The EMBO Journal vol.10 no.9 pp.2577-2582, 1991
The Doublesex proteins of Drosophila melanogaster bind
directly to a sex-specific yolk protein gene enhancer
Kenneth C.Burtis, Karen T.Coschiganol, Bruce
S.Baker2 and Pieter C.Wensink'
Department of Genetics, University of California, Davis, CA 95616,
'Department of Biochemistry and The Rosenstiel Center, Brandeis
University, Waltham, MA 02254-91 10 and 2Department of Biological
Sciences, Stanford University, Stanford, CA 94305, USA
Communicated by R.Nothiger
The doublesex (dsx) gene of Drosophila melanogaster
polypeptides, whose synthesis is regulated by alternative
sex-specific splicing of the primary dsx transcript. The
alternative splicing of the dsx mRNA is the last known
step in a cascade of regulatory gene interactions that
involves both transcriptional and post-transcriptional
Genetic studies have shown that the
products ofthe dsx locus are required for correct somatic
sexual differentiation of both sexes, and have suggested
that each dsx product functions by repressing expression
of terminal differentiation genes specific to the opposite
sex. However, these studies have not shown whether the
dsx gene products function directly to regulate the
expression of target genes, or indirectly through another
regulatory gene. We report here that the male- and
female-specific DSX proteins, expressed in E.coli, bind
directly and specifically in vitro to three DNA sequences
located in an enhancer region that regulates female-
specific expression of two target genes, the yolk protein
genes 1 and 2. This result suggests strongly that dsx is
a final regulatory gene in the hierarchy of regulatory
genes controlling somatic sexual differentiation.
differentiation of the fruit fly, Drosophila melanogaster, is
controlled by numerous regulatory genes, often arranged in
hierarchical cascades. The end result of these regulatory
hierarchies is to control the expression of the genes that
actually define morphogenetic cell properties, first termed
genes by Garcia-Bellido (1975) and more
commonly described as downstream or target genes. The
identities of these target genes and the mechanisms by which
their expression is controlled by the final regulatory genes
in the hierarchies are largely unknown, leaving a significant
gap in our understanding ofDrosophila differentiation at the
molecular level. In this manuscript we report the first
evidence of a direct molecular linkage between the regulatory
hierarchy controlling somatic sexual differentiation and a
specific set of target genes.
Oxford University Press
Somatic sexual differentiation
controlled by a well characterized hierarchy of regulatory
genes (reviewed in Baker, 1989; Slee and Bownes, 1990;
Steinmann-Zwicky et al., 1990). The initial setting of this
hierarchy as either male or female involves transcriptional
controls (Salz et al., 1989; Erickson and Cline, 1991) while
the transmission of sexual identity down the cascade is by
the regulation ofRNA processing (Boggs et al., 1987; Bell
et al., 1988; Nagoshi et al., 1988; Burtis and Baker, 1989;
Sosnowski et al., 1989). Genetic experiments suggest that
negative control (reviewed by Baker and Ridge, 1980; Baker
and Belote, 1983; Wolfner, 1988; Steinmann-Zwicky et al.,
1990). However, to date there has been no direct evidence
regarding the molecular mechanism by which the information
transmitted down this hierarchy is passed on to the target
genes that mediate thefinal steps in somatic sexual
* The doublesex (dsx) gene appears to be at the bottom of
the hierarchy and thus may directly regulate target genes.
The primary transcript of the dsx gene is alternatively spliced
in males and females to yield sex-specific mRNAs which
encode male-specific and female-specific polypeptides (Burtis
and Baker, 1989). The two dsx proteins are identical for the
first 397 amino acids (aa), but have unique carboxy termini
of 152 aa (male) and 30 aa (female). It has been proposed
that the male-specific product of the dsx locus, DSXM,
represses the expression of female-specific genes in males
and that the DSXF protein, perhaps in conjunction with the
product of the intersex gene, represses the expression of
male-specific genes in females (Baker and Ridge, 1980;
Nothiger et al.,
1987; Burtis and Baker,
repression could occur through direct binding of DSX
proteins to target genes.
