Download Development and evolution of the insect mushroom bodies: towards

Arthropod Structure & Development 32 (2003) 79–101
www.elsevier.com/locate/asd
Development and evolution of the insect mushroom bodies: towards the
understanding of conserved developmental mechanisms in a
higher brain center
Sarah M. Farris*, Irina Sinakevitch
Arizona Research Laboratories Division of Neurobiology, University of Arizona, 611 Gould-Simpson Building, Tucson, AZ 85721, USA
Received 6 January 2003; accepted 10 March 2003
Abstract
The insect mushroom bodies are prominent higher order neuropils consisting of thousands of approximately parallel projecting intrinsic
neurons arising from the minute basophilic perikarya of globuli cells. Early studies described these structures as centers for intelligence and
other higher functions; at present, the mushroom bodies are regarded as important models for the neural basis of learning and memory. The
insect mushroom bodies share a similar general morphology, and the same basic sequence of developmental events is observed across a
wide range of insect taxa. Globuli cell progenitors arise in the embryo and proliferate throughout the greater part of juvenile development.
Discrete morphological and functional subpopulations of globuli cells (or Kenyon cells, as they are called in insects) are sequentially
produced at distinct periods of development. Kenyon cell somata are arranged by age around the center of proliferation, as are their
processes in the mushroom body neuropil. Other aspects of mushroom body development are more variable from species to species, such as
the origin of specific Kenyon cell populations and neuropil substructures, as well as the timing and pace of the general developmental
sequence.
q 2003 Elsevier Ltd. All rights reserved.
Keywords: Acheta; Apis; Axon reorganization; Drosophila; Neurogenesis; Periplaneta
1. Introduction
The mushroom bodies were first described in honeybees
by Felix Dujardin (Dujardin, 1850). Later descriptions by
Flögel (1876) established the defining characteristics of
insect mushroom bodies: lobed structures in the supraesophageal ganglion consisting of the parallel fibers of
hundreds or thousands of tiny basophilic perikarya called
globuli cells. Mushroom body like structures have been
identified in species of Annelida, Onychophora and many
arthropod groups (Holmgren, 1916; reviewed in Strausfeld
et al. (1998)).
Felix Dujardin (1850) suggested that the mushroom
bodies were involved in ‘intelligent’ behavior due to their
particular prominence in the brains of the social Hymenoptera. More recent studies support this suggestion,
showing their involvement in context-dependent multimodal sensory integration (Schildberger, 1984; Li and
* Corresponding author. Tel.: þ 1-520-621-9668; fax: þ1-520-621-8282.
E-mail address: [email protected] (S.M. Farris).
1467-8039/03/$ - see front matter q 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S1467-8039(03)00009-4
Strausfeld, 1997, 1999), the prediction and monitoring of
motor behavior (Mizunami et al., 1998b; Okada et al., 1999)
and certain types of learning and memory (Zars et al., 2000;
Pascual and Préat, 2001). Vowles (1964) was the first to
show that mechanical lesions near the mushroom bodies of
the ant perturbed performance in a maze learning paradigm
based on olfactory discrimination. A possible role for the
mushroom bodies in olfactory and other types of learning
and memory led to the discovery of the first anatomical
mushroom body mutants in the fruit fly (Heisenberg et al.,
1985). The honeybee has also served as an important model
organism in learning and memory studies, due to its rich
repertoire of complex, learning-dependent behaviors in
nature as well as its tractability in controlled learning and
memory paradigms using the proboscis extension reflex (for
review, Menzel, 2001).
Despite many anatomical and physiological studies, few
general principles of mushroom body organization have
been proposed to explain the structure and function of these
centers. Developmental studies of the mushroom bodies
80
S.M. Farris, I. Sinakevitch / Arthropod Structure & Development 32 (2003) 79–101
have been performed in many insect species; however, such
studies have tended to proceed more or less independently.
An important exception is the early work of A.A. Panov,
published in Russian and often overlooked today, which
provides some of the first detailed accounts of mushroom
body neurogenesis and comparative aspects of mushroom
body development (Panov, 1957, 1966).
Comparative developmental studies have gained increasing popularity due to the insight that such an approach
can provide about the evolution of specific structures.
Fig. 1. Anatomy of the insect mushroom bodies. (A) Frontal section of the mushroom body of one hemisphere of the brain in the honeybee A. mellifera at the
level of the medial lobe (M), medial to the left. In this species the Kenyon cell somata (Kc) can be subdivided into three concentric subpopulations (oc, nc, ic)
based on size and location about each calyx (C). Kenyon cells provide dendrites into the calyces and their axons continue into the short pedunculus (P), where
they bifurcate to form the medial and vertical lobes. (B) Frontal section of the honeybee mushroom body at the level of the vertical lobe (V), which projects
from the pedunculus towards the anterior surface of the brain. (C) Frontal section of the mushroom body of one hemisphere of the brain in the American
cockroach P. americana. The same basic morphology is observed as in the honeybee, with Kenyon cells processes forming two calyces (only one visible) and
bifurcating in the pedunculus to form a medial and vertical lobe. Unlike A. mellifera, the cockroach vertical lobe projects dorsally rather than anteriorly, and a
distinct lobelet (L) made up of class III Kenyon cells is observed. (D) Sagittal section of the honeybee mushroom body, anterior to the left. Cc, central complex;
oc, outer compact Kenyon cells; nc, non-compact Kenyon cells; ic, inner compact Kenyon cells. Scale bars ¼ 100 mm.
S.M. Farris, I. Sinakevitch / Arthropod Structure & Development 32 (2003) 79–101
Consequently, developmental studies of the mushroom
bodies have again taken up the type of comparative
approach used by Panov, combining conventional histological methods with newer immunohistochemical analyses
to elucidate a variety of aspects of mushroom body
development in widely divergent insect species. The results
of such studies can be compared with those performed in
genetically tractable insects in order to determine molecules
and pathways involved in basic developmental processes
such as axon outgrowth and branching. Results from studies
of crickets, honeybees, cockroaches and flies reveal
ontogenetic events that are highly conserved across taxa,
suggesting fundamental organizational rules.
2. Anatomy of the insect mushroom bodies
The basic organization of the mushroom bodies is
retained in all insects with the exception of the Archaeognatha (Fig. 1). Mushroom body intrinsic neurons, or
Kenyon cells, are located in the dorso-posterior brain and
are easily recognizable by their tightly packed, cytoplasmpoor soma. The number of Kenyon cells can range from
thousands to hundreds of thousands depending on the
species. Kenyon cell neurites project antero-ventrally into
the brain, producing proximal to the cell body one or more
dendritic arborizations in neopteran insects. The neuropil
formed by these arborizations is termed the calyx. Although
numerous Kenyon cell subtypes have been identified in the
insects, most appear to fall into two broad categories:
clawed (Class II) and spiny or otherwise (Class I). The cupor bulb-shaped calyx receives olfactory afferents in most
species, although in the Hymenoptera it also receives visual
afferents. Each mushroom body may contain one or two
calyces; in most insects with two calyces, the Kenyon cell
populations and afferent input to each calyx appears to be
equivalent (Gronenberg, 2001). In the Orthoptera (locusts,
grasshoppers, katydids, crickets), however, each calyx
comprises a distinct population of Kenyon cells and as
such receives a separate set of afferents (Schürmann, 1973;
Weiss, 1977, 1981; Malaterre et al., 2002).
