Regulation of Body Pigmentation by the
Abdominal-B Hox Protein and Its
Gain and Loss in Drosophila Evolution
Sangyun Jeong,1Antonis Rokas,2and Sean B. Carroll1,*
1Howard Hughes Medical Institute and Laboratory of Molecular Biology, University of Wisconsin, 1525 Linden Drive,
Madison, WI 53706, USA
2Microbial Genome Analysis and Annotation, The Broad Institute of MIT and Harvard, 7 Cambridge Center,
Cambridge, MA 02141, USA
events underlying trait modification have not
been elucidated. Pigmentation of the posterior
male abdomen is a recently acquired trait in
the Drosophila melanogaster lineage. Here, we
directly activates expression of the yellow pig-
mentation gene in posterior segments. ABD-B
regulation of pigmentation evolved through the
ulatory element of the yellow gene of a common
ancestor of sexually dimorphic species. Within
the melanogaster species group, male-specific
pigmentation has subsequently been lost by at
least three different mechanisms, including the
mutational inactivation of a key ABD-B binding
site in one lineage. These results demonstrate
how Hox regulation of traits and target genes
is gained and lost at the species level and have
general implications for the evolution of body
form at higher taxonomic levels.
Many animal bodies are constructed of serially homolo-
gous parts, such as segments, somites, vertebrae, and
appendages. In the course of evolution, the number and
morphology of these structures have undergone dramatic
Understanding the genetic regulatory mechanisms that
govern the differentiation of serially homologous parts is
central to understanding both the development and the
evolution of animal forms.
several types of serially homologous structures (reviewed
in Carroll et al., 2005). A large body of evidence has im-
plicated changes in the regulation of Hox genes (Warren
et al., 1994; Averof and Akam, 1995; Burke et al., 1995;
Averof and Patel, 1997; Stern, 1998; Belting et al., 1998;
Telford and Thomas,1998;CohnandTickle, 1999;Abzha-
nov et al., 1999; Mahfooz et al., 2004) and of the target
genes they control (Warren et al., 1994; Carroll et al.,
1995; Weatherbee et al., 1999; Lewis et al., 2000;
Tomoyasu et al., 2005) in the evolution of animal diversity.
However, despite their prominent roles in development
and evolution, knowledge of the scope and mechanisms
crucial respects (Mahaffey, 2005; Pearson et al., 2005).
First, while Drosophila melanogaster is the best studied
model species, only a modest number of direct Hox-regu-
lated target genes have been identified (Pearson et al.,
2005). Second, nearly all identified target genes encode
signaling proteins or transcription factors that act by regu-
lating the expression of other genes (Pearson et al., 2005).
Third, for some Hox proteins, such as Abdominal-B, no
direct target genes have been characterized. And finally,
despite the inference that the sets of target genes regu-
lated by individual Hox proteins have diverged between
homologous structures (Warren et al., 1994; Carroll,
1995; Weatherbee et al., 1999), the evolutionary gain or
loss of Hox regulation of individual genes has not been
directly demonstrated at the molecular level.
Understanding how Hox-regulated target genes and
traits evolverequires both the identificationof direct target
lution of the Hox-target interaction can be reconstructed.
Toward this end, we have analyzed the development and
evolutionofa Hox-regulated traitinDrosophila.InD.mela-
nogaster, the male has fully pigmented tergites in the fifth
and sixth abdominal segments (A5 and A6), whereas the
female’s tergites have only a narrow pigment stripe (Fig-
ures 1B and 1C). This sexually dimorphic pigmentation
pattern is controlled by a genetic regulatory circuit involv-
ing the Hox gene Abd-B. Loss-of-function mutations of
Cell 125, 1387–1399, June 30, 2006 ª2006 Elsevier Inc. 1387
gain-of-function alleles, such as Abd-BMcp, cause the ex-
pansion of pigmentation to the A4 segment (Celniker
et al., 1990; Hopmann et al., 1995; Figure 1E) or even to
tion factors BAB1 and BAB2 (Couderc et al., 2002), and
doublesex (dsx), which encodes a transcription factor
with a male- (DSXm) and a female-specific (DSXf) isoform
(Burtis, 1993), function to repress male-type pigmentation
in the female abdomen. Loss of these gene functions
causesectopicpigmentationof thefemaleA5 andA6seg-
ments (Baker and Ridge, 1980; Couderc et al., 2002;
Figure 1D). The sexually dimorphic pigment pattern de-
pends upon regulatory interactions among the Abd-B,
bab, and dsx genes (Kopp et al., 2000). Genetic analyses
suggest that ABD-B has dual functions in promoting
male-specific pigmentation: It appears to positively regu-
late the melanin pattern, and it represses the expression
of both bab genes in the A5 and A6 segments. In females,
the repressive action of ABD-B on bab is overcome by the
DSXfprotein, which promotes sufficient bab expression to
suppress posterior pigmentation (Kopp et al., 2000).
In the genus Drosophila, many species show differ-
ences in pigmentation traits. Pigmentation of the posterior
male abdomen is a trait found in many members of the
melanogaster species group but not in several other major
groups. The dimorphic regulation of bab expression is
closely correlated with dimorphic pigmentation (Kopp
et al., 2000) as well other pigmentation patterns (Gompel
and Carroll, 2003). It is not known, however, which regula-
tory interactions among Abd-B, bab, dsx, and pigmenta-
tion genes are direct and which are indirect.
Here, we show through biochemical and transgenic
analyses that ABD-B directly activates the expression of
the yellow gene in the male A5 and A6 segments through
binding sites in a specific cis-regulatory element (CRE) of
the yellow gene. Furthermore, we demonstrate that ABD-
B regulation of the yellow gene evolved in a common an-
cestor of the melanogaster species group and that muta-
tional inactivation of a key ABD-B binding site occurred
within the D. kikkawai lineage that has lost dimorphic pig-
mentation. These results demonstrate that Hox proteins
do directly regulate terminal phenotypes and that the
gain and loss of Hox regulation occur at the level of indi-
vidual species through modifications of CREs.
The cis-Regulatory Region Controlling yellow
Expression in the Abdomen
The Yellow protein is expressed in the pupal epidermis in
a striped pattern near the A/P compartment boundary in
all abdominal segments of developing females, in seg-
ments A1–A4 of developing males, and in a broad intense
pattern throughout the anterior of segments A5 and A6 of
males. This pattern foreshadows the melanin pattern of
the adult flies (Wittkopp et al., 2002a). In order to dissect
the regulation of yellow expression in the abdomen, we
first identified cis-regulatory sequences necessary for ac-
curate gene expression.
