Oncogene (2001) 20, 8299 ± 8307
ã 2001 Nature Publishing Group All rights reserved 0950 ± 9232/01 $15.00
www.nature.com/onc
Emc, a negative HLH regulator with multiple functions in
Drosophila development
Sonsoles Campuzano*,1
1
Centro de BiologõÂa Molecular Severo Ochoa, Cantoblanco, 28049 Madrid, Spain
Expression and functional analyses of Emc have
demonstrated that it is a prototype for a protein required
for multiple processes in development. Initially characterized as a negative regulator of sensory organ
development, it was later found to regulate many other
developmental processes and cell proliferation. Its ability
to block the function of bHLH proteins by forming
heterodimers, which are ine€ective in DNA binding,
accounts for the role of Emc in preventing the acquisition
of several cell fates which are under the control of bHLH
proteins. However, while maintaining this repressive
molecular mechanism, emc also appears to act as a
positive regulator of di€erentiation. Oncogene (2001)
20, 8299 ± 8307
Keywords: Drosophila; extramacrochaetae; HLH; transcription factor; development
Identi®cation of extramacrochaetae
According to the usual nomenclature system of
Drosophila genes, the extramacrochaetae (emc) gene
was named after the ®rst discovered phenotype
associated to its loss-of-function mutations, i.e., the
development of extra sensory organs (SOs) on the
adult ¯y. The larval and adult epidermis of Drosophila
contains thousands of SOs that are distributed
according to a stereotyped pattern. Each SO is formed
by the progeny of a single cell, the sensory mother cell
(SMC), which, during the third larval instar or early
pupal stages, is singled out from the population of
imaginal discs cells or abdominal histoblasts (Bate,
1978; Bodmer et al., 1989; GarcõÂ a-Bellido and
Merriam, 1971; Hartenstein and Posakony, 1989).
Genetic analysis of loss and gain-of-function mutations
in the achaete-scute complex (AS-C) indicated that it
was involved in the commitment of epidermal cells to
di€erentiate as SOs and that the AS-C might be under
some type of negative regulation (GarcõÂ a-Bellido and
SantamarõÂ a 1978; GarcõÂ a-Bellido, 1979). Back in the
early eighties, a genetic screen was performed to
identify genes that might negatively regulate the ASC (Botas et al., 1982). The rationale of the screening
was that in the presence of extra doses of the AS-C,
*Correspondence: S Campuzano; E-mail: [email protected]
which by their own do not alter the pattern of SOs, a
recessive mutation in a gene encoding a repressor of
AS-C would become dominant and promote the
generation of extra SOs. That is, only one copy of
that gene would not produce enough protein to repress
several copies of AS-C and allow a normal pattern of
SOs. Mutations that ful®lled this criterion were found
in two loci: hairy (h) and the novel emc (Botas et al.,
1982). It was further observed that loss of emc
function, in a wild type background for the AS-C,
caused development of extra SOs, whereas AS-C
de®ciencies suppressed this phenotype (Moscoso del
Prado and GarcõÂ a-Bellido, 1984a,b). Taken together,
these results suggested that the generation of extra SOs
in emc mutants depended on the activity of the AS-C
and not on the modi®cation of a parallel developmental pathway. The molecular analysis of AS-C and
emc con®rmed this hypothesis and provided a
molecular basis for the negative regulation of AS-C
by emc.
Emc titrates away proneural proteins
The achaete (ac) and scute (sc) genes encode proteins
which contain the basic ± helix ± loop ± helix domain
(bHLH) characteristic of a family of transcriptional
regulators (Murre et al., 1989a; Villares and Cabrera,
1987; see Massari and Murre, 2000, for a recent
review). Biochemical analysis of mammalian bHLH
proteins involved in myogenesis, the MyoD family, and
of immunoglobulin enhancer binding bHLH proteins
E12 and E47, showed that the HLH domain allows
these proteins to form homodimers or heterodimers
with other members of the family, while the basic
region is required for binding to speci®c DNA
sequences (Murre et al., 1989b). ac and sc are
expressed at the right place and time to promote
SMC determination, as anticipated by the genetic
analyses. Both genes are coexpressed in the wing
imaginal discs, the anlagen of most of the thorax and
the wings, in clusters of cells located at precisely
determined positions (Cubas et al., 1991; Romani et
al., 1989; Skeath and Carroll, 1991) (Figure 1a). SMCs
are born within these clusters of ac ± sc expressing cells
(Cubas et al., 1991; Skeath and Carroll, 1991) (Figure
1a). Moreover, removal of speci®c proneural clusters in
ac or sc mutants removes the corresponding SMCs and
Multiples requirements for Emc in development
S Campuzano
8300
Figure 1 Relationship between the pattern of expression of sc (a) and emc (b, c) in the imaginal wing disc and the pattern of
SMCs. (a) Imaginal wing disc from an enhancer trap line that expresses lacZ (green) in the SMCs. The distribution of sc mRNA,
revealed by in situ hybridization, is shown in purple. Note that SMCs are born within the clusters of sc expressing cells. (b) Pattern
of expression of emc. It is fairly complementary to that of sc. Note minima of emc expression at the regions where the SMCs for the
DC and SC bristles develop. (c) Expression of emc (purple) and distribution of SMCs (green) in a Hairy wing imaginal wing disc.
(Only the prospective wing region is shown). Although sc is ubiquitously expressed in Hairy wing discs (Balcells et al., 1988),
expression of emc is not modi®ed (compare with b). Note the development of ectopic SMCs at the posterior wing margin (PWM)
that arise within a region of minimal emc expression. Positions of some SMCs is indicated, namely those that will give rise to the
anterior notopleural bristle (ANP); anterior postalar bristle (APA); anterior wing margin bristles (AWM); dorsocentral bristles
(DC); giant sensilum of the radius (GSR); sensilla of the dorsal radius (Rd); scutellar bristles (SC) and sensilla of the tegula (Td).
