|
|
|
|||||||
|
|||||||||
J. Biol. Chem., Vol. 276, Issue 50, 47296-47302, December 14, 2001
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
From the San Raffaele Scientific Institute, via Olgettina
58, 20132 Milan, Italy and Faculty of Medicine, Universitá
Vita-Salute San Raffaele, via Olgettina 58, 20132 Milan, Italy
Received for publication, August 9, 2001, and in revised form, September 20, 2001
Sox proteins, a subclass of high mobility
group box proteins, govern cell fate decisions by acting both as
classical transcription factors and architectural components of
chromatin. We aimed to demonstrate that the DNA bending activity of Sox
proteins is essential to regulate gene expression. We focused on mouse
Sox2, which participates in the transactivation of the Fgf4
(fibroblast growth factor
4) gene in the inner cell mass of the blastocyst. We
generated six substitutions in the high mobility group box of Sox2. One
mutant showed a reduced DNA bending activity on the Fgf4
enhancer (46° instead of 80°), which resulted in more powerful
transactivation compared with the wild type protein. We then selected
two single-base mutations in the Fgf4 enhancer that make
the DNA less bendable by the Sox2 protein. Again, a different DNA bend
(0° and 42° instead of 80°) resulted in a different activation of
transcription, but in this case reduced bending corresponded to
decreased transcription. We found that the opposite effect on
transcription of similar DNA bending angles is due to a 20°
difference in the relative orientation of the DNA bends, proving that a
correct three-dimensional geometry of enhanceosome complexes is
necessary to promote transcription.
Transcriptional activation of eukaryotic genes requires the
assembly of multiprotein complexes on promoter and enhancer sequences. The right combination of factors binding to specific DNA sequences ensures specificity, and the contacts between the different proteins increase the stability of the complex. However, the various
transcription factors are often bound to DNA at nonadjacent sites, and
direct contacts between them are possible only if they are juxtaposed through the deformation of the DNA helix. Long DNA fragments have sufficient flexibility and can form loops, bringing in close contact proteins bound to distant sites. On the contrary, fragments smaller than 500 bp1 are rather stiff
(1), and the formation of multiprotein complexes can require specific
proteins that introduce bends in DNA. Proteins of the HMG box family
have DNA bending activity and are thus expected to act as architectural factors.
The HMG box family is composed of chromatin proteins and transcription
factors that contain the DNA binding domain of the same name (2, 3).
All of these proteins share the ability to recognize distorted DNA
structures and introduce bends in DNA. However, although chromatin
proteins HMGB1 and 2 bind to DNA in a sequence-independent manner (4),
transcriptional regulators such as the sex-determining factor SRY (5)
are sequence-specific.
SRY is one of the best characterized DNA-bending proteins. It is
encoded on the Y chromosome and promotes the development of male
gonads. Through its HMG box, it recognizes sites similar to the
sequence A/TAACAAA/T (6), and, upon binding DNA from the minor groove,
it bends it at an 83° angle (7). A mutation that significantly
reduces the DNA bending activity of human SRY causes sex reversal (8).
However, a direct correlation between the bending and transactivational
activities of SRY could never be definitively established, because so
far no SRY target gene has been identified with certainty, precluding
direct experiments on transcriptional control.
To prove beyond doubt that the bending activity of HMG box proteins has
a functional role in the regulation of transcription, we focused our
attention on Sox2. Sox (SRY-related HMG box)
proteins are stage- and tissue-specific transcription factors that
specify cell fate during development (9). Sox2 contains an HMG box that
recognizes and bends the same DNA sites as SRY and has a transactivation domain at its C terminus (10). Sox2 starts to be
expressed at the blastocyst stage, in the inner cell mass; there, in collaboration with the octamer-binding protein Oct-3 (11), it
binds to the enhancer of the Fgf4 (fibroblast
growth factor 4) gene and
determines its expression (12, 13).
We generated six different mutants in the HMG box of Sox2. We found one
mutant with reduced bending activity; this mutant activates
Fgf4 transcription more than the wild type
protein. We also mutated the Sox2 binding site and selected for
single-nucleotide substitutions that render the DNA molecule stiffer
toward the action of wild type Sox2; again we found that different DNA
bending resulted in a different level of transcription. In this case, reduced bending decreased transcription. When we compared
the complexes obtained using the mutated protein and the mutated
binding site, we found a 20° difference in the relative orientation
of the DNA bends. Taken together, these data prove that a correct three-dimensional geometry of nucleoprotein complexes is crucial for
transcriptional activation by HMG box proteins.
Bacterial Expression Vectors and Protein
Purification--
Plasmid pGex-Sox2.HMG was constructed by cloning
into the EcoRI-SalI sites of plasmid pGex-4T1
(Amersham Pharmacia Biotech), a 387-bp fragment of mouse Sox2 cDNA
(nucleotides 449-835; Ref. 10) coding for the Sox2 HMG box plus 26 amino acids upstream and 24 amino acids downstream. The fragment was
obtained by PCR from mouse genomic DNA using primers Sox2for
(5'-GGGAATTCCAGCAAGCTTCGGGGGG-3') and Sox2rev
(5'-AAGTCGACTAGCTCGCCATGCTGTTCCC-3'). To introduce mutations in the
Sox2 HMG box, pGex-Sox2.HMG was used as template in two-step PCR
mutagenesis, using Sox2for and Sox2rev as external primers and six
pairs of internal mutagenic primers (substi- tutions in bold type):
M47Ifor: 5'-GAGGCCCATAAACGCCTTC-3', and M47Irev:
5'-GAAGGCGTTTATGGGCCTC-3'; N48Qfor:
5'-GAGGCCCATGCAAGCCTTC-3', and N48Qrev:
5'-GAAGGCTTGCATGGGCCTC-3'; R58Kfor:
5'-GGGGCAGAAGCGTAAGATGG-3', and R58Krev:
5'-CCATCTTACGCTTCTGCCCC-3'; K89Ifor:
5'-GAGACCGAGATACGGCCGT-3', and K89Irev:
5'-ACGGCCGTATCTCGGTCTC-3'; F92Sfor:
5'-CGGCCGTCCATCGACGAG-3', and F92Srev:
5'-CTCGTCGATGGACGGCCG-3', and Y112Ffor:
5'-CGGATTATAAATTCCGGCCGC-3', and Y112Frev:
5'-GCGGCCGGAATTTATAATCCG-3'. The final PCR products
were cloned into the EcoRI-SalI sites of pGex-4T1
to obtain the mutant pGex-Sox2.HMG plasmids. Sequences were confirmed by double-stranded sequencing. Plasmid pGex-Oct-3.POU was kindly provided by Hans Schöler (University of Pennsylvania).
