|
|
|
|||||||
|
|||||||||
(Received for publication, November 1, 1995; and in revised form, January 18, 1996) From the
The nuclear hormone receptors regulate target gene expression in
response to hormones of extracellular origin. The DNA binding
specificity of these receptors therefore plays the critical role of
defining the precise repertoire of target genes that respond to a given
hormone. We report here an analysis of the DNA binding specificity of
the thyroid hormone receptor (c-ErbA protein) and that of an oncogenic
derivative, the v-ErbA protein. These otherwise closely similar
proteins exhibit quite divergent DNA sequence specificities at multiple
positions within the DNA binding site. The thyroid hormone receptor
(c-ErbA protein) exhibits a particularly broad DNA specificity, whereas
the v-ErbA protein is comparatively quite specific. Intriguingly, these
differences in DNA recognition largely map to an N-terminal receptor
domain not traditionally implicated in DNA binding, and are further
influenced by heterodimer formation with retinoid X receptors. We
propose that the N terminus of nuclear hormone receptors plays an
critical role in DNA recognition by altering the conformation of the
receptor domains that make the actual base-specific contacts. The nuclear hormone receptors are a family of ligand-regulated
transcription factors that mediate cellular responses to a broad range
of small hydrophobic
hormones(1, 2, 3, 4, 5) .
Members of the family include the steroid receptors, the retinoid acid
receptors (RARs), ( The DNA recognition properties of each receptor play the critical
role of defining the specific repertoire of target genes that respond
to a given hormone. Most receptors bind to DNA as protein dimers, with
each receptor molecule binding to a ``half-site,'' a
conserved 6-8-nucleotide DNA
sequence(7, 8, 9, 10, 11) .
Recognition of the sequence of each half-site has generally been
believed to be mediated exclusively by a zinc-finger motif within the
center of each receptor (Fig. 1) (12, 13, 14) . Amino acids in the P-box helix
within this zinc-finger motif make direct contact with bases in the
major grove of the DNA half-site, and altering amino acids in the P-box
can alter the half-site specificity of the
receptor(14, 15, 16, 17, 18, 19, 20, 21, 22) .
An additional
Figure 1:
Schematic of the c-Erb A and v-Erb A
proteins. Schematic representations of c- and v-Erb A proteins are
presented (A) with the DNA binding and hormone binding domains
indicated. Alterations in v-Erb A relative to c-Erb A are also shown,
including N- and C-terminal deletions, 13 internal amino acid
substitutions (vertical bars), and N-terminal
retroviral-derived Gag sequences. Amino acids involved in the novel DNA
recognition properties of v-Erb A that were altered in this study
(mutants C32Y and S61G) are highlighted. Although having no known
effect on DNA half-site specificity, codon 78 (a lysine in c-Erb A and
a threonine in v-Erb A) was also altered in our studies to produce a
v-Erb A allele (S61G/T78K) with a fully c-Erb A-like zinc finger domain (18) . The relevant amino acid sequences are also presented (B). (c) denotes the c-Erb A sequence, and (v) denotes the v-Erb A sequence.
Aberrant forms of
nuclear hormone receptors are involved in several forms of cancer. The
v-erbA oncogene, for example, is a mutated copy of the host
cell gene (c-erbA) for
T The analogous N-terminal domains
of RARs, RXRs, and retinoid orphan receptors also play a critical role
in DNA half-site specificity (32, 33, 34, 35, 36) .
Unfortunately, this region of the nuclear receptors has not been
modeled in structural studies, leaving the precise molecular basis of
its actions unclear. Unlike the zinc-finger domain, it is doubtful that
the N terminus makes direct base-specific contacts with the DNA
half-site(14) . We describe here a detailed dissection of
the effects of these N-terminal receptor determinants on half-site
recognition. We report that the contributions of the N terminus to DNA
recognition specificity are at least as potent as those of the
zinc-finger domain itself and can broadly influence sequence
recognition over multiple locations in the half-site; we propose a
conformational model to explain these effects. We also demonstrate that
interactions with RXR, a heterodimeric partner for many nuclear hormone
receptors, has a differential effect on the DNA half-site recognition
properties of v-ErbA and c-ErbA, indicating that processes influencing
heterodimer formation by nuclear hormone receptors may also alter their
target gene specificity.
