| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Pediatric Surgical Research Laboratories (E.U., C.M.H., N.M.,
P.K.D.) Massachusetts General Hospital Harvard Medical
School Boston, Massachusetts 02114
Department of
Biochemistry (A.J.R., B.L., L.M., N.B.P., C.-Y.K., G.C., J.T.R.,
M.A.W.) Case Western Reserve University Cleveland, Ohio
44106-4935
Institut de Genetique Humaine (F.P.) UPR
Centre Nationale de la Recherche Scientifique 1142
Montpellier 34396, France
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
- and ß-subunits in vitro, dissociation of
the functional glycoprotein heterodimer is slow relative to its
physiological lifetime (1). Further, insulin analogs of similar
affinity for the insulin receptor have been observed to exhibit
different mitogenic potency. Although the postreceptor signaling
apparatus is incompletely understood, the overall strength of
insulin’s signal for DNA synthesis correlates with the lifetime
of the hormone-receptor complex rather than its fractional occupancy
(2, 3). In this paper we investigate whether the principle of kinetic
regulation can extend to the function of an endocrine transcription
factor. A model is provided by human SRY, the testis-determining factor
encoded by the Y chromosome (4, 5). Our strategy, uncovered in the
course of investigating mutations in SRY (6, 7), is similar to that
employed in studies of insulin (2, 3). We compare SRY analogs with
similar specific DNA affinities but differing kinetics of DNA
association and dissociation. Prolonging the lifetime of a bent
protein-DNA complex is shown to correlate with enhanced potency in a
cotransfection model of SRY-dependent transcriptional activation (6, 8). Our results suggest that the kinetic stability of component DNA
bends can regulate the ability of an enhanceosome to support multiple
rounds of transcriptional initiation (Fig. 1
A). To our knowledge, this is the first
report of possible kinetic regulation by a specific transcription
factor.
|
B; Refs. 10, 11, 12).
Mutations in the HMG box are associated with 46,XY pure gonadal
dysgenesis and sex reversal (sites highlighted in Fig. 1C;
Refs. 6, 10, 13, 14, 15). The bound DNA structure, in part resembling
A-DNA, is similar to that induced by the TATA-binding protein TBP (Fig. 1F; Refs. 16, 17). DNA bending is in each case associated with
insertion of a nonpolar cantilever side chain between base pairs (Fig. 1E). Insertion disrupts base stacking but not base pairing (18). Human
SRY contains Ile at the cantilever position, residue 13 of the HMG box
consensus sequence. Cantilever side chains otherwise conserved among
SRY and related SOX sequences are Phe or Met (Fig. 1B). Met occurs in
Lef1/Tcf1 HMG boxes (12) where its role is similar to that of I13 in
SRY (10). The inserted side chain is in each case flanked by a Phe,
which abuts the DNA backbone (Fig. 1E). Substitution of either side
chain in SRY by Ala abolishes specific DNA-binding activity (19). A
mutation in SRY leading to a foreshortened and polar cantilever (I13T)
is associated with XY sex reversal and impairs specific DNA binding
(6). The variant cantilever accelerates dissociation of the I13T
protein-DNA complex by at least 50-fold leading to a proportionate
reduction in its association constant. These impaired kinetic and
equilibrium properties may each contribute to the failure of testicular
differentiation. Analogous partial intercalation occurs among diverse
protein-DNA complexes (20, 21, 22, 23, 24, 25, 26). Protein-directed DNA bending is
proposed to facilitate assembly of DNA-multiprotein preinitiation
complexes [enhanceosomes and repressosomes (27, 28)] giving rise to
architectural gene regulation (29, 30). Experimental distinction between kinetic and equilibrium mechanisms requires identification of amino-acid substitutions that affect one property but not the other. Such substitutions are rare because the two properties are linked: the equilibrium association constant (Ka) is equal to the quotient between rates of association and dissociation (kon/koff). Studies of insulin signaling thus used substitutions that confer compensating changes in rates of receptor binding and release (2, 3). To find such a compensating substitution in SRY, we were guided by an analogy between the HMG box and the chemistry of DNA intercalating agents (31). The latter are organic molecules containing aromatic moieties; examples include daunorubicin, doxorubicin, and adriamycin, agents widely employed in cancer chemotherapy (32, 33, 34, 35, 36, 37, 38, 39, 40). Three considerations informed our structural intuition. First, these drugs insert into DNA to disrupt base stacking but not base pairing, a process reminiscent of insertion of a cantilever side chain (31, 39, 40); second, intercalative agents exhibit a wide range of rate- and association constants, modulated in part by the structure and number of aromatic intercalating moieties (32, 33, 34, 35); third, and most important, transcriptional blockade in vitro and cytotoxicity in vivo correlate with dissociation rates of drug-DNA complexes (36, 37, 38). Kinetic regulation of transcriptional elongation by a DNA-binding agent suggests the possibility of analogous kinetic regulation of transcriptional initiation.
Would an aromatic cantilever in a bent SRY-DNA complex, like that in a
chemotherapeutic agent, modulate its kinetic and functional properties?
