|
|
|
|||||||
|
|||||||||
(Received for publication, July 10,
1995; and in revised form, August 1, 1995) From the
We examine the association of transcription factor TFIIIA with
RNA-DNA heteroduplexes containing sequences from the Xenopus
borealis 5 S rRNA gene. Under conditions where TFIIIA selectively
binds to 5 S rRNA or the internal control region of the 5 S rRNA gene,
no specific association of TFIIIA with DNA-RNA heteroduplexes
containing either strand of 5 S DNA could be detected. We discuss our
results with respect to specific models of TFIIIA recognition of the
internal control region and of DNA-RNA hybrids by zinc finger proteins.
TFIIIA is a positive transcription factor for the 5 S ribosomal
RNA gene (Engelke et al., 1980; Pelham and Brown, 1980). The
protein has a modular structure comprised of nine zinc finger domains
(Hanas et al., 1983; Smith et al., 1984; Ginsberg et al., 1984; Miller et al., 1985) that are linked to
a carboxyl-terminal domain that interacts with other transcription
factors (Hayes et al., 1989; Mao and Darby, 1993). The nine
zinc fingers bind as a linear array along the internal control region
(ICR) ( Gottesfeld and colleagues have clearly established that
distinct clusters of zinc fingers within TFIIIA have different
affinities for DNA sequences compared with RNA sequences (Liao et
al., 1992; Clemens et al., 1993, 1994; McBryant et
al., 1995). These studies indicate that TFIIIA utilizes distinct
domains to confer specificity to the interaction with 5 S DNA and 5 S
RNA; however, all zinc finger domains contribute to the stability of
association with nucleic acid. Other experiments that examine the
interaction of TFIIIA with 5 S rRNA are consistent with the protein
recognizing RNA through different mechanisms than the recognition of
DNA (Darby and Joho, 1992; Theunissen et al., 1992; Romaniuk et al., 1987; You et al., 1991; Sands and Bogenhagen,
1987, 1991). Since TFIIIA can recognize both DNA and RNA with sequence
selectivity it is possible that TFIIIA might recognize RNA-DNA
heteroduplexes with comparable sequence selectivity. In this work we
directly examine the interaction of TFIIIA with DNA-RNA heteroduplexes
containing either strand of the 5 S rRNA gene as DNA. In contrast with
earlier speculations (Rhodes and Klug, 1986; Fairall et al.,
1989; Shi and Berg, 1995), we find that TFIIIA does not have
sequence-selective interactions with these specific RNA-DNA
heteroduplexes under the same conditions that it selectively binds to 5
S rRNA or the 5 S rRNA gene.
Figure 1:
Preparation of ``sense''
or ``antisense'' RNA-DNA heteroduplexes containing the Xenopus 5 S rRNA gene sequence. The scheme illustrates the
organization of the plasmid pXP10, where the arrowedregion represents the 5 S rRNA gene and the hatchedarea the ICR to which TFIIIA binds. The alternative
construct used shows pXP14 contains the same 5 S rRNA sequence, but it
is inverted (Wolffe et al., 1986). The arrow at 1 indicates the start of the SP6 polymerase
transcript.
Figure 2:
Gel retardation assays showing TFIIIA
binding to both DNA-RNA heteroduplexes and to DNA and 5 S rRNA. A, DNA and 5 S rRNA. TFIIIA binding to a 298-bp
fragment of the Xenopus 5 S rRNA gene (DNA, lanes
1-6) or to 5 S rRNA (RNA, lanes 7-12) is
shown. 5 ng of end-labeled nucleic acid (lanes1 and 7) was incubated with 0.15 ng (lanes2 and 8), 0.3 ng (lanes3 and 9), 0.6 ng (lanes4 and 10), 1.2 ng (lanes5 and 11), and 2.4 ng (lanes6 and 12) of TFIIIA. Complexes were resolved on a 0.7%
agarose gel. The resolution of free nucleic acid (Free) and
the binding of single (1) or multiple TFIIIA molecules (2, 3) to the fragments is indicated, as are the
generation of nucleoprotein aggregates. B, DNA-RNA
heteroduplex. 5 ng of ``sense'' (XP14) (lanes1-5) or ``antisense'' (XP10) (lanes
6-10) RNA-DNA heteroduplex (lanes1 and 6) was incubated with 15 ng (lanes2 and 7), 30 ng (lanes3 and 8), 60 ng (lanes4 and 9), and 120 ng (lanes5 and 10) of TFIIIA.