Glutathione S-transferase (GST) fusion proteins were
expressed in Escherichia coli and purified on
glutathione-Sepharose (Amersham Pharmacia Biotech) according to the
manufacturer's protocol. The GST moiety was separated from Sox2.HMG by
incubating the fusion protein bound to glutathione-Sepharose beads for
2 h at room temperature with thrombin (final concentration, 2 units/µl in phosphate-buffered saline). The Sox2 HMG boxes (wt and
mutant) released in the buffer were quantified by Coomassie staining in
comparison with a known marker and stored at Electrophoretic Mobility Shift Assay--
The
Determination of the DNA-Protein Complex Dissociation
Constants--
The amount of active protein in the preparations was
determined by competitive EMSA, adding increasing quantities (0-400
nM) of unlabeled probe to binding mixture containing a
small amount of labeled Fgf probe (0.1 nM) and 150 nM Sox2 HMG box. Incubation and electrophoresis were
carried out as described in the previous paragraph. The radioactivity
present in the bands was measured by exposing the dried gel to
PhosphorImager screens (Molecular Dynamics). The concentration of
active protein was deduced from the concentration of unlabeled probe
that starts to displace the labeled probe.
To determine the dissociation constants of the protein-DNA complexes,
wt and mutant Sox2 HMG boxes were titrated into binding mixtures
containing a fixed amount (0.1 nM) of probe. The samples were electrophoresed, and the radioactivity was measured as described in the previous paragraph. Under conditions of protein excess, the
dissociation constant is equivalent to the concentration of polypeptide
that binds 50% of the probe.
Calculation of DNA Bend Parameters--
For circular permutation
analysis, plasmids containing wt and mutant Fgf4 enhancers
(pBend2-Fgf4, pBend2-Fgf4.mut4 and pBend2-Fgf4.mut11) were prepared by
insertion of annealed synthetic oligonucleotides (Bend-Fgf4:
5'-CTAGAACTCTTTGTTTGGATGCTAATGGGA-3' and
5'-TCGATCCCATTAGCATCCAAACAAAGAGTT-3'; Bend-Fgf4.mut11:
5'-CTAGAACTCTTTATTTGGATGCTAATGGGA-3' and
5'-TCGATCCCATTAGCATCCAAATAAAGAGTT-3'; and Ben2-Fgf4.mut4:
5'-CTAGAACTCTTTGGTTGGATGCTAATGGGA-3' and
5'-TCGATCCCATTAGCATCCAACCAAAGAGTT-3') between XbaI and
SalI restriction sites in plasmid pBend2. Preparation of
probes, electrophoresis, and data analysis were carried out as
previously reported (7).
The series of phasing vectors was constructed by cloning the
oligonucleotide pairs described above into the
EcoRI-SalI sites of pSB10 81-86 vectors (15).
Probes were prepared by end-labeling RsaI-PvuII
fragments with 32P. Each probe was incubated in binding
mixtures containing wt or N48Q Sox2 HMG box in concentrations equal to
the KD of the complexes. After 10 min of incubation
in ice, the samples were electrophoresed at 4 °C in 10%
polyacrylamide gels at 10 V/cm for 18-20 h. The relative mobilities
(complex mobility divided by probe mobility) were plotted as a function
of the length of the spacer between the Sox2 binding site and the
intrinsic DNA bend, and the best fit to a cosine function (phasing
function) was determined (Prism software). The relative orientations of DNA bends were determined from the minima of the phasing functions. The
helical periodicity was assumed to be 10.5 bp/turn of the DNA helix.
Selection of Mutant Sox2 Binding Sites--
Sequences recognized
by the wt Sox2 HMG box, but intrinsically less bendable than the wt
sequence, were selected from a pool of 131-bp DNA molecules containing
mutant Sox2-binding sites in the middle of the fragment (D probe).
Plasmid pBend2-Fgf4 was used as template in two-step PCR muta-
genesis using as external primers bend.for (5'-AGATATCCAGCTGCCC-3')
and bend.rev (5'-GGATATCTTTAAACTCGAG-3') and a pair of
degenerated primers, where one nucleotide of the Sox2 binding site was
individually substituted by the other three nucleotides: fgf1,
CTAGAACTCVTTGTTTGG and CCAAACAABGAGTTCTAG; fgf2, CTAGAACTCTVTGTTTGG and
CCAAACABAGAGTTCTAG; fgf3, GAACTCTTVGTTTGGATG and
CCAAACBAAGAGTTCTAG; fgf4, GAACTCTTTHTTTGGATG and
CATCCAAADAAAGAGTTC; fgf5, GAACTCTTTGVTTGGATG and
CATCCAABCAAAGAGTTC; fgf6, AACTCTTTGTVTGGATGC and
CATCCABACAAAGAGTTC; and fgf7, CTTTGTTVGGATGCTAATG
and CATTAGCATCCBAACAAAG, where H = A/C/T, B = C/G/T, V = A/C/G, and D = A/T/G.