PWB, posterior wing blade; v3, vein 3
SOs (Cubas et al., 1991; Romani et al., 1989; Skeath
and Carroll, 1991) and a generalized distribution of
Ac ± Sc products (in Hairy wing mutants) promotes
Oncogene
development of extra SOs in ectopic positions (Balcells
et al., 1988; Skeath and Carroll, 1991). Since AS-C
proteins confer on cells the ability to become neural
Multiples requirements for Emc in development
S Campuzano
precursors, the AS-C genes have been called proneural
genes and the clusters of cells expressing them are
referred to as the proneural clusters (Ghysen and
Dambly-ChaudieÁre, 1989; Romani et al., 1989).
Another bHLH protein Daughterless (Da), previously
identi®ed by its role in sex determination (Cline, 1989)
acts synergistically with Ac/Sc to build the SO pattern
(Caudy et al., 1988a,b; Dambly-ChaudieÁre et al., 1988).
HLH proteins have been classi®ed according to tissue
distribution, dimerization capabilities and DNA-binding site speci®city (Massari and Murre, 2000). Due to
their restricted pattern of expression, Ac and Sc belong
to HLH class II. On the contrary, Da belongs to class I
HLH proteins, which show a generalized distribution
(Massari and Murre, 2000). Usually, class II HLH
proteins activate transcription in the form of heterodimers with class I HLH proteins (Massari and Murre,
2000). Thus, the molecular structure of As ± Sc and Da
as well as their genetic interactions (Caudy et al.,
1988a,b; Dambly-ChaudieÁre et al., 1988; Villares and
Cabrera, 1987) indicated that they are transcriptional
regulators that, in combination, control the activity of
genes involved in SMC commitment. In agreement
with this hypothesis, AS-C proteins and Da can form
heterodimers that bind to speci®c DNA sites in vitro
(Cabrera et al., 1987; Murre et al., 1989b).
emc also encodes a transcription factor of the HLH
family, but the Emc protein lacks the basic region that
is involved in the interaction with DNA (Ellis et al.,
1990; Garrell and Modolell, 1990) and thus, it belongs
to class V of HLH proteins (Massari and Murre, 2000).
Heterodimers formed by a bHLH protein and an HLH
protein devoid of the basic region are unable to bind to
DNA (Davis et al., 1990; Voronova and Baltimore,
1990). Since it was already demonstrated that proneural proteins speci®cally bind to DNA in the form of
heterodimers (Murre et al., 1989b), a plausible
hypothesis for the regulatory interactions between
Emc and AS-C was put forward (Figure 2a). Emc
would antagonize proneural ac ± sc function by sequestering proneural bHLH proteins in complexes incapable of e€ective interaction with DNA, rather than by
a direct repression of AS-C transcription (Ellis et al.,
1990; Garrell and Modolell, 1990, reviewed in Garrell
and Campuzano, 1991). This mechanism would
account for the dose dependent interactions between
emc, as, sc and da, which suggested that a ®nely tuned
balance of proneural proteins and of Emc was required
for proper SMC determination. It also explained how
the gain-of-function emc mutation Achaetous (Ach)
caused loss of SOs (GarcõÂ a-Alonso and GarcõÂ a-Bellido,
1988): in Ach imaginal discs there is a higher than
normal accumulation of emc mRNA and, presumably,
of Emc protein (Cubas and Modolell, 1992). This
mechanism for EMC action was further supported by
the molecular analysis of mouse myogenesis. Emc
shows a high degree of similarity to the murine family
of Inhibitor of Di€erentiation (Id) proteins (Benezra et
al., 1990; Ellis et al., 1990; Garrell and Modolell, 1990,
reviewed in Norton, 2000). Like EMC, IDs are HLH
proteins that lack the basic domain (Benezra et al.,
8301
Figure 2 Possible mechanisms for Emc action. (a) Depicts the
situation in an SMC. Sc and Da form heterodimers that are able
to bind to DNA. The black rectangles indicate that the binding
domain of this heterodimer ®ts into the binding site present in the
promoter of the target gene. Sc/Emc and Da/Emc heterodimers
are unable to bind to DNA (indicated by the white rectangle).
Commitment to the SMC fate takes place in cells like that shown
in (a) in which the concentration of active heterodimers exceeds
that of ine€ective heterodimers, thus allowing activation of target
gene promoter. In this case Emc plays a repressive role. A similar
mechanism applies to the activation of Sxl in females. (b and c)
represent two possible mechanisms for Emc to function as an
activator of certain processes. (b) Represents a mesodermal cell in
which Twist (twi) homodimers activate transcription of myogenic
genes, while Twist/Da heterodimers, although able to bind to
other promoters as indicated by the hatched rectangle, do not
active such genes. The presence of Emc may be required for
somatic muscle development to titrate away Da, allowing the
formation of Twist homodimers. (c) A hypothetical situation in
which only A/B heterodimers are able to activate the target gene
promoter. The ability of the C protein to sequester B in the form
of heterodimers unable to bind to this promoter, is counteracted
by the interaction of Emc with C. Note that a similar mechanism
may operate in case the productive interaction between (a and b)
takes lace after promoter binding. See text for more details about
(a and b)
Oncogene
Multiples requirements for Emc in development
S Campuzano
8302
Oncogene
1990) and, similarly to the AS-C regulation of Emc, Id
negatively regulated myogenesis through the formation
of non-DNA binding complexes with the myogenic
bHLH factors (Benezra et al., 1990). The biochemical
demonstration that Id can physically interact with
members of the bHLH family and attenuate their
binding to DNA in vitro, and that in cultured cells Id
can inhibit the activation by MyoD of a structural
muscle-speci®c gene enhancer, the muscle creatin
kinase enhancer (Benezra et al., 1990), provided strong
support for the proposed mechanism of action of emc.
