Neuron, Vol. 17, 27–41, July, 1996, Copyright 1996 by Cell Press
Control of Daughter Cell Fates during Asymmetric
Division: Interaction of Numb and Notch
Ming Guo, Lily Yeh Jan, and Yuh Nung Jan
Howard Hughes Medical Institute
and Departments of Physiology and Biochemistry
University of California, San Francisco
San Francisco, California 94143-0724
atonal,arefirstexpressed inclustersof ectodermalcells
to endow these cells with the competence to adopt
neuronal fates (reviewed by Ghysen and Dambly-Chau-
diere, 1988; Campuzano and Modolell, 1992). One sen-
sory organ precursor (SOP) is then singled out from
each cluster through a process of “mutual inhibition”
mediated by “neurogenic genes” including Notch and
Delta (reviewed by Campos-Ortega, 1988; Artavanis-
Tsakonas and Simpson, 1991; Ghysen et al., 1993). To
form an es organ, the SOP gives rise to two distinct
daughter cells, IIa and IIb (see Figure 3A). The IIa cell
then divides to produce the hair cell (tricogen) and the
socket cell (tormogen), the outer support cells. Shortly
and the sheath cell (thecogen), the inner cells (Bate,
1978; Hartenstein and Posakony, 1989). A different se-
ries of asymmetric divisions allows the SOP of a cho
organ (see Figure 3D) to give rise to the neuron and
three different support cells (the sheath cell, the cap
and Bodmer, 1995).
Notchand Deltahave beenshown tomediate cell–cell
communication in eye development (Cagan and Ready,
1989), muscle development (Corbin et al.,1991), oogen-
esis (Ruoholaet al.,1991; Xuetal.,1992), and neurogen-
esis (Lehmann et al., 1983) in Drosophila. Homologs of
Notch and Delta have been characterized in various
species rangingfrom worm tohuman (reviewedby Arta-
vanis-Tsakonas et al., 1995). Notch encodes a trans-
membraneprotein with EGF-like repeatsin theextracel-
lular domain and tandem ankyrin repeats in the
intracellular domain (Wharton et al., 1985; Kidd et al.,
1986). The Notch product has been postulated to act
as a receptor, whereas the Delta gene product, another
transmembrane protein with EGF repeats (Vaessin et
al., 1987; Kopczynski et al., 1988), may function as a
ligand for the Notch receptor (Heitzler and Simpson,
1991; Rebay et al., 1991).
Cell–cell interaction is required not only for singling
out SOPs, butalso for theproper cell fate determination
of SOPprogeny,as indicated bystudiesof temperature-
sensitive alleles of Notch and Delta. Reduction of Notch
or Deltafunction duringSOP progenyformation inpupal
development causes an adult es organ (bristle) to con-
tain four neurons and no support cells (Hartenstein and
Posakony,1990; Parks and Muskavitch, 1993).A slightly
earlier temperature shift, presumably prior to IIa and IIb
division, leads to two neurons and two sheath cells in
Deltatsflies (Parks and Muskavitch, 1993). Thus, it was
postulated that Delta specifies four distinct progeny
cells in a stepwise fashion during the SOP division and
the IIb cell division (Parks and Muskavitch, 1993).
Besidescell–cell interaction,a cell-intrinsicfactor, the
numb gene product, is also essential for cell fate deter-
mination of SOP progeny. numb encodes a protein with
a motif called the phosphotyrosine-binding (PTB) do-
main (Kavanaugh and Williams, 1994) or phosphotyro-
sine interaction (PID) domain (Bork and Margolis, 1995).
Loss of numb function transforms the IIb cell into the
IIa cell during embryogenesis (Uemura et al., 1989), and
During development of the Drosophila peripheral ner-
vous system, a sensory organ precursor (SOP) cell
undergoesroundsofasymmetric divisionsto generate
four distinct cells of a sensory organ. Numb, a mem-
brane-associated protein, is asymmetrically segre-
gated into one daughter cell during SOP division and
acts as an inherited determinant of cell fate. Here, we
show that Notch, a transmembrane receptor mediat-
ing cell–cell communication, functions as a binary
switch in cell fate specification during asymmetric di-
esis. Moreover, numb negatively regulates Notch,
probably through direct protein–protein interaction
that requires the phosphotyrosine-binding (PTB) do-
main of Numb and either the RAM23 region or the
very C-terminal end of Notch. Notch then positively
regulates a transcription factor encoded by tramtrack
(ttk). This leads to Ttk expression in the daughter cell
that does not inherit Numb. Thus, the inherited deter-
minant Numb bestows a bias in the machinery for
cell–cell communication to allow the specification of
distinct daughter cell fates.
In asymmetricdivision, an important processinthe gen-
eration of cell diversity during development, a mother
cell produces two daughter cells that ultimately adopt
distinct fates. Two mechanisms may be responsible for
making the two daughter cells different. One, the cell-
intrinsic mechanism, involves an inherited determinant
that is asymmetrically segregated to one daughter cell
at the time of cell division. The other, the cell-extrinsic
mechanism,entailscommunication ofthedaughter cells
with each other or with surrounding cells (reviewed by
Horvitz and Herskowitz, 1992). In the Drosophila periph-
eral nervous system (PNS), both mechanisms are used
to generate the distinct cell types of a sensory organ
(Uemura et al., 1989; Hartenstein and Posakony, 1990;
Parks and Muskavitch, 1993; Rhyu et al., 1994). It thus
providesan opportunityto assesshow thesetwo mech-
ification during asymmetric divisions.
Two main types of sensory organs in the larval and
adult Drosophila PNS, the externalsense (es) organ and
chodotonal (cho) organ, are formed in a progressive
process during embryogenesis and pupal development
(Bodmer and Jan, 1987; Campos-Ortega and Harten-
stein, 1985; Ghysen et al.,1986). The “proneural genes,”
including those in the achaete–scute complex and
overexpression of numb results in reciprocal cell fate
transformation (Rhyuet al.,1994). Immunocytochemical
studies have shown that Numb is asymmetrically local-
ized to one pole of the SOP and segregated to one of
the daughter cells (Rhyu et al., 1994; Knoblich et al.,
1995). Thus, asymmetrically distributed Numb confers
distinctdaughtercell fates(Rhyuet al.,1994).Asymmet-
ric divisions of the SOP and its daughter cells that give
rise to the adult es organ (bristle) also depend on numb
function (Rhyu et al., 1994); the cell fate transformation
issimilartothephenotype causedbyreduction of Notch
function. Similarly, asymmetric division of the MP2 neu-
ral precursor in the central nervous system (CNS) de-
pends on asymmetric localization of Numb (Spana et
numb exerts its function at least in part via a down-
stream target gene called tramtrack (ttk) (Guo et al.,
1995). ttk encodes two zinc finger proteins due to alter-
native splicing (Harrison and Travers, 1990; Brown et
al., 1991; Read and Manley, 1992). The 69 kDa Ttk pro-
teinhasbeenshowntoact asatranscriptional repressor
intheprocessofsegmentation (Brown et al.,1991;Read
et al., 1992). Like numb, ttk acts as a genetic switch.
Loss of ttk function causes the SOP daughter cell IIa to
be transformed into IIb, a phenotype opposite to the
lossofnumb function phenotype.Overexpression of ttk,
on the other hand, results in the same cell fate transfor-
mation as does loss of numb function (Guo et al., 1995).
Genetic and immunocytochemical studies indicate that
ttk is negatively regulated by numb (Guo et al., 1995).
Several questions are raised by the dependence of
asymmetric division on a gene encoding a cell-intrinsic
determinant(Numb) as wellas genes such as Notch and
Delta, which mediatecell–cell interaction.How does the
information derived from cell-intrinsic signals interface
with signals arising from cell–cell interaction? How are
these different instructions to the two daughter cells
we firstshowthat Notchplays a critical role inasymmet-
ric divisions during embryogenesis. We then demon-
strate that genetically Notch is most likely negatively
regulated by numb, and biochemically Notch binds to
Numb. This direct protein–protein interaction requires
either the RAM23 region or the very C-terminal end of
Notch and the PTB domain of Numb. Finally, we have
identified ttk as a downstream target gene of Notch.
Takentogether, we proposethat Ttkactsasa readout to
integrate theinformationderived from both cell-intrinsic
mechanism and cell–cell communication; the inherited
determinant Numb sets up or biases the direction of
cell–cell interaction to allow the specification of distinct
daughter cell fates during asymmetric divisions.
Ghysen et al., 1993), the role of Notch in asymmetric
divisions of the SOP has only been examined in the
adult (Hartenstein and Posakony, 1990), but not in the
embryo.Hence,we characterized theembryonicpheno-
types due toreduction of Notch function or overexpres-
sion of Notch and showed that they corresponded to
reciprocal daughter cell fate transformations.
