNA concentrations of 1.0-500.0 µg/ml
(results not shown). Thus, in contrast to the second-order kinetics of
formation of a multimolecular d(CGG) tetraplex (Fry and
Loeb, 1994) the observed compact forms of d(CGG) oligomers
are generated at a zero-order kinetics, suggesting that these
structures are unimolecular. It is notable that when incubated for
>30 min at 4 °C in the presence of 100 mM KCl, the
24-mer d(CGG) sequence displays in a nondenaturing gel a major rapidly
migrating band and a minor electrophoretically retarded band that
presumably represents a multimolecular tetraplex (Fry and Loeb, 1994)
(results not shown). It thus appears that an equilibrium between a
folded unimolecular form of d(CGG) and a multistranded
tetraplex form of such a sequence is determined by the concentration of
DNA, the presence of salt, and the time of incubation. oligomers unfold at elevated temperatures was
substantiated by absorbance thermal denaturation analysis. As shown in Fig. 2, heating of every examined d(CGG) oligomer
elicits hyperchromicity at 257 mµ, indicating the melting of these
DNA fragments.
oligomers. Oligomer DNA, 2.0-4.0
µg in 1.0 ml of TE buffer pH 8.0, 100 mM KCl, was boiled
for 3 min and annealed for additional 30 min at 4 °C. Changes in
the absorbance of the DNA at 257 mµ as a function of temperature
were measured as described under ``Material and Methods.''
The curve depicts the increase in absorbance of each oligomer relative
to its absorption at 20 °C normalized to
1.0.
variants, we irradiated the annealed DNA with UV
light to form covalently cross-linked folded species. A unimolecular
tetrahelical form of the Tetrahymena telomeric sequence Tet,
d(TG) was shown to become
cross-linked by UV-irradiation through the generation of thymidine
dimers (Williamson et al., 1989). To examine the formation of
cross-linked folded d(CGG) structures, oligomers H3T2 and
H3T3 were constructed that contain d(CGG) tracts
interspersed by thymidine runs (Table 1). End-labeled H3T2 or
H3T3 oligodeoxynucleotides were UV irradiated, and the formation of
covalently cross-linked compact forms of these oligomers was detected
by discerning species that retain rapid migration in a denaturing gel
(see ``Experimental Procedures''). As shown in Fig. 3,
H3T3 that was UV-irradiated in a buffer solution with no added salt or
in the presence of 200 mM chloride salts of
Na, K
, Cs
, or
Rb
displays similar amounts of a form that migrates in
a denaturing sequencing gel ahead of the main band of denatured
oligomer. This form is generated at a lower efficiency when H3T3 is
UV-irradiated in the presence of 200 mM NH
Cl (Fig. 3), and it is undetectable when this oligomer is
irradiated in the presence of 10 mM Mg (not
shown). Comparable results are obtained by UV irradiation of H3T2
oligomer, except that a rapidly migrating cross-linked form of this
sequence is generated efficiently in the presence of
NH
ions (data not presented). Hence,
d(CGG)
tracts interspersed by runs of thymidines form
compact structures that can be cross-linked by UV irradiation. Unlike
quadruplex DNA, which requires specific cations for its formation and
stabilization (Sen and Gilbert, 1991; Williamson, 1992) the compact
forms of d(CGG) appear to be generated independently of the
type or the presence of a cation.Chemical Probing of Folded Forms of d(CGG)
To investigate the structure of the folded forms
of d(CGG) Oligomers oligomers, we first probed them with DEPC, which
detects unpaired purine residues in DNA (Herr, 1985; Furlong and
Lilley, 1986; Huertas et al., 1993; Murchie and Lilley, 1994).
Oligomers were incubated with DEPC in a TBE buffer solution, pH 8.3,
containing 100 mM KCl, at either 4 °C (folding conditions)
or at 80 °C (denaturing conditions). After the reaction was
terminated and DEPC was removed, the DNA was exposed to piperidine to
cleave DEPC-modified unpaired guanine residues, and product DNA
fragments were resolved by sequencing gel electrophoresis (see
``Experimental Procedures''). In defining the reaction
conditions, we found that bacteriophage M13 universal primer is
modified and cleaved to the same extent when treated by DEPC at 4 or 80
°C, indicating a comparable reactivity of DEPC under the
nondenaturing and denaturing temperatures (results not shown).