Thus, biochemical analyses in mammalian cells and
genetic data in ¯ies converged into the discovery of a
conserved mechanism of regulation of the function of
bHLH proteins. Afterwards, it was shown that, indeed,
EMC can form heterodimers with proteins of the AS-C
and Da and inhibit transcriptional activation of target
genes by interfering with their DNA binding ability
(Cabrera et al., 1994; MartõÂ nez et al., 1993; Van Doren
et al., 1991).
Although Emc regulates proneural protein function,
it has been suggested that Emc may repress AS-C
transcription indirectly since in emc mutant imaginal
wing discs there is an increased accumulation of Sc
(Cubas et al., 1991; Cubas and Modolell, 1992). It
has been proposed that the e€ect of Emc on ac ± sc
transcription results from its interference with ac and
sc self and cross-activation (MartõÂ nez et al., 1993;
Van Doren et al., 1992). Considering that more
recent evidence argues strongly against such crossregulatory interactions occurring between the endogenous ac and sc genes (GoÂmez-Skarmeta et al.,
1995), alternatives have to be sought to explain the
deregulation of sc in emc mutants. A possibility is
the promiscuity of Id-family proteins that, not only
interfere with the function of bHLH proteins, but
can modify the DNA binding ability (either inhibiting
or enhancing it) of homeoproteins, ETS proteins and
others, and can interfere with the activity of the Rb
protein (reviewed in Norton, 2000). Thus, the
derepression of sc in emc mutants may result from
Emc interfering with the regulation of sc by members
of the prepattern, a postulated combination of
transcription factors that regulate ac/sc in the
imaginal disc (reviewed by Modolell and Campuzano,
1998).
The targeted disruption of both Id1 and Id3 genes in
mice has demonstrated that these proteins are required
for several developmental processes, including neurogenesis (Lyden et al., 1999). Expression of several
proneural genes is expanded in such double mutants
(Lyden et al., 1999). This observation is consistent with
experiments where enforced expression of Id genes
inhibited both the activity and expression of bHLH
proneural proteins. Given the transcriptional hierarchies that operate within the proneural bHLH family
(Kageyama and Nakanishi, 1997), it seems likely that
Ids interfere with the regulation of a given proneural
bHLH gene by interacting with the product of another
bHLH gene located further up in the hierarchy (Lyden
et al., 1999).
Emc provides positional information for SO patterning
The distribution of Ac/Sc in the imaginal discs
pre®gures the pattern of SOs, since their precursors,
the SMCs, always arise from proneural clusters (Cubas
et al., 1991; Skeath and Carroll, 1991) (Figure 1a).
However, it has been found that generalized expression
of sc under the control of a heat shock promoter in
¯ies devoid of endogenous Ac and Sc promotes
development of bristles located at normal positions.
This suggested the presence in the imaginal discs of
topological cues other than the patterned distribution
of Ac and Sc (RodrõÂ guez et al., 1990). The spatial
distribution of emc RNA suggested that Emc could
provide some of these cues (Cubas and Modolell, 1992;
Van Doren et al., 1992). In the wing imaginal disc, emc
expression is roughly complementary to that of ac/sc
(Figure 1b). Moreover, SMCs appear not only within
proneural clusters but also within minima of emc
expression, a situation that also holds for SMCs
developing after generalized sc expression (Cubas and
Modolell, 1992) (Figure 1c). Expression of emc is
independent of AS-C (Cubas and Modolell, 1992).
These observations support a model in which the
interactions between proneural proteins and Emc re®ne
the spatial distribution of cells with the ability to
become SMCs. Within proneural clusters, the levels of
Emc would establish a threshold for the amount of Ac/
Sc proteins necessary for a cell to become an SMC.
Only cells with levels of proneural proteins sucient to
titrate Emc and to activate the downstream genes of
the neural di€erentiation pathway would be candidates
to become SMCs (reviewed in Garrell and Campuzano,
1991). These cells constitute a `proneural ®eld' (Cubas
and Modolell, 1992). Interaction among them,
mediated by the Notch signaling pathway, is further
required to restrict the number of cells that ®nally
di€erentiate as SMCs (reviewed in Artavanis-Tsakonas
et al., 1999). The remaining cells of the proneural
clusters would di€erentiate as epidermis.
Embryonic functions of emc
A current theme in development is the reiterative use of
the same protein for very di€erent processes. For
instance, the AS-C proteins and Da are required in
neurogenesis, sex determination and myogenesis (Carmena et al., 1995; Garrell and Campuzano, 1991; Jan
and Jan, 1993). Similarly, emc is required for multiple
processes during Drosophila embryonic and postembryonic development other than the regulation of the
pattern of bristles. Only in some cases it has been
possible to relate emc requirements with the function of
particular bHLH proteins, suggesting that Emc may
interfere with the function of non-bHLH proteins.
The widespread but not ubiquitous distribution of emc
transcripts in the embryo and the embryonic lethality of
emc null mutants (GarcõÂ a-Alonso and GarcõÂ a-Bellido,
1988) prompted the analysis of the embryonic functions
of emc (Cubas et al., 1994; Ellis, 1994).
Multiples requirements for Emc in development
S Campuzano
Emc is supplied maternally as shown by the
accumulation of emc mRNA in the ovaries and in
the syncytial blastoderm stage (Cubas et al., 1994; Ellis,
1994). This early expression corresponds to a requirement of Emc for sex determination. Sex is determined
at this stage by the transcriptional activation of the Sex
lethal (Sxl) gene, which is on in females and o€ in
males. The ®rst zygotic transcription of Sxl depends on
the X : A ratio, namely, the ratio between the number
of X-chromosomes (X) and autosomal sets (A) (Baker,
1989) and on the action of the bHLH protein Da
(Cline, 1989). It was determined genetically that the
X : A ratio is assessed by the interaction of `numerator
elements' present in the X-chromosomes with `denominator autosomal elements' (Cline and Meyer, 1996).
