Current Biology

Review

The Evolution of Insect Metamorphosis

James W. Truman
Department of Biology and Friday Harbor Laboratories, University of Washington, 620 University Road, Friday Harbor, WA 98250, USA
Correspondence: trumanj@janelia.hhmi.org
https://doi.org/10.1016/j.cub.2019.10.009

The evolution of insect metamorphosis is one of the most important sagas in animal history, transforming
small, obscure soil arthropods into a dominant terrestrial group that has profoundly shaped the evolution
of terrestrial life. The evolution of flight initiated the trajectory towards metamorphosis, favoring enhanced
differences between juvenile and adult stages. The initial step modified postembryonic development, result-
ing in the nymph–adult differences characteristic of hemimetabolous species. The second step was to com-
plete metamorphosis, holometaboly, and occurred by profoundly altering embryogenesis to produce a larval
stage, the nymph becoming the pupa to accommodate the deferred development needed to make the adult.
These changing life history patterns were intimately linked to two hormonal systems, the ecdysteroids and
the juvenile hormones (JH), which function in both embryonic and postembryonic domains and control the
stage-specifying genes Krüppel homolog 1 (Kr-h1), broad and E93. The ecdysteroids induce and direct molt-
ing through the ecdysone receptor (EcR), a nuclear hormone receptor with numerous targets including a
conserved transcription factor network, the ‘Ashburner cascade’, which translates features of the ecdyste-
roid peak into the different phases of the molt. With the evolution of metamorphosis, ecdysteroids acquired
a metamorphic function that exploited the repressor capacity of the unliganded EcR, making it a hormone-
controlled gateway for the tissue development preceding metamorphosis. JH directs ecdysteroid action,
controlling Kr-h1 expression which in turn regulates the other stage-specifying genes. JH appears in basal
insect groups as their embryos shift from growth and patterning to differentiation. As a major portion of
embryogenesis was deferred to postembryonic life with the evolution of holometaboly, JH also acquired a
potent role in regulating postembryonic growth and development. Details of its involvement in broad expres-
sion and E93 suppression have been modified as life cycles became more complex and likely underlie some
of the changes seen in the shift from incomplete to complete metamorphosis.

Introduction for hormone extraction and chemical characterization. Ensuing

Amphibians and insects provide most people with their first
contact with metamorphosis. The striking transformation of a
tadpole into a frog is relatively rapid but one can track its
progression from day to day as its limbs extend and its tail is re-
sorbed. Metamorphosis in insects, though, seems more myste-
rious as their external exoskeleton hides the gradual changes
occurring inside, and one only sees the results when the old
outer cuticle is abruptly shed, such as when the butterfly
emerges from its chrysalis. Metamorphosis is an ancient life his-
tory trait that extends deep to the roots of the amphibians and
many groups of invertebrates. Insects are unusual, though, in
that their metamorphosis is a derived trait that evolved well after
they colonized the land. Insect orders approximating different
steps in the evolution of metamorphosis exist today, so that
comparative studies allow a piecing together of the molecular,
developmental and endocrine changes that accompanied this
transition.
Insect metamorphosis is under hormonal control, which has
been a subject of experimental study for almost a century. The
importance of insects as competitors for food and fiber as well
as their impact on human and animal health spurred extensive
research into their physiology and development, and especially
into how they control their metamorphosis. The early stages of
this research typically used large insects which facilitated surgi-
cal manipulations and the isolation of sufficient amounts of tissue
genetic and molecular studies, though, gradually shifted to
Drosophila melanogaster, which still serves as an unrivaled plat-
form for gene discovery. However, many derived features of
Drosophila, such as an invariant number of larval instars and
an almost complete replacement of larval cells by those of the
adult at metamorphosis, have resulted in a de-emphasis or
loss of some aspects of the hormonal control seen in less derived
insects. Consequently, despite the fact that work on Drosophila
is responsible for the discovery of virtually all the genes involved
in hormone reception, their downstream effects, and stage spec-
ificity of action, the broader functions of such genes often re-
mained obscure for decades and were finally shown in insects
with less derived life history patterns. For example, work on
the flour beetle, Tribolium castaneum, provided the proof that
the Methoprene tolerant gene (Met) does, indeed, encode the re-
ceptor for juvenile hormone (JH), a key hormone that suppresses
metamorphosis [1,2]. Also, studies on this beetle and the cock-
roach Blattella germanica showed that the helix-turn-helix tran-
scription factor Eip93F (E93) is a master stage-specifying gene
that stands on top of a gene cascade that produces the adult
phenotype [3].
Besides Drosophila, this discussion draws on other insect
model systems such as Bombyx mori, Manduca sexta, Tribolium
castaneum, Oncopeltus fasciatus, Pyrrhocoris apterus and
Blattella germanica, for which genomic information and gene

R1252 Current Biology 29, R1252–R1268, December 2, 2019 ª 2019 Elsevier Ltd.
Current Biology

Review

Ametabolous Hemimetabolous Holometabolous

Palaeoptera
Palaeopte
era Polyneoptera Paraneoptera

Mantophasmatodea
smatodea
Ephemeroptera
ptera
natha
Archaeognatha

Grylloblattidae
idae

Raphidioptera
tera
Phasmatodeadea

tera
Thysanoptera

Hymenoptera
era

era
Siphonaptera
era
Megaloptera
Phthiraptera
ra
Dermaptera ra

Lepidoptera
ra
Embioptera a

Coleoptera
a

a
Trichoptera
a
Zygentoma

a
Neuroptera
Plecopteraa
Orthopteraa

a
Mecoptera
Hemiptera
a
Psocodea
Mantodea
Zoraptera

Blattodea
Odonata

Diptera
0
N

66
Age [millions of years ago]

145

200

250
300

360
Holometabola
420
Pterygota
485
Current Biology

Figure 1. Phylogeny of insects showing the major types of development.

The figure shows how the three major types of insect development, ametabolous, hemimetabolous and holometabolous, map onto insect phylogeny, with
examples of the immature and adult stages for each. The asterisk indicates neometabolous forms that have independently evolved a life history with a larva, pupa
and adult. For the major orders the width of the boxes show the number of families through time. Phylogeny from [134,135] with family distributions adapted from
[14] based on [136].

suppression techniques can be woven into a rich history of endo- process termed ecdysis. Consequently, the growth of the juve-
crine experimentation. Although these few systems do not begin nile alternates between a period of feeding and growth (the inter-
to cover the amazing range of variation that one sees in the in- molt) and a molt. The span from one ecdysis to the next is termed
sects, it is heartening that highly derived life history strategies, an instar. In most insects, there is a characteristic number of in-
such as female neoteny in Strepsiptera [4,5] and scale insects stars between hatching and metamorphosis, but the decision to
[6] or the complete metamorphosis that evolved in parallel in begin metamorphosis typically depends on reaching a charac-
thrips [7], utilize the same molecular switches that regulate teristic species-specific size [8]. Therefore, suboptimal nutrition
more traditional life history patterns. This review explores the or injuries or disease may result in the intercalation of additional
developmental changes that accompanied the evolution of the molts until this size is achieved. Insects are extremely diverse in
insect larval and pupal stages. It then focuses on the two main their feeding strategies, however, and in some species the larva
hormonal systems that regulate insect molting and metamor- can opt for an earlier instar to start metamorphosis if food condi-
phosis and considers how these systems evolved to support tions are poor (for example [9]).
the increasing complexity of insect life histories. Finally, it exam- By any criteria, insects are a spectacularly successful terres-
ines our growing knowledge of the function of the stage specifi- trial group with diverse life history strategies and 28 extant orders
cation genes that regulate the phenotype of each major life stage (Figure 1). The true insects and other hexapods such as spring-
and how these genes and their control have changed with the tails (Collembola) and diplurans (Diplura) diverged from their
increasing complexity of insect life histories. aquatic crustacean ancestors and were part of the arthropod in-
vasion of land that began over 450 million years ago. The life his-
Insect Life History Strategies tory of these earliest insects was similar to those of the jumping
Because their rigid exoskeleton constrains growth and changes bristletails (Archaeognatha) and silverfish (Zygentoma) of today.
in morphology, an insect’s life history is punctuated by molts, These primitively wingless insects have an ametabolous lifestyle
during which a new cuticle is formed and the old one shed, a which shows direct development: the hatchling is a miniature

Current Biology 29, R1252–R1268, December 2, 2019 R1253


version of the adult and its morphology is essentially unchanged
during its growth to an adult. Like most other arthropods, the
adults in these ametabolous orders continue to molt, typically
alternating a molt with a bout of reproduction [10].
The insects are the only invertebrates to have evolved pow-
ered flight and they owned the skies for over 100 million years un-
til the Pterosaurs appeared in the late Triassic. Flight provided a
means to avoid predation, to move rapidly between dispersed
food patches, and to undertake long-range dispersals. Early or-
ders of winged insects, such as the extinct Palaeodictyoptera
and Megasecoptera, were also considered ametabolous
because their juveniles had small articulated winglets that even-
tually became large enough to support flight in later instars
[11,12]. The evolution of wings and flight, though, posed devel-
opmental problems, the solution for which provided the opportu-
nity for the first step into metamorphosis [13]. Instead of growing
as small articulated winglets, wing growth in juveniles took place
in immobile wing pads, which then morphed into articulated
wings at the adult molt. A second, wing-related change evident
in today’s winged insects (the Pteryogota) was that the adult
molt became a terminal molt. The cessation of adult molting
was likely related to making the wing as light as possible. Weight
reduction occurs by the death of the epidermal cells that secrete
the cuticle of the wing membrane after adult emergence. How-
ever, without these cells, the cuticle of the wing membrane
cannot be made again. The only extant insects that molt their
wings are the mayflies (Ephemeroptera), a very basal order
which has a winged subimago stage that quickly molts to the
winged adult. The subimagos have thick wings and are very
awkward flyers. Interestingly, the loss of adult molting seems
to be irreversible: two parasitic orders of insects, the lice (Pthir-
aptera) and the fleas (Siphonaptera), are wingless as adults but
neither have reacquired adult molting.
The evolution of flight, then, ushered in the strategy of incom-
plete metamorphosis, also known as hemimetabolous develop-
ment. The immature stage is called a nymph, and nymphs
typically resemble the adult but lack wings. External wing pads
appear during nymphal growth, but they do not transform into
functional wings until the adult molt. Nymphs and adults usually
occupy similar habitats, exploit similar food sources, but often
have different cuticle types. While nymphs and adults usually
have similar morphologies, in orders like the Odonata (dragon-
flies) and Ephemeroptera (mayflies) that have aquatic nymphs
that transform into aerial adults, the two stages may show
marked differences in body form. The high mobility of the hemi-
metabolous adult likely set the stage for the step to complete
metamorphosis.
Evolution of the hemimetabolous lifestyle primarily involved
alterations in postembryonic development. As in their
ametabolous ancestors, the basic adult body plan was laid
down during embryogenesis, although some degree of delayed
development, such as adding new rows of visual units (omma-
tidia) to the growing eye, occurs in some groups. The shift to
holometabolous development, though, involved a fundamental
alteration of embryonic development to produce a larval stage
that bore little or no resemblance to the adult. Adaptations of
the larval and adult stages were developmentally uncoupled,
which allowed the evolution of a larval stage with reduced
motility but enhanced capacity to exploit food resources without
R1254 Current Biology 29, R1252–R1268, December 2, 2019
Current Biology
Review
affecting adult specializations for dispersal and reproduction
[14]. This life history innovation resulted in the rapid diversifica-
tion and radiation within the Holometabola [15], which includes
the orders that have the most species (Coleoptera), the most
biomass (Hymenoptera), and the greatest impacts on agriculture
(Lepidoptera) and on human and animal health (Diptera).
In holometabolous insects, the transition from larva to the
adult takes place through the nonfeeding, transitional stage
known as the pupa. The oldest unequivocal fossils of holometab-
olous larvae are from about 320 million years ago [16]. Curiously,
in this era of giant insects, the earliest holometabolans were
quite small. Their small size may have allowed them to exploit
niches that were unavailable or not productive for their larger
relatives. A drop in ancient oxygen levels may have subsequently
doomed the giants but would have had little effect on the
small holometabolous forms and may have opened the door
to their radiation [16]. Some hemimetabolous groups, such
as thrips and some hemipterans, have independently evolved a
holometabolous-like lifestyle, a condition sometimes called
‘neometaboly’ [17].
The Larva of Insects with Complete Metamorphosis
Ideas about insect metamorphosis extend back to Aristotle
[18,19] and were later built on by the anatomist Sir William Har-
vey, who considered the larva as the product of an ‘imperfect
egg’, from which the animal hatched before acquiring its spe-
cies-typical form. The pupa then served as a second egg in
which the proper form of the animal was constructed from the re-
mains of the larva. Later refinement of these ideas held that the
pupa corresponded to the nymph of hemimetabolous forms
[20,21], but these ideas for the correspondence of the life stages
in holometabolous and hemimetabolous insects have been
challenged by other researchers who considered larvae and
nymphs to be equivalent stages [22,23]. They held that the
pupa is a modified last nymphal stage [22] or that the last
nymphal molt splits into a two-step process with the pupa form-
ing the first step [23]. Discussions of these latter ideas can be
found in other papers [24,25], but we found that the idea of the
larva arising from a modified embryonic stage fits better with
the available developmental, endocrine and molecular data
[26–28]. I will use the terms ‘larvae’ and ‘nymphs’ for the imma-
ture feeding stages of holometabolous and hemimetabolous in-
sects, respectively.
During embryogenesis, regions of the embryo organize into
primordia destined to form legs, antennae, eyes and so on. In
hemimetabolous insects, such primordia grow and pattern
themselves as they form miniature versions of the adult struc-
ture. The embryos of holometabolous species have the same
set of primordia, but clonal experiments with Drosophila demon-
strated that only part of each primordium is utilized to make a
reduced, larval structure, while the remainder is set aside to
contribute later to the adult counterpart [29]. This pattern is nicely
illustrated for the compound eye (Figure 2A) [30,31]. In hemime-
tabolous embryos, the eye primordium begins organization with
a row of ommatidia on its posterior border, followed by addition
of successive rows as a morphogenetic wave progresses
anteriorly across the primordium. The nymph hatches with a
well-developed compound eye, although some species retain
a primordium along the anterior margin to add additional rows
Current Biology