The seven PCR products were mixed together, obtaining a pool of 21 different DNA molecules containing mutant Sox2 binding sites in the
middle. The pool was end-labeled with 32P and used as probe
in a band shift assay with wt Sox2 HMG box in the conditions described
above. After autoradiography, shifted bands corresponding to lower
mobility complexes were cut out from the gel, and DNA was electroeluted
and reamplified by PCR. The PCR product was end-labeled with
32P and used again as probe in a band shift assay. Four
rounds of selection/amplification were performed to enrich the pool of
less bendable molecules, and finally the molecules were cloned into the
EcoRV restriction site of pBluescript (Stratagene) and
sequenced. Two of the mutations (mut4 and mut11) were inserted into
pBend2-Fgf4 and tested in circular permutation assays.
Transient Transfection Assays--
Plasmid pT81-Fgf4 was
constructed by cloning into the BamHI-KpnI sites
of pT81-luc (ATCC) a 286-bp fragment (nucleotides 2939-3224) containing the Fgf4 enhancer (14), obtained by PCR from
mouse genomic DNA with primers enhFgf.for
(5'-GCGGATCCTTAGCTCGCTTCAGG-3') and enhFgf.rev:
(5'-GCGGTACCGAGCCACCAGACAGAAAG-3'). To introduce in the Fgf4
en- hancer the nucleotide substitutions that decrease DNA flexibility,
pT81-Fgf4 was used as template in two-step PCR mutagenesis using
enhFgf.for and enhFgf.rev together with the two oligonucleotide pairs
mut4 (5'-GAAAACTCTTTGGTTGGATGC-3' and 5'-CTTTTGAGAAACCAACCTACG-3') and
mut11 (5'-GAAAACTCTTTATTTGGATGC-3' and 5'-CTTTTGAGAAATAAACCTACG-3').
The final PCR products were cloned into
BamHI-KpnI sites of pT81-luc, obtaining the
pT81-Fgf4mut4 and pT81-Fgf4mut11.
Plasmid pSG5-Sox2 was constructed by cloning into the BamHI
site of pSG5 (Stratagene) a 977-bp fragment (nucleotides 406-1383) corresponding to the Sox2 coding region (10). To construct variants of
pSG5-Sox2, HindIII-SacII fragments obtained from
each mutant pGex-Sox2.HMG were cloned into the same sites of wt
pSG5-Sox2, substituting the wt sequence.
HeLa and 3T3 cells (~1.5 × 105 cells in a 6-cm
dish) were transfected by the calcium phosphate method with various
combinations of reporter plasmids (pT81-Fgf4, pT81-Fgf4mut4, and
pT81-Fgf4mut11), wt and mutant Sox2 expression plasmids, and pCMV-OCT3.
PGK- Mutagenesis of the Sox2 HMG Box--
The DNA binding and bending
activities of HMG box proteins are completely specified by their HMG
box. We therefore studied the interaction between the Sox2 HMG box and
its cognate DNA target site.
We generated six point mutations within the HMG box to find mutant
proteins that bend DNA with different angles. The sequences of Sox2 and
SRY HMG boxes are very similar (about 85% identity); we then used the
structure of the human SRY HMG box-DNA complex (16) as a model to
identify the Sox2 residues involved in contacts to DNA. In SRY, five
amino acids (Met-64, Asn-65, Phe-67, Ile-68, and Trp-98) form a
T-shaped wedge in direct contact with the central base pairs of the DNA
binding site. They mediate the widening of the minor groove and the
bending of the helix toward the major groove. Mutation of Met-64 alters
DNA bending and causes sex reversal (8). We generated in Sox2 the
mutations M47I, K89I, and F92S, which correspond to SRY mutations that
cause sex reversal; we also selected in Sox2 Asn-48, adjacent to
Met-47; Tyr-112, corresponding to a SRY residue that contacts the 5'
part of the binding site; and Arg-58, which corresponds to a SRY
residue that forms salt bridges to the phosphate backbone. These
residues were replaced by conservative substitutions (Asn to Gln, Arg
to Lys, and Tyr to Phe) to avoid radical modifications of the
protein-DNA interaction surfaces. We produced wt and mutant HMG boxes
in E. coli (see "Experimental Procedures"), and we
tested in vitro their DNA binding and bending activities.
Analysis of DNA Binding Activity of wt and Mutant Sox2--
We
then determined the dissociation constant (KD) of
the DNA-protein complexes formed by the wt and mutant Sox2 HMG boxes
with the Fgf4 enhancer (only the concentration of active protein was considered). The seven polypeptides were titrated on a
limiting amount of probe; KD is equal to the
concentration of active protein that binds 50% of the probe. Fig.
1 shows titrations with wt HMG box and
mutant N48Q; the KD values of all polypeptides are
summarized in Table I. The
KD determined for wt Sox2 HMG box (1.5 × 10 DNA Bending Activity of wt and Mutant HMG Boxes--
The effect of
the mutations on the DNA bending activity was assayed by circular
permutation analysis (CPA). This is a standard method to investigate
distortions in DNA, based on the observation that a fragment containing
a bend near the end migrates faster than a fragment with a bend in the
middle (18). The Fgf4 binding site was cloned into the
pBend2 vector between two repeated 121-bp units, and seven circularly
permutated fragments were obtained by restriction enzyme digestion
(Fig. 2A) and used in EMSA as probes with wt and mutant HMG boxes. On the basis of three independent experiments we estimated that wt Sox2 induces an 80° deviation of the
DNA axis from linearity. This value is close to that determined for SRY
(83°; Ref. 7). Mutants M47I, R58K, K89I, F92S, and Y112F distort DNA
exactly as the wt protein, whereas N48Q is significantly different
(46°; Fig. 2B). A summary of the biochemical properties of
the seven polypeptides is presented in Table I.
Effect of Oct-3 on Sox2 DNA Binding and Bending
Activities--
The transcriptional activation of Fgf4 gene
is mediated by the synergistic action of Sox2 and Oct-3 (12, 13). The
two proteins bind contiguous target sites in the Fgf4
enhancer and interact with each other through the POU domain and the
HMG box. We then checked the influence of Oct-3 on the DNA binding and bending activities of the Sox2 HMG box. Titration and circular permutation experiments were performed again in the presence of the
Oct-3 DNA-binding region, composed by the POU-specific domain and the
POU homeodomain. In these conditions, a ternary complex is formed by
the Sox2 HMG box, the Oct-3 POU domain, and DNA. The apparent DNA
binding affinity of both wt and N48Q HMG box is increased by the
presence of the Oct-3 POU domain (data not shown). On the other hand,
Oct-3 has little effect on the DNA bending activity of Sox2 (Fig.
