ArticlePDF Available

Struhl, G. Genes controlling segmental specification in the Drosophila thorax. Proc. Natl Acad. Sci. USA 79, 7380-7384

Authors:

Abstract and Figures

The roles of three homeotic genes, Ubx+, Scr+, and Antp+, in the Drosophila thorax have been studied by determining the cellular phenotypes of mutations resulting in loss of gene function. The principal results are: (i) The Scr+ and Ubx+ genes are required in the prothorax and metathorax, respectively; in the absence of these genes, both segments develop like the mesothorax. (ii) The Antp+ gene is required in all three thoracic segments: in its absence, parts of the mesothorax are transformed into corresponding parts of the antenna, and similar transformations to antenna are found in the prothorax and metathorax if the Scr+ and Ubx+ genes also are absent. (iii) Loss of the Ubx+ gene early in embryogenesis, but not later, leads to the inappropriate activity of the Scr+ gene in the meso- and metathorax. Results i and ii argue strongly that segmental determination is specified in a combinatorial fashion in the head and thorax by the selective activities of the Scr+, Ubx+, Antp+, and putative head-determining genes. Result iii suggests that a product of the Ubx+ gene also plays an early, regulatory role in ensuring the correct spatial expression of the Scr+ gene during subsequent development.
Phenotypes of mutant clones. (A-D) Sibling Ubx1 and Scrl3A Ubx' clones induced in the posterior compartments of the first and second legs around the blastoderm stage (mitotic recombination in Scrl3A Ubx1 e1/Ki Sb63b M(3)w'24 cells can give rise to Scrl3A Ubx' clones if it occurs proximal to the Ki locus or to Kinked Ubx' clones if it occurs distal to the Scr locus). (A) Kinked Ubx' clone in the posterior femur of the first leg showing the normal bristle pattern. (x80.) (B) Scrl3A Ubx' clone in the posterior femur of the first leg that forms the bristle pattern normally found in the posterior femur of the second leg (compare with D). (x80.) (C) Kinked Ubx' clone in the second leg that forms the bristle pattern normally found in the first leg (compare withA). (x80.) (D) Scrl3A Ubx1 clone in the second legformingthe normal bristle pattern. (x80.) (E)ScrClAntpN8+RC3 Ubx1 clone induced in the anterior compartment of the first leg during the late first instar. In addition to normal second leg structures (e.g., s, the sternopleura bristles), this clone also forms tissue characteristic of the first (I), second (II), and third (E) antennal segments. (x 150.) (F)AntpN8+RC3 Ubx' clone induced in the anterior compartment of the first leg around the blastoderm stage. This clone forms an abnormal pattern of bristles and is associated with fusion of the femur and tibia. The remainder of the clone forms normal structures in the proximal and distal portions of the leg, including the sex comb (arrow). (x 100.) (G) Detail of the clone inE showing sensilla characteristic of the normal third antennal segments. (x200.)
… 
Content may be subject to copyright.
Proc.
NatL
Acad.
Sci.
USA
Vol.
79,
pp.
7380-7384,
December
1982
Developmental
Biology
Genes
controlling
segmental
specification
in
the
Drosophila
thorax
(homeosis/insect
segments/determination/clonal
analysis)
GARY
STRUHL*
Medical
Research
Council
Laboratory
of
Molecular
Biology,
Hills
Road,
Cambridge
CB2
2QH,
England
Communicated
by
M.
F.
Perutz,
August
25,
1982
ABSTRACT
The
roles
of
three
homeotic
genes,
Ubx+,
Scr+,
and
Antp+,
in
the
Drosophila
thorax
have
been
studied
by
deter-
mining
the
cellular
phenotypes
of
mutations
resulting
in
loss
of
gene
function.
The
principal
results
are:
(i)
The
Scr+
and
Ubx+
genes
are
required
in
the
prothorax
and
metathorax,
respectively;
in
the
absence
of
these
genes,
both
segments
develop
like
the
me-
sothorax.
(ii)
The
Antp+
gene
is
required
in
all
three
thoracic
seg-
ments:
in
its
absence,
parts
of
the
mesothorax
are
transformed
into
corresponding
parts
of
the
antenna,
and
similar
transformations
to
antenna
are
found
in
the
prothorax
and
metathorax
if
the
Scr+
and
Ubx+
genes
also
are
absent.
(iii)
Loss
of
the
Ubx+
gene
early
in
embryogenesis,
but
not
later,
leads
to
the
inappropriate
activity
of
the
Scr+
gene
in
the
meso-
and
metathorax.
Results
i
and
ii
argue
strongly
that
segmental
determination
is
specified
in
a
com-
binatorial
fashion
in
the
head
and
thorax
by
the
selective
activities
of
the
Scr+,
Ubx+, Antp+,
and
putative
head-determining
genes.
Result
iii
suggests
that
a
product
of
the
Ubx+
gene
also
plays
an
early,
regulatory
role
in
ensuring
the
correct
spatial
expression
of
the
Scr+
gene
during
subsequent
development.
Insects
are
composed
of
a
series
of
head,
thoracic,
and
abdom-
inal
segments,
each
displaying
unique
and
sometimes
dramat-
ically
different
characteristics.