Figure 5:
Footprinting analysis of TFIIIA binding to
both DNA-RNA heteroduplexes and to DNA. A, TFIIIA binding to
5` end-labeled ``sense'' (XP14) and ``antisense''
(XP10) heteroduplexes was analyzed by DNase I digestion as described
under ``Materials and Methods.'' The lanes are
labeled as follows: lanes2 and 6, DNase I
digestion of free heteroduplex; lanes3 and 4 or 7 and 8, digestion of heteroduplex with one
and two TFIIIA molecules bound to the fragment, respectively. Lanes1 and 2 show labeled G cleared fragments in a
Maxam-Gilbert G ladder. B, TFIIIA binding to double-stranded
DNA of the same sequence present in the heteroduplexes. The lanes are
labeled as follows: -, DNase I digestion of free DNA (lanes10 and 13); +, DNase I digestion of the
same fragment bound by a single TFIIIA molecule (lanes11 and 14). Lanes9 and 12 are of
labeled G cleared fragments in a Maxam-Gilbert G ladder. The position
of the ICR is indicated.
Figure 3:
Binding titration for TFIIIA binding to
either ``antisense'' (
Figure 4:
Specificity and strength of TFIIIA binding
to DNA-RNA heteroduplexes. A, TFIIIA binding to either a
fragment of the 5 S rRNA gene (DNA, lanes 9-16)
or 5 S rRNA (5S rRNA, lanes 1-8). The complexes
formed by mixing 5 ng of nucleic acid and 3 ng of TFIIIA (lanes
2-10) were incubated with 10-fold (lanes 3, 6, 11, and 14), 100-fold (lanes 4, 7, 12, and 15), and 500-fold excesses (lanes 5, 8, 13, and 16) of either specific (Sp. containing the 5 S rRNA
sequence) or nonspecific DNA (NSp. pBR322) and separated on a
0.7% agarose gel. The resolution of free nucleic acid (Free)
and the binding of single (1) or multiple TFIIIA molecules to
the fragments is indicated. B, 5 ng of DNA-RNA heteroduplex
(``antisense'') (lane1) was incubated with
TFIIIA (120 ng) to generate a mixed population of TFIIIA-heteroduplex
complexes (lane2). Binding was then assessed upon
titration of specific (Sp.) or nonspecific competitor DNA (NSp.) such that the samples contained a 1-fold (lanes3 and 7), 10-fold (lanes4 and 8), 100-fold (lanes5 and 9), or
500-fold excess (lanes6 and 10). The
resultant nucleoprotein complexes were separated on a 0.7% agarose gel. C shows the mobility of naked nucleic acid, and T shows the mobility of the complex of TFIIIA with nucleic acid
without competitor.
Our
final assessment of TFIIIA interaction with the RNA-DNA heteroduplexes
was to examine whether sequence-selective binding occurred by DNase I
footprinting. Our strategy was to carry out the DNase I cleavage
reaction after mixing TFIIIA either with the RNA-DNA heteroduplex or
with DNA as a control before resolving the nucleoprotein complexes on a
non-denaturing gel (Hayes and Wolffe, 1992). The nucleoprotein complex
of interest was then eluted from the gel, and the radiolabeled nucleic
acid was isolated and resolved on a denaturing polyacrylamide gel to
reveal any footprint. Resolution of the first complex formed when a
single molecule of TFIIIA is bound to 5 S DNA reveals a clear footprint
over the ICR (Fig. 5B, lanes 9-14).