Regulation of the yolk protein genes provides an ideal
system in which to look for a direct molecular interaction
between the DSX proteins and a target gene. The yolk
protein genes are expressed sex-specifically in a tissue found
in both sexes, the adult fat bodies. Molecular genetic
experiments have demonstrated that proper expression of
these genes requires the continuous action of the sex
determination hierarchy (Belote et al., 1985). Moreover,
germline transformation experiments have shown that a 127
bp fat body enhancer (FBE) of the yolk protein genes 1 and
2 (yp] and yp2) is likely to be the target of sex-specific
regulation because this enhancer is sufficient to direct the
female-specific transcription characteristic of the yp genes
in adult fat bodies (Shepherd et al., 1985; Garabedian etal.,
In this manuscript we report that the male- and female-
specific protein products (DSXM and DSXF) of the dsx
locus, expressed in E. coli, bind specifically to the FBE,
demonstrating a direct molecular interaction between the sex
determination hierarchy and a target gene. The observation
that both the male- and female-specific DSX proteinsinteract
K.C.Burtis et al.
with the FBE is discussed with respect to prior genetic and
Gel mobility shift assays using doublesex proteins
overexpressed in E.coli
Coding sequences from male- and female-specific dsx
cDNAs (Burtis and Baker, 1989) were inserted into the
expression vector pT7-7 (Tabor and Richardson, 1985) in
order to overexpress the dsx polypeptides in E.coli under
the control of the T7410 promoter. In the experiments
described in this paper, the expression constructs utilized
a T7-derived initiation codon, resulting in the production of
a fusion protein containing 10 additional amino acids at the
amino-terminus as described in Materials and methods.
Subsequent experiments with constructs expressing DSX
polypeptides without additional amino acids have yielded
identical results (data not shown). Soluble extracts were
prepared from cells carrying either the pT7-7 vector plasmid
alone (control extract), the pT7-7 vector containing a cDNA
encoding the DSXM polypeptide (male extract), or the
pT7-7 vector containing a cDNA encoding the DSXF
polypeptide (female extract). Migration of the female and
male proteins on denaturing SDS -polyacrylamide gels
indicate apparent molecular weights of52 kDa for the female
fusion protein and 67 kDa for the male fusion protein
(Figure la), somewhat larger than the molecular weights of
45.8 kDa (female) and 58.5 kDa (male) calculated from the
amino acid sequence.
Initial evidence for direct interaction between DSX
proteins and DNA sequences containing the FBE was
obtained by gel mobility shift assays. Aliquots ofeach extract
were incubated with a mixture of four end-labeled restriction
fragments; a large fragment derived from vector sequences,
and three smaller fragments derived from sequences located
between -888 and -161 nt relative to the transcriptional
start site of the yolk protein 1 gene. The smallest of these
three fragments (extending from -347 to -161) includes
sequences (-322 to -196) previously identified a sex-,
enhancer of the yp genes
(Garabedian et al., 1986). As seen in Figure lb, incubation
with either the DSXM or DSXF extracts, but not the control
extract, resulted in a substantial reduction in the amount of
unbound DNA present in the smallest restriction fragment
(containing the FBE), but not in the quantity of unbound
DNA in the two adjacent yolk protein gene restriction
fragments. Thus, DSXM and DSXF are sequence-specific
DNA binding proteins in vitro. Mobility shift assays using
a 4-fold lower concentration of protein and the FBE-
Further evidence for the specificity ofthis binding has been
derived from experiments involving the addition of a
competitor oligonucleotide (containing sequences from -309
to -285, a binding site identified by subsequent DNase I
footprinting analysis; see below) to the binding reaction
(Figure ld). The appearance of shifted complexes in the
reactions containing control extract is unaffected by the
presence of a 1000-fold molar excess of this competitor,
indicating that these are due to non-specific interactions
between E.coli proteins and labeled DNA. However, the
Fig. 1. DSX protein interaction with yp DNA. (a) Expression of DSX
proteins in E. coli. Cells containing the pT7-7 vector alone (lane 1), the
pT7-7:DSXF construct (lane 2) or the pT7-7:DSXM construct (lane 3)
were boiled in loading buffer and electrophoresed on a 9% denaturing
polyacrylamnide gel (Laemmnli, 1970). The Mr of protein standards (in
thousands) are indicated at the left. Positions of DSXM (in) and DSXF
(f) are indicated by markers at the right; both migrate moreslowly
than predicted by amino acid sequence. (b) Gel mobility shiftassay
with several yp restriction fragments. The three smallest restriction
fragments contain sequences from -888 to -667 (221 bp), -666 to
-347 (319 bp) and -346 to -161 (185 bp); all relative to theypl
transcription initiation site. The largest fragment (4.4 kb) contains
pBR322 vector linked to yp sequences (-160 to -89). The smallest
band contains the 127 bp FBE (-322 to -196) (Shepherd et al.,
1985; Garabedian et al., 1986).