Review

A Embryo
Nymph Adult
Grasshopper

Hatch
Birth photoreceptors

0 50 100%
E1 Growth
E2
E3
Embryo Larva Pupa Adult
Moth

50 Growth* Morphogenesis Differentiation

0 E1 100%
E2 pp
E3
Day 0 Day 4 Isomorphic Morphogenetic
B C Growth: phase phase
p *
Eye ** Molt

Primordium size
* *
* *
0h
0h Fee
e
eed
ed 4d
ed
Feed 4d
Wing
Eye
Wing
L3 L4 L5 Pupa

Current Biology

Figure 2. Modification of embryonic and postembryonic development associated with the evolution of the larva.
(A) Comparison of eye development in a generalized hemimetabolous insect (grasshopper) with that in a holometabolous insect (moth). The embryonic timeline
(% of development) shows the time of deposition of the three embryonic cuticles (black line: epicuticle; red line: procuticle). Photoreceptors start being born at
similar times, but moth embryos make only the first few photoreceptors while grasshoppers show extended photoreceptor production. The unused portion of the
eye primordium (light orange) in the caterpillar is maintained through larval growth and eventually forms the eye imaginal disc that forms the adult compound eye.
Adapted from [28] and based on [31,47]. Timing of embryonic molting [38,52]. (B) Caterpillars of Manduca sexta have some imaginal primordia that invaginate
early to form an imaginal disc (the wing disc, in green) and do not contribute to the larva, while others (red) are associated with their larval counterpart and make
larval cuticle but in the last larval instar they dedifferentiate and invaginate to make a late-forming imaginal disc. The confocal images of propidium iodide stained
whole mounts showing the eye primordium (p, dotted outline) at day 0 of the last larval instar when it is part of the epidermal monolayer and at Day 4 when it has
invaginated to form the eye disc; the asterisk indicates larval photoreceptors just behind the primordium. Through the same period the wing disc transforms
from a simple sack to a complexly folded structure with incipient wing veins. Adapted from [41]. (C) Imaginal primordia that become early-forming (green) versus
late-forming (red) discs show different patterns of isomorphic growth: as an early forming disc they grow through both the intermolt and molt periods while in the
late-forming strategy growth occurs only in the early phase of each larval molt. Under both strategies growth shifts into a morphogenetic mode early in the final
larval instar.

at subsequent nymphal molts. In embryos of holometabolous
species, by contrast, the eye primordium makes only an initial,
posterior set of photoreceptors which differentiate into the sim-
ple lateral eyes (stemmata) characteristic of larvae. The anterior
portion of the primordium remains dormant during larval growth
but, in the last larval instar, undergoes morphogenesis into the
eye imaginal disc [31].
The legs of Lepidoptera present a more complex metamorphic
challenge, because the animal makes two different types of
complex leg during its lifetime: the caterpillar has shortened
thoracic legs modified for grasping, while the adult moth or but-
terfly has typical walking legs. The formation of a walking leg,
either during embryogenesis in a cricket [32,33] or postem-
bryonically in a Drosophila leg imaginal disc, occurs through a
similar pattern of recruitment of proximodistal patterning genes
[34]. This patterning cascade is arrested at an intermediate point
in embryos of the moth Manduca sexta, producing the modified
caterpillar leg [35]. The caterpillar leg, though, retains a remnant
of the embryonic leg primordium [36,37], which undergoes
extensive proliferation and morphogenesis in the last larval stage
and combines with some of the larval cells as they complete the
patterning program to make the adult walking leg [37].
The larval form, then, arises from the arrest of ancestral pat-
terns of tissue growth and patterning. The structures that are
organized prior to the arrest are adapted to the larval form, while
the remainder of the primordium is carried forward to contribute
to the adult system (Figure 2A). The point at which the embryonic
programs arrest to form the larval structure and the degree to
which larval cells then contribute to the adult version varies ac-
cording to the species and organ system considered [27].
Insect embryos typically make three embryonic cuticles, and
these are well developed in hemimetabolous forms [38]. The first
embryonic cuticle is delicate and is deposited with the comple-
tion of segmentation and establishment of the embryonic
primordia. The embryo then enters a phase of rapid growth
and patterning which is largely finished at dorsal closure when

Current Biology 29, R1252–R1268, December 2, 2019 R1255


it deposits a second embryonic cuticle, called the pronymphal
cuticle. This pronymph stage has the general shape of the
nymph, although its proportions may be constrained by the con-
fines of the eggshell. Its cuticle differs from nymphal cuticle and
may bear specializations for escape from the shell and the ovipo-
sition site. Then follows tissue differentiation and maturation and
the deposition of the first nymphal cuticle. We proposed that the
pronymph was the forerunner of the larval stage [26] because the
patterning processes that were interrupted in making the larva
were those that occur during pronymph formation. In addition,
some pronymphs show unique behavioral adaptations, such
as body crawling movements that allow them to escape from
the oviposition site [39], and such stage-specific specializations
may have been important pre-adaptations for the evolution of the
larva [26]. We further argued that the embryos of holometabolans
underwent only two molts during embryogenesis [26]; however,
subsequent work [38] showed that a number of holometabolous
insect species also make three embryonic cuticles, although the
second cuticle is somewhat reduced and lacks endocuticle.
The Pupa and Imaginal Discs and Primordia
As described above, the origin of the pupal stage has been a
controversial issue over the years. In support of the classic
view that the pupa corresponds to the hemimetabolous nymph,
we proposed [26] that the immature stage responsible for growth
shifted from the nymph to the pronymph as the latter trans-
formed into a larva that was capable of feeding. The number of
pronymphal/larval instars increased to accommodate growth,
while the number of nymphal instars decreased as the nymph
became a nonfeeding transitional stage, now recognized as
the pupa. The pupa then provided a mold or template for the
body of the adult, in which the adult structures subsequently
differentiated and matured, but it also developed adaptations
to ensure its survival, such as a characteristic cuticle type. If pu-
pation occurs within a protected environment, then the pupal
cuticle is typically thin, transparent and unadorned; pupae
exposed to the environment, such as the chrysalis of butterflies,
have a hardened cuticle that may be adorned with extensions
and coloration that aid in its camouflage or protect it from attacks
by other arthropods.
The pupa is typically formed from reprogrammed larval cells
as well as cells from imaginal primordia (for example [37]). The
relative contribution from these two depends on the organ and
species concerned. An imaginal primordium is a collection of
larval cells with a latent embryonic capacity that allows them to
partially dedifferentiate and undergo growth and morphogenesis
to form part or all of an adult structure. Imaginal primordia are
usually associated with the corresponding larval structures, but
their cells remain diploid while the surrounding larval cells may
become highly polyploid. The ancestral condition for such
primordia is likely similar to that seen for the antennal, leg and
eye primordia of Lepidoptera (Figure 2B,C) [27,40]. These
primordia are integral parts of the larval epidermis, making a
new larval cuticle at each molt and adding layers of endocuticle
through the intermolt. Early in the last larval instar, though, they
stop making endocuticle, detach from the larval cuticle, partially
dedifferentiate and initiate proliferation and morphogenesis [41].
As they grow, they often invaginate into the body cavity to
become late-forming imaginal discs.
R1256 Current Biology 29, R1252–R1268, December 2, 2019
Current Biology
Review
A more derived condition is seen when the imaginal primor-
dium does not contribute to the larva but invaginates into the
body, where it becomes an early forming imaginal disc. Early
forming discs are not found in the beetle Tribolium, although
they are found in more derived groups of beetles [26]. In Lepi-
doptera, the only early forming discs are the wing and genital
discs (Figure 2B), while in higher Diptera, such as Drosophila,
virtually all adult structures come from early forming discs [29].
The only epidermal cells retaining an ancestral pattern of devel-
opment are the abdominal histoblasts, small collections of
diploid cells that make larval and then pupal cuticle before they
proliferate and spread over the abdomen to form the adult
cuticle.
The imaginal primordia show two different phases of growth
(Figure 2B,C). In the preterminal larval instars, their growth is
sensitive to nutritional conditions and proportional to that of
the larva as a whole. Through this period the early-forming discs
have a distinct growth advantage over their late-forming coun-
terparts (Figure 2C). At best, they secrete only a thin epicuticle
[40] and can proliferate through both the molt and intermolt pe-
riods. The primordia that transform into discs in the last larval
instar, however, are integral parts of the larval epidermis in pre-
terminal instars. As such, they make cuticle through the latter
part of the molt and the intermolt and their proliferation is
restricted to a brief burst at the start of each molt. In the last
instar both types of primordia shift to a morphogenetic program
that renders their growth independent of external nutrient input
[41,42].
Ecdysteroids and the Control of Molting and
Metamorphosis
The ecdysteroids are polyhydroxylated steroids that are made
from dietary cholesterol or related plant sterols [43] (Figure 3A).
Their primary function is to induce molting and they serve this
function throughout the arthropods [44,45]. In insects, circu-
lating ecdysteroids typically come from the prothoracic glands,
although in some species other tissues such as the epidermis,
the oenocytes, and the adult gonads also have this capacity.
The prothoracic glands secrete ecdysone which is then further
hydroxylated to 20-hydroxyecdysone (20E) by ecdysone 20-
monooxygenase in peripheral tissues [43,46]. 20E is normally
considered to be the active form of the hormone, but ecdysone
can function in its own right early in molting or metamorphic
programs (for example [47]). The control of ecdysone secretion
from the prothoracic glands is complex, but the two major
players are peptide hormones from brain neurosecretory cells:
the prothoracicotropic hormone and the insulin-like peptides
[48]. The functions of ecdysteroids in the control of insect
metamorphosis have striking parallels with those of the thyroid
hormones in directing the metamorphosis of amphibians
[49,50]. These parallels will be emphasized throughout this
section.
As seen in Figure 3A, ecdysteroid levels are low through the in-
termolt but show a major peak during each molt. As ecdysone
20-monooxygenase is upregulated during molts [43,46], periph-
eral tissues rather than the prothoracic glands determine the
steroid composition during the peak. Typically, ecdysone pre-
dominates at the start of the peak with 20E becoming predomi-
nant later (for example [51]; Figure 3A).
Current Biology

Review

A PTTH Peripheral tissues B

insulins
DNA Ligand
Brain Ecdysone 20 EcR-A A/B binding binding
Prothoracic
Glands monooxygenase Activation C D E F
EcR-B A/B Activation
20E
E
Ecdysteroid titer

9-cis RA
Other EcR RXR or EcR EcR

Coleoptera
EcR RXR or EcR EcR

Lepidoptera lipid
L4 L5 Pupa Adult
Days & diptera EcR USP
Co-repressors Co-activators
C 20E 20E
Larval tissues EcR USP EcR USP
EcR-B1 B1 B1
EcRE EcRE
Closed chromatin Open chromatin

Co-repressors Co-activators Other signaling pathways?

20E 20E
EcR USP EcR USP
EcR-A A A
EcRE EcRE RE
Imaginal tissues
Closed chromatin Open chromatin
Current Biology

Figure 3. The ecdysteroid system of insects.