Figure 2:
Determination of the half-site
recognition specificity of the v-ErbA and T
Each of the 25 possible
permutated half-sites were synthesized as a direct repeat with a 4-base
spacer (DR-4), an arrangement found in many physiological response
elements for c-ErbA(8, 9) . We then tested the ability
of c-ErbA or v-ErbA to bind to these elements in vitro using
an electrophoretic mobility shift assay. An autoradiograph of a typical
assay is shown in Fig. 2, whereas the results from multiple
assays were quantified, averaged, and are displayed in Fig. 3.
Several observations are of note. (a) Although c-ErbA bound
best to the consensus TAAGGTCA half-site, an extensive set of
variations of this consensus were also strongly recognized by the
receptor. This was particularly evident for the base substitutions at
half-site positions -2, -1, +1, +4, +5, and
+6, most of which were accommodated by c-ErbA with relatively
modest effects on binding. (b) In contrast, most substitutions
in the consensus half-site at positions +2 and +3 were highly
destabilizing for c-ErbA binding (with the exception of the +3T
substitution, which retained moderate T
Figure 3:
Half-site recognition properties of
T
The
v-ErbA protein demonstrated a very different, and generally more
restrictive, DNA specificity pattern from that of c-ErbA. V-ErbA bound
the consensus half-site element relatively strongly. However, v-ErbA
bound weakly, or not at all, to many of the substitutions at +4,
+5, and +6 that were strongly bound by c-ErbA (Fig. 3B; presented as a difference plot in Fig. 3C). v-ErbA, in common with c-ErbA, also failed to
bind to most of the base substitutions at positions +2 and
+3. However, substitutions at -1 proved an intriguing
exception and were recognized as well, or better, by v-ErbA as by
c-ErbA. In fact, our results indicate that the optimal v-ErbA half-site
sequence was TGAGGTCA, in contrast to the optimal c-ErbA site,
TAAGGTCA, but consistent with results obtained by a randomized binding
site selection procedure(42) . Half-site sequences that
exhibited differential recognition by v- and c-ErbA in these direct DNA
binding studies were subsequently re-analyzed in DNA competition
experiments with essentially the same results (Table 1). We
conclude that differences in DNA recognition between v-ErbA and c-ErbA
extend throughout the DNA half-site and are manifested as a more
restrictive v-ErbA recognition of bases +3, +4, +5, and
+6, and a broader recognition of bases at -1.
Figure 4:
Half-site recognition properties of
mutant v-Erb A proteins. Electrophoretic mobility shift assays were
performed as in Fig. 2and Fig. 3, but using the C32Y
v-Erb A N-terminal domain mutant (A), the S61G/T78K v-Erb A
zinc-finger mutant (B), or the triple C32Y/S61G/T78K v-Erb A
mutant (C). The total radiolabel migrating as bound complex
was quantified and is expressed relative to binding of the consensus
TAAGGTCA half-site DR-4 element (100%). The same results, expressed as
a difference plot (binding by wild-type v-Erb A minus binding by each
mutant v-Erb A), are also shown (D-F). At least two
separate experiments were performed for each oligonucleotide
substitution; the average and standard deviation are
plotted.