Genetic analysis has identified a region of the yellow lo-
cus required for gene function in the abdomen (Geyer and
Corces, 1987), and molecular studies have identified dif-
ferent regions of the yellow gene that govern expression
in several structures, including the body, wings, and bris-
viously described 1.4 kb CRE (Wittkopp et al., 2002b),
domen: The contrast between expression in the more an-
terior portion of each segment and the stripe near the A/P
compartment boundary was diminished, and the higher
levels of expression in male segments A5 and A6 was
less pronounced (data not shown). To determine whether
additional sequences flanking this CRE could contribute
to the fidelity of reporter-gene expression, we tested the
activity of a 2.6 kb fragment located 269 bp from the 50
end of the yellow transcription unit (Figure 1A). This region
includes some regulatory sequences that contribute to
gene expression in the wing. This element, termed the
wing/body (wb) element, drove robust, sexually dimorphic
expression of enhanced green fluorescent protein (EGFP)
in the pupal abdomen and a sharp, striped pattern in the
(Figures 1F and 1G), thus recapitulating the native Yellow
If the sexually dimorphic pattern of the wb element is
regulated by Abd-B and bab, its expression should be
modified differently in mutants for these genes. In females
with one mutant copy of the bab locus, the activity of the
wbelement wasfully derepressed inA6and partially dere-
pressed in A5, which correlated with the male-type adult
pigmentation pattern of the same genotype (Figures 1D
and 1H). In addition, the wb element responded to the
Abd-BMcpgain-of-function allele—male-specific expres-
sion of EGFP was expanded to the A4 segment in mutant
element and, hence, the yellow gene are transcriptionally
regulated, directly or indirectly, by both Abd-B and bab
Identification of an ABD-B-Responsive Element
In order to determine whether regulation of yellow might
be direct, we sought to identify regions within the wb ele-
ment that were ABD-B-responsive. We first determined
whether smaller regions of the wb element were sufficient
to drive sexually dimorphic expression of a reporter gene.
A 1.6 kb portion of the wb element lacking the 1.0 kb 50-
most sequence that contains the wing CRE (Geyer and
Corces, 1987; Wittkopp et al., 2002b; Gompel et al.,
2005) drove elevated EGFP expression throughout the
A1–A4 segments as well as robust expression in A5 and
A6 in males (Figures 2A and 2B). We shall refer to this
1.6 kb element as the body CRE.
1388 Cell 125, 1387–1399, June 30, 2006 ª2006 Elsevier Inc.
We then made a series of deletion constructs for six
subregions of the body CRE (BED1–BED6) to determine
which subregion (or subregions) was necessary or suffi-
cient to drive sexually dimorphic expression (Figure 2A).
Only BED3 males lost the male-specific expression of
EGFP in the A5 and A6 segments, demonstrating an es-
sential role of this region (Figures 2A and 2C). To deter-
mine whether the BE3 cis-regulatory region is also suffi-
cient for the response to ABD-B, we made a set of
constructs that contained BE3 and/or adjacent regions
and examined reporter expression in the pupal abdomen
(Figure 2A). All five BE3-containing constructs, including
BE1–3, BE2–3, BE3, BE3–4, and BE3–6, express EGFP
in a male-specific pattern at varying levels and respond
to the Abd-BMcpallele (right column in Figure 2A; Figures
2D–2I). The BE3 region alone drove weak reporter expres-
strong expression (Figure 2H) that was responsive to Abd-
is required for the response to ABD-B and that the BE3–6
region contains the cis-regulatory sequences required for
robust male-specific activation of the yellow gene.
Figure 1. Expression of the yellow Pigmentation Gene Is Transcriptionally Modulated by Abd-B and bab
(A) Organization of the D. melanogaster yellow locus. Arrow indicates the position of transcription initiation. The boundaries of the wing/body CRE are
indicated. The exons are represented by solid black rectangles. Below this map, the wing/body (wb) construct is depicted as a solid bar.
(B–E) Abdominal cuticles are displayed with the dorsal tergites to the right. Segments A4, A5, and A6 are indicated.
(B) Abdominal tergites of D. melanogaster females display a posterior black stripe widened at the dorsal midline of each segment.
(C) Abdominal tergites of D. melanogaster males exhibit male-specific pigmentation of the entire posterior two segments as well as pigmentation
stripes in the A1 to A4 segments as in the female.
(D) D. melanogaster babAr07/+ females display male-type pigmentation in A5 and A6 segments.
(E) Ectopic activity of Abd-B in D. melanogaster Abd-BMcp/+ males leads to the expansion of posterior male pigmentation into the A4 segment.
(F–I) Confocal images of the dorsal abdomen of transgenic pupae expressing the nuclear EGFP reporter protein (green).
(F) Six abdominal segments of a wb/+ female express EGFP in a striped pattern corresponding to the adult female pigmentation pattern (B).
(G)Inwb/+males,thenon-sex-specific stripedexpressionaswellasrobustmale-specificexpressionofEGFPcorrespondprecisely totheadultmale
pigmentation pattern (C).
(H) The posterior two segments of a wb/babAr07female robustly express EGFP in male-specific pattern, reflecting ectopic pigmentation of the same
(I) In wb/Abd-BMcpmales, ectopic expression of EGFP in the A4 segment is consistent with the pigmentation pattern of the same genotype (E).
Cell 125, 1387–1399, June 30, 2006 ª2006 Elsevier Inc. 1389
Figure 2. Identification of a Male-Specific Regulatory Element in the yellow body CRE
(A) Summary of reporter constructs and their activities. The 1.6 kb D. melanogaster body CRE is divided into six subregions, BE1 to BE6, from which
aseriesofdeletion constructs werederived.Intherightcolumn, thesymbol +or?represents thepresenceorabsenceofmale-specificexpression of
EGFP in the A5 and A6 segments. Superscript ‘‘a’’ indicates weak expression of the reporter.
(B) The wild-type body CRE drives elevated reporter expression throughout the A1–A4 segments as well as in segments A5 and A6.
(C) In BED3 males, male-specific pigmentation is lost; the segmental striped expression of EGFP is retained.
(D) In BE1–3 males, reporter expression is very similar to that of the intact body CRE reporter (B).
(E) In BE2–3 males, there is weak expression of EGFP in A5 and A6.