The X-linked genes sisterless-a (sis-a), which encodes a
bZIP protein (Erickson and Cline, 1993), and sisterlessb, (sis-b) later identi®ed as scute, correspond to
numerator elements (Cline, 1988; Erickson and Cline,
1991; Torres and SaÂnchez, 1989, 1991). The critical role
of Sc and Da (both bHLH proteins) in sex determination suggested that denominator elements could be
negative regulators of bHLH protein activity, similar
to Emc, encoded by autosomal genes and thus present
in equal amounts in males and females. Only in
females, where Sc concentration is twice that in males,
enough active Sc/Da heterodimers would be present to
overcome the inhibitory e€ects of the putative
autosomal-linked negative regulators and to activate
Sxl (Figure 2a). (Da, encoded by an autosomal gene, is
present in equal amounts in both sexes). Indeed,
overexpression experiments performed in early embryos indicated that Sxl activation depended on the
relative concentrations of bHLH X-linked proteins and
HLH autosomal-linked proteins (Parkhurst et al.,
1990). These data were con®rmed by the identi®cation
of emc and deadpan, a class VI HLH protein, as
denominator elements (Younger-Shepherd et al., 1992).
Hence, the choice of male vs female depends, similar to
the commitment to be an SMC or an epidermoblast,
on the relative concentration of bHLH and HLH
proteins (Figure 2a).
During the subsequent stages of embryogenesis, emc
is expressed in complex patterns and, accordingly,
several developmental processes are a€ected in emc
mutants (Cubas et al., 1994; Ellis, 1994). It is
noteworthy that emc mRNA is undetectable in the
ectodermal layer from which neuroblasts segregate
indicating that, in early embryonic neurogenesis, emc
does not regulate proneural function. Consistently, the
early patterns of proneural clusters and neural
precursors appear normal in emc mutants. Note,
however, that emc is expressed at later stages in the
developing central nervous system. Correspondingly,
before the start of germ ban retraction, emc mutants
display disorganized spatial distributions of neuroblasts
and their descendants and, later, anomalous axonal
patterns (Cubas et al., 1994).
emc mRNA accumulates in or adjacent to regions
that will undergo morphogenetic movements such as
the cephalic and ventral furrows, the amnioproctodeal
invagination and the invaginating stomodeum.
Although emc mutant embryos do not hatch and their
cuticle displays multiple alterations (Cubas et al.,
1994), no clear relationship has been established
between the pattern of expression of emc and the emc
mutant phenotype.
Especially informative is the role of emc in the
development of the Malpighian tubules, the excretory
system of the ¯y, and of the trachea. Both Malpighian
tubules and trachea are tubular structures that develop
in two steps. In the ®rst one, there is cell proliferation
in certain areas of the hindgut or of the ectoderm,
respectively, to provide the correct number of cells that
will form the ®nal structure. In the second one, there
are cell rearrangements, which do not involve cell
proliferation, but cell migration and changes of cell
shape, which result in formation and elongation of the
tubules (A€olter and Shilo, 2000; Skaer, 1993). In the
case of the Malpighian tubules, an enlarged cell, the tip
cell, appears at the distal part of each developing
tubule and directs and patterns cell proliferation
through activation of the Epidermal Growth Factor
Receptor (EGFR), (Baumann and Skaer, 1993). The
tip cell arises from a cluster of ac expressing cells and
maintains ac expression after its singling out (Hoch et
al., 1994). emc expression precedes and accompanies
formation of the Malpighian tubules and, accordingly,
emc mutations drastically a€ect their development.
Thus, in emc null mutants, tubules remain as rudiments
and contain a reduced number of cells (Cubas et al.,
1994; Ellis, 1994). In hypomorphic emc mutants, each
tubule rudiment has several extra tip cells (Cubas et al.,
1994; Carrera et al., 1998) but these extra cells do not
enhance cell proliferation. Moreover, the tubules do
not elongate. These results suggest, ®rst, a role of emc
in the control of cell proliferation; second, a role in the
allocation of the tip cell fate, where emc appears to
repress ac function, and third, a role of emc in the cell
rearrangements that lead to the elongation of the
tubules.
With regard to tracheal development, emc is
expressed in cells surrounding and including the
invaginating tracheal pits, the rudiments of the tracheal
tree (Cubas et al., 1994; Ellis, 1994). In emc mutants,
the ®nal structure of the tracheal tree is often altered
(Cubas et al., 1994). Similar phenotypes are found in
mutants for breathless (btl), an FGFR-type transmembrane receptor required for the proper migration of the
tracheal cells (Lee et al., 1996). Interestingly, Btl
protein levels are reduced in emc mutants (Cubas et
al., 1994). Moreover, expression of btl depends on the
localized expression of the bHLH-PAS transcription
factor trachealess (thl) (Ohshiro and Saigo, 1997). Thus
emc, a negative regulator of the function of bHLH
proteins, and the bHLH-PAS Thl are coexpressed in
the tracheal pit cells and both positively regulate btl.
The molecular mechanism for this positive cooperation
is at present unknown, but Figure 2 shows several
possibilities for Emc modes of action. Interestingly,
embryos from Id17/7 ± Id37/7 mice exhibit vascular
malformations and the absence of branching and
8303
Oncogene
Multiples requirements for Emc in development
S Campuzano
8304
sprouting of blood vessels into the neuroectoderm
(Lyden et al., 1999). This suggests parallel functions of
emc and Id in the formation of tubular structures.