additional non-specific competitor. Labeled DNA was incubated with
control extract (lane 1), DSXF extract (lane 2) or DSXM extract
(lane 3). (c) Gel mobility shift assay with only theFBE-containing
restriction fragment. A 4-fold reduced extract concentration (0.5 jul per
reaction) was used. Labeled DNA was incubated with control extract
(lane 1), DSXF extract (lane 2) or DSXM extract (lane 3). (d)
Competition by synthetic oligonucleotide. 2.5 nmol of a synthetic
competitor oligonucleotide (1000-fold molar excess over the
radiolabeled fragment; see text) was added to the reactions shown in
lanes 2 and 4. Reactions included either control extract (lanes
2), DSXM extract (lanes 3 and 4), or no extract (lane 5).
1 /Ag of pUC18 DNA was added as an
Binding of doublesex to yolk protein gene enhancer
shift of the FBE-containing fragment by the extract
containing DSXM protein is completely eliminated in the
presence of competitor oligonucleotide (Figure ld, lane 4),
indicating that the interaction between DSXM protein and
the FBE-containing fragment is specific.
Several explanations are possible for the heterogeneous
mobility of bound complexes seen in these experiments.
These include binding of multiple molecules ofDSX to one
fragment, binding of altered forms of the protein (e.g.
proteolytic fragments), and loss of protein from DNA during
Fig. 2. DNase I footprints of DSX protein interaction with FBE DNA. The indicated extracts (C, control; M, DSXM; F, DSXF) diluted 104-fold
(lanes 2, 6, 10), 103-fold (lanes 3, 7, 11) 102-fold (lanes 4, 8, 12), or 10-fold (lanes 5, 9, 13, 16-18) were incubated with radiolabeled FBE and
analyzed by DNase I footprinting. No extract was added to reactions for lanes 1, 14, 15. The AG lane is an (A+G) chemical degradation
sequencing ladder of the FBE fragment (Maxam and Gilbert, 1980). Footprints are indicated to the right and their endpoints and sequences are
shown in Figure 4.
K.C.Burtis et al.
the course of electrophoresis. We show below that at least
three DSX molecules can bind simultaneously to this DNA
involving the addition ofcompetitor oligonucleotide after the
formation of bound complexes (K.C.Burtis, unpublished
data) have indicated a rapid dissociation rate for the DSXM
protein (essentially complete release after 5 min), suggesting
that loss of complex during electrophoresis
is also a
Localization of DSX binding sites by DNase I
Binding sites were localized within the enhancer region by
DNase I footprinting assays. Increasing amounts of protein
radiolabeled enhancer DNA (Figure 2). Control extract gave
no footprints at any of the concentrations tested (lanes 2-5,
16). However, extract from cells overproducing DSXM
protein gave three footprints (lanes 6-9, 17). A single
footprint appeared at low protein concentration (footprint A,
lane 7). A 10-fold increase in protein concentration yielded
a second footprint (footprint B, lane 8). A further 100-fold
increase in protein (data not shown) or a further 10-fold
increase in protein and 25-fold decrease in non-specific
competitor DNA (lane 17) yielded footprint C. Extract from
cells overproducing DSXF protein gave footprinting results
indistinguishable from those with DSXM extracts (Figure 2,
lanes 10- 13, 18).