(A) Ecdysone (E) is made by the prothoracic glands and secreted mainly under the influence of two brain hormones, the prothoracicotropic hormone (PTTH) and
the insect insulins. E is then converted to 20-hydroxyecdysone by peripheral tissues. The ecdysteroid titer is based on Manduca sexta [51,137] and shows total
ecdysteroids for the larval and pupal molts but ecdysone and 20E separately for the molt to the adult. (B) The structure of the ecdysone receptor (EcR) showing the
different domains of the protein. Most species express two EcR isoforms that vary in the transactivation ability of their A/B domain [58]. Depending on insect
group, EcR can either act as a homodimer or heterodimerize with the retinoid X receptor (RXR). RXR binds ligands such as 9-cis-retinoic acid (9-cis-RA) but there
are variants that lack the ligand binding domain or bind a constitutive structural lipid. The latter form is called Ultraspiracle (USP). The EcR in the latter group
cannot form homodimers. Adapted from [58]. (C) Prior to metamorphosis in Drosophila, larval and imaginal tissues express different EcR isoforms. In both cases,
the unliganded EcR–USP complex binds to its response element (EcRE) and binds to corepressors that suppress transcription. Ecdysteroid binding leads to
opening of chromatin structure and gene activation. The two major isoforms differ in this capacity and this difference may support different functions for ec-
dysteroids in larval versus imaginal tissues. EcRE, ecdysone response element; RE, response element for other transcription factors.

The deposition of a new cuticle, be it during the embryonic or
postembryonic period, requires a pulse of ecdysteroids. Each of
the three embryonic cuticles produced during grasshopper
embryogenesis is associated with an ecdysteroid peak [52]. In
Drosophila, embryos that lack the ecdysone biosynthetic en-
zymes go through embryogenesis but do not make a larval
cuticle and have a ghostly appearance that resulted in such
mutants being called the ‘Halloween Group’. This collection of
mutants subsequently provided the tools for dissection of the
ecdysone biosynthetic pathway [53]
The Ecdysone Receptor
Both ecdysteroids and thyroid hormones act through members
of the Group II subfamily of nuclear hormone receptors: the thy-
roid hormone receptor (TR) in the case of amphibians, and the
ecdysone receptor (EcR) in insects [54]. The structure of EcR is
typical for this subfamily (Figure 3B): it has two C4 zinc fingers
in the carboxyl domain to mediate DNA binding, the ecdyste-
roid-binding ‘E’ domain, and an amino-terminal A/B region that
is differentially spliced to make two major isoforms, EcR-A and
EcR-B [55,56]. These two isoforms are found throughout the in-
sects [57], but it is not known if their functions differ with the
different types of metamorphic development.
As with other members of this nuclear receptor subfamily,
some insect EcRs can homodimerize or form heterodimers
with another nuclear receptor, the retinoid X receptor (RXR). In
more basal insects RXR binds fatty acids such as 9-cis-retinoic
acid, as seen in its vertebrate homolog. EcR–RXR has under-
gone considerable evolution within the insects (Figure 3B) [58].
The ligand-binding domain of RXR was apparently lost in bee-
tles. In the Diptera and Lepidoptera, RXR has been modified
to bind a structural lipid and is called Ultraspiracle (USP), while
EcR has evolved an expanded dimerization interface that
facilitates making EcR–USP dimers but prevents formation of
EcR homodimers [58]. Because most of the model systems for
studying ecdysteroid action come from these two orders, we
know a lot about the functions of EcR–USP heterodimers, but
little about the EcR homodimers or effects of ligand binding
to RXR.
The EcR–USP heterodimer is typically in the nucleus and
bound to DNA (Figure 3C) [59]. Without ligand, EcR–USP recruits
nucleosome remodeling proteins to compact the chromatin and
cause local repression. With binding of 20E, coactivators such as
histone acetylases are exchanged for the co-repressors to
‘open’ chromatin structure and facilitate transcription. The

Current Biology 29, R1252–R1268, December 2, 2019 R1257


A Intermolt Molt
Ecdysteroid titer
Preparatory High Ecdysial
phase 20E phase
Decline
E or Cuticle
low 20E of 20E
induction
phase
B 20E
High 20E
Low 20E Early genes
(High threshold)
(or E?)
E74A, E75B
Early/late
DHR3, DHR4
Early genes
(Low threshold)
Late
e.g. E74B
βFtz-F1
Current Biology
Figure 4. The relationship of phases of the molting process with the
ecdysteroid titer and the various components of the ‘Ashburner
cascade’.
(A) The ecdysteroid requirements for the three major phases of the molting
process. (B) The components of the transcription factor cascade that are
active during the corresponding phases of the molt and their relationship to the
ecdysteroid titer. The genes are examples of transcription factors from each
class.
ligand-binding domain mediates core repression or activation
based on whether or not ligand is present. The two major EcR
isoforms differ in both their functionality and their tissue distribu-
tion. In Drosophila melanogaster, EcR-B1 has a ligand-indepen-
dent activation domain that is lacking in EcR-A. EcR-B1 is found
through the larval stages with EcR-A joining it in the last instar
and into metamorphosis [56,60], but at the latter time they are
separated into different cell types with larval cells expressing
EcR-B1 and imaginal cells only EcR-A [56].
Ecdysteroids and Growth and Molting
The action of ecdysteroids in causing molting of the cuticle is an
ancient function of these hormones that precedes their involve-
ment in metamorphosis. The insect cuticle is multilayered and
made primarily of protein and fibers of chitin, a polymer of
N-acetyl-glucosamine [61,62]. Even in the largest insect, the
epidermis that secretes the cuticle is only a single cell thick.
This monolayer produces the cuticle that lies above it.
The process of molting the cuticle can be divided into a
number of phases [63] which I have condensed to three.
These phases are orchestrated by different features of the ec-
dysteroid pulse (Figure 4A). The rising ecdysteroid titer evokes
the preparatory phase that involves a commitment component
which determines the nature of the molt (for example larval
versus pupal [64]) and the cellular events of detachment of the
epidermis from the overlying cuticle (apolysis), the secretion of
molting fluid into the intervening space, cell division, and the dif-
ferentiation of new epidermal organelles such as a sensory hair
and its socket. These events can be evoked by either low levels
of 20E or ecdysone (for example, [47]). As ecdysone is the major
R1258 Current Biology 29, R1252–R1268, December 2, 2019
Current Biology
Review
component at the start of the ecdysteroid pulse [51], it is likely the
form responsible for these events.
This is followed by the cuticle induction phase, starting with
secretion of an outer epicuticle and then the procuticle made
of protein and chitin. Whether the cuticle will be rigid or flexible
depends on its protein composition. The outer layers of the pro-
cuticle (the exocuticle) eventually become heavily crosslinked,
making them extremely hard and resistant to degradation. The
inner layers (the endocuticle) continue to be deposited even after
the old cuticle is shed and the new one expanded. Ecdysone or
low levels of 20E cannot evoke cuticle production; rather, high
levels of 20E are required (for example [47]). This difference is
important, as the peripheral tissues are responsible for the con-
version of ecdysone to 20E [43], thereby controlling the length of
the preparatory phase and the onset of cuticle production. An
extended preparatory phase is especially important for the
molt from the pupa to the adult to allow sufficient time for the dif-
ferentiation and growth needed to make many adult structures.
The premature termination of this phase and precocious cuticle
production are seen when lepidopteran pupae are injected with a
high dose of 20E or other ecdysteroid agonists [65]: they show a
‘hyperecdysonism’ response, becoming malformed adults with
naked cuticle and small eyes because the high 20E forced cuticle
deposition before the completion of scale formation and eye
growth.
The ecdysial phase is the last one of the molt and involves the
degradation and shedding of the old cuticle and expansion and
tanning of the new one. During this phase, proteases and
chitinases degrade the old endocuticle, leaving the highly
cross-linked, but thin exocuticle. A cascade of neuropeptides
then orchestrates a behavioral sequence to rupture and shed
the old cuticle (ecdysis sensu stricto), and then expand and
crosslink the new one [66]. These events require the withdrawal
of ecdysteroids. This requirement for the withdrawal of 20E is
why insects treated with ecdysteroid mimics that are not readily
metabolized typically die late in the molt without shedding their
old cuticle [67].
The different phases of the molt have characteristic patterns of
gene expression (Figure 4B). These were first identified in the
context of ecdysone’s effects on metamorphosis using cultured
salivary glands of the midge Chironomous tentans [68]. Addition
of ecdysone to larval glands evoked a stereotyped temporal
sequence of ‘puffs’ — regions of expanded chromatin denoting
transcriptional activity — along their giant polytene chromo-
somes, providing the first evidence that steroid hormones might
directly activate gene transcription. Comparison of puffing pat-
terns with and without inhibitors of protein synthesis indicated
that some genes were direct steroid targets while others were
downstream of this initial set. Quantitative analysis of these puffs
and their steroid requirements by Michael Ashburner and col-
leagues [69], using Drosophila salivary glands, identified a
network of gene interactions often called the ‘Ashburner
cascade’. While many of these genes are tissue specific and
encode effector genes, there is a core set of transcription factors
that are expressed in all tissues, at each molt, and across spe-
cies [70–72]. Many of these transcription factors are orphan nu-
clear receptors and they constitute a genetic circuit that interacts
with and interprets that ecdysteroid titer, translating it into tem-
poral programs of gene activity needed for the different phases
Current Biology
Review
of the molt [72]. Not all of the details of the interactions in this cir-
cuit have been resolved, but enough is known about interactions
of individual components to appreciate its overall function.
As seen in Figure 4B, different parts of the cascade are asso-
ciated with the different phases of the molt. The early genes,
which are directly induced by ecdysteroids, include low-
threshold and high-threshold classes [73]. The low-threshold
class genes, such as E74B, appear to be associated with the
preparatory phase whereas the high threshold class genes, like
E74A and E75B, are associated with the cuticle initiation phase.
Members of the high-threshold class also suppress the low-
threshold genes, thereby terminating the preparatory phase.
Consequently, the length of the preparatory phase is relatively
plastic and can be maintained by the appropriate hormone con-
ditions as long as the high threshold genes are not induced. The
high steroid titers also evoke the activation of early-late genes,
such as Drosophila Hormone Receptor 3 (DHR3), that have
some steroid sensitivity but depend on the early genes for their
full activation [74].
Output genes, such as those for cuticle proteins, are targets of
these transcription factors, as well the ecdysteroids (for example
[75–77]). A full picture of how all of the members of the cascade
act to direct the complex program of cuticle assembly has yet to
be made. Genes like DHR3 feed forward to activate late genes
like b-Ftz-F1, expression of which also requires the removal of
20E [78]. b-Ftz-F1 is involved in some later cuticle gene induction
(for example [76]), but it also induces production of the receptor
for ecdysis triggering hormone, a key hormone that controls the
shedding of the cuticle [79]. Hence, maintaining elevated 20E
levels at the end of the molt prevents b-Ftz-F1 expression and
the shedding of the old cuticle.
Ecdysteroids and Metamorphosis
The ecdysteroids also drive metamorphic development in a
manner analogous to that of thyroid hormone and amphibian
metamorphosis [49,50]. Amphibian tadpoles begin in a pre-
metamorphosis phase characterized by nutrient-dependent
growth of the tadpole, including its limb buds. The secretion of
thyroxine (T4) by the thyroid gland causes prometamorphosis
characterized by TRa expression, accelerated growth and
patterning of limb buds, and the induction of TRb. The devel-
oping tissues subsequently acquire deiodinase activity permit-
ting the local conversion of T4 to triiodothyronine (T3) [80].
Only T3 is capable of causing metamorphic climax and it acts
through TRb to evoke the dramatic events of tail loss and the
eruption of the forelimbs.
For insects, the period from hatching until the start of the last
larval stage is analogous to amphibian premetamorphosis —
both larval tissues and imaginal primordia grow in a nutrient-
dependent fashion through this period. This ‘premetamorpho-
sis’ status is largely maintained by the presence of JH (see
below) and restricts ecdysteroid action to that of molting.
When the growth of the larva or nymph exceeds its threshold
size for metamorphosis [81], its next molt will be to the last larval
or nymphal instar. In the last larval instar of the Holometabola,
tissues undergo a switch called pupal commitment [64]. As in
prometamorphosis in amphibians, during pupal commitment
gene programs switch from simple growth to preparing tissues
for metamorphosis. These changes are seen in larval tissues
(for example [82]) and in imaginal discs and primordia that
switch to morphogenetic growth and pupal patterning [42].
This commitment to metamorphosis is irreversible and it may
occur in a mosaic fashion at different times across the tissues
of the larva.
In Drosophila, this period also involves a dramatic increase in
EcR, but by the start of metamorphosis, the imaginal discs and
the larval cells express different receptor isoforms [56]. These
different receptor isoforms reflect different roles for ecdysteroids
in the metamorphosis of these two classes of cells (Figure 3C).
Larval tissues show strong gene activation to initiate metamor-
phosis, and the lack of ecdysteroids or EcR or the presence of
a dominant-negative form of EcR (EcRDN) [83] prevents
metamorphosis. These tissues express EcR-B1, which has an
additional activation domain that supports high levels of gene in-
duction in response to 20E [59]. The metamorphic development
of imaginal discs, by contrast, is blocked by the lack of ecdyste-
roids or the presence of EcRDN, but occurs if EcR is removed,
even if steroid is absent [84].
For imaginal tissues, therefore, ecdysteroids mainly provide a
permissive signal that allows ongoing programs of tissue
patterning and cell specification to progress. These cells express
EcR-A, which lacks an amino-terminal activation domain. As
suggested in Figure 3C, the action of EcR-A likely emphasizes
the function of EcR–USP to support local chromatin remodeling,
thereby providing access for other developmental control path-
ways to drive the premetamorphic changes [85]. Indeed, a major
function of ecdysteroids in imaginal tissues is to remove the
repression imposed by its own receptor system. Consequently,
mutations that block ecdysone production result in much more
severe developmental deficits than do mutants that remove
either EcR or USP (for example [55]). Interestingly, a similar phe-
nomenon is seen in some tissues in Xenopus tadpoles in which
the removal of TRa allows some tissues to complete metamor-
phosis in the absence of thyroid hormone [86].
Juvenile Hormone and Control of the Nature of the Molt
The JHs (Figure 5A) control the entry into metamorphosis and
allow the ecdysteroid pulses to evoke different types of molts
[87]. The JHs are a family of sesquiterpenes, the most common
being JH-III, epoxymethylfarnesoate. Depending on species or
stage, the structure of JH varies in the number of the epoxide
groups and methyl versus ethyl side branches. JH is synthesized
by paired cephalic glands, the corpora allata. It has a wide variety
of functions, the chief of which being the control over the pro-
gression through the different life stages during growth and
metamorphosis [87]. It may also regulate aspects of reproduc-
tion [88,89], as well as controlling pigment polymorphisms, caste
determination in social insects, behavioral control in worker hon-
eybees, and some seasonal polymorphisms [63]. I will confine
my discussion of JH to its actions on growth and metamor-
phosis.
JH is a lipid-soluble hormone and readily penetrates the
cuticle. Low doses applied topically can locally suppress meta-
morphosis in the underlying epidermis, while the rest of the in-
sect moves ahead in its life history. This was strikingly illustrated
by the insect physiologist V.B. Wigglesworth, who wrote his
initials in nymphal cuticle on the back of metamorphosing
Rhodnius by the judicious application of a JH-containing extract
[90]. The ability of JH and its mimics to act through surface
Current Biology 29, R1252–R1268, December 2, 2019 R1259

