INTEGR. COMP. BIOL., 43:508–521 (2003)
Evolution of Insect Eye Development:
First Insights from Fruit Fly, Grasshopper and Flour Beetle1
MARKUS FRIEDRICH2
Department of Biological Sciences, Wayne State University, 5047 Gullen Mall, Detroit, Michigan 48202
SYNOPSIS. The molecular genetic dissection of Drosophila eye development led to the exciting discovery of
a surprisingly large panel of genes and gene activities, which are functionally conserved across phyla. Little
effort has yet been made towards pinpointing non-conserved gene functions in the developing Drosophila
eye. This neglects the fact that Drosophila visual system development is a highly derived process. The comparative analysis of Drosophila eye development within insects can be expected to enhance resolution and
accuracy of between phyla comparisons of eye development, and to reveal molecular developmental changes
that facilitated the evolutionary transition from hemimetabolous to holometabolous insect development. Here
we review aspects of early Drosophila eye development, which are likely to have diverged from the situation
in more primitive insects, as indicated by results from work in the flour beetle Tribolium castaneum and the
grasshopper Schistocerca americana.
INTRODUCTION
The power of biological model systems equals the
ease of genetic analysis multiplied by applicability to
human biology. The fruit fly Drosophila melanogaster
has performed beyond expectation under this equation.
Originally chosen for reasons of affordability and short
generation time rather than similarity to vertebrate
physiology, the analysis of the complete Drosophila
genome delivered ultimate proof that this model species also positions usefully close to humans on the tree
of life if shared genes and gene functions are taken as
measure of relatedness (Rubin et al., 2000). The surprising evolutionary insights that emanated from the
molecular dissection of Drosophila compound eye development constitute a paradigm case of radical
change of comparative perspective. Considering the
fundamental differences between the Drosophila compound eye and the vertebrate lens eye, an independent
origin of insect and vertebrate visual organs seemed
certain (Salvini-Plawen and Mayr, 1977). This premolecular view was revolutionized by the discovery
in Walter Gehring’s laboratory that the orthologs of
the paired class transcription factor Pax6 are essential
for eye development in both fly and vertebrates (Quiring et al., 1994). Subsequent studies revealed that homologs of three equally conserved gene families eyes
absent (eya), sine oculis (so) and dachshund (dac) interact with Pax6 in an evolutionarily conserved eye
specification regulatory network (for review see Desplan, 1997; Pappu and Mardon, 2002). The finding of
similarities in the function of the proneural transcription factor atonal (ato) and the signaling molecule
hedgehog (hh) in the developing retina of flies and
vertebrates raised the possibility that the conservation
of animal eye development extends to processes down-
stream of eye specification (Kumar, 2001; Brown et
al., 2001; Neumann and Nuesslein-Volhard, 2000).
The nature of Drosophila eye development conservation has been a topic of lively discussion (Halder et
al., 1995; Hanson and Van Heyningen, 1995; Gehring,
1996; Tomarev, 1997; Kumar, 2001; Pichaud et al.,
2001; Pichaud and Desplan, 2002). Less attention has
been paid to the fact that Drosophila eye development
also offers the opportunity to unravel molecular developmental changes underlying the emergence of
evolutionary novelty. Drosophila represents one of the
most derived modes of insect eye development to the
eyes of an entomologist (Fig. 1). This is in part due
to the derived life cycle and in part to unique morphological modifications of the juvenil instars. As is
typical for holometabolous insects, the fruitfly develops through a series of specialized postembryonic
growth stages. These larval instars lack most elements
of the adult body plan including the compound eyes.
The complex transformation to adult morphology begins in the last larval instar and completes during the
resting stage of the pupa. The development of the adult
Drosophila eye is thus a postembryonic process. In
primitive insects, however, much of the adult retina
develops already during embryogenesis (Fig. 1). The
immature instars or nymphs of hemimetabolous insects
such as grasshopper or cockroach for instance are born
with adult like morphology. Elaboration of the adult
morphology during the final nymphal molt requires
only the extension of wings in addition to differentiation of functional genitalia. The first instar nymph
possesses a pair of fully functional compound eyes,
which are enlarged between the subsequent growth
molts and maintained into the adult. Morphological
and molecular data firmly establish that holometabolous development evolved from within hemimetabolous insects (Kristensen, 1995). The postembryonic
mode of Drosophila eye development is thus clearly
an evolutionarily derived process.
Evolutionary change of morphology requires
change of developmental programs (Carroll, 1994).
1 From the Symposium Comparative and Integrative Vision Research presented at the Annual Meeting of the Society for Integrative and Comparative Biology, 4–8 January 2003, at Toronto, Canada.
2 E-mail: [email protected]
508
EVOLUTION
OF INSECT
EYE DEVELOPMENT
509
FIG. 1. Life cycle and time course of compound eye development in Drosophila, Tribolium and Schistocerca. Time windows of compound
eye retina specification, determination and differentiation indicated by open, grey and black arrows respectively. Specification and determination
arrows in Tribolium box are hypothetical (see text). Cladogram to the right of the boxes indicates phylogenetic relationships between the three
species.
Evidently, the evolution of holometabolous insects
must have involved modifications of embryonic development to generate the derived morphology of the
larva. It is also reasonable to assume that the postembryogenic differentiation of adult structures starting
from the larval body plan involved modifications of
ancestrally embryonic differentiation processes. Thus,
the evolutionary transition from embryonic to postembryonic eye development most likely involved not
only a temporal shift of the onset of adult retina differentiation, but also considerable modifications of
embryonic and postembryonic visual system development. The extent to which the deeply conserved molecular genetic control mechanisms of Drosophila eye
development are embedded in derived patterning
mechanisms is unclear. This is because our knowledge
of insect eye development outside the genus Drosophila is fairly limited. Accounts on eye development in
insect species other than Drosophila are scattered
(Meinertzhagen, 1973; Bate, 1978; Trujillo-Cenoz,
1985; Egelhaaf, 1988; Melzer and Paulus, 1994; Friedrich et al., 1996; Champlin and Truman, 1998). A major focus has been the comparison of the sequence of
cell differentiation events in the developing retina,
which is highly conserved in insects and crustaceans
(Melzer et al., 2000; Hafner and Tokarski, 2001). This
review discusses first evidence for evolutionary divergence of molecular genetic patterning events, which
lead up to the onset of differentiation of the compound
eye retina in Drosophila. Examples will be drawn
from ongoing comparative analyses in the American
desert locust Schistocerca americana, a hemimetabolous insect, and the flour beetle Tribolium castaneum,
a primitive holometabolous insect.
Drosophila, Tribolium and Schistocerca: Three ways
to make an insect eye
Drosophila represents one of the most complex
modes of insect compound eye development (for review see Wolff and Ready, 1993). First complications
stem from the fact that the Drosophila larva is acephalic. It lacks head appendages and is furnished with
an internalized head skeleton instead (Fig. 1). As a
consequence, the larval head components are replaced
with newly differentiated adult head capsule tissues
during postembryonesis. The cuticle skeleton of the
adult head develops from specialized ectodermal sacs,
the imaginal discs. These may be considered secondary embryonic fields that are set aside during early
embryogenesis and remain buried deep inside the body
during larval development. The compound eye retina
develops from the eye-antennal imaginal disc, which
is a derivative of the embryonic visual primordium
(Fig. 2A). Only little developed at the beginning of
postembryogenesis the eye-antennal imaginal disc proceeds through continuous growth and differentiation
during larval development. By the second larval instar,
the anterior part, which will give rise to the antenna,
510
MARKUS FRIEDRICH
FIG. 2. Different stages of compound eye morphogenesis in Drosophila, Tribolium and Schistocerca. A. Lateral view of Drosophila stage
13 embryonic head according to Chang et al. (2001). B. Lateral view of Tribolium embryonic head at germband retraction stage according to
Ullmann (1966) and Heming (1982). C. Lateral view of Schistocerca embryonic head at 40% of development. D. Third instar Drosophila
antennal-eye imaginal disc. E. Lateral view of one day old Tribolium pupal head. F. Lateral view of Schistocerca first nymphal instar head.
Arrows indicate position of the morphogenetic furrow. Ommatidial preclusters represented by grey circles. ant 5 antenna, cly 5 clypeolabrum,
eyd 5 eye disc, eyl 5 eye lobe, eyp 5 eye placode, ilo 5 inner optic lobe (lobula), ley 5 larval eye, man 5 mandible, max 5 maxilla, lab
5 labium, olo 5 outer optic lobe (medulla and lamina), pro 5 protocerebrum; anterior is left and dorsal is up.
can be morphologically discriminated from the posterior part, which will form the retina and additional
head cuticle areas. Close to middle of the third and
last larval instar, retina differentiation starts at the posterior margin of the eye disc. This process is marked
by formation of the morphogenetic furrow, which refers to the front of differentiation where cells are uniformly shorter and distally constricted forming a conspicuous indentation along the dorsoventral axis of the
eye disc. The furrow moves from its posterior start
point towards anterior (Fig. 2D). Posterior to the furrow, ommatidial preclusters emerge in a regular array
anticipating the regularity of the adult retina. Photoreceptor cells join the ommatidial preclusters first, followed by cone cells and pigment cells. This early
phase of retina cell determination and differentiation
continues until the furrow has reached its final destination at about 10 hours after pupation when the eye
disc everts. Once the entire eye field has been established, the retina cells undergo terminal differentiation
in a concerted manner. Hallmarks of this process are
the elimination of surplus cells by apoptosis, synthesis
of screening pigments and elaboration of the photoreceptor cell rhabdomeres. Approximately two thirds
of the posterior eye disc proper differentiates as retina
while the anterior third and parts of the peripodial
membrane, which is the second tissue layer of the saclike disc, develop into adjacent head cuticle elements
(Haynie and Bryant, 1986).