Generally, Kenyon cell processes project along the inner
surface of the calyx and at its center they funnel into the
neck of the pedunculus. Several separate tracts of neurites
can be visible in the most proximal region of the
pedunculus, but these tracts typically fuse as their projection
progresses antero-ventrally. Kenyon cell axons in the
pedunculus can give rise to both presynaptic specializations
(varicosities) and postsynaptic specializations (spines) and
can receive afferent neurons (Li and Strausfeld, 1997, 1999;
Strausfeld, 2002).
After projecting through the pedunculus, Kenyon cell
axons typically branch, although class II Kenyon cells can
provide exceptions as described below. Although conventionally referred to as the output region of the mushroom
bodies, Kenyon cell axons in the lobes have swellings and
81
spines and receive afferents as well as provide inputs onto
the dendrites of efferent neurons (Li and Strausfeld, 1997).
This is particularly evident in taxa lacking calyces, in which
the lobes are necessarily the only part of the mushroom body
to receive afferents (Strausfeld et al., 1998). Kenyon cell
processes in the lobes are therefore not axons in the strictest
sense, but for simplicity these ‘axons’ will be referred to as
such in this account.
Most Kenyon cell axons bifurcate with the resulting
branches projecting at approximately right angles to one
another. This bifurcation results in vertical (dorsal, a) and
medial (b) lobes that are typical of all insects. An exception
is seen in vespid wasps, where mushroom bodies appear to
have a single lobe due to the particularly convoluted
trajectory of the ‘medial’ axons branches (Ehmer and Hoy,
2000). A more common exception is observed in clawed
Kenyon cells, which in certain adult holometabolous insects
provide just a single vertically or medially directed axon
(Pearson, 1971; Lee et al., 1999; Strausfeld, 2002). In some
insects, different populations of Kenyon cells may form
several separate lobe systems that are designated by various
combinations of characters. In the fruit fly Drosophila
melanogaster Meigen (Diptera, Drosophilidae) three such
lobe systems are present, and are referred to as a/b, a0 /b0 and
g (Crittenden et al., 1998).
3. Early development and neurogenesis
3.1. Embryonic development of the mushroom bodies
Mushroom body development typically begins in the
embryo. In hemimetabolous insects, in which immature
nymphs behave in many ways like the adult insect, the
mushroom bodies of the newly hatched nymph resemble
tiny versions of those in the adult (Farris and Strausfeld,
2001; Malaterre et al., 2002). In the Holometabola,
however, the appearance of the mushroom bodies at
hatching is greatly variable from species to species.
Newly hatched larval Diptera such as D. melanogaster
and Phormia regina Meigen (Diptera, Calliphoridae) have
well-defined mushroom bodies with a calyx, pedunculus
and lobes (Gundersen and Larsen, 1978; Armstrong et al.,
1998). The monarch butterfly Danaus plexippus plexippus
L. (Lepidoptera, Danaidae), however, has at hatching only a
small number of Kenyon cells with axons forming a thin
pedunculus (Nordlander and Edwards, 1970); and the larva
of the honeybee Apis mellifera L. (Hymenoptera, Apidae)
has only progenitor cells at hatching and no identifiable
Kenyon cells until the third or fourth instar (Panov, 1957;
Farris et al., 1999).
Mushroom body progenitor cells (mushroom body
neuroblasts; MBNBs) are unique in that they are smaller
than other protocerebral neuroblasts, and may form discrete
glial-delimited aggregates in the dorsal protocerebrum
(Panov, 1957; Masson, 1970; Malun, 1998; Farris et al.,
82
S.M. Farris, I. Sinakevitch / Arthropod Structure & Development 32 (2003) 79–101
1999) (Fig. 2). In Diptera and Lepidoptera, in which MBNB
number is typically low, groups of progenitors are more
loosely bound and the progeny of each individual MBNB
may be discerned (Panov, 1957; Nordlander and Edwards,
1970; Gundersen and Larsen, 1978; Ito and Hotta, 1992).
The MBNBs appear to generate only mushroom body
intrinsic neurons (Kenyon cells) and glia (Ito and Hotta,
1992; Ito et al., 1997).
Neuroblast clusters are the first signs of the mushroom
bodies in the developing insect brain. The number of MBNB
aggregates per hemisphere is equal to the number of calyces
in the adult mushroom bodies (Panov, 1957); the number of
MBNBs per calyx ranges from one in the moth Ephestia
kuehniella Zeller (Lepidoptera, Pyralidae; Schrader, 1938 as
cited in Nordlander and Edwards (1970)) to as many as 500
in A. mellifera (Farris et al., 1999). In many species, each
MBNB or MBNB cluster eventually resides within its own
calyx as development proceeds. One partial exception to
this rule occurs in the walkingstick Carausius morosus
Brunner (Phasmida, Phasmatidae), in which two embryonic
calyces fuse into one but retain two separate MBNBs
(Malzacher, 1968). Although the adult A. mellifera mushroom bodies are composed of two calyces, Malun (1998)
reported that only a single MBNB cluster per hemisphere is
present in the first instar larva, with the second cluster
arising at the third larval instar. Panov (1957) and Farris
et al. (1999), however, identified two distinct MBNB
clusters per hemisphere in the first instar larva.
In the hemimetabolous insect Acheta domestica L.
(Orthoptera, Gryllidae; often incorrectly cited as Acheta
domesticus), MBNB clusters and associated cell bodies of
ganglion mother cells (GMCs) are first identified subsequent
to the appearance of the surrounding protocerebral neuroblasts, at about 57% development (Malaterre et al., 2002). In
Periplaneta americana L. (Dictyoptera, Blattidae) the
appearance of MBNBs occurs somewhat earlier, as early
Fig. 2. Developing mushroom bodies in the pupal brain of the ant Atta
cephalotes L. (Hymenoptera, Formicidae). The mushroom body neuroblasts (Nb) form two large aggregates, each residing within a single
developing calyx (C). Kenyon cell bodies (Kc) surround the neuroblasts. P,
pedunculus; M, medial lobe. Scale bar ¼ 50 mm.
as 30 –36% development (Salecker and Boeckh, 1995;
Farris and Strausfeld, unpublished observations). It is likely,
however, that the very first MBNBs arise significantly
earlier than has been reported, but specific markers for these
neuroblasts have not been developed in cricket and
cockroach. The progenitors are therefore not recognizable
until their numbers and those of their progeny are enough to
form a distinct aggregate. Therefore the origin of MBNBs in
hemimetabolous insects has yet to be determined.
In the holometabolous D. melanogaster, molecular
markers for MBNBs have been identified and used to
determine the exact origin of these progenitors during
embryogenesis. The genes eyeless and dachshund are
expressed in a proneural cluster of 10 –12 neuroblasts per
hemisphere, part of the Pc3 neuroblast group (Noveen et al.,
2000; Younossi-Hartenstein et al., 1996). This equivalence
group gives rise to four MBNBs per hemisphere at
embryonic stage 9 (Noveen et al., 2000). The MBNBs are
among the first progenitors to delaminate in the developing
brain, and begin producing progeny immediately. Interestingly, these first progeny never express mushroom body
markers and do not appear to contribute to these neuropils in
a manner typical of Kenyon cells (Noveen et al., 2000).
Production of Kenyon cells, as identified by their expression
of specific markers, does not occur until around stage 15.
This appears to contrast with the findings of Ito et al. (1997)
that the MBNBs generate only mushroom body intrinsic
neurons and glia; however, only MBNB progeny born after
larval hatching were analyzed in this study. The identity and
fate of the first MBNB progeny in the D. melanogaster
embryo is unknown at present.
The basic sequence of embryonic mushroom body
development is consistent between hemimetabolous and
holometabolous insects. The onset of Kenyon cell production is marked by the accumulation of minute, closely
packed soma surrounding or ventral to each MBNB cluster.