Elimination of Notch function throughout develop-
ment results in a significant overproduction of neuronal
precursors (Lehmann et al., 1983). The general disorga-
nization of the PNS in these null mutants, however,
makes it difficult to ascertain any cell fate alterations
within the lineage of a single sensory organ. To over-
come this difficulty, a temperature-sensitive allele of
Notch (Notchts) was used. Embryos of Notchtswere first
raised at the permissive temperature (18?C) to allow the
normal development of the embryo and then shifted to
the nonpermissive temperature (30?C) at certain devel-
opmental stages to disrupt the function of Notch (see
SOPs are singled out and then divide asymmetrically
during 4–7 hr of embryogenesis at 25?C (Bodmer et al.,
1989). Shifting Notchtsembryos to the nonpermissive
temperature at stages corresponding to 4–5 hr of em-
bryogenesis at 25?C resulted in significant overproduc-
tion of neuronsin mostembryos,as revealedbystaining
with a neuronal marker MAb22C10 (Zipursky et al.,
1984).Usinganti-Cutantibody,which recognizesall four
cells of the es organ (Blochlinger et al., 1990), we exam-
ined thesite of emergence of normally a single es organ
and found that the number of Cut-positive cells in-
creased by a factor of two to three (data not shown).
Thus, most likely supernumerary SOPs emerged from
the disruption of Notch function at this early stage due
to failure of singling out of SOPs.
WhenNotchtsembryoswere shiftedtothe nonpermis-
sive temperature at stages corresponding to 5–7 hr of
embryogenesis at25?C, we observed phenotypesof cell
fate transformation within bothes and cho lineages. We
focused our studies in the dorsal-most region of an
abdominal hemisegment, which normally harbors two
es organs, and doubly stained the embryos with anti-
Cut antibody and MAb22C10. In wild type, there are
eight Cut-positive cells and two MAb22C10-positive
neurons. Among younger embryos in a Notchtsembryo
collection, which would have their Notch function dis-
rupted at an earlier stage, some had ten to twelve Cut-
positivecells; mostof thesecellsexpressed MAb22C10.
The rest of the embryos had a complement of eight
Cut-positive cells, though up to six of these cells also
expressed MAb22C10. It thus appears that even after
the singling out of the SOPs, the number of neurons
within anesorganincreaseswith thedisruptionof Notch
function. Thiscould beinterpretedasa cellfatetransfor-
mation during asymmetric divisions of SOP, analogous
to what hasbeen observed in the adult bristle formation
(Hartenstein and Posakony, 1990).
To test this further, we doubly stained the Notchts
embryos with MAb22C10, a neuronal marker, and anti-
Prospero antibody, which labels the nuclei of sheath
cells in both es and cho organs (Vaessin et al., 1991).
We focused our studies on the older embryos in the
collection that would have their Notch function dis-
rupted at a later stage. In the wild-type embryo, the two
Reduction of Notch Function during SOP
Division Causes Support Cells to Be
Transformed into Neurons
Although there have been extensive studies of Notch
function in the process of lateral inhibition to single out
a neuronal precursor from a proneural cluster(Campos-
Ortega, 1988; Artavanis-Tsakonas and Simpson, 1991;
Interaction of Notch with numb
Figure 1. Reduction
Transforms Sheath Cells to Neurons
(A–D) Two simple es organs at the dorsal-
mostregion ofanabdominal segmentinwild-
type embryo. (A) A schematic drawing of the
two isolated es organs, with each circle rep-
resenting anindividualcell. Green circles cor-
respond to the MAb22C10 staining that rec-
ognizes the cytoplasm of neurons shown in
(B) and (D), whereas red dots correspond to
the anti-Prospero staining that labels the nu-
clei of sheath cells shown in (C) and (D). In
wild type, two neurons (B) (two arrows in [D])
and two sheath cells (C) can be observed
(bracket in [D]).
(E–G) In Notchts, four neurons ([E] and four
arows in [G]) can be detected without the
associated sheath cells ([F] and bracket
(H) Lateral region of an abdominal segment
where five cho organs are aligned side by
side. In this region, there are also other cells,
one v? cho organ and three es organs. Simi-
larly to (A), green circles and red dots corre-
spond to the MAb22C10 and anti-Prospero
staining respectively. she, sheath cell; neu,
neuron; lig, ligament cell; cap, cap cell.
(I–K) In wild type, five cho neurons (I and K)
have five sheath cells associated with them
(J and K).
(L–N) In Notchts, the number of MAb22C10-
positive neurons increases (L and N), while
the number of anti-Prospero positive sheath
cells decreases (M and N). Moreover, the
lent to that of the sheath cell reduction (N).
isolated dorsal-most es organs have two MAb22C10-
positive neurons and two Prospero-expressing sheath
cells (Figures 1A–1D). In contrast, the Notchtsembryos
contained four neurons without associated sheath cells
at the corresponding location (Figures 1E–1G), indicat-
ing that the sheath cells are transformed into neurons
(see Figure 3B).As expected from such a cell fatetrans-
formation, staining with anti-Cut antibody revealed four
stronglystaining cells(hairs or sockets)and fourweakly
staining cells(neurons or sheathcells) (data not shown).
neurons toProspero-positive sheath cellsin cho organs
wasobserved (Figures1L–1N). This cellfatetransforma-
tion from support cells to neurons in both es and cho
organs further substantiates the role of Notch in cell
fate specification during asymmetric divisions. These
embryonic phenotypes were then used to investigate
the epistatic genetic relationships between Notch and
ttk, as well as between Notch and numb (see below).
to that due to loss of Notch function during the process
of singling out of neuronal precursors, it is suggested
that the Notch intracellular domain corresponds to a
constitutively active form of the Notch receptor (Lieber
et al., 1993; Rebay et al., 1993; Struhl et al., 1993).
We used theUAS–GAL4 system(Brand and Perrimon,
1993) and studied the effect of overexpression of the
constitutively active form of Notch (NotchACT) (Doherty
et al., 1996) that contains a truncated Notch (with the
extracellular domain deleted but with the transmem-
ing SOP divisions. Transgenic flies carrying UAS–
NotchACTweremated to transgenicflies carrying a GAL4
enhancer–trap line, 1407 GAL4, which expresses the
reportergene inall neurons of the PNS (Luo etal., 1994).
The embryos carrying UAS–NotchACTas wellas the 1407
GAL4 line would then express NotchACTprimarily in neu-
rons. At the dorsal-most region of abdominal segments
of these embryos,anti-Cut staining showed theexpres-
sion pattern of four strongly stained and four weakly
stained cells in approximately 90% of the segments
(data not shown). Double labeling of both MAb22C10
and anti-Prospero, however, revealed a cell fate trans-
formation inat least one of these two es organsin about
30%–40%of theabdominal segments examined.A typi-
cal example is shown in Figures 2A–2C, in which the es
organ on the top contains two anti-Prospero-positive
cells but no neurons, suggestive of a transformation
Overexpression of Activated Notch Results
in the Reverse Cell Fate Transformation
When a truncated form of Notch (intracellular domain)
is ubiquitously overexpressed during the stage when
neuronalprecursors are singled out, there isa reduction
of neuronal precursors (Lieber et al., 1993; Rebay et al.,
1993; Struhl et al., 1993; Lyman and Yedvobnick, 1995;
Bang et al., 1995).Since these phenotypes are opposite
Figure 2. Overexpression of NotchACTTrans-
forms Neurons to Sheath Cells
(A–C) Of the two dorsal-most es organs, the
top es organ (the bracket on the left in [C])
has two anti-Prospero positive sheath cells
(B), without the associated MAb22C10-posi-
tive neuron (A). (C) The lower es organ (the
bracket on theright in [C]) isnot transformed.
(D–F) In the lateral region of an abdominal
segment, the number of anti-Prospero posi-
tive sheath cells increases (E and F), while
the number of MAb22C10-positive neurons
decreases (D and F). Moreover, the overpro-
duction of sheath cells matches in number
with the reduction of neurons (F).
from neurons to sheath cells (Figure 3C). The lower es
organ, on the other hand, appears normal. In the cho
organs at the lateral region, we also observed an in-
crease in the number of Prospero-positive sheath cells
anda concurrentdecrease inthenumber ofMAb22C10-
positiveneurons (Figures2D–2F),indicative ofa cell fate
transformation from neurons to sheath cells (Figure 3F).
In this region, at least three of the five cho organs in
most segments showed transformation. Thus, overex-
pression of NotchACTresults in a cell fate transformation
oppositeto that caused by reduction of Notch function.
cell into the socket cell. The four sockets most likely
arise from a transformation of the IIb cell into the IIa
cell, followed with a transformation of the hair cell into
the socket cell (Figure 4E). Overexpression of NotchWT
using the same GAL4 enhancer–trap line caused about
20%of thebristlesto formdoublesockets (Figure4C) or
four sockets (Figure 4D). Thus, NotchACToverexpression
as compared with those due to overexpression of the
wild-type Notch product.
Sensitive Genetic Interaction between
numb and Notch
The effects of reducing Notch function on adult bristle
formation (Hartenstein and Posakony, 1990) are similar
to those due to overexpression of numb (Rhyu et al.,
1994). To determine whether Notch and numb function
in the same genetic pathway, we first tested to see
whether there is a synergistic enhancement of the adult
bristle phenotypes due toboth reduction of Notch func-
tion and numb overexpression.