Eye imaginal disc formation as seen in Drosophila
is not an obligatory feature of postembryonic eye development in holometabolous insects. More basal dipteran and holometabolous species develop through a
eucephalic larva, which carries a fully developed head
capsule equipped with antennal and mouthpart appendages. The larva of the flour beetle Tribolium castaneum exemplifies this level of evolutionary organization (Fig. 1). In this case, metamorphosis is significantly less dramatic. The differentiation of the adult
body plan proceeds through reinitiated growth and terminal differentiation of larval structures, which from
this perspective function as adult organ primordia. The
adult antenna for instance develops via dramatic
growth and further differentiation of the larval antenna. The adult Tribolium retina, however, is not formed
by further differentiation of the larval eyes. It seems
to develop de novo in a region of the late larval head,
which corresponds to the future field of the retina in
the adult head capsule. This ectodermal tissue compartment has been termed ‘‘eye placode’’ (Fig. 2E)
EVOLUTION
OF INSECT
(Marshall, 1928; Friedrich et al., 1996; Friedrich and
Benzer, 2000). The beginning of Tribolium retina differentiation is marked by the initiation of the morphogenetic furrow at the posterior margin of the eye placode. The morphogenetic furrow passes through the eye
placode in anterior direction, coming to halt shortly
before the antenna. The timing of Tribolium retina differentiation corresponds very closely to that in Drosophila (Fig. 1). The progression of the morphogenetic
furrow starts during the second half of the last larval
instar and continuous through approximately the first
third of pupation after which terminal differentiation
initiates. It is important to note that the Tribolium eye
placode area becomes fully elaborated in the embryo
together with all other essential compartments of the
adult head (Fig. 2B). In the more primitive Tribolium,
adult head compartment and for the most part organ
primordia formation have remained embryonic. It is
separated in time from the postembryonic differentiation of the adult retina. This marks a fundamental difference to the continuous postembryonic morphogenesis of the Drosophila head from internalized imaginal
discs, and the integration of Drosophila retina differentiation into this derived process.
Eye development in hemimetabolous insects such as
the grasshopper Schistocerca americana is a repetitive
multi-step process which starts in the embryo and extends through postembryogenesis (Fig. 1) (Anderson,
1978; Friedrich and Benzer, 2000). Almost one third
of the adult Schistocerca compound eye retina is of
embryonic origin. This embryonic fraction is formed
from the ectoderm of the lateral-most tissue compartments of the embryonic head, the eye lobes (Fig. 2C).
The morphogenetic furrow initiates in the posterior
margin of the eye lobe ectoderm shortly before 35%
of embryogenesis. With the morphogenetic furrow
progressing anteriorly, the embryonic retina continues
to extend in anterior direction throughout much of the
second half of embryogenesis, which however is also
characterized by the terminal differentiation of the retina. Most of the eye field of the first instar nymph is
thus fully differentiated and functional. Except for the
anterior margin where cell proliferation and differentiation continuously reinitiate between the postembryonic growth molts thereby enlarging the embryonic
retina field to the size of the adult retina in several
postembryonic increments (Figs. 1 and 2F) (Anderson,
1978).
The evolution of postembryonic eye development
must have required the elaboration of mechanisms
which prevent the onset of adult retina differentiation
in the embryo, maintain the prospective retina field
during early postembryogenesis, and coordinate the
postembryogenetic differentiation of the adult retina
with the complex process of metamorphosis. These
changes of development represent ground state changes as they are prerequisite for the postembryonic development of the compound eye retina in holometabolous species in general. Additional modifications
must have evolved in the lineage leading to Drosoph-
EYE DEVELOPMENT
511
ila, which, as indicated, concerned the postembryonic
development of the prospective retinal field and the
coordination of this process with the de novo development of the other adult head primordia in the eyeantennal imaginal disc. Yet further modifications of
embryonic visual system development are likely to
have been enforced in Drosophila by the evolution of
its extreme mode of short germ development in combination with the evolution of the acephalic larval head
morphology (Melzer and Paulus, 1989). It is thus not
surprising that the molecular control of Drosophila eye
development differs from that in more primitive species at numerous steps preceding the final differentiation of the eye field.
PATTERNING OF THE EMBRYONIC VISUAL ANLAGE
Specification of the precursor embryonic tissue
which gives rise to the visual system is the first step
in the sequence of events leading to the formation of
the insect eye. In Drosophila, all visual system components map to a single unpaired primordium straddling the dorsal midline in the anterior head region of
the blastoderm embryo, which is traditionally considered the nonsegmental tip of the embryo (acron) and
in addition includes the anlagen for the first brain neuromer, the protocerebrum (Green et al., 1993; Dumstrei et al., 1998; Namba and Minden, 1999). The Drosophila visual primordium is molecularly defined by
expression of the homeobox gene so (Cheyette et al.,
1994; Serikaku and O’Tousa, 1994; Chang et al.,
2001). Homologs of so (Six3/6) are also involved in
specifying the visual primordium in other animal phyla
(Oliver et al., 1995; Jean et al., 1999). While the transcriptional code for Drosophila visual primordium
specification seems highly conserved, the mechanisms
responsible for its initial spatial regulation may be of
more recent evolutionary origin. The expression of so
in the visual primordium depends on activation by the
TGF-b signaling factor decapentaplegic (dpp), which
is expressed in a dorsoventral gradient in the early
Drosophila blastoderm embryo (Chang et al., 2001).
Maximal Dpp levels in the dorsal midline of the embryo initiate expression of the HOX-3 orthologous
transcription factor zerknuellt (zen) which represses so
expression along the dorsal midline (Fig. 3). At the
same time, lower Dpp levels continue to activate so
expression along the lateral sides of the embryo. It has
been noted that the resulting split of the visual anlage
resembles the morphogenesis of the vertebrate anterior
plate (Chang et al., 2001). In this case, however, homologs of the signaling factor gene hh, induce the median split of the anterior brain anlage (Ekker et al.,
1995).
A possible explanation for this discrepancy is that
the long germ insect Drosophila exhibits considerable
differences in the formation of the embryonic anlagen
compared to the ancestral mode of short germ development. This raises the possibility that evolutionarily
derived anterior head patterning mechanisms in Drosophila replaced ancestral mechanisms. Indeed, two
512
MARKUS FRIEDRICH
RETINAL FATE COMMITMENT (I): EYE
FIG. 3. Derived role of Drosophila dpp and zen in visual primordium patterning. A. Schematic dorsal view of visual system compartments in the anterior head of the Drosophila embryo early extended germband stage. Zen expression (grey) is induced by Dpp
along the dorsal midline. Visual primordium specifying genes (black
area) are repressed by Zen and activated by Dpp at a distance (after
Chang et al., 2001). B. Schematic dorsal view of early heart-stage
germband of Schistocerca gregaria. Dpp and Zen (grey) are coexpressed in the necklace cells, which mark the boundary between
amnion and serosa cells (after Dearden and Akam, 2001). Stomodeal
opening indicated by circle positioned on midline. Visual primordium indicated by black areas; Anterior is up.
lines of evidence from anterior patterning gene expression in primitive short germ insects such as Tribolium and Schistocerca suggest that midline patterning of the Drosophila head is controlled by derived
mechanisms (Fig. 3). In both of these species zen is
expressed in the extra-embryonic serosa and amnionic
membrane tissue but not in the germband proper (Falciani et al., 1996; Dearden et al., 2000). The peripheral
location of this expression domain excludes an involvement of zen in patterning the embryonic midline
of these species. Furthermore, the expression of dpp
overlaps with that of zen in the serosa and amnionic
membrane of Tribolium and Schistocerca. This suggests that the activation of zen by dpp is conserved
during extraembryonic tissue specification (SanchezSalazar et al., 1996; Dearden and Akam, 2001). The
spatial regulation of its expression pattern however
makes it seem unlikely that dpp is involved in midline
patterning using alternative transcriptional mediators.
Second, genes involved in patterning the anterior
Drosophila head such as tailless (tll), orthodenticle
(otd), and the segmentation gene wingless (wg) are
typically expressed in circumferential domains straddling the dorsal ectoderm midline like so (Nagy and
Carroll, 1994; Chang et al., 2001). With the onset of
gastrulation, these circumferential stripes break up into
isolated expression elements. In most cases, expression
ceases in the dorsal midline reminiscent of dorsal midline so repression by dpp/zen. In Tribolium however
the earliest detectable blastoderm expression patterns
of wg, tll, and otd are already split in the dorsal midline (Nagy and Carroll, 1994; Li et al., 1996; Schroder
et al., 2000). The same is true for the expression of
wg in the orthopteran species Acheta domesticus and
Schistocerca gregaria (Niwa et al., 2000; Dearden and
Akam, 2001). Taken together, these data support the
existence of ancestral midline patterning mechanisms
that might have been strongly modified or lost during
Drosophila evolution.