To test this possibility, we exploit a fortuitous experiment of nature:
the only difference between the SRY HMG boxes of human beings and
chimpanzees is the cantilever substitution I13F (41, 42). Comparison of
these HMG boxes demonstrates that the I13F substitution enhances
SRY-directed expression of a reporter gene; such expression is absent
in control studies of I13T and I13A variants. Despite the I13F
variant’s enhanced activity, the substitution appears not to alter the
structure, DNA-binding, or DNA-bending properties of the SRY HMG box in
accord with a previous study (42). However,
1H-nuclear magnetic resonance (NMR) studies of
dissociation kinetics indicate that the lifetime of an F13 complex is
prolonged relative to that of an I13 complex. Consistent with studies
of transcriptional blockage by DNA intercalative drugs, the extent of
kinetic stabilization is similar to the extent of transcriptional
enhancement. Accordingly, we suggest that an aromatic cantilever more
effectively "locks" the bent SRY-DNA complex in a specific
transcriptional preinitiation complex (Fig. 1A). This model predicts
the occurrence of a novel class of clinical variants: bent but unlocked
HMG box-DNA complexes with native affinity but decreased lifetime.
Kinetic control of the assembly or disassembly of macromolecular
complexes may be a general feature of regulation in diverse biological
settings.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
; lane k in Fig. 2A). Such activation is specific to SRY
and its consensus target sequences [5'-ATTGTT-3' (44, 45, 46, 47)]
rather than to a structure-specific DNA-binding activity as indicated
by the inactivity of human HMG2; further, architectural transcription
factors with different DNA sequence specificities [Lef-1 and IRE-ABP
with consensus sequences 5'-TTTGAA-3' (12, 44)] are only
weakly active in this assay (Fig. 2A). Analysis of deletions and
base substitutions in the MIS promoter’s HMG-box binding
site [5'-TTTGTG; base pairs –48 through –56 relative to the start
site of transcription (47)] has shown that activation by SRY persists,
presumably through multiple 5'-ATTGTT-3' sites in the vector and
reporter.
|
|
); assays were repeated at least three times to assess
statistical significance. Substitutions that reduce specific DNA
binding by 10-fold or more in vitro (M9A, F12A, I13A, and
I13T) exhibit no detectable transcriptional activation. A substitution
that impairs specific DNA binding by 2.5-fold (M9I; Ref. 15) exhibits
low but significant activity at high concentrations of vector DNA
(1.0–2.0 µg) whereas at lower concentrations any differences from
wild type are not significant. Because M9I and I13T are associated with
gonadal dysgenesis (10), human phenotypes (male or female) correlate
with maintenance or attenuation of transcriptional activity and,
in turn, with changes in specific DNA affinity (high or low).
Remarkably, the chimpanzee substitution I13F enhances transcriptional
activation by 40-fold, which is about 4-fold greater than that
stimulated by native SRY (Fig. 2). No differences in protein expression
were detected in the transfected cells by Western blot using a murine
polyclonal antiserum specific for human SRY (Fig. 3C). To investigate the origins of the
I13F variant’s enhanced activity, native and variant HMG boxes were
purified and characterized in vitro. Assays for DNA binding,
bending, structure, stability, and kinetics are described in turn
below.
|
A). In each case a higher-order complex is formed with increasing
protein concentration, previously shown to reflect a DNA-dependent
protein-protein interaction (48). The two domains exhibit similar GMSA
affinities at 4 and 20 C. Sequence specificity was investigated by
systematic GMSA shift of variant 15-bp probes (18) containing the four
possible base pairs 5' to the insertion site (5'-AXTGTT-3';
X = A, T, C, or G) or 3' to the insertion site
(5'-ATXGTT-3'; X = A, T, C or G). These studies
(illustrated in Fig. 3B for the 3'-X site) demonstrate that
the I13F substitution does not alter nucleotide preference at either
DNA position. The I13F substitution likewise conferred no detectable
change in binding to a four-way DNA junction (data not shown). These
results strongly suggest that the enhanced transcriptional
activity of the I13F variant is not due to enhanced affinity for the
consensus SRY target sequence, variant sequences, or nonspecific
prebent DNA.
DNA-Bending Properties Are Similar
Because altered DNA bending may in principle be associated with
enhanced or decreased gene-regulatory activity independently of DNA
binding affinities (15), SRY-induced DNA bending was measured by
permutation gel electrophoresis (PGE) (Ref. 49 ; Fig. 4). The DNA site was
5'-CCCATTGTTCTCT-3'. Human and chimpanzee domains exhibit
almost identical patterns of position-dependent electrophoretic
mobilities. Because interpretation can depend on gel composition,
experiments were performed with both 8% and 10% polyacrylamide.