Whereas heat shock during larval development
causes flies carrying one or two copies of the hs-numb
transgene to exhibit double hairs, as a result of trans-
formation of the socket cell into the hair cell (Rhyu et
al., 1994), without heat shock these flies did not have
any bristle phenotype (Figure 5B). Flies carrying only
one copy of the functional Notch gene, Notch55e11/?,
also showed no obvious hair or socket abnormalities
(Figure 5C). By contrast, even in the absence of any
heat shocktreatment, Notch55e11/?;hs-numb/? flieshad
double hairswithoutsockets at theanterior wingmargin
(Figure 5D). This phenotype was further enhanced in
Notch55e11/? flies that carried two copies of hs-numb
(Figure 5E). In addition to finding double hairs with no
sockets in place of a normal bristle (Figure 5F), we also
observed double hairs right next to a normal looking
bristle (Figure 5G) and two sets of double hairs right
Overexpression of the Wild-Type Notch Results
in Cell Fate Transformation Similar to That
Caused by Activated Notch
If overexpressing the truncated Notch, NotchACT, indeed
reveals the activity of the Notch receptor upon activa-
tion, it should produce phenotypes qualitatively similar
to those due to overexpression of the wild-type Notch
(NotchWT). It has been reported that overexpression of
wild-type Notch via the heat shock promoter does not
result in any significant phenotypes (Lieber et al., 1993;
Rebay et al., 1993; Fortini et al., 1993). We found that
Notch overexpression due to UAS–NotchWTin combina-
tion with various GAL4 enhancer–trap lines resulted in
a very mild phenotype of transformation of neurons to
sheath cells in the embryo (data not shown), but much
stronger phenotype in adult es organs (bristles).
Overexpression of NotchACTin the 109-68 GAL4 en-
hancer–trap line, which expresses GAL4 in all four cells
of an adult es organ, caused 90% of the bristles in
the notum to form double sockets (Figure 4A) or, less
frequently, triple sockets or four sockets (Figure 4B).
due to overexpression of the intracellular domain of
Notch (Lieber et al., 1993; Rebay et al., 1993; Struhl et
al., 1993; Lyman and Yedvobnick, 1995; Bang et al.,
1995) and is indicative of a transformation of the hair
Interaction of Notch with numb
Figure 3. Wild-Type and Proposed SOP Cell Lineages in the Embryo due to Reduction of Notch Function or NotchACTOverexpression
(A–C) Cell lineages for the es organ in wild type (A), Notchts(B), and NotchACToverexpression (C). Within each lineage, we have boxed two
daughters whose fates are indicated to be affected by Notch. In the Notchts(B), the sheath cell is transformed into the neuron, yielding two
neurons. Following NotchACToverexpression (C), the neuron is transformed into the sheath cell.
(D–F) Cell lineages for the cho organ in wild type (D), Notchts(E), and NotchACToverexpression (F). We have modified drawings of the cho
lineage according to Brewster and Bodmer (1995). Lig, the ligament cell; ect, the ectodermal cell. In a cho organ, two neurons are produced
at the expense of the sheath cell in the Notchts(E). Following NotchACToverexpression (F), two sheath cells are formed coincident with the
loss of the neuron. Both reduction of Notch function and overexpression of NotchACTappear to disrupt the normally asymmetric division,
producing two identical daughter cells. Only the effects of reduction of Notch function and NotchACToverexpression on the formation and
differentiationof theIIblineage inthe esorgan andof thechIIIlineage inthecho organ aredepicted inthisfigure;similar cell fatetransformations
in the IIa lineage and between IIa and IIb have been observed but are not indicated here.
next to each other (Figure 5H). Two independent Notch
alleles and two independent insertion lines of hs-numb
weretested, and similar synergistic enhancement of the
bristlephenotype wasobserved (data not shown). Thus,
itis unlikely that the observed phenotypes could be due
to the site of insertion of the hs-numb transgene or
therefore represent a strong synergistic effect of reduc-
elevating Numb expression due to the basal activity of
hs-numb without heat shock; neither alone generated
any detectable phenotype. This suggests, though does
not prove, that numb and Notch function in the same
organs. Compared with the wild type (Figures 6A and
6B), loss of numb function reduced the number of neu-
rons to less than 10% of that in the wild type (Uemura
et al., 1989), as revealed by the neuronal marker
MAb22C10 (Figures 6C and 6D). Reduction of Notch
function, on theother hand, caused a neuronal overpro-
duction as described earlier (Figures 6E and 6F). The
double mutants of numb and Notchtsshowed a range
of phenotypes; 10% of the embryos demonstrated a
neuronal overproduction approaching the Notchtsphe-
notype (Figures 6G and 6H). The rest of the embryos
did not exhibit as extensive an overproduction, but the
number of neurons was much larger than that in numb
null mutants (Figures 6I and 6J). Since a reduction of
Notch function partially suppressed the numb null phe-
notypes, Notch might act downstream of numb and ap-
pears to be negatively regulated by numb. We then ex-
amined Numb expression when Notch function is
disrupted. The asymmetric Numb localization was still
present in at least some of the dividing precursor cells
in both CNS and PNS of the Notch55e11null mutant em-
downstream of numb.
Epistatic Relationship between Notch and numb
To investigate the epistatic relationship between numb
and Notch, we examined whether reduction of Notch
function is capable of suppressing the phenotypes re-
sulting from loss of numb function. Embryos homozy-
gous for the null mutation numb1and Notchtswere
shifted to the nonpermissive temperature at the time of
cell fate specification within the lineage of the sensory
Figure 4. Overexpression of NotchACTand NotchWTResults in Similar Phenotypes of Multiple Sockets
(A) Double sockets phenotypes of both macrochaetes and microchaetes on the notum due to NotchACToverexpression.
(B) Four sockets observed in a NotchACT-expressing fly.
(C) Double sockets due to overexpression of NotchWTobserved at a lower frequency.
(D) Four sockets were also seen in NotchWToverexpression.
(E) The proposed lineage for the four socket phenotype observed in both NotchACToverexpression and NotchWToverexpression. There is a
cell fate transformation first from the IIb cell to the IIa cell and then from the hair cell to the socket cell.
Direct Protein–Protein Interaction between
Notch and Numb
Since we detected a very strong synergistic interaction
in transheterozygous flies carrying Notch55e11/?; hs-
numb/? (without heat shock) (see Figure 5), we won-
dered whetherthere could be directphysical interaction
between these two proteins.We assessedthis possibil-
ity first using the yeast two-hybrid interaction assay
(Bartel et al., 1993). Full-length as well as fragments of
Numb or Notch were fused with either the LexA DNA-
binding domain or the GAL4 transcriptional activation
domain.If Numb and Notch physically interact, cotrans-
formation of these two fusion proteins in yeast cells
would result in the reconstitution of the “hybrid” tran-
scriptionfactorthatcan beassessedby the?-galactosi-
dase activity. As described below, this assay revealed
interaction between Notch and Numb. These two pro-
teins were then subdivided into fragments to localize
the regions involved in this protein–protein interaction.
To identify the portions of the Notch protein involved
in the interaction with Numb, we subdivided the Notch
intracellular domain into three regions,A, M, and P (Fig-
ure7),and assayed theirinteractionswith Numb individ-
ually. Two regions, A and P, interacted with Numb,
whereas the M region did not (Figure 7A and Table 1).
This interaction appears to be specific, since neither
Numb nor Notch-A bound to the control protein Lamin.
Moreover, Suppressor of Hairless (Su(H)), a known
Notch-binding protein (Tamura et al., 1995; Hsieh et al.,
1996),bound toNotch-A butnot Notch-P(Table 1).Both
regions A and P display a high degree of amino acid
conservation among vertebrate Notch homologs (Wein-
master et al., 1991, 1992; Bierkamp and Campos-
Ortega, 1993) and include several functional motifs. Re-
thePESTregion; ithasbeensuggestedtoplay animpor-
tant role in Notch functioning (Xu et al., 1990; Lieber et
al., 1993). Region A includes RAM23, a Su(H)-binding
region, and ankyrin repeats, a Deltex-binding region
(Diederich etal., 1994; Matsunoet al.,1995). Since Su(H)
and possibly Deltex areinvolved in thecell fate decision
in the es lineage (Ashburner, 1982; Schweisguth and
Posakony, 1994; Matsuno et al., 1995), we wondered
whether Numb binds to RAM23 and/or ankyrin repeats.
As shown in Table 1 and Figure 7A, Numb bound to
RAM23, but not ankyrin repeats of Notch.
Two linesof evidence implicate theN-terminalportion
butnot theC-terminalportion ofNumb inNotchbinding.
As described later, a glutathione S-transferase (GST)
fusion protein of Notch bound to Numb-N but not to
Numb-C (Figure 7C). In the yeast two-hybrid assay,
Notch-A also showed interaction with Numb-N but not
Numb-C (Figure 7B and Table 1). Consistent with the
finding, Notch-A also bound to Numb-P, which includes
Numb-N and is conserved between Drosophila Numb
and mouse Numb (Zhong et al., 1996 [this issue of Neu-
ron]; Figure 7B and Table 1). We then tested whether
the PTB domain alone, part of Numb-N, is sufficient to
confer the binding to Notch-A. Indeed, we detected an
interaction as shown in Figure 7B and Table 1.
To confirm the physical interaction between Notch
and Numb, we used the in vitro binding assay as an
independent test. The Notch intracellular domain was
fused to GST, and this fusion protein was expressed in
bacteria. The GST–Notch fusion protein was immobi-
lized on glutathione–Sepharose beads and then mixed
with in vitro translated35S-labeled Numb-N or Numb-C.