Compared to the zinc-finger swap, just
changing codon 32 in the v-ErbA N terminus to that of c-ErbA (a mutant
denoted C32Y; Fig. 1) had a greater overall effect on DNA
recognition, significantly enhancing recognition of the +3T,
+4C, +4G, +5G, +5T, +6C, +6G, and
+6T substitutions, and decreasing binding to -1G (Fig. 4A). In these aspects the specificity of the C32Y
v-ErbA mutant nearly paralleled that of c-ErbA (compare Fig. 3A and 4A, and compare the difference
plots in Fig. 4D and 3C). Nonetheless, the
C32Y mutation alone failed to confer a strong, c-ErbA-like recognition
of the +4A or +5A substitutions. Indeed, a triple v-ErbA
mutation, denoted C32Y/S61G/T78K, significantly enhanced recognition of
the +4A and +5A substitutions and yielded an overall DNA
recognition pattern indistinguishable from that of c-ErbA (compare Fig. 4C and 3A, and compare the difference
plots in Fig. 4F and 3C). Parallel results
were obtained in DNA competition experiments (Table 1). We
conclude that virtually all the differences in DNA recognition noted
for v- and c-ErbA are accounted for by amino acid differences in the N
terminus and zinc-finger domains of these proteins, and that the
N-terminal domain plays a dominant role in this phenomenon.
Figure 5:
Effects of heterodimer formation with RXR
on DNA half-site specificity of T
Surprisingly, formation of heterodimers with RXR had
non-identical effects on v-ErbA relative to c-ErbA. In common with
c-ErbA, heterodimerization of v-ErbA with RXR enhanced binding to
half-site substitutions at the -2, +4, +5, and +6
positions (Fig. 5B). In contrast, however,
heterodimerization with RXR had much less of an effect on recognition
of the +1, +2, and +3 half-site positions by v-ErbA than
by c-ErbA (Fig. 5B). As a consequence, certain
differences between v- and c-ErbA in half-site recognition (at
positions -1, +5, and +6) were muted by
heterodimerization with RXR, whereas novel differences in half-site
recognition (at positions +1, +2, and +3) were created
(compare the difference plots in Fig. 5C and Fig. 3C).
How can this promiscuity
of DNA sequence recognition be reconciled with the specific base
contacts seen in structural studies? Studies with non-cognate elements
suggest that the precise receptor-DNA contacts can vary with modest
changes in orientation or conformation of the receptor accommodating
non-consensus sequences(12, 45, 46) . It is
also likely that the specific base pair-receptor contacts, typically
hydrogen bonds or van der Waals' interactions, contribute only
modestly to the change in free energy of receptor upon DNA binding,
compared with the stronger, sequence-independent ionic contacts between
the receptor and the DNA
backbone(12, 45, 46) . In fact, base-specific
interactions may function chiefly to prevent receptor binding to
inappropriate DNA sequences (by ``clashing'' sterically or
otherwise with the amino acids in the receptor P- and A-boxes) rather
than by contributing positively to binding by stabilizing contacts to
the correct base pairs.
How might the N-terminal region
be capable of such effects? The N-terminal domain has not been included
in the structural studies reported to date, and therefore its precise
actions must remain speculative. It appears unlikely, however, that the
v-ErbA N terminus can make direct contact with the bases in the AGGTCA
core hexanucleotide, precluding a direct effect on base pair
recognition. We also disfavor a model in which the N terminus mediates
its effects indirectly by altering the conformation of the target DNA;
although both v-ErbA and c-ErbA induce DNA bending, the extent of the
bend is independent of the nature of the N terminus. (
Volume 271,
Number 18,
Issue of May 3, 1996 pp. 10800-10805
©1996 by The American Society for Biochemistry and Molecular Biology, Inc.
UNEXPECTED DIVERSITY IN HALF-SITE SPECIFICITY CONTROLLED BY
NON-ZINC-FINGER DETERMINANTS (*)
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
FOOTNOTES
ACKNOWLEDGEMENTS
REFERENCES
)retinoid X receptors (RXRs), and the
thyroid hormone receptors (T
Rs). Additional diversity is
generated within individual receptor classes by the expression of
multiple receptor isoforms; for example, TRs are encoded by
two separate loci, 
and 
(3, 4) . Despite
this diversity, the different nuclear hormone receptors share a common
mode of action, functioning by binding to specific DNA sequences and
regulating expression of nearby target
genes(1, 2, 3, 4, 6, 7) . -helix, the A-box located at the C-terminal extreme
of the zinc-finger motif, makes minor groove contacts with bases at the
5` end of the half-site(14, 23) .