3). Two additional sets of bands appear.
The middle one corresponds to HMG box-POU domain-DNA complexes, and the upper one corresponds to two HMG boxes bound to DNA. The bending angle calculated for the ternary complex (77°) is very close
to that estimated for the HMG box-DNA complex (80°). We conclude that
the bending activity of Sox2 is not significantly affected by
Oct-3.
The N48Q Mutation Enhances the Transcriptional Activity of
Sox2--
We then correlated the N48Q bending activity with its
transcriptional activity. 3T3 and HeLa cells were transiently
co-transfected with fixed amounts of Oct-3 expressing vector and a
reporter plasmid driven by the Fgf4 enhancer (Fig.
4A), together with increasing amounts of expression vectors for either wt or N48Q Sox2 (full-length proteins). Wt Sox2 determines, at the maximum level, a 7-fold induction
of luciferase expression (Fig. 4B), whereas the mutant N48Q
reaches a 21-fold induction. At low concentrations, N48Q is actually
less effective than the wt protein, and this is probably due to its
reduced DNA binding affinity. At high concentrations, all DNA target
sites are presumably saturated even by the mutant, and differences in
the binding affinities become irrelevant, whereas the transcriptional
activities must depend exclusively on the bending activity of Sox2. We
conclude that the bending activity of Sox2 affects transcriptional
control directly.
DNA Mutations That Affect DNA Bending by Sox2--
To further
confirm the correlation between DNA bending and transcription, we used
a complementary approach. We generated a population of mutated Sox2
binding sites (Fig. 5A) by
substituting individually every nucleotide to find DNA molecules that
are intrinsically stiffer and less bendable by wild type Sox2 protein.
By PCR we obtained a mixed population of 131-bp fragments containing
the mutated Sox2 binding sites in the middle, and we used it as probe in EMSA with wt Sox2. Because of the position of the site, the migration of DNA-protein complexes is very sensitive to differences in
the bending angle. We recovered from the gel the complexes with higher
mobility than the wt sequence, containing less distorted DNA fragments.
After several cycles of selection and amplification we ended up with
fragments with reduced bendability that were cloned and sequenced. The
mutant sequences were inserted in the pBend-Fgf4 plasmid and compared
in CPA. Mutant 11 substitutes the fourth nucleotide of the binding site
(A instead of G), and this change abolishes distortion by Sox2 (Fig.
5C). In mutant 4 the fifth nucleotide is G instead of a T,
and this reduces the bending induced by wt Sox2 to 46° (Fig.
5B).
By itself, this result confirms that the DNA sequence determines the
general architecture of nucleoprotein complexes, and even slight
variations can affect severely the overall geometry. We then tested
whether a different DNA bend (obtained through modification of the
binding site rather than the mutation of the binding protein) could
result in a difference in transcription.
Reduced DNA Flexibility Affects Transactivation Mediated by
Sox2--
The nucleotide substitutions reducing DNA flexibility (mut4
and mut11) were inserted in the reporter plasmid pT81-Fgf4 and tested
in transient co-transfections of wt Sox2 and Oct-3. Fig. 6 shows that a different DNA bend
determines a different activation of transcription. In this case, a
smaller distortion leads to a reduced activation rather than an
increase as seen with N48Q Sox2 protein. The planar angles formed by wt
Sox2 protein on the mut4 site and by N48Q Sox2 protein on the wt site
are similar, as measured by CPA. To unravel this apparent paradox we
made the hypothesis that the three-dimensional geometries of the
complexes N48Q/Fgf4 and wtSox2/Fgf4mut4 were
different. CPA is a good tool to detect general protein-induced
perturbations of DNA, including static bending, increased DNA
flexibility, and aberrant protein structures; however, it gives no
information concerning the direction of the DNA bend. A more specific
method for the study of protein-directed bends is phasing analysis,
which allows delineation of the direction of a DNA distortion in space
(19).
Phasing Analysis--
Phasing analysis utilizes as probes in EMSA
DNA fragments having an intrinsic bend at different distances from the
site of induced bending. The spacing between the intrinsic and the
induced bends is incrementally varied over one turn of the DNA helix
(Fig. 7A); thus, there is a
spacing where the two bends are in phase and form a complementary
larger angle and a spacing where the two bends are out of phase and
have the effect of straightening the DNA fragment. The mobility of the
complexes in a polyacrylamide gel is slowest when the two bends add up
(cis-isomer) and highest when the two bends counteract each
other (trans-isomer).
We compared by phasing analysis the complexes wtSox2/Fgf4,
N48Q/Fgf4, and wtSox2/Fgf4mut4 (Fig.
7B). We plotted the relative mobilities of the complexes as
a function of the linker length (Fig. 7C), and we determined
for each complex the spacer length that gives the lowest complex
mobility. The minima for N48Q/Fgf4 and for
wtSox2/Fgf4 are indistinguishable, indicating that the bends
point in the same direction. On the contrary, a significant change was
observed in the case of wtSox2/Fgf4mut4, the spacer length
giving the cis isomer differs by 0.6 bp in comparison with wtSox2/Fgf4. Assuming an average of 10.5 bp/turn of the DNA
helix, this corresponds to a 20° difference in the relative
orientations of the DNA bends. These data prove that the mutations in
Sox2 protein (N48Q) and the Fgf4 enhancer (mut4) determine
different alterations in the architecture of the corresponding
nucleoprotein complexes, compared with the wt complex.