Yet,
in
the
absence
of
particular
homeotic
genes,
one
or
more
segments
can
be
transformed
into
other
segments
(1-4).
Thus,
segments
are
homologous
units
in
which
diverse
cell
patterns
develop
under
discrete
genetic
controls.
Of
the
many
homeotic
genes
that
have
been
described
in
Drosophila,
several
have
the
distinguishing
feature
that
they
must
be
active
in
some
segments
and
inactive
in
others
[e.g.,
genes
of
the
bithorax-complex
(1,
2)
and
the
Antp+
(5-7)
and
Scr+
(8-10,
6)
genes].
Such
segment-specific
genes
appear
to
dictate
the
particular
developmental
pathways
followed
by
the
different
segments.
Recently,
a
second
class
of
homeotic
genes
has
been
identified
whose
members
appear
to
be
required
in
most,
or
all,
segments.
The
products
of
two
of
these
genes
[Pc+
(2,
11)
and
esc+
(12)]
appear
to
be
required
for
the
selective
expression
of
the
bithorax-complex
genes
and
possibly
other
segment-specific
genes.
Thus,
these
gene
products
may
have
regulatory
roles
in
ensuring
the
correct
expression
of
segment-
specific
homeotic
genes.
In
this
report
I
describe
the
interdependent
roles
of
three
segment-specific
homeotic
genes
(Ubx+,
Antp+,
and
Scr+)
in
the
cells
giving
rise
to
the
adult
thorax
of
Drosophila.
The
re-
sults
argue
strongly
that
these
genes
act
in
a
combinatorial
fash-
ion
to
specify
segmental
determination.
In
addition,
they
sug-
gest
that
the
products
of
some
of
these
genes
also
have
regulatory
roles
in
controlling
the
selective
expression
of
others.
METHODS
Genotypes
Employed.
Except
where
otherwise
indicated,
all
mutations
and
chromosomes
are
described
in
ref.
13.
The
Ubx'
mutation
behaves
as
a
hypomorph,
resulting
in
the
loss
of
most,
but
probably
not
all,
of
the
wild-type
gene
function
(14,
15).
The
AntpNs+RC3
mutation
is
an
apparent
null
allele
of
the
Antp+
gene
(7).
The
Scrl3A
mutation
is
a
recessive
lethal
allele
kindly
provided
by
C.
Nusslein-Volhard.
The
ScrCl
mu-
tation
was
induced
on
a
chromosome
already
bearing
the
AntpNS+RC3
mutation
with
the
mutagen
ethyl
methanesulfonate
(unpublished
data).
Both
the
Scrl3A
and
ScrCl
mutations
cause
embryonic
lethality
when
trans
to
other
Scr
mutations;
also,
such
mutant
embryos
show
the
characteristic
homeotic
phe-
notype
resulting
from
loss
of
the
Scr+
gene
described
previously
(6).
Analysis
of
Mutant
Clones.
Marked
clones of
cells
homo-
zygous
for
mutations
of
the
Scr,
Antp,
and
Ubx
loci
were
ob-
tained
as
described
below
for
the
triple
mutant
combination
ScrCl
AntpNs+RC3
Ubx'.
A
chromosome
carrying
all
three
mu-
tations
as
well
as
the
ell
mutation
was
constructed,
and
embryos
or
larvae
of
the
genotype
Scrcl
AntpNs+RC3
Ubx'
e"/Ki
Sb63b
M(3)w"24
were
irradiated
at
appropriate
times
after
egg
laying
(3
±
1
hr
=
blastoderm
stage
and
24-48
hr
=
first
larval
instar;
all
experiments
were
performed
at
25°C).
All
the
mutations
are
located
on
the
right
arm
of
the
third
chromosome;
because
Ki
is
positioned
closest
to
the
centromere,
mitotic
recombinations
that
occur
between
Ki
and
the
centromere
can
give
rise
to
homozygous
Ki+
Sb+
M(3)w+
cells,
which
must
also
be
homozy-
gous
for
the
Scrl3A,
AntpNs+RC3,
Ubxl,
and
ell
mutations.
The
Kinked,
Stubble
b,
and
Minute(3)w"2"
mutations
are
dominant
and
affect
bristle
morphology;
the
ebony"
mutation
is
recessive
and
causes
cuticle,
especially
bristles,
to
be
heavily
pigmented.
Thus,
clones
of
homozygous
cells
bear
blackened,
but
otherwise
normal,
bristles,
in
contrast
to
surrounding
heterozygous
cells
that
bear
gnarled,
stunted
bristles.
In
addition
to
affecting
bris-
tle
morphology,
the
M(3)w"24
mutation
also
causes
heterozy-
gous
cells
to
divide
at
slower
than
normal
rates
(16).
Conse-
quently,
homozygous
Ki+
cells
carrying
two
copies
of
the
M(3)w+
allele
are
able
to
grow
faster
than
surrounding
hetero-
zygous
cells
(17)
and
hence,
can
form
large
portions
of
the
adult
compartments
(18)
to
which
they
contribute.
Clones
were
identified
either
under
the
dissecting
or
com-
pound
microscope
by
the
appearance
of
marked
bristles
and
because
they
frequently
disrupted
the
normal
morphology.