Identical analysis of both DNA-RNA heteroduplexes does not reveal a
footprint (Fig. 5A, lanes 1-8). Thus
although TFIIIA will bind to either heteroduplex, this interaction is
not sequence selective. The early analysis of TFIIIA interactions with the 5 S RNA
gene suggested that TFIIIA might bind preferentially in a
sequence-specific manner to the non-coding DNA strand (Sakonju and
Brown, 1982). Subsequent studies have clearly established that such
selective interactions can occur for both TFIIIA and other zinc finger
proteins within the context of double helical DNA (Fairall et
al., 1986; Pavletich and Pabo, 1991). The suggestion that
comparable strand- and sequence-specific interactions detected in the
binding of zinc finger proteins to DNA-RNA heteroduplexes (Shi and
Berg, 1995) might also occur with TFIIIA is not supported by our
results (Fig. 3Fig. 4Fig. 5). We find that TFIIIA
will associate with a DNA-RNA heteroduplex (Fig. 2) but that
this interaction is relatively weak and nonspecific compared with
binding to 5 S DNA or RNA (Fig. 3Fig. 4Fig. 5)
(Clemens et al., 1993). The lack of sequence-specific
recognition of heteroduplexes containing either strand of 5 S DNA is
consistent with previous data indicating that TFIIIA recognizes the
internal control region as B-type DNA (Gottesfeld et al.,
1987; Hayes et al., 1990). Since DNA-RNA heteroduplexes
approximate to an A-type conformation (Hall, 1993), our results provide
yet further evidence against the proposal that equivalent recognition
of DNA and RNA might be mediated by the same zinc fingers within TFIIIA
(Rhodes and Klug, 1986; Fairall et al., 1986, 1989). Our
results are entirely consistent with the definition of distinct domains
of TFIIIA that interact with DNA or RNA in a sequence-selective manner
(see Clemens et al.(1993), Darby and Joho(1992), and
Theunissen et al.(1992)). Proteins such as the histones
will not interact stably with RNA-DNA heteroduplexes (Dunn and
Griffith, 1980; Hovatter and Martinson, 1987). Our results suggest that
the interaction of TFIIIA with the 5 S RNA gene will be destabilized by
the generation of any RNA-DNA heteroduplex during transcription. Thus
the generation of RNA-DNA heteroduplexes might provide a means of
destabilizing nucleoprotein complexes within the chromosome. It should,
however, be noted that in our experiments we generate very long
heteroduplexes (298 bp) whereas in vivo any heteroduplex
formed during transcription would be much shorter. TFIIIA in isolation
is displaced from DNA by transit of bacteriophage RNA polymerase
(Campbell and Setzer, 1991) yet remains bound to DNA within the intact
transcription complex (Bogenhagen et al., 1982; Wolffe et
al., 1986). Presumably multiple contacts between TFIIIA and other
transcription factors that bind outside of the transcribed region
facilitate the retention of TFIIIA on the 5 S rRNA gene during the
transcription process (Wolffe and Morse, 1990; Kassavetis et
al., 1990).
Volume 270,
Number 39,
Issue of September 29, pp. 22665-22668, 1995
©1995 by The American Society for Biochemistry and Molecular Biology, Inc.
)of the 5 S rRNA gene (Smith et al., 1984;
Vrana et al., 1988). The exact mode of interaction of the zinc
fingers with the ICR has been controversial (Fairall et al.,
1992; Fairall and Rhodes, 1992). A feature of early models was the
suggestion that A-type DNA could be involved in TFIIIA binding (Rhodes
and Klug, 1986; Fairall et al., 1989). Other experiments
suggested that TFIIIA recognizes B-type DNA (Gottesfeld et
al., 1987; Hayes et al., 1990). In addition, the path of
DNA is known to be distorted through interaction with TFIIIA (Schroth et al., 1989; Bazett-Jones and Brown, 1989). The solution of
the crystal structure of the zinc finger protein Zif 268 indicated that
zinc finger proteins of the TFIIIA class can interact with a B-type
helix (Pavletich and Pabo, 1991; Nekludova and Pabo, 1994). However,
the contacts made with the double helix by Zif 268 are predominantly to
a single strand of the helix. It has recently been suggested that a
combination of an A-type DNA structure and predominant recognition of a
single strand of the double helix might allow TFIIIA to bind in a
sequence-selective manner to a DNA-RNA heteroduplex (Shi and Berg,
1995).