The neuropil of the embryonic mushroom bodies first
appears as a thin bundle of fibers extending into the
protocerebrum, which gradually thickens into a pedunculus
(Tettamanti et al., 1997; Noveen et al., 2000; Malaterre et al.,
2002; Farris and Strausfeld, unpublished observations). In
D. melanogaster the first Kenyon cell axons enter the
protocerebrum along the fasciclin II expressing pioneer cell
P41 (Noveen et al., 2000; Kurusu et al., 2002). The lobes
arise from the pedunculus shortly afterwards. The medial
lobe typically appears slightly before the vertical lobe
(Tettamanti et al., 1997; Noveen et al., 2000; Kurusu et al.,
2002; Malaterre et al., 2002) indicating a delay in Kenyon
cell branching during initial axonal outgrowth. The calyx is
the last neuropil region to form, appearing late in
embryogenesis or in the early nymph in the hemimetabolous
insects (Afify, 1960, as cited in Edwards, 1969); Panov,
1966; Farris and Strausfeld 2000; Kurusu et al. 2002;
Malaterre et al. 2002), but typically after larval hatching in
the holometabolous insects (Nordlander and Edwards, 1970;
Tettamanti et al., 1997; Farris et al., 1999) (Fig. 3). As
S.M. Farris, I. Sinakevitch / Arthropod Structure & Development 32 (2003) 79–101
growth continues in P. americana and A. mellifera, each
MBNB cluster resides within the walls of each of the four
calyces, eventually settling onto the ventral surface of the
calyx for the remainder of mushroom body development
(Farris and Strausfeld, 2001). In A. domestica, D. plexippus
plexippus and D. melanogaster this organization is reversed,
with the neuroblasts residing at the roof of the protocerebrum and separated from the more ventrally situated calyx
by an ever-increasing volume of Kenyon cell bodies (Ito and
Hotta, 1992; Malaterre et al., 2002). The relative location of
MBNBs within the developing calyces appears to be
characteristic of each insect order (S.M. Farris, unpublished
observations).
The basic sequence of early developmental events
described above is also followed in insects, such as D.
plexippus plexippus and A. mellifera, in which MBNBs are
present at larval hatching, but Kenyon cell production and
process outgrowth is mostly or entirely post-embryonic
(Farris et al., 1999). Two major differences in the general
developmental pattern are evident in these larval derived
mushroom bodies, perhaps as a result of the relative delay of
onset. First, the number of MBNBs increases 10-fold or
more through the larval instars as a result of symmetrical
divisions (Panov, 1957; Nordlander and Edwards, 1970;
Farris et al., 1999). In comparison, MBNB number in D.
melanogaster is known to stay constant throughout development (Truman and Bate, 1988), and in A. domestica only
a twofold increase in number is observed between hatching
83
and adulthood (Cayre et al., 1996; Malaterre et al., 2002). In
D. plexippus plexippus MBNB number increases from 3 to
30 per cluster (Nordlander and Edwards, 1970), while in A.
mellifera an initial 25 – 45 MBNBs per cluster in the first
instar larva generate a total of 2000 MBNBs in the brain of
the late fifth instar (Farris et al., 1999). It is proposed in the
case of A. mellifera that this enormous number of NBs is
necessary to generate, in ten days of larval and pupal
development, the estimated 170,000 Kenyon cells per
hemisphere that make up the honeybee mushroom bodies,
which are some of the largest described among the insects
(Witthöft, 1967; Farris et al., 1999).
The second difference in developmental events in insects
with larval derived mushroom bodies is that the formation
of the medial and vertical lobes does not appear to be
staggered. Nordlander and Edwards (1968, 1970) report
only that the lobes of D. plexippus plexippus grow
progressively thicker during development. The authors
note, however, that the calyx is the last neuropil component
to develop, exactly as is observed in embryonically derived
mushroom bodies. In A. mellifera as well, no apparent
disjunction in medial and vertical lobe formation is
observed but again calyx formation is clearly delayed
(Farris et al., 1999). Even in the fourth instar larva in which
Kenyon cell production has just begun, both a medial and a
vertical lobe are clearly present (Farris and Strausfeld,
unpublished data). It is possible, however, that were larval
brains examined at finer intervals during the critical stage in
Fig. 3. Early stages of mushroom body development in hemimetabolous and holometabolous insects. Frontal sections. (A) In the early embryo of P. americana
two neuroblast clusters (Nb) in the dorsal protocerebrum are surrounded by Kenyon cell bodies (Kc). Kenyon cell processes form a distinct pedunculus (P), but
no calyx at this early stage. (B) The fifth instar larva of A. mellifera has a similar mushroom body morphology, in which the pedunculus is clearly visible
beneath the neuroblast clusters, but the calyx has not yet formed. Scale bars A ¼ 25 mm, B ¼ 50 mm.
84
S.M. Farris, I. Sinakevitch / Arthropod Structure & Development 32 (2003) 79–101
neuropil development, the medial and vertical lobes would
be seen to arise at slightly staggered intervals as in the
embryonically derived mushroom bodies.
3.2. Post-embryonic neurogenesis and its regulation
Whether generating Kenyon cells or additional neuroblasts, the MBNBs are continuously mitotically active
throughout the greater portion of brain development (Panov,
1962; Truman and Bate, 1988). In D. melanogaster, the
brain of the newly hatched larva contains only a few
mitotically active progenitors, these being the optic
Anlagen, the lateral neuroblast (which contributes to the
antennal lobe) and the four MBNBs (Ito and Hotta, 1992;
Truman and Bate, 1988). This scarcity of neurogenesis at
larval hatching has allowed for relatively selective chemical
ablation of the mushroom bodies by feeding first instar
larvae doses of the DNA synthesis inhibitor hydroxyurea
(deBelle and Heisenberg, 1994). In all holometabolous
insects surveyed thus far the MBNBs do not enter a
quiescent stage but appear to divide continuously for the
duration of larval development (Nordlander and Edwards,
1970; Truman and Bate, 1988; Ito and Hotta, 1992; Datta,
1995; Farris et al., 1999). Mitotic activity continues until the
mid- to late pupal stage (Nordlander and Edwards, 1970;
Truman and Bate, 1988; Ito and Hotta, 1992; Farris et al.,
1999; Ganeshina et al., 2000). In A. mellifera the MBNBs
undergo programmed cell death in the mid-pupal stage and
disappear before adult eclosion (Fahrbach et al., 1995;
Farris et al., 1999; Ganeshina et al., 2000). The MBNBs also
appear to be lost during the pupal stage in D. melanogaster,
with scattered neurogenesis in the adult mushroom bodies
attributed to the last divisions of persisting GMCs (Technau,
1984; Ito and Hotta, 1992). Persistence of MBNBs and adult
neurogenesis in the mushroom bodies of holometabolous
insects has been reported only in species of Coleoptera
(Bieber and Fuldner, 1979; Cayre et al., 1996).
In hemimetabolous insects MBNB activity is also
constant throughout the juvenile stages and adult neurogenesis is more widespread (Cayre et al., 1996; Farris and
Strausfeld, 2001; Malaterre et al., 2002). MBNB activity in
the adult brain does not always produce Kenyon cells, as
newborn cells in the adult mushroom bodies of P.
americana and Locusta migratoria L. (Orthoptera, Acrididae) express glial cell markers (Cayre et al., 1996).