Clones
in
several
parts
of
the
fly
that
normally
carry
few,
or
no,
large
bristles
were
difficult
to
detect
and
hence,
may
have
been
missed.
In
addition,
because
of
the
large
numbers
of
flies
in-
volved
(ca.
15,000),
it
was
not
possible
to
screen
all
parts
of
all
flies
with
equal
thoroughness.
For
these
reasons,
the
frequen-
cies
of
clones
in
different
parts
of
the
fly,
or
ofclones
of
different
*
Present
address:
Dept.
of
Biochemistry
and
Molecular
Biology,
Har-
vard
University,
7
Divinity
Ave.,
Cambridge,
MA
02138.
7380
The
publication
costs
of
this
article
were
defrayed
in
part
by
page
charge
payment.
This
article
must
therefore
be
hereby
marked
"advertise-
ment"
in
accordance
with
18
U.
S.
C.
§1734
solely
to
indicate
this
fact.
Downloaded by guest on February 19, 2021
Proc.
Natl.
Acad.
Sci.
USA
79
(1982)
7381
genotypes
in
the
same
part,
cannot
be
usefully
compared.
All
clones
identified
under
the
dissecting
microscope
were
subsequently
examined
under
the
compound
microscope.
RESULTS
The
independent
roles
of
the
Ultrabithorax+
(Ubx+)
and
An-
tennapedia+
(Antp')
genes
have
been
studied
previously
by
inducing
marked
clones
of
cells
homozygous
or
hemizggous
for
mutations
of
either
locus
(e.g.,
Ubx1
and
AntpNs+Rc
)
in
em-
bryos
or
larvae
that
are
otherwise
heterozygous
(1,
7,
15,
19).
Here,
this
approach
has
been
applied
to
the
Sex
combs
reduced'
(Scr')
gene
(8-10)
and
to
all
possible
combinations
of
the
Antp',
Scr',
and
Ubx+
genes
by
generating
marked
clones
of
the
following
mutant
genotypes:
Scrl3A,
Scrcl
Ant
Ns+RC3
Scr3A
Ubx1,
AntpNs+RC3
Ubxl,
and
Scrcl
AntpNs+R
3
Ubx1
The
phenotypes
of
such
clones
are
summarized
in
Table
1,
il-
lustrated
in
Fig.
1,
and
described
below.
Ubx1
Clones.
Clones induced
after
8
hr
following
fertilization
are
normal
in
the
pro-
and
mesothoracic
segments
but
transform
the
dorsal
and
ventral
appendages
of
the
metathorax
(haltere
and
third
leg)
into
their
counterparts
in
the
mesothorax
(wing
and
second
leg)
(1,
15,
19).
Clones
induced
earlier
are
identical
except
that
they
transform
the
posterior
compartments
of
both
the
second
and
third
legs
into
the
posterior
compartment
of
the
first
leg
(refs.
15
and
20;
e.g.,
Fig.
1).
Thus,
the
Ubx+
gene
is
required
in
the
metathorax
to
specify
meta-
as
opposed
to
me-
sothoracic
development.
In
addition,
it
appears
to
have
an
early
function
in
portions
of
the
posterior
meso-
and
metathorax
in
preventing
prothoracic
development.
Scrl3A
Clones.
The
Scr+
gene,
which
maps
close
to
the
Antp+
gene,
has
been
defined
by
a
series
of
dominant
and
re-
cessive
mutations
(8-10).
Though
most
of
these
mutations
result
in
lethality
when
homozygous,
one
mutant
allele,
ScrEdRl8,
is
viable
when
trans
to
other
recessive
mutations
of
the
gene;
such
mutant
flies
show
transformations
of
the
first
leg
to
second
leg
and
of
the
proboscis
to
maxillary
palp,
but
the
meso-
and
me-
tathoracic
segments
are
normal
(10).
Clones
of
cells
homozygous
for
the
Scr
3Amutation
were
induced
around
the
blastoderm
stage
or
during
subsequent
development.
These
clones
are
nor-
mal
in
both
the
meso-
and
metathoracic
segments
but
cause
autonomous
transformation
of
all
portions
of
the
first
leg
into
corresponding
portions
of
the
second
leg
(Fig.
1).
A
few
Scrl3A
clones
were
observed
in
the
dorsal
derivative
of
the
prothorax,
the
humerus.
However,
these
clones
were
small,
marking
one,
or
at
most
a
few
bristles,
and
hence,
difficult
to
interpret.
Fi-
nally,
clones
in
the
proboscis
autonomously
transformed
the
labial
palps
into
maxillary
palps,
but
clones
in
the
eye-antenna
were
normal.
These
results
indicate
that
the
Scr'
gene
is
not
required
either
in
the
meso-
or
metathorax
or
in
the
eye-an-
tenna
but
that
it
is
required
in
at
least
the
ventral
prothorax
for
specifying
pro-
as
opposed
to
mesothoracic
development.
In
addition,
they
indicate
that
the
Scr+
gene
is
required
in
the
proboscis
for
specifying
labial
as
opposed
to
maxillary
devel-
opment.
Scrl3A
Ubxl
Clones.
The
phenotype
of
Scr
BA
Ubx'
clones
induced
around
the
blastoderm
stage
and
during
larval
devel-
opment
is
the
additive
phenotype
of
Scrl3A
and
Ubx1
clones
induced
during
larval
development
(Fig.