As shown in Figure 7C, Numb-N, but not Numb-C, was
retainedwith GST–Notch onbeads.Neither Numb-Nnor
Numb-C was retained with GST protein alone on the
cellular domain to bind Numb-N in vitro as well as in
yeast two-hybrid assay suggests that direct protein–
Interaction of Notch with numb
Figure 5. SynergisticInteraction between Reductionof NotchFunc-
tion and Slight Elevation of Numb Expression
The bristles in anterior wing margin of wild type (A), hs-numb (B),
Notch55e11/? (C), Notch55e11/?; hs-numb/? (D), and Notch55e11/?; hs-
numb (E–H). The stout bristles are in focus. Bristles in hs-numb (B)
and Notch55e11/? (C) are essentially wild type in appearance (A). In
Notch55e11/?; hs-numb/? (D), there is a phenotype of “double hairs”
(arrowhead) at low frequency. In Notch55e11/?; hs-numb, there is a
high frequencyof abnormal bristles (E), including “double hairs with
no socket” (arrowhead in [F]), “double hairs next to a bristle with a
normallooking bristle” (arrowheadin [G]), as well as “four hairs next
to one another” (arrowhead in [H]).
Figure 6. Epistasis of numb and Notch: Reduction of Notch Func-
tion Suppresses the Neuronal Phenotype of numb Null Allele
MAb22C10stainingofawild-typeembryo(A andB),a numb1embryo
(C andD), a Notchtsembryo (E andF), and thedoublemutant numb1;
Notchtsembryos (G–J). Whole-mount embryos are in(A), (C), (E), (G),
and (I), while the dorsal cluster (thin bracket) and lateral cluster
(thick bracket) of two abdominal segments are shown at higher
magnification in (B), (D), (F), (H), and (J). Compared with the wild
type (A and B), the numb1embryo shows a severe reduction in the
number of neurons (C and D), whereas the Notchtsembryo exhibits
a significant increase in the number of neurons (E and F). About
10% of the double mutant numb1; Notchtsembryos such as the one
in (G) and (H) show neuronal overproduction almost to the same
extent as Notchts(E and F), whereas the majority of the double
mutants such as the one in (I) and (J) shows the “intermediate
phenotype,” the number of neurons is more than that in numb1
embryo (C and D), but fewer than that in Notchtsembryo (E and F).
protein interaction mayprovide a molecularmechanism
for the negative regulation of Notch by numb.
ttk Acts Downstream of Notch
Since Notch and ttk are both downstream of numb and
cause similar phenotypes in the cell fate specification
in the es organ lineage, we asked whether Notch and
ttk act in the same genetic pathway and, if so, whether
Notch acts upstream of ttk.
We first examined whether there is any alteration of
theTtkexpression due toeither reductionofNotch func-
tion or overexpression of NotchACT. We focused our in-
vestigation on the IIb lineage of the es organ, because
Ttk is expressed in the sheath cell that also expresses
Prospero, but not in the other daughter cell of IIb, the
body (Figures 8D–8F). Thus, Notch function is required
not only to specify the sheath cell but also to express
Ttk in this daughter cell of an asymmetric division.
In a reciprocal experiment, we examined Ttk expres-
sion in embryos with NotchACToverexpression (Figures
8G–8L). In these embryos, we found that often one es
organ in the dorsal-most region showed transformation
(Figures 8G–8I). Anti-Prospero antibody labeled two
sheath cells in this es organ, including one that arose
from a transformation of the neuron into a sheath cell
(Figure8H). Thesuperimposition ofanti-Ttk staining and
anti-Prospero staining reveals that Ttk is expressed in
all the sheath cells, including those transformed from
neurons (Figure 8I). Similarly, in the cho organs, Ttk is
expressed in all the sheath cells, including those that
are transformed from neurons (Figures 8J–8L). Thus,
activated Notch is able toturn on Ttkexpression incells
that normally do not express Ttk. The up- and down-
regulation of Ttk expression due to overexpression of
activated Notch and reduction of Notch function, re-
spectively, indicates that ttk is very likely downstream
of Notch and is positively regulated by Notch.
Totest thishypothesis further,we askedwhethercon-
stitutively active Notch requiresttk inmediating cell fate
transformation. We therefore overexpressed NotchACTin
the ttk loss-of-function mutant. Compared with the wild
type (see Figure 6A), loss of ttk resulted in an overpro-
duction of neurons (Figure 9B), whereas NotchACTex-
pression resulted in a reduction of neurons (Figure 9A).
We found that overexpression of NotchACTin ttk mutant
background resulted in an overproduction of neurons
(Figure 9C). The extent of this neuronal overproduction
is indistinguishable from that due to loss of ttk function;
without ttk function, NotchACTfails to transform neurons
into sheath cells. Therefore, the capability of NotchACT
to transform neurons into sheath cells depends on nor-
mal ttk function, suggesting that ttk acts downstream
Figure 7. Direct Protein–Protein Interaction between Numb and
(A) and (B) depict the region of Numb andNotch used in two-hybrid
(A) Notch intracellular domain. The RAM23 region, the ankyrin re-
peats (ANK), the OPA repeats, and the PEST region are indicated.
The fusionproteincontainingregions A, P, or RAM23 binds to Numb
(indicated by a plus sign), whereas region M and ANK fail to bind
(indicated by a minus sign) (see Table 1).
(B) Numb protein. The conserved region between Drosophila Numb
and mouse Numb is illustrated as a striped area and includes the
PTB domain. Regions P, N, and PTB, but not C, bind to Notch-A
(see Table 1).
(C) In vitro binding of [35S]Numb-N but not [35S]Numb-C to GST–
with the control GST protein (lane 1), or GST–Notch fusion protein
(lane 2), as well as glutathione–Sepharose beads. Glutathione–
Sepharose beadswereused to isolateGST-containingprotein com-
plex, which werethen analyzedby SDS–PAGE and autoradiography
(top, [35S]Numb-N; bottom, [35S]Numb-C). An aliquot of [35S]Numb-N
or [35S]Numb-C was run on the same gel (lane 3) to indicate the size
of the in vitro translated Numb protein fragments.
Notch Specifies Distinct Daughter Cell Fates
during Multiple Asymmetric Divisions
in the Formation of Es and Cho Organs
Notch has been shown to play an important role during
theprocess of singling out of SOPs fromproneural clus-
ters, a cell fate decision between neuronal fate and epi-
dermal fate. In this paper, by studying the effect of both
reduction of Notch function and overexpression of
NotchACTand NotchWT, we show that Notch serves as a
binary switch between two daughter cells during SOP
divisions in both the es organ and the cho organ.
Notch appears to be important in specifying sheath
cell fate in the IIb cell lineage of es organs and the chIII
cell lineage of cho organs. Reduction of Notch function
transforms sheath cells to neurons, whereas overex-
pression of NotchACTresults in the opposite cell fate
transformation in the embryo. Moreover, Notch is also
involved in the choice between the hair and the socket
cellfates. Inadult esorgans, reduction ofNotchfunction
in Notchtsmutants results in twinning (two hairs without
socket) (data not shown), whereas overexpression of
neuron that expresses anti-Elav (Guo et al., 1995; Ra-
maekersetal., submitted).Figures8A–8F show thedou-
ble labeling of Notchtsembryos using anti-Ttk and anti-
Elav antibodies. In the dorsal-most region that normally
contains two es organs and hence two neurons, we
detected four Elav-expressing neurons (Figure 8B). As
shown earlier, thetwo extra neurons arose froma trans-
formationofthesheathcellsinto neurons.We foundthat
Ttk was expressed in most of the cells in the epidermis
of the Notchtsembryo (Figure 8A), but not in neurons,
including the supernumerary neurons derived from
transformationof sheathcells(Figure8C).Similar obser-
vations were made in the lateral cho organs. None of
theneurons,including those sheathcellsthat had trans-
formed into neurons, were recognized by anti-Ttk anti-
Interaction of Notch with numb
Table 1. Notch and Numb Interact in the Yeast Two-Hybrid Assay
Activation Domain Fusion
3. Vector alone
8. Vector alone
10. Vector alone
12. Vector alone
20. Vector alone
25. Vector alone
31. Vector alone
Plasmids containing fusion protein with LexA–DNA-binding domain and plasmids containing fusion protein with GAL4 activation domain were
cotransformed into L40 yeast strain containing the lacZ reporter gene under the control of GAL4 elements. The ?-galactosidase activity was
monitored semiquantitatively by filter lift assay: double plus, turning blue in 1.5 hr; plus, turning blue in 1.5–6 hr; minus/plus, some or all
colonies turning blue in 6–24 hr; minus, remaining white over 24 hr.
seems to promote the socket cell fate in the IIa lineage
and the sheath cell fate in the IIb lineage.
Notch also appears to be involved in the choice be-
tween the IIa and IIb cell fates during SOP divisions.