In this study we investigated the role of the bending activity of
the Sox2 transcription factor in the transcriptional activation of the
Fgf4 gene. We first quantified the affinity of the Sox2 HMG
box to its target site. The estimated KD = 1.5 × 10 These results indicate that the interactions with DNA of SRY and Sox2
are similar. Taking advantage of this similarity, we introduced
mutations into the Sox2 HMG box that were presumed to alter its DNA
bending activity, using SRY as a guide. The most dramatic effect was
obtained by substituting Asn-48 with Gln (N48Q mutant); the bending
activity was strongly reduced. The substituted Asn corresponds to one
of the five amino acids of SRY that enter into the minor groove and
induce the distortion of the DNA helix (16). The side chain of this Asn
is involved in hydrogen bonding and electrostatic interactions with the
fourth and fifth base pairs of the binding site. The Asn An unexpected result came from mutant M47I. In SRY, the corresponding
substitution causes a reduction in the bending activity of the protein,
which phenotypically results in sex reversal (8). The mutation in Sox2,
on the contrary, has no effect. This is even more surprising if we
consider that the adjacent Asn is indeed involved in DNA bending.
Obviously, the fine details of the interactions of SRY and Sox with DNA
are not completely identical.
Our goal was to prove that the ability of Sox2 to activate
transcription is dependent on its ability to bend DNA. Indeed we found
that the reduction of the bending activity in mutant N48Q determines a
3-fold increase of the protein transactivation potential, indicating
that at least one of the mechanisms used by Sox2 to promote
transcription is based on its capacity to manipulate DNA geometry. At
first glance, the finding that reduced bending corresponds to increased
transcription seems surprising. A similar result, however, was found in
the case of IHF, an E. coli protein that activates
transcription at several promoters by inducing a 180° DNA bend, thus
promoting DNA-RNA polymerase interactions (22). In a random mutation
screening, mutants of IHF that maximize transcriptional activation were
identified; a reduction in the bending activity corresponded to
stronger activation (23). Three-dimensional models showed that wt IHF
bends DNA more sharply than is necessary for DNA back looping to occur,
and the optimal bending angle is much less than the 180° turn imposed
by wt IHF. Even though the molecular mechanisms underlying IHF- and
Sox2-mediated transactivation are different, our data can have a
similar explanation; the distortion induced by wt Sox2 might be larger
than optimal to promote contacts among the proteins involved in
Fgf4 transactivation.
One might ask why evolution selected a protein that is nonoptimal in
its capacity for transcriptional activation. Sox2 regulates expression
of In the second part of our work, we confirmed the relation between DNA
bending and transcriptional activity by modulating DNA flexibility. The
sequence of the Sox2 binding site affects the flexibility of the DNA
molecule; single nucleotide substitutions reduce (mut4) or completely
abolish (mut11) the ability of wt Sox2 protein to distort DNA. The mut4
sequence is bent by wt Sox2 by an angle similar to the one produced by
the mutant N48Q protein on the wild type sequence, as measured by CPA.
Unexpectedly, the substitutions in the protein and in the DNA site had
opposite effects on transcription. We explained this apparent paradox
by means of phasing analysis; we found that the mutations in the protein and in the binding sites affect the three-dimensional geometry
of the DNA-protein complex in different ways. In particular, the
orientation of the directed DNA bend induced by N48Q is essentially identical to that induced by wt Sox2. On the contrary, when mut4 is
bound by the wt protein, the bend points 20° away from the direction
of the wild type bend, most likely preventing the correct assembly of
the nucleoprotein complex on the Fgf4 enhancer.
In conclusion, in this work we have directly confirmed the hypothesis
that an essential component of the action of Sox proteins is DNA
bending. Furthermore, such bending must be precise both in amplitude
and in direction to provide a correct framework for the rest of the
transcriptional machinery to operate.
We thank Dr. H. R. Schöler for
providing reagents. We are very grateful to Dr. S. Guazzi for
preliminary experiments.
*
This work was supported by grants from Associazione
Italiana Ricerca Cancro and the European Union program (to M. E. B.).The costs of publication of this
article were defrayed in part by the
payment of page charges. The article
must therefore be hereby marked
"advertisement" in
accordance with 18 U.S.C. Section
1734 solely to indicate this fact.
Published, JBC Papers in Press, October 2, 2001, DOI 10.1074/jbc.M107619200
The abbreviations used are:
bp, base pair(s);
aa, amino acid;
CPA, circular permutation analysis;
GST, glutathione
S-transferase;
HMG, high mobility group;
PCR, polymerase
chain reaction;
wt, wild type;
EMSA, electrophoretic mobility shift
assay.
Spatially Precise DNA Bending Is an Essential Activity of the
Sox2 Transcription Factor*
ABSTRACT
TOP
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
REFERENCES
INTRODUCTION
TOP
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
REFERENCES
EXPERIMENTAL PROCEDURES
TOP
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
REFERENCES
80 °C in
phosphate-buffered saline containing 0.1 mM dithiothreitol
and 0.05% Nonidet P-40.
-32P-labeled Fgf probe was made of annealed
oligonucleotides Fgf-probe.for (5'-ctagAACTCTTTGTTTGGATGCTAATGGGA-3')
and Fgf-probe.rev (5'-tcgaTCCCATTAGCATCCAAACAAAGAGTT-3'). The Fgf
probe reproduces the sequence of the murine Fgf4 enhancer
(14) and contains adjacent Sox2 and Oct-3 binding sites. Purified
polypeptides were incorporated in the indicated amounts, in a final
volume of 10 µl. Binding buffers, electrophoresis, and
autoradiography were as desribed (7). The Oct-3 POU domain alone
has a low solubility after removal of the GST moiety; thus, the binding
reactions with the Oct-3 POU domain were performed by incubating,
first, the Sox2 HMG box with ~500 ng of GST-OCT-3.POU bound to
glutathione-Sepharose in 8 µl of binding buffer, allowing the two
proteins to interact. After 45 min at 4 °C, the probe (0.1 nM) was added together with thrombin (0.01 unit) to remove
the GST, and after 40 min at room temperature the samples were applied
to the gel.