R-1(24, 25) . V-ErbA has sustained a
number of mutations relative to its TR progenitor (Fig. 1); as a result, v-ErbA is impaired in hormone binding and
in transcriptional activation and appears to function in the cancer
cell as a constitutive
repressor(26, 27, 28, 29, 30) .
In addition v-ErbA exhibits a distinct DNA specificity from that of
c-ErbA; although both v-ErbA and c-ErbA efficiently recognize
AGGTCA half-sites, only the c-ErbA polypeptide also recognizes AGGACA
half-sites(18, 20, 31, 32) . This
alteration in DNA specificity is due, in part, to the presence of a
serine at codon 61 in the P-box of v-ErbA that is a glycine in c-ErbA (Fig. 1) and is consistent with structural studies that place
codon 61 proximal to the fourth base of the
half-site(12, 13, 14, 18, 19, 20, 31) .
Unexpectedly, however, additional receptor domains N-terminal to the
zinc-finger also strongly influence the half-site specificity of the
ErbA polypeptides(18, 20, 32) . The amino
acid at codon 32, a cysteine in v-ErbA and a tyrosine in c-ErbA (Fig. 1), is particularly critical, and only by changing the
identity of both codon 32 and codon 61 is it possible to exchange the
DNA recognition properties of v-ErbA and
c-ErbA(18, 32) .
Source of Proteins and Oligonucleotide
Probes
The wild-type v-ErbA, avian c-ErbA-1, and avian
RXR
receptors were obtained as nuclear extracts from recombinant
baculovirus-infected Sf9 cells(18) . The C32Y, S61G/T78K, and
C32Y/S61G/T78K mutant v-ErbA polypeptides were isolated as glutathione S-transferase fusion proteins from Escherichia coli transfected with the appropriate pGEX recombinant plasmid
vectors(32, 37) ; no significant differences were
observed in the DNA binding specificities of ErbA proteins synthesized
in the bacterial system compared to the same polypeptide encoded in the
baculovirus system (data not shown). The different oligonucleotide
probes employed in the binding experiments were synthesized on a
Milligen Cyclone Plus using phosphoramidite chemistry. Each probe was
created as two complementary oligonucleotides with four base overhangs (e.g. for the consensus, 5`-TCGAA TAAGG TCAAA TAAGG TCAGA G-3`
and 5`-TCGAC TCTGA CCTTA TTTGA CCTTA T-3`), the oligonucleotides were
annealed, and the double-stranded DNAs were radiolabeled by fill-in
using [-
P]dGTP and Klenow fragment of E. coli DNA polymerase.DNA Binding Assays
Standard electrophoretic
mobility shift assays were employed as described
previously(18, 38) . Briefly, the radiolabeled
oligonucleotide probe (typically 1-3.5 
10
counts/min, approximately 20-30 ng of DNA) was incubated
with the protein of interest in 15 µl of binding buffer (10 mM Tris-Cl, pH 7.5, 50 mM KCl, 5% glycerol, 13.3 mg/ml
bovine serum albumin, and 133 µg/ml poly(deoxyinosine
deoxycytosine)) for 15 min at room temperature. Any protein-DNA
complexes that formed were then resolved by electrophoresis through a
4.5% polyacrylamide, 0.12% bisacrylamide gel at 200 V for 75 min. The
resulting electrophoretograms were dried, visualized by
autoradiography, and quantified by use of a Betascan analyzer. In the
absence of RXRs, c-ErbA and v-ErbA generally produced both monomeric
and dimeric complexes, probably due to the low cooperativity with which
the c-Erb A isoform forms homodimers(7) . Perhaps as a
consequence of this low cooperativity, the ratio of monomers to dimers
formed by a given ErbA protein did not significantly vary from one
response element to another ( Fig. 2and quantification data not
shown). Competition assays were performed by binding the receptor of
interest to 27 ng of the radiolabeled consensus sequence in the
presence of differing amounts (0, 125, 500, or 1000 ng) of the
unlabeled competing DNA.