VERSUS ANTENNA
The Drosophila visual system consists of several
components. The information received by the regular
sensory array of the facetted retina is processed in the
optic neuropil layers lamina, medulla, and lobula,
which together constitute the optic lobe. In addition,
the fruit fly possesses three ocelli, comparatively simple photosensory organs that are centered between the
compound eyes at the dorsal midline of adult head,
and an extra set of simple larval eyes, the Bolwig organs. Except for the ocelli, the tissues of all Drosophila visual system components derive from compartmentalization of the early embryonic visual anlage at
stage 12 of Drosophila embryonic development (Fig.
2A). At the stage of visual primordium partitioning,
the population of cells, which will give rise to the primordium of the adult retina, the eye imaginal disc, is
characterized by specific expression of the master regulatory gene ey (Chang et al., 2001; Kumar and Moses, 2001c). This would suggest early commitment of
the Drosophila eye disc, in line with the fate mapping
supported view that the adult Drosophila retina is determined during embryogenesis (Postlethwait and
Schneiderman, 1971; Younossi-Hartenstein et al.,
1993). However, partitioning of precursor tissue may
not be correlated with terminal commitment for organ
fate. The differentiation of the Drosophila retina is the
endpoint of a series of discrete stages of tissue commitment. This was discovered studying of the role of
Notch (N) and Epidermal growth factor receptor (Egfr)
signaling during early eye-antennal disc development,
which revealed that the posterior part of the eye-antennal imaginal disc remains competent to adopt antennal fate until the second half of the second larval
instar (Fig. 1) (Kumar and Moses, 2001a). At this
stage, the commitment decision of eye versus antenna
fate is under antagonistic control of the two signaling
pathways with N signaling promoting retina fate versus Egfr signaling promoting antenna fate. Consistent
with the experimental evidence for postembryonic determination of the retina field primordium, the regulatory genes essential for Drosophila retina development are not coexpressed in the eye-antennal disc before the second larval instar. Induction of Drosophila
retinal fate results from the concerted action of seven
transcriptional regulators: the recently duplicated Drosophila Pax6 orthologs ey and twin of eyeless (toy),
the related paired domain protein eye gone (eyg), and
the transcription factors dac, eya, so and optix (reviewed in Kumar, 2001; Pappu and Mardon, 2002).
Only toy and eyg are coexpressed with ey in the eyeantennal disc primordium following compartmentalization of the embryonic visual anlage (Kumar and Moses, 2001c). This step controls the specification of the
retinal precursor tissue, which is followed by postembryonic determination. In a consistent manner, also
the delay of antenna primordium determination correlates with absence of the antenna fate specifying
EVOLUTION
OF INSECT
transcription factor distal-less (dll) in the embryonic
antennal anlage (Kumar and Moses, 2001a).
From a comparative perspective it is important to
note that the absence of adult retina and antenna differentiation in the Drosophila embryo correlates with
the lack of expression of transcription factors, which
are essential for the determination of the retina and
antenna primordia. In non-holometabolous insects
such as Schistocerca, the primordia of both the adult
antenna and eye are formed during embryonic development (Fig. 2C). This implies that all of the required
determination genes need to be expressed already in
the respective embryonic anlagen. Transcription factors, which are known to function in Drosophila antenna determination such as dll, spalt-major (salm) and
extradenticle (exd) (Dong et al., 2002), are indeed expressed in the embryonic grasshopper antenna (Friedrich, unpublished observation). One prediction from
the Drosophila model is therefore that also the eye
determination genes are coexpressed in the posterior
eye lobe margin of the grasshopper prior to the initiation of retina differentiation. Preliminary analyses of
eya and so expression during grasshopper embryonic
development have yielded results that are consistent
with such scenario (Dong and Friedrich, unpublished).
Hypothesis building regarding the onset of eye determination transcription factor coexpression during
visual system development in Tribolium is less
straightforward. Two scenarios are conceivable. Evidently, the primordia of the adult antennae are formed
during embryogenesis in the form of the larval antennae, which also holds for most other head structures
in the eucephalic Tribolium (Fig. 2B). By analogy, the
compound eye retina primordium, i.e., the eye placode,
may also be determined in the embryo but prevented
from initiating differentiation. The determination step
may involve coexpression of the retinal determination
genes in the embryonic retinal primordium. However,
morphological or molecular evidence for embryonic
determination of the Tribolium eye placode is yet
missing. Alternatively, the shift of retina determination
network gene coexpression into postembryogenesis
seen in Drosophila may represent a mechanism, which
evolved early in holometabolous insects to preclude
embryonic differentiation of adult retina. For this hypothesis to hold true, the members of the Tribolium
eye specification transcription factor network should
not be coexpressed during embryonic development but
in the eye placode during postembryonic development
(Fig. 1). The comprehensive analysis of eye determination gene expression in Tribolium and Schistocerca
will provide the data necessary for the correct evolutionary interpretation of postembryonic retina determination in Drosophila.
AXIS AND COMPARTMENT SPECIFICATION
The Drosophila eye-antennal imaginal disc proceeds through continuous patterning and growth during larval development. Fundamental patterning steps
concern the establishment of the anteroposterior and
EYE DEVELOPMENT
513
FIG. 4. Expression of wg and fng during dorsoventral compartment
formation in the early Drosophila eye disc. A. First instar eye-antennal imaginal disc. B. Third instar eye disc. Area filled with diagonal grey bars rising from left to right indicates fng expression
domain. Area filled with diagonal grey bars rising from right to left
indicates wg expression domain. Orientation of ommatidial clusters
after rotation indicated by arrows. Position of morphogenetic furrow
indicated by lateral indentations. Dorsal is up and anterior is left.
dorsoventral axes (Lee and Treisman, 2001). In addition, the retina field is divided into dorsal and ventral
compartments, which meet exactly along the midline
of the disc forming a border that has been termed the
equator. In developmental terms, the equator is an N
signaling based organizing center, which stimulates
growth of the eye disc prior to the onset of retina differentiation, and provides signals necessary for planar
cell polarity patterning in the differentiating retina
(Cho and Choi, 1998; Dominguez and de Celis, 1998;
Papayannopoulos et al., 1998; Reifegerste and Moses,
1999; Cho et al., 2000). Morphologically, the Drosophila equator becomes manifest by a 90 degree rotation of the developing ommatidia, which occurs in
mirror image orientation in the dorsal and ventral half
of the retina field (Fig. 4B) (Dietrich, 1909). The dorsoventral compartmentalization of the Drosophila retina field also affects early patterning and growth of the
eye disc. Genetic ablation of the Lobe gene for instance leads to ventral specific loss of retina growth
(Chern and Choi, 2002). Loss of homothorax (hth)
gene activity leads to ectopic retina differentiation in
the ventral but not dorsal compartment of the eye disc
(Pichaud and Casares, 2000). The zinc finger transcription factor teashirt (tsh) represses retina differentiation
in the ventral compartment but promotes retina differentiation in the dorsal compartment (Singh et al.,
2002).
The occurrence of similar dorsoventral compartment
pattern elements in the retina of other arthropod species as well as the general need for tissue growth in
the developing retina would lead one to expect that
formation of the equatorial organizing center by dorsoventral compartment formation is a conserved aspect
of Drosophila eye disc development (Friedrich et al.,
1996). This hypothesis can be tested as the gene network involved in dorsoventral compartment formation
is well understood (Cho and Choi, 1998; Dominguez
and de Celis, 1998; Papayannopoulos et al., 1998; Cho
et al., 2000). One of the first steps involves the dorsal
514
MARKUS FRIEDRICH
compartment specific activation of wg expression by
the transcription factor pannier (pnr) (Maurel-Zaffran
and Treisman, 2000). Wg in turn activates expression
of the homeobox transcription factor mirror (mirr) the
function of which is to suppress the expression of
fringe (fng) in the dorsal part of the disc (Yang et al.,
1999). As a consequence, fng expression is restricted
to the ventral compartment. This leads to specific activation of N signaling along the midline where fng
enhances the activation of N by its ligand Delta (Dl)
while suppressing N activation by the second N ligand
Serrate (Ser) in the ventral part of the disc. In summary, dorsal expression of pnr, wg and mirr, versus
ventral expression of fng are essential elements of dorsoventral compartment formation in the Drosophila
eye disc (Fig. 4).
Interestingly, wg is not expressed in a dorsalized
manner throughout development of the early grasshopper embryonic retina although the polar expression of
wg at the anterior retina margin is conserved (Fig. 6)
(Friedrich and Benzer, 2000). Complementary to this,
the grasshopper eye lobe also lacks ventrally restricted
expression of fng while its expression in and anterior
to the furrow is conserved (Dong and Friedrich, unpublished; Dearden and Akam, 2000). These results
suggest the absence of an N signaling based equatorial
organizing center in the grasshopper retina. Although
surprising at first glance, this is consistent with the
lack of evidence for planar cell polarity patterning in
the grasshopper retina (Wilson et al., 1978). Furthermore, while the N pathway is an essential upstream
growth activator in the Drosophila eye disc, many additional signaling factors stimulate proliferation in the
Drosophila eye disc such as Dpp, Egfr and Wg (Burke
and Basler, 1996; Halfar et al., 2001; Lee and Treisman, 2001). It is therefore conceivable that different
signaling pathways support retina growth in grasshopper without participation of N. To determine if the N
signaling based dorsoventral compartment formation is
a derived aspect of Drosophila eye imaginal disc development it will be necessary to analyze the relevant
genes in other insect species with dorsoventral specific
pattern elements. Shared genetic mechanisms would
indicate the function of an evolutionarily conserved
patterning mechanism that was lost during the evolution leading to grasshopper. Lack of dorsoventral compartment specific expression of wg and fng on the other hand would suggest that dorsoventral patterning
mechanisms evolved multiple times independently.