-gal was used as normalizing vector. The total amount of
transfected DNA was adjusted to 9 µg with empty pBluescript. The
amounts of plasmids used in each experiment are reported in the legends
of the figures.
RESULTS
TOP
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
REFERENCES
8 M) is similar to those estimated
for SRY and other Sox proteins (8, 17). The DNA affinities of mutants
M47I, K89I, F92S, and Y112F were indistinguishable from that of the wt,
whereas mutants N48Q and R58K showed 8- and 12-fold reductions,
respectively. Thus, all of the mutants bind the Fgf4 target
site, and only two have slightly reduced affinity.
View larger version (31K):
[in a new window]
Fig. 1.
Determination of the affinity of wt and
mutant Sox2 HMG boxes to the Fgf4 enhancer.
A, increasing amounts of each polypeptide, indicated at the
tops of the lanes, were incubated in 10 µl of
binding buffer with a fixed amount of Fgf probe (0.1 nM).
EMSA and quantification of the bands are described under
"Experimental Procedures." Arrow F, free Fgf probe;
arrow B, wt and mutant HMG box/DNA complex. B,
fraction of bound Fgf probe as a function of active protein
concentration. The KD corresponds to the protein
concentration that binds 50% of the probe (mean of at least three
independent experiments with each polypeptide).
DNA affinity and bending activity of Sox2 mutants
View larger version (24K):
[in a new window]
Fig. 2.
Mutation N48Q reduces Sox2 DNA bending
activity. A, DNA fragments used as probes in CPA.
Plasmid pBend2-Fgf4, containing the wt Fgf4 enhancer
(black box) flanked by tandemly repeated DNA sequences, was
cleaved at the restriction sites indicated in the map. The DNA
fragments (designated A-G) all contain circular
permutations of the same sequence of 146 bp. B, circular
permutation analysis of DNA bending induced by wt and mutant Sox2 HMG
boxes. 3 fmol of labeled DNA probes were incubated in 10 µl of
binding buffer with wt and mutant HMG boxes (wt, 25 nM;
N48Q, 300 nM). The retarded bands running behind the main
protein-DNA complex band are complexes containing more than one
polypeptide molecule/DNA molecule. C, calculation of DNA
bending angles. The mobility of the protein-DNA complexes
(Rbound) was normalized to the mobility of the
corresponding free probe (Rfree). The distance
of the center of the Sox2 binding site from the 5' end of the fragment
was divided by the total length of the probe (flexure displacement,
D/L). The plotted points were interpolated with
quadratic functions: y = 1.508x2 1.516x + 0.907 (r2 = 0.994) for wt
Sox2, and y = 0.470x2 0.500x + 0.777 (r2 = 0.952) for N48Q.
The first and second order parameters of the equation are in close
agreement and yield an estimate of flexure angles of 80° for wt Sox2
and 46° for N48Q.
View larger version (57K):
[in a new window]
Fig. 3.
Oct-3 does not affect Sox2 DNA bending
activity. Wild type Sox2 HMG box (25 nM) was incubated
at 4 °C with or without 1 µM of Oct-3 POU domain and
circular permutation probes (0.3 nM) (see "Experimental
Procedures" for details). Arrow Fr, free DNA; arrow
B1, HMG box-DNA complex; arrow B2; HMG box-POU
domain-DNA ternary complex; arrow B3, complex with two HMG
boxes binding to the same DNA molecule. The bending angles induced by
the Sox2 HMG box alone (B1 complexes) and by the HMG box in cooperation
with the Oct-3 POU domain (B2 complexes) are not significantly
different.
View larger version (22K):
[in a new window]
Fig. 4.
Mutation N48Q increases the transactivational
potential of Sox2 protein. A, diagram representing the
reporter plasmid pT81-Fgf4. The Fgf4 enhancer, containing
Sox2, Oct-3, and Sp1 binding sites, was cloned in front of the TK
promoter and luciferase coding region. B, transcriptional
activities of wt and N48Q Sox2 proteins. 3T3 fibroblasts were
transiently transfected with Sox2 expression plasmids (0, 0.2, 0.5, 1, 2.5, 3, 3.5, and 4 µg) and pT81-Fgf4 reporter plasmid (2 µg).
PGK- -gal plasmid expression was used for internal normalization. All
transfections were carried out in triplicate, in three independent
experiments.
View larger version (112K):
[in a new window]
Fig. 5.
Single-nucleotide substitutions in the Sox2
binding site reduce DNA flexibility. A, single
nucleotides in the Sox2 binding site of the Fgf4 enhancer
were individually substituted, obtaining a pool of 21 mutants. From
these, mut4 and mut11 were selected as having reduced flexibility.
B and C, circular permutation analysis of DNA
bending induced by wt Sox2 HMG box in mut4 (B) and mut11
(C). 3 fmol of circularly permutated DNA probes
(A-G) containing either mutation were incubated in 10 µl
of binding buffer with wt protein (25 nM). Bending angles
were calculated as described in the legend to Fig. 3.
View larger version (16K):
[in a new window]
Fig. 6.
Mutations mut4 and mut11 decrease
Sox2-mediated transactivation of the Fgf4
enhancer. 3T3 fibroblasts were transfected with 2 µg of
either wt pT81-Fgf4, pT81-Fgf4mut4, or pT81-Fgf4mut11, and increasing
amounts of wt pSG5-Sox2 (0, 0.2, 0.5, 1, 2.5, 3, 3.5, and 4 µg). All
transfections were carried out in triplicate, in three independent
experiments, as described in the legend to Fig. 4.
View larger version (33K):
[in a new window]
Fig. 7.