Reduction of Notch function during formation of the
adult es organs results in the formation of four neurons
at the expense of three support cells (Hartenstein and
These phenotypes most likely are produced by first a
switch of cell fate between IIa and IIb and then a switch
of cell fate between hair and socket or between neuron
and sheath. In other words, Notch function is required
firsttospecify IIa and then to specifysocket (a daughter
of IIa) and sheath cell (a daughter of IIb). Reducing or
removing Notch function results in two IIb cells and
their altered divisions leading to the production of four
neurons, whereas elevating Notch activity leads to two
IIa cells, which then divide abnormally and give rise to
four sockets. Thus, Notch appears to be required for all
three asymmetric divisions of the es organ and for at
least the final division in the cho organ lineage. The
asymmetry is disrupted by either reduction of Notch
function or Notch overexpression, resulting in symmet-
The Interaction between Notch and numb
Like Notch, numb is involved in multiple asymmetric
divisions during the formation of a sensory organ (Rhyu
et al., 1994). There are several possible scenarios con-
cerning interactions between numb and Notch. Notch
could act upstream of numb, regulating the asymmetric
localization of Numb in the SOP, the segregation of
Numb to one of the daughter cells, or Numb activity.
Alternatively, Numb, once segregated into one of the
daughter cells, could bias subsequent cell–cell interac-
that numb and Notch function in parallel to ensure the
proper cell fate specification. The possibility that Notch
acts entirely upstream of numb has been ruled out, be-
cause a reduction of Notch function in double mutants
of Notchtsand numb null mutation partially suppressed
the numb mutant phenotype. Since these double mu-
tants eliminate numb function and yet some of the em-
bryos show overproduction of neurons as do Notchts
embryos,the action ofNotch does not appear torequire
numb gene function. Rather, manifestation of the numb
mutantphenotype dependsonNotch geneactivity, indi-
cating that numb is upstream of Notch and negatively
regulates Notch. In this scenario, loss of numb function
may leadto anincrease of Notch activityinthe daughter
cell that normally inherits Numb. In the double mutant
Figure 9. Loss of ttk Function Prevents NotchACTfrom Transforming
Neurons to Sheath Cells
MAb22C10 staining of an embryo expressing NotchACT(A), a ttkosn
embryo (B), and a ttkosnembryo expressing NotchACT(C). Compared
with the wild type (Figure 6A), there is a reduction in the number of
sensory neurons in NotchACTexpressing embryos (A), whereas loss
of ttk function causes an overproduction of sensory neurons (B).
The ttkosnembryo that expresses NotchACT(C) shows the neuronal
overproduction phenotype indistinguishable from that due to ttk
loss of function (B).
Figure 8. Ttk Expression Is Altered in Opposing Ways in Notchts
Mutants and Flies Expressing NotchACT
(A–C) Double labeling of Notchtsembryos using anti-Ttk 88 kDa (A),
anti-Elav(B) antibodies, andthe overlay of these two images (C) for
two dorsal-most es organs. In contrast with the presence of two
neurons in wild type, we detected four Elav-expressing neurons
(bracketed in [B]). As shown earlier, the two extra neurons arose
from a transformation of the sheath cells intoneurons. Although Ttk
is expressed in most of the cells in the epidermis of the Notchts
embryo(A),it isnotexpressedinanyneurons, includingthesupernu-
merary neurons derived from transformation of sheath cells (C).
(D–F) The double labeling of Notchtsembryos using anti-Ttk 88 kDa
(D), anti-Elav (E) antibodies, and the overlay of these two images
(F) in five lateraland one v? cho organs.Similarly, none of the super-
numerary neurons (E), including those sheath cells that had trans-
(G–I)Double labelingofembryos with NotchACToverexpression using
anti-Ttk 88 kDa (G), anti-Prospero (H) antibodies, and the overlay
of these two images (I) for the two dorsal-most es organs. Often,
one es organ shows transformation. The arrow in (H) points to two
sheath cells, including one that is transformed from a neuron. (Top
es organ is not transformed.) Ttk is expressed in all sheath cells (G
and I), including one that is transformed from a neuron.
(J–L)Double labelingofembryoswith NotchACToverexpression using
anti-Ttk 88 kDa (J), anti-Prospero (K) antibodies, and the overlay of
these two images (L) for the lateral and v? cho organs. Similarly, all
sheath cells (K), including those that are transformed from neurons,
express Ttk (J and L).
inthedouble mutant. Consistentwith thepossibilitythat
numbis upstreamof Notch,Numbis stillasymmetrically
localized in Notch null alleles. We also detected a very
strong synergistic enhancement of the effects due to a
partial loss of Notch function and a slight elevation of
numb expression. While these findings suggest that
Notch acts downstream of numb, as proposed pre-
viously by Posakony (1994), Rhyu (1994), and Jan and
Jan (1995), they do not exclude the possibility of more
complex and elaborate schemes of action of Notch and
numb. For example, numb could also be involved in
an additional pathway parallel to the Notch signaling
pathway. Models that incorporate our findings are illus-
trated inFigure 10. Numb,depicted inred, isasymmetri-
cally distributed to one pole of the SOP during SOP
divisionand segregatedinto oneof thedaughtercellsso
as to bias or set up the direction of cell–cell interaction,
causing that cell to adopt the cell fate of IIb. This may
be achieved in at least two ways, as indicated by the
following two models. In model one (left), cell–cell com-
munication occurs between the two daughter cells.
Numb may suppress the Notch (in brown) activity di-
rectly, possibly via direct protein–protein interactions
between Notch and Numb. Numb may also suppress
Notch activity indirectly by regulating Delta (in green)
that carries the null mutation numb1and the hypomor-
phic allele Notchts, the residual Notch activity is likely
to be higher than that in the Notchtssingle mutant. This
could account for the milder Notch mutant phenotype
Interaction of Notch with numb
Figure 10. Models for the Interaction among
numb, ttk, and Notch
Each circle represents an individual cell.
Whereas circles with gray shading represent
epidermal cells surrounding the SOP and its
daughter cells, circles with yellow shading
illustrate the SOP and its daughter cells. The
membrane-associated Numb (nb), depicted
in red, is segregated into one daughter cell
soas tobias or setupthedirectionofcell–cell
interaction. This may be achieved in at least
two ways, as indicated by the following two
models. In model one (left), cell–cell commu-
nication occurs between the two daughter
cells. WhereastheNotch activity isillustrated
in orange, with the reduced activity shown as
the light orange, the Delta activity is demon-
strated in green, with the reduced activity in
light green. Numb may suppress the Notch
activity directly, or indirectly by regulating
Delta activity, which renders the Numb-con-
taining daughter cell less effective in receiv-
ing signals andmore effective insending sig-
nals to its sister cell. In model two (right),
the cell–cellsignaling mediatedby Notch and
Delta occurs between SOP daughter cells
and the surrounding epidermal cells which
express Delta. The suppression of Notch by numb results in a difference in Notch activity between two daughter cells. These two models are
not mutually exclusive. In either model, the Notch activity in the cell without Numb would be sufficiently high to cause Ttk expression (in blue),
whereas the Notch activity in the other daughter cell is suppressed by Numb to such an extent that it can no longer induce Ttk expression.
activity. Suppression of Notch activity by Numb renders
the Numb-containing daughter cell less effective in re-
ceiving signals and more effective in sending signals to
its sister cell, so that it adopts the cell fate of IIb and
its sister adopts the cell fate of IIa. In model two (right),
the cell–cell signaling mediated by Notch and Delta oc-
curs between SOP daughter cells and the surrounding
epidermal cells that express Delta. The segregation of
Numbinto one of thedaughter cellscauses Notch activ-
ity in that cell to be suppressed. The daughter cell that
lacks Numb has higher Notch activity and becomes IIa,
whereas the daughter cell that has inherited Numb has
lower Notch activity, thereby assuming the default cell
fateof IIb. These two models arenot mutually exclusive.
In either model, the Notch activity in the IIa cell would
be sufficiently high to cause Ttk expression (in blue),
whereastheNotch activityin IIbissuppressed byNumb
to such an extent that it can no longer induce Ttk ex-
that Numb may suppress Notch activity through direct
protein–protein interaction. One can envision a variety
ofscenarios as tohowthis maytakeplace. For instance,
the binding of Numb to Notch may prevent Notch from
being activated. Alternatively, Numb could inhibit the
coupling of activated Notch with its downstream ef-
fectors, such as Su(H), orprevent the activation of Su(H)
by Notch. Two regions of Notch show interactions with
in different species. The RAM23 region of Notch physi-
cally interacts with Su(H) (Tamura et al., 1995; Hsieh et
al., 1996) as well as Numb, whereas Notch-P binds to
Numb but not Su(H). Recently, Dishevelled has been
shown to interact with Notch C-terminus (Axelrod et al.,
1996). It will be interesting to see whether Dishevelled
binds to Notch-P, as does Numb.
We have identified the PTB domain of Numb as the
Notch-binding domain. PTB domain in other proteins
including Shcbindstophosphotyrosine inproteinssuch
as EGF receptor and NGF receptor and is involved in
signaling through tyrosine phosphorylation (Kavanaugh
and Williams, 1994; Kavanaugh et al., 1995; reviewed
by vanderGeer andPawson, 1995;Laminetet al.,1996).
This raises the possibility that signaling from Numb to
Notch may be regulated by tyrosine phosphorylation.