Emc is also involved in the development of
mesodermal derivatives. The mesoderm develops from
the ventral-most cells of the blastoderm. These cells
require Twist (Twi), a bHLH protein, for its proper
speci®cation (reviewed by Baylies et al., 1998). This
region invaginates during gastrulation and is later
subdivided into somatic and visceral mesoderm which
give rise, respectively, to the body wall muscles
(somatic muscles) and the muscles of the gut (visceral
muscles) (reviewed in Baylies et al., 1998). emc is
expressed at blastoderm stage in the prospective
mesoderm; in the mesodermal layer during gastrulation
and later, in the muscle attachment sites (apodemes)
(Cubas et al., 1994). Accordingly, emc mutations
profoundly a€ect the pattern of somatic muscles, the
defects ranging from loss of some muscles to large
alterations of this pattern, with muscles forming large
masses detached from the epidermis (Cubas et al.,
1994). The latter phenotype may be attributed to
defective speci®cation of the apodemes. The function
of emc in this process may be related to that of Delilah,
a bHLH protein that is expressed almost exclusively in
the apodemes (Armand et al., 1994). The alteration of
the pattern of somatic muscle in emc mutants appears
to result from de®cient twi function. First, levels of
Twist are decreased in emc embryos (Cubas et al.,
1994). Second, Emc may be indirectly required for Twi
function (Castanon et al., 2001) (Figure 2b). Twist can
form homodimers and heterodimers with Da in vitro
(Castanon et al., 2001). Interestingly, in vivo analysis
indicates that whereas Twist homodimers specify
mesodermal fate, Twist-Da heterodimers repress expression of genes required for somatic myogenesis
(Castanon et al., 2001). Considering these data and the
previous observation that Emc can heterodimerize with
Da (Cabrera et al., 1994; MartõÂ nez et al., 1993; Van
Doren et al., 1991), during somatic muscle development Emc would be required to relieve inhibition of
Twist by Da, by sequestering Da as Emc-Da
heterodimers (Figure 2b).
emc is also expressed in the cells of the visceral
mesoderm while they are spreading over the midgut
and indeed, cells of the visceral mesoderm showed a
clear migration defect in emc mutants (Cubas et al.,
1994).
In summary, the analysis of the embryonic functions
of emc has shown that emc not only represses
di€erentiation but it plays positive roles, like favoring
muscle development, although mechanistically it behaves as a negative regulator of other proteins.
Emc represses the initiation of neural di€erentiation in
the eye
The initial steps of di€erentiation in the Drosophila eye,
similar to the determination of the SMCs, depend on
the activity of bHLH proteins, which is counteracted
Oncogene
by HLH repressors. The eye of Drosophila comprises
about 750 ommatidia, each composed of eight
photoreceptor and several support cells. These cells
di€erentiates in the eye imaginal disc in a stereotyped
order behind the so-called morphogenetic furrow (MF)
(Ready et al., 1976). The proneural bHLH gene atonal
(ato) is required, in combination with Da, for the
speci®cation of R8 (Jarman et al., 1994), the ®rst
photoreceptor to develop and the one that subsequently recruits the other photoreceptors into the
developing ommatidium (Freeman, 1997). ato is ®rst
weakly expressed in a dorsoventral stripe of cells
anterior and within the MF, and later is upregulated,
®rst in proneural clusters, and, ®nally, in the R8 cells
(Dokucu et al., 1996). The expression and activity of
ato is negatively regulated by Emc and Hairy, a bHLH
transcriptional repressor, which are expressed anterior
to the MF (Brown et al., 1995). Thus, whereas clones
of cells lacking either h or emc function alone have
only subtle e€ects on photoreceptor cell development,
combined loss of emc and h causes upregulation of ato
and the subsequent di€erentiation of photoreceptor
cells, in much anterior positions than in the adjacent
wild type tissue (Brown et al., 1995). These observations suggest that a reduction on emc and h levels is a
major prerequisite for the initiation of eye development. This is precisely the case. Recently it has been
shown that downregulation of h and emc is controlled
by the Notch signaling pathway, which is activated in
the eye disc cells anterior and close to the MF in
response to Hedgehog secreted by the already di€erentiated photoreceptors (Baonza and Freeman, 2001).
emc and vein di€erentiation
The Drosophila wing displays a stereotyped pattern of
®ve longitudinal veins and two transverse veins. This
pattern is established in the imaginal wing disc by the
combined action of the Notch and EGFR signaling
pathways (reviewed in de Celis, 1998). Expression of
veinlet/rhomboid (ve/rho) in prevein regions helps to
locally activate the EGFR/Ras pathway, which is
instrumental to promote vein formation. EGFR/Ras
activates Delta in the veins, which binds to and
activates Notch at the cells ¯anking the veins. The
Notch pathway in turn activates transcription of
E(spl)mb, which encodes a bHLH protein that
represses expression of ve/rho and prevents the
acquisition of vein fate in the cells ¯anking the proveins. The shape and position of emc clones in the wing
and the pattern of expression of emc suggests a role for
this gene in vein di€erentiation. Thus, clones of cells
hypomorphic for emc are elongated, frequently appear
to run along the veins and they also can di€erentiate
ectopic veins of normal thickness (Baonza and GarcõÂ aBellido, 1999; Baonza et al., 2000; de Celis et al., 1995;
GarcõÂ a-Alonso and GarcõÂ a-Bellido, 1988). In the pupal
wing emc is expressed at highest levels in the intervein
cells ¯anking the veins (Baonza et al., 2000). These are
the cells where Notch signaling is maximally active and
Multiples requirements for Emc in development
S Campuzano
indeed, expression of emc coincides with that of
E(spl)mb (Baonza et al., 2000). Moreover, Baonza et
al. (2000) demonstrated that Notch signaling is
required from the beginning of pupal development to
establish emc transcription pattern. This regulatory
interaction may account for the observation that,
although the phenotype of emc clones is di€erent to
that associated to loss-of-function alleles of Notch,
which cause the formation of thicker veins but not the
appearance of ectopic veins, however, when Notch
signaling is compromised, vein di€erentiation becomes
very sensitive to reductions in emc function (Baonza et
al., 2000). Moreover, both Emc and E(spl)mb
cooperate in the repression of ve (Baonza et al.,
2000). Since Emc and E(spl)mb do not physically
interact (Baonza et al., 2000), Emc may titrate a
protein that inhibits the transcriptional activity of an
unidenti®ed factor which co-operates with E(spl)mb in
the repression of ve/rho (Figure 2c,d). These observations and those presented above concerning eye
development, are the ®rst indications of how the
function of an Id-family protein, Emc, is integrated
with that of an upstream signaling pathway, the Notch
pathway, which plays a crucial and evolutionarily
conserved role in the control of cell determination
(reviewed in Artavanis-Tsakonas et al., 1999).