Different orientation of the DNA bends in
N48Q/Fgf4 and wtSox2/Fgf4mut4
complexes. A, DNA fragments used as probes in phasing
analysis contain either the wt or mut4 Fgf4 enhancer
(black box) separated from the (AT)6 intrinsic
bend (red line) by a spacer of variable length (open
boxes). The drawing represents the orientation in space of the
fixed bends. The numbers indicate the distance in base pairs
between the centers of the Sox2 binding site and the (AT)6
intrinsic bend. The total length of the fragments varies from 310 to
320 bp. B, phasing analysis. 3 fmol of labeled DNA probes
were incubated in 10 µl of binding buffer with either wt Sox2 or
mutant N48Q HMG boxes (wt, 25 nM; N48Q, 300 nM). C, calculation of the relative orientations
of the directed bends. The normalized complex mobilities (from
at least five independent experiments) were plotted as a function of
the length of the linker spacer, and the best fit to a cosine function
was determined. The minima of the phasing functions for
N48Q/Fgf4 and wtSox2/Fgf4mut4 differs by 0.6 bp,
which corresponds to a 20° difference in the relative orientation of
the directed bends, assuming a helical DNA periodicity of 10.5 bp. The
minima for N48Q/Fgf4 and wtSox2/Fgf4
coincide.
DISCUSSION
TOP
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
REFERENCES
8 M indicates a quite strong DNA binding
activity, comparable with that observed for other HMG box transcription
factors, such as SRY, LEF1, and Sox5 (8, 20, 21). We then calculated
that, upon binding, the HMG box induces an 80° distortion of the DNA double helix. We also proved that Oct-3, which cooperates with Sox2 in
Fgf4 regulation, does not affect the DNA bending activity of Sox2.
Gln
substitution adds a methylene group to the side chain, increases its
bulk, and thus affects the induced DNA bending. The DNA binding
affinity is also reduced 8-fold.
- and 
-crystallin in the lens (24), osteopontin during
hypoblast formation (25), and probably several other genes. The
expression of different genes might have different architectural
demands, and the entity of the distortion imposed by wt Sox2 is
possibly a compromise that allows a sufficient expression of all its
target genes. A different explanation is also possible; an increased
expression of Fgf4 is not necessarily an advantage for the
organism. Fgf4 is a secreted protein that acts within complex
regulatory pathways, in which intracellular and extracellular signals
cooperate to define the body plan of the embryo (26). Increased
Fgf4 expression could disrupt the delicate equilibrium that
regulates intercellular communication.
ACKNOWLEDGEMENTS
FOOTNOTES
To whom correspondence should be addressed. Fax:
39-0226434861; E-mail: bianchi.marco@hsr.it.
ABBREVIATIONS
REFERENCES
TOP
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
REFERENCES
1.
Travers, A. A.,
Ner, S. S.,
and Churchill, M. E. A.
(1994)
Cell
77,
167-169[CrossRef][Medline]
[Order article via Infotrieve]
2.
Bianchi, M. E.,
and Beltrame, M.
(1998)
Am. J. Hum. Genet.
63,
1573-1577[CrossRef][Medline]
[Order article via Infotrieve]
3.
Grosschedl, R.,
Giese, K.,
and Pagel, J.
(1994)
Trends Genet.
10,
94-100[CrossRef][Medline]
[Order article via Infotrieve]
4.
Thomas, J. O.,
and Travers, A. A.
(2001)
Trends Biochem. Sci.
26,
167-174[CrossRef][Medline]
[Order article via Infotrieve]
5.
Goodfellow, P. N.,
and Lovell-Badge, R.
(1993)
Annu. Rev. Genet.
27,
71-92[CrossRef][Medline]
[Order article via Infotrieve]
6.
Harley, V. R.,
Lovell-Badge, R.,
and Goodfellow, P. N.
(1994)
Nucleic Acids Res.
22,
1500-1501 7.
Ferrari, S.,
Harley, V. R.,
Pontiggia, A.,
Goodfellow, P. N.,
Lovell-Badge, R.,
and Bianchi, M. E.
(1992)
EMBO J.
11,
4497-4506[Medline]
[Order article via Infotrieve]
8.
Pontiggia, A.,
Rimini, R.,
Harley, V. R.,
Goodfellow, P. N.,
Lovell-Badge, R.,
and Bianchi, M. E.
(1994)
EMBO J.
13,
6115-6124[Medline]
[Order article via Infotrieve]
9.
Pevny, L. H.,
and Lovell-Badge, R.
(1997)
Curr. Opin. Genet. Dev.
7,
338-344[CrossRef][Medline]
[Order article via Infotrieve]
10.
Collignon, J.,
Sockanathan, S.,
Hacker, A.,
Cohen-Tannoudji, M.,
Norris, D.,
Rastan, S.,
Stevanovic, M.,
Goodfellow, P. N.,
and Lovell-Badge, R.
(1996)
Development
122,
509-520[Abstract]
11.
Schöler, H. R.,
Ruppert, S.,
Suzuki, N.,
Chowdhury, K.,
and Gruss, P.
(1990)
Nature
344,
435-439[CrossRef][Medline]
[Order article via Infotrieve]
12.
Yuan, H.,
Corbi, N.,
Basilico, C.,
and Dailey, L.
(1995)
Genes Dev.
9,
2635-2645 13.
Ambrosetti, D.-C.,
Basilico, C.,
and Dailey, L.
(1997)
Mol. Cell. Biol.
17,
6321-6329[Abstract]
14.
Brookes, S.,
Smith, R.,
Thurlow, J.,
Dickson, C.,
and Peters, G.
(1989)
Nucleic Acids Res.
17,
4037-4045 15.
Drak, J.,
and Crothers, D. M.
(1991)
Proc. Natl. Acad. Sci. U. S. A.
88,
3074-3078 16.
Werner, M. H.,
Huth, J. R.,
Gronenborn, A.,
and Clore, G. M.
(1995)
Cell
81,
705-714[CrossRef][Medline]
[Order article via Infotrieve]
17.
Connor, F.,
Wright, E.,
Denny, P.,
Koopman, P.,
and Ashworth, A.
(1995)
Nucleic Acids Res.
17,
3365-3372
18.
Wu, H.-M.,
and Crothers, D. M.
(1984)
Nature
308,
509-513[CrossRef][Medline]
[Order article via Infotrieve]
19.
Zinkel, S. S.,
and Crothers, D. M.