1).
Thus,
in
the
thorax,
doubly
mutant
cells
transform
the
pro-
and
metathorax
into
mesothorax
but
do
not
affect
the
mesothorax
itself.
This
result
indicates
that
the
transformation
of
portions
of
the
meso-
and
metathorax
to
prothorax
caused
by
early
loss
of
the
Ubx+
gene
(15,
20)
only
occurs
if
the
Scr'
gene
is
present
and
hence,
sug-
gests
that
this
transformation
results
from
the
expression
of
the
Scr+
gene
in
the
meso-
and
metathorax
where
it
should
nor-
mally
be
inactive.
Table
1.
Phenotypes
and
numbers
of
mutant
clones
Genotypet
Petp
and
number§
and
time
Jhenotype
X-ray
of
clone
Eye-
Pro-
Leg
1
Leg
2
Leg
3
Hal-
dose,
induction,
hr
antenna
boscis
Ant.:Post.
Ant.:Post.
Ant.:Post.
Wing
tere
Flies,
no.
radsl
wild
type
EA
P
1
2
3
W
H
-
Ubx:4
+
2
EA
P
1
2:1*
2:1*
W
W
Refs.
15,20
Ubx:48
+
4
EA
P
1
2 2
W
W
Refs.
15,
19,20
Scr
EA
M
2
2
3
W
3
±
1
&
48
±
4
45
25
27:13
60:9
33:13
21
-
4,280
750,
1,000
Scr
Ubx
EA
M
2
2
2
W W
3
±
1
5 2
5:12
22:23
20:15
8
10
2,325
750
48
±
4
13
4
7:3
19:7
13:1
36
6
1,500
1,000
Antp
EA
P
(1)
A*
(3)
(W)
-
3
±
1
&
24-48
61
23
7(13):33(40)
49(64):27(31)
16(27):4(10)
10(18)
-
3,090
500,
1,000
ScrAntp
Ubx
EA
M
A*
A*
A*
(W)
W
3
±
1
&
48
±
4
64
16
15:12
30(35):11
30(41):17
11(20)
24
2,200
500,
1,000
AntpUbx
EA
-
(1)
A*
A*
(W)
W
3
±
1
1
1(2):0
2:2
5:0
0
0
250
750
48
±
4
12
-
3(9):2(5)
11(12):9
15:2
2
14
475
1,000
Scr
Antp
EA
M
A*
A*
(3)
(W)
-
3
±
1
&
48
±
4
18
1
10:5
4(8):1
2(3):0(2)
3(6)
1,400
750,
1,000
t
Genotypes
are
shown
in
boldface;
time
of
clone
induction
(hours
after
egg
laying)
is
given
in
standard
type
underneath.
Full
genotypes
are
given
in
Results.
t
Phenotypes
are
shown
in
boldface.
Ant.
and
Post. refer
to
the
anterior
and
posterior
compartments
of
the
legs
which
are
treated
separately
when
they
differ
in
phenotype.
1*
=
prothoracic
pattern
formed
in
proximal
posterior
leg
instead
of
mesothoracic
pattern.
A*
=
antennal
structures
or
antennal
and
mesothoracic
structures
formed.
Parentheses
[e.g.,
(W)]
=
abnormal
or
abbreviated
cuticular
patterns
formed.
M
=
maxillary
tissue
formed
instead
of
labial
tissue
in
the
proboscis.
§
Numbers
of
clones
having
the
phenotype
shown
in
boldface
are
given
in
standard
type
underneath.
Clones
in
the
anterior
and
posterior
com-
partments
of
the
legs
are
treated
separately.
Numbers
in
parentheses
indicate
total
numbers
of
clones
when
only
some
of
the
clones
showed
the
phenotype
indicated
in
boldface
and
the
remainder
appeared
normal.
NOne
rad
=
0.01
gray.
Developmental
Biology:
Struhl
Downloaded by guest on February 19, 2021
7382
Developmental
Biology:
Struhl
Vj
/
--.-
\
9
an")
'\
D
An
-
/.
.-7/.
'
N
b'
(
it
.)
\,
-
--
!L
\
_-
-,
_.,
'
...
-
-J
-
e
i
/
A
1-
N
I
,
L
\
i--
S
A,
ll
.
.
k
= b"!;,
~~~~~~~~~~~~~~~~~~~..-P.
....................
1s
'it
.'._
~~~F
I-I
ij
l
.
i'
\
\
D
4v
E
1
G
a
FIG.
1.
Phenotypes
of
mutant
clones.
(A-D)
Sibling
Ubx1
and
Scrl3A
Ubx'
clones
induced
in
the
posterior
compartments
of
the
first
and
second
legs
around
the
blastoderm
stage
(mitotic
recombination
in
Scrl3A
Ubx1
e1/Ki
Sb63b
M(3)w'24
cells
can
give
rise
to
Scrl3A
Ubx'
clones
if
it
occurs
proximal
to
the
Ki
locus
or
to
Kinked
Ubx'
clones
if
it
occurs
distal
to
the
Scr
locus).
(A)
Kinked
Ubx'
clone
in
the
posterior
femur
of
the
first
leg
showing
the
normal
bristle
pattern.