The PTB domain between Drosophila Numb and mouse
Numb shares 63.3% amino acid identity (Zhong et al.,
1996). We note also that similar protein–protein interac-
tion has been observed between mouse Numb and
mouseNotch1 (Zhonget al.,1996), suggestingan evolu-
tionarily conserved mechanism.
ttk Is Regulated by Both Notch and numb
We have previously demonstrated that ttk is involved in
multiple asymmetricdivisionsin bothesand cho organs
(Guo etal., 1995).Transformation ofneurons into sheath
cells is caused either by ttk overexpression or by
NotchACToverexpression in the presence of normal ttk
gene function. In the absence of ttk function, NotchACT
overexpression fails to transform neurons into sheath
cells. This suggests a requirement of ttk for the sheath
cell fate, though loss of zygotic ttk function does not
cause sheath cells to be transformed into neurons (Guo
et al., 1995), presumably due to the presence of redun-
dant pathways. The dependence of NotchACTaction on
ttk also indicates that ttk is a downstream target of
Notch. Indeed, whereas Ttk is normally expressed in
only one of the IIb daughters, reduction of Notch func-
tion apparently prevented ttk expression in either
daughter cell, and overexpression of NotchACTleads to
abnormal ttkexpression inbothdaughtercells. Previous
studies have shown that ttk is negatively regulated by
For examining the numb localization, 4.5–6.5 hr embryos of
Notch55e11/FM6flies were doublystainedusing bothanti-Asense and
anti-Numb antibodies. Notch mutant embryos were identified as
those embryos that have overproduction of neurons in the CNS.
numb(Guo etal.,1995).Thisisconsistentwith themodel
shown in Figure 10 in which ttk is positively regulated
byNotch, which isin turnnegatively regulated by numb.
We propose that Ttk, as a transcription factor, acts as
a readout to integrate signals derived both from a cell-
intrinsic determinant and from cell–cell communication.
Besides ttk, other downstream targets of Notch have
been identified, such as Su(H) and Enhancer of split
(E(spl)). Su(H) and Notch exhibit direct protein–protein
interaction in the yeast two-hybrid assay (Fortini and
Artavanis-Tsakonas, 1994; Tamura et al., 1995; Hsieh et
al., 1996). In tissue culture cells, the binding of Delta
and Notch allows the translocation of Su(H) from the
nas, 1994). Su(H) then acts as a transcriptional activator
to regulate directly E(spl) (Lecourtois and Schweiguth,
1995; Baileyand Posakony, 1995), which isa gene com-
plex of seven related genes that encode basic helix-
loop-helix proteins characteristic of transcriptional reg-
ulators (Delidakis and Artavanis-Tsakonas, 1992; Knust
et al., 1992). It is interesting to note that the overexpres-
sion of NotchACTor Su(H) results in phenotypes that are
slightly different from the phenotype due to the overex-
pression of E(spl). Whereas overexpression of NotchACT
and Su(H) results in predominantly double sockets
(Schweisguth and Posakony, 1994), overexpression of
the m5 and m8 genes in the E(spl) complex results in
duplicated bristles, double sockets as well as other ab-
errant outer support cells (Tata and Hartley, 1995). The
phenotype due to overexpression of ttk is similar to that
of overexpressionof E(spl) (Guo et al.,1995; Ramaekers
et al., submitted). It would be interesting to examine
the interaction between ttk, Su(H), and E(spl) and to
investigate how Notch signaling eventually leads to the
cell fate decision.
Temperature Shift of NotchtsEmbryos
wNotchtsstock was kept at 18?C. The developmental time in 18?C
was estimated to be twice the amount of time at 25?C (Giniger et
al., 1993). Embryos collected between 8–14 hr at 18?C were put into
either a 30?C incubator, or a water bath at 32?C. Embryos were kept
at these nonpermissive temperatures for 5 hr. After the temperature
shift, embryos were placed in an 18?C incubator for 0.5–4 hr prior
Immunofluorescence and Confocal Microscopy
Embryos were fixed and stained as described by Guo et al. (1995).
A Zeiss microscope and a Bio-Rad MRC-600 confocal were used
to view and obtain the images. All the antibodies used in this study
have been described previously: MAb22C10 (Zipursky et al., 1984),
rat or rabbit anti-Cut antibody (Blochlinger et al., 1990), anti-Pros-
pero antibody (Vaessin et al., 1991), anti-Asense antibody (Brand et
al., 1993), guinea pig anti-Ttk 88 kDa antibodies (Read et al., 1992),
anti-Elav (Mab44C11) (Bier et al., 1988), and anti-Numb antibody
(Rhyu et al., 1994).
and mounted in Hoyer’s solution (Ashburner, 1989).
Yeast Two-Hybrid Interaction Assay
Fragments of the Notch cDNA were inserted into either the pBHA
vector (obtained from C. T. Chien), pGAD424 vector, or pGADGH
vector (obtained from Clontech). Region A includes the area from
the SalI site to the PstI site (amino acids 1792–2269), or the area
from the beginning of the intracellular domain to the EcoRI site
(amino acids 1766–2263). Region M spans the area from the EcoRI
site (amino acid 1767) to the SalI site (amino acid 2612). Region P
includes the C-terminal fragment from amino acid 2602 to the end
of the protein. RAM23 construct contains amino acids 1766–1897,
whereas the ANK construct aminoacids 1895–2109. The aminoacid
coordinates shown here are as in Wharton et al. (1985). The numb
full-length cDNA was inserted into either the pGAD vector or the
pBHA vector (generated by C. T. Chien). Numb-P starts from the
beginningof theproteintothePstI site(aminoacids 1–330).Numb-N
spans amino acids 1–223 (the BamHI site), whereas Numb-C spans
amino acids 224 to the end of the protein. pBHA Numb-N, pBHA
Numb-C, and pBHA Numb-P were obtained from C. T. Chien and
S. W. Wang. Full-length Su(H) was cloned into pGAD424 (generated
by S. W. Wang).
The yeast transformation was performed according to Bartel et
al. (1993). The ?-galactosidase enzymatic activity was evaluated
by filter-lifting assay. Nitrocellulose filters were used to lift up the
colonies, andthen frozenin liquid nitrogen. Subsequently, thefilters
were overlaid on the Whatman paper presoaked with 5–45 ?l of
X-Gal in3 mlof Z buffer.Coloniesproducing ?-galactosidaseturned
blue with time. For each set of experiments, a single colony from
thetransformantscarrying eitherthetwofusionproteinsto betested
or proper controls (Figure 7B) was then streaked to the same plate
andassayed againto ensure theequivalence ofexperimental condi-
Genetics and Drosophila Strains
Drosophila strains were raised on standard cornmeal-yeast agar
medium at room temperature or 25?C except those with Notchts
For overexpressing the NotchACTin ttk mutant background, males
with a genotype of yw; 1407 GAL4; ttkosn/TM6, Ubx weremated with
females with a genotype of yw; UAS–NotchACTttkosn/TM3, Sb. The
embryos werecollected and stainedwith markers. All the ttkmutant
should also have NotchACTexpression driven by 1407 GAL4. The ttk
mutant embryos canbe unambiguously identifiedas thoseembryos
that have defects in head involution and dorsal closure (Guo et al.,
For examining the dosage-sensitive interaction between
Notch55e11/? and hs-numb, waNotch55e11/FM6 females were crossed
to w males or yw, hs-numb (line 2.4C) males, respectively. Female
progenywith Bar?eyeshape, thus witha genotypeof waNotch55e11/?
or waNotch55e11/?;hs-numb/?, werecollected, andtheir phenotypes
wereexamined. For examining the phenotype of the waNotch55e11/?;
hs-numb flies, waNotch55e11/?; hs-numb/CyO stock was first con-
structed. waNotch55e11/?; hs-numb flies were unambiguously identi-
fied as those with wing notching (due to the haploinsufficiency of
theNotch55e11allele) andstraight wings.All crosses describedabove
were set up in parallel. The same experiment was performed using
homozygotes of ndGrell2, an independent viable Notch allele, andone
or two copies of hs-numb (line 11.1) located on the third chromo-
some. Similar synergistic enhancement was observed (see text).
For constructing the double mutants between numb and Notch,
we constructed flies with a genotype of wNotchts, numb1pr cn Bc/
CyO,P[T8–lacZ]. Embryosfrom thisstockwerecollected anddouble
serum. Double mutants were identified as those embryos that show
no anti-Numb epidermal staining nor anti-?-galactosidase staining.
GST Fusion Proteins and In Vitro Protein Binding Assays
The whole region of Notch intracellular domain (1766–2703) was
fused in frame with GST in GST-4T vector (Pharmacia). For protein
expression, overnight bacterial culture was diluted and grown until
OD at260nm reached 1.0.After addingIPTG, theculture was grown
for another 3–7 hr. The bacteria were then harvested and lysed by
sonication in PBS in the presence of protease inhibitors (aprotinin,
leupeptin and pepstatin). Commassie bluestaining showed the pro-
duction of the protein whose size corresponds to the predicted
fusion between GST and Notch intracellular domain.
35S-labeled Numb was synthesized in vitro from pNAC-Numb-N
or pNAC-Numb-C (generated by C. P. Shen from the same regions
of Numb used in two-hybrid assays) using the TNT-coupled rabbit
reticulocyte lysate system (Promega); 10 ?l of this lysate containing
35S-labeledNumb-N or Numb-C wasmixed withGST fusionproteins
Interaction of Notch with numb
immobilized on the glutathione–Sepharose 4B beads (Pharmacia),
and the mixture was incubated at room temperature for 1–2 hr. The
beads were then washed with cold 0.05% Triton X-100 in PBS and
collected by centrifugation. Torelease the boundprotein,the beads
were boiled in gel sample buffer and
Bodmer, R., and Jan, Y.N. (1987). Morphological differentiation of
the embryonic peripheral neurons in Drosophila. Roux’s Arch. Dev.