Emc regulates cell proliferation
Cooperation of emc and the Notch pathway also
operates in the control of cell proliferation. In the wing
disc, cell lacking emc function do not survive but those
harboring strong hypomorphic emc alleles can proliferate, albeit to a reduced rate compared to wild type
cells (Baonza and GarcõÂ a-Bellido, 1999; de Celis et al.,
1995; GarcõÂ a-Alonso and GarcõÂ a-Bellido, 1988). Loss
of Notch function similarly reduces cell viability (de
Celis and GarcõÂ a-Bellido, 1994). Moreover, in the wing
anlage cells doubly mutant for both emc and Notch
have extremely poor viability suggesting that emc and
the Notch pathway cooperate to promote cell proliferation (Baonza et al., 2000). This interaction does
not rely on Notch controlling emc transcription. Thus,
it has been proposed that emc and Notch signaling act
in parallel, possibly on the same set of downstream
genes, to promote cell proliferation (Baonza et al.,
2000).
In cultured ®broblasts, Id proteins positively control
cell cycle progression at mid-late G1 phase by several
ways (reviewed in Norton, 2000). Namely, Id proteins
can relieve the inhibitory interaction of pRB with
E2F, a factor required for the transcription of S phase
genes. Additionally, Ids interfere with the transcriptional activation by bHLH proteins of the cyclindependent kinase inhibitor p21WAF1/CIP1. Since mechanisms controlling cell cycle progression are evolutionarily conserved, it is safe to speculate that Emc
may act similarly in controlling the cell cycle in ¯ies.
Con®rmation of this hypothesis awaits future experimentation.
Regulation of transcription and translation of Emc
8305
The above results indicate that the pattern of
expression of emc and the level of accumulation of
Emc protein must be ®nely controlled during development. Although this is a relevant issue, not much is
known of the regulation of emc. The clustering of emc
regulatory mutations, mostly P-element insertions, in a
rather short stretch of DNA (Ellis et al., 1990; Garrell
and Modolell, 1990) argues against the existence of
extended cis regulatory regions. However, the fact that
emc mutations associated to chromosomal breakpoints
map at considerable distance from the emc promoter,
suggests otherwise (Garrell and Modolell, 1990).
Importantly, emc transcription is not regulated by the
AS-C (Cubas and Modolell, 1992). As indicated above,
Notch pathway regulates emc expression, although in
opposite ways in di€erent tissues: it represses emc
transcription in the eye disc (Baonza and Freeman,
2001) while activates it at the prospective wing margin
region of the wing imaginal disc and at the vein/
intervein boundaries in pupal wings (Baonza et al.,
2000). Although the direct e€ector is unknown, in the
latter case it is known to be distinct from E(spl)mb
(Baonza et al., 2000).
The polychaetoid (pyd) mutation (also called Tamou,
a Japanese term for `hairy') gives rise to ectopic SOs
resembling the phenotype of insuciency of emc
(Chen et al., 1996; Takahisa et al., 1996). Moreover,
Pyd and emc mutations act synergistically in the
development of extra veins and SOs (Takahisa et al.,
1996). pyd ectopic bristles are due to the formation of
additional SMCs (Chen et al., 1996; Takahisa et al.,
1996) which appear to result from reduced transcription of emc in pyd mutant discs (Takahisa et al.,
1996). Pyd is a PDZ domain protein whose sequence
is closely related to mammalian Zona occludens
proteins ZO-1, ZO-2 and ZO-3, members of the
membrane-associated guanylate kinase homologue
(MAGUK) family (reviewed in Dimitratos et al.,
1999). Since Pyd is associated with septate and
adherens junctions (Takahisa et al., 1996; Wie and
Ellis, 2001), its role in emc regulation cannot be a
direct one. The presence of multiple protein ± protein
interaction domains in PDZ proteins allows them to
act as sca€olding components to organize protein
complexes (Dimitratos et al., 1999). Accordingly, it
has been suggested that Pyd participates in a signaling
pathway that ultimately activates emc expression
(Takahisa et al., 1996).
Analysis of another mutation that a€ects the adult
pattern of SOs, Bearded, provided indirect evidence
for a postranscriptional regulation of emc. Bearded
contains in their 3' untranslated region (UTR) several
copies of two heptanucleotide sequence motifs, the
Brd box (AGCTTTA) and the GY box (GTCTTCC).
Mutant Brd transcripts that are truncated in the 3'
UTR due to a transposon insertion and consequently
lack two of the three Brd boxes and the GY box, are
present at elevated steady-state levels. This is consistent with these motifs functioning to destabilize the
Oncogene
Multiples requirements for Emc in development
S Campuzano
8306
wild-type transcript (Leviten et al., 1997). The Brd
box may also a€ect translational eciency (Lai and
Posakony, 1997). emc contains in its 3' UTR a single
Brd box and four copies of a GY box-related
sequence (GTTTTCC) (Lai and Posakony, 1997). A
role of these sequences in emc mRNA stability is
supported by the Achaetous mutation: This gain-offunction mutation is associated with the insertion of
the transposable element Tirant in the ®rst exon of
emc (Garrell and Modolell, 1990) and levels of emc
transcripts are increased in Achaetous imaginal discs
(Cubas and Modolell, 1992). It is unknown how these
RNA sequences regulate translation and/or mRNA
stability: regulatory proteins may bind to these
sequences and regulate messenger activity or stability
or, alternatively, formation of RNA : RNA heteroduplexes may regulate mRNA function (Lai and
Posakony, 1998). The Emc transcript may contain
other domains susceptible of postranscriptional regulation. Thus, in vitro RNA binding experiments
demonstrate that Hel-N1, a human protein highly
similar to the product of the Drosophila elav gene, can
bind to the 3' UTR of Id mRNA. Hel-N1 binding
sequences are also present in the 3' UTR of the
Drosophila emc mRNA (King et al., 1994). Elav is an
RNA binding protein known to regulate splicing of
several neural speci®c genes (Lisbin et al., 2001). It
would thus be interesting to determine whether Elav,
which accumulates in the neurons (Robinow and
White, 1988), plays any role in the downregulation
of Emc which appears necessary for neural development to proceed.