Estimated DNA bends in the native complex are 80o
and 85o, respectively (inequivalent estimates
presumably reflect limitations of the PGE method). Such sharp bending
is consistent with prior biochemical studies (45). Binding of human and
chimpanzee domains yields similar patterns of flexure-dependent
electrophoretic mobilities.
|
2o) in DNA bend angle; this estimate is
unaffected by gel composition. Because PGE can be associated with
technical artifacts (50), however, SRY-induced DNA bending was also
evaluated by phase-modulation fluorescence resonance energy transfer
(FRET) (51). This technique employs a DNA probe containing a
fluorescent donor (fluorescein) at the 5'-end of one strand and
an acceptor (tetramethylrhodamine; TAMRA) at the 5'-end of the other,
each tethered by a flexible hexanyl linker. Bending is detected as
enhanced FRET efficiency due to a decreased end-to-end distance. The
frequency dependence of fluorescence phase and modulation enables donor
lifetimes to be obtained in the presence and absence of the acceptor
(Table 2A). Such lifetimes permit
assessment of mean distance changes and modeling of distance
distributions (see Discussion).
|
A). Similar
changes occur on binding of TBP to its DNA site (51). Data may be
fitted to a single mean lifetime or with improved fit to two discrete
lifetimes (Table 2A). In the free DNA the mean lifetime of the donor in
the absence of acceptor is 4.44 nsec and is essentially unchanged on
protein binding. The presence of the acceptor leads to a reduction in
the donor’s mean lifetime to 2.56 nsec in the free DNA. Enhanced FRET
on protein binding yields mean lifetimes of 1.79 nsec in the native
complex and 1.81 nsec in the I13F complex. The difference (
0.02
ns) is in the same direction as implied by PGE but is within
experimental error. Corresponding FRET efficiencies are 0.42 (free DNA)
and 0.59 (each complex). Similar FRET efficiencies in the two complexes
imply similar mean induced DNA bends and similar distributions of DNA
bend angles. Although the extent of FRET enhancement is consistent with
a bend angle of 80–85o (51), quantitative
interpretation is limited by possible confounding effects of DNA
unwinding.
|
Comparison of two-dimensional (2D) NMR spectra reveals the new F13
aromatic spin system, which exhibit properties expected of an
unhindered side chain at a protein surface: negligible secondary
shifts, motional narrowing, and absence of nonlocal nuclear Overhauser
effects (NOEs). Specific DNA complexes exhibit a similar correspondence
of NMR spectra (Fig. 6). The
upfield methyl resonances of V5 and L46 provide probes for the
DNA-bound structure of the minor and major wings, respectively (10). As
expected, the unusual aliphatic cantilever resonances of I13 [shifted
upfield by the ring currents of flanking DNA base pairs (18)] are
absent in the variant complex. Together, the spectroscopic results
suggest that the enhanced activity of the I13F protein is not due to a
nonlocal change in the structure of the free or bound HMG box.
|
).
Respective concentrations of guanidine at which 50% of the domain is
unfolded (Cmid; column 2 of Table 3) are essentially identical. A slight
difference in free energies (G
; column 4)
is suggested, but inferred differences are within experimental error
(G 0.25 ± 0.26 kcal/mol). Thermal
stabilities, evaluated by measuring the temperature dependence of both
CD ellipticity (at helix-sensitive wavelength 222 nm) and Trp
fluorescence, are also similar (unfolding midpoints 40 ± 1 C and
42 ± 1 C, respectively; Table 1). Because of the breadth of the
unfolding transition in the range 30–50 C (47), fractional unfolding
at physiological temperature (37 C) would not differ significantly.
These in vitro results are in accord with the similar
biological expression stabilities of native and I13F variant SRY in
cell culture (see above).
|
Imino resonances of a 15-bp DNA duplex containing a central
5'-ATTGTT-3' target site exhibit large changes in chemical shift on
specific binding of the SRY domain. Titration with the I13F domain
(spectra c–e in Fig. 7A) reveals that
free and bound DNA are in slow exchange on the
1H-NMR chemical-shift time scale as previously
described in the native complex (18). Arrows in Fig. 7A–c
indicate discrete DNA imino resonances from a specific complex in
equilibrium with free DNA. Prominent is the anomalous position of the
T8 imino resonance, which is upfield among the guanine imino resonances
(Fig. 7A–e). I13 and F13 complexes exhibit significant differences in
imino 1H-NMR signatures (Fig. 7). These
differences are consistent with the ring current of an aromatic
cantilever: i.e. upfield differences are observed at
flanking thymidine imino and adenine H2 positions (hence above and
below the F13 aromatic ring), whereas smaller downfield differences
occur at the edge of the insertion site flanking (side-chain protons of
F12 and W43). This pattern implies planar (rather than perpendicular)
insertion of F13 between base pairs, consistent with the shape of the
cavity induced in the DNA and analogous to planar intercalative agents
(34, 39, 40). The slow-exchange condition provides a lower bound
on the lifetime of the bound state: slow exchange between resonances
separated by 0.1 ppm (40 Hz at a 1H-NMR resonance
frequency of 400 MHz) implies a lifetime greater than 25 msec.
|
, B and C, for native and
variant complexes) connect the diagonal resonances of a given DNA
proton in the free and bound states. Representative traces
(inset in Fig. 7, B and C) enable comparison of the relative
extent of exchange. Whereas in the wild-type complex the cross-peak
intensities are 
65% of those of diagonal resonances, in the
variant complex their intensities are significantly lower (15–20%;
asterisks in Fig. 7, B and C). Native exchange cross-peaks
are consistent with a lifetime similar to the NOESY mixing time whereas
the lifetime of the chimpanzee complex is prolonged 3- to 4-fold.