A further indication that the DNA binding specificities and
affinities of the two DSX proteins are very similar is given
by the ionic strength sensitivity of the footprints (Record et
1981). As shown in Figure 3, DSXM and DSXF
binding decreased in parallel at all sites as the concentration
of KCl was increased. We conclude that both of the DSX
proteins present in E. coli extracts bind to three sites in the
FBE and that each protein binds DNA in a very similar if
not identical manner. This similarity strongly suggests that
the DNA binding domain of the DSX proteins lies in the
amino acid sequence common to the two proteins.
A comparison ofthe three binding sites revealed a potential
9 bp consensus recognition sequence, CTACAAAGT. Four
sequences match this consensus with homologies ranging
from seven out of nine to nine out of nine nucleotides
(Figure 4). The consensus copies occur at similar positions
in the three footprints, four nucleotides from the 5' ends and
five to six nucleotides from the 3' ends of the footprints
(Figure 4), indicating that the consensus sequences are likely
to be involved in the binding reaction. Two copies occur
in the larger footprint B, indicating that it may be composed
of two binding sites.
constant amount of
Direct regulation of yolk protein gene expression by
evidence that a product of the dsx locus is a direct link
between the elaborate cascade of regulation controlling
somatic sexual differentiation in Drosophila and the sex-
specific enhancer of a
demonstrate unequivocally that the proteins encoded by the
dsx gene are sequence-specific DNA binding proteins, and
that a high-affinity DSX binding site lies within the 127 bp
sequence previously found to be a sex-specific, tissue-
specific, and stage-specific enhancer of the yolk protein
genes. Previous experiments have suggested that DSXM
negatively regulates expression of the yolk protein genes in
male fat bodies (Postlethwait et al., 1980; Bownes and
Nothiger, 1981; Ota etal., 1981; Belote et al., 1985; Burtis
and Baker, 1989). We now propose that this repression
Fig. 3. Ionic strength dependence of DSX footprints. The indicated
extract (M, DSXM; F, DSXF) was incubated with radiolabeled FBE
DNA in the presence of increasing concentrations of KCI and analyzed
as in Figure 2. No extract was added to reactions for lanes 1 and 10.
The final KCI concentrations in the binding reactions were 50 mM
(lanes 1, 2, 6, 10), 100 mM (lanes 3, 7), 200 mM (lanes 4, 8) and
400 mM (lanes 5, 9). Reactions examining footprint C (upper panel)
contained 40 ng of the non-specific competitor and 10-fold diluted
extract. Reactions examining footprints B and A (lower panel)
contained 1000 ng of non-specific competitor and 100-fold diluted
Binding of doublesex to yolk protein gene enhancer
occurs because DSXM protein binds directly to the FBE in
males, interfering with enhancer action and thus repressing
the expression of the yolk protein genes. The determination
of whether this mechanism is also used in the repression of
other female-specific genes in males must await identification
of additional target genes.
Repressors have been found with increasing frequency to
play an important role in the regulation of eukaryotic gene
expression, in many cases through interactions with positive
enhancer elements (reviewed by Levine and Manley, 1989;
Stenlund and Botchan, 1990). The mechanism of repression
may involve direct competition for the binding sites of
activator proteins, protein-protein interactions leading to
interference with activator function but not binding, or direct
interference with the basal transcriptional machinery.
mechanisms is involved in the regulation of yp expression
by DSX, the location of the binding site would favor one
of the first two possibilities.
Determination of whether the DSX proteins act directly
to regulate the expression of other sex-specific target genes
is dependent on the identification and isolation ofthese genes.