Biol. 196, 69–77.
Bodmer, R., Carretto, R., and Jan, Y.N. (1989). Neurogenesis of the
peripheral nervous system in Drosophila melanogaster embryos:
DNA replication patterns and cell lineages. Neuron 3, 21–32.
Bork, P., and Margolis, B. (1995). A phosphotyrosine interaction
domain. Cell 80, 693–694.
Brand, A.H., and Perrimon, N. (1993). Targeted gene expression as
a means of altering cell fates and generating dominant phenotypes.
Development 118, 401–415.
Brand, M., Jarman, A.P., Jan, L.Y., and Jan, Y.N. (1993). asense is
a Drosophilaneuralprecursor gene andiscapableof initiatingsense
organ formation. Development 119, 1–17.
Brewster, R.,andBodmer, R. (1995). Originandspecification of type
II sensory neurons in Drosophila. Development 121, 2923–2936.
Brown, J.L., Sonoda, S., Ueda, H., Scott, M., and Wu, C. (1991).
Repression of the Drosophila fushi tarazu (ftz) segmentation gene.
EMBO J. 10, 665–674.
Cagan,R.L.,andReady,D.F. (1989).Notchis requiredforsuccessive
cell decisions in the developing Drosophila retina. Genes Dev. 3,
Campos-Ortega, J.A.(1988).Cellular interactionsduringearlyneuro-
genesis in Drosophila melanogaster. Trends Neurosci. 11, 400–405.
Campos-Ortega, J.A., and Hartenstein, V. (1985). The Embryonic
Development of Drosophila melanogaster (New York: Springer-
Campuzano,S., andModolell, J.(1992).Patterning oftheDrosophila
nervous system: the achaete–scute gene complex. Trends Genet.
Corbin, V., Michelson, A.M., Abmayr, S.M., Neel, V., Alcamo, E.,
Maniatis, T., and Young, M.W. (1991). A role for the Drosophila
neurogenic genes in mesoderm differentiation. Cell 67, 311–323.
Delidakis, C., and Artavanis-Tsakonas, S. (1992). The Enhancer of
split [E(spl)] locus of Drosophila encodes seven independent helix-
loop-helix proteins. Proc. Natl. Acad. Sci. USA 89, 8731–8735.
Diederich, R.J., Matsuno, K., Hing, H., and Artavanis-Tsakonas, S.
(1994). Cytosolic interaction between deltex and Notch ankyrin re-
peats implicates deltex in the Notch signaling pathway. Develop-
ment 120, 473–481.
Doherty, D., Feger, G., Younger-Shepherd, S., Jan, L.Y., and Jan,
Y.N. (1996). Delta is a ventral to dorsal signal complementary to
Serrate, another Notch ligand, in Drosophila wing formation. Genes
Dev. 10, 421–434.
Fortini, M.E., and Artavanis-Tsakonas, S. (1994). The Suppressor of
Hairless protein participates in Notch receptor signaling. Cell 79,
Fortini, M.E., Rebay, I., Caron, L.A., and Artavanis-Tsakonas, S.
(1993). An activated Notch receptor blocks cell fate commitment in
the developing Drosophila eye. Nature 365, 555–557.
Ghysen, A., and Dambly-Chaudiere, C. (1988). From DNA to form:
the achaete–scute complex. Genes Dev. 7, 723–733.
Ghysen, A., Dambly-Chaudiere, C., Aceves, E., Jan, L.Y., and Jan,
Y.N. (1986).Sensory neurons andperipheral pathways inDrosophila
embryos. Roux’s Arch. Dev. Biol. 195, 281–289.
Ghysen, A., Dambly-Chaudiere, C., Jan, L.Y., and Jan, Y.N. (1993).
Cell interactions and gene interactions in peripheral neurogenesis.
Genes Dev. 7, 723–733.
Giniger, E., Jan, L.Y., and Jan, Y.N. (1993). Specifying the path of
the intersegmental nerve of the Drosophila embryo: a role for Delta
and Notch. Development 117, 431–440.
Guo, M., Bier, E., Jan, L.Y., and Jan, Y.N. (1995). tramtrack acts
downstream of numb to specify distinct daughter cell fates during
asymmetric cell divisions in the Drosophila PNS. Neuron 14,
Harrison, S., and Travers, A. (1990). The tramtrack gene encodes a
Drosophila finger protein that interacts with the ftz transcriptional
regulatory region andshows a novel embryonic expression pattern.
EMBO J. 9, 207–216.
analyzed by SDS–
Wethank E. Giniger for constructing andinjectingthe UAS–NotchACT
construct, M. Rothenberg for the technical assistance with two-
hybrid system, and C. T. Chien, S. W. Wang, and C. P. Shen for
various DNA constructs. We express special thanks to L. Seugnet
and M. Haenlin in the lab of P. Simpson for providing us with the
UAS–NotchWTflies, E. Spana and C. Doe for communicating their
results prior to publication, Artavanis-Tsakonas’s lab for pMtNMg
(Notch) cDNA, F. Schweisguth for Su(H) cDNA, and D. Read for
providing anti-Ttk antibodies. We are very grateful to W. M. Zhong,
E. Frise, and S. Barbel for assistance with the artwork of figures
and C. Bargmann, W. M. Zhong, B. Hay, D. Doherty, E. Frise, and
C.P. Shen forhelpfuldiscussions andcomments onthemanuscript.
Supported by a National Institute of Mental Health grant to the
Silvio Conte Center for Neuroscience Research at the University of
California, San Francisco. L. Y. J. and Y. N. J. are Investigators of
the Howard Hughes Medical Institute.
The costs of publication of this article were defrayed in part by
the payment of page charges. This article must therefore be hereby
marked “advertisement” in accordance with 18 USC Section 1734
solely to indicate this fact.
Received April 8, 1996; revised June 19, 1996.
Artavanis-Tsakonas, S., and Simpson, P. (1991). Choosing a cell
fate: a view from the Notch locus. Trends Genet. 7, 403–408.
Artavanis-Tsakonas,S., Matsuno, K., andFortini, M.E. (1995). Notch
signaling. Science 268, 225–232.
Ashburner, M. (1982). The genetics of a small autosomal region of
Drosophila melanogaster containing the structural gene for alcohol
dehydrogenaseIII: hypomorphic mutations affectingthe expression
of Hairless. Genetics 101, 447–459.
Ashburner, M. (1989). Drosophila: A Laboratory Handbook (Cold
Spring Harbor, New York: Cold Spring Harbor Laboratory Press).
Axelrod,J.D., Matsuno,K.,Artavanis-Tsakonas,S., andPerrimon,N.
(1996). Interaction between wingless and Notch signaling pathways
mediated by Dishevelled. Science 271, 1826–1832.
Bailey, A., and Posakony, J. (1995). Suppressor of Hairless directly
activates transcription of Enhancer of split complex genes in re-
sponse to Notch receptor activity. Genes Dev. 9, 2609–2622.
Bang, A., Bailey, A., and Posakony, J. (1995). Hairless promotes
stablecommitmentto thesensoryorgan precursor cell fatebynega-
tively regulates the activity of the Notch signaling pathway. Dev.
Biol. 172, 479–494.
Bartel,P., Chien, C.T., Sternglanz,R.,andFields,S. (1993).Using the
two-hybridsystem to detect protein–protein interactions. In Cellular
Interactions in Development: A Practical Approach, D.A. Hartley,
ed. (Oxford: Oxford University Press), pp. 153–179.
Bate, C.M. (1978). Development of sensory systems in arthropods.
In Handbook of Sensory Physiology, M. Jacobson, ed. (New York:
Springer-Verlag), pp. 1–53.
Bier, E., Ackerman, L., Barbel, S., Jan, L.Y., and Jan, Y.N. (1988).
Identificationandcharacterization of aneuron-specific nuclearanti-
gen in Drosophila. Science 240, 913–916.
Bierkamp, C., and Campos-Ortega, J.A. (1993). A zebrafish homo-
logue of the Drosophila neurogenic gene Notch and its pattern of
transcription during early embryogenesis. Mech. Dev. 43, 87–100.
Blochlinger,K., Bodmer, R., Jan, L.Y., andJan, Y.N. (1990). Patterns
of expression of Cut, a protein required for external sensory organ
development in wild-type and cut mutant Drosophila embryos.
Genes Dev. 4, 1322–1331.
Hartenstein, V., and Posakony, J.W. (1989). Development of adult
sensilla on the wing and notum of Drosophila melanogaster. Devel-
opment 107, 389–405.
Hartenstein, V., and Posakony, J.W. (1990). A dual function of the
Notch gene in Drosophila sensillum development. Dev. Biol. 142,
Heitzler, P., and Simpson, P. (1991). The choice of cell fate in the
epidermis of Drosophila. Cell 64, 1083–1092.
