screens. Firstly, since most genes in Drosophila are not haploinsufficient, only homozygous mutant embryos will reveal a phenotype. This can involve significant extra work since one must
isolate and maintain mutant alleles over wild-type balancer chromosomes and then identify homozygotes using transgenic markers
on the balancers such as lacZ transgenes. Secondly, due to maternal
gene product, only a few genes will exhibit early zygotic phenotypes
[7]. For example, despite known roles for RhoA in cellularization
[8] and ventral furrow formation [9], zygotic RhoA mutant
embryos exhibit only late embryonic phenotypes such as defects
in dorsal closure [10]. In addition, although RNAi through the
female germline can generate early embryonic phenotypes [11],
strong tissue-specific RNAi at such early stages is not effective.
A later EMT event that is better suited to genetic screens is
wing disc eversion. This occurs during the early stages of metamorphosis when imaginal discs undergo major morphogenetic movements to create the adult body. Topologically, imaginal discs are
epithelial sacs with their apical surfaces oriented inwards. Hence,
during eversion, discs must somehow turn themselves inside out.
Traditionally this was thought to involve a “turning-out-yourpocket” mechanism [12]. This model was overturned in 2004 by
Pastor-Pareja et al. [13], however, when it was shown that one side
of the disc, the squamous peripodial epithelium (PE), undergoes a
partial EMT (pEMT) event that creates perforations through which
the rest of the wing disc can emerge. The morphogenetic movements of the PE have been subsequently well documented using
live imaging ex vivo [14, 15]. Recently, a similar mechanism has
been demonstrated for leg discs [16], suggesting that all imaginal
discs may use this method.
Wing eversion presents as a useful system for identifying genes
involved in EMT in other systems, as it exhibits many of the classic
hallmarks of EMT. PE cells express matrix metalloproteases
(MMPs) to break down the basement membrane, lose apico-basal
polarity, delocalize cell-cell junction proteins such as E-Cadherin
(E-Cad), and extend F-Actin rich protrusions as they invade the
overlying larval epidermal cells. The wing discs from either side of
the larva subsequently undergo an epithelial migration, eventually
meeting up at the dorsal midline where they fuse to form the final
pupal/adult epithelium [17]. PE cells also undergo apoptosis as the
retraction of the epithelium proceeds but our group and others
[13] have found that eversion will still take place when one inhibits
apoptosis with p35, so it appears separable from the pEMT itself.
Mutations in genes required for eversion such as the Jun
N-terminal Kinase (JNK) pathway and MMPs result in easily scored
adult phenotypes such as missing wings and thoracic clefts [18–
20]. Also, since eversion occurs late in development, tissue-specific
RNAi is very effective. Disc eversion, therefore, provides an accessible system to screen for, and analyze, genes involved in EMT.
116
Sofia Golenkina et al.
isolate and maintain mutant alleles over wild-type balancer chromosomes and then identify homozygotes using transgenic markers
on the balancers such as lacZ transgenes. Secondly, due to maternal
gene product, only a few genes will exhibit early zygotic phenotypes
[7]. For example, despite known roles for RhoA in cellularization
[8] and ventral furrow formation [9], zygotic RhoA mutant
embryos exhibit only late embryonic phenotypes such as defects
in dorsal closure [10]. In addition, although RNAi through the
female germline can generate early embryonic phenotypes [11],
strong tissue-specific RNAi at such early stages is not effective.
A later EMT event that is better suited to genetic screens is
wing disc eversion. This occurs during the early stages of metamorphosis when imaginal discs undergo major morphogenetic movements to create the adult body. Topologically, imaginal discs are
epithelial sacs with their apical surfaces oriented inwards. Hence,
during eversion, discs must somehow turn themselves inside out.
Traditionally this was thought to involve a “turning-out-yourpocket” mechanism [12]. This model was overturned in 2004 by
Pastor-Pareja et al. [13], however, when it was shown that one side
of the disc, the squamous peripodial epithelium (PE), undergoes a
partial EMT (pEMT) event that creates perforations through which
the rest of the wing disc can emerge. The morphogenetic movements of the PE have been subsequently well documented using
live imaging ex vivo [14, 15]. Recently, a similar mechanism has
been demonstrated for leg discs [16], suggesting that all imaginal
discs may use this method.
Wing eversion presents as a useful system for identifying genes
involved in EMT in other systems, as it exhibits many of the classic
hallmarks of EMT. PE cells express matrix metalloproteases
(MMPs) to break down the basement membrane, lose apico-basal
polarity, delocalize cell-cell junction proteins such as E-Cadherin
(E-Cad), and extend F-Actin rich protrusions as they invade the
overlying larval epidermal cells. The wing discs from either side of
the larva subsequently undergo an epithelial migration, eventually
meeting up at the dorsal midline where they fuse to form the final
pupal/adult epithelium [17]. PE cells also undergo apoptosis as the
retraction of the epithelium proceeds but our group and others
[13] have found that eversion will still take place when one inhibits
apoptosis with p35, so it appears separable from the pEMT itself.
Mutations in genes required for eversion such as the Jun
N-terminal Kinase (JNK) pathway and MMPs result in easily scored
adult phenotypes such as missing wings and thoracic clefts [18–
20]. Also, since eversion occurs late in development, tissue-specific
RNAi is very effective. Disc eversion, therefore, provides an accessible system to screen for, and analyze, genes involved in EMT.
116
Sofia Golenkina et al.