Members of the orthopteran family Gryllidae and the
cockroach Diploptera punctata Eschscholtz (Dictyoptera,
Blaberidae), however, continue adding Kenyon cells to the
mushroom bodies during adult life (Cayre et al., 1994, 1996;
Gu et al., 1999). In the case of A. domestica MBNB activity
after adult eclosion is extensive, contributing up to 20% of
the total Kenyon cell volume in 50 day old adults (Malaterre
et al., 2002).
Adult neurogenesis in insect mushroom bodies is
regulated by the actions of juvenile hormone (JH) and
ecdysteroids, as are many other aspects of nervous system
development. In A. domestica, in which neurogenesis in the
mushroom bodies continues for the duration of adult life,
mitotic activity of MBNBs decreases in response to
naturally occurring and artificially induced increases in
circulating ecdysone titers, and increases in the presence of
JH (Cayre et al., 1994, 1997a, 2000b). Both hormones
influence polyamine metabolism (Cayre et al., 1995,
1997a), and the polyamine putrescine has been shown to
be an important mediator of the mitogenic effect of JH
(Cayre et al., 1997b). Different polyamines have subsequently been proven to specifically induce neurogenesis
(putrescine) or cell differentiation (spermine and spermidine) in cultured A. domestica Kenyon cells (Cayre et al.,
2001).
In the blaberid cockroach D. punctata, the pattern of
adult mushroom body neurogenesis and its hormonal
control differs greatly from that of A. domestica. In D.
punctata adults, mitotic activity does not continue indefinitely but rather declines gradually after adult eclosion and
eventually ceases at about eight days of age (Gu et al.,
1999). During this short period of mitotic activity in D.
punctata JH has no apparent effect on neurogenesis in the
mushroom bodies. The ecdysteroid 20-HE enhances
proliferation, and 20-HE application in older insects can
also prolong MBNB activity beyond eight days of age (Gu
et al., 1999), a reverse effect from what is observed in A.
domestica. The contradictory roles of JH and ecdysteroids in
the two insect species may reflect the disparity in duration of
MBNB activity after adult eclosion. For example, novel
responses to JH and ecdysteroids may be necessary for the
MBNBs of A. domestica, which continue to divide through
multiple cycles of hormone fluctuations that are correlated
with the reproductive cycle.
Adult neurogenesis in A. domestica is also regulated by
interaction with the environment. In an invertebrate
correlate of the classic ‘enriched environment’ studies in
the rat (Volkmar and Greenough, 1972; Greenough and
Volkmar, 1973), housing of A. domestica nymphs in
complex environments was shown to increase Kenyon cell
production by MBNBs in the newly eclosed adult
(Scotto-Lomassese et al., 2000). The effect of environmental complexity on neurogenesis appears to be directly
mediated by sensory input, as antennal ablations and
removal of visual stimuli by painting the eyes removes the
capacity for increased neurogenesis in enriched rearing
conditions (Scotto-Lomassese et al., 2002). Environmental
regulation of MBNB activity is likely to be an important
component of mushroom body plasticity in A. domestica,
but the sparse distribution of adult neurogenesis across the
remaining insect orders indicates that this is not the case for
most insects. In these insects mushroom body plasticity
must instead be accomplished entirely by the modification
of existing neurons. For example, increased mushroom body
neuropil volume and Kenyon cell outgrowth in response to
the acquisition of foraging experience has been observed in
A. mellifera, in which adult neurogenesis is absent (Withers
S.M. Farris, I. Sinakevitch / Arthropod Structure & Development 32 (2003) 79–101
et al., 1993; Durst et al., 1994; Fahrbach et al., 1995;
Withers et al., 1995; Farris et al., 2001). Coss et al. (1980)
also described morphological changes in Kenyon cell
dendritic spines that appeared to be associated with foraging
behavior. A subsequent paper by Brandon and Coss (1982)
reported that these changes could be rapidly induced by a
single orientation flight. The limitations of the experimental
methods employed in these studies, however, prevent
definitive conclusions about the effects of foraging on
dendritic spine morphology. A later study on age- and
experience-related changes in honeybee Kenyon cell
dendrites failed to find any robust changes in spine
morphology after the first day of adult life (Farris et al.,
2001).
3.3. Generation and organization of Kenyon cell
subpopulations
Neurogenesis patterns during development provide clues
about the organization of Kenyon cells in the adult
mushroom bodies. In the mushroom bodies of A. mellifera,
three Kenyon cell types can be easily identified by the size
and location of their cell bodies (Panov, 1957; Mobbs, 1982)
(Fig. 1). They are arranged concentrically, such that one
subpopulation with small cell bodies resides in the center of
the calyx (inner compact), and the rest of the calyx is filled
by large cell bodies (non-compact). Another subpopulation
of Kenyon cells with small cell bodies lies outside of the
calyx (outer compact; after the terminology of Farris et al.,
1999). The characteristic cell body sizes are apparent during
their development, where it is observed that the outer
compact cells are produced first by the MBNBs, followed by
the non-compact (Panov, 1957; Farris et al., 1999). The
MBNBs of A. mellifera, as in other insects, reside in the
center of the calyx, but as development progresses their size
and number decreases and the center of the calyx becomes
populated by the last born inner compact Kenyon cells.
These observations in the bee revealed that Kenyon cell
bodies are arranged by birthdate with respect to the MBNB
cluster, and thus with respect to each calyx.
Experiments in which developing insects were treated
with markers of mitotic activity such as H3 thymidine or 50 bromo-2-deoxyuridine (BrdU) have confirmed these conclusions for A. mellifera and other insects. (Nordlander and
Edwards, 1970; Millikin and Stark, 1990; Ito and Hotta,
1992; Cayre et al., 1994; Farris et al., 1999; Cayre et al.,
2000a; Farris and Strausfeld, 2001; Malaterre et al., 2002).
Treatment with H3 thymidine or BrdU initially labels the
MBNBs and GMCs, but at progressively longer intervals
after treatment the marker is diluted out of the MBNBs by
repeated divisions. In these long term labeling experiments,
Kenyon cells produced at the time of treatment are observed
occupying correspondingly greater distances from the
proliferative center (Fig. 4). The labeled Kenyon cells are
separated from the MBNBs by the unlabeled cell bodies of
Kenyon cells born afterwards, pushed farther and farther
85
away in an apparently passive fashion as more neurons are
produced. The resulting organization of Kenyon cell bodies
is concentric and birthdate dependent, with the last born
Kenyon cells residing at the very center of the calyx, often
in the space formerly occupied by the MBNBs in insects
without adult neurogenesis, as in A. mellifera. The first born
Kenyon cells have their soma at the outermost margins of
the cell body region, even lying outside of the calyx cup in
insects such as P. americana and again in A. mellifera
(Farris et al., 1999; Farris and Strausfeld, 2001). Recent
studies in D. melanogaster in which single cell and
neuroblast clones were induced at different times in
development have again confirmed that Kenyon cells born
earlier in development occupy positions increasingly more
distal to the MBNBs as development proceeds (Kurusu et al.,
2002). This ordering of Kenyon cells about the calyx has
therefore been observed in all insects surveyed so far,
whether hemimetabolous or holometabolous. Birthdate
dependent ordering of cell soma with respect to the calyx,
generated by the passive movements of Kenyon cells away
from the MBNBs, is a highly conserved aspect of mushroom
body organization.