(F) In BE3 males, reporter expression is male specific but very weak.
(G) In BE3–4 males, there is a low level of EGFP expression in A5 and A6.
(H) In BE3–6/+ males, there is robust expression of EGFP in a male-specific pattern.
(I) In BE3–6/Abd-BMcpmales, male-specific expression of EGFP fully extends to the A4 segment.
1390 Cell 125, 1387–1399, June 30, 2006 ª2006 Elsevier Inc.
Direct Regulation of the yellow Gene by ABD-B
We next examined whether the ABD-B protein directly ac-
tivates the yellow gene through interaction with binding
sites in the body CRE. We systematically searched for
ABD-B binding sites in vitro by DNase I footprinting of all
domain (HD). We identified four sites that were strongly
ing sites were clustered together within the BE3 region
(Figure 3A; BS4, BS5, BS6, and BS7). BS5 contains
tifs (which is the preferred core binding site for ABD-B;
Ekker et al., 1994), but the other two sites do not contain
any potential core sequence for Hox proteins (Ekker
et al., 1994) (Figure 3B). In order to confirm ABD-B binding
were performed using four overlapping oligonucleotides
containing pairs of putative sites. The short sequence be-
binding site sequence in the context of either the BS45 or
the BS67 oligonucleotide (Figure 3B). Analysis of ABD-B
binding to the wild-type and mutated oligonucleotides
indicated that only mutations in BS5 and BS7 reduced
ABD-B binding (Figure 3C). The footprinting of BS4 and
BS6 appears then to be a consequence of ABD-B binding
to the adjacent bona fide BS5 and BS7.
To test whether BS5 and BS7 are required for the func-
tion of the CRE in vivo, we introduced mutations into the
core sequences of both binding sites within the BE3–6
element (KO[ABD-B]; Figure 3B). Disruption of these sites
expression in segments A5 and A6 (compare Figures 3D
and 3E), demonstrating that the yellow gene is directly
regulated by the ABD-B Hox protein.
Evolution of Posterior Pigmentation
in the melanogaster Species Group
expression and make an informative choice of species for
further study, we first considered the evolutionary history
of male-specific pigmentation in Drosophila. The mela-
nogaster species group includes three major clades: the
Oriental lineage (to which D. melanogaster belongs), the
montium subgroup, and the ananassae subgroup (see
Figure S1 in the Supplemental Data available with this arti-
cle online). Ancestral character reconstruction indicates
state in the Oriental lineage (posterior probability of di-
morphically pigmented ancestor 97% ± 2%; Figure S2)
and the melanogaster species group (probability of dimor-
phic ancestor 84% ± 8%; Figure S2). Furthermore, the
male-specific repression of bab expression is widespread
throughout all three clades (Kopp et al., 2000), suggesting
that dimorphic regulation of bab was present in the com-
mon ancestor of the entire melanogaster species group.
members of the obscura, willistoni, and saltans groups, do
male-specific pigmentation appears to have arisen once
in the melanogaster species group, and the absence of
male-specific pigmentation in species such as D. bipecti-
sequent losses of the trait.
Thistraithistoryshouldbereflected inthemolecular ge-
netic architecture of trait formation. The identification of
the ABD-B-regulated body CRE of the yellow gene offers
the opportunity to trace when ABD-B regulation was
Therefore, we selected several potentially informative
species for further study based upon their pigmentation
patterns and phylogenetic relationships (Figure 4). These
included D. biarmipes, another member of the Oriental
lineage that is dimorphically pigmented; D. santomea,
losses of the trait in the Oriental, montium, and ananassae
subgroups, respectively; and D. subobscura, a member of
the obscura group (the sister clade of the melanogaster
species group) that exhibits intense monomorphic pig-
mentation of all abdominal segments.
Evolution of the yellow body CRE and Dimorphic
In order to determine how the function of the body CRE
may have changed during Drosophila evolution, we iso-
lated orthologous body CREs from D. santomea, D. kikka-
wai,D.bipectinata, andD.subobscura;the entire50region
of the D. biarmipes yellow gene, which includes the body
CRE, was isolated previously (Gompel et al., 2005). We
then tested CRE activity when linked to the EGFP reporter
and transformed into D. melanogaster. Reporter expres-
sion driven by the D. biarmipes 50region (50ybia, Figures
5A and 5F) was sexually dimorphic, as we expected if the
mon origin and respond to the same transcription factors.
Surprisingly, the D.santomea (san body, Figures5Band
also drove sexually dimorphic expression of EGFP, even
though in these species abdominal pigmentation is either
monomorphic (D. bipectinata) or absent (D. santomea).
to transcriptional regulatory inputs present in D. mela-
nogaster. Indeed, the expression of the san body and bip
body reporter constructs responded to the Abd-BMcpmu-
tation (Figures 5L and 5M). Based upon the phylogenetic
relationships among these species (Figure 4), the most
parsimonious explanation for the shared responsiveness
of the D. santomea, D. biarmipes, and D. bipectinata
CREs to Abd-B is that the common ancestor of all of
these species possessed an ABD-B-responsive CRE and
therefore, as we inferred above, was likely to have been
However,while D.kikkawai isdescended fromthe same
ancestor, the kik body CRE drives only monomorphic re-
porter expression (Figures 5D and 5I) and does not re-
spond to ectopic Abd-B activity (Figure 5N). Therefore,
Cell 125, 1387–1399, June 30, 2006 ª2006 Elsevier Inc. 1391
Figure 3. Direct Activation of the yellow body CRE by ABD-B
(A) DNase I footprinting was performed with labeled BE3 subregion-containing fragments and recombinant ABD-B HDand GST proteins. Amounts of
each protein used were as follows: lane 2, no protein; lane 3, 1 mg GST; lane 4, 150 ng ABD-B; lane 5, 300 ng ABD-B; lane 6, 600 ng ABD-B. A G+A
sequencing ladder is shown in lanes 1 and 7. Black boxes indicate the regions protected by ADB-B HD proteins.
(B) Oligonucleotides andtheirmutated sequences used inEMSA studies.EMSA experimentswithfour overlapping oligonucleotides shown at thetop
(BS45, BS55, BS56, and BS67) revealed that the region between BS5 and BS6 is not required for ABD-B binding. Thus, mutated sequences for each
putative binding site were introduced into either BS45 or BS67 oligonucleotides shown in the bottom half of the panel.