Perspectives
More than 10 years have passed since the molecular
cloning of emc and Id, and we have gained some
insights on the function of emc during Drosophila
development. Emc provides a clear example of how cell
determination relies on the cell's ability to respond to
subtle di€erences in the concentration of regulatory
proteins. However, many unresolved questions remain.
We do not know yet which bHLH proteins Emc does
interact with in many of the development processes
a€ected by emc mutations. We even do not know
whether all of these Emc-interacting proteins are
bHLH proteins. Such proteins can be identi®ed by
means of the two-hybrid system and later should be
functionally characterized by genetic analyses. Furthermore, genetic screens can be devised to look for genes
that interact with emc, based for instance on the
suppression or enhancement of the adult phenotype of
the emc gain-of-function mutation Achaetous. Some of
these approaches are currently under way and we can
foresee the identi®cation of new Emc partners that may
help to shed a light into parallel processes carried out
by mammalian Id proteins.
Acknowledgments
I would like to thank J Modolell, M Ruiz-GoÂmez and J de
Celis for comments on the manuscript. Support from
DGICYT, Comunidad Auto noma de Madrid and an
institutional grant from FundacioÂn RamoÂn Areces to the
Centro de BiologõÂ a Molecular Severo Ochoa is acknowledged.
References
A€olter M and Shilo BZ. (2000). Curr. Opin. Cell Biol., 12,
731 ± 735.
Armand P, Knapp AC, Hirsch AJ, Wieschaus EF and Cole
MD. (1994). Mol. Cell. Biol., 14, 4145 ± 4154.
Artavanis-Tsakonas S, Rand MD and Lake RJ. (1999).
Science, 284, 770 ± 776.
Baker B. (1989). Nature, 340, 521 ± 524.
Balcells L, Modolell J and Ruiz-GoÂmez M. (1988). EMBO J.,
7, 3899 ± 3906.
Baonza A, de Celis JF and GarcõÂ a-Bellido A. (2000).
Development, 127, 2383 ± 2393.
Baonza A and Freeman M. (2001). Development, 128, 3889 ±
3898.
Baonza A and GarcõÂ a-Bellido A. (1999). Mech. Dev., 80,
133 ± 146.
Bate M. (1978). Handbook of Sensory Physiology. Vol 1X.
M Jacobson (ed). New York, Springer.
Baumann P and Skaer H. (1993). Development, (Suppl): 65 ±
76.
Baylies MK, Bate M and Ruiz-GoÂmez M. (1998). Cell, 93,
921 ± 927.
Benezra R, Davis RL, Lockshon D, Turner DL and
Weintraub H. (1990). Cell, 61, 49 ± 59.
Bodmer R, Carretto R and Jan YN. (1989). Neuron, 3, 21 ±
32.
Botas J, Moscoso del Prado J and GarcõÂ a-Bellido A. (1982).
EMBO J., 1, 307 ± 310.
Oncogene
Cabrera CV, Alonso MC and Huikeshoven H. (1994).
Development,120, 3595 ± 3603.
Brown NL, Sattler CA, Paddock SW and Carroll SB. (1995).
Cell, 80, 879 ± 887.
Cabrera CV, MartõÂ nez-Arias A and Bate M. (1987). Cell, 50,
425 ± 433.
Carmena A, Bate M and JimeÂnez F. (1995). Genes Dev., 9,
2373 ± 2383.
Carrera P, Abrell S, Kerber B, Walldorf U, Preiss A, Hogh
M and JaÈckle H. (1998). Proc. Natl. Acad. Sci. USA, 95,
10779 ± 10784.
Castanon I, Von Stetina S, Kass J and Baylies MK. (2001).
Development, 128, 3145 ± 3159.
Caudy M, Grell EH, Dambly-ChaudieÁre C, Ghysen A, Jan
LY and Jan YN. (1988a). Genes Dev., 2, 843 ± 852.
Caudy M, Vassin H, Brand M, Tuma R, Jan LY and Jan YN.
(1988b). Cell, 55, 1061 ± 1067.
Chen C-M, Freedman JA, Bettler DR, Manning SD, Giep
SN, Steiner J and Ellis HM. (1996). Mech. Dev., 57, 215 ±
227.
Cline TW. (1988). Genetics, 119, 829 ± 862.
Cline TW. (1989). Cell, 59, 231 ± 234.
Cline TW and Meyer BJ. (1996). Annu. Rev. Genet., 30, 637 ±
702.
Cubas P and Modolell J. (1992). EMBO J., 11, 3385 ± 3393.
Cubas P, de Celis JF, Campuzano S and Modolell J. (1991).
Genes Dev., 5, 996 ± 1008.
Multiples requirements for Emc in development
S Campuzano
Cubas P, Modolell J and Ruiz-GoÂmez M. (1994). Development, 120, 2555 ± 2566.
Dambly-ChaudieÁre C, Ghysen A, Jan LY and Jan YN.
(1988). Roux's Arch. Dev. Biol., 197, 419 ± 423.
Davis RL, Cheng PF, Lassar AB and Weintraub H. (1990).
Cell, 60, 733 ± 746.
de Celis JF. (1998). Int. J. Dev. Biol., 42, 335 ± 343.
de Celis JF and GarcõÂ a-Bellido A. (1994). Mech. Dev., 46,
109 ± 122.
de Celis JF, Baonza A and GarcõÂ a-Bellido A. (1995). Mech.
Dev., 53, 209 ± 221.
Dimitratos SD, Woods DF, Stathakis DG and Bryant PJ.
(1999). BioEssays, 21, 912 ± 921.