In A. mellifera there are at least three Kenyon cell
subpopulations with characteristic cell body morphologies
(Mobbs, 1982; Strausfeld, 2002). These concentric
populations are an obvious indicator that the Kenyon
cells are not an isomorphic pool of neurons. Kenyon cells
further vary with respect to their dendritic and axonal
morphology, neuropeptide contents and other aspects of
gene expression (Mobbs, 1982; Kucharski et al., 1998;
Farris et al., 1999; Strausfeld et al., 2000; Takeuchi et al.,
2001; Strausfeld 2002). Analyses of the progeny of
individual MBNBs in D. melanogaster show that all
MBNBs contribute to the generation of each Kenyon cell
subtype (Ito et al., 1997). As D. melanogaster possesses
only four MBNBs per hemisphere, the mushroom bodies
are thus composed of four identical ‘clonal units’, each
containing all Kenyon cell subpopulations and each
arising from a single neuroblast (Ito et al., 1997). This
quadripartite structure of the mushroom bodies of D.
melanogaster has also been revealed by the expression
patterns of P[GAL4] lines exhibiting mushroom body
labeling (Yang et al., 1995). In the calliphorid fly P.
regina, in which the mushroom bodies are generated by
five MBNBs, the resulting neuropil displays a fivefold structure (Gundersen and Larsen, 1978); however,
immunohistochemical studies of another calliphorid fly,
Phaenicia sericata Meigen (Diptera, Calliphoridae) reveal
four-fold symmetry in the mushroom bodies (Sinakevitch
and Strausfeld, 2002). MBNB number has not been
determined in P. sericata; if four MBNBs are present as
in D. melanogaster, a similar four-fold symmetry of the
mushroom bodies may be expected. Further comparative
studies will therefore be necessary to clarify the
relationship between neuroblast number and neuropil
organization in the Diptera.
86
S.M. Farris, I. Sinakevitch / Arthropod Structure & Development 32 (2003) 79–101
Kenyon cell subpopulations are each born during a
distinct developmental period. This is reflected by the
organization of their cell bodies into discrete groups
around the calyx. Due to their distinct cell body
morphology, Kenyon cell subpopulations in A. mellifera
suggest that the generation of one cell type begins only
when the generation of the previous cell type is
completed (Panov, 1957; Farris et al., 1999). For
example, the onset of generation of non-compact cells
is clearly marked by the appearance of their large cell
bodies in the prepupal stage; treatment of prepupae with
BrdU labels only this cell population (Farris et al.,
1999). In A. domestica, all of the large Kenyon cells
that supply the posterior calyx are born prior to
hatching, as are the Class II and Class III Kenyon
cells of P. americana (Panov, 1966; Farris and
Strausfeld, 2000; Farris and Strausfeld, 2001; Malaterre
et al., 2002; Farris and Strausfeld, 2003). Analysis of
multicellular and single Kenyon cell clones generated at
different time points in mushroom body development
have confirmed the sequential generation of intrinsic
neuron subtypes in D. melanogaster (Armstrong et al.,
1998; Lee et al., 1999). Specifically, Kenyon cells that
make up the g lobe of the adult are generated from
embryogenesis until the middle of the third instar.
Those providing the a0 /b0 lobes are generated in the late
larval stage, and those providing the a/b lobes appear
after puparium formation. As these transitions coincide
with the prepupal ecdysone peak and the pupariation
event (also a period of high ecdysone), respectively, Lee
et al. (1999) have suggested that cell type switching by
neuroblasts may be under control of this hormone. The
sequence of lobe neuropil formation is thus agedependent as well, and represents another highly
conserved characteristic of the developmental organization of the mushroom bodies (See ‘Process Outgrowth
and Plasticity’ below). The early-generated Kenyon
cells of the D. melanogaster g lobe, also known as
clawed or Class II Kenyon cells, have been identified in
a number of insect species and are typically the first
born mushroom body intrinsic neurons (Malzacher,
1968; Farris et al., 1999; Farris and Strausfeld, 2001;
Malaterre et al., 2002). The later born Class I or spiny
Kenyon cells appear to make up the remainder of the
mushroom bodies. Each Kenyon cell subpopulation is
thus defined not only in by the shared morphology of its
Fig. 4. Kenyon cells born at different times in development demonstrating
age-dependent concentricity in the adult honeybee mushroom bodies.
Frontal sections of a single mushroom body calyx. (A) BrdU label applied
to feeding fifth instar larva is localized in the outer compact cells (oc,
arrows), indicating that these cells were born at the time of BrdU treatment.
(B) BrdU applied to the D1 pupa marks the transition from production of
non-compact (nc) to inner compact (ic) Kenyon cells (arrows). (C) BrdU
labeling of the D5 pupa marks the last-born inner compact Kenyon cells at
the center of the calyx (arrow). Scale bars ¼ 50 mm.
S.M. Farris, I. Sinakevitch / Arthropod Structure & Development 32 (2003) 79–101
87
neurons but by their time of generation during mushroom body development.
3.4. Conserved and divergent aspects of neurogenesis in the
mushroom bodies
The basic progression of Kenyon cell generation during
mushroom body development is conserved across a wide
variety of insect taxa. Kenyon cells are produced by 2 –4
aggregates of MBNBs, the number of which corresponds to
the total number of calyces. Cell bodies are concentrically
arranged by birthdate around the proliferative center, such
that the earliest born cells are found at the outer edges of the
calyx and the latest at the center, displacing the MBNBs in
insects which lose these progenitors later in life. Distinct
Kenyon cell types are produced during discrete blocks of
time at specific periods in development, with the apparently
ubiquitous clawed (g) Kenyon cells typically being
produced first.
An important difference in the general progress of
neurogenesis between described species is the timing of
specific developmental events. It is possible that these
differences in the developmental timeline reflect the variety
of early life histories and behavioral repertoires of the
species studied (Farris et al., 1999; Farris and Strausfeld,
2001). In hemimetabolous insects the nymph must search
for food and shelter immediately after hatching, thus
necessitating the presence of functioning higher brain
regions at an early stage in development. The embryonic
stage is thus elongated and the brain resembles a smaller
version of that of the adult at hatching. Holometabolous
insect larvae, in contrast, typically hatch on or in their
source of food and shelter and perform few non-feeding or
defensive behaviors within the first few days. Such simple
behaviors may not require the mushroom bodies; indeed,
sensory centers such as the eyes and antennal lobes are often
rudimentary in newly hatched larvae, and the necessity of a
functioning higher sensory integration center is thus
negligible at this time. Mushroom body development can
thus be relatively incomplete at hatching and can be
protracted into the larval and the even more behavior-poor
pupal stage. This developmental delay is taken to its
extreme in the social Hymenoptera, which are completely
cared for by adult workers and correspondingly have in their
early instars poorly developed brains in which Kenyon cell
production does not begin until several days after hatching
(Panov, 1957; Farris et al., 1999) (Fig. 5).
Other differences in neurogenesis patterns involve the
number of MBNBs and the duration of their proliferative
activity. Such variables are presumably tailored to the
unique structural characteristics of the mushroom bodies of
an individual insect, which may in turn be influenced by
behavioral ecology. For example, the largest numbers of
MBNBs are found in the social Hymenoptera, which have
large mushroom bodies consisting of hundreds of thousands
of Kenyon cells. Perhaps, as was initially proposed by
Fig. 5. Brain of a newly hatched hemimetabolous insect compared with that
of a newly hatched holometabolous insect. (A) Frontal section of the brain
of the first instar cockroach nymph revealing well-developed mushroom
bodies as evidenced by the prominent vertical (V) and medial (M) lobes.
The central complex (Cc) and antennal lobe (Al) neuropils are also visible
at this early stage of juvenile development. (B) Frontal section of the brain
of the first instar honeybee larva showing a relatively undifferentiated
protocerebral neuropil (Pc) with no identifiable mushroom bodies or central
complex. The antennal and optic lobe neuropils have not yet formed, the
latter represented only by the dense optic anlagen (Oa). Scale bars
A ¼ 100 mm, B ¼ 50 mm.