Gel-shift studies of the 33P-labeled 15-bp duplex
in NMR buffer indicate similar apparent binding of the two domains at
4, 20, and 37 C. In each case apparent binding is substantially weaker
than that observed using a 36-bp probe (see Fig. 3, A and B). At 37
C and 100 nM protein concentrations, for example,
the percent shift is only 4 ± 1%. It is likely that the
intensity of the shifted bands is attenuated by rapid dissociation of
native and variant complexes on the time scale of electrophoresis.
A dissociation constant is the ratio of the rates of association and
dissociation. Because apparent affinities are similar, the prolonged
lifetime of the I13F domain indicates a similarly prolonged association
rate. Comparison with kinetic studies of a non-sequence-specific HMG
box [HMG1-A recognition of a cisplatin-DNA adduct (52)] implies that
sequence-specific DNA binding is significantly slower (both in
association and dissociation) than structure-specific DNA binding. |
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
A. In this article we have compared SRY analogs to assess a
proposed chemical and functional analogy between the HMG box and
aromatic DNA-binding drugs (31). On intercalative DNA binding, such
drugs disrupt base stacking but seldom base pairing and often induce
features of A-DNA. Our strategy exploits a fortuitous experiment of
nature, the aromatic cantilever of chimpanzee SRY (F13; Refs. 41, 42). Although the equilibrium properties of I13 and F13 domains are
similar, an aromatic cantilever enhances potency in a model
transcriptional assay in association with enhanced kinetic
stability.
Phase-Modulation FRET Assay Implies Well Defined Bend Angle
Electrophoretic anomalies have been widely used as probes of DNA
bending. Interpretation of PGE may be confounded, however, by geometric
features of a protein-DNA complex (50). As a control for such
artifacts, we have employed a complementary FRET assay. FRET efficiency
reflects the distance between 5'-ends of the DNA, which is a function
of both DNA bending and DNA unwinding. A similar assay has been
described in studies of TBP (51). The present studies are in accord
with this and previous PGE studies of HMG-box complexes (15, 42, 45).
Electrophoretic- and FRET methods both indicate that the I13F
substitution has at most a small effect on induced DNA bending; a
decrement of 2o as estimated by PGE is
within the error of the FRET measurements. Assuming no change in
orientation factor 
2, corresponding mean FRET
distances are 58.9 Å ± 1 Å in the free DNA (using a measured
Ro value of 55.9 Å) and 51.1 Å ± 1 Å in the
complexes (using a measured Ro value of 54.3 Å).
The reduction in mean distance is hence about 8 ± 2 Å, which is
consistent with a bend angle of 80–85o. The
single mean-lifetime and single mean-distance approximations neglect
the actual distribution of distances in solution that arise from
motions of the linkers, dyes, and DNA. These approximations are
nonetheless of qualitative value and have been successfully applied in
FRET studies of HMG-1a (52). Although the relationship between
end-to-end distance and bend angle is not straightforward (due to
variable DNA unwinding), the FRET observations corroborate PGE results
to demonstrate that the DNA-bending properties of the two proteins are
similar and so unlikely to be the source of the variant’s enhanced
transcriptional activity.
Neither PGE nor single-lifetime FRET analysis distinguishes between
static and dynamic changes in DNA bending. A mutation in a protein may
change the mean DNA bend angle, for example, without change in
fluctuations. Alternatively, a mutation may induce dynamic instability
or imprecision in DNA bending, leading to a larger distribution of
populated angles. By enabling an estimate of distance distributions,
FRET techniques can, in principle, provide insight into this
distinction (56). Thus, the present frequency-domain data of
donor-acceptor samples are much better fit by two discrete lifetimes
than by a single lifetime (Table 2A).
These lifetimes can be interpreted according to a gaussian distribution
of distances (51, 56). This analysis, which also assumes that isotropic
orientation factor 2 equals 2/3, in each case,
suggests that binding of the I13 and F13 domains induces a 9 Å ± 1 Å
decrease in end-to-end distance (difference between <r> values for
free and bound DNA) with no significant change in widths of the
distributions (Fig. 5B and Table 2B). The widths are due predominantly
to reorientation of the linkers. These results are comparable to two
prior FRET studies of DNA bending induced by nonspecific HMG boxes (52, 57).
In the absence of an adequate cantilever side chain, would a variant
SRY-DNA complex be sharply bent and if so, would significant
fluctuations occur in bend angle? In the future this question may be
addressed through studies of additional analogs. It would be of
interest to find a variant cantilever sufficient to provide near-native
specific binding affinity (or equivalently, free energy G) but
associated with nonlocal dynamic instability. The plausible existence
of such analogs arises from consideration of enthalpy-entropy
compensation (58): reduced enthalpy of side-chain insertion (H) may
be compensated by the enhanced entropy of a fluctuating complex (S).