R/c-ErbA
proteins by electrophoretic mobility shift assay. A series of
radiolabeled oligonucleotide DR+4 probes, consisting of single
base substitutions in the consensus half-site element TAAGGTCA, were
tested for the ability to bind to TR/c-ErbA (A) or
v-ErbA (B) in an electrophoretic mobility shift assay. Each
base substitution is identified, below the panel, as to location and
the nature of the base substitution; thus in our nomenclature,
-2A represents a AAGGTCA half-site sequence and +2T refers
to a TAATTCA half-site sequence, whereas +2G (or +4T, etc.)
is simply the consensus TAAGGTCA half-site. The locations of free
probe, probe bound to receptor monomers (arrowhead), and probe
bound to receptor dimers (arrows) are indicated. An asterisk indicates a non-specific probe-derived band observed
even in the absence of protein (data not
shown).
The Different DNA Binding Specificities of v-ErbA and
c-ErbA Extend over Multiple Positions within the DNA
Half-site
We wished to further explore the mechanisms by which
the N terminus of the ErbA protein influences DNA recognition
specificity. We first tested if the influence of the N-terminal and
P-box amino acids on DNA recognition was localized to the fourth base
position analyzed previously, or if their effects extended over
additional positions in the half-site. We began with an optimized
consensus half-site (TAAGGTCA), systematically altered each position to
the alternative three bases, and tested the ability of each substituted
DNA sequence to bind to ErbA. Our use of an 8-base half-site in the
current study follows from the recent demonstration that alterations in
bases 5` to the traditionally defined hexanucleotide half-site can also
influence DNA
recognition(21, 39, 40, 41, 42) .
These 5` bases are probably recognized through interactions with the
receptor A-box helix(14, 23) . To allow comparison
with previous work, we number the original hexanucleotide half-site
sequence as +1 thorough +6, and denote the two 5` bases as
the -2 and -1 positions.R binding).
R/c-ErbA and v-ErbA proteins: quantified data.
Electrophoretic mobility shift assays were performed as in Fig. 2using either the c-ErbA protein (panel A) or the
v-ErbA protein (panel B). The total radiolabel migrating as
bound complex was quantified and is expressed relative to binding of
the consensus TAAGGTCA half-site DR+4 element (= 100%). At
least two separate experiments were performed for each oligonucleotide
substitution; the average and standard deviation are plotted. To
facilitate comparison of the c-ErbA and v-ErbA binding specificity, the
results were also expressed as a difference plot (binding by v-ErbA
minus binding by c-ErbA (panel C).
The N-terminal Domain, Not the Zinc-finger region, Plays
the Dominant Role in Defining the Different Half-site Specificities of
v- and c-ErbA
We next examined the relative contributions of the
N-terminal and zinc-finger domains to the differing DNA half-site
specificities of c- and v-ErbA. The zinc-finger domain of the nuclear
hormone receptors mediates most of the known protein/base pair contacts (12, 13, 14) . Nonetheless, complete
replacement of the v-ErbA zinc-finger domain with that of c-ErbA (a
mutant denoted S61G/T78K, Fig. 1) produced only relatively
modest effects on the half-site specificity of the v-ErbA protein
(compare Fig. 3B and 4B). Most evident was an
increase in v-ErbA binding to the +4A substitution, although some
enhancement of binding to the +3T, +4C, +5A, +5G,
+5T, and +6G substitutions was also observed. A difference
plot (probe bound by wild-type v-ErbA minus that bound by the mutant; Fig. 4E) is also presented to allow comparison to Fig. 3C.