(x80.)
(B)
Scrl3A
Ubx'
clone
in
the
posterior
femur
of
the
first
leg
that
forms
the
bristle
pattern
normally
found
in
the
posterior
femur
of
the
second
leg
(compare
with
D).
(x80.)
(C)
Kinked
Ubx'
clone
in
the
second
leg
that
forms
the
bristle
pattern
normally
found
in
the
first
leg
(compare
withA).
(x80.)
(D)
Scrl3A
Ubx1
clone
in
the
second
legformingthe
normal
bristle
pattern.
(x80.)
(E)ScrClAntpN8+RC3
Ubx1
clone
induced
in
the
anterior
compartment
of
the
first
leg
during
the
late
first
instar.
In
addition
to
normal
second
leg
structures
(e.g.,
s,
the
sternopleura
bristles),
this
clone
also
forms
tissue
characteristic
of
the
first
(I),
second
(II),
and
third
(E)
antennal
segments.
(x
150.)
(F)AntpN8+RC3
Ubx'
clone
induced
in
the
anterior
compartment
of
the
first
leg
around
the
blastoderm
stage.
This
clone
forms
an
abnormal
pattern
of
bristles
and
is
associated
with
fusion
of
the
femur
and
tibia.
The
remainder
of
the
clone
forms
normal
structures
in
the
proximal
and
distal
portions
of
the
leg,
including
the
sex
comb
(arrow).
(x
100.)
(G)
Detail
of
the
clone
inE
showing
sensilla
characteristic
of
the
normal
third
antennal
segments.
(x200.)
AntpN8+RC3
Clones.
The
phenotype
of
homozygous
AntpNs+RC3
clones
has
been
described
(7).
Whether
induced
around
the
blastoderm
stage
or
during
larval
development,
such
clones
transform
portions
of
the
second
leg
(both
anterior
and
posterior
compartments)
into
corresponding
portions
of
the
antenna.
However,
some
portions
of
the
leg,
notably
the
distal
tarsus,
develop
normally
even
when
mutant.
Some
clones
in
the
wing
and
mesonotum
as
well
as in
the
first
and
third
legs
show
ab-
normal
or
abbreviated
patterns
of
bristles,
but
others
appear
normal
(Table
1;
Fig.
1).
Finally,
all
clones
in
the
head
are
nor-
mal.
These
results
led
to
the
conclusion
that
the
Antp+
gene
is
required
in
all
three
thoracic
segments,
though
only
in
the
ventral
second
leg
could
the
gene
be
said
to
be
required
for
specifying
one
determined
state
(mesothorax)
as
opposed
to
an-
other
(eye-antenna)
(7).
ScrCl
AntpNs+RC3
Ubx'
Clones.
The
phenotype
of
triply
mu-
tant
cells
induced
either
around
the
blastoderm
stage
or
during
subsequent
development
is
the
additive
phenotype
of
AntpNs+RC
and
Scrl3A
Ubx1
clones-i.e.,
such
clones
transform
corre-
sponding
portions
of
all
three
legs
into
antennal
structures
(as
in
AntpNs+RC3
clones
in
the
second
leg),
whereas
remaining
portions
of
these
legs,
which
are
not
transformed
into
antenna,
develop
as
in
the
mesothorax
(Fig.
1).
The
phenotypes
of
triply
mutant
cells
in
the
wing
and
mesonotum
are
indistinguishable
from
that
of
equivalent
AntpNs+RC3
clones;
triply
mutant
cells
in
the
haltere
and
metanotum
formed
wing
and
mesonotum
structures.
Finally,
ScrCl
AntpNs+RC3
UbxJ
clones
were
com-
pletely
normal
in
the
eye-antenna
but
transformed
the
labial
palps
of
the
proboscis
into
maxillary
palps.
These
results
suggest
that
the
Antp+
gene
is
required
in
all
three
thoracic
segments
for
specifying
one
determined
state
(mesothorax)
as
opposed
to
another
(eye-antenna)
but
that
this
requirement
is
obscured
in
the
pro-
and
metathorax
by
the
activities
of
the
Scr'
and
Ubx+
genes.
Scrcl
AntpNs+RC3
Clones.
ScrCl
AntpNs+RC3
clones
induced
in
the
head
and
pro-
and
mesothorax
during
embryonic
and
lar-
val
development
were
indistinguishable
in
phenotype
from
tri-
ply
mutant
clones
induced
in
the
same
segments.
Clones
in-
duced
in
the
metathorax
were
indistinguishable
from
AntpNs+RC3
clones
in
the
metathorax.
AntpN8+RC3
UbX'
Clones.
AntpNs+RC3
Ubxl
clones
induced
in
the
meso-
and
metathorax
during
larval
development
were
indistinguishable
in
phenotype
from
triply
mutant
clones
in-
duced
in
the
same
segments.
Clones
induced
in
the
head
and
prothorax
were
indistinguishable
from
AntpNs+RC3
clones
in
the
head
and
prothorax
(Fig.
1).
Clones
induced
at
the
blastoderm
stage
behaved
like
clones
induced
during
larval
development
with
one
curious
exception.