Such analogs would enable the regulatory importance of precise and
stable DNA bending to be tested. A signature of such a variant complex
would be an anomalously broad distribution of end-to-end distances.
Mutation Enhances Transcriptional Potency and Kinetic Stability
The present study extends previous studies of SRY (13, 14, 15, 42) by correlating phenotypes and biochemical activities with
transcriptional regulation. Although the mechanism of SRY-dependent
transcriptional activation in our assay is unknown and not specific to
MIS, our vector-reporter system nonetheless provides a model
of ATTGTT-dependent transcriptional activation. Our study has
compared functional proteins (human SRY and I13F chimpanzee variant)
and nonfunctional proteins (clinical variants M9I and I13T and control
mutations F12A and I13A). Whereas residues I13 and F13 are each
compatible with the male phenotype in a primate, a polar and
foreshortened cantilever (I13T) or altered hydrophobic wedge (M9I)
blocks testicular differentiation (6, 10, 13, 14, 15). These and other
substitutions with decreased or undetectable specific DNA binding
exhibit decreased or undetectable SRY-dependent transcriptional
activation (Table 2).
Because the chimpanzee SRY HMG box exhibits native DNA affinity,
bending, and nucleotide specificity at the insertion site, we
hypothesize that the variant’s enhanced activity in the cotransfection
assay is due to enhanced kinetic stability: the bent protein-DNA
complex is more stably maintained once formed in a preinitiation
complex. Physical evidence for slower dissociation (relative
to the native complex) is provided by 1H-NMR.
Quantitative implications for the kinetics of protein binding and
release may be estimated as follows. The change in barrier height
corresponding to a factor of 4 in on- or off rates
(G+) would be modest (0.8 kcal/mol). A
lifetime of 200 msec at 40 C (consistent with the native NMR data)
would imply a dissociation rate of 5
sec-1. Assuming an
equilibrium constant of 50 nM under these
conditions, this would imply an association rate of
108
M-1sec-1.
This is 10-fold slower than the rate of association of HMG1 to a
prebent DNA cisplatin adduct at 25 C (koff 30
sec-1 and
kon 109
M-1sec-1;
Ref. 52). We propose that the rate-determining step in SRY binding
occurs in the pathway of HMG box-directed DNA reorganization and that
the protein cantilever participates in this step. The analogous kinetic
barrier in the adducted HMG1 domain complex is presumably reduced by
distortion of the free platinated DNA. Stopped-flow kinetic studies of
the binding of HMG-1a to cisplatin-modified DNA has recently shown that
although the association rate is near the diffusion limit, significant
differences in association rates (but not dissociation rates) are
observed when the DNA sequences are varied at the site of platination
(59). The present study, by contrast, highlights a difference in
dissociation rates. The slower dissociation rate for specific SRY-DNA
complexes is consistent with studies of lymphoid enhancer factor 1
(LEF-1) ( by competition kinetic GMSA (estimated lifetime 120 min or
koff 1.1 x
10-4
sec-1 at 20 C; 60).
Surface plasmon resonance studies of HMG1 and HMG2 (which each contain
two HMG box domains) indicate that bidentate binding to a 30-bp duplex
exhibits kinetic constants of 2 x 104
M-1
sec-1
(kon) and 6–7 x
10-2
sec–1 (koff) (61).
Intriguingly, the prolongation of I13F variant’s lifetime is similar
to its extent of enhanced transcriptional activation. Although this
could be a coincidence, the absence of a plausible alternative
mechanism suggests a model of kinetic regulation in which the net
accumulation of mRNA is controlled not by the equilibrium constant of
the SRY-DNA complex but by its rate of dissociation from an activated
preinitiation complex. This model further posits that, of the many
protein-DNA interactions in a preinitiation complex, the lifetime of
the SRY-DNA complex can be the limiting parameter. We imagine that once
SRY dissociates and the DNA unbends, multiple other protein-DNA and
protein-protein interactions are weakened (Fig. 1A). Accordingly, we
propose that kinetic stability of the bent SRY-DNA complex enables
maintenance of the preinitiation complex through multiple rounds of
transcription. Although the physiological significance of this
observation is not clear (as the present cotransfection assay does not
use a bona fide SRY target promoter and, in any case, both I13 and F13
ultimately results in male differentiation in the appropriate primate),
it is possible that SRY-responsive genes in vivo will be
found to exhibit diverse promoter/enhancer structures with distinct
rate-limiting steps in respective pathways of preinitiation complex
assembly and disassembly.
We caution that the present data are only correlative. Confounding effects in the intracellular milieu of our gonadal cotransfection assay (such as differential nuclear targeting, and differential binding to bystander proteins in the cell) cannot be excluded. Because the substitutions at position 13 are distant from possible phosphorylation sites and nuclear localization signals, however, such effects seem unlikely. It is also possible that the I13 and F13 domains have different affinities for a 5'-ATTGTT-3' site within a nucleosome, a structural target not tested in the present study. Despite these caveats, a kinetic constraint on SRY activity would be broadly consistent with kinetic control in other nonequilibrium systems as discussed below.