RXR Heterodimerization Further Broadens the Half-site
Recognition Properties of Both c-ErbA and v-ErbA, but in a
Non-equivalent Fashion
Many nuclear receptors can form
heterodimers with other members of the nuclear hormone receptor family
(reviewed in Refs. 43 and 44). In fact, RXRc-ErbA heterodimers
exhibit higher overall DNA binding affinity and enhanced target gene
activation than do c-ErbA homodimers, and may be the dominant form of
c-ErbA in living cells(43, 44) . We therefore tested
the effect of RXRc-ErbA heterodimerization on DNA half-site
specificity. Interestingly, heterodimerization with RXR further
extended the already broad DNA recognition properties of c-ErbA to
allow recognition of virtually all the single base substitutions
tested. Base substitutions at -2, -1, +4, +5 and
+6 that were moderately well recognized by c-ErbA homodimers were
now strongly recognized by RXRc-ErbA heterodimers (Fig. 5A). Furthermore, base substitutions at +1,
+2, and +3 that exhibited no detectable binding by c-ErbA
homodimers were now strongly recognized by the RXRc-ErbA
heterodimer (Fig. 5A). The ability of the
RXRc-ErbA heterodimer to bind to each permutated half-site is
expressed relative to its ability to bind the consensus sequence; thus
our results reflect actual alterations in sequence specificity. All
binding experiments were performed in DNA probe excess, under
conditions such that only the binding of receptor heterodimers was
measured; no binding of RXR homodimers was detected under these
conditions.
R/c-ErbA and v-ErbA.
Electrophoretic mobility shift assays were performed as in Fig. 2and 3, but using RXRc-ErbA heterodimers (panel
A) or RXRv-ErbA heterodimers (panel B) in place of
the homodimers previously described. The total radiolabel migrating as
bound complex was quantified and is expressed relative to binding of
the consensus TAAGGTCA half-site DR+4 element (= 100%). At
least two separate experiments were performed for each oligonucleotide
substitution; the average and standard deviation are plotted. To
facilitate comparison, the probe bound by the RXRErbA heterodimer
in each panel (open bars) is compared to the probe bound by
the analogous ErbA homodimer (closed bars). A difference plot
(binding by RXRv-ErbA minus binding by c-ErbARXR) is also
shown (panel C).
The c-ErbA Protein Exhibits a Broad Ability to
Accommodate Half-site Variations That Diverge from the Optimal
Consensus Sequence
Structural modeling indicates that the base
pairs in the DNA half-site are contacted through a web of multiple
interactions with amino acids in both the P- and A-box of the
receptor(12, 13, 14) . These multiple
protein-base-specific DNA contacts, both direct and through immobilized
water molecules, might be expected to impose a strict DNA sequence
specificity on these receptors. In practice, however, nuclear hormone
receptors exhibit a relatively promiscuous DNA sequence specificity and
conversely can accept a surprising variety of P- and A-box amino acid
substitutions without abolishing DNA
binding(18, 19, 20, 21, 22) .
In fact, as demonstrated here, c-ErbA homodimers exhibit a near
absolute specificity for the consensus half-site sequence only at
positions +2 and +3 (TAAGGTCA), whereas a broad range of
substitutions are permissible at nearly all other positions. Although
obtained with artificial elements, these results are consistent with
the considerable sequence variation found in naturally occurring
response elements(8, 9) .v-ErbA Exhibits a Restrictive Half-site Specificity That
Maps to a Single Base Change Outside of the Traditionally Defined DNA
Binding Domain
Although structurally closely related to c-ErbA,
the v-ErbA protein exhibits a much more narrow DNA sequence specificity
over much of the half-site sequence. Intriguingly, although there is a
difference in the P-box between v- and c-ErbA, at codon 61, alteration
of this amino acid produces only modest effects on DNA recognition. In
fact, in a crystallographic analysis(14) , this codon does not
make any base-specific contacts with the consensus half-site, although
it potentially can make unfavorable contacts with incorrect half-sites (19, 20, 22) . On the other hand, alteration
of codon 32 in the N terminus of the receptor (and outside of the
previously defined DNA binding domain) dramatically broadens the
sequence recognition properties of v-ErbA to more closely approximate
those of c-ErbA; mutation of both N-terminal and zinc-finger domains
confers a fully c-ErbA specificity.