While the mechanisms related to dorsoventral patterning are suspect of evolutionary change, establishment of the anterior posterior axis in the Drosophila
eye disc is likely to follow ancient paths. The primary
determinant of the anteroposterior axis is the expression of wg, which changes from dorsal specific expression in the second larval instar to a pair of dorsal
and ventral domains at the anterior margin of the eye
field with beginning of the third instar (Fig. 4). As Wg
inhibits furrow initiation, movement and neuronal differentiation, the initiation of retina differentiation is
forced to occur at maximal distance to these expression domains at the midline of the posterior eye lobe
margin (Ma and Moses, 1995; Treisman and Rubin,
1995). Furthermore, ectopic activation of Wg signaling
can enforce anterior compartment identity to the posterior of the retinal eye field (Lee and Treisman, 2001).
Polar expression of wg in front of the retina field consistent with this patterning function was found conserved in several arthropod species ranging from
Schistocerca and Tribolium to crustacean and myriapod species (Friedrich and Benzer, 2000; Niwa et al.,
2000; Duman-Scheel et al., 2002; Hughes and Kaufman, 2002).
INITIATION OF DIFFERENTIATION:
WITH OR WITHOUT DPP?
The initiation of retina differentiation in the Drosophila eye disc has been genetically dissected into
two discrete phases (Kumar and Moses, 2001b). Phase
one or ‘‘furrow birth,’’ which involves the primordial
activation of morphogenetic furrow formation and
photoreceptor differentiation at the posterior eye disc
margin, requires activity of the N, EgfR and Hh signaling pathways (Dominguez and Hafen, 1997; Borod
and Heberlein, 1998; Kumar and Moses, 2001b). The
second phase begins with the continuous initiation of
differentiation along the disc margins necessary for
lateral increase of the eye field size. This process,
dubbed ‘‘reincarnation,’’ depends on input from the
Egfr and Dpp signaling pathways (Chanut and Heberlein, 1997; Kumar et al., 1998; Curtiss and Mlodzik,
2000). In the grasshopper eye lobe, furrow initiation
occurs also at the posterior margin centered at the midline followed by anterior progression of the furrow and
lateral extension of the eye field (Fig. 2C). Considering
the conserved spatial dynamics of retina differentiation
it seems likely that the same regulatory mechanisms
are at work in the grasshopper eye. However, comparative expression pattern analyses of the key player dpp
indicate partial evolutionary divergence at the molecular regulatory level. In Drosophila, dpp is strongly
expressed along the posterior and lateral disc margins
before furrow birth and continues to be expressed
along the lateral eye disc margins during furrow reincarnation (Blackman et al., 1991; Cho et al., 2000).
This is consistent with functional evidence for Dpp
involvement in furrow initiation (Blackman et al.,
1991; Chanut and Heberlein, 1997; Pignoni and Zipursky, 1997). In situ hybridization experiments in
Schistocerca failed to detect expression of dpp in the
posterior and lateral grasshopper eye lobe ectoderm
margins prior to and following furrow initiation, which
argues against a role of dpp in both phases of grasshopper furrow initiation (Friedrich and Benzer, 2000).
It has been proposed that the underlying cause for
this discrepancy between grasshopper and Drosophila
eye development may be the divergence of mechanisms that control the expression of wg in the eye field
(Friedrich and Benzer, 2000). In Drosophila, dpp is
essential for transforming the dorsal compartment spe-
EVOLUTION
OF INSECT
cific expression of wg into the anterior polar domains.
The dorsal expression domain of the morphogenetic
furrow inhibitor wg extends initially to the posterior
margin of the disc during the dorsoventral compartment patterning phase (Fig. 4A) (Cho et al., 2000).
During the late second larval instar, wg becomes repressed at the posterior margin by dpp. In the third
instar eye disc, dpp continues to be required for repression of wg transcription in both the dorsal and ventral margin (Royet and Finkelstein, 1997; Cho et al.,
2000). A similar wg expression pattern change is not
observed during grasshopper eye lobe development
where wg is expressed in anterior polar domains from
very early on (Friedrich and Benzer, 2000). The wg
and dpp expression patterns described in the developing Tribolium eye lobes match that in the Schistocerca eye lobes (Nagy and Carroll, 1994; SanchezSalazar et al., 1996). The lack of dpp expression in
the grasshopper eye lobes prior to the initiation of the
morphogenetic furrow thus correlates with a fundamental difference regarding the emergence of the conserved expression of wg in the anterior eye field of
primitive insects. It is therefore possible that the requirement of dpp for furrow initiation along the lateral
margins of the Drosophila eye disc evolved in conjunction with its apparently derived role of suppressing
wg from these regions.
Recent studies however demonstrated that dpp promotes morphogenetic furrow initiation by activating
the expression of retina determination genes such as
eya (Curtiss and Mlodzik, 2000). Provided the respective Dpp signal is not secreted by extraretinal tissues
in the grasshopper (see below), or replaced by related
TGF-ß related signal transduction pathways, the differences between the regulatory networks controlling
furrow initiation in Drosophila and Schistocerca might
even be deeper.
PROGRESSION OF DIFFERENTIATION
Once initiated, the furrow moves from posterior to
anterior laying out the regular array of developing ommatidial precursor clusters in the posterior eye field.
In Drosophila, the progression of this dynamic differentiation border is maintained by the combined action
of at least three signal transduction pathways, which
coordinate the transcriptional control of retina differentiation (Bessa et al., 2002) (Fig. 5). The transcription
factors Hth, Tsh and Ey, which are coexpressed in the
anterior disc region, have been proposed to interact
physically and keep the anterior eye disc tissue in an
undifferentiated growth state (Bessa et al., 2002). The
long-range signaling factor Dpp, expressed in the furrow, instructs cells far anterior to the furrow to advance from this state into a preproneural state (PPN),
which is characterized by repression of retina differentiation antagonist hth and expression of the proneural transcription factor daughterless (da). Precocious neural differentiation in the PPN zone is prevented by the helix-loop-helix transcription factor
hairy (h), which is also activated by dpp. Exit from
EYE DEVELOPMENT
515
FIG. 5. Signaling and transcription factor domains along the anteroposterior axis of the differentiating retinal field in Drosophila,
Tribolium and Schistocerca. Empty bars: Signaling factors. Bright
bars: Furrow progression activating transcription factors. Black bars:
Furrow progression inhibiting transcription factors. I: Anterior most
eye disc region developing into head cuticle. II: Anterior eye disc
region developing into retina. PPN: Preproneural zone. MF: morphogenetic furrow. IV: Differentiating retina.
the PPN state is induced via activation of the N signaling pathway by the short-range factor Dl, which is
also expressed in the furrow. This leads to repression
of the neuronal specification antagonists extramachrochaete (emc) and h, allowing for expression of ato
in the furrow. The initial uniform expression of ato in
the furrow refines to R8 photoreceptor cells posterior
to the furrow. This is followed by rapid recruitment of
additional photoreceptors cells to the newly emerging
ommatidia. The positive feedback loop between differentiation and induction of furrow progression is
closed by the expression of Hh in the maturing photoreceptor cells, which diffuses anteriorly and activates
N and dpp transcription in the furrow.
As the basic cellular morphology of the morphogenetic furrow is conserved in diverse arthropods there
is little reason to suspect evolutionary divergence of
the regulatory network driving furrow progression.
516
MARKUS FRIEDRICH
This is furthermore suggested by the strikingly similar
involvement of hh and ato in vertebrate eye development indicating evolutionary conservation at a deep
phylogenetic level (Neumann and Nuesslein-Volhard,
2000; Brown et al., 2001). Nonetheless, comparative
expression analysis in the developing retina of Schistocerca and Tribolium suggests partial evolutionary diversification (Fig. 5). In Drosophila dpp begins to
clear from the posterior margin and to increase expression in the morphogenetic furrow after furrow initiation (Blackman et al., 1991). In Schistocerca dpp is
expressed at low levels throughout the anterior eye
field ectoderm in addition to an expression domain
posterior to the furrow, which is associated with the
area of the eye lobe retina where the photoreceptor
axons project from the retina into the lamina (Friedrich
and Benzer, 2000). This finding has implications regarding possible mechanisms involved in grasshopper
furrow progression by comparison to Drosophila.
Grasshopper dpp may either instruct the entire anterior
retina field to enter the PPN state or carry out an entirely different function. Ongoing studies in our lab
show that the homeodomain transcription factor exd,
the essential dimerization partner of hth, is expressed
neither in the grasshopper eye nor in regions immediately adjacent to it. This implies the absence of hth
mediated transcriptional control in the anterior of the
grasshopper eye lobe (Dong and Friedrich, unpublished). The lack of a Drosophila PPN related transcriptional switch of hth expression in the grasshopper
eye field points to the possibility that the furrow progression promoting function of dpp is not conserved.