One
of
the
two
clones
obtained
in
the
posterior
compartment
of
the
second
leg
differentiated
pro-
thoracic
structures
in
the
proximal
leg
and
antennal
structures
in
the
distal
leg.
Normally,
AntpNS+RC3
clones
are
able
to trans-
form
only
the
second
leg
into
antenna.
Hence,
the
distal
portion
of
this
clone
behaved
like
an
AntpN$+Rc3
clone
in
the
distal
sec-
ond
leg,
even
though
the
proximal
portion
differentiated
pro-
thoracic
structures.
This
result
suggests
that
Ubx}
clones
in-
duced
around
the
blastoderm
stage
may
transform
only
proximal
portions
of
the
posterior
second
leg
into
corresponding
portions
of
the
first
leg.
Proc.
Natl.
Acad.
Sci.
USA
79
(1982)
f
-
-1
k-./
Downloaded by guest on February 19, 2021
Proc.
Natl.
Acad.
Sci.
USA
79
(1982)
7383
DISCUSSION
'The
principal
results
to
be
considered
are:
(i)
The
prothorax
and
metathorax
require
the
activities
of
the
Scr'
and
Ubx+
genes,
respectively,
throughout
development;
in
the
absence
of
both
genes,
both
segments
develop
like
the
mesothorax.
(ii)
All
three
thoracic
segments
require
the
activity
of
the
Antp+
gene.
In
the
absence
of
this
gene,
part
of
the
mesothorax
is
transformed
into
corresponding
antennal
structures,
and
similar
transformations
to
antenna
are
observed
in
the
prothorax
and
metathorax
when
the
Scr+
and
Ubx+
genes
also
are
absent.
(iii)
Loss
of
the
Ubx+
gene
early
in
development,
but
not
later,
causes
posterior
por-
tions
of
the
meso-
and
metathorax
to
develop
as
in
the
prothorax
(15,
20).
This
phenotype
is
only
observed
when
the
Scr+
gene
is
present.
Result
i
provides
further
evidence
for
the
view
(1,
2,
4,
7)
that
the
mesothorax
is
an
epigenetic
ground
state
in
which
none
of
-the
segment-determining
genes
either
is
active
or
required
and
that
the
more
specialized
segments
of
the
head,
thorax,
and
abdomen
are
elaborations
of
this
ground
state
specified
by
the
activities
of
homeotic
genes..
In
contrast,,
results
ii
and
iii
in-
dicate
that
both
the
Antp+
and
Ubx+
genes
are
required
for
mesothoracic
development
and
hence,
challenge
this
hypothesis.
The
concept
of
a
ground
state
has
been
useful
in
interpreting
the
phenotypes
of
many
homeotic
genes
(1,
2, 4,
7).
Moreover,
it
makes
good
sense
in
evolutionary
terms,
because
insects
al-
most
certainly
arose
from
more
primitive,
millipede-like
ances-
tors
composed
mostly
of
similar-leg-bearing
segments
(21-23).
Therefore,
instead
of
discarding
this
concept,
I
suggest
the
fol-
lowing,
more
complex
interpretation
of
the
Ubx
and
Antp
phe-
notypes.
In
the
case
of
the
Ubx
phenotype,
result
iii
suggests
that
the
Ubx+
gene
may
be
required
in
the
posterior
mesothorax
only
for
preventing
the
Scr+
gene
from
being
expressed
inap-
propriately.
This
suggestion
is
supported
by
result
i,
because
in
the
absence
of
both
the
Ubx+
and
Scr+
genes,
the
mesothorax
develops
normally.
A
similar
argument
has
been
made
.for
the
role
of
the
Antp+
gene:
assuming
that
the
eye-antenna
segment
Head
Antp
Scr
Ubx
Scr
Ub
T2
Bady
IEA
Ant
ead@OOO
Ant
OO**@
~x
OO
O@
EA
1
2
3
FIG.
2.
Proposed
roles
(Left)-
and
realms
of
action
(Right)
of
the
Antp',
Scr',
and
Ubx+
genes.
The
prothorax
(T1),
metathorax
(T3),
and
eye-antenna
(EA)
of
the
adult
fly
are
elaborations
of
the
ground
state
mesothorax
(T2)
specified
by
the
continuous
activities
of
the
Scr',
Ubx+,
and
putative
head-determining
genes,
respectively
(shown
as
arrows
leading
from
T2
to
T1, T3,
and
EA)
during
embryonic
and
larval
development.
In
addition
to
its
role
in
specifying
metathoracic
devel-
opment,
the
Ubx+
gene
also
encodes
a
product
that
has
a
transient
early
role
(15,
20)
in
preserving
the
ground
state
by
preventing
in-
appropriate
activity
of
the
Scr+
gene
or
gene
product
during
subse-
quent
development
[curved
arrow
(broken
line)
blocking
the
activity
of
the
Scr+
gene].
Similarly,
the
Antp+
gene
product
is
required
to
preserve
the
ground
state
by
preventing
inappropriate
activity
of
the
head-determining
genes
or
gene
products
[curved
arrow
(solid
line)
blocking
the
activity
of
the
head-determining
genes],
though
unlike
the
Ubx
gene
product,
it
appears
to
be
required
continuously
(7).