The most probable candidates for direct regulation include
other genes regulated continuously by the sex determination
hierarchy during adult life, such as the genes regulating male
courtship behavior (Belote and Baker, 1987), and genes
controlling the determination of sex-specific tissues, which
genetic evidence suggests are regulated by the hierarchy at
various times during larval and pupal development (Belote
and Baker, 1982). The regulation of genes expressed in sex-
specific tissues of adult flies may be less direct. It has been
established in some cases that adult expression of these genes
is no longer responsive to the status ofthe sex determination
hierarchy (reviewed by Wolfner, 1988), and that the critical
period during which the hierarchy acts to determine the
expression of these genes coincides with the period during
which the morphological development of the tissue occurs,
days before the genes are expressed (Chapman and Wolfner,
1988). It appears that these genes are directly controlled by
tissue-specific rather than sex-specific mechanisms and are
not direct targets of dsx regulation, although the possibility
cannot be excluded that a transitory direct interaction with
the DSX proteins during the critical period results in a stable
expression state (Wolffe and Brown, 1988).
Site-specific binding of DSXF to the FBE
An unpredicted observation in these experiments is the in
vitro binding of DSXF to the FBE,
experiments have suggested no role for DSXF in regulating
yolk protein gene expression in females. In considering this
unexpected result, it is important to note that DSXF and
DSXM are identical for the first 397 amino acids, but have
it remains to be determined which of these
Fig. 4. Footprint sequences. The sequences protected in each footprint
are shown. Homologies to the proposed consensus are underlined and
their first nucleotides and matches to the consensus are indicated.
unique carboxy termini of 152 aa in DSXM and 30 aa in
DSXF. It seems most likely that the shared 397 amino acid
region mediates the specific binding we observe in vitro.
Thus the different repression effects observed in vivo are
likely to be due to the unique carboxy termini. We propose
two hypotheses to explain the different in vivo activities.
In the first hypothesis, the two proteins bind to the same
set of sites in vivo but have different effects. In this model
the unique domain of DSXM interferes with the function of
an activator protein bound to female-specific enhancers and
the unique domain of DSXF interferes with the function of
an activator bound to male-specific enhancers. A prediction
of this model is that binding ofDSXF in vivo to the binding
sites in the FBE would not lead to repression ofthe yp genes,
which is crucial since both DSXF and the yp genes are
expressed in females. It also predicts that if DSXF were
expressed inappropriately in males,
productively to the FBE binding sites, preventing DSXM
from binding to the FBE and thus preventing the repression
ofyp expression. Previous experiments have indeed shown
that expression of small quantities of DSXF in male flies,
produced from a cDNA copy of the gene introduced into
the genome by P-element mediated transformation, results
in a significant derepression ofypI mRNA expression (Burtis
and Baker, 1989). A precedent for inhibitory domains on
repressor proteins has been found with the yeast protein ax2,
which binds to operator sequences adjacent to a-specific
genes. Binding of ca2 interferes with the expression of a-
specific genes by blocking the activation of these genes
mediated by the MCM-I protein (Keleher et al., 1988). Hall
and Johnson (1987) constructed several mutant versions of
the MAToz2 gene with in-frame deletions. These genes
encoded mutant ca2 proteins with internal deletions, which
still displayed tight binding to operator sequences in vitro
(and probably in vivo), but failed to repress an operator-
regulated gene in vivo, thus indicating that binding and
repression were mediated by different domains of the a2
An alternative hypothesis is that DSXF binds to the FBE
in vitro but not in vivo. In this model, the unique domain
of DSXF interacts with another protein or is modified post-
translationally, altering its binding specificity and directing
it to an alternative set of target genes. One candidate for
an interacting protein is the product of the intersex gene,
which has been shown by genetic studies to be required
specifically in females in addition to the dsx gene product
for proper regulation of somatic sexual differentiation (for
review see Baker and Belote, 1983; Wolfner, 1988). It is
possible that an interaction between dsx and intersex gene
products may lead to a heterodimer with an altered DNA
binding specificity, presumably leading to specific binding
of sequences adjacent to male-specific genes. Again, there
is a precedent in the mating type genes of S.cerevisiae, in
which interaction of the MATal gene product (al protein)
with the ct2 protein leads to a heterodimer that no longer
recognizes operators adjacent to a-specific genes, but rather
binds to operators regulating haploid-specific genes (Goutte
and Johnson, 1988).