(C) EMSAs were performed with increasing amounts of ABD-B HD protein (17 ng, 50 ng, 150 ng, and 150 ng) and the wild-type or mutated oligonu-
shifts were abolished completely. Band shifts were not detected in control lanes using 2 mg of GST protein (GST, +). Mutations of BS5 and BS7,
termed BS45[m5] and BS67[m7], respectively, strongly reduce ABD-B binding to the probes. The two sets of mutations that knock out the BS5
and BS7 sites (named KO[ABD-B]) were introduced into wb and BE3–6 elements ([B], bottom).
(D and E) Disruption of two ABD-B binding sites within BE3–6 causes a complete loss of male-specific expression of EGFP (compare [D]
1392 Cell 125, 1387–1399, June 30, 2006 ª2006 Elsevier Inc.
the absence of male-specific pigmentation in D. kikkawai
could be due to the loss of ABD-B regulation, whereas
the losses of pigmentation in D. bipectinata and D. santo-
mea must have a different mechanistic basis. In order to
trace the fate of ABD-B regulation of the body CRE in de-
tail, we examined the CREs from these species for ABD-B
binding sites, focusing in particular on a region that is
shared in all species and contains the functional ABD-B
sites identified in D. melanogaster (Figure 6A).
Evolution of ABD-B Binding Sites in the body CRE
When compared to the BS5 and BS7 sequences of
D. melanogaster, base substitutions have occurred within
these sites in some species (Figure 6A). We tested the ef-
fect of sequence divergence on the DNA binding affinity of
ABD-B HD protein through EMSA experiments with oligo-
nucleotides containing each binding site. We found that
binding by ABD-B were correlated with the divergent
function of the CREs when tested in D. melanogaster.
kawai, D. santomea, and D. melanogaster and differs only
at a single base in D. biarmipes (Figure 6A). ABD-B binds
to all four of these sequences equally well (Figure 6B,
left). However, in D. kikkawai, BS7 contains two base sub-
stitutions (Figure 6A) that diminish ABD-B binding
(Figure 6B, right). In order to test whether these base
changes within the ABD-B binding sites of the D. kikkawai
body CRE could account for the loss of male-specific re-
porter expression in A5 and A6, the TT/CC point muta-
tions were introduced into BS7 of the D. melanogaster
BE3–6 element. We found that the mutated CRE com-
pletely lacked activity in the abdomen (Figure 6C). Thus,
mutation of these sites alone is sufficient to inactivate an
otherwise functional element. In order to test whether
these substitutions account fully for the unresponsiveness
of the D. kikkawai CRE to Abd-B, we also made the recip-
rocalsubstitution oftheD.melanogasterTTsequence into
the CC sites of the D. kikkawai CRE. These substitutions
indicates that other functional sites (e.g., for cofactors)
have also diverged in the D. kikkawai CRE, which should
laxed. Together, these reciprocal substitutions indicate
that mutations in ABD-B BS7 and other sites in the
D. kikkawai CRE contributed to the evolutionary loss of
ABD-B regulation of this element.
The evolutionary loss of male-specific pigmentation has
occurred by different mechanisms in the D. bipectinata
and D. santomea lineages. The bipectinata body CRE
drove male-specific reporter expression in A5 and A6,
and BS7 was bound well by ABD-B, despite a single
that another regulator of pigmentation, the product of the
bric ? a brac2 (bab2) locus—a repressor of pigmentation—
is not expressed dimorphically in D. bipectinata as it is in
Figure 4. Frequent Losses of Male-Spe-
cific Pigmentation in the melanogaster
Inmost species within the melanogastergroup,
male-specific pigmentation is present (as in D.
melanogaster and D. biarmipes), whereas most
species outside of this group lack sexually
dimorphic pigmentation in the A5 and A6 seg-
ments. Losses of male-specific pigmentation
are observed in several lineages of the mela-
nogaster species group, including D. santomea
and D. bipectinata, which lack pigmentation,
and D. kikkawai, which is monomorphically
pigmented in a striped pattern. Phylogenetic
tree of the seven species studied here is based
on data presented in Figure S1. D. subobscura
and D. pseudoobscura are in the obscura spe-
cies group, which is sexually monomorphic
and displays intense abdominal pigmentation.
Cell 125, 1387–1399, June 30, 2006 ª2006 Elsevier Inc. 1393
(Kopp et al., 2000). Thus, we deduce that the change in
bab2 regulation in the D. bipectinata lineage can account
for the loss of male-specific pigmentation in that lineage.
In D. santomea, however, BAB2 expression is dimor-
phic (Gompel and Carroll, 2003). Furthermore, note that
BS5 and BS7 are perfectly conserved in the D. santomea
body CRE (Figures 6A and 6B), which drove dimorphic re-
porter expression in D. melanogaster (Figure 5G). Thus,
the loss of pigmentation in this species is not due to evo-
lutionary changes in bab or yellow but must be due to the
evolution of other loci, which is consistent with genetic
analysis (Llopart et al., 2002; Carbone et al., 2005).
The inference that the common ancestor of D. kikkawai,
D. santomea, and D. bipectinata had an ABD-B-respon-
sive CRE places the origin of ABD-B regulation at or be-
fore the common ancestor of the melanogaster species
group. To further delimit the origin of ABD-B regulation
of the yellow gene, we examined the activity of the
D. subobscura body CRE in D. melanogaster. This CRE
drove broad expression of the reporter in all abdominal
segments, with, unexpectedly, slightly elevated expres-
sion in the male A5 and A6 segments (Figures 5E and
5J). The elevated expression raised the possibility that
the D. subobscura CRE might also be responsive to
Abd-B. This was indeed the case, as reporter expression
was modified in the Abd-BMcpmutant (Figure 5O). Align-
ment of the body CREs from the five melanogaster group
species and D. subobscura, as well as D. pseudoobscura,
revealed that, while there is extensive sequence diver-
gence between the two species groups’ body CREs, sev-
eral shared blocks of colinear sequence exist (Figure S3).
These findings indicate that the common ancestor of the
obscura and melanogaster groups possessed a body
CRE that may have also been ABD-B responsive. How-
ever, because no significant sequence similarity was
found between these body CREs and the body CREs of
D. willistoni, D. virilis, or D. grimshawi, and because these
Figure 5. Evolution of yellow body CRE Function
All flies examined are heterozygous for transgenic reporter constructs. In (K)–(O), males are heterozygous for both the transgene and the Abd-BMcp
(A, F, andK) The entire 50upstream region of D. biarmipes (50ybia)drives sexually dimorphic expression of EGFP (A andF) andis responsive toectopic
Abd-B activity (K).