Similarly, the patterns of DEPC modification of the shorter 8-mer and
11-mer d(CGG) oligodeoxynucleotides (Fig. 4), reveal
that all of the guanine residues in these DNA fragments are cleaved to
a similar extent at 4 and 80 °C. It thus appears that these shorter
oligomers fold only minimally or randomly at 4 °C. By contrast,
only selected clusters of guanine residues are modified by DEPC in the
15-, 24-, and 33-mer oligodeoxynucleotides (Fig. 4A). A
reproducible pattern of DEPC modification of these oligomers identifies
within each d(CGG) tract a central or epicentral cluster of
four to five residues, which is modified by DEPC and thus constitutes
an unpaired loop. Conversely, all the other guanine residues remain
protected and are located, therefore, in paired regions (Fig. 4B). These results sustain the previous
indications for the formation of folded structures by d(CGG) tracts (Fig. 1, Fig. 2, and Fig. 3). The
presence of a single DEPC-modified region in the compacted oligomers
indicates that these tracts probably fold back to form a hairpin-like
structure with a central or epicentral single strand loop.
oligomers. A, denaturing
polyacrylamide gel electrophoregram of DEPC-treated d(CGG)
oligomers. End-labeled 8-, 11-, 15-, 24- and 33-mer
oligodeoxynucleotides, 0.20-0.25 µg of DNA in 40 µl of TE
buffer, pH 8.0, containing 100 mM KCl, were heat-denatured as
described under ``Experimental Procedures.'' The oligomers
were either annealed and treated with DEPC at 4 °C or heated and
exposed to DEPC at 80 °C. The DEPC-treated DNA was cleaved by
piperidine and electrophoresed through a denaturing 12% polyacrylamide,
8.0 M urea gel as detailed under ``Experimental
Procedures.'' Bck, piperidine cleavage of oligomers that
were not exposed to DEPC; solid arrowheads, strong cleavage; open arrowheads, weak cleavage. B, allocation of
DEPC-modified nucleotide tracts within the d(CGG)
sequence. The solid-line frame marks a strongly cleaved
region, and the broken line demarcates the weakly cleaved
nucleotides.
sequences of the FMR1 gene in normal subjects (Kunst and
Warren, 1994). We inquired, therefore, whether or not the introduction
of thymidine clusters into a d(CGG) stretch alters its
pattern of folding. Annealed or denatured 15-mer and H1T2 oligomers
were modified by DEPC or KMnO, which detect unpaired
guanine or thymidine residues, respectively. As seen in Fig. 5,
the region that loops-out at 4 °C is located at a similar position
in the two oligomers. However, the two thymidine residues introduced
into H1T2 remain unpaired at 4 °C and merge with the unpaired
guanines to form a larger loop. To broaden this inquiry, the 33-mer,
H3T1, H3T2, and H3T3 oligomers were similarly probed by DEPC and
KMnO for regions of unpaired guanine and thymidine
residues, respectively. Patterns of DEPC-modification shown in Fig. 6A indicate that the introduction of three runs of one,
two, or three thymidine residues each into the 33-mer DNA does not
alter the position of the primary unpaired loop. Note that the patterns
of KMnO modification demonstrate that whereas the
central thymidine cluster merges with the unpaired guanines to form a
larger loop, the remaining two thymidine runs form two additional
unpaired regions that do not include guanine residues (Fig. 6B). However, all three unpaired regions in H3T3
contain thymidine residues with only a negligible participation of
guanines in the looped-out tracts (Fig. 6, A and B and the scheme in Fig. 6C). It appears,
therefore, that the symmetry of the primary fold in d(CGG) remains unaltered whether or not thymidine residues are
introduced. Yet, as additional tracts of thymidine are added, they may
form auxiliary loops that do not include guanine residues.