Dujardin (1850), the behavioral demands of social behavior
necessitate the development of a particularly complex
sensory integration center. The persistence of MBNB
activity in the imago of some insect species may similarly
be influenced by life history, perhaps being most common in
long-lived species that spend a significant portion of their
lives in the adult stage. Future comparative studies of
mushroom body neurogenesis will be necessary in order to
determine how insect life history and neurogenesis in a
complex brain region are correlated.
88
S.M. Farris, I. Sinakevitch / Arthropod Structure & Development 32 (2003) 79–101
4. Process outgrowth and plasticity
4.1. Formation of the pedunculus and lobes
The axons forming the layered pedunculus and lobes
transmute from a concentric to a laminar organization as
they proceed distally from the cell soma. Kenyon cell
neurites and axons are organized concentrically within the
calyx and/or proximal pedunculus, but typically unfold into
a flat laminar organization in the distal pedunculus and
lobes. The point at which this concentric to laminar
transition occurs varies from species to species. In P.
americana and A. mellifera the transition to laminae occurs
almost immediately beneath the calyx (Strausfeld and Li,
1999; Strausfeld 2002), while in D. melanogaster the
concentric organization unfolds further distal, forming three
or more individual lobe systems that are arranged anterior to
posterior as are laminae in larger insects (Strausfeld et al.,
2003). In A. domestica and Schistocerca gregaria Forskal
(Orthoptera, Acrididae) the axons of the vertical lobe
remain concentrically organized, while those in the medial
lobe unfold into a more typical laminar arrangement
(O’Shea et al., 1998; Malaterre et al., 2002).
Annuli and laminae in the pedunculus and lobes are
added sequentially during mushroom body development.
Annuli are arranged by decreasing age from the outside of
the neuropil inwards (Yang et al., 1995; Tettamanti et al.,
1997; Armstrong et al., 1998; Kurusu et al., 2002; Malaterre
et al., 2002), while laminae are arranged from anterior
(oldest) to posterior (newest; Farris and Strausfeld, 2001;
Farris and Strausfeld, unpublished data). The axons of
newborn Kenyon cells enter the pedunculus and lobes via a
Fig. 6. Taurine immunoreactivity (A) –(C) and aspartate immunoreactivity (D), (E) in the juvenile and adult cockroach reveal the gradual addition of laminae in
the lobes. Sagittal sections. Taurine immunoreactivity in the medial lobe of the first instar nymph (A), fourth instar nymph (B) and the P. americana adult (C)
shows an increase in number of immunoreactive laminae with increasing age. Aspartate immunoreactivity in the adult vertical lobe (D) and the fourth instar medial
lobe (E) shows a different pattern of laminar staining from that of taurine, but a similar increase in the number of laminae with age. In all cases, the ingrowth lamina
(arrows) contains relatively less taurine than Kenyon cells of other laminae; however, aspartate content in this lamina during development can be more variable.
Immunostaining performed as described by Farris and Strausfeld (2001) and Sinakevitch et al. (2001). Scale bars A, B, C, E ¼ 20 mm, D ¼ 25 mm.
S.M. Farris, I. Sinakevitch / Arthropod Structure & Development 32 (2003) 79–101
discrete tract and are phenotypically quite distinct from
mature Kenyon cells. This ‘core’ is found in the center of
concentric neuropils, and is shifted gradually more posterior
as the concentric arrangement unfolds into laminae in the
distal pedunculus and lobes. In laminar neuropils, the core
forms the posteriormost ‘ingrowth’ lamina (Farris and
Strausfeld, 2001; Kurusu et al., 2002; Strausfeld et al., 2003)
In P. americana, the anteriormost and earliest identified
lamina in the lobes (termed the g layer) is composed of the
branched axons of the embryonically produced clawed
Kenyon cells (Strausfeld and Li, 1999; Farris and Strausfeld, 2001; Sinakevitch et al., 2001). Additional laminae are
added to the posterior edge of the lobe one by one
throughout nymphal development via the ingrowth lamina
(Farris and Strausfeld, 2001; Sinakevitch et al., 2001) (Fig.
6). In D. melanogaster, the axons of the early-born clawed
Kenyon cells make up the outer sheath of the concentric
pedunculus, with progressively younger annuli added
internally due to the ingrowth of axons through the
centrally-located core (Armstrong et al., 1998; Verkhusha
et al., 2001; Kurusu et al., 2002). Further distal, the
pedunculus unfolds into three major lobe systems (g, a0 /b0
and a/b) that are arranged by birthdate from anterior to
posterior (Crittenden et al., 1998; Lee et al., 1999). The g
lobe occupies the most anterior position exactly as does the
g layer in the laminar medial and vertical lobes of larger
insects such as A. mellifera and P. americana (Strausfeld
and Li, 1999; Strausfeld 2002).
Kenyon cells with axons in the ingrowth core or lamina
are distinct from those forming older laminae, likely due to
their immature nature. The axons occupying the core are
characteristically delicate, and in P. americana are decorated with numerous growth cone-tipped filaments (Technau and Heisenberg, 1982; Farris and Strausfeld, 2001;
Sinakevitch et al., 2001; Kurusu et al., 2002). Newborn
Kenyon cells do not express most markers common to
mature Kenyon cells (Armstrong et al., 1998; Crittenden
et al., 1998; Farris and Strausfeld, 2001; Sinakevitch et al.,
2001). The extending axons, however, contain large
amounts of f-actin and therefore have a high affinity for
the mushroom toxin phalloidin (Kurusu et al., 2002). In the
D. melanogaster calyx, four tracts are observed, one from
each neuroblast clonal group; these four tracts fuse into one
core as they progress into the pedunculus and lobes. As
would be expected, application of this non-species specific
label to the developing mushroom bodies of P. americana
and other insects invariably identifies Kenyon cell axons
making up a discrete ingrowth lamina or core (Fig. 7(C), (D)
and (G)). As in D. melanogaster, a number of smaller tracts
are typically seen to exit the calyces, fusing distally into a
single tract as they progress deeper into the pedunculus.
Exceptions to the fusing of core fibers into a single tract
occur in some Orthoptera and Coleoptera. In coleopteran
species, the core fibers from each calyx remain as two
separate tracts throughout the pedunculus and lobes
(Bretschneider, 1914; Larsson et al., 2002). In the mush-
89
room bodies of S. gregaria the core splits into six separate
tracts in the vertical lobe and forms a broad ingrowth lamina
in the posterior medial lobe (O’Shea et al., 1998; S.M. Farris
and I. Sinakevitch, unpublished observations). Regardless
of the specific morphology, the core region of each species
occupies the same relative region of the pedunculus and
lobes throughout development, therefore providing the basis
for an age-dependent organization of axons in the mushroom body lobes of a wide variety of insect taxa.
4.2. Unusual characteristics of core Kenyon cells
Newborn Kenyon cells are therefore distinct from mature
Kenyon cells, but there is some evidence that they are
already taking on complex functions at this early stage in
maturation, and perhaps contributing to the organization of
other neurons. In P. americana, many of the axons in the
ingrowth lamina produce a dense system of collaterals that
extend roughly perpendicular to the long axis of the medial
and vertical lobes (Farris and Strausfeld, 2001) (Fig.