)Instead, we suggest that the receptor N terminus acts
indirectly by influencing the tertiary conformation of the zinc-finger
P- and A-box helices. We propose that although the P- and A-box amino
acids make the actual discriminatory contacts with bases in the DNA,
the N terminus can alter the precise position or orientation of the
zinc-finger recognition helices and thus can define in a global fashion
the exact base contacts made, and the stringency of the protein-DNA
interaction. Perhaps, due to these interactions with the N terminus,
the P-box helix of c-ErbA is slightly tilted away from the 3` end of
the half-site relative to the P-box helix of v-ErbA, resulting in the
enhanced ability of c-ErbA to accommodate a broader range of base
substitutions over these positions. In fact, the zinc-finger domain is
known to exhibit some flexibility in its location relative to the DNA
major groove(45) . Furthermore, a conformational coupling
between the receptor N terminus and the zinc-finger domain is
consistent with previous mutagenesis studies, and with the apparent
proximity of these two domains, as suggested by the limited
crystallographic data available(14, 32) .RXR Heterodimer Formation Alters the DNA Half-site
Specificity of v- and c-ErbA, but in a Non-equivalent
Manner
Many nuclear hormone receptors can bind to DNA both as
homodimers, and as heterodimers with other members of the receptor
family (reviewed in (43) and (44) ). RXRc-ErbA
heterodimers exhibit an increased overall affinity for DNA and elevated
transcriptional activation properties compared to those of homodimers.
We demonstrate here that heterodimerization with RXRs can also alter
the half-site specificity of c-ErbA and v-ErbA. For c-ErbA, this is
manifested as a further broadening of half-site specificity, including
a novel recognition of half-site substitutions (at +2 and +3)
that are not detectably bound by c-ErbA homodimers. Perhaps reflecting
a reciprocal phenomenon, the inhibitory effects of receptor P-box amino
acid substitutions on DNA binding can also be counteracted by RXR
heterodimerization. ()It is possible that the inherently
high overall affinity for DNA of the RXRc-ErbA heterodimer simply
overwhelms the destabilizing effects of these response element or amino
acid substitutions. However, it is striking that recognition of certain
half-site substitutions is relatively unaffected by RXR
heterodimerization, whereas recognition of others is greatly enhanced.
This was particularly evident for v-ErbA, which as an RXR heterodimer
exhibits significantly enhanced binding to base substitutions at
+4 to +6, but not to substitutions at +2 and +3.
Notably, the N-terminal domain of ErbA forms part of the heterodimer
interface with RXR(14) ; perhaps RXR influences half-site
specificity in the heterodimer through this interaction with the ErbA N
terminus, altering, in turn, the conformation of the ErbA zinc-finger
domain.v-ErbA, c-ErbA, and Target Genes in the Cancer
Cell
It is intriguing that the DNA recognition properties of the
v-ErbA oncoprotein homodimer are both broader (at -1) and more
narrow (at +4, +5, and +6) than those of the
c-ErbA/TR progenitor. v-ErbA acts in the cancer cell by
interfering with erythroid differentiation, apparently by binding to
and repressing target genes normally activated by c-ErbA, RARs, and
possibly
RXRs(26, 27, 28, 29, 30, 38, 47) .
The extended recognition specificity of v-ErbA at the -1 position
may contribute to its ability to compete for, and repress target genes
for heterologous receptors such as RARs. Conversely, the more narrow
specificity of v-ErbA at +4 to +6 may prevent repression of
c-ErbA-responsive genes critical for cell proliferation(18) .