Interestingly, Drosophila dpp is not essential for furrow progression as hh on its own can induce the PPN
state although at a slower pace (Burke and Basler,
1996; Greenwood and Struhl, 1999). It has been proposed that the expression of dpp in the Drosophila
morphogenetic furrow evolved to accelerate furrow
progression in this fast developing species (Friedrich
and Benzer, 2000; Bessa et al., 2002). This idea was
based on the assumption of an otherwise largely conserved network of transcriptional control of retina differentiation. The correlated lack of hth/exd in the
grasshopper eye lobe however raises the possibility
that the divergence of dpp expression is connected to
more fundamental differences in the early control of
retina differentiation. This may be related to the evolutionarily derived dynamic determination of head versus retina fate in the Drosophila eye disc.
RETINAL FATE COMMITMENT (II):
EYE VERSUS HEAD CUTICLE
Coexpression of the seven essential eye specification master genes during the second larval instar poises cells in the Drosophila eye disc to adopt retinal fate
(Kumar and Moses, 2001a). However, only approximately two thirds of the posterior disc will differentiate into retina while cells in the anterior margin give
rise to adjacent head cuticle elements (Haynie and
Bryant, 1986). Consistent with this, the domain of eye
specification transcription factor coexpression is centered in the posterior of the eye disc and does not
extend through the anterior disc (Bessa et al., 2002).
At the time of furrow progression, for instance, the
expression of dac and eya does not extend more anteriorly than to the anterior border of the PPN zone
due to hth mediated repression. The expression domain
of hth on the other hand which initially covers the
entire early eye-antennal imaginal disc becomes progressively restricted to the posterior eye disc due to
repression by Dpp emanating from the moving furrow
(Pichaud and Casares, 2000; Bessa et al., 2002). There
are thus two commitment stages through which cells
of the eye disc pass. During the second larval instar
cells become committed to participating in retina over
antenna associated head compartment formation. During the third larval instar, cells adopt retina fate if they
fall within the reach of the morphogenetic furrow
while developing into head cuticle components otherwise. At this stage, retina fate is promoted in the posterior disc by the combined action of the dpp, hh, N
and Egfr signaling pathways. The specification of head
cuticle fate is in the realm of the anterior polar expression domains of wg, which besides activating hth
expression in the anterior eye disc field, is a negative
regulator of photoreceptor differentiation itself (Ma
and Moses, 1995; Treisman and Rubin, 1995; Pichaud
and Casares, 2000). In addition, there is evidence that
Wg signaling can induce head cuticle morphogenesis
as opposed to retina differentiation (Royet and Finkelstein, 1997; Baonza and Freeman, 2002). In summary, wg is not only essential for determining the anterior border to the progression furrow but also for the
morphogenesis of major parts of the dorsal (vertex)
and ventral (gena) Drosophila head. Consistent with
this, the respective eye disc anlagen reside within the
expression domains of hth and wg (Pichaud and Casares, 2000).
Although the expression of wg in two polar domains
anterior to the developing retina is highly conserved,
the fate map of wg expressing cells in the developing
grasshopper head shows significant differences with
that in the Drosophila eye disc (Fig. 6). Already the
early lateral wg expression domains reside within the
anterior eye field of the eye lobe but do not extend
into areas outside the eye lobes. The discrepancy is
most obvious with regards to the dorsal head regions,
which give rise to the lateral and median ocelli. These
lie within the wg expression domains in Drosophila
but are remote from the wg expression domains in the
grasshopper eye lobe. Further support for evolutionary
divergence of wg related patterning of the adult Drosophila head is indicated by the lack of Wg signaling
downstream target gene expression in the grasshopper
head. The transcription factor engrailed, which is activated by wg via otd, is required for ocelli formation
in Drosophila but not expressed in the differentiating
ocelli of the grasshopper embryo (unpublished observation) (Royet and Finkelstein, 1995). The expression
of wg in the grasshopper eye lobe is thus compatible
EVOLUTION
OF INSECT
EYE DEVELOPMENT
517
FIG. 6. Comparison of Drosophila and Schistocerca dorsal head fate map. Left: Left half of grasshopper embryonic head at 40% of development. Middle: Morphology of generalized dorsal insect head. Right: Drosophila third instar antennal-eye imaginal disc. Grey shaded areas:
wg expression domains. Homologous ocelli indicated by arrows. ant 5 antennal disc, eyd 5 eye disc, eyl 5 eye lobe, fro 5 frons, ver 5
vertex.
with a role in negatively regulating furrow progression
and thereby organizing the anterior border of the retina, but not with patterning adjacent head cuticle regions to the extent it is occurring in Drosophila. This
difference in wg patterning functions correlates with
the fact that partitioning of head versus retina field by
formation of the eye lobes has been completed before
furrow initiation during grasshopper embryonic head
development. The integration of retina differentiation
and adult head cuticle partitioning under antagonistic
control by wg and dpp represents a derived aspect of
Drosophila eye disc development.
CONTROL AT A DISTANCE (I):
EXTRARETINAL SIGNALING SOURCES
The antagonism of anterior head cuticle versus retina differentiation is only one example for the coordination of retina differentiation with adult head development in the Drosophila eye-antennal imaginal
disc. To embrace the entire scope of this patterning
aspect, it is important to recall the double layer nature
of the sac-like eye-antennal imaginal disc. The apical
surface of the disc proper, the posterior compartment
of which forms the retina, is overlaid by the peripodial
membrane, which contributes a considerable part of
the posterior dorsal head cuticle (Haynie and Bryant,
1986). There is thus not only a need to coordinate
growth and differentiation between the anterior and
posterior pole of the disc proper but also between the
peripodial membrane and the disc proper. Recent studies have provided evidence of regulatory communication between these tissue layers by exchange of signaling factors across the lumen of the disc. Hh, Dpp
and Wg for instance are expressed in the peripodial
membrane of the early eye-antennal disc from where
they instruct dorsoventral compartment formation in
the disc proper (Cho et al., 2000). Similarly, Fng and
Ser have been reported to be expressed in the peripodial membrane instead of the disc proper in the context of dorsoventral compartment formation (Gibson
and Schubiger, 2000). Information exchange seems to
proceed in both directions as Dpp expressed in the disc
proper furrow has been reported to maintain cell survival in the peripodial membrane (Gibson et al., 2002).
Although results from these studies differ in some detail, Ser for instance has also been described as disc
proper located target of thh signaling (Cho et al.,
2000), the evidence for signal exchange between disc
proper and peripodial membrane is overall consistent
and compelling. Active and passive transport mechanisms have been proposed to be involved in the exchange of signaling molecules between the two tissues.
Dpp, expressed in the furrow of the disc proper has
been found to accumulate in the disc lumen (Gibson
et al., 2002). Cellular processes reaching from the peripodial membrane to the disc proper have been proposed to transport Hh, Wg and Dpp protein (Cho et
al., 2000; Gibson and Schubiger, 2000).
From an evolutionary perspective, the regulatory interactions between peripodial membrane and disc
proper represent a derived patterning aspect of the
Drosophila eye-antennal disc. In the more primitive
Tribolium eye placode no peripodial membrane is
formed and the differentiating retina cells contact the
cuticle of the larval head (Friedrich et al., 1996). There
is thus no candidate non-retinal signaling source facing
the Tribolium eye placode. Likewise, no peripodial
membrane equivalent tissue exists in the developing
grasshopper embryonic head. Nonetheless, extraembryonic tissues of the grasshopper embryo could serve
as analogous extraretinal patterning sources. From gastrulation to about 35% of development, the amnionic
membrane covers most of the grasshopper embryo
reaching from the dorsal margins ventrally. The eye
lobe ectoderm in particular develops in contact with
the amnionic membrane (Friedrich and Benzer, 2000).
Interestingly, dpp is expressed in the dorsal edges of
the grasshopper amnionic membrane from where it
could diffuse to the eye lobes (personal observation).
The amnionic membrane will therefore have to be considered as potential signaling source in future analyses
of eye development in the grasshopper and non-holometabolous insects in general.
CONTROL AT A DISTANCE (II):
HORMONAL REGULATION OF RETINA DIFFERENTIATION
In addition to local signals hormonal instructions
affect the progression of the Drosophila morphoge-
518
MARKUS FRIEDRICH
netic furrow from a distance. During postembryogenesis, ecdysone is secreted from the larval ring glands
and metabolized into the hormone 20-hydroxyecdysone (20E) in peripheral tissues. Each of the three larval molts in Drosophila is associated with a transient
peak in 20E concentration, which instructs the epidermal cells to secrete new cuticle. These 20 levels peaks
occur in the presence of a second hormonal mediator,
juvenil hormone (JH), which preserves the growth
character of the molts. JH levels drop during the last
larval instar and a moderate ecdysone peak induces the
larva to stop food uptake and to prepare for pupation.