Preparation of RNA-DNA Heteroduplex
RNA-DNA
heteroduplexes containing either the ``sense'' or
``antisense'' RNA strand were prepared by initial generation
of the appropriate RNA from EcoRI-digested pXP14 and pXP10
plasmids, respectively (Wolffe et al., 1986). These plasmids
contain a single copy of the Xenopus 5 S rRNA gene oriented
next to a SP6 polymerase promoter such that in vitro transcription generates antisense 5 S rRNA from pXP10 and sense 5
S rRNA from pXP14. Briefly, RNA was produced by incubation of 5 µg
of EcoRI-digested plasmid with 10 mM dithiothreitol,
0.5 mM NTPs (Pharmacia Biotech Inc.), 50 units of
human placental ribonuclease inhibitor (Life Technologies, Inc.), and
20 units of SP6 RNA polymerase (Promega) in a total volume of 100
µl. A 1-h incubation at 42 °C typically generated 3 µg of
RNA. The DNA template was then removed by addition of 2 units of
RNase-free DNase I (Promega) and incubated for 10 min at 37 °C, and
the sample was subsequently phenol/chloroform-extracted and
precipitated. The RNA was annealed to 1 µg of an appropriate
5`-labeled 20-mer primer (pXP10, 5`-CGGGATCCGGCTGGGCCCCC-3`; pXP14,
5`-CGGGATCCATCTGTTCGGGG-3`) by resuspension in 50 mM KCl, 25
mM Hepes, pH 7.5, and heating to 90 °C followed by slow
cooling to 50 °C over an hour. Following precipitation the primer
was extended using reverse transcriptase (Life Technologies, Inc.) at
42 °C for 60 min. After phenol/chloroform extraction and
precipitation, the DNA-RNA heteroduplex was separated and eluted from
6% native polyacrylamide gels.Purification of TFIIIA
7 S storage particles and
TFIIIA were purified as described by Smith et al.(1984). In
brief, immature ovary homogenate was fractionated on glycerol
gradients, bound to diethylaminoethyl cellulose, and eluted on a salt
gradient. The 7 S particle fractions were adjusted to 0.1 M KCl in 50 mM HEPES (pH 7.5), 5 mM
MgCl
, 1 mM dithiothreitol, 10 µM ZnCl, 20% glycerol (buffer A). RNase A was added to
the mixture (50 µg of enzyme/mg of protein), which was incubated
for 5 min, and then the volume was brought up 2-fold with buffer A
containing 10 M urea (to a final urea concentration of 5 M). The mixture was then loaded onto a 2-ml (bed volume)
Bio-Rex 70 column, and TFIIIA was eluted with an increasing
concentration of KCl (the protein eluting at a final concentration of 1 M). No other proteins were detectable by SDS-polyacrylamide
gel electrophoresis and silver staining.TFIIIA Gel Shifts
TFIIIA gel shifts were performed
as described previously (Lee et al., 1993). Briefly, 5 ng of
end-labeled heteroduplex, DNA, or RNA was incubated with 5-120 ng
of TFIIIA in a total volume of 10 µl of binding buffer (20 mM Hepes, pH 7.5, 70 mM NH
Cl, 7 mM MgCl, 10 µM ZnCl, 10 mM dithiothreitol, 3% (v/v) glycerol, 20 µg/ml bovine serum
albumin). Reactions were incubated at room temperature for 15 min and
loaded directly onto 0.8% agarose gels in 0.5 TB buffer (45 mM Tris borate, pH 8.0) with applied voltage of 4 V/cm for 3 h at
room temperature. EDTA was omitted in all binding and electrophoresis
buffers to avoid denaturing TFIIIA. In the competition studies, 5 ng of
end-labeled heteroduplex and 15 ng of TFIIIA was co-incubated with
either 5-2500 ng of EcoRI-HindIII fragment from
pXP10 (``specific DNA'') or pBR322 (``nonspecific
DNA'') and allowed to equilibrate for 15 min prior to loading onto
agarose gels. The experiments examining TFIIIA binding to DNA or RNA
utilized an identical protocol for either a 298-bp EcoRI-HindIII DNA fragment from pXP10 or in vitro transcribed 5 S rRNA.DNase I Footprinting
TFIIIA complexes with either
DNA or heteroduplex were digested with DNase I prior to separation on
0.8% agarose gels. The approximate concentration of DNase I (Boehringer
Mannheim) was determined empirically and performed at room temperature
for 4 min prior to direct loading onto the gel. After electrophoresis
the wet gel is placed on x-ray film, and the nucleoprotein complexes of
interest excised and eluted. Following phenol/chloroform extraction and
precipitation the samples are resuspended in formamide buffer and
resolved on 8% denaturing acrylamide gels.