it might bind non-
Materials and methods
Expression of dsx proteins
DSXM and DSXF proteins were over-expressed
expression plasmid pT7-7 (Taborand Richardson, 1985).DSXF or DSXM
coding sequences from cDNAs AC16 and AC32 (Burtis and Baker, 1989)
in Ecoli using the
K.C.Burtis et al. Download full-text
were inserted into the vector between the EcoRI and HindHI or ClaI sites,
respectively. The resulting constructs encoded proteins with 10 additional
N-terminal amino acids (MARIRSEAGI) fused to
polypeptides. Expression was induced in 25 ml of cells in mid-log phase
using M13 phage mGP1-2 harboring the T7-7 RNA polymerase gene as
described (Tabor and Richardson, 1985). Cells were then pelleted and
resuspended in 4 ml of buffer Z-100 (100 mM KCI, 25 mM HEPES pH
7.9, 12.5 mM MgCl2, 1 mM DTT, 0.1% NP40, 10% glycerol,
PMSF) with 2 mM benzamidine, 0.514g/ml leupeptin, 0.7 ,ug/ml pepstatin
added. Cell lysates were prepared using a French press (two passes at 18 000
p.s.i.). Lysates were then centrifuged at 27 000 g, 10 min. The supematant
was brought to 40% saturation with ammonium sulfate and precipitate
collected by centrifugation (12 000 g, 10 min). The pellet was resuspended
at I to 10 mg/ml in buffer Z-50 (same as buffer Z-100, except 50 mM KCI),
and dialysed twice against 100 vol of buffer Z-50 over a 12 h period.
full length DSX
Gel mobility shift assays
Interactions between the DSX proteins and yp DNA were assayed by gel
mobility shift assays (Fried and Crothers, 1981; Gamer and Rezvin, 1981).
Standard binding reactions contained (in the order added) 4 1d of 5 x buffer
Z-50, 2jig poly[(dIdC):(dIdC)], 2A1 (-2 pg total protein) of extract
(control, DSXF, or DSXM ), and 2.5 fmol of each of the 32P-end-labeled
restriction fragments, in a total vol of 20 M1. Reactions were incubated at
room temperature for 20 min, then electrophoresed in non-denaturing 4%
acrylamide gels (30:0.8 ratio acrylamide:bisacrylamide, 25 mM Tris,
190 mM glycine,
1 mM EDTA, pH 8.5).
DNase I footprinting
The DNase I footprinting method (Galas and Schmitz, 1978) as modified
by Heberlein et al. (1985) was further modified as follows: 25 11 of extract
(diluted as indicated from 10 mg/ml), 1 ng (7 fmol) of 5' 32P-end-labeled
FBE (-322 to - 196) DNA fragment (label on coding strand), and indicated
amount of non-specific competitor (poly[(dIdC):(dIdC)]) were incubated
in 12.5 mM HEPES (pH 7.6), 0.05 mM EDTA, 6.25 mM MgCl2, 5%
glycerol, 0.5 mM DTT, 50 mM KC1, 2% polyvinyl alcohol in a final volume
of50 11 on ice, 26 min. Following DNase I digestion, reactions were stopped
by adding 90A1of 500 mM NaCl, 0.1% SDS, 20 mM EDTA, 12 mg/ml
Keleher,C.A., Goutte,C. andJohnson,A.D. (1988) Cell, 53, 927-936.
Laemmli,U.K. (1970) Nature, 227, 680-685.
Levine,M. andManley,J.L. (1989) Cell, 59, 405-408.
Maxam,A.M. andGilbert,W. (1980) MethodsEnzymol., 65, 499-560.
Nagoshi,R.N., McKeown,M., Burtis,K.C., Belote,J.M. and Baker,B.S.
(1988) Cell, 53, 229-236.
Nothiger,R., Leuthold,M., Andersen,N., Gerschwiler,P., Gruiter,A.,
Keller,W., Leist,C., Ota,T., Fukunaga,A., Kawabe,M. and Oishi,K.