A B
Brain Hemimetabolous Holometabolous
JH titer
CA JH-III
N-4 N-5 (last) Adult L-4 L-5 (last) Pupa
C JH D E JH
JH
hsp83
met
JH Kr-h1
tai met
JHRE
egfp dsRNA TcMet (RNAi) N5
precociously formed miniature pupae. From [1], copyright (2007) National Academy of Sciences, U.S.A. (E) Examples from Pyrrhocoris apterus showing that
RNAi-mediated knockdown of components in the JH response pathway in the 4 nymphal instar lead to formation of precocious adults rather than 5th instar
nymphs. Images from [97].
contact has made this hormone a major target for developing
specific agents for insect control [67].
The general pattern of the JH titer is fairly consistent during
embryonic and postembryonic life. During embryogenesis, JH
is sometimes present in the newly laid egg but it is rapidly metab-
olized and is absent through early to mid-embryogenesis, but
then appears after dorsal closure as the tissues of the fully-
grown embryo mature and the last embryonic cuticle is depos-
ited [91–93]. The subsequent growth of both larvae and nymphs
occurs in an environment of high circulating JH (Figure 5B). In
hemimetabolous insects, JH levels are high through the penulti-
mate nymphal stage, but then drop during the molt to the last
nymphal stage. JH is then absent through the molt to the adult.
In holometabolous insects, the pattern is similar in that JH is pre-
sent as larvae progress through their instars, but then disappears
early in the last larval instar to set the stage for the formation of
the pupa. JH then transiently reappears during the prepupal ec-
dysteroid pulse that directs the molt to the pupa, but then disap-
pears to allow adult differentiation.
The JH receptor
The developmental effects of JH functions are mediated through
an intracellular receptor from the basic helix-loop-helix Per/Arnt/
Sim (bHLH-PAS) transcription factor family [94]. It is related to
the vertebrate aryl-hydrocarbon receptor, but is the only family
member co-opted to serve as a hormone receptor. It is encoded
by the Methoprene tolerant (Met) gene, which was first identified
in Drosophila in a screen for mutants resistant to the JH mimic
methoprene [95]. Establishing that Met was a JH receptor was
complicated in Drosophila because this fly has a second recep-
tor gene, germ cell expressed (gce), which gave rise to Met by
gene duplication [96]. Demonstration that Met, indeed, mediates
the anti-metamorphic actions of JH came from insects that have
only one Met gene, such as the beetle Tribolium and the bug
Pyrrhocoris (Figure 5D,E). In both, Met knock-down by RNA
interference (RNAi) caused both premature metamorphosis
and the loss of sensitivity to exogenous JH [1,97]. Definitive proof
that Met and Gce are JH receptors then came using Drosophila
R1260 Current Biology 29, R1252–R1268, December 2, 2019
Current Biology
Review
Figure 5. The juvenile hormone (JH) system
of insects.
(A) JH is synthesized and released from the paired
Corpora Allata (CA) situated posterior to the brain.
JH-III is the most widely occurring JH. (B) Typical
JH titers for a hemimetabolous and a holometab-
Adult olous insects based on titers from locusts [138],
and the moth Manduca sexta [139,140]. Shaded
regions are periods of the molt. (C) The pathway
Met Kr-h1 for the developmental actions of JH [after 94]. JH
X X binds to its receptor, Met, which is bound to
cytoplasmic hsp83 in the absence of ligand. JH-
Met then translocates to the nucleus where it
dimerizes with Taiman (Tai) and binds to JH
response elements (JHRE). The main JH target for
developmental functions is the Krüppel-homolog
1 (Kr-h1) transcription factor. (D) Pupae of the
beetle Tribolium castaneum resulting from larvae
injected with double-stranded RNA in the 4th larval
Precocious adults Adult instar. Larvae receiving a control construct (egfp)
Current Biology undergo normal subsequent larval molts and
make normal sized pupae; those receiving Met
RNAi, mimicked the effects of JH removal and
th
genetics [98]. Furthermore, Met was shown to bind JH and JH
mimics at physiological concentrations, and the interaction of
the ligand with residues in the ligand-binding pocket has been
resolved for Gce [99].
Met (or Gce) heterodimerizes with Taiman, another bHLH-PAS
family member also known as Steroid Response Coactivator
(SRC) or b-Ftz-F1 Interacting Steroid Coactivator (FISC) [94]
(Figure 5C). This interaction is intriguing since Taiman also par-
ticipates in ecdysone action as part of a coactivator complex
with EcR–USP. The main target of JH–Met is the gene Krüppel
homolog 1 (Kr-h1), which encodes a C2H2 zinc finger transcrip-
tion factor. In many species, the removal of Kr-h1 mimics both
the loss of JH and the loss of Met (Figure 5E). JH is also thought
to have an as-yet unidentified membrane receptor that may be
especially important for some of the reproductive functions of
JH [94]. If this membrane receptor has a role in metamorphosis,
it is likely supportive to Met/Gce, as removal of the latter fully
mimics removal of JH.
The Actions of JH
The embryonic actions of JH are poorly understood. JH typically
appears during embryogenesis around the time of dorsal closure
as the embryo shifts from growth and patterning to differentiating
the tissues of the first instar nymph or larva [26]. Embryos vary
greatly, though, in their sensitivity to exogenous JH given before
this time. They appear to be insensitive to JH through segmenta-
tion and establishment of limb buds and other primordia, a
period that is concluded by the first embryonic molt. In embryos
from more basal groups, such as grasshoppers and crickets, JH
treatment after this period arrests morphogenesis and the next
ecdysone pulse induces premature tissue differentiation
[26,33]. This action of JH in suppressing tissue patterning is in
line with its morphostatic action on the imaginal primordia
described below. In contrast to what is seen with basal hemime-
tabolous species, on holometabolous embryos JH treatment has
comparatively mild effects, primarily being confined to suppres-
sion of morphogenetic movements (blastokinesis) within the egg
[100]. Indeed, most of the morphogenetic processes that are
Current Biology
Review
sensitive to JH treatment in grasshopper embryos have been
shifted into the postembryonic realm in the Holometabola, where
they still display their JH sensitivity.
One embryonic action of JH is opposite to what was ex-
pected from its postembryonic effects. As described below,
JH is generally considered to have a ‘status quo’ effect on
molting, causing the insect to produce the same type of cuticle
that it had had before. In embryonic crickets and grasshoppers,
though, JH treatment after the E1 molt results in a jump to a
nymphal (E3) molt, rather than repeating the E1 molt [26,33].
The reverse experiment of chemically preventing JH production
in embryos of the cockroach, Nauphoeta cinerea, had the
opposite effect of blocking the appearance of nymphal features
[101]. In Blattella germanica, the use of maternal RNAi to sup-
press the production of JH or its receptor produced a variety of
phenotypes, one of which being an arrest as an apparent pro-
nymph [102]. More needs to be done, but in the embryos of
these hemimetabolous insects JH appears to serve to promote
nymph development.
JH has two main categories of action during postembryonic
life. Its classic status quo action is linked to the ecdysteroid pulse
that causes a molt [103]. The presence or absence of JH at the
beginning of the ecdysteroid pulse affects the ‘commitment’ of
the cells, thereby determining whether the molt will be a progres-
sive molt to the next stage or a status quo molt repeating the
same stage [64]. Commitment occurs early in the preparatory
phase so that required cell divisions and the reprogramming of
response genes occurs before the high levels of 20E cause
cuticle production. A modified relationship is seen during the
larval–pupal transition, in which pupal commitment of larval cells
is caused by the small ecdysteroid pulse that terminates feeding
and induces a search for a pupation site. The pupal molt itself is
delayed for hours to days until the occurrence of the large prepu-
pal peak of ecdysteroids.
A second class of JH action has to do with controlling the form
that the insect will take at the next molt, the morphostatic action
of JH [104]. As detailed above, the larva carries imaginal
primordia either as invaginated imaginal discs or as regions of
diploid cells that have the capacity to partially dedifferentiate
and undergo pupal morphogenesis. The tonic presence of JH
in the intermolt periods in preterminal larval instars acts to pre-
vent these discs and primordia from initiating premature
morphogenesis [41].
In hemimetabolous insects, JH disappears late in the molt to
the last nymphal stage to allow the formation of the adult
(Figure 5B). Holometabolous insects also show the clearing of
JH from the circulation of last stage larvae. In the latter, though,
JH reappears during the prepupal ecdysteroid pulse. The reap-
pearance of JH has two functions. In some insects it ensures
that imaginal tissues will undergo a pupal molt rather than
skipping directly to the adult [103]. A second function relates
to the fact that some of the proliferation and cell patterning for
adult structures begins in the late larva and extends through
the pupal molt into the early phases of adult development,
when they can be turned off by high levels of 20E. Similar high
20E levels occur during the prepupal peak and the presence of
JH prevents these from permanently switching off proliferation
and patterning [105]. JH appears to play a similar role during pu-
pariation in Drosophila [106].
Variations in Requirements for JH
Insect larvae vary dramatically in how they respond to implanta-
tion of active corpora allata or treatment with exogenous JH.
Many beetle larvae and caterpillars respond by undergoing addi-
tional larval molts and becoming giant larvae. By contrast, the
same treatments will not cause larvae of Diptera to make an ex-
tra larval instar. In the former species, the number of larval instars
is somewhat plastic and can be modified by nutritional manipu-
lations, while in the latter the larval instar number is fixed. It has
become clear recently, though, that even in the species that
show classic JH-mediated plasticity, not all of the instars are
sensitive to JH [107–109].
Classic experiments in a blood-sucking bug, Rhodnius
prolixus, and the greater wax moth, Galleria mellonella, showed
that a first instar nymph or the epidermis from first instar larvae
could be caused to directly undergo metamorphosis if exposed
to the hormonal environment of the last nymphal or larval stage.
These results were the foundation of the idea that the presence
of JH was essential for every larval or nymphal molt. Recently,
though, it was shown that larvae of Bombyx mori which lacked
the JH receptor or the enzymes to make JH nevertheless pro-
gressed through the first two larval molts and only initiated meta-
morphosis after the 3rd larval instar [107,108]. These studies, and
similar ones in the bug Pyrrhocoris apterus [108], show that JH is
not required for the first few molts of the juvenile.
These results suggest that a metamorphosis-promoting factor
is released after the first few molts and, once it appears, JH is
then needed to suppress metamorphosis and maintain the juve-
nile stage. The instar at which this metamorphosis-promoting
factor first appears may determine whether or not there is JH-
controlled plasticity in the number of instars. For example, one
might speculate that in Drosophila such a factor does not appear
until after the larva has crossed the threshold size for metamor-
phosis and, hence, the last larval stage has already been estab-
lished. There is an increasing list of peptides that influence insect
growth [110], such as ones from brain, from the fat body, and
from growing imaginal discs. How these might relate to meta-
morphosis-promoting factors is not clear but will likely provide
an intriguing story for the future.
The progression through larval instars was thought to be very
different in nematodes as compared to insects, as in the former
the progression is controlled by a hard-wired genetic circuit of
heterochronic genes [111], while insects were thought to use
hormonal switches to progress through their series. It now ap-
pears that the early set of instars in insects are likely also
‘hard-wired’ while the last few instars, in which most of the
growth occurs, have a hormone-regulated plasticity to deal
with nutritional challenges. We need to reexamine if some of
the heterochronic genes of nematodes might also function in in-
sects to regulate the early, ‘hard-wired’ portion of their larval/
nymphal series.
Stage-Specification Genes and Life History Strategies
An important advance over the past decade has been the appre-
ciation of the role of stage-specification genes. By analogy to the
role of the homeotic genes that establish differences along the
spatial axis of the insect, the stage-specification genes establish
stage differences along its temporal axis. They are the embodi-
ment of ‘commitment’ and they determine the character of the
Current Biology 29, R1252–R1268, December 2, 2019 R1261
 Current Biology

Review

Ai ii B
Br
+ +

E93
+
L
Br/E93
P
E93
A L
X
Br/E93
L-A mosaic
JH
Kr-h1
iii iv Ecdysteroids Low 20E
MIFs?