This stage is the wandering stage during which retina
differentiation is initiated. The next ecdysone peak,
which induces the pupal molt, is accompagnied by rising JH levels. Juvenil hormone levels drop in the early
pupa. A final pupal 20E level peak in the absence of
JH initiates the terminal differentiation of adult structures including the retina (for review see Riddiford,
1993). Molecular genetic analyses have revealed that
the progression of the morphogenetic furrow and the
early differentiation of photoreceptor clusters are sensitive to 20E levels as well as to the presence of components of the ecdysteroid signal transduction machinery. Conditional genetic ecdysteroid depletion in D.
melanogaster causes stall of furrow progression, incorrect initiation of R8 founder cells, and loss of neuronal differentiation (Brennan et al., 1998). Somewhat
conflicting results have been obtained by in vitro culturing studies in which morphogenetic furrow progression and photoreceptor differentiation could be observed in the absence of ecdysteroid supplements (Li
and Meinertzhagen, 1997). Nonetheless, the involvement of ecdysteroid signaling in the control of furrow
progression has been further substantiated by the finding that the furrow also stalls in tissue mutant for the
ecdysteroid signaling immediate early gene Broad
complex (BR-C) isoform Z2 (Brennan et al., 2001).
The canonical ecdysteroid receptor complex consisting
of the ecdysteroid receptor (EcR) and the RXR homolog Ultraspiracle (Usp) seems not to be involved in
the transduction of the signal as the furrow is not affected in EcR mutant tissue (Brennan et al., 2001).
Loss of Usp however causes acceleration of furrow
progression associated with ectopic expression of the
BR-C isoform protein Z1 anterior to the furrow. The
discrepancy compared to the ecdysteroid depletion
phenotype is explained by the fact that Usp acts as
repressor when unbound, but as transcriptional activator in the ligand bound situation (Zelhof et al., 1997;
Ghbeish and McKeown, 2002). The biological significance of the hormonal control of early retina differentiation in Drosophila remains to be determined. One
possibility is that the hormonal clues are necessary to
align the multitude of organ differentiation events with
each other and the organism’s growth progress. The
need for hormonal regulation of early retina differentiation is obvious for a second holometabolous species,
in which the effect of ecdysteroid levels on retina development has been carefully investigated, the tobacco
hornmoth Manduca sexta (Champlin and Truman,
1998). It is the transient halt of development during
pupal diapause in Manduca, which is controlled by the
reversible block of early retina development at below
threshold ecdysteroid levels. This responsiveness of
the Manduca retina changes during the final pupal ecdysteroid surge, which irreversibly terminates the progression of the morphogenetic furrow and triggers terminal differentation throughout the retina field in a
concerted manner. The final pupal ecdysteroid peak
has the same consequences in Manduca and Drosophila including rhabdomere formation, pigment synthesis
and lens cuticle secretion from cone cells. In both systems, these developmental responses can be triggered
in cultured retinas by application of 20E (Li and Meinertzhagen, 1995; Champlin and Truman, 1998). In
combination, these data suggest that the ecdysone dependence of terminal retina differentiation is an ancestral aspect of postembryonic eye development in
holometabolous insects. The regulatory effect of ecdysteroid signaling on early retina differentiation may
also trace back to the origin of holometabolous insects
but obviously experienced lineage specific modifications.
The shared ecdysteroid signaling dependence of
postembryonic retina differentiation in Manduca and
Drosophila raises the question if the hormonal control
of eye development evolved in the ancestral lineage of
holometabolous insects, or was inherited from hormonal mechanisms already involved in the control of
retina differentiation in primitive insects. Ecdysteroid
level peaks have long been known to induce molting
both during embryogenesis and postembryogenesis in
hemimetabolous insects including orthopterans such as
Schistocerca (Lagueux et al., 1979). Less information
existed regarding possible involvement in additional
differentiation processes. Cell proliferation at the postembryonic morphogenetic furrow in Schistocerca has
been reported to reach highest levels between molts
while being silent during molting, which is induced by
ecdysteroid level peaks (Anderson, 1978). These data
suggest that ecdysteroids may have an inhibiting effect
on furrow progression during grasshopper postembryogenesis. In culturing experiments of embryonic
Schistocerca eye lobes, however, supplementing 20E
stimulated cell proliferation and the rate of morphogenetic furrow progression (Dong et al., 2003). Also
aspects of terminal differentiation such as screening
pigment synthesis were significantly enhanced by 20E
application. At the same time, base levels of furrow
progression, cell proliferation and screening pigment
synthesis could be observed in culture even when ecdysteroid signaling was blocked by application of ecdysteroid antagonist Cucurbitacin B suggesting that
ecdysteroid signaling is not essential for retina differentiation in Schistocerca (Dong et al., 2003). In combination, the available data from Schistocerca indicate
that retinal development is not dependent on but sensitive to ecdysteroid levels both during embryogenesis
and postembryogenesis in primitive insects. The strin-
EVOLUTION
OF INSECT
gent control of early retina differentiation in species
such as Manduca may have evolved by modification
of preexisting mechanisms of hormonal control. These
mechanisms may already have been related to the developmental control of diapause, which is also observed in primitive insects (Tawfik et al., 2002). The
critical dependence of terminal differentiation on elevated ecdysteroid levels during pupation on the other
hand has to be added to the list of derived aspects of
Drosophila eye development when compared to primitive insects.
SUMMARY AND PERSPECTIVES
Given the infancy of molecular genetic studies of
eye development in non-Drosophila insects our understanding of insect eye development evolution is by
necessity still incomplete and preliminary. Nonetheless, the data accumulated so far seem to build a strong
case for the hypothesis that relatively recent evolutionary modifications affected many steps of the molecular developmental control of Drosophila retina formation. Some generalization may be helpful to structure the diversity of observations presented. The evolution of rapid embryogenesis in higher flies via an
extreme form of long germband development affected
the early patterning of the embryonic visual primordium. The restriction of retina differentiation to late
stages of postembryonic development, which separates
holometabolous insects from primitive hemi- and ametabolous insects, involved changes in the timing, and
perhaps logic, of retina primordium determination, and
modifications of the hormonal control of retina development. The evolution of postembryonic adult head
primordium formation from internalized imaginal
discs, which allowed the emergence of the acephalic
larva typical for brachyceran flies, is likely to have
enforced the widest range of developmental modifications. These may concern the control of initiation
and progression of retina differentiation, the developmental communication between retina and non-retina
tissues, and the dynamic partitioning of retina versus
adjacent head cuticle compartments. A comprehensive
analysis of eye development in Schistocerca and Tribolium holds promise of elucidating the apparently
highly eventful evolutionary history of Drosophila visual system development. In addition, it will be necessary to extend the comparison to crustaceans to verify the ancestral status of processes operating in primitive insects by outgroup comparison (Hafner and Tokarski, 1998; Melzer et al., 2000; Harzsch and
Walossek, 2001).
ACKNOWLEDGMENTS
I am grateful to the members of the lab for reading
the manuscript and three anonymous reviewers for
helpful comments. This research was funded by NSF
grants DBI-0070099 and DBI-0091926.
REFERENCES
Anderson, H. 1978. Postembryonic development of the visual system of the locust, Schistocerca gregaria. I. Pattern of growth
EYE DEVELOPMENT
519
and developmental interactions in the retina and optic lobe. J.
Embryol. Exper. Morphol. 45:55–83.
Baonza, A. and M. Freeman. 2002. Control of Drosophila eye specification by Wingless signalling. Development 129:5313–22.
Bate, C. M. 1978. Development of sensory systems in Arthropods.
In M. Jacobson (ed.), Handbook of sensory physiology, pp. 1–
53. Springer Verlag, Heidelberg, New York.
Bessa, J., B. Gebelein, F. Pichaud, F. Casares, and R. S. Mann. 2002.
Combinatorial control of Drosophila eye development by eyeless, homothorax, and teashirt. Genes Dev. 16:2415–27.
Blackman, R. K., M. Sanicola, L. A. Raftery, T. Gillevet, and W.
M. Gelbart. 1991. An extensive 39 cis-regulatory region directs
the imaginal disk expression of decapentaplegic, a member of
the TGF-beta family in Drosophila. Development 111:657–66.
Borod, E. R. and U. Heberlein. 1998. Mutual regulation of decapentaplegic and hedgehog during the initiation of differentiation
in the Drosophila retina. Dev. Biol. 197:187–97.
Brennan, C. A., M. Ashburner, and K. Moses. 1998. Ecdysone pathway is required for furrow progression in the developing Drosophila eye. Development 125:2653–64.
Brennan, C. A., T. R. Li, M. Bender, F. Hsiung, and K. Moses. 2001.
Broad-complex, but not ecdysone receptor, is required for progression of the morphogenetic furrow in the Drosophila eye.
Development 128:1–11.
Brown, N. L., S. Patel, J. Brzezinski, and T. Glaser. 2001. Math5 is
required for retinal ganglion cell and optic nerve formation.
Development 128:2497–508.
Burke, R. and K. Basler. 1996. Hedgehog-dependent patterning in
the Drosophila eye can occur in the absence of Dpp signaling.
Dev. Biol. 179:360–8.
Carroll, S. B. 1994. Developmental regulatory mechanisms in the
evolution of insect diversity. In M. Akam, P. Holland, P. Ingham, and G. Wray (eds.), The evolution of developmental
mechanisms. Development 1994 Supplement, pp. 217–223. The
Company of Biologists Ltd., Cambridge.
Champlin, D. T. and J. W. Truman. 1998. Ecdysteroids govern two
phases of eye development during metamorphosis of the moth,
Manduca sexta. Development 125:2009–2018.