It is intriguing in this regard to note the different effects of RXR
heterodimerization on the DNA specificity of c-ErbA and v-ErbA. By
enhancing v-ErbA recognition of base substitutions at +4 to
+6, RXR heterodimerization would permit v-ErbA to bind to, and
repress certain sets of c-ErbA target genes that are not repressed by
v-ErbA homodimers. Conversely, by enhancing c-ErbA, but not v-ErbA,
recognition of half-site substitutions at +2 and +3, RXR
heterodimerization would create novel targets for c-ErbA activation
that cannot be repressed by v-ErbA homo- or heterodimers. At the high
levels of v-ErbA expression observed in the cancer cell, it is likely
that both homodimeric and heterodimeric v-ErbA species co-exist, and
that the precise balance between these species has an important
influence in determining the effects of v-ErbA on the neoplastic
phenotype.
)R, thyroid hormone
receptor.))
We are especially grateful to H. W. Chen for providing
reagents, assistance, and helpful discussion, to J. Hamaguchi for
providing Sf9 cell extracts, and to P. Romaniuk for sharing information
and many excellent suggestions prior to publication.
©1996 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. Lee and M. L. Privalsky Heterodimers of Retinoic Acid Receptors and Thyroid Hormone Receptors Display Unique Combinatorial Regulatory Properties Mol. Endocrinol., April 1, 2005; 19(4): 863 - 878. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. M. Yen Physiological and Molecular Basis of Thyroid Hormone Action Physiol Rev, July 1, 2001; 81(3): 1097 - 1142. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Hauksdóttir and M. L. Privalsky DNA Recognition by the Aberrant Retinoic Acid Receptors Implicated in Human Acute Promyelocytic Leukemia Cell Growth Differ., February 1, 2001; 12(2): 85 - 98. [Abstract] [Full Text]
|
|
|
|
 |
|
 S. Mihara, N. Suzuki, S. Wakisaka, S. Suzuki, N. Sekita, S. Yamamoto, N. Saito, T. Hoshino, and T. Sakane Effects of Thyroid Hormones on Apoptotic Cell Death of Human Lymphocytes J. Clin. Endocrinol. Metab., April 1, 1999; 84(4): 1378 - 1385. [Abstract] [Full Text]
|
|
|
|
|
|
P. Bartunek and M. Zenke Retinoid X Receptor and c-erbA/Thyroid Hormone Receptor Regulate Erythroid Cell Growth and Differentiation Mol. Endocrinol., September 1, 1998; 12(9): 1269 - 1279. [Abstract] [Full Text]
|
|
|
|
 |
|
 C. Tzagarakis-Foster and M. L. Privalsky Phosphorylation of Thyroid Hormone Receptors by Protein Kinase A Regulates DNA Recognition by Specific Inhibition of Receptor Monomer Binding J. Biol. Chem., May 1, 1998; 273(18): 10926 - 10932. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. Hadzic, I. Habeos, B. M. Raaka, and H. H. Samuels A Novel Multifunctional Motif in the Amino-terminal A/B Domain of T3Ralpha Modulates DNA Binding and Receptor Dimerization J. Biol. Chem., April 24, 1998; 273(17): 10270 - 10278. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. C. Thurmond and A. G. Goodridge Characterization of Thyroid Hormone Response Elements in the Gene for Chicken Malic Enzyme. FACTORS THAT INFLUENCE TRIIODOTHYRONINE RESPONSIVENESS J. Biol. Chem., January 16, 1998; 273(3): 1613 - 1622. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Liu, A. Takeshita, T. Nagaya, A. Baniahmad, W. W. Chin, and P. M. Yen An Inhibitory Region of the DNA-Binding Domain of Thyroid Hormone Receptor Blocks Hormone-Dependent Transactivation Mol. Endocrinol., January 1, 1998; 12(1): 34 - 44. [Abstract] [Full Text]
|
|
|
|
|
|
F. D. Urnov and A. P. Wolffe An Array of Positioned Nucleosomes Potentiates Thyroid Hormone Receptor Action in Vivo J. Biol. Chem., June 1, 2001; 276(23): 19753 - 19761. [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 |