D JH 20E
P2
C L8 L9 L10
JH/Met
L
X
Br/E93
L L L
X
P E93 A
Kr-h1 Early genes
L (High thresh)
C Br PN
L1 Early/late
N1 L1
E93 Broad Early genes
(Low thresh)
JH Late
Kr-h1
LF
LF Larval
Pupal
NF Pu Adult
Pu
Effector gene sets
E93
A A

Current Biology

Figure 6. Role of the stage-specification genes in directing insect life histories.

(A) RNAi-based knock-down of broad or E93 in Tribolium castaneum alters progression through the life cycle. (i) Relationship of broad (Br) and E93 to the normal
progression from larva (L) to pupa (P) to adult (A). (ii) Removal of Broad causes the last instar larva to molt to a larval-adult mosaic rather than to a pupa. Images in
(Ai) and (Aii) republished with permission of The Company of Biologists, from [119] ª 2008. (iii) Knockdown of E93 in the larva results in multiple larval molts rather
that forming a pupa (C). Adapted from [125]. (iv) Knockdown of E93 in the pupa results in a second pupa (P2) rather than an adult (from [3]). (B) Levels of expression
of broad (Br), E93, and Krüppel homolog 1 (Kr-h1) in Tribolium castaneum showing how their levels relate to JH and molting peaks of ecdysteroids (red and black
bars) and to interactions amongst the various genes (levels from [125,128]). Other metamorphosis-inducing factors (MIFs) may be involved in broad and E93
induction in some last larval instar tissues. (C) Summary of the shift in expression of Kr-h1, broad, and E93 in hemimetabolous (such as cockroach) and basal
(Tribolium) and derived (higher flies) holometabolous insects. Broad expression is a feature of nymphal (N) and pupal (Pu) stages and is maintained by JH acting
through Kr-h1. The preceding pronymph (PN) and larval (L) stages are characterized by little or no broad expression. E93 determines the adult stage in all cases
and serves as the gateway for metamorphosis for hemimetabolous species and basal holometabolans, but its gateway function is taken over by broad in derived
Holometabola. (D) Intersection of the pathways downstream of JH and of ecdysteroids to determine the stage-specific nature of the larval, pupal and adult molts.
The boxed genes are components of the Ashburner cascade that are common to all molts.

ensuing molt. These genes, summarized in Figure 6, typically
include Kr-h1, broad and Ecdysone-inducible protein 93F
(E93). All were initially identified in Drosophila and associated
with ecdysteroid action during metamorphosis. The function of
broad to establish the pupal stage was initially demonstrated in
Drosophila [112,113], but the scope of the functions of E93 and
Kr-h1 was only revealed by work on insects with less derived
development.
In holometabolous insects, Kr-h1, broad, and E93 direct the
genetic programs associated with the larva, pupa, and adult,
respectively. Broad (also called Broad Complex, Br-C) was
the first to be discovered and its action is best known. It be-
longs to the Bric-a-brac-Tramtrack-Broad zinc finger (BTB-
ZF) family of transcription factors. Depending on species, there
may be four to six main isoforms that share a common BTB
domain but vary in their paired C2H2 zinc fingers at their
carboxyl terminus [114]. The BTB domain can interact with
both co-repressors and co-activators as well as supporting
chromatin remodeling [115]. The precise interactions that
Broad uses, though, are not well understood. Broad protein ap-
pears as larval tissues become committed to pupal develop-
ment [116] and Drosophila mutants lacking all broad isoforms
stay as last stage larvae and do not initiate their premetamor-
phic programs [112]. Experiments in Drosophila show that
Broad suppresses both larval- and adult-related programs
[113,117] while activating genes specific to the pupal stage.
This dual action of Broad is strikingly evident in Tribolium, in
which Broad knock-down by RNAi results in the last stage larva
becoming a larva-adult mosaic rather than a pupa (Figure 6A)
[118,119].

R1262 Current Biology 29, R1252–R1268, December 2, 2019

Current Biology
Review
In hemimetabolous insects, broad is expressed during the
nymphal instars but then disappears in the last instar when the
nymph transforms into the adult [120–122]. This expression
pattern would seem to support the proposal that the nymphal
stage corresponds to the holometabolous pupa, but RNAi-
based knockdown of broad in early stage nymphs of the bugs
Oncopeltus fasciatus and Pyrrhocoris apterus and the cock-
roach Blattella germanica did not cause precocious metamor-
phosis. It did, though, prevent the nymphal wing pads from
shifting to a premetamorphic pattern of enhanced growth
[97,120,122], suggesting an ancestral role for broad in controlling
wing growth and morphogenesis. However, recent work on the
cricket, Gryllus bimaculatus, showed that knock-down of Broad
does, indeed, cause precocious metamorphosis [123]. The dif-
ferences between the effects in crickets versus the other species
may reflect differences in the role of broad or differences in the
effectiveness of RNAi-mediated gene knockdown. The nymphal
wing pads have the highest levels of broad expression [122] and
an incomplete knock-down in some species might affect the
wings but not the rest of the animal. Further work is obviously
needed to establish whether broad is the stage-specification
gene for the nymph.
E93 encodes a helix-turn-helix transcription factor. It was orig-
inally studied in Drosophila because it was a stage-specific early
puff that first appeared in the late prepupal salivary gland [124].
Gene knockdown experiments in both hemimetabolous and
holometabolous insects showed that E93 specifies the adult
stage [3]. In the absence of E93, nymphal cockroaches stay as
permanent nymphs. Similarly, in the beetle Tribolium the adult
form is not made if E93 expression is suppressed. Suppression
of E93 in the last instar larva results in continued larval molting
[125], while knock-down in the pupa results in the molt to another
pupal stage [3] (Figure 6A). E93 functions as a pioneer protein,
opening up sites throughout the chromatin to provide accessi-
bility to the gene sets needed to make the adult [126]. It also re-
presses broad expression by closing the chromatin around the
gene [126].
Kr-h1 is often also considered as the stage-specification gene
for the nymphal and larval states and, therefore, provides a
reason for equating nymphs and larvae [24,97,127]. Its involve-
ment in larval and nymphal molting, though, is in the context of
its function in mediating the action of JH [128,129]. The early
larval molts that do not require JH also do not require Kr-h1
[108], so Kr-h1 is not needed for larval molting per se. Also,
Kr-h1 later switches to controlling the formation of the pupa in
the cases in which JH reappears in the prepupa to prevent pre-
cocious adult formation (for example [130]).
Kr-h1, broad, and E93 expression are all sensitive to hormonal
signals and they also interact with each other (Figure 6B). As
noted above, Kr-h1 is the principal target of JH and levels of
Kr-h1 protein are often used as a proxy for the JH titer. Broad
and E93 expression are induced by ecdysteroids but they may
also be induced by other factors that appear during the last larval
instar (for example [41]). Kr-h1 suppresses E93 expression in
both cockroaches [24] and Tribolium [130]. The relationship of
JH/Kr-h1 to broad expression, though, is complex. JH maintains
broad expression in nymphs [120, 122], but in holometabolous
larvae JH inhibits the appearance of Broad [116]. But once larval
cells become pupally committed by broad induction, they now
act in a ‘nymphal’ fashion, because JH now maintains broad
expression. Indeed, pupae treated with JH maintain broad
expression and undertake a second pupal molt rather than
becoming an adult [113]. Depending on species, Broad and
E93 may inhibit each other [113,130]
Evolutionary Implications of the Stage-Specification
Genes and their Control
There has been a continuing controversy over how the nymphal
and adult stages of hemimetabolous forms evolved into the
larval, pupal and adult stages of the Holometabola. Broad and
E93 provide important insights into this controversy but, on first
sight, they lead to somewhat conflicting conclusions. The broad
data support the homology of the nymph to the pupa. The hemi-
metabolous pronymph and the holometabolous larva express
relatively little broad, but this gene is then expressed prominently
in the nymph and the pupa, respectively. The important issue
discussed above is whether broad acts as a stage-specification
gene for the nymph, as it does for the pupa. Resolution of this
point is needed to establish broad’s ancestral function and will
provide a clearer insight into the origin of the pupal stage.
The role of E93 has presented a somewhat different picture. It
is clear that E93 determines the adult stage in both hemimetab-
olous and holometabolous insects. In hemimetabolous nymphs,
JH, acting through Kr-h1, suppresses E93 and thereby maintains
the nymphal state. In the absence of Kr-h1, E93 is expressed and
causes adult differentiation. This has been referred to as the
MEKRE93 pathway [24], with E93 providing the gateway into
metamorphosis. This is also the case for the beetle Tribolium
castaneum. The drop in JH and Kr-h1 levels in the last stage larva
allows E93 expression, but the result is a pupa rather than an
adult. Referenced to the initial induction of E93, the beetle shows
a two-molt progression of larva to pupa and then pupa to adult
rather than only a single molt to the adult. This is consistent
with the idea that the pupa arose from the terminal molt being
divided into two parts to accommodate the transition from larva
to adult [23]. In the beetle, though, E93 expression is transient
because JH and Kr-h1 reappear during the larval–pupal transi-
tion to suppress E93 and allow broad to be expressed to cause
pupal differentiation. The disappearance of JH after pupal
ecdysis then allows E93 expression, the suppression of broad,
and adult differentiation.
Although E93 expression provides the entry into metamor-
phosis in Tribolium, it apparently does not serve this role in
more derived Holometabola, such as Drosophila. E93 still deter-
mines the adult stage but it is not expressed until late in the pre-
pupal period [131]. Besides specifying the pupal stage in flies,
broad has taken over the role of metamorphic gatekeeper as
shown by broad null mutants arresting as premetamorphic
larvae [112].
This shift from E93 to broad in controlling the entry into meta-
morphosis may explain the variation seen within the Holometa-
bola in the need for JH/Met/Kr-h1 signaling to make a pupa.
While lack of Kr-h1 expression during the larval–pupal transition
of Tribolium results in imaginal primordia undergoing adult,
rather than pupal, differentiation, a similar lack of JH in many
lepidopteran larvae, such as Bombyx mori, or in Drosophila,
does not. In Tribolium, the use of E93 to initiate metamorphosis
means that it must then be suppressed by the JH/Met/Kr-h1
Current Biology 29, R1252–R1268, December 2, 2019 R1263