(B, G, and L) The D. santomea body CRE (san body) drives sexually dimorphic expression of EGFP (B and G) and responds to Abd-B (L), even though
the flies are unpigmented.
(C, H, and M) The D. bipectinata body CRE (bip body) drives sexually dimorphic expression of EGFP (C and H) and responds to Abd-B (M).
(D, I, and N) The striped pattern of EGFP expression in both sexes of D. kikkawai body (kik body) transgene corresponds to the adult pigmentation
pattern (D and I). The kik body element does not respond to Abd-B (N).
(E, J, and O) The D. subobscura body CRE (sub body) drives EGFP expression in the female abdomen (E) and the male abdomen (J) with slightly
elevated expression in A5 and A6 that responds to Abd-B (O).
1394 Cell 125, 1387–1399, June 30, 2006 ª2006 Elsevier Inc.
elements showed no sign of responsiveness to Abd-B in
D. melanogaster (data not shown), ABD-B regulation of
yellow appears not to predate the common ancestor of
the obscura and melanogaster groups.
cura CRE might be direct, we inspected the sequences of
BS5 and BS7 in the obscura group. Both of the obscura
group species’ CREs contained numerous substitutions
in BS5 (Figure 6A) that abolished binding by ABD-B
(Figure 6B). Both species also contained the same three
base substitutions in BS7 (Figure 6A), which was bound
weakly by ABD-B (Figure 6B). To test whether the subobs-
cura BS7 might be functional in vivo, we inserted these
three base substitutions into the BS7 of the D. mela-
nogaster CRE and tested the activity of the mutated ele-
porter expression in A5 and A6, indicating that the BS7 is
functional and could account for the ABD-B responsive-
ness of the D. subobscura CRE (Figure 6E). This observa-
tion is consistent with an origin of some degree of ABD-B
regulation of the body CRE in a common ancestor of the
obscura and melanogaster species groups.
B in vitro but was inactive in the context of the D. mela-
nogaster CRE. The D. subobscura BS7 sequence was
also bound weakly in vitro but was active in vivo. The
most likely explanation for this difference in activity con-
cerns the orientation of ABD-B binding sites in each BS7
quences is eliminated and one remains (Figure 6A). In the
D. subobscura sequence, both TTAT core sequences are
absent, but the T/A substitution creates a new TTAT
core sequence on the opposite strand (Figure 6A). Thus,
while the affinity of ABD-B binding appears equivalent
(Figure 6B), the protein is contacting different bases in a
different orientation, which clearly has different conse-
quences for CRE function in vivo (Figures 6C and 6E).
We have shown that the Abdominal-B Hox protein directly
activates the expression of the yellow pigmentation gene
through specific binding sites in the body CRE. Further-
tationappears tohave originated inacommonancestorof
the melanogaster and obscura species group and was
subsequently lost in the D. kikkawai lineage, at least in
part through mutational inactivation of a key ABD-B
Figure 6. Evolutionary Changes in ABD-
B Binding Sites of the yellow body CRE
(A) Sequence alignment of potential ABD-B
binding sites of seven species. BS5 and BS7
are highlighted in black, and base substitutions
are shown in red. Nucleotides that are com-
pletely conserved across six species are high-
lighted in green. The left panel depicts the phy-
logenetic relationships of D. melanogaster (D.
mel), D. santomea (D. san), D. biarmipes (D. bia),
D. kikkawai (D. kik), D. bipectinata (D. bip), D.
subobscura (D. sub), and D. pseudoobscura
tides containing the BS5 or BS7 sequences of
six species. EMSAs were performed with in-
creasing amounts of the ABD-B HD protein
(17 ng, 50 ng, and 150 ng). Note the reduced
binding of ABD-B to BS5 of D. subobscura
and BS7 of D. kikkawai.
duced are indicated below each panel.
(C) In BE3–6[CC] males, male-specific expres-
sion of EGFP is completely lost.
(D) In D. kikkawai[TT] males, no expression of
EGFP is restored.
sion of EGFP is strongly reduced.
Cell 125, 1387–1399, June 30, 2006 ª2006 Elsevier Inc. 1395
ing of the architecture and evolution of Hox-regulated ge-
netic circuits and how Hox proteins regulate the formation
and diversification of body patterns.
Hox Regulation of Traits and Target Genes
The functions of homeotic genes in the development of
body parts have long been viewed as operating at the
top tier of a hierarchical process. In the conceptual frame-
work first formulated by Garcia-Bellido (1975), individual
Hox genes act as ‘‘selector’’ genes that direct the differ-
entiation of initially similar developmental fields along
one of several alternative paths. They were envisioned
to operate by regulating the activity of teams of ‘‘realiza-
tor’’ genes that determine the size and shape of individual
structures. The reaper gene, which encodes a cell-death-
promoting protein, is one such realizator that affects seg-
ment shape (Lohmann et al., 2002). It has not been shown
previously whether realizator genes also include structural
genes involved in terminal cell differentiation. Most direct
Hox-regulated target genes identified thus far encode
transcription factors and signaling proteins that act by
controlling the expression of other genes (Pearson et al.,
2005). Our results show that Hox regulation of morpholog-
ical traits is not strictly hierarchical in that direct Hox reg-
ulation of genetic circuits extends to the level of terminal
We did not anticipate that ABD-B regulation of pigmen-
tation patterns would be mediated by the direct control of
pigmentation genes. Throughout the melanogaster spe-
cies group, sexually dimorphic pigmentation patterns cor-
relate with dimorphic regulation of the bab locus (Kopp
we initially inferred that bab and perhaps other regulatory
loci were interposed between Abd-B and pigmentation
genes. The available evidence suggests that BAB regu-
lates yellow indirectly, through the regulation of other
Evolution of ABD-B Regulation and Male-Specific
At least three regulatory interactions have evolved for the
segments: the direct control of pigmentation genes by
ABD-B, the sexually dimorphic regulation of bab, and
Our analysis of the evolution of the body CRE sequence
and function and previous studies of bab expression
(Kopp et al., 2000; Gompel and Carroll, 2003) enable us
to map the evolutionary gain and loss of these interactions
onto the history of the species considered here (Figure 7).
To do so, we must first infer the state of regulatory interac-
tions in common ancestors of these species.