The
realmsof
action
of
theAntp+,
Scr+,
Ubx+,
and
"Heacd
genes
in
the
eye-
antenna
and
thoracic
segments
(1,
2,
3)
are
indicated
in
matrix
form,
each
row
representing
a
gene
and
each
column
representing
a
segment.
Closed
circles,
gene
product
must
be
active in
at
least
part,
if
not
all,
of
the
segment.
Open
circles,
gene
product
must
be
absent
or
inactive.
Circles
with
dots,
gene
product
must
be
active
early
but
is
absent
or
inactive
subsequently.
is
an
elaboration of
the
ground
state
specified
by
one
or
more
head-determining
genes,
the
phenotypes
of
dominant
and
re-
cessive
mutations
of
the
Antp+
gene
suggest
that
it
is
required
in
ventral
portions
of
the
mesothorax
only
to
prevent
inappro-
priate
activity
of
the
head-determining
genes
or
their
products
+00
+0001
+@000
+'0000
+00'
+oo
+0o-o*e
+00
0
+0'0
+'0.00
0000
-0000
-00
-000
+0
0
00
-0000
EA
1
21*2:1*
EA
l
2
2
EA
2 2
3
EA
2
2
2
Ubx-early
Head
Antp
Scr
Ubx
Ubx-
later
Scr
Scr
Ubx
-
early
Antp
Ubx-
later
Antp
Scr
Antp
Scr
Ubx
-
early
FIG.
3.
Combinatorial
roles
of
the
Antp',
Scr',
Ubx+,
and
putative
head-determining
genes
in
specifying
segmental
determination
in
the
thorax
and
eye-antenna.
The
genetic
and
phenotypic
consequences
of
removing
these
genes
singly
or
in
combination
are'
diagramed
in
matrix
form
as
in
the
wild-type
case
shown
in
Fig.
2
(lightly
shaded
circles
and
half
circles
indicate
inappropriate
activity
of
gene
product;
full
descriptions
of
phe-
notypes
are
given
in
Results
and
Table
1
and
full
genotypes
are
given
in
Results).
Segments
in
which
gene
loss
has
no
effect
on
the
code
word
of
active
and
inactive
gene
products
develop
normally
and
are
not
boxed;
segments
in
which
gene
loss
results
in
an
inappropriate
code
word
are
homeoti-
cally
transformed
(boxes);
segments
in
which
a
nonsense
code
word
is
created
develop
abnormally
(boxes
with
dotted
lines).
Early
or
later
times
of
gene
loss
areindicated
where
relevant.
+
0-'0
_D
Q
.....
,
_
+
*
_Se
O
@,
0:+00
-0o0ooo
-o
-0.0I0I0:
-0000
+0-.'OIo:
+0Q.
-0000
+00
a
-O00C
+0
)1-
oAo(l)
e:
-
o3)
At
-
0
EA
(1)
(3)
EA
1)
A*
EA
A
(3)
EA
Pg
P
Antp
Developmental
Biology:
Struhl
Downloaded by guest on February 19, 2021
7384
Developmental
Biology:
Struhl
(7).
This
interpretation
predicts
that
in
the
absence
of
both
the
Antp+
and
head-determining
genes,
the
mesothorax
should
develop
normally.
However,
this
prediction
cannot
be
tested
until
mutations
that
inactivate
the
putative
head-determining
genes
are
isolated
In
summary,
both
the
Antp+
and
Ubx+
genes
may
act
in
the
mesothorax
only
as
regulatory
functions
that
pre-
serve
the
ground
state
by
preventing
the
inappropriate
activity
of
other
homeotic
genes
or
their
products.
According
to
this
hypothesis,
the
mesothorax
should
develop
normally
in
the
si-
multaneous
absence
of
the
bithorax-complex,
Antp',
Scr',
and
putative
head-determining
genes
and
hence,
correspond
to
an
epigenetic
ground
state.
The
proposed
roles
and
realms
of
action
of
the
Antp+,
Scr',
Ubx+,
and
putative
head-determining
genes
are
outlined
in
Fig.
2.
As
shown
in
Fig.
3,
mutations
that
cause
one
segment
to
express
a
combination
of
homeotic
gene
products
normally
expressed
by
another
segment
result
in
clear
homeotic
trans-
formations
(e.g.,
Ubxl,
Scrl3A);
however,
mutations
that
gen-
erate
novel
combinations
of
homeotic
gene
products
not
nor-
mally
found
in
any
segment
cause
the
development
of
abnormal
cell
patterns
(e.g.,
AntpNS+`
3
in
the
pro-
and
metathorax).
These
findings
argue
strongly
that
segmental
determination
is
specified
in
a
combinatorial
fashion-mutations
such
as
Scrl3A
and
Ubx}
resulting
in
inappropriate
code
words
and
the
AntpNs+RC3
mutation
sometimes
resulting
in
nonsense
code
words.
A
priori
it
is
not
possible
to
predict
the
phenotypic
con-
sequences
of
a
nonsense
code
word.
However,
as
in
the
case
of
the
AntpNs+RC3
mutation
in
the
pro-
and
metathorax,
it
may
be
possible
to
convert
a
nonsense
word
into
a
sense
word
by
eliminating
other
homeotic
genes
(Scr+
or
Ubx+),
thereby
lead-
ing
to
a
predictable
homeotic
phenotype.