pathway to allow broad expression and pupal development
[125]. With the delay of E93 expression in Drosophila, though,
JH is no longer needed during the larval–pupal transition to sup-
press precocious adult differentiation. JH is still present at that
period in flies, though, because of its second function of prevent-
ing the high 20E levels of the pupal molt from terminating meta-
morphic patterning programs [105,106]
Further Insights from Derived Life Cycles
Kr-h1, broad and E93 are also key genes involved in the highly
derived life cycles seen in some insects. Within the hemimetab-
olous orders, a holometabolous-like life history has evolved in
thrips (order Thysanoptera) and in some Hemiptera: the scale in-
sects (Coccidae), mealybugs (Pseudococcidae), and whiteflies
(Aleyrodidae). Thrips have two larval stages that feed, followed
by two to three non-feeding pupal stages before making the
adult. Broad expression is low in the first larval stage, but then
rises in the second larval instar in preparation for the production
of the propupa (the first pupal instar) and continues during the
production of the pupa [7]. In the hemipterans, only the males
show a complex metamorphosis. As in thrips, in the Japanese
mealy bug, Planococcus kraunhiae, broad transcripts appear
as the feeding ‘larva’ molts to a prepupa and then to a pupa
[132]. In both thrips [133] and mealy bugs [6], the expression of
E93 accompanies the expression of broad as they enter the pu-
pal series, but then E93 carries on alone in making the adult.
This parallel evolution of a holometabolous-like life cycle in the
Neometabola, a group of insects that are clearly members of the
hemimetabolous clade, may provide valuable insights into two
aspects of the changes that gave rise to the Holometabola.
Firstly, their metamorphosis from their larval to their adult stages
does not involve one pupa-like instar, but rather two to three
such instars. This differs from the idea that the terminal nymphal
instar became specialized as the homometabolous pupa (for
example [22]). The neometabolous strategy is more like our pro-
posal that the conversion of the pronymph into a feeding larval
stage resulted in selection pressure to concentrate feeding activ-
ity in the larva and make the nymph a transition stage [26]. This
transformation was then followed by a reduction in the number
of nymphal (pupal) instars to a single pupal instar evident in Hol-
ometabola today. Thrips may also be in the process of such a
reduction with one suborder (the Tubulifera) having three pupa-
like instars while the other (Terebrantia) has reduced the pupa
to two instars.
Although the males of the Hemiptera that show a neometabo-
lous lifestyle make a mobile, winged male, the females are neo-
tenous and remain sessile and larva/nymph-like. Typically, these
females reproduce asexually for part of the year and only start
producing males under times of stress. Expression of E93 is
very low in such females, associated with their failure to assume
an adult form [6].
Within the Holometabola, neoteny has also been developed in
the Strepsiptera, an order of parasitic insects that are related to
beetles and are parasitic on wasps and Hemiptera. The female
remains larviform and never leaves the host while the male
undergoes a normal metamorphosis. The male shows the ex-
pected expression of broad and E93 during its metamorphic pro-
gression while the female expresses little, if any, of these two
transcription factors [4,5].
R1264 Current Biology 29, R1252–R1268, December 2, 2019
Current Biology
Review
Conclusions
The lives of insects are dominated by molts which segment their
life history into discrete instars that allow growth and changes in
morphology. Each molt is initiated and coordinated by a large
pulse of ecdysteroids that act through a conserved genetic cir-
cuit, the Ashburner cascade, that translates features of the ste-
roid pulse into the synthesis of a new cuticle and the shedding
of the old one (Figure 6D). The character of the molt, though, is
determined by stage-specification genes that select the effector
gene sets that are acted on by the cascade. In the Holometabola,
the specification genes E93, and likely also broad, function as
pioneer proteins that open up large stretches of chromatin allow-
ing access to the transcription factors needed for adult and
pupal differentiation, respectively. JH, acting through the tran-
scription factor Kr-h1, exerts the major control over the stage-
specification genes. How the developmental hormones and
stage-specification genes relate to the evolution of the nymph
and adult pattern of hemimetabolous insects into the larva/
pupa/adult pattern of the Holometabola has been a persisting
issue of controversy.
The evolution of the homoletabolous larva involved
fundamental alterations in embryogenesis. In hemimetabolous
orders, embryogenesis typically generates a miniature approxi-
mation of the adult form. In the Holometabola, by contrast,
embryogenesis is altered to produce a larva whose form has
little relationship to that of the adult. Ancestral programs of
growth and patterning are truncated, and the resulting struc-
tures adapted for use in the new larval form. This truncation
also resulted in persisting primordia that undergo deferred
morphogenesis at the end of larval growth to make the first
approximation of the adult in the form of the pupa. The phases
of embryonic growth and patterning altered to make the larva
were those that occur during a cryptic embryonic stage, the
pronymph, that precedes the formation of the nymph and the
normal appearance of JH. Given the sensitivity of the pronymph
stage to precocious JH exposure, a heterochronic advance-
ment of JH into this earlier stage may have caused the
patterning arrest needed to form the larva. The pronymph
shows low levels of broad compared to the nymph and this
feature of little or no broad expression in the larva is consistent
with its origin from the pronymph stage.
Although there is conflicting evidence as to whether or not
broad is really a nymphal stage-specification gene, its expres-
sion is clearly a feature of nymphs (Figure 6C). JH, through
Kr-h1, maintains broad expression through the nymphal instars.
A similar relationship of Broad to JH is also evident in the holo-
metabolous pupa. Once larval cells become pupally committed
and begin expressing broad, JH can no longer turn off that
expression and, indeed, JH is necessary for broad expression
during the pupal molt to prevent precocious adult differentiation.
Moreover, treatment of pupae with JH supports a reinduction of
broad resulting in second pupa rather than an adult. What we do
not understand, though, is how JH suppresses broad expression
in larval stages while it promotes broad expression in nymphs.
The answer may lie in the control of broad expression in the pro-
nymph, a relationship that has not been examined.
E93 determines the adult stage in both hemimetabolous and
holometabolous insects, but its relationship to both broad and
JH (via Kr-h1) is complex and different features support
Current Biology
Review
alternative views to the origin of the pupa. The nymph-adult tran-
sition is simple: JH, acting through Kr-h1, suppresses E93 and
maintains broad; the decline in JH in the last nymphal instar al-
lows E93 expression, thereby both suppressing broad and
causing adult differentiation. This relationship is somewhat
modified in basal holometabolans, as represented by Tribolium.
Here the suppression of E93 by JH–Kr-h1 in the larva also con-
trols the entry to metamorphosis and this relationship is taken
as support that the larval–pupal transition corresponds to the
nymph–adult transition [24,97]. However, the repression of E93
by JH–Kr-h1 is seen again in the following molt of the pupa to
the adult. The pupa–adult molt is more like the last nymphal
molt, because Broad is now present and JH–Kr-h1 functions in
the pupa to maintain broad and inhibit E93. Including Broad in
the relationship indicates that the pupa–adult molt corresponds
to the nymph–adult molt.
To finally interpret the relationships that are seen at metamor-
phosis, though, we need more information on the hormonal regu-
lation of the embryonic molts of more basal species. Hormonal
relationships have continued to change after the evolution of
complete metamorphosis, as seen in highly derived groups
such as the higher Diptera (Figure 6C). In these insects the entry
to metamorphosis has now been given over to broad and E93
expression is delayed until the end of the pupal stage at the start
of adult differentiation. This shift has diminished the role of the
prepupal JH peak, and simplified the endocrine regulation over
metamorphosis.
ACKNOWLEDGEMENTS
I am grateful to Lynn Riddiford for many discussions during the preparation of
this article and for critical readings of the manuscript. Important feedback on
the manuscript was also provided by Marek Jindra, Joel Kingsolver, Stuart
Reynolds and anonymous reviewers. I thank Marek Jindra for supplying the
images for Figure 5D,E. Many of the ideas arose during the course of research
funded by National Science Foundation, the National Institutes of Health, and
the Howard Hughes Medical Institute.
REFERENCES
1. Konopova, B., and Jindra, M. (2007). Juvenile hormone resistance
gene Methoprene-tolerant controls entry into metamorphosis in
the beetle Tribolium castaneum. Proc. Natl. Acad. Sci. USA 104,
10488–10493.
2. Charles, J.P., Iwema, T., Epa, V.C., Takaki, K., Rynes, J., and Jindra, M.
(2011). Ligand-binding properties of a juvenile hormone receptor, Metho-
prene-tolerant. Proc. Natl. Acad. Sci. USA 108, 21128–21133.
3. Ureña, E., Manjón, C., Franch-Marro, X., and Martı́n, D. (2014). Transcrip-
tion factor E93 specifies adult metamorphosis in hemimetabolous and
holometabolous insects. Proc. Natl. Acad. Sci. USA 111, 7024–7029.
4. Erezyilmaz, D.F., Hayward, A., Huang, Y., Paps, J., Acs, Z., Delgado,
J.A., Collantes, F., and Kathirithamby, J. (2014). Expression of the pupal
determinant broad during metamorphosis and neotenic development of
the strepsipteran Xenos vesparum. PLoS One 9, e93614.
5. Chafino, S., López-Escardó, D., Benelli, G., Kovac, H., Casacuberta, E.,
Franch-Marro, X., Kathirithamby, J., and Martı́n, D. (2018). Differential
expression of the adult specifier E93 in the strepsipteran Xenos
vesparum Rossi suggests a role in female neoteny. Sci. Rep. 8, 14176.
6. Vea, I.M., Tanaka, S., Tsuji, T., Shiotsuki, T., Jouraku, A., and Minakuchi,
C. (2019). E93 expression and links to the juvenile hormone in hemipteran
mealybugs with insights on female neoteny. Insect Biochem. Mol. Biol.
104, 65–72.
7. Minakuchi, C., Tanaka, M., Miura, K., and Tanaka, T. (2011). Develop-
mental profile and hormonal regulation of the transcription factors broad
and Krüppel homolog 1 in hemimetabolous thrips. Insect Biochem. Mol.
Biol. 41, 125–134.
8. Nijhout, H.F., and Callier, V. (2014). Developmental mechanisms of body
size and wing-body scaling in insects. Annu. Rev. Entomol. 60, 141–156.
9. Chafino, S., Ureña, E., Casanova, J., Casacuberta, E., Franch-Marro, X.,
and Martı́n, D. (2019). Upregulation of E93 gene expression acts as the
trigger for metamorphosis independently of the threshold size in the bee-
tle Tribolium castaneum. Cell Rep. 27, 1039–1049.
10. Rousset, A., and Bitsch, C. (1993). Comparison between endogenous
and exogenous yolk proteins along an ovarian cycle in the firebrat
Thermobia domestica (Insecta.Thysanura). Comp. Biochem. Physiol.
104B, 33–44.
11. Kukalova-Peck, J. (1991). Fossil history and the evolution of hexapod
structures. In The Insects of Australia, Vol. 1 (Carlton: Melbourne Univer-
sity Press), 141–17.
12. Prokop, J., Percharová, M., Nel, A., Hörnschemeyer, T., Krzeminska, E.,
Krzeminski, W., and Engel, M.S. (2017). Paleozoic nymphal wing pads
support dual model of insect wing origins. Curr. Biol. 27, 263–269.
13. Belles, X. (2019). The innovation of the final moult and the origin of insect
metamorphosis. Philos. Trans. R. Soc. Lond. B 374, 20180415.
14. Yang, A.S. (2001). Modularity, evolvability, and adaptive radiations: a
comparison of hemi- and holometabolous insects. Evol. Dev. 3, 59–72.
15. Rainford, J.L., Hofreiter, M., Nicholson, D.B., and Mayhew, P.J. (2014).
Phylogenetic distribution of extant richness suggests metamorphosis is
a key innovation driving diversification in insects. PLoS One 9, e109085.
16. Nel, A., Roques, P., Nel, P., Prokin, A.A., Bourgoin, T., et al. (2013). The
earliest known holometabolous insects. Nature 503, 257–261.


17. Sehnal, F., Svácha, P., and Zrzavy, J. (1996). Evolution of insect meta-
morphosis. In Metamorphosis: Postembryonic Reprogramming of
Gene Expression in Amphibian and Insect Cells, L.I. Gilbert, J.R. Tat,
and B.G. Atkinson, eds. (San Diego, CA: Academic Press), pp. 3–58.
18. Erezyilmaz, D.F. (2006). Imperfect eggs and oviform nymphs: a history of
ideas about the origins of insect metamorphosis. Integr. Comp. Biol. 46,
795–807.
19. Reynolds, S.E. (2019). Cooking up the perfect insect: Aristotle’s transfor-
matopnal idea about the complete metamorphosis of insects. Philos.
Trans. R. Soc. Lond. B 374, 20190074.
20. Berlese, A. (1913). Intorno alle metamorfosi degli insetti. Redia 9,
121–136.
21. Imms, A.D. (1937). Recent Advances in Entomology (Philadelphia, PA:
Blakiston).
22. Hinton, H.E. (1963). The origin and function of the pupal stage. Proc. R.
Entomol. Soc. London Ser. A 38, 77–85.
orie de la nymphe des Insectes Hol-
23. Poyarkoff, E. (1914). Essai d’une the
taboles. Arch. Zool. Exp. Gen. 54, 221–265.
ome
24. Belles, X., and Santos, C.G. (2014). The MEKRE93 (Methoprene tolerant-
Krüppel homolog 1-E93) pathway in the regulation of insect metamor-
phosis, and the homology of the pupal stage. Insect Biochem. Mol.
Biol. 52, 60–68.
25. Jindra, M. (2019). Where does the pupa come from? The timing of
juvenile hormone signaling supports homology between stages of hemi-
metabolous and holometabolous insects. Philos. Trans. R. Soc. B 374,
20190064.
26. Truman, J.W., and Riddiford, L.M. (1999). The origins of insect metamor-
phosis. Nature 401, 447–452.
27. Truman, J.W., and Riddiford, L.M. (2002). Endocrine insights into the
evolution of metamorphosis in insects. Annu. Rev. Entomol. 47, 467–500.
28. Truman, J.W., and Riddiford, L.M. (2019). The evolution of insect meta-
morphosis: a developmental and endocrine view. Philos. Trans. R. Soc.
B 374, 20190070.
Current Biology 29, R1252–R1268, December 2, 2019 R1265