Based upon the distribution of ABD-B binding sites,
tion, weinfer thatallthreeregulatory interactions existed in
the common ancestor of the melanogaster group, and
perhaps in a common ancestor of the melanogaster and
obscura groups. The reason for the uncertainty between
thesetwonodes(green arrowsinFigure7) isthesurprising
evidence for ABD-B regulation of the body CRE in D. sub-
obscura (and D. pseudoobscura). From mature adult phe-
notypes alone, one would not infer any sexually dimorphic
However, in one member of the obscura species group (D.
theposteriorsegments (data notshown). Coupled withthe
discoveryof a weakbutfunctional ABD-B BS7binding site
and ABD-B responsiveness of the D. subobscura body
CRE, these observations suggest that ABD-B regulated
yellow expression and pigmentation in an ancestor of the
obscura group. These results underscore the potential for
knowledge of the underlying regulatory architecture of trait
formation to inform evolutionary reconstructions beyond
what is possible with phenotypes alone.
The absence of sexually dimorphic posterior pigmenta-
tion and the monomorphic expression of body CREs in
groups outside of the melanogaster and obscura groups
place the origin of male-specific pigmentation to within
the clade consisting of the melanogaster and obscura
species groups. The mechanism for the origin of elevated
posterior pigmentation is evident from the functional dis-
section of the body CRE here—namely, via the gain of
ABD-B binding sites and sites for any obligatory cofactors
in a CRE that controlled expression of the yellow gene in
the abdomen. This scenario for the gain of ABD-B regula-
tion is conceptually identical to that which we have put
forth for the co-option of extant transcription factors and
CREsin the evolution of novel patternsof geneexpression
in the wing (Gompel et al., 2005).
for the loss of male-specific pigmentation (Figure 7). First,
binding site BS7 is sufficient to inactivate an otherwise
functional element. Second, in D. bipectinata, a change
in the regulation of the bab locus from dimorphic to mono-
morphic expression is sufficient to account for the evolu-
tionary loss of male-specific pigmentation (while the
body CRE has retained responsiveness to Abd-B). And
third, in D. santomea, the sexually dimorphic activity of
theD. santomea body CREinD. melanogaster, theperfect
bab expression pattern in D. santomea indicate that none
of these three regulatory interactions have changed in D.
santomea; rather, evolutionary changes at other loci have
been involved in the loss of pigmentation (Llopart et al.,
tion patterns suggests that evolutionary loss of parts of
regulatory circuits may be fairly common (Prud’homme
et al., 2006).
Microevolution, Macroevolution, and Hox Genes
One of the longstanding goals of evolutionary biology has
been to identify genetic events responsible for morpho-
logical change and to elucidate how changes at the mo-
lecular level translate into phenotypic diversity. In recent
1396 Cell 125, 1387–1399, June 30, 2006 ª2006 Elsevier Inc.
years, two distinct approaches have been taken to meet
this goal. The first has been comparative studies of gene
expression during development, which have identified
correlations between the deployment of regulatory genes,
particularly Hox genes, and differences in morphology.
The second approach has been direct genetic analysis
of intraspecific variation and interspecific differences.
For many years, these two approaches have been far
apart because of their different scales of analysis. Most
comparative studies have focused on slowly evolving
traits among higher taxa, and genetic analyses have
been necessarily restricted to species that produce fertile
hybrids, among which the type and extent of morpholog-
tion has remained open in some minds as to whether the
mechanisms that produce morphological variation and
divergence at the species level are sufficient to explain
the larger-scale divergences at higher taxonomic levels.
Figure 7. Model for the Evolutionary Gain and Loss of Posterior Pigmentation in the melanogaster Species Group
The pigmentation patterns, BAB expression patterns, and body CRE-driven reporter expression patterns in A5 and A6 are mapped onto a phylogeny
of the species studied here. In the right column, the presence or absence of Abd-B-mediated activation of yellow expression is represented by + or ?
symbols. We infer that ABD-B regulation of yellow evolved in a common ancestor of the melanogaster and obscura species groups or the mela-
nogaster species group (green arrows to black bars). We further suggest three different molecular mechanisms underlying the loss of male-specific
pigmentation in different monomorphic lineages that descended from a common ancestor with male-specific pigmentation: a loss of dimorphic bab
regulation in D. bipectinata (black oval), a loss of ABD-B regulation of yellow (and probably other pigmentation genes) in D. kikkawai (teal oval), and
a loss of pigmentation through other loci in D. santomea. If the common ancestor of the melanogaster and obscura groups had male-specific
pigmentation, as we have inferred, then the pigmentation patterns of the obscura group represent another mode of modification of this pattern
that included some reduction of ABD-B regulation.
Cell 125, 1387–1399, June 30, 2006 ª2006 Elsevier Inc. 1397
morerapidly evolvingcharacters.Because ofthemajorin-
fluence thatthe comparative study of Hox genesand Hox-
regulated structures has had to date in evolutionary devel-
opmental biology, the demonstration of the evolution of
related Drosophila species should help address concerns
about the extrapolation of microevolutionary processes to
pigmentation and the molecular mechanisms of the evolu-
tionary gain and loss of Hox target regulation among
closely related species, the very sort of process that has
been proposed to explain the large-scale divergence of
homologous structures over much greater evolutionary
distances (Warren et al., 1994; Weatherbee et al., 1999;
Drosophila Strains and Culturing
We used the CantonSstrain as the wild-type form of D. melanogaster.
The babAr07allele is null for both bab1 and bab2 genes (Couderc et al.,
of ABD-B proteins in the A4 segment (Duncan, 1987). D. santomea, D.
subobscura, andD. kikkawai were obtained from J.Coyne,N. Gompel,
and J. David, respectively. D. biarmipes and D. bipectinata were ob-
tained from the Tucson Stock Center. The D. melanogaster yw strain
was used as the P element transformation host. All flies were reared
according to standard procedures.
Abdominal Cuticle Preparations
Flies (3–4 days old) were killed in ethanol and incubated in 10% KOH
solution for 1 hr at 65ºC. Adult abdominal cuticles were cut along the
dorsal midline and mounted with a drop of Hoyer’s medium (Ash-
burner, 1989). The cuticles were flattened with a coverslip and incu-
bated for 3 hr at 65ºC.