(1987)
Nature
328,
178-181[CrossRef][Medline]
[Order article via Infotrieve]
20.
Giese, K.,
Amsterdam, A.,
and Grosschedl, R.
(1991)
Genes Dev.
5,
2567-2578 21.
Connor, F.,
Cary, P. D.,
Read, C. M.,
Preston, N. S.,
Driscoll, P. C.,
Denny, P.,
Crane-Robinson, C.,
and Ashworth, A.
(1994)
Nucleic Acids Res.
22,
3339-3346 22.
Rice, P. A.,
Yang, S.,
Mizuuchi, K.,
and Nash, H. N.
(1996)
Cell
87,
1295-1306[CrossRef][Medline]
[Order article via Infotrieve]
23.
Engelhorn, M.,
and Geiselmann, J.
(1998)
Mol. Microbiol.
30,
431-441[CrossRef][Medline]
[Order article via Infotrieve]
24.
Kamachi, Y.,
Uchikawa, M.,
Collignon, J.,
Lovell-Badge, R.,
and Kondoh, H.
(1998)
Development
125,
2521-2535[Abstract]
25.
Botquin, V.,
Hess, H.,
Fuhrmann, G.,
Anastassidis, C.,
Gross, M. K.,
Vriend, G.,
and Schöler, H. R.
(1998)
Genes Dev.
12,
2073-2090 26.
Basilico, C.,
and Moscatelli, D.
(1992)
Adv. Cancer Res.
59,
115-165[Medline]
[Order article via Infotrieve]
Copyright © 2001 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:
|
|
 |
|
  S. Takiya, S. Saito, T. Yokoyama, D. Matsumoto, T. Aizawa, M. Kamiya, M. Demura, and K. Kawano DNA-Binding Property of the Novel DNA-Binding Domain STPR in FMBP-1 of the Silkworm Bombyx mori J. Biochem., July 1, 2009; 146(1): 103 - 111. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Leprivier, D. Baillat, A. Begue, B. Hartmann, and M. Aumercier Ets-1 p51 and p42 isoforms differentially modulate Stromelysin-1 promoter according to induced DNA bend orientation Nucleic Acids Res., May 21, 2009; (2009) gkp307v1. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. J. Formeister, A. L. Sionas, D. K. Lorance, C. L. Barkley, G. H. Lee, and S. T. Magness Distinct SOX9 levels differentially mark stem/progenitor populations and enteroendocrine cells of the small intestine epithelium Am J Physiol Gastrointest Liver Physiol, May 1, 2009; 296(5): G1108 - G1118. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Pei Regulation of Pluripotency and Reprogramming by Transcription Factors J. Biol. Chem., February 6, 2009; 284(6): 3365 - 3369. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. A. Atcha, A. Syed, B. Wu, N. P. Hoverter, N. N. Yokoyama, J.-H. T. Ting, J. E. Munguia, H. J. Mangalam, J. L. Marsh, and M. L. Waterman A Unique DNA Binding Domain Converts T-Cell Factors into Strong Wnt Effectors Mol. Cell. Biol., December 1, 2007; 27(23): 8352 - 8363. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 O. V. Taranova, S. T. Magness, B. M. Fagan, Y. Wu, N. Surzenko, S. R. Hutton, and L. H. Pevny SOX2 is a dose-dependent regulator of retinal neural progenitor competence. Genes & Dev., May 1, 2006; 20(9): 1187 - 1202. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. T. Ghashghaei, J. Weber, L. Pevny, R. Schmid, M. H. Schwab, K. C. K. Lloyd, D. D. Eisenstat, C. Lai, and E. S. Anton The role of neuregulin-ErbB4 interactions on the proliferation and organization of cells in the subventricular zone PNAS, February 7, 2006; 103(6): 1930 - 1935. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Pérez-Lago, M. Salas, and A. Camacho A precise DNA bend angle is essential for the function of the phage {phi}29 transcriptional regulator Nucleic Acids Res., January 7, 2005; 33(1): 126 - 134. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. E. Mysiak, C. Wyman, P. E. Holthuizen, and P. C. van der Vliet NFI and Oct-1 bend the Ad5 origin in the same direction leading to optimal DNA replication Nucleic Acids Res., December 1, 2004; 32(21): 6218 - 6225. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. A. Carr, J. Mead, and A. K. Vershon {alpha}1-induced DNA bending is required for transcriptional activation by the Mcm1-{alpha}1 complex Nucleic Acids Res., April 26, 2004; 32(8): 2298 - 2305. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. C. Williams Jr., M. Cai, and G. M. Clore Molecular Basis for Synergistic Transcriptional Activation by Oct1 and Sox2 Revealed from the Solution Structure of the 42-kDa Oct1{middle dot}Sox2{middle dot}Hoxb1-DNA Ternary Transcription Factor Complex J. Biol. Chem., January 9, 2004; 279(2): 1449 - 1457. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Zhang, T. Basta, E. D. Jensen, and M. W. Klymkowsky The {beta}-catenin/VegT-regulated early zygotic gene Xnr5 is a direct target of SOX3 regulation Development, December 1, 2003; 130(23): 5609 - 5624. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Zhang and M. L. Dufau Repression of the Luteinizing Hormone Receptor Gene Promoter by Cross Talk among EAR3/COUP-TFI, Sp1/Sp3, and TFIIB Mol. Cell. Biol., October 1, 2003; 23(19): 6958 - 6972. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. A. Wray, M. W. Hahn, E. Abouheif, J. P. Balhoff, M. Pizer, M. V. Rockman, and L. A. Romano The Evolution of Transcriptional Regulation in Eukaryotes Mol. Biol. Evol., September 1, 2003; 20(9): 1377 - 1419. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. Schlierf, A. Ludwig, K. Klenovsek, and M. Wegner Cooperative binding of Sox10 to DNA: requirements and consequences Nucleic Acids Res., December 15, 2002; 30(24): 5509 - 5516. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| All ASBMB Journals | Molecular and Cellular Proteomics |
| Journal of Lipid Research | ASBMB Today |