7(A) –(C)). These collaterals are only produced by Kenyon
cells with axons in the ingrowth lamina, and are readily
visualized with phalloidin as are the main axons. Ingrowth
lamina collaterals apparently identical to those in P.
americana can be observed in other members of the
Blattaria and in basal Isoptera, and a dense network of
fibers reminiscent of the dictyopteran collateral system is
observed to surround the core in A. domestica (S.M. Farris,
unpublished observations). In the holometabolous insects,
core fiber collaterals are particularly evident in the medial
lobe of A. mellifera, and minute collaterals have been
observed in D. melanogaster (Strausfeld et al., 2003). These
structures therefore appear to represent a common morphological characteristic of newborn Kenyon cells. The
function of collaterals is unknown, although they have
been proposed to serve as guidance cues for lobe extrinsic
neurons, which must constantly reorganize their processes
in order to accommodate Kenyon cells added throughout
nymphal development in hemimetabolous insects. The
collaterals, which extend from the ingrowth lamina to the
anterior surface of the lobes where extrinsic neurons enter,
may provide a scaffold along which these neurons can grow
and make contact with newly added Kenyon cell axons
(Farris and Strausfeld, 2001). Whether collaterals perform a
similar function in holometabolous insects, in which
mushroom body development occurs over a much shorter
time span, remains to be determined.
Although axons of the ingrowth lamina or core are
typically characterized by an apparent lack of expression of
most mushroom body markers, they do show a high affinity
for the putative neurotransmitter glutamate (Sinakevitch
et al., 2001) (Fig. 7(B), (E), (F) and (H)). In P. americana,
the ingrowth lamina and collateral system is strongly
labeled by anti-glutamate antibody; such labeling is not
observed in any other Kenyon cell population. Antiglutamate immunoreactivity is also observed in core fibers
90
S.M. Farris, I. Sinakevitch / Arthropod Structure & Development 32 (2003) 79–101
S.M. Farris, I. Sinakevitch / Arthropod Structure & Development 32 (2003) 79–101
of the holometabolous insects A. mellifera and D.
melanogaster, (Strausfeld et al., 2003) indicating that this
is a widespread characteristic of newborn Kenyon cells.
Interestingly, in A. domestica, glutamate immunoreactivity
is found in a population of Kenyon cells forming the
posterior calyx (Schürmann et al., 2000) as well as in fibers
making up only the outer ring of the core. Unlike the other
insects surveyed, the centralmost part of the core in A.
domestica shows no affinity for anti-glutamate antibody.
Due to the immature nature of its constituent axons, the
ingrowth lamina or core consists of a constantly revolving
population of neurons. Axons extending from newborn
Kenyon cells into the core are soon pushed into the
periphery by still younger fibers. The production of
collaterals and glutamate seems, therefore, to be a transient
stage in the maturation of Kenyon cells. Such a dramatic
change in phenotype may be considered a transdifferentiation-like event, similar to that seen in the radial glia of the
mammalian brain (Chanas-Sacre et al., 2000). The distinct
morphology and putative neurotransmitter profile observed
in ingrowth lamina and core Kenyon cells indicates that they
serve a unique function prior to their final differentiation
into the more typical morphology of Kenyon cells
occupying the remainder of the lobes. Further investigations
into the immature phenotype of Kenyon cells will be
necessary to determine whether these cells are serving as
transient guidance cues for extrinsic neurons or are playing
other roles required for their proper integration into the
mushroom body circuitry.
Studies of D. melanogaster mutants with axon branching
defects in the mushroom bodies provide evidence that some
Kenyon cells may perform guidance functions, at least for
other intrinsic neurons. The mushroom bodies of flies with
mutant copies of the Dscam gene (the D. melanogaster
homologue of the mammalian Down syndrome cell
adhesion molecule gene) show a variety of axonal branching
and guidance defects, often leading to the loss of the medial
or vertical lobe (Wang et al., 2002). Individual mutant
neurons produce supernumerary axonal branches that often
fail to project into both lobes. The morphology of wild type
Kenyon cells in mosaic mushroom bodies depends on the
time in development when the mutant phenotype is revealed
91
in a subset of Kenyon cells. When homozygous mutant
clones are induced before the onset of a0 b0 Kenyon cell
production, loss of the medial or vertical lobe is often
observed. In these insects, the wild-type a/b neurons
produced later in development by non-recombinant neuroblasts continue to follow the mutant projection pattern. A
similar mutant phenotype and cell non-autonomous effect
was observed after induction of Rac GTPase (Ng et al.,
2002) and trio (Awasaki et al., 2000) homozygous mutant
neuroblast clones. It thus appears that Kenyon cells in the
mushroom bodies of D. melanogaster play a significant role
in regulating proper branching and outgrowth in subsequently produced intrinsic neurons.
4.3. Formation of the calyces
Many authors have noted that development of the calyx
lags after that of the lobes in most insects, indicating that
Kenyon cells produce axons first, and then dendrites.
Calyces are entirely absent in basal insects like the Odonata
(dragonflies and damselflies) and in those with secondarily
rudimentary antennae and antennal lobes such as the aquatic
dytiscid beetles (Strausfeld et al., 1998). In these insects,
afferent input necessarily enters the pedunculus and lobes
directly. The late addition of calyces in development
together with their absence in basal insects has led to
speculation that these neuropils are relatively recent
additions to the insect mushroom bodies (Strausfeld et al.,
1998). Expression patterns of P[GAL4] lines, which exhibit
either 4-fold (corresponding to each neuroblast) or 2-fold
symmetry, indicate that two of the MBNBs are more closely
associated to each other (Yang et al., 1995). In the
pedunculus, the four fiber tracts exiting the calyx initially
fuse into two tracts, and finally into a single axon bundle (Ito
et al., 1997; Strausfeld et al., 2003). The single calyx of
insects such as D. melanogaster is therefore believed to
represent the fusion of two calyces, each derived from two
neuroblasts (Ito et al., 1997). It has also been proposed,
however, that most double calyces are derived from the
duplication of a structure similar to the single anterior calyx
of the Orthoptera, which is made up of the clawed and spiny
Kenyon cells commonly observed in most insect species
Fig. 7. Morphological and immunochemical characteristics of ingrowing Kenyon cell axons across insect taxa. The ingrowth lamina or core is indicated by an
arrow in all panels. (A)–(C) are sagittal sections, (D)–(H) are frontal sections. (A) In the adult cockroach (P. americana), Golgi impregnated sagittal sections
through the medial lobe reveal axons of the ingrowth lamina and the associated collateral system reaching across the width of the lobe. (B) The ingrowth lamina
of the adult cockroach exhibits glutamate-like immunoreactivity as does the collateral system. (C) Ingrowth lamina axons and collaterals in the cockroach
nymph also show a strong affinity for phalloidin (green), which labels f-actin. Antibodies directed against the catalytic subunit of the D. melanogaster protein
kinase A (anti-DC0; purple) reveal the laminae of the remainder of the medial lobe. (D) As in the cockroach, the developing brain of the cricket A. domestica
shows strong DC0-like immunoreactivity (purple) in the medial (M) and vertical (V) lobes and phalloidin labeling of the core (green). (E), (F) Glutamate-like
immunoreactivity is observed in the core fibers of the cricket vertical (E) and medial (F) lobes, again reminiscent of the staining pattern observed in the
cockroach. In the holometabolous honeybee A. mellifera (G), a similar pattern of DC0 (purple) and phalloidin (green) immunostaining is observed in the D3
pupal mushroom bodies. Phalloidin-labeled neurites stream around the periphery of the neuroblast aggregate (Nb) and enter the penduculus in a single tract.