(1981) Genetics, 99, 429-441.
Postlethwait,J., Bownes,M. andJowett,T. (1980)Dev. Biol., 79, 379-387.
Record,M.T.Jr, Mazur,S.J., Melancon,P., Roe,J.-H., Shaner,S.L. and
Unger,L. (1981) Annu. Rev. Biochem., 50, 997-1024.
Salz,H.K., Maine,E.M., Keyes,L.N., Samuels,M.E., Cline,T.W. and
Schedl,P. (1989) Genes Dev., 3, 708-719.
Shepherd,B., Garabedian,M.J., Hung,M.-C. and Wensink,P.C. (1985) Cold
SpringHarborSymp. Quant. Biol., 50, 521-526.
Slee,R. and Bownes,M. (1990) Quart. Rev. Biol., 65, 175-204.
Sosnowski,B.A., Belote,J.M. andMcKeown,M. (1989) Cell, 58, 449-459.
Steinmann-Zwicky,M., Amrein,H. andNothiger,R. (1990)Adv. Genet.,
Stenlund,A. and Botchan,M. (1990) GenesDev., 4, 123-136.
Tabor,S. and Richardson,C.C. (1985) Proc. Natl. Acad. Sci. USA, 82,
Wolffe,A.P. and Brown,D.D. (1988) Science, 241, 1626-1632.
Wolfner,M.F. (1988) Trends Genet., 4, 333-337.
Received onMay 8, 1991
We thank Kim O'Donnell, Marie Lossky, Scott Erdman, and Susan Hardin
for critical comments on this manuscript, and Michael Lisbin for noting
the consensus homology in footprint C. This research was supported by
grants from the National Science Foundation and the Searle Scholars
Program/The Chicago Community Trust (to K.C.B.) and the National
Institutes of Health (to B.S.B. and P.C.W.).
Baker,B.S. (1989) Nature, 340, 521 -524.
Baker,B.S. and Belote,J.M. (1983) Annu. Rev. Genet., 17, 345-393.
Baker,B.S. and Ridge,K.A. (1980) Genetics, 94, 383-423.
Bell,L.R., Maine,E.M., Schedl,P. and Cline,T.W. (1988) Cell, 55,
Belote,J.M., Handler,A.M., Wolfner,M.F., Livak,K.J. and Baker,B.S.
(1985) Cell, 40, 339-348.
Belote,J.M. and Baker,B.S. (1982) Proc. Natl. Acad. Sci. USA, 79.
Belote,J.M. and Baker,B.S. (1987) Proc. Natl. Acad. Sci. USA, 84,
Boggs,R.T., Gregor,P., Idriss,S., Belote,J.M. and McKeown,M. (1987)
Cell, 50, 739-747.
Bownes,M. and Nothiger,R. (1981) Mol. Gen. Genet., 182, 222-228.
Burtis,K.C. and Baker,B.S. (1989) Cell, 56, 997-1010.
Chapman,K.B. and Wolfner,M.F. (1988) Dev. Biol., 126, 195-202.
Erickson,J.W. and Cline,T.W. (1991) Science, 251, 1071-1074.
Fried,M. and Crothers,D.M. (1981) Nucleic Acids Res., 9, 6505-6525.
Galas,C. and Schmitz,A. (1978) Nucleic Acids Res., 5, 3157-3170.
Garabedian,M.J., Shepherd,B.M. and Wensink,P.C. (1986) Cell, 45,
Garcia-Bellido,A. (1975) In Cell Patterning. Ciba Symposium, Elsevier,
Amsterdam, Vol. 29, pp. 161-178.
Garner,M.M. and Rezvin,A. (1981) Nucleic Acids Res., 9, 3047-3060.
Goutte,C. and Johnson,A.D. (1988) Cell, 52, 875-882.
Hall,M.N. and Johnson,A.D. (1987) Science, 237, 1007-1011.
Heberlein,U., England,B. and Tjian,R. (1985) Cell, 41, 965-977.