29. Cohen, S.M. (1993). Imaginal disc development. In The Development of
Drosophila melanogaster, Vol. 2, M. Bate, and A. Martinez-Arias, eds.
(Plainview, NY: Cold Spring Harbor Press), pp. 747–841.
30. Paulus, H.F. (1986). Evolutionswege zum Larvalauge der Insekten-ein
€ren Lateralaugen
Modell für die Entstehung und die Ableitung der ozella
der Myriapoda von Fazettenaugen. Zool. Jahrb. Syst. 113, 353–371.
31. Liu, Z., and Friedrich, M. (2004). The Tribolium homologue of glass and
the evolution of insect larval eyes. Dev. Biol. 269, 36–54.
32. Inoue, Y., Mito, T., Miyawaki, K., Matsushima, K., Shinmyo, Y., Heanue,
T.A., Mardon, G., Ohuchi, H., and Noji, S. (2002). Correlation of expres-
sion patterns of homothorax, dachshund, and Distal-less with the proxi-
modistal segmentation of the cricket leg bud. Mech. Dev. 113, 141–148.
33. Erezyilmaz, D.F., Riddiford, L.M., and Truman, J.W. (2004). Juvenile hor-
mone acts at embryonic molts and induces the nymphal cuticle in the
direct-developing cricket. Dev. Genes Evol. 214, 313–323.
34. Kojima, T. (2004). The mechanism of Drosophila leg development along
the proximodistal axis. Dev. Growth Differ. 46, 115–129.
35. Tanaka, K., and Truman, J.W. (2007). Molecular patterning mechanism
underlying metamorphosis of the thoracic leg in Manduca sexta. Dev.
Biol. 305, 539–550.
36. Kim, C.-W. (1959). The differentiation center inducing development from
larval to adult leg in Pieris brassicae (Lepidoptera). J. Embryol. Exp. Mor-
phol. 7, 572–582.
37. Tanaka, K., and Truman, J.W. (2005). Development of the adult leg
epidermis in Manduca sexta: contribution of different larval cell popula-
tions. Dev. Genes Evol. 215, 78–89.
38. Konopova, B., and Zrzavý, J. (2005). Ultrastructure, development, and
homology of insect embryonic cuticles. J. Morph. 267, 339–362.
39. Bernays, E.A. (1971). The vermiform larva of Schistocerca gregaria
(Forskål): form and activity (Insects, Orthoptera). Z. Morph. Tiere 70,
183–200.
40. Svácha, P. (1992). What are and what are not imaginal discs: reevaluation
of some basic concepts (Insecta, Holometabola). Dev. Biol. 154,
101–117.
41. Truman, J.W., Hiruma, K., Allee, J.P., MacWhinnie, S.G.B., Champlin,
D.T., and Riddiford, L.M. (2006). Juvenile hormone is required to couple
imaginal disc formation with nutrition in insects. Science 312, 1385–1388.
42. Mirth, K.C., Truman, J.W., and Riddiford, L.M. (2009). The Ecdysone re-
ceptor controls the post-critical weight switch to nutrition-independent
differentiation in Drosophila wing imaginal discs. Development 136,
2345–2353.
43. Lafont, R., Dauphin-Villemant, C., Warren, J.T., and Rees, H. (2012). In
Ecdysone chemistry and Biochemistry, Insect Endocrinology, and L.I.
Gilbert, eds. (Amsterdam: Elsevier Press), pp. 106–176.
44. Krishnakumaran, A., and Schneiderman, H.A. (1970). Control of molting
in mandibulate and chelicerate arthropods by ecdysones. Biol. Bull.
139, 520–538.
45. Qu, Z., Kenny, N.J., Lam, H.M., Chan, T.F., Chu, K.H., Bendena, W.G.,
Tobe, S.S., and Hui, J.H. (2015). How did arthropod sesquiterpenoids
and ecdysteroids arise? Comparison of hormonal pathway genes in
noninsect arthropod genomes. Genome Biol. Evol. 7, 1951–1959.
46. Smith, S.L., Bollenbacher, W.E., and Gilbert, L.I. (1983). Ecdysone 20-
monooxygenase activity during larval-pupal development of Manduca
sexta. Mol Cell Endocrinol. 31, 227–251.
47. Champlin, D.T., and Truman, J.W. (1998). Ecdysteroids govern two
phases of eye development during metamorphosis of the moth,
Manduca sexta. Development 125, 2009–2018.
48. Yamanaka, N., Rewitz, K.F., and O’Connor, M.B. (2013). Ecdysone con-
trol of developmental transitions: lessons from Drosophila research.
Annu. Rev. Entomol. 58, 497–516.
49. Tata, J.R. (2000). Autoinduction of nuclear hormone receptors during
metamorphosis and its significance. Insect Biochem. Mol. Biol. 30,
645–651.
R1266 Current Biology 29, R1252–R1268, December 2, 2019
Current Biology
Review
50. Brown, D.D., and Cai, L. (2007). Amphibian metamorphosis. Dev. Biol.
306, 20–33.
51. Warren, J.T., and Gilbert, L.I. (1986). Ecdysone metabolism and distribu-
tion during the pupal-adult development of Manduca sexta. Insect Bio-
chem. 16, 65–82.
52. Lagueux, M., Hetru, C., Goltzene, F., Kappler, C., and Hoffmann, J.A.
(1979). Ecdysone titre and metabolism in relation to cuticulogenesis in
embryos of Locusta migratoria. J. Insect Physiol. 25, 709–723.
53. Rewitz, K.F., O’Connor, M.B., and Gilbert, L.I. (2007). Molecular evolution
of the insect Halloween family of cytochrome P450s: phylogeny, gene or-
ganization and functional conservation. Insect Biochem. Mol. Biol. 37,
741–753.
54. Laudet, V., and Gronemeyer, H. (2002). The Nuclear Receptors Facts-
book. (London: Academic Press).
55. Koelle, M.R., Talbot, W.S., Segraves, W.A., Bender, M.T., Cherbas, P.,
and Hogness, D.S. (1991). The Drosophila EcR gene encodes an ecdy-
sone receptor, a new member of the steroid receptor superfamily. Cell
67, 59–77.
56. Talbot, W.S., Swyryd, E.A., and Hogness, D.S. (1993). Drosophila tissues
with different metamorphic responses to ecdysone express different
ecdysone receptor isoforms. Cell 73, 1323–1337.
57. Watanabe, T., Takeuchi, H., and Kubo, T. (2010). Structural diversity and
evolution of the N-terminal isoform-specific region of ecdysone receptor-
A abd -B1 isoforms in insects. BMC Evol. Biol. 10, 40.
58. Hill, R.J., Billas, I.M., Bonneton, F., Graham, L.D., and Lawrence, M.C.
(2013). Ecdysone receptors: from the Ashburner model to structural
biology. Annu. Rev. Entomol. 58, 251–271.
59. Riddiford, L.M., Cherbas, P., and Truman, J.W. (2000). Ecdysone recep-
tors and their biological actions. Vitam. Horm. 60, 1–73.
60. Sullivan, A.A., and Thummel, C.S. (2003). Temporal profiles of nuclear re-
ceptor gene expression reveal coordinate transcriptional responses dur-
ing Drosophila development. Mol. Endocrinol. 17, 2125–2137.
61. Andersen, S.O., Højrup, P., and Roepstorff, P. (1995). Insect cuticular
proteins. Insect Biochem. Mol. Biol. 25, 153–176.
62. Moussian, B. (2010). Recent advances in understanding mechanisms of
insect cuticle differentiation. Insect Biochem. Mol. Biol. 40, 363–375.
63. Nijhout, H.F. (1994). Insect Hormones (Princeton, NJ: Princeton Univer-
sity Press).
64. Riddiford, L.M. (1976). Hormonal control of insect epidermal cell commit-
ment in vitro. Nature 259, 115–117.
65. Williams, C.M. (1968). Ecdysone and ecdysone-analogues: their assay
and action on diapausing pupae of the Cynthia silkworm. Biol. Bull.
134, 344–355.
66. White, B.H., and Ewer, J. (2014). Neural and hormonal control of postec-
dysial behaviors in insects. Annu. Rev. Entomol. 59, 363–381.
67. Dhadiallah, T.S., Carlson, G.R., and Le, D.P. (1998). New insecticides
with ecdysteroidal and juvenile hormone activity. Annu. Rev. Entomol.
43, 545–569.
€nderungen in
68. Clever, U., and Karlson, P. (1960). Induktion von puff-vera
den Speicheldrüsenchromosomen von Chironomus tentans durch
Ecdyson. Exp. Cell Res. 20, 623–626.
69. Ashburner, M., Chihara, C., Meltzer, P., and Richards, G. (1974). Tempo-
ral control of puffing activity in polytene chromosomes. Cold Spring
Harb. Symp. Quant. Biol. 38, 655–662.
70. Hiruma, K., and Riddiford, L.M. (2010). Developmental expression of
mRNA for epidermal and fat body proteins and hormonally regulated
transcription factors in the tobacco hornworm, Manduca sexta.
J. Insect Physiol. 56, 1390–1395.
71. Mane-Padrós, D., Borràs-Castells, F., Belles, X., and Martı́n, D. (2012).
Nuclear receptor HR4 plays an essential role in the ecdysteroid-triggered
gene cascade in the development of the hemimetabolous insect Blattella
germanica. Mol. Cell. Endocrinol. 348, 322–330.
Current Biology
Review
72. King-Jones, K., and Thummel, C.S. (2005). Nuclear receptors – a
perspective from Drosophila. Nat. Rev. Genet. 6, 311–323.
73. Karim, F.D., and Thummel, C.C. (1991). Ecdysone coordinates the timing
and amounts of E74A and E74B transcription in Drosophila. Genes Dev.
5, 1067–1079.
74. White, K.P., Hurban, P., Watanabe, T., and Hogness, D.S. (1997). Coor-
dination of Drosophila metamorphosis by two ecdysone-induced nuclear
receptors. Science 276, 114–117.
75. Charles, J.P. (2010). The regulation of expression of insect cuticle protein
genes. Insect Biochem. Mol. Biol. 40, 205–213.
76. Wang, H.B., Moriyama, M., Iwanaga, M., and Kawasaki, H. (2010). Ecdy-
sone directly and indirectly regulates a cuticle protein gene, BMWCP10,
in the wing disc of Bombyx mori. Insect Biochem. Mol. Biol. 40, 453–459.
77. Cui, H.Y., Lestradet, M., Bruey-Sedano, N., Charles, J.P., and Riddiford,
L.M. (2009). Elucidation of the regulation of an adult cuticle gene Acp65A
by the transcription factor Broad. Insect Mol. Biol. 18, 421–429.
78. Woodard, C.T., Baehrecke, E.H., and Thummel, C.S. (1994). A molecular
mechanism for the stage specificity of the Drosophila prepupal genetic
response to ecdysone. Cell 79, 607–615.
79. Cho, K.H., Daubnerová, I., Park, Y., Zitnan, D., and Adams, M.E. (2014).
Secretory competence in a gateway endocrine cell conferred by the nu-
clear receptor bFTZ-F1 enables stage-specific ecdysone responses
throughout development in Drosophila. Dev Biol. 385, 253–262.
80. Brown, D.D. (2005). The role of deiodinases in amphibian metamor-
phosis. Thyroid 15, 815–821.
81. Nijhout, H.F., Riddiford, L.M., Mirth, C., Shingleton, A.W., Suzuki, Y., and
Callier, V. (2014). The developmental control of size in insects. Wiley In-
terdiscip. Rev. Dev. Biol. 3, 113–134.
82. Lepesant, J.A., Levine, M., Garen, A., Lepesant-Kejzlarvoa, J., Rat, L.,
and Somme-Martin, G. (1982). Developmentally regulated gene expres-
sion in Drosophila larval fat bodies. J. Mol. Appl. Genet. 1, 371–383.
83. Cherbas, L., Hu, X., Zhimulev, I., Belyaeva, E., and Cherbas, P. (2003).
EcR isoforms in Drosophila: testing tissue-specific requirements by tar-
geted blockade and rescue. Development 130, 271–284.
, C., Antoniewski, C., and Truman, J.W. (2005).
84. Schubiger, M., Carre
Ligand-dependent de-repression via EcR/USP acts as a gate to coordi-
nate the differentiation of sensory neurons in the Drosophila wing. Devel-
opment 132, 5239–5248.
85. Uyehara, C.M., and McKay, D.J. (2019). Direct and widespread role for
the nuclear receptor EcR in mediating the response to ecdysone in
Drosophila. Proc. Natl Acad. Sci. USA 116, 9893–9902.
86. Buchholz, D.R., and Shi, Y.-B. (2018). Dual function model revised by thy-
roid hormone receptor alpha knockout frogs. Gen. Comp. Endocrinol.
265, 214–218.
87. Jindra, M., Palli, S.R., and Riddiford, L.M. (2013). The juvenile hormone
signaling pathway in insect development. Annu. Rev. Entomol. 58,
181–204.
88. Raikhel, A.S., Brown, M.R., and Belles, X. (2005). Hormonal control of
reproductive processes. In Comprehensive Molecular Insect Science,
Vol 3, L.I. Gilbert, K. Iatrou, and S.S. Gill, eds. (Amsterdam: Elsevier
Press), pp. 433–491.
89. Wyatt, G.R., and Davey, K.G. (1996). Cellular and molecular actions of ju-
venile hormone. II. Roles of juvenile hormone in adult insects. Adv. Insect
Physiol. 26, 1–155.
90. Wigglesworth, V.B. (1958). Some methods for assaying extracts of the ju-
venile hormone in insects. J. Insect Physiol. 2, 73–84.
91. Bergot, B.J., Jamieson, G.C., Ratcliff, M.A., and Schooley, D.A. (1980).
JH Zero: new naturally occurring insect juvenile hormone from devel-
oping embryos of the tobacco hornworm. Science 210, 336–338.
92. Bergot, B.J., Baker, F.C., Cerf, D.C., Jamieson, G., and Schooley, D.A.
(1981). Qualitative and quantitative aspects of juvenile hormone titers in
developing embryos of several insect species: Discovery of a new JH-
like substance extracted from eggs of Manduca sexta. In Juvenile Hor-
mone Biochemistry. Action, Agonism and Antagonism, G.E. Pratt, and
G.T. Brookes, eds. (Amsterdam: Elsevier/North-Holland Biomedical
Press), pp. 33–45.
93. Temin, G., Zander, M., and Roussel, J.P. (1986). Physico-chemical
(GC-MS) measurements of juvenile hormone III titres during embryogen-
esis of Locusta migratoria. Int. J. Invert. Reprod. 9, 105–112.
s, X., and Shinoda, T. (2015). Molecular basis of juvenile
94. Jindra, M., Belle
hormone signaling. Curr. Opin. Insect Sci. 11, 39–46.
95. Wilson, T.G., and Fabian, J. (1986). A Drosophila melanogaster mutant
resistant to a chemical analog of juvenile hormone. Dev. Biol. 118,
190–201.
96. Baumann, A., Fujiwara, Y., and Wilson, T.G. (2010). Evolutionary diver-
gence of the paralogs Methoprene tolerant (Met) and germ cell expressed
(gce) within the genus Drosophila. J. Insect Physiol. 56, 1445–1455.
97. Konopova, B., Smykal, V., and Jindra, M. (2011). Common and distinct
roles of juvenile hormone signaling genes in metamorphosis of holome-
tabolous and hemimetabolous insects. PLoS One 6, e28728.
98. Jindra, M., Uhlirova, M., Charles, J.P., Smykal, V., and Hill, R.J. (2015).
Genetic evidence for function of the bHLH-PASS protein Gce/Met as a
juvenile hormone receptor. PLoS Genet. 11, e1005394.
99. Bittova, L., Jedlicka, P., Dracinsky, M., Kirubakaran, P., Vondrasek, J.,
Hanus, R., and Jindra, M. (2019). Exquisite ligand stereoselectivity of a
Drosophila juvenile hormone receptor contrasts with its broad agonist
repertoire. J. Biol. Chem. 294, 410–423.
100. Riddiford, L.M., and Williams, C.M. (1967). The effects of juvenile hor-
mone analogues on the embryonic development of silkworms. Proc.
Natl. Acad. Sci. USA 57, 595–601.
101. Brüning, E., and Lanzrein, B. (1987). Function of juvenile hormone III in
embryonic development of the cockroach, Nauphoeta cinerea. Int. J.
Invert. Reprod. 12, 29–44.
102. Fernandez-Nicolas, A., and Belles, X. (2017). Juvenile hormone signaling
in short germ-band hemimetabolan embryos. Development 144, 4637–
4644.
103. Williams, C.M. (1961). The juvenile hormone. II. Its role in the endocrine
control of molting, pupation, and adult development in the Cecropia silk-
worm. Biol. Bull. 121, 572–585.
104. Truman, J.W., and Riddiford, L.M. (2007). The morphostatic actions of ju-
venile hormone. Insect Biochem. Mol. Biol. I37I, 761–770.
105. Champlin, D.T., Reiss, S.E., and Truman, J.W. (1999). Hormonal control
of ventral diaphragm myogenesis during metamorphosis of the moth
Manduca sexta. Dev. Genes Evol. 209, 265–274.
106. Riddiford, L.M., Truman, J.W., Mirth, C.K., and Shen, Y.C. (2010). A role
for juvenile hormone in the prepupal development of Drosophila
melanogaster. Development 137, 1117–1126.
107. Daimon, T., Kozaki, T., Niwa, R., Kobayashi, I., Furuta, K., Namiki, T.,
Uchino, K., Banno, Y., Katsuma, S., Tamura, T., et al. (2012). Precocious
metamorphosis in the juvenile hormone-deficient mutant of the silkworm,
Bombyx mori. PLoS Genet. 8, e1002486.
108. Daimon, T., Uchibori, M., Nakao, H., Sezutsu, H., and Shinoda, T. (2015).
Knockout silkworms reveal a dispensable role for juvenile hormones in
holometabolous life cycle. Proc. Natl. Acad. Sci. USA 112, E4226–E4235.
109. Smykal, V., Daimon, T., Kayukawa, T., Takaki, K., Shinoda, T., and Jin-
dra, M. (2015). Importance of juvenile hormone signaling arises with
competence of insect larvae to metamorphose. Dev. Biol. 390, 221–230.
110. Droujinine, I.A., and Perrimon, N. (2016). Interorgan communication path-
ways in physiology: focus on Drosophila. Annu. Rev. Genet. 23, 539–570.
111. Faunes, F., and Larrain, J. (2016). Conservation in the involvement of het-
erochronic genes and hormones during developmental transitions. Dev.
Biol. 416, 3–17.
Current Biology 29, R1252–R1268, December 2, 2019 R1267