Chang, T., J. Mazotta, K. Dumstrei, A. Dumitrescu, and V. Hartenstein. 2001. Dpp and Hh signaling in the Drosophila embryonic
eye field. Development 128:4691–704.
Chanut, F. and U. Heberlein. 1997. Role of decapentaplegic in initiation and progression of the morphogenetic furrow in the developing Drosophila retina. Development 124:559–67.
Chern, J. J. and K. W. Choi. 2002. Lobe mediates Notch signaling
to control domain-specific growth in the Drosophila eye disc.
Development 129:4005–13.
Cheyette, B. N., P. J. Green, K. Martin, H. Garren, V. Hartenstein,
and S. L. Zipursky. 1994. The Drosophila sine oculis locus
encodes a homeodomain-containing protein required for the development of the entire visual system. Neuron 12:977–96.
Cho, K. O., J. Chern, S. Izaddoost, and K. W. Choi. 2000. Novel
signaling from the peripodial membrane is essential for eye disc
patterning in Drosophila. Cell 103:331–42.
Cho, K. O. and K. W. Choi. 1998. Fringe is essential for mirror
symmetry and morphogenesis in the Drosophila eye. Nature
396:272–6.
Curtiss, J. and M. Mlodzik. 2000. Morphogenetic furrow initiation
and progression during eye development in Drosophila: the
roles of decapentaplegic, hedgehog and eyes absent. Development 127:1325–36.
Dearden, P. and M. Akam. 2000. A role for Fringe in segment morphogenesis but not segment formation in the grasshopper, Schistocerca gregaria. Dev. Genes Evol. 210:329–36.
Dearden, P. K. and M. Akam. 2001. Early embryo patterning in the
grasshopper, Schistocerca gregaria: Wingless, decapentaplegic
and caudal expression. Development 128:3435–44.
Dearden, P., M. Grbic, F. Falciani, and M. Akam. 2000. Maternal
expression and early zygotic regulation of the Hox3/zen gene
in the grasshopper Schistocerca gregaria. Evol. Dev. 2:261–70.
Desplan, C. 1997. Eye development: Governed by a dictator or a
junta? Cell 91:861–864.
520
MARKUS FRIEDRICH
Dietrich, W. 1909. Die Facettenaugen der Dipteren. Z. Wiss. Zool.
92:465–539.
Dominguez, M. and J. F. de Celis. 1998. A dorsal/ventral boundary
established by Notch controls growth and polarity in the Drosophila eye. Nature 396:276–8.
Dominguez, M. and E. Hafen. 1997. Hedgehog directly controls
initiation and propagation of retinal differentiation in the Drosophila eye. Genes Dev. 11:3254–64.
Dong, P. D., J. S. Dicks, and G. Panganiban. 2002. Distal-less and
homothorax regulate multiple targets to pattern the Drosophila
antenna. Development 129:1967–74.
Dong, Y., L. Dinan, and M. Friedrich. 2003. The effect of manipulating ecdysteroid signaling on embryonic eye development in
the locust Schistocerca americana. Dev. Genes Evol. 213:587–
600.
Duman-Scheel, M., N. Pirkl, and N. H. Patel. 2002. Analysis of the
expression pattern of Mysidium columbiae wingless provides
evidence for conserved mesodermal and retinal patterning processes among insects and crustaceans. Dev. Genes Evol. 212:
114–23.
Dumstrei, K., C. Nassif, G. Abboud, A. Aryai, and V. Hartenstein.
1998. EGFR signaling is required for the differentiation and
maintenance of neural progenitors along the dorsal midline of
the Drosophila embryonic head. Development 125:3417–26.
Egelhaaf, A. 1988. Evidence for the priming role of the central retinula cell in ommatidium differentiation of Ephestia kuehniella.
Roux’s Arch. Dev. Biol. 197:184–189.
Ekker, S. C., A. R. Ungar, P. Greenstein, D. P. von Kessler, J. A.
Porter, R. T. Moon, and P. A. Beachy. 1995. Patterning activities
of vertebrate hedgehog proteins in the developing eye and brain.
Curr. Biol. 5:944–55.
Falciani, F., B. Hausdorf, R. Schroder, M. Akam, D. Tautz, R. Denell, and S. Brown. 1996. Class-3 hox genes in insects and the
origin of zen. Proc. Nat. Acad. Sci. U. S. A. 93:8479–8484.
Friedrich, M. and S. Benzer. 2000. Divergent decapentaplegic expression patterns in compound eye development and the evolution of insect metamorphosis. J. Exp. Zool. (Mol. Dev. Evol.)
288:39–55.
Friedrich, M., I. Rambold, and R. R. Melzer. 1996. The early stages
of ommatidial development in the flour beetle Tribolium castaneum (Coleoptera, Tenebrionidae). Dev. Genes Evol. 206:
136–146.
Gehring, W. J. 1996. The master control gene for morphogenesis
and evolution of the eye. Genes Cells 1:11–5.
Ghbeish, N. and M. McKeown. 2002. Analyzing the repressive function of ultraspiracle, the Drosophila RXR, in Drosophila eye
development. Mech. Dev. 111:89–98.
Gibson, M. C., D. A. Lehman, and G. Schubiger. 2002. Lumenal
transmission of decapentaplegic in Drosophila imaginal discs.
Dev. Cell 3:451–60.
Gibson, M. C. and G. Schubiger. 2000. Peripodial cells regulate
proliferation and patterning of Drosophila imaginal discs. Cell
103:343–50.
Green, P., A. Y. Hartenstein, and V. Hartenstein. 1993. The embryonic development of the Drosophila visual system. Cell And
Tissue Research 273:583–598.
Greenwood, S. and G. Struhl. 1999. Progression of the morphogenetic furrow in the Drosophila eye: The roles of Hedgehog,
Decapentaplegic and the Raf pathway. Development 126:5795–
808.
Hafner, G. S. and T. R. Tokarski. 1998. Morphogenesis and pattern
formation in the retina of the crayfish Procambarus clarkii. Cell
Tissue Res. 293:535–50.
Hafner, G. S. and T. R. Tokarski. 2001. Retinal development in the
lobster Homarus americanus. Comparison with compound eyes
of insects and other crustaceans. Cell Tissue Res. 305:147–58.
Halder, G., P. Callaerts, and W. J. Gehring. 1995. New perspectives
on eye evolution. Curr. Opin. Genet. Dev. 5:602–9.
Halfar, K., C. Rommel, H. Stocker, and E. Hafen. 2001. Ras controls
growth, survival and differentiation in the Drosophila eye by
different thresholds of MAP kinase activity. Development 128:
1687–96.
Hanson, I. and V. Van Heyningen. 1995. Pax6: More than meets the
eye. Trends Genet. 11:268–72.
Harzsch, S. and D. Walossek. 2001. Neurogenesis in the developing
visual system of the branchiopod crustacean Triops longicaudatus (LeConte, 1846): corresponding patterns of compoundeye formation in Crustacea and Insecta? Dev. Genes Evol. 211:
37–43.
Haynie, J. L. and P. J. Bryant. 1986. Development of the eye-antenna
imaginal disc and morphogenesis of the adult head in Drosophila melanogaster. J. Exp. Zool. 237:293–308.
Heming, B. S. 1982. Structure and development of the larval visual
system in embryos of Lytta viridana Leconte (Coleoptera, Meloidae). Journal of Morphology 172:23–43.
Hughes, C. L. and T. C. Kaufman. 2002. Exploring myriapod segmentation: the expression patterns of even-skipped, engrailed,
and wingless in a centipede. Dev. Biol. 247:47–61.
Jean, D., G. Bernier, and P. Gruss. 1999. Six6 (Optx2) is a novel
murine Six3-related homeobox gene that demarcates the presumptive pituitary/hypothalamic axis and the ventral optic stalk.
Mech. Dev. 84:31–40.
Kristensen, N. P. 1995. Forty years’ insect phylogenetics. Zool. Beitr.
N.F. 36:83–124.
Kumar, J. P. 2001. Signalling pathways in Drosophila and vertebrate
retinal development. Nat. Rev. Genet. 2:846–57.
Kumar, J. P. and K. Moses. 2001a. EGF receptor and Notch signaling act upstream of Eyeless/Pax6 to control eye specification.
Cell 104:687–97.
Kumar, J. P. and K. Moses. 2001b. The EGF receptor and Notch
signaling pathways control the initiation of the morphogenetic
furrow during Drosophila eye development. Development 128:
2689–97.
Kumar, J. P. and K. Moses. 2001c. Expression of evolutionarily conserved eye specification genes during Drosophila embryogenesis. Dev. Genes Evol. 211:406–14.
Kumar, J. P., M. Tio, F. Hsiung, S. Akopyan, L. Gabay, R. Seger, B.
Z. Shilo, and K. Moses. 1998. Dissecting the roles of the Drosophila EGF receptor in eye development and MAP kinase activation. Development 125:3875–85.
Lagueux, M., C. Hetru, F. Coltzene, C. Kappler, and J. A. Hoffmann.
1979. Ecdysone titre and metabolism in relation to cuticulogenesis in embryos of Locusta migratoria. J. Insect Physiol. 25:
709–723.