of the 15-mer and H1T2 oligomers. A, denaturing
polyacrylamide gel electrophoregram of DEPC- and
KMnO-treated oligomers. End-labeled oligomers,
0.20-025 µg of DNA in 40 µl of TE buffer, pH 8.0, 100
mM KCl, were heat-denatured, annealed at 4 °C, or heated
at 80 °C and treated by DEPC or KMnO at these
respective temperatures as described under ``Experimental
Procedures.'' Piperidine cleavage and 12% polyacrylamide, 8.0 M urea denaturing gel electrophoresis were conducted as
described for Fig. 4. Bck, piperidine cleavage of
oligomers not treated by DEPC or KMnO; solid
arrowheads, strong cleavage; open arrowheads, weak
cleavage. B allocation of DEPC- and KMnO-modified
nucleotide tracts within the 15-mer and H1T2 sequences. The solid
line frame marks a strongly cleaved region, and a broken line demarcates the weakly cleaved
nucleotides.
of d(CGG) oligomers containing thymidine
tracts of different length. A, denaturing polyacrylamide gel
electrophoregram of DEPC-treated oligomers. End-labeled 33-mer, H3T1,
H3T2, and H1T2 oligomers, 0.20-0.25 µg of DNA in 40 µl of
TE buffer, pH 8.0, 100 mM KCl were denatured, annealed at 4
°C, or heated at 80 °C and reacted with DEPC at these
respective temperatures as described under ``Experimental
Procedures.'' The oligomers were cleaved by piperidine and
electrophoresed through a 12% polyacrylamide, 8.0 M urea
denaturing gel as described in the legend to Fig. 4. Bck, piperidine cleavage of oligomers not treated by DEPC; solid arrowheads, strong cleavage; open arrowheads,
weak cleavage. B, denaturing polyacrylamide gel
electrophoregram of KMnO
-treated oligomers. End-labeled
33-mer, H3T1, H3T2, and H1T2 oligomers were treated with
KMnO, cleaved by piperidine, and electrophoresed through a
12% polyacrylamide denaturing gel under conditions detailed under
``Experimental Procedures.'' Bck, piperidine
cleavage of oligomers not treated by KMnO; solid
arrowheads, strong cleavage; open arrowheads, weak
cleavage. C, allocation of DEPC- and KMnO-modified
nucleotide tracts within the examined sequences. A solid line frame marks the strongly cleaved region, and the broken line demarcates the weakly cleaved
nucleotides.
Guanine Residues in Folded d(CGG)
By acting as both donors
and acceptors of Hoogsteen hydrogen bonds that involve the N-7 position
of each residue, four guanines can be arrayed cyclically to form a
guanine quartet. This N-7 hydrogen bonding renders the guanine residues
in the quartet resistant to methylation by DMS (Sen and Gilbert, 1991;
Williamson, 1992). However, guanine pairs in a Hoogsteen-bonded
double-stranded structure are readily alkylated as the N-7 position of
only one residue participates in a hydrogen bond, whereas the other
remains free and accessible to DMS. To examine whether or not the
compact forms of different d(CGG) Oligomers
Are Not Protected against Methylation oligomers represent
unimolecular tetrahelices, methylation protection analysis of the
folded structures was conducted. We probed the 33-mer DNA fragment,
which displays a single unpaired region ( Fig. 4and Fig. 6) and H3T2 and H3T3, which have three unpaired regions
each (Fig. 6). The oligomers were exposed to DMS at 4 °C
(folding conditions) or at 90 °C (denaturing conditions) and
subsequently treated with piperidine as detailed under
``Experimental Procedures.'' As Fig. 7shows, the
compact form of the telomeric sequence Ter that folds at 4 °C and
in the presence of potassium ions to form a unimolecular tetraplex
(Balagurumoorthy et al., 1992; Lee and Lee, 1993; Murchie and
Lilley, 1994) remains largely protected against methylation. By clear
contrast, all of the guanine residues in folded 33-mer, H3T2, and H3T3
become methylated (Fig. 7). It appears, therefore, that the
compact forms of these oligomers are not tetrahelices but rather
back-folded hairpin structures that are stabilized by either
Watson-Crick or by Hoogsteen hydrogen bonds (see
``Discussion'').