Horvitz, H.R., and Herskowitz, I. (1992). Mechanisms of asymmetric
cell division: two Bs or not two Bs, that is the question. Cell 68,
Hsieh, J.J.-D., Henkel, T., Salmon, P., Robey, E., Peterson, M.G.,
and Hayward, S.D. (1996). Truncated mammalian Notch1 activates
CBF1/RBPJk-repressed genes by a mechanism resembling that of
Epstein–Barr virus EBNA2. Mol. Cell. Biol. 16, 952–959.
Jan, Y.N., and Jan L. Y. (1995). Maggot’s hair and bug’s eye: role
of cell interactions and intrinsic factors in cell fate specification.
Neuron 14, 1–5.
Kavanaugh, W.M., and Williams, L.T. (1994). An alternative to SH2
domains forbinding tyrosine-phosphorylatedproteins.Science266,
Kavanaugh, W.M., Turck, C.W., and Williams, L.T. (1995). PTB do-
main binding to signaling proteins through a sequence motif con-
taining phosphotyrosine. Science 268 , 1177–1179.
Kidd, S., Kelley, M.R., and Young, M.W. (1986). Sequence of the
Notch locus of Drosophila melanogaster: relationship of the en-
coded protein to mammalian clotting and growth factors. Mol. Cell.
Biol. 6, 3094–3108.
Knoblich, J., Jan, L.Y., and Jan, Y.N. (1995). Localization of Numb
andProspero via a novel mechanism for asymmetric protein segre-
gation. Nature 377, 624–627.
Knust, E., Schrons, H., Grawe, F., and Campos-Ortega, J.A. (1992).
Sevengenes oftheEnhancer ofsplit complex ofDrosophila melano-
gaster encode helix-loop-helix proteins. Genetics 132, 505–518.
Kopczynski, C.C., Alton, A.K., Fechtel, K., Kooh, P.J., and Muska-
vitch, M.A.T. (1988). Delta, a Drosophila neurogenic gene, is tran-
scriptionallycomplex andencodesaproteinrelated toblood coagu-
lation factors and epidermal growth factor of vertebrates. Genes
Dev. 2, 1723–1735.
Laminet, A.A., Apell, G., Conroy, L., and Kavanaugh, W.M. (1996).
Affinity, specificity, and kinetics of the interaction of the SHC phos-
photyrosine binding domain with asparagine-X-X-phosphotyrosine
motifs of growth factor receptors. J. Biol. Chem. 271, 264–269.
Lecourtois, M., andSchweiguth,F. (1995). The neurogenic Suppres-
sor of Hairless DNA-binding protein mediates the transcriptional
activationoftheEnhancer ofsplit complex genestriggeredby Notch
signaling. Genes Dev. 9, 2598–2608.
Lehmann, R., Jimenez, F., Dietrich, U., and Campos-Ortega, J.A.
(1983). On thephenotype and development of mutants of early neu-
rogenesis in Drosophila melanogaster. Roux’s Arch. Dev. Biol. 192,
Lieber, T., Kidd, S., Alcamo, E., Corbin, V., and Young, M.W. (1993).
Antineurogenic phenotypes induced by truncated Notch proteins
indicate a role in signal transduction and may point to a novel func-
tion for Notch in nuclei. Genes Dev. 7, 1949–1965.
Luo,L., Liao,Y.J.,Jan, L.Y.,andJan, Y.N. (1994).Distinct morphoge-
netic functions of similar small GTPases: Drosophila Drac1 is in-
volved in axonal outgrowth and myoblast fusion. Genes Dev. 8,
Lyman, D. and Yedvobnick, B. (1995). Drosophila Notch receptor
activity suppresses Hairless function during adult external sensory
organ development. Genetics 141, 1491–1505.
Matsuno, K., Diederich, R.J., Go, M.J., Blaumueller, C.M., and Arta-
vanis-Tsakonas, S. (1995). Deltex acts as a positive regulator of
Notchsignaling throughinteractions with theNotchankyrin repeats.
Development 121, 2633–2644.
Parks, A., and Muskavitch, M.A.T. (1993). Delta function is required
for bristle organ determination and morphogenesis in Drosophila.
Dev. Biol. 157, 484–496.
Posakony, J.W. (1994). Nature versus nurture: asymmetric cell divi-
sions in Drosophila bristle development. Cell 76, 415–418.
Read, D., and Manley, J. (1992). Alternatively spliced transcripts of
the Drosophila tramtrack gene encode zinc finger proteins with
distinct DNA binding specificities. EMBO J. 11, 1035–1044.
Read, D., Levine, M., and Manley, J. (1992). Ectopic expression of
theDrosophilatramtrackgeneresults inmultiple embryonicdefects,
including repression of even-skipped and fushi tarazu. Mech. Dev.
Rebay, I., Fleming, R.J., Fehon, R.G., Cherbas, L., Cherbas, P., and
Artavanis-Tsakonas, S. (1991). Specific EGF repeats of Notch medi-
ate interactions with Delta and Serrate: implications for Notch as a
multifunctional receptor. Cell 67, 687–699.
Rebay, I., Fehon, R.G., and Artavanis-Tsakonas, S. (1993). Specific
truncations of Drosophila Notch define dominant activated and
dominant negative forms of the receptor. Cell 74, 319–329.
Rhyu, M. (1994). How sister cells acquire different fates: the role of
numb in asymmetric divisions. PhD thesis, University of California,
San Francisco, California.
Rhyu, M., Jan, L., and Jan, Y.N. (1994). Asymmetric distribution of
numb proteinduring divisionof thesensory organprecursor confers
distinct fates to daughter cells. Cell 76, 477–491.
Ruohola, H., Bremer, K.A., Baker, D., Swedlow, J.R., Jan, L.Y., and
Jan, Y.N.(1991).Role ofneurogenicgenes inestablishmentoffollicle
cell fate and oocyte polarity during oogenesis in Drosophila. Cell
Schweisguth, F., and Posakony, J.W. (1994). Antagonistic activities
of Suppressor-of-Hairless and Hairless control alternative cell fates
in the Drosophila adult epidermis. Development 120, 1433–1441.
Spana, E., Kopczynski, C., Goodman, C., and Doe, C. (1995). Asym-
metriclocalizationofnumb autonomouslydetermines siblingneuron
identity in the Drosophila CNS. Development 121, 3187–3195.
Struhl, G.,Fitzgerald,K.,andGreenwald,I.(1993). Intrinsicactivity of
the lin-12 and Notch intracellular domains in vivo. Cell 74, 331–345.
Tamura,K.,Taniguchi,Y., Minoguchi,S., Saka,T., Tun,T., Furukawa,
T., andHonjo, T.(1995).Physical interactionbetweena noveldomain
of the receptor Notch and the transcription factor RBP-Jk/ Su(H).
Curr. Biol. 5, 1416–1423.
Tata, F., and Hartley, D.A. (1995). Inhibition of cell fate in Drosophila
by Enhancer of split genes. Mech. Dev. 51, 305–315.
Uemura, T., Shepherd, S., Ackerman, L., Jan, L.Y., and Jan, Y.N.
(1989). numb, a gene required in determination of cell fate during
sensory organ formation in Drosophila embryos. Cell 58, 349–360.
Wharton, K.A., Johansen, K.M., Xu, T., and Artavanis-Tsakonas, S.
(1985). Nucleotide sequence from the neurogenic locus Notch im-
plies a gene product that shares homology with proteins containing
EGF-like repeats. Cell 43, 567–581.
Vaessin, H., Bremer, K., Knust, E., and Campos-Ortega, J.A. (1987).
The neurogenicgene DeltaofDrosophilamelanogaster isexpressed
in neurogenic territories and encodes a putative transmembrane
protein with EGF-like repeats. EMBO J. 6, 3431–3440.
Vaessin, H., Grell, E., Wolff, E., Bier, E., Jan, L.Y., and Jan., Y.N.
(1991). prospero is expressed in neuronal precursors and encodes
a nuclear protein that is involved in the control of axonal outgrowth
in Drosophila. Cell 67, 941–953.
van der Geer, P., and Pawson, T. (1995). The PTB domain: a new
protein module implicated in signal transduction. Trends Biochem.
Sci. 20, 277–280.
Weinmaster, G., Roberts, V.J., and Lemke, G. (1991). A homolog
of Drosophila Notch expressed during mammalian development.
Development 113, 199–205.
Weinmaster, G., Roberts, V.J., and Lemke, G. (1992). Notch2: a
second mammalian Notch gene. Development 116, 931–941.
Xu, T.,Rebay, I.,Fleming,R.J., Scottgale,T.N.,andArtavanis-Tsako-
nas, S. (1990). The Notch locus and the genetic circuitry involved
in early Drosophila neurogenesis. Genes Dev. 4, 464–475.
Xu, T., Caron, L.A., Fehon, R.G., and Artavanis-Tsakonas, S. (1992).
Interaction of Notch with numb
The involvement of theNotch locus in Drosophila oogenesis. Devel-
opment 115, 913–922.
Zhong, W.M., Feder, J.N., Jiang, M.M., Jan, L.Y., and Jan, Y.N.
(1996). Asymmetric localization of a mammalian Numb homolog
during mouse cortical neurogenesis. Neuron 17, this issue.
Zipursky,S.L., Venkatesh, T.R., Teplow, D.B., andBenzer, S. (1984).
ies as molecular probes. Cell 36, 15–26.