Is Kinetic Control a General Biological Phenomenon?
The signature of kinetic control is a correlation between rates
(rather than equilibrium constants) and a functional endpoint. Examples
in endocrinology include the kinetic trapping of gonadotropin
heterodimers (1) and kinetic control of mitogenic signaling by insulin
(2, 3). Kinetic control has also been established among serpin
proteases and has physiological relevance (62). The misfolding of
proteins and subsequent trapping of non-ground-state forms within
aggregates are proposed as pathogenic steps in prion-related
encephalopathies (63). Kinetic control of a processive enzymatic
process may likewise be effected by rates as exemplified by
transcriptional elongation: rapidly dissociating drug-DNA complexes do
not hinder elongation even if compensated by rapid association rates
leading to similar equilibrium constants. Model studies have used T7
RNA polymerase (a small single-subunit polymerase) and the multisubunit
bacterial RNA polymerase of Escherichia coli (36, 37, 38).
Slowly dissociating intercalative agents cause prolonged pausing of the
polymerase preceding the site of intercalation, in turn enhancing the
probability of transcriptional termination with release of the enzyme
from the transcription bubble. The faster and less stable T7 enzyme
terminates at each slowly dissociating ligand complex. Actinomycin D,
echinomycin, mithracycin, and noglamycin provide examples of slowly
dissociating drugs (t1/2 > 5 min) whereas
adriamycin provides an example of a rapidly dissociating drug
(t1/2 < 1 sec). A metaphor of a train (the
processive RNA polymerase) being derailed by a car on the tracks (the
intercalative complex) is supported by the correlation between
diminished bioactivity and creep rate or lateral diffusion of the drug
along the DNA. Enhanced lateral diffusion by an intercalative agent
(visualized as a car outracing the train) correlates with impaired
bioactivity even if the overall rate of drug-DNA dissociation is slow
(35).
Summary
Structure-function relationships in the SRY HMG box have been
inferred from comparative studies of clinical variants (6, 10, 13, 14, 15, 44, 45). By correlating phenotype with biochemical activities, such
studies have suggested that SRY’s genetic function requires specific
DNA binding (44, 45) and DNA bending (15). The present study has
focused on comparison of primate SRY HMG boxes (42), which differ only
by the presence of an aliphatic- (human, gorilla, orangutan, baboon,
and marmoset) or aromatic (common and pygmy chimpanzees) cantilever.
This comparison was motivated by an analogy proposed between
DNA-bending proteins and intercalative DNA-binding drugs (31). The
results highlight the importance of the cantilever mechanism and
suggest a new avenue of investigation, kinetic control of
transcription. In the future it would be of interest to extend this
analysis to M13 (the most common cantilever among mammalian Sry and SOX
sequences; Fig. 1D). Characterization of this and diverse side chains
of chemical interest may further test the proposed relationship between
kinetic stability and transcriptional activation.
Future dissection of the structural, kinetic, and thermodynamic factors underlying SRY-dependent gene regulation will require identification of physiological target genes and biochemical reconstitution of sex- and tissue-specific preinitiation complexes. It is likely that variant HMG boxes with well defined biochemical properties will provide valuable tools to uncover target genes and explore mechanisms of regulation. Our hypothesis of kinetic control predicts the occurrence of novel clinical mutations in HMG boxes that accelerate rates of specific DNA association and dissociation but do not impair DNA affinity. The SRY system may provide an opportunity to investigate a nonequilibrium mechanism of transcriptional regulation.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Site-Directed Mutagenesis
Substitutions were introduced in phage M13 mp19RF by
oligonucleotide-directed mutagenesis (6) and recloned by PCR into an
expression plasmid (18). Constructions were verified by DNA
sequencing.
Four-Way DNA Junctions
Four 30-bp oligonucleotides J1 were prepared as described
(7). One strand was labeled with 32P, annealed
overnight with the other three strands from 80 C to 4 C in 50
mM Tris-HCl (pH 7.5), 5 mM dithiothreitol, 10
mM MgCl2, and 50 mg/ml BSA.
Efficiency of junction formation was more than 80% as estimated by
12% PAGE.
Cell Culture and Western Blot
Cell line CH34 (6) was cultured in DMEM containing 5%
heat-inactivated FBS and 1x penicillin/streptomycin (Life Technologies, Inc., Gaithersburg, MD) at 37 C under 5%
CO2. Cotransfections were performed using the
lipofectAMINE 2000 kit (Life Technologies, Inc.). Briefly,
after cells were washed twice with 1x PBS, SRY expression plasmids
were cotransfected with pXGH5 (6) in 1:1 ratio as an internal control
for transfection efficiency and expression level. After overnight
cotransfection, fresh medium was added. For Western blots, identical
aliquots of cell extracts were electrophoresed on 10% SDS-PAGE. The
gel was blotted on nitrocellulose membrane (Bio-Rad Laboratories, Inc. Hercules, CA). Hybridization solutions containing anti-HMG
domain antibody SE75 at a concentration of 1:100 or antihuman-GH
antibody (obtained from Sigma, St. Louis, MO) at a
concentration of 1:100, respectively, were hybridized onto the
membrane. After goat antirabbit IgG was incubated for 1 h, a
chemiluminescent detection system (Bio-Rad Laboratories, Inc.) was used for specific protein detection. Antibody SE75 is
specific for the HMG box of human SRY and was raised as described (64).