oligomers. Oligomers, 0.20-0.25
µg of DNA in 40 µl of TBE buffer, pH 8.3, 100 mM KCl
were heat-denatured for 3 min at 90 °C and annealed for 30 min at 4
°C. After the addition of 0.1 volume of 1.0 mg/ml
dG
carrier DNA and 0.1 volume of 0.5% DMS freshly
diluted in TBE buffer, pH 8.3, 100 mM KCl, the samples were
either incubated for 15 min at 4 °C or for 45 s at 90 °C.
Termination of the methylation reaction, piperidine cleavage, and
denaturing gel electrophoresis were conducted as described under
``Experimental Procedures.''
tract in fragile X cells and the
ensuing blocking of FMR1 transcription and its delayed
replication, requires a better understanding of the alternative
structures of the d(CGG) sequence. We reported previously
that d(CGG) oligomers readily form in vitro a
multistrand Hoogsteen-bonded tetrahelical structure that is stabilized
by cytosine C-5 methylation (Fry and Loeb, 1994). This observation led
us to speculate that single strand tracts of d(CGG), which
become exposed during replication or transcription, might also fold
back to form Hoogsteen-bonded unimolecular hairpin or tetraplex
domains. Such structures could act in vivo to precipitate
slippage during replication that could result in d(CGG) expansion, and they may also block the progression of the
transcription or replication machineries. readily loop
back under physiological conditions in vitro to form hairpin
structures. Unimolecular hairpin or tetraplex structures of telomeric
DNA display an increased electrophoretic mobility in a nondenaturing
gel and conversely, their electrophoretic migration slows when they
become thermally denatured (Williamson et al., 1989). In
analogy, we find that whereas heat-denatured d(CGG) oligomers migrate in a nondenaturing gel at rates inversely
proportional to their length, their relative mobility becomes
anomalously accelerated under annealing conditions (Fig. 1).
Taken as evidence for the folding of d(CGG) into more
compact forms, this observation is strengthened by the demonstration of
UV hyperchromicity of thermally denaturing d(CGG) oligomers (Fig. 2). Similar hyperchromicity is displayed by unfolding
telomeric DNA sequences (Scaria et al., 1992). Finally, as was
shown in the past for telomeric DNA (Williamson et al., 1989),
compact forms of d(CGG) runs interspersed by thymidine
clusters can be covalently cross-linked by UV light (Fig. 3). are unimolecular
is indicated by their faster mobility relative to the unfolded
oligomers ( Fig. 1and Fig. 3) and by the zero-order
kinetics of their formation (see ``Results''). Such
monomolecular compact forms may represent either fold-back hairpin
structures or an intrastrand tetrahelix. Formation and stabilization of
tetraplex DNA require Na or K
cations
(Sen and Gilbert, 1991; Williamson, 1992). As the compact forms of
d(CGG)
are generated and maintained to a comparable extent
with or without various cations ( Fig. 1and Fig. 3), it
is unlikely that they represent quadruplex structures. The modification
of all of the guanines in folded d(CGG) oligomers by DMS (Fig. 7) further argues against their arrangement in a tetraplex
conformation. Finally, DEPC modification provides direct evidence for
the folding of d(CGG) into a hairpin structure rather than
into a tetrahelix. Had the folded structures been unimolecular
quadruplexes, three unpaired loops that serve as hinges should have
been observed. Instead, a single region of unpaired nucleotides is
discerned at a central or epicentral location within several
d(CGG) oligomers (Fig. 4). Insertion of three
thymidine clusters consisting of one, two, or three residues each into
a d(CGG) tract does not change the location of this primary
looped-out region ( Fig. 5and Fig. 6). Insertion of three
thymidine clusters into the d(CGG) sequence introduces two
additional regions of unpaired thymidines, which do not include