Lee, J. D. and J. E. Treisman. 2001. The role of Wingless signaling
in establishing the anteroposterior and dorsoventral axes of the
eye disc. Development 128:1519–1529.
Li, C. and I. A. Meinertzhagen. 1995. Conditions for the primary
culture of eye imaginal discs from Drosophila melanogaster. J.
Neurobiol. 28:363–80.
Li, C. and I. A. Meinertzhagen. 1997. The effects of 20-hydroxyecdysone on the differentiation in vitro of cells from the eye imaginal disc from Drosophila melanogaster. Invert. Neurosci. 3:
57–69.
Li, Y., S. J. Brown, B. Hausdorf, D. Tautz, R. E. Denell, and R.
Finkelstein. 1996. Two orthodenticle-related genes in the shortgerm beetle Tribolium castaneum. Dev. Genes Evol. 206:35–
45.
Ma, C. and K. Moses. 1995. wingless and patched are negative
regulators of the morphogenetic furrow and can affect tissue
polarity in the developing Drosophila compound eye. Development 121:2279–89.
Marshall, W. S. 1928. The development of the compound eye of the
confused flour beetle, Tribolium castaneum Jacq. Tans. Wis.
Acad. Sci. Arts Lett. 23:611–630.
Maurel-Zaffran, C. and J. E. Treisman. 2000. pannier acts upstream
of wingless to direct dorsal eye disc development in Drosophila.
Development 127:1007–16.
Meinertzhagen, I. A. 1973. Development of compound eye and optic
lobe in insects. In D. Young (ed.), Developmental neurobiology
of arthropods, pp. 51–104. Cambridge University Press, Cambridge.
Melzer, R. R., C. Michalke, and U. Smola. 2000. Walking on insect
paths? Early ommatidial development in the compound eye of
EVOLUTION
OF INSECT
the ancestral crustacean, Triops cancriformis. Naturwissenschaften 87:308–11.
Melzer, R. R. and H. F. Paulus. 1989. Evolutionswege zum Larvalauge der Insekten—Die Stemmata der höheren Dipteren und
ihre Abwandlung zum Bolwig-Organ. Z. Zool. Syst. Evolutionsforsch. 27:200–245.
Melzer, R. R. and H. F. Paulus. 1994. Postlarval development of
compound eyes and stemmata of Chaoborus crystallinus (de
Geer, 1776) (Diptera, Chaoboridae). Stage-specific reconstructions within individual organs of vision. Internat. J. Insect
Morph. Embryol. 23:261–274.
Nagy, L. M. and S. Carroll. 1994. Conservation of wingless patterning functions in the short-germ embryos of Tribolium castaneum. Nature 367:460–463.
Namba, R. and J. S. Minden. 1999. Fate mapping of Drosophila
embryonic mitotic domain 20 reveals that the larval visual system is derived from a subdomain of a few cells. Dev. Biol. 212:
465–76.
Neumann, C. J. and C. Nuesslein-Volhard. 2000. Patterning of the
zebrafish retina by a wave of sonic hedgehog activity. Science
289:2137–9.
Niwa, N., Y. Inoue, A. Nozawa, M. Saito, Y. Misumi, H. Ohuchi,
H. Yoshioka, and S. Noji. 2000. Correlation of diversity of leg
morphology in Gryllus bimaculatus (cricket) with divergence in
dpp expression pattern during leg development. Development
127:4373–4381.
Oliver, G., A. Mailhos, R. Wehr, N. G. Copeland, N. A. Jenkins, and
P. Gruss. 1995. Six3, a murine homologue of the sine oculis
gene, demarcates the most anterior border of the developing
neural plate and is expressed during eye development. Development 121:4045–55.
Papayannopoulos, V., A. Tomlinson, V. M. Panin, C. Rauskolb, and
K. D. Irvine. 1998. Dorsal-ventral signaling in the Drosophila
eye. Science 281:2031–4.
Pappu, K. and G. Mardon. 2002. Retinal specification and determination in Drosophila. In K. Moses (ed.), Drosophila eye development, pp. 5–19. Springer Verlag, Berlin, Heidelberg.
Pichaud, F. and F. Casares. 2000. homothorax and iroquois-C genes
are required for the establishment of territories within the developing eye disc. Mech. Dev. 96:15–25.
Pichaud, F. and C. Desplan. 2002. Pax genes and eye organogenesis.
Curr. Opionion Genet. Dev. 12:430–434.
Pichaud, F., J. Treisman, and C. Desplan. 2001. Reinventing a common strategy for patterning the eye. Cell 105:9–12.
Pignoni, F. and S. L. Zipursky. 1997. Induction of Drosophila eye
development by decapentaplegic. Development 124:271–8.
Postlethwait, J. H. and H. A. Schneiderman. 1971. A clonal analysis
of development in Drosophila melanogaster: Morphogenesis,
determination and and growth in the wild type antenna. Dev.
Biol. 24:477–519.
Quiring, R., U. Walldorf, U. Kloter, and W. J. Gehring. 1994. Homology of the eyeless gene of Drosophila to the small eye gene
in mice and Aniridia in humans. Science 265:785–9.
Reifegerste, R. and K. Moses. 1999. Genetics of epithelial polarity
and pattern in the Drosophila retina. Bioessays 21:275–85.
Riddiford, L. M. 1993. Hormones and Drosophila development. In
M. Bate and A. Martinez-Arias (eds.), The development of Drosophila melanogaster, pp. 899–939. Cold Spring Harbor Laboratory Press, Cold Spring Harbor.
EYE DEVELOPMENT
521
Royet, J. and R. Finkelstein. 1995. Pattern formation in Drosophila
head development: The role of the orthodenticle homeobox
gene. Development 121:3561–72.
Royet, J. and R. Finkelstein. 1997. Establishing primordia in the
Drosophila eye-antennal imaginal disc: The roles of decapentaplegic, wingless and hedgehog. Development 124:4793–800.
Rubin, G. M., M. D. Yandell, J. R. Wortman, G. L. Gabor Miklos,
C. R. Nelson, I. K. Hariharan, M. E. Fortini, P. W. Li, R. Apweiler, W. Fleischmann, J. M. Cherry, and S. Henikoff. 2000.
Comparative genomics of the eukaryotes. Science 287:2204–
15.
Salvini-Plawen, L. and E. Mayr. 1977. On the evolution of photoreceptors and eyes. Evol. Biol. 10:207–263.
Sanchez-Salazar, J., M. T. Pletcher, R. L. Bennett, S. J. Brown, T. J.
Dandamudi, R. E. Denell, and J. S. Doctor. 1996. The Tribolium
decapentaplegic gene is similar in sequence, structure, and expression to the Drosophila dpp gene. Dev. Genes & Evol. 206:
237–246.
Schroder, R., C. Eckert, C. Wolff, and D. Tautz. 2000. Conserved
and divergent aspects of terminal patterning in the beetle Tribolium castaneum. Proc. Natl. Acad. Sci. U. S. A. 97:6591–6.
Serikaku, M. A. and J. E. O’Tousa. 1994. sine oculis is a homeobox
gene required for Drosophila visual system development. Genetics 138:1137–50.
Singh, A., M. Kango-Singh, and Y. H. Sun. 2002. Eye suppression,
a novel function of teashirt, requires Wingless signaling. Development 129:4271–80.
Tawfik, A. I., Y. Tanaka, and S. Tanaka. 2002. Possible involvement
of ecdysteroids in embryonic diapause of Locusta migratoria.
J. Insect Physiol. 48:743–749.
Tomarev, S. I. 1997. Pax-6, eyes absent, and Prox 1 in eye development. Int. J. Dev. Biol. 41:835–42.
Treisman, J. E. and G. M. Rubin. 1995. wingless inhibits morphogenetic furrow movement in the Drosophila eye disc. Development 121:3519–27.
Trujillo-Cenoz, O. 1985. The eye: Development, structure and neural
connections. In G. A. Kerkut and L. I. Gibert (eds.), Comprehensive insect physiology and pharmacology, pp. 171–223. Pergamon Press, Oxford.
Ullmann, S. L. 1967. The development of the nervous system and
other ectodermal derivatives in Tenebrio molitor L. (Insecta,
Coleoptera). Phil. Trans. R. Soc. London B 252:1–24.
Wilson, M., P. Garrard, and S. McGinness. 1978. The unit structure
of the locust compound eye. Cell. Tiss. Res. 195:205–226.
Wolff, T. and D. Ready. 1993. Pattern formation in the Drosophila
retina. In M. Bate and A. Martinez-Arias (eds.), The development of Drosophila melanogaster, pp. 1277–1326. Cold Spring
Harbor Laboratory Press, Cold Spring Harbor.
Yang, C. H., M. A. Simon, and H. McNeill. 1999. mirror controls
planar polarity and equator formation through repression of
fringe expression and through control of cell affinities. Development 126:5857–66.
Younossi-Hartenstein, U. Tepass, and V. Hartenstein. 1993. Embryonic origin of the imaginal discs of the head of Drosophila
melanogaster. Roux’s Arch. Dev. Biol. 203:60–73.
Zelhof, A. C., N. Ghbeish, C. Tsai, R. M. Evans, and M. McKeown.
1997. A role for Ultraspiracle, the Drosophila RXR, in morphogenetic furrow movement and photoreceptor cluster formation. Development 124:2499–2506.