Cross-reactivity to SOX HMG box of a 67 kDa can occur (P. Berta,
personal communication). The present Western blots contained no bands
less than 40 kDa in the control transfection (thus excluding detection
of rat Sry) and only a single band of size appropriate to human SRY in
the SRY- and I68F-SRY-transfected lanes.
DNA-Binding Assay
Oligonucleotides were purchased from Oligos, Etc. (Wilsonville,
OR). 1) The duplex probe was labeled with 32P or
33P, annealed, and analyzed using the
gel-retardation assay (18). Each reaction contained 2.5–160
nM protein (see caption to Fig. 4A) and less than 1
nM labeled DNA in 10 mM potassium phosphate (pH
7.0), 50 ng/ml BSA, 50 mM KCl, 4 mM
dithiothreitol, and 2.5 mM MgCl2; the
reaction was incubated for 1 h on ice. Only specific binding is
observed under these conditions. 2) For analysis of four-way DNA
junctions, reactions contained 25–250 nM protein and 4
nM labeled junction in the same buffer. The 36-bp probe
contains the DNA sequence
5'-CATACTGCGGGGGTGATTGTTCAGGATCATACTGCG-3' and complement
(target site underlined). The parent 15-bp probe (also used
for NMR studies) contains the sequence 5'-GTGATTGTTCAG-3' and
complement. To assess sequence specificity, variant duplexes
5'-GTGAXTGTTCAG-3' and 5'-GTGATXGTTCAG-3' (X
= other base pairs) with respective complements were similarly
prepared.
Electrophoretic DNA-Bending Assay
Double-stranded oligonucleotides containing consensus SRY
binding site 5'-CCCATTGTTCTCT-3' and complement were cloned
between XbaI and SalI sites of vector pBend2
(65). Probes of equal length (147 bp) with the binding site at varying
distance from ends (distances of bend center from 5'-end are 120, 95,
79, 51, 47, and 27 bp) were generated by PCR, and 5'-labeled with
33P-ATP using T4 polynucleotide kinase. Ten
microliter binding reactions contained 50 mM KCl,
20 mM Tris-HCl, pH 7.4, 5
mM MgCl2, 50 ng
poly(dI-dC), 10% glycerol, approximately 1 nM
33P-labeled DNA probe, and 60
nM SRY-HMG box domain. After 1–2 h incubation on
ice, samples were run on polyacrylamide gels in 0.5x (0.045
M) Tris-borate buffer containing EDTA (TBE)
at 10 V/cm.
Steady-State Tryptophan Fluorescence and Stability
Measurements
Steady-state fluorescence spectra and automated guanidine
titrations were obtained using an Aviv spectrofluorimeter (model
ATF105, Aviv Instruments, Lakewood, NJ). Intrinsic Trp fluorescence was
measured at an emission frequency of 390 nm (slit width 5 nm) after
excitation at 270 nm (slit width 2 nm) at 4 C. Denaturation studies of
the human domain as monitored by CD at 222 nm yielded similar results.
Human and chimpanzee SRY domains were made 1 µM in 140
mM KCl, 10 mM potassium phosphate (pH 7.4) in
the titrating cuvette. The same concentration of SRY was used in the
titrant reservoir containing 7.2 M guanidine-HCl in the
above buffer. Experimental curves were fitted by nonlinear least
squares to the equations: G = Gu +
m x [guanidine] = –RT x ln K, where
Gu is the free energy of unfolding
extrapolated to zero denaturant concentration and K is the equilibrium
constant between native and unfolded states of the protein (66).
Phase-Modulation Fluorescence and FRET
Time-resolved measurement were performed on a SPEX Fluorolog
-2 Lifetime Spectrofluorometer (Hoboken, NJ). The light
source was a xenon lamp with excitation wavelength of 490 nm. The
frequency of the excitation beam was modulated in a Pockel cell between
20 and 200 MHz. Sample response was monitored at the emission
wavelength of fluorescein (518 nm) by a photomultiplier (PMT) through a
Corion 520 band-pass filter (ThermoCorion Inc., Franklin, MA) at
magic angle. Time-resolved measurement were performed on 15
µM samples. The reference was LUDOX TM-50 colloidal
silica suspension (DuPont Merck Pharmaceutical Co.,
Wilmington, DE) obtained from Aldrich Chemical Co., Inc.
(Milwaukee, WI). Measurements were performed at 4 C. Fitting analysis,
performed on Globals Unlimited software (Laboratory for Fluorescence
Dynamics, University of Illinois, Urbana-Champaign, IL),
indicated that fluorescein’s spectral response could be best described
with double-exponential curve with two distinct lifetimes contributing
to phase-response (51). Fitting parameter space was changed to
target parameters representing physical invariant of system, after
which a complete system could be specified. For fluorescein-rhodamine
pair energy transfer was assumed to give rise to two lifetimes.