contiguous guanines ( Fig. 5and Fig. 6). These oligomers
remain fully accessible to DMS (Fig. 7), and thus it appears
that these oligomers also maintain a hairpin structure rather than a
tetraplex conformation. The weight of the evidence provided here
points, therefore, at a similarity between the d(CGG) tracts and telomeric DNA sequences, which are capable of forming
back-folded hairpin structures at a physiological range of salt
concentrations, temperatures and pH values (Sundquist and Klug, 1989;
Sen and Gilbert, 1990; Balagurumoorthy and Brahmachari, 1994; Choi and
Choi, 1994). tracts, is corroborated by the works of Gacy et
al.(1995) and Chen et al.(1995), which were published as
this work was completed. Employing NMR analysis, these authors
demonstrated that d(CGG), as well as the complementary
d(CCG) sequence (Chen et al., 1995) and other
trinucleotide sequences that expand in human disease (Gacy et
al., 1995), form hairpin structures. Calculations indicate that
the stability of d(CGG) hairpin increases linearly with
length and that a threshold stability is attained at lengths conforming
with those of expanded d(CGG) sequence in individuals
afflicted with the fragile X syndrome (Gacy et al., 1995). oligomers (Fig. 4) or in tracts interspersed with
thymidine clusters ( Fig. 5and Fig. 6), we constructed
models for the structure of d(CGG) fold-back hairpins. The
scheme presented in Fig. 8A suggests that the 24-mer DNA
folds back asymmetrically to form a hairpin structure that is
stabilized by five guanine-guanine Hoogsteen base pairs. By contrast,
the 33-mer folds back with a nearly perfect symmetry to form a hairpin
structure that is maintained by four guanine-guanine Hoogsteen
base pairs and/or 10 guanine-cytosine Watson-Crick pairs (Fig. 8A). Modification by DEPC of purines in the
folded 33-mer d(CGG) tract cannot reveal whether or not the cytosine
residue at position 14 from the 5` end (marked *C in Fig. 8A), is paired or not. Hence, an alternative
structure of folded 33-mer can be envisioned in which this residue
remains unpaired and the guanine at position 19 pairs with a guanine at
position 13. A resulting hairpin structure is less symmetric than the
one depicted in Fig. 8A, having a dinucleotide 3`
unpaired tail and being stabilized by eight Watson-Crick and/or five
Hoogsteen hydrogen bonds (model not shown). Since the electrophoregram
shown in Fig. 8A fails to distinguish between a single-
or dinucleotide 3` overhang, it cannot be determined which alternative
folded form exists. A generally similar model for the hairpin structure
of a d(CGG) oligomer, based on NMR analysis was presented
recently by Gacy et al.(1995). A stereo model of d(CGG)
constructed on the basis of NMR analysis implicates both
G
C and
G
G in the stabilization of the stem
of the hairpin (Chen et al., 1995). Hence, it appears that the
schematic representation of the hairpin structures in Fig. 8A is sterically plausible as attested by
independent NMR analysis and model construction.
oligomers. A, schemes of the
hairpin forms of the 24- and 33-mer d(CGG)
fragments. The model is based on the identification by
DEPC-modification of looped regions of unpaired guanines in the
annealed oligomers (Fig. 4). Watson-Crick and non-Watson-Crick
bonds are marked by three dashes and two Xs,
respectively. The absence of a bridging notation indicates lack of base
pairing. The marking of pairing in the scheme does not indicate that
any specific bond is actually formed. The notation *C in the
model of folded 33-mer marks a cytosine residue that may or may not be
paired (see ``Discussion''). B, schemes of
alternative folded forms of the H3T3 oligomer. The models are based on
the identification by DEPC and KMnO
modification of
unpaired guanines and thymidines, respectively, in the annealed
oligomers (Fig. 6).