It
has
been
convenient
to
classify
homeotic
genes
involved
in
segmental
determination
into
two
groups:
(i)
"selector"
genes
(24)
that
specify
segmental
determination
throughout
devel-
opment
[e.g.,
genes
of
the
bithorax-complex
(1,
2)
and
Antp
(5-7)
and
Scr
(6,
8-10)
loci]
and
(ii)
regulatory
genes
(2,
11,
25)
[e.g.,
the
Pc+
(2,
11)
and
esc+
(12)
genes]
that
are
required
in
all
segments
for
establishing
and
maintaining
the
segment-spe-
cific
expression
of
segmental
selector
genes.
However,
results
ii
and
iii
challenge
this
simple
classification
because
they
sug-
gest
the
possibility
that
the products
of
some
segment-specific
selector
genes
may
themselves
have
regulatory
roles
in
initi-
ating
or
maintaining
the
selective
expression
of
other
segment-
specific
genes.
I
thank
P.
A.
Lawrence
for
excellent
advice
and
enthusiastic
en-
couragement
throughout
this
work,
C.
Nusslein-Volhard
for
generously
providingthe
Scr1
mutation,
P.
Johnston
for
technical
assistance,
and
several
of
my
colleagues in
the
Cell
Biology
Division
of
the
Laboratory
of
Molecular
Biology
for
critical
readings
of
manuscript.
I
also
thank
the
Thomas
C.
Usher
Fund
and
the
Medical
Research
Council
of
Great
Britain
for
financial
support
and
laboratory
facilities.
1.
Lewis,
E.
B.
(1963)
Am.
Zool.
3,
33-56.
2.
Lewis,
E.
B.
(1978)
Nature
(London)
276,
565-570.
3.
Tazima,
Y.
(1964)
The
Genetics
of
the
Silkworm
(Logos,
London).
4.
Garcia-Bellido,
A.
(1977)
Am.
Z7ool
17,
613-629.
5.
Denell,
R.
E.,
Hummels,
K.
R.,
Wakimoto,
B.
T.
&
Kaufman,
T.
C.
(1981)
Dev.
Biol
81,
43-50:
6.
Wakimoto,
B. T.
&
Kaufman,
T.
C.
(1981)
Dev.
Biol
81,
51-64.
7.
Struhl,
G.
(1981)
Nature
(London)
292,
635-638.
8.
Kaufman,
T.
C.,
Lewis,
R.
&
Wakimoto,
B.
T.
(1980)
Genetics
94,
115-133.
9.
Lewis,
R.
A.,
Kaufman,
T.
C.,
Denell,
R.
E.
&
Tellerico,
P.
(1980)
Genetics
95,
367-381.
10.
Lewis,
R.
A.,
Wakimoto,
B.
T.,
Denell;
R.
E.
&
Kaufman,
T.
C.
(1980)
Genetics
95,
383-397.
11.
Duncan,
I.
&
Lewis,
E.
B.
(1982)
in
Developmental
Order:
Its
Origin
and
Regulation,
ed.
Subtelny,
S.
(Liss,
New
York),
pp.
533-554.
12.
Struhl,
G.
(1981)
Nature
(London)
293,
36-41.
13.
Lindsley,
D.
L.
&
Grell,
E.
L.
(1968)
Carnegie
Inst.
Washington
PubI
627.
14.
Morata,
G.
&
Kerridge,
S.
(1980)
in
Development
and
Neuro-
biology
of
Drosophila,
eds.
Siddiqi,
O.,
Babu,
P.,
Half,
L.
M.
&
Hall,
J.
C.
(Plenum,
New
York),
pp.
141-154.
15.
Kerridge,
S.
&
Morata,
G.
(1982)
J.
Embryol
Exp.
Morphol
68,
211-234.
16.
Ferrus, A.
(1975)
Genetics
79,
589-599.
17:
Morata,
G.
&
Ripoll,
P.
(1975)
Dev.
Biol
42,
211-221.
18.
Garcia-Bellido,
A.,
Ripoll,
P.
&
Morata,
G,
(1973)
Nature
(Lon-
don)
New
Biol
245,
251-253.
19.
Morata,
G.
&
Garcia-Bellido,
A.
(1976)
Wilhelm
Roux's
Arch.
179,
124-143.
20.
Morata,
G.
&
Kerridge,
S.
(1981)
Nature
(London)
290,
778-781.
21.
Snodgrass,
R.
E.
(1938)
Smithson.
Misc.
Collect.
97,
No.
6.
22.
Tiegs,
0.
E.
&
Manton,
S.
M.
(1958)
BioL
Rev.
33,
255-337.
23.
Manton,
S.
M.
(1973)J.
Zool
171,
111-130.
24.
Garcia-Bellido,
A.
(1975)
in
Cell
Patterning,
ed.
Brenner,
S.
(Elsevier/North-Holland,
Amsterdam.),
pp.
161-182.
25.
Capdevila,
M.
P.
&
Garcia-Bellido,
A.
(1981)
Wilhelm
Roux's
Arch.
Dee.
Biol
190,
339-350.
Proc.
Nad
Acad.
Sci.
USA
79
(1982)
Downloaded by guest on February 19, 2021
<