Reporter Constructs of D. melanogaster
All reporter plasmids for transgenic flies were made by polymerase
chain reaction (PCR) and/or restriction-enzyme-based strategies. For
the wb and body constructs, a 2.6 kb fragment located between
?2869 and ?269 and a 1.6 kb fragment located between ?1867 and
?269 from the transcription start site (+1), respectively, were sub-
cloned into the pH-Stinger vector carrying nuclear EGFP (Figure 1A
and Figure 2A; Barolo et al., 2004). Serially deleted constructs for six
subregions of the body element (BED1–BED6) and seven overlapping
constructs were made as represented in Figure 2. Each subregion is
indicated in Figure S3.
Imaging EGFP Signal
Forimaging EGFPexpressioninthepupal abdomens, pupae(85–90 hr
after puparium formation) were dissected from their pupal cases and
mounted in Halocarbon 700 oil. Oil-saturated pupae were imaged on
a Bio-Rad MRC 1024 confocal microscope.
Protein Production and DNA Binding Assays
The DNA sequence encoding the ABD-B HD was amplified by PCR
from an Abd-B cDNA plasmid (a gift from W. McGinnis) and the follow-
ing EcoRI or XhoI site-containing primers: ABD-B HD forward primer,
50-CGGAATTCGTGTCCGTCCGGAAAAAGCGCA-30; reverse primer,
50-CGCTCGAGTCAGTTGTTGTTGTTCTGCTGATTG-30. Amplified pro-
ducts were subcloned into the pGEX-4T-1 expression vector (Pharma-
cia Biotech, Inc.) and then verified by sequencing. The purification of
recombinant proteins was performed as described previously (Fran-
gioni and Neel, 1993). The final concentrations of N-laurylsarcosine
and Triton X-100 in STE buffer were 1.5% and 2.0% respectively. The
purified ABD-B HD proteins were used to conduct DNase I footprinting
experiments as described previously (Hersh and Carroll, 2005), except
that the protein-DNA binding reaction was carried out in 13 footprint
buffer (25 mM HEPES [pH 7.9], 50 mM KCl, 0.5 mM DTT, 6.25 mM
MgCl2, 0.025 mM EDTA, 8.5% glycerol). For EMSA assays, forward
and reverse oligonucleotides were annealed and end labeled with
Klenow enzyme and [a-32P]dNTP. Binding reactions were performed
in 13 footprint buffer containing 20 mg/ ml poly(dI-dC) at 4ºC for 30
min. DNA-protein complexes were analyzed by electrophoresis on
a 5% nondenaturing gel. Forward and reverse oligonucleotides used
for EMSA are summarized in Table S2.
Cloning and Mutagenesis of Orthologous yellow Enhancer
Genomic DNA was purified from D. santomea, D. bipectinata, and D.
kikkawai using the QIAamp DNA Mini Kit (QIAGEN). The orthologous
body enhancer elements were amplified by PCR with the following
degenerate primers: forward primer, 50-AKCGTKWGKCAATTATGCC
ARAGAG-30; reverse primer, 50-TAKCGCTCCTGVAGYTTGTAGGC-30.
Degenerate PCRproductsweresubcloned intothepGEM-TEasyvec-
tor (Promega), and five clones from four independent PCR reactions
were sequenced. Based on sequence alignment using DIALIGN,
each orthologous body element was subcloned into pH-Stinger vector
in a multistep process. For the sub body construct, the PCR reaction
was performed with a D. subobscura genomic plasmid (Wittkopp
et al., 2002b) and the following XbaI or BamHI site-containing primers:
forward primer, 50-GCTCTAGATCCGAAACAAGTGCAATTTCT-30; re-
verse primer, 50-CGGGATCCTTGCTCTCTGAGGCAGTTTTT-30. Am-
plified product was subcloned into pH-Stinger vector and verified by
sequencing. Mutations of the native body elements shown in Figures
6C–6E were generated by PCR. For all constructs, a minimum of
four independent lines were examined.
Supplemental Data include Supplementary Experimental Procedures,
Supplementary References, three figures, and two tables and can be
found with this article online at http://www.cell.com/cgi/content/full/
We thank W. McGinnis and P. Wittkopp for DNA clones; J. Coyne, N.
Gompel, and J. David for stocks; B. Hersh for technical comments; K.
Vaccaro for technical assistance; and B. Prud’homme and T. Williams
for suggestions on the manuscript. S.J. was supported by the Korea
Science and Engineering Foundation, and the work was supported
by the Howard Hughes Medical Institute (S.B.C).
Received: February 6, 2006
Revised: March 24, 2006
Accepted: April 25, 2006
Published: June 29, 2006
Abzhanov, A., Popadic, A., and Kaufman, T.C. (1999). Chelicerate Hox
genes and the homology of arthropod segments. Evol. Dev. 1, 77–89.
Ashburner, M. (1989). Drosophila: A Laboratory Manual (Cold Spring
Harbor, NY, USA: Cold Spring Harbor Laboratory Press).
Averof, M., and Akam, M. (1995). Hox genes and the diversification of
insect and crustacean body plans. Nature 376, 420–423.
Averof, M., and Patel, N.H. (1997). Crustacean appendage evolution
associated with changes in Hox gene expression. Nature 388, 682–
1398 Cell 125, 1387–1399, June 30, 2006 ª2006 Elsevier Inc.
Baker, B.S., and Ridge, K.A. (1980). Sex and the single cell. I. On the
action of major loci affecting sex determination in Drosophila mela-
nogaster. Genetics 94, 383–423.
Barolo, S., Castro, B., and Posakony, J.W. (2004). New Drosophila
transgenic reporters: insulated P-element vectors expressing fast-
maturing RFP. Biotechniques 36, 436–440, 442.
Belting, H.G., Shashikant, C.S., and Ruddle, F.H. (1998). Modification
of expression and cis-regulation of Hoxc8 in the evolution of diverged
axial morphology. Proc. Natl. Acad. Sci. USA 95, 2355–2360.
Burke, A.C., Nelson, C.E., Morgan, B.A., and Tabin, C. (1995). Hox
genes and the evolution of vertebrate axial morphology. Development
Burtis, K.C. (1993). The regulation of sex determination and sexually
dimorphic differentiation in Drosophila. Curr. Opin. Cell Biol. 5, 1006–
Carbone, M.A., Llopart, A., deAngelis, M., Coyne, J.A., and Mackay,
T.F. (2005). Quantitative trait loci affecting the difference in pigmenta-
tion between Drosophila yakuba and D. santomea. Genetics 171, 211–
Carroll, S.B. (1995). Homeotic genes and the evolution of arthropods
and chordates. Nature 376, 479–485.