Dokucu ME, Zipursky SL and Cagan RL. (1996). Development, 122, 4139 ± 4147.
Ellis HM. (1994). Mech. Dev., 47, 65 ± 72.
Ellis HM, Spann DR and Posakony JW. (1990). Cell, 61,
27 ± 38.
Erickson JW and Cline TW. (1991). Science, 251, 1071 ±
1074.
Erickson JW and Cline TW. (1993). Genes Dev., 9, 1688 ±
1702.
Freeman M. (1997). Development, 124, 261 ± 270.
GarcõÂ a-Bellido A. (1979). Genetics, 91, 491 ± 520.
GarcõÂ a-Alonso LA and GarcõÂ a-Bellido A. (1988). Roux's
Arch. Dev. Biol., 197, 328 ± 338.
GarcõÂ a-Bellido A and Merriam J. (1971). Develop. Biol., 24,
61 ± 87.
GarcõÂ a-Bellido A and SantamarõÂ a P. (1978). Genetics, 88,
469 ± 486.
Garrell J and Campuzano S. (1991). BioEssays, 13, 493 ± 498.
Garrell J and Modolell J. (1990). Cell, 61, 39 ± 48.
Ghysen A and Dambly-ChaudieÁre C. (1989). Trends Genet.,
5, 251 ± 255.
GoÂmez-Skarmeta JL, RodrõÂ guez I, MartõÂ nez C, CulõÂ J,
FerreÂs-Marco MD, Beamonte D and Modolell J. (1995).
Genes Dev., 9, 1869 ± 1882.
Hartenstein V and Posakony JW. (1989). Development, 107,
389 ± 405.
Hoch M, Broadie K, JaÈckle H and Skaer H. (1994).
Development, 120, 3439 ± 3450.
Jan YN and Jan LY. (1993). Cell, 75, 827 ± 830.
Jarman AP, Grell EH, Ackerman L, Jan LY and Jan YN.
(1994). Nature, 369, 398 ± 400.
Kageyama R and Nakanishi S. (1997). Curr. Opin. Genet.
Dev., 7, 659 ± 665.
King PH, Levine TD, Fremeau Jr RT and Keene JD. (1994).
J. Neurosci., 14, 1943 ± 1952.
Lai EC and Posakony JW. (1997). Development, 124, 4847 ±
4856.
Lai EC and Posakony JW. (1998). Cell, 93, 1103 ± 1104.
Lee T, Hacohen N, Krasnow M and Montell D. (1996).
Genes Dev., 10, 2912 ± 2921.
Leviten MW, Lai EC and Posakony JW. (1997). Development, 124, 4039 ± 4051.
Lisbin MJ, Qiu J and White K. (2001). Genes Dev., 15, 2546 ±
2561.
Lyden D, Young AZ, Zagzag D, Yan W, Gerald W, O'Reilly
R, Bader BL, Hynes RO, Zhuang Y, Manova K and
Benezra R. (1999). Nature, 401, 670 ± 677.
MartõÂ nez C, Modolell J and Garrell J. (1993). Mol. Cell.
Biol., 13, 3514 ± 3521.
Massari ME and Murre C. (2000). Mol. Cell. Biol., 20, 429 ±
440.
Modolell J and Campuzano S. (1998). Int. J. Dev. Biol., 42,
275 ± 282.
Moscoso del Prado J and GarcõÂ a-Bellido A. (1984a). Roux's
Arch. Dev. Biol., 193, 246 ± 251.
Moscoso del Prado J and GarcõÂ a-Bellido A. (1984b). Roux's
Arch. Dev. Biol., 193, 242 ± 245.
Murre C, McCaw PS and Baltimore D. (1989a). Cell, 56,
777 ± 783.
Murre C, McCaw PS, Vaessin H, Caudy M, Jan LY, Jan YN,
Cabrera CV and Baltimore D. (1989b). Cell, 58, 537 ± 544.
Norton JD. (2000). J. Cell Sci., 113, 3897 ± 3905.
Ohshiro T and Saigo K. (1997). Development, 124, 3975 ±
3986.
Parkhurst SM, Bopp D and Ish-Horowicz D. (1990). Cell,
63, 1179 ± 1191.
Ready DF, Hanson TE and Benzer S. (1976). Dev. Biol., 53,
217 ± 240.
Robinow S and White K. (1988). Dev. Biol., 126, 294 ± 303.
RodrõÂ guez I, HernaÂndez R, Modolell J and Ruiz-GoÂmez M.
(1990). EMBO J., 9, 3583 ± 3592.
Romani S, Campuzano S, Macagno E and Modolell J.
(1989). Genes Dev., 3, 997 ± 1007.
Skaer H. (1993). The development of Drosophile melanogaster. MartõÂ nez-Arias A and Bate M (eds). Cold Spring
Harbor Lab. Press. New York. Cold Spring Harbor.
Skeath JB and Carroll SB. (1991). Genes Dev., 5, 984 ± 995.
Takahisa M, Togashi S, Suzuki T, Kobayashi M, Murayama
A, Kondo K, Miyake T and Ueda R. (1996). Genes Dev.,
10, 1783 ± 1795.
Torres M and SaÂnchez L. (1989). EMBO J., 8, 3079 ± 3086.
Torres M and SaÂnchez L. (1991). Development, 113, 715 ±
722.
Van Doren M, Ellis HM and Posakony JW. (1991).
Development, 113, 245 ± 255.
Van Doren M, Powell PA, Pasternak D, Singson A and
Posakony JW. (1992). Genes Dev., 6, 2592 ± 2605.
Villares R and Cabrera CV. (1987). Cell, 50, 415 ± 424.
Voronova A and Baltimore D. (1990). Proc. Natl. Acad. Sci.
USA, 87, 4722 ± 4726.
Wie X and Ellis HM. (2001). Mech. Dev., 100, 217 ± 231.
Younger-Shepherd S, Vaessin H, Bier E, Jan LY and Jan
YN. (1992). Cell, 70, 911 ± 922.
8307
Oncogene