112. Kiss, I., Beaton, A.H., Tardiff, J., Fristrom, D., and Fristrom, J.W. (1988).
Interactions and developmental effect of mutations in the Broad-
Complex of Drosophila melanogaster. Genetics 118, 247–259.
113. Zhou, X., and Riddiford, L.M. (2002). Broad specifies pupal development
and mediates the ‘status quo’ action of juvenile hormone on the pupal-
adult transformation in Drosophila and Manduca. Development 129,
2259–2269.
114. Chaharbakhshi, E., and Jemc, J.C. (2016). Broad-complex, tramtrack,
and bric-à-brac (BTB) proteins: critical regulators of development. Gen-
esis 54, 505–518.
, G.G. (1998). Crystal structure of the
115. Ahmad, K.F., Engel, C.K., and Prive
BTB domain of PLZF. Proc. Natl. Acad. Sci. USA 95, 12123–12128.
116. Zhou, B., Hiruma, K., Shinoda, T., and Riddiford, L.M. (1998). Juvenile
hormone prevents ecdysteroid-induced expression of Broad Complex
RNAs in the epidermis of the tobacco hornworm, Manduca sexta. Dev.
Biol. 203, 233–244.
117. Riddiford, L.M., Hiruma, K., Zhou, X., and Nelson, C.A. (2003). Insights
into the molecular basis of the hormonal control of molting and metamor-
phosis from Manduca sexta and Drosophila melanogaster. Insect Bio-
chem. Mol. Biol. 33, 1327–1338.
118. Konopova, B., and Jindra, M. (2008). Broad-Complex acts downstream
of Met in juvenile hormone signaling to coordinate primitive holometabo-
lan metamorphosis. Development 135, 559–568.
119. Suzuki, Y., Truman, J.W., and Riddiford, L.M. (2008). The role of Broad in
the development of Tribolium castaneum: implications for the evolution of
the holometabolous insect pupa. Development 135, 569–577.
120. Erezyilmaz, D.F., Riddiford, L.M., and Truman, J.W. (2006). The pupal
specifier broad directs progressive morphogenesis in a direct-devel-
oping insect. Proc. Natl. Acad. Sci. USA 103, 6925–6930.
121. Piulachs, M.-D., Pagone, V., and Belles, X. (2010). Key roles of the Broad-
Complex gene in insect embryogenesis. Insect Biochem. Mol. Biol. 40,
468–475.
122. Huang, J.-H., Lozano, J., and Belles, X. (2013). Broad-complex functions
in postembryonic development of the cockroach Blattella germanica
shed new light on the evolution of insect metamorphosis. Biochim. Bio-
phys. Acta 1830, 2178–2187.
123. Ishimaru, Y., Tomonari, S., Watanabe, T., Noji, S., and Mito, T. (2019).
Regulatory mechanisms underlying the specification of the pupal-homol-
ogous stage in a hemimetabolous insect. Philos. Trans. R. Soc. Lond. B
374, 20190225.
124. Baehrecke, E.H., and Thummel, C.S. (1995). The Drosophila E93 gene
from the 93F early puff displays stage- and tissue-specific regulation
by 20-hydroxyecdysone. Dev. Biol. 171, 85–97.
125. Chafino, S., Ureña, E., Casanova, J., Casacuberta, E., Franch-Marro, X.,
and Martin, D. (2019). Upregulation of E93 gene expression acts as the
trigger for metamorphosis independently of the threshold size in the bee-
tle Tribolium castaneum. Cell Rep. 27, 1039–1049.
126. Uyehara, C.M., Nystrom, S.L., Niederhuber, M.J., Leatham-Jensen, M.,
Ma, Y., Buttitta, L.A., and McKay, D.J. (2017). Hormone-dependent con-
trol of developmental timing through regulation of chromatin accessi-
bility. Genes Dev. 31, 862–875.
R1268 Current Biology 29, R1252–R1268, December 2, 2019
Current Biology
Review
127. Lozano, J., and Belles, X. (2011). Conserved repressive function of Krüp-
pel homolog 1 on insect metamorphosis in hemimetabolous and holome-
tabolous species. Sci. Rep. 1, 163.
128. Minakuchi, C., Namiki, T., and Shinoda, T. (2009). Krüppel homolog 1, an
early juvenile hormone-response gene downstream of Methoprene-
tolerant, mediates its anti-metamorphic action in the red flour beetle Tri-
bolium castaneum. Dev. Biol. 325, 341–350.
129. Minakuchi, C., Namiki, T., Yoshiyama, M., and Shinoda, T. (2008). RNAi-
mediated knockdown of juvenile hormone acid O-methyltransferase
gene causes precocious metamorphosis in the red flour beetle Tribolium
castaneum. FEBS J. 275, 2919–2931.
130. Ureña, E., Chafino, S., Manjón, C., Franch-Marro, X., and Martin, D.
(2016). The occurrence of the holometabolous pupal stage requires the
interaction between E93, Krüppel-homolog 1 and Broad-Complex.
PLoS Genet. 12, e1006020.
131. Mou, X., Duncan, D.M., Baehrecke, E.H., and Duncan, I. (2012). Control
of target gene specificity during metamorphosis by the steroid response
gene E93. Proc. Natl. Acad. Sci. USA 109, 2949–2954.
132. Vea, I.M., Tanaka, S., Shiotsuki, T., Jouraku, A., Tanaka, T., and Minaku-
chi, C. (2016). Differential juvenile hormone variations in scale insect
extreme sexual dimorphism. PLoS One 11, e0149459.
133. Suzuki, Y., Shiotsuki, T., Jouraku, A., Miura, K., and Minakuchi, C. (2019).
Personal communication.
134. Wang, Y.-H., Engel, M.S., Rafael, J.A., Wu, H.-Y., Redei, D., Xie, Q.,
Wang, G., Liu, X.-G., and Bu, W.-J. (2016). Fossil record of stem groups
employed in evaluating the chronogram of insects (Arthropoda: Hexa-
poda). Sci. Rep. 6, 38939.
135. Wipfler, B., Letsch, H., Frandsen, P.B., et al. (2019). Evolutionary history
of Polyneoptera and its implications for our understanding of early
winged insects. Proc. Natl. Acad. Sci. USA 116, 3024–3029.
136. Labandeira, C. C. 1994. A compendium of fossil insect families. Milwau-
kee Public Museum Contributions, no. 88: 1–71.
137. Bollenbacher, W.E., Smith, S.L., Goodman, W., and Gilbert, L.I. (1981).
Ecdysteroid titer during larval-pupal-adult development of the tobacco
hornworm, Manduca sexta. Gen. Comp. Endocrinol. 44, 302–306.
138. Huibregtse-Minderhoud, L., Van den Hondel-Franken, M.A.M., Van der
Kerk-Van Hoof, A.C., Biessels, H.W.A., Salemink, C.A., Van der Horst,
D.J., and Beenakkers, A.M.T. (1980). Quantitative determination of the ju-
venile hormones in the haemolymph of Locusta migratoria during normal
development and after implantation of corpora allata. J. Insect Physiol.
26, 627–631.
139. Baker, F.C., Tsai, L.W., Reuter, C.C., and Schooley, D.A. (1987). In vivo
fluctuation of JH, JH acid, and ecdysteroid titer, and JH esterase activity
during development of fifth stadium Manduca sexta. Insect Biochem. 17,
989–996.
140. Fain, M.J., and Riddiford, L.M. (1975). Juvenile hormone titers in the he-
molymph during later larval development of the tobacco hornworm,
Manduca sexta (L.). Biol. Bull. 149, 506–521.