The Förster distances for free and bound DNA were measured to be
56 Å and 54 Å, respectively, and the distance between chromophores
was assumed to vary according to normal distribution. Parameter
optimization was performed to minimize overall
2. Error levels of 0.2 (phase) and 0.01
(modulation) are assumed for analysis. FRET measurements were obtained
in 50 mM KCl and 50 mM Tris-borate (pH
7.4).
CD Spectroscopy
Spectra were obtained using an Aviv spectropolarimeter equipped
with thermister temperature control for automated analysis of thermal
melting. Samples were observed in a 1-mm pathlength quartz cuvette in a
buffer consisting of 10 mM potassium phosphate (pH 7.4) and
50 mM KCl.
NMR Spectroscopy
Spectra were observed at 400 and 500 MHz using Varian
spectrometers (Varian Instruments Inc., Palo Alto, CA) at
Harvard Medical School and The University of Chicago. Spectra of
exchangeable resonances were obtained in H2O
using laminar-shifted shaped pulses in the absence of solvent
presaturation as described (18). Spectra of free domains were obtained
at 25 C in 10 mM deuterated acetic acid (pH 4.5) and 140
mM KCl; such spectra are similar in the pH range 4–8.
Spectra of complexes were obtained at 40 C in 10 mM
potassium phosphate (pH 6.0) and 50 mM KCl.
|
ACKNOWLEDGMENTS |
|---|
We thank E. Haas and V. Itah (Bar Ilan University, Ramat Gan, Israel) for measurements of fluorescent Ro values and advice regarding FRET methods; W. Olson and A. R. Srinivasan (Rutgers University, Newark, NJ) for discussion of FRET design and linker dynamics; P. Berta (Institut de Genetique Humaine, Montpellier, France) for anti-SRY antisera; D. Love and P. E. Wright (Scripps Research Institute, La Jolla, CA) for CURVES analysis of Lef1-DNA and SRY-DNA structures; W. Jia for assistance with NMR measurements and processing; G. L. Waneck for advice regarding site-directed mutagenesis; and L. Labeots and M. Mellody for assistance with DNA binding studies.
|
FOOTNOTES |
|---|
This work was supported in part by grants from the NIH to P.K.D. (Grant CA-17393) and M.A.W. (Grant CA-63485).
1 Present Address: Urawa Municipal Hospital, Department of Pediatric
Surgery, 2460 Mimuro Yrawa-shi, Saitama 336-8522 Japan. 
2 These authors contributed equally.
Received for publication September 13, 2000. Revision received January 2, 2001. Accepted for publication January 4, 2001.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
helices. Science 5186:763–770
This article has been cited by other articles:
 |
|
 |
|
  K. L. Brady, S. N. Ponnampalam, M. J. Bumbulis, and D. R. Setzer Mutations in TFIIIA That Increase Stability of the TFIIIA-5 S rRNA Gene Complex: UNUSUAL EFFECTS ON THE KINETICS OF COMPLEX ASSEMBLY AND DISSOCIATION J. Biol. Chem., July 22, 2005; 280(29): 26743 - 26750. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. A. Komar, S. R. Gross, D. Barth-Baus, R. Strachan, J. O. Hensold, T. Goss Kinzy, and W. C. Merrick Novel Characteristics of the Biological Properties of the Yeast Saccharomyces cerevisiae Eukaryotic Initiation Factor 2A J. Biol. Chem., April 22, 2005; 280(16): 15601 - 15611. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Niimi, Y. Hayashi, S. Futaki, and K. Sekiguchi SOX7 and SOX17 Regulate the Parietal Endoderm-specific Enhancer Activity of Mouse Laminin {alpha}1 Gene J. Biol. Chem., September 3, 2004; 279(36): 38055 - 38061. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Gallego Sequence-dependent nucleotide dynamics revealed by intercalated ring rotation in DNA-bisnaphthalimide complexes Nucleic Acids Res., July 7, 2004; 32(12): 3607 - 3614. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Baud, E. Margeat, S. Lumbroso, F. Paris, C. Sultan, C. Royer, and N. Poujol Equilibrium Binding Assays Reveal the Elevated Stoichiometry and Salt Dependence of the Interaction between Full-length Human Sex-determining Region on the Y Chromosome (SRY) and DNA J. Biol. Chem., May 17, 2002; 277(21): 18404 - 18410. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. Li, W. Zhang, G. Chan, A. Jancso-Radek, S. Liu, and M. A. Weiss Human Sex Reversal Due to Impaired Nuclear Localization of SRY. A CLINICAL CORRELATION J. Biol. Chem., November 30, 2001; 276(49): 46480 - 46484. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. T. MacLaughlin, J. Teixeira, and P. K. Donahoe Perspective: Reproductive Tract Development--New Discoveries and Future Directions Endocrinology, June 1, 2001; 142(6): 2167 - 2172. [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