Carroll, S.B. (2000). Endless forms: the evolution of gene regulation
and morphological diversity. Cell 101, 577–580.
Carroll, S.B., Weatherbee, S.D., and Langeland, J.A. (1995). Homeotic
genes and the regulation and evolution of insect wing number. Nature
Carroll, S.B., Grenier, J.K., and Weatherbee, S.D. (2005). From DNA to
Diversity: Molecular Genetics and the Evolution of Animal Design,
Third Edition (Malden, MA, USA: Blackwell Science Press).
Proc. Natl. Acad. Sci. USA 90, 1566–1570.
Celniker, S.E., Sharma, S., Keelan, D.J., and Lewis, E.B. (1990). The
molecular genetics of the bithorax complex of Drosophila: cis-regula-
tion in the Abdominal-B domain. EMBO J. 9, 4277–4286.
and axial patterning in snakes. Nature 399, 474–479.
Couderc, J.L., Godt, D., Zollman, S., Chen, J., Li, M., Tiong, S., Cram-
ton, S.E., Sahut-Barnola, I., and Laski, F.A. (2002). The bric ? a brac
locus consists of two paralogous genes encoding BTB/POZ domain
proteins and acts as a homeotic and morphogenetic regulator of
imaginal development in Drosophila. Development 129, 2419–2433.
Duncan, I. (1987). The bithorax complex. Annu. Rev. Genet. 21, 285–
Ekker, S.C., Jackson, D.G., von Kessler, D.P., Sun, B.I., Young, K.E.,
and Beachy, P.A. (1994). The degree of variation in DNA sequence
recognition among four Drosophila homeotic proteins. EMBO J. 13,
Frangioni, J.V., and Neel, B.G. (1993). Solubilization and purification of
enzymatically active glutathione S-transferase (pGEX) fusion proteins.
Anal. Biochem. 210, 179–187.
Garcia-Bellido, A. (1975). Genetic control of wing disc development in
Drosophila. Ciba Found. Symp. 29, 161–178.
Geyer, P.K., and Corces, V.G. (1987). Separate regulatory elements
mental transcription of the yellow locus in Drosophila melanogaster.
Genes Dev. 1, 996–1004.
Gompel, N., and Carroll, S.B. (2003). Genetic mechanisms and con-
straints governing the evolution of correlated traits in drosophilid flies.
Nature 424, 931–935.
Gompel, N., Prud’homme, B., Wittkopp, P.J., Kassner, V.A., and Car-
roll, S.B. (2005). Chance caught on the wing: cis-regulatory evolution
and the origin of pigment patterns in Drosophila. Nature 433, 481–487.
Hersh, B.M., and Carroll, S.B. (2005). Direct regulation of knot gene
expression by Ultrabithorax and the evolution of cis-regulatory ele-
ments in Drosophila. Development 132, 1567–1577.
Hopmann, R., Duncan, D., and Duncan, I. (1995). Transvection in the
pendent interactions in trans. Genetics 139, 815–833.
Kopp, A.,Duncan,I., Godt,D., andCarroll, S.B. (2000). Genetic control
and evolution of sexually dimorphic characters in Drosophila. Nature
Lewis, D.L., DeCamillis, M., and Bennett, R.L. (2000). Distinct roles of
the homeotic genes Ubx and abd-A in beetle embryonic abdominal
appendage development. Proc. Natl. Acad. Sci. USA 97, 4504–4509.
ila santomea. Evolution Int. J. Org. Evolution 56, 2262–2277.
Lohmann, I., McGinnis, N., Bodmer, M., and McGinnis, W. (2002). The
Drosophila Hox gene deformed sculpts head morphology via direct
regulation of the apoptosis activator reaper. Cell 110, 457–466.
Mahaffey, J.W. (2005). Assisting Hox proteins in controlling body form:
are there new lessons from flies (and mammals)? Curr. Opin. Genet.
Dev. 15, 422–429.
Mahfooz, N.S., Li, H., and Popadic, A. (2004). Differential expression
patterns of the hox gene are associated with differential growth of
insect hind legs. Proc. Natl. Acad. Sci. USA 101, 4877–4882.
Pearson, J.C., Lemons, D., and McGinnis, W. (2005). Modulating Hox
Prud’homme, B.,Gompel, N.,Rokas, A.,Kassner, V.A., Williams,T.M.,
Yeh, S., True, J.R., and Carroll, S.B. (2006). Repeated morphological
evolution through cis-regulatory changes in a pleiotropic gene. Nature
Stern, D.L. (1998). A role of Ultrabithorax in morphological differences
between Drosophila species. Nature 396, 463–466.
Stern, D.L. (2000). Evolutionary developmental biology and the prob-
lem of variation. Evolution Int. J. Org. Evolution 54, 1079–1091.
Telford, M.J., and Thomas, R.H. (1998). Expression of homeobox
genes shows chelicerate arthropods retain their deutocerebral seg-
ment. Proc. Natl. Acad. Sci. USA 95, 10671–10675.
Tomoyasu, Y., Wheeler, S.R., and Denell, R.E. (2005). Ultrabithorax is
required for membranous wing identity in the beetle Tribolium casta-
neum. Nature 433, 643–647.
Warren, R.W., Nagy, L., Selegue, J., Gates, J., and Carroll, S. (1994).
Evolution of homeotic gene regulation and function in flies and butter-
flies. Nature 372, 458–461.
Weatherbee, S.D., and Carroll, S.B. (1999). Selector genes and limb
identity in arthropods and vertebrates. Cell 97, 283–286.
Weatherbee, S.D., Nijhout, H.F., Grunert, L.W., Halder, G., Galant, R.,
Selegue, J., and Carroll, S. (1999). Ultrabithorax function in butterfly
wings and the evolution of insect wing patterns. Curr. Biol. 9, 109–115.
Wittkopp, P.J., True, J.R., and Carroll, S.B. (2002a). Reciprocal func-
tions of the Drosophila Yellow and Ebony proteins in the development
and evolution of pigment patterns. Development 129, 1849–1858.
Wittkopp, P.J., Vaccaro, K., and Carroll, S.B. (2002b). Evolution of
yellow gene regulation and pigmentation in Drosophila. Curr. Biol.
Cell 125, 1387–1399, June 30, 2006 ª2006 Elsevier Inc. 1399