Experimental Strategy
We prepared heteroduplexes
by initially transcribing linearized DNA templates using bacteriophage
SP6 RNA polymerase to generate RNA molecules of 298 nucleotides (Wolffe et al., 1986). These single-stranded RNA molecules were
annealed to DNA oligonucleotides near their 3` termini, which were then
extended by reverse transcriptase to generate RNA-DNA heteroduplexes of
285 bp in length, which were purified on non-denaturing polyacrylamide
gels (Fig. 1). Where necessary RNA or DNA were radiolabeled
either by the inclusion of radioactive RNA precursors during
transcription by SP6 RNA polymerase or by end labeling of the
oligonucleotide, respectively.
TFIIIA Binds to 5 S rRNA, 5 S DNA, and to Both 5 S
RNA-DNA Heteroduplexes
We initially established conditions under
which the interaction of TFIIIA with 5 S rRNA, 5 S DNA, and 5 S RNA-DNA
heteroduplexes could be detected by a gel retardation assay (Lee et
al., 1993). An increasing excess of TFIIIA relative to 5 S DNA led
to the accumulation of a single complex and then multiple complexes
until an aggregate appears (Fig. 2A, lanes
1-6). This reflects the initial association of a single
molecule of TFIIIA with the internal control region (Smith et
al., 1984; see Fig. 5later), followed by the nonspecific
sequestration of additional TFIIIA molecules (Daly and Wu, 1989).
Similar results are obtained on analysis of TFIIIA binding to 5 S rRNA (Fig. 2A, lanes 7-12) (see also Romaniuk et al.(1987), Sands and Bogenhagen(1987), Darby and
Joho(1992), and Clemens et al. (1993)). It should be noted
that the higher order complexes observed with both 5 S DNA and 5 S RNA
at high TFIIIA concentrations could have been eliminated by the
inclusion of additional nonspecific competitor DNA in the binding
reactions (not shown). Under these binding conditions, the stable
association of TFIIIA with RNA-DNA heteroduplexes containing either
strand of the 5 S rRNA gene as DNA could be detected (Fig. 2B, lanes 1-10). A single complex
appears followed by other complexes that are increasingly retarded in
their mobility through the gel. We conclude that TFIIIA will bind to
RNA-DNA heteroduplexes containing 5 S DNA and RNA sequences. We next
examined the specificity of this interaction.
TFIIIA Binds to 5 S RNA-DNA Heteroduplexes
Nonspecifically
We examine the nature of TFIIIA association with
5 S RNA-DNA heteroduplexes in three ways. We initially determined the
binding affinity of TFIIIA to both 5 S heteroduplexes in comparison to
the interaction with 5 S DNA. TFIIIA interacts with both heteroduplexes
equivalently (Fig. 3). Quantitation of binding indicates that
TFIIIA binds to both heteroduplexes with dissociation constants of
2
10M in comparison to the
interaction with DNA that has a dissociation constant of 2
10M (see also Clemens et
al.(1993)). We next determined the specificity of disruption of
the complex of TFIIIA with the heteroduplex using specific and
nonspecific DNA sequences as competitors (Fig. 4). TFIIIA was
specifically competed off either RNA (Fig. 4A, lanes 1-8) or DNA (Fig. 4A, lanes
9-16) using specific 5 S DNA sequences as a competitor,
while no specific competition is detectable when TFIIIA is bound to a
heteroduplex containing the non-coding DNA strand of the 5 S rRNA gene (Fig. 4B, lanes 1-10). Similar results
were obtained with heteroduplexes containing the coding strand of the 5
S rRNA gene (not shown). We suggest that TFIIIA does not form a
specific complex with the 5 S heteroduplex that can be more efficiently
competed with specific compared with nonspecific DNA. This reflects the
relatively low affinity of TFIIIA for the RNA-DNA heteroduplex.
) or ``sense'' DNA-RNA
heteroduplexes (⊡). Autoradiographs of gel mobility shifts were
scanned with a laser densitometer, and the fraction of bound nucleic
acid was plotted against the TFIIIA concentration used. A titration
curve of TFIIIA binding to DNA is included for comparison
(
).
)
©1995 by The American Society for Biochemistry and Molecular Biology, Inc.
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
ABSTRACT
This article has been cited by other articles:
|
|
 |
|
  B. S. Shastry Transcription factor IIIA (TFIIIA) in the second decade J. Cell Sci., March 1, 1996; 109(3): 535 - 539. [Abstract] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| All ASBMB Journals | Molecular and Cellular Proteomics |
| Journal of Lipid Research | ASBMB Today |