sequence to form Hoogsteen-bonded
tetrahelix (Fry and Loeb, 1994) or back-folded hairpin structures (Gacy et al.(1995) and this communication) may have a biological
significance. First, the exposure of long stretches of single-stranded
d(CGG) during DNA replication and the formation of a stable
hairpin structure by this tract could ensue slippage and trinucleotide
expansion. Second, hairpin formation could block replication,
transcription, and perhaps translation. It was shown recently that the
progression in vitro of AMV reverse transcriptase along
insulin-like growth factor II mRNA is blocked at a guanine-rich region
that forms a quadruplex domain and two stable hairpins (Christiansen et al., 1994). Similarly, the progression in vitro of
a variety of DNA polymerases along a guanine-rich region in the
globin promoter DNA is arrested at an intrastrand tetraplex-forming
sequence (Woodford et al., 1994). Last, expansion of the
trinucleotide (CGG) repeat stalls the progression of the 40 S ribosomal
subunit during the translation of FMR1 mRNA (Feng et
al., 1995). Hence, it may be argued that the exposure of stretches
of expanded (CGG) tracts in DNA or mRNA and their folding
into hairpin structures might block replication and transcription or
translation, respectively. The formation of (CGG) secondary
structures in FMR1 DNA or mRNA could thus contribute to the
observed delay of FMR1 replication in fragile X cells (Hansen et al., 1993), to the silencing of its transcription (Pierreti et al., 1991; Hansen et al., 1992; Sutcliffe et
al., 1992), and to the interrupted translation of its mRNA (Feng et al., 1995). Last, evidence indicates that cytosine
methylation by methyltransferase is enhanced by hairpins formed by
human c-Ha-ras telomere-like sequence and d(CCG) trinucleotide
repeat (Smith et al., 1994). Thus, the folding of d(CGG) may also be instrumental in the hypermethylation of the
d(CGG) region in DNA of fragile X syndrome cells. and the resulting inhibition of DNA
synthesis and transcription may not occur in cells of normal and
``premutation'' individuals that have a shorter and unstable
d(CGG) stretch (Gacy et al., 1995). These shorter
tracts may also be stabilized in a rigid conformation by single strand
d(CGG) binding proteins of the type that was recently
identified in HeLa cells (Richards et al., 1993). The
expansion of the d(CGG) stretch in individuals afflicted
with the fragile X syndrome could, however, defeat the capacity of such
binding proteins to coat it, and the naked d(CGG) stretches
may fold-back to form stable hairpin domains.
)
©1995 by The American Society for Biochemistry and Molecular Biology, Inc.
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  K. Usdin The biological effects of simple tandem repeats: Lessons from the repeat expansion diseases Genome Res., July 1, 2008; 18(7): 1011 - 1019. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Khateb, P. Weisman-Shomer, I. Hershco-Shani, A. L. Ludwig, and M. Fry The tetraplex (CGG)n destabilizing proteins hnRNP A2 and CBF-A enhance the in vivo translation of fragile X premutation mRNA Nucleic Acids Res., September 27, 2007; 35(17): 5775 - 5788. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Napierala, A. Bacolla, and R. D. Wells Increased Negative Superhelical Density in Vivo Enhances the Genetic Instability of Triplet Repeat Sequences J. Biol. Chem., November 11, 2005; 280(45): 37366 - 37376. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Fukuda, M. Katahira, E. Tanaka, Y. Enokizono, N. Tsuchiya, K. Higuchi, M. Nagao, and H. Nakagama Unfolding of higher DNA structures formed by the d(CGG) triplet repeat by UP1 protein Genes Cells, October 1, 2005; 10(10): 953 - 962. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Yafe, S. Etzioni, P. Weisman-Shomer, and M. Fry Formation and properties of hairpin and tetraplex structures of guanine-rich regulatory sequences of muscle-specific genes Nucleic Acids Res., May 20, 2005; 33(9): 2887 - 2900. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Khateb, P. Weisman-Shomer, I. Hershco, L. A. Loeb, and M. Fry Destabilization of tetraplex structures of the fragile X repeat sequence (CGG)n is mediated by homolog-conserved domains in three members of the hnRNP family Nucleic Acids Res., August 9, 2004; 32(14): 4145 - 4154. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Kim, Y. M. Kwon, J. S. Kim, H. Lee, J.-H. Park, Y. M. Shim, J. Han, J. Park, and D.-H. Kim Tumor-Specific Methylation in Bronchial Lavage for the Early Detection of Non-Small-Cell Lung Cancer J. C |
ABSTRACT