(H) Glutamate immunoreactivity is seen in the core of the D3 honeybee pupa is similar to that observed in the hemimetabolous cockroach and cricket.
Glutamate and DC0 immunostaining was performed as described in Sinakevitch et al. (2001) and Farris and Strausfeld (2003), respectively. Oregon green
conjugated phalloidin (Molecular Probes) was applied to DC0-stained preparations at a 1:500 concentration and incubated overnight at 4 8C. Scale bars A, B,
C, E ¼ 20 mm, D ¼ 100 mm, F ¼ 50 mm, G ¼ 10 mm, H ¼ 5 mm.
92
S.M. Farris, I. Sinakevitch / Arthropod Structure & Development 32 (2003) 79–101
(Weiss, 1981). Again, further comparative studies of calyx
anatomy and development in a wider range of insect species
will be necessary to determine with certainty the origins of
this neuropil.
Calycal development has been studied in relatively less
detail than lobe development. In D. melanogaster, little has
been reported aside from the fact that the neuropil steadily
increases in size until pupariation, at which point massive
degeneration and subsequent regrowth into the adult structure
occurs (Lee et al., 1999). Like the cell bodies themselves, the
calyces are divided into four equivalent units, each unit being
supplied by the Kenyon cells produced by a single neuroblast
(Yang et al., 1995). Golgi impregnation studies have revealed
several distinct dendritic morphologies among Kenyon cells
of the dipteran species D. melanogaster and Musca domestica
L. (Diptera, Muscidae; Strausfeld, 1976; Strausfeld et al.,
2003). The developmental organization of specific subtypes of
Kenyon cell dendrites in D. melanogaster has not been
described to date.
The development of subdivisions of the calyces and their
constituent Kenyon cells has been best described in A.
mellifera. In this insect the calyces, like the cell somata, can
be divided morphologically into three main subdivisions,
the lip, collar and basal ring (Mobbs, 1982). Although
sequentially arranged from dorsal to ventral in the calyx,
these three regions are not produced in their entirety strictly
by birth-date as are the cell soma and lobe regions. The
calyx is first visible in the prepupal honeybee. Panov (1957)
reported that the basal ring forms first in development, in
contrast to Farris et al. (1999) who reported that this
neuropil region differentiates last, during the third to fourth
day of the pupal stage (D3 –D4). In his account, Panov
describes the early born outer compact Kenyon cells as
providing dendrites only to the basal ring, which may have
led him to conclude that the initial rudiments of the calyx
present in the prepupa must correspond to this region. It has
subsequently been shown, however, that outer compact
Kenyon cells are in fact clawed Kenyon cells, which in A.
mellifera will eventually contribute to all three regions of
the calyx (Strausfeld, 2002). A similar organization is
observed in D. melanogaster (Yang et al., 1995). The
dendrites of subsequently born non-compact (spiny) cells
are therefore not stacked on top of the basal ring as proposed
by Panov, but rather are integrated within the early calyx.
Both accounts agree that the lip and collar region are first
distinguishable at D2 –D3 of the pupal stage and appear to
arise simultaneously (Panov, 1957; Farris et al., 1999) (Fig.
8). The late formation of the basal ring coincides with the
birth of the inner compact cells whose dendrites, along with
those of the outer compact cells, make up this neuropil
region (Farris et al., 1999; Strausfeld, 2002). Olfactory
projection neurons, the primary afferents to the lip region of
the calyx, are already present in the dorsal calyx by the first
day of the pupal stage (Schröter and Malun, 2000). The
differentiation of calycal regions according to afferent
innervation patterns may therefore occur even before
Fig. 8. Successive stages of calyx development in the honeybee pupa.
Frontal sections, dorsal to the top. (A) The calyces (C) are visible in the D1
pupa as balls of neuropil residing below each neuroblast aggregate (Nb). All
three Kenyon cell populations (oc, nc, ic) can be identified from this stage
onwards. (B) Increased calyx growth and decreased size of the neuroblast
aggregate is apparent in the D2 pupa. The division between the lip and
collar regions of the calyces (arrow) are first visible at this stage. Glial cells
entering the ventral calyx may indicate the onset of differentiation of the
basal ring (arrowheads). (C) The calyces clearly begin to resemble those of
the adult, with distinct lip (arrow) and collar regions, in the D3 pupa.
Growth of this neuropil is accompanied by a steady decline in size of the
neuroblast aggregates, whose constituent cells undergo programmed cell
death by D5– D6. Glial cells outline the developing basal ring region
(arrowheads). Ic, inner compact Kenyon cells; nc, non-compact Kenyon
cells; oc, outer compact Kenyon cells. Scale bars ¼ 50 mm.
S.M. Farris, I. Sinakevitch / Arthropod Structure & Development 32 (2003) 79–101
pupation, and must be mediated by the clawed Kenyon cells
whose dendrites are the sole constituents of the prepupal
calyx (Panov, 1957).
Prior to adult eclosion, the dendritic branches of spiny
Kenyon cells in the lip and collar region of the A. mellifera
mushroom bodies are decorated with many short, filamentous fibers (Farris et al., 2001). In the adult these structures
are replaced by stout spines, indicating that the filaments
may be immature spines that obtain their mature morphology only after adult eclosion.
In P. americana, most Kenyon cells produce several
dendritic arborizations along their entire dorsal – ventral
axis, rather than being confined to a discrete calycal subregion as in A. mellifera. Exact correlations between cell
birthdate and the positions of individual dendritic arbors are
not immediately evident, although the last-born Kenyon
cells do tend to provide arborizations to the more ventral
regions of the calyx (Mizunami et al., 1998a; Farris and
Strausfeld, 2001). The dendrite-producing neurites, however, are contained within discrete age dependent layers in
the parallel fiber containing inner calyx wall (zona interna of
Weiss, 1974). Kenyon cells whose cell bodies are ventral in
the calyx and whose axons make up the most osterior
laminae in the lobes (and are thus the most recently born)
send their neurites along the innermost margin of the zona
interna (Mizunami et al., 1998a; Sinakevitch et al., 2001).
Neurites occupy progressively deeper positions in the zona
interna as their age increases, and the neurites of clawed
Kenyon cells are so deep as to pass directly through the
synaptic neuropil of the calyx (Mizunami et al., 1998a;
Strausfeld and Li, 1999). A newly discovered class of
Kenyon cells in the cockroach, termed Class III, perpetuate
this trend to its extreme by projecting along the outer
surface of the calyces before entering the pedunculus (Farris
and Strausfeld, 2003). As expected, these cells appear to be
the first produced mushroom body intrinsic neurons in the P.
americana embryo.
In A. domestica, the unusual calyx configuration of the
adult has a correspondingly unique developmental story.
Unlike the double calyces of most insects which appear to
be made up of equivalent groups of Kenyon cells, the
anterior and posterior (or accessory) calyces of the
Orthoptera are formed from different intrinsic neuron
subtypes (Panov, 1966; Schürmann, 1973; Weiss, 1981;
Malaterre et al., 2002). The posterior calyx arises before the
anterior calyx in embryonic development (Panov, 1966;
Malaterre et al., 2002). The first formed posterior calyx
appears to be composed of a population of Kenyon cells
that, based on morphological similarities, may be homologous to the embryonically-produced Class III Kenyon cells
of some Dictyoptera (Schürmann, 1973; Farris and
Strausfeld, 2003). This structure also contains clawed
Kenyon cells in A. domestica (Schürmann, 1973), although
Weiss (1981) reported that these cells contribute instead to
the anterior calyx in the acridid Orthoptera. The anterior
calyx is also composed of dendrites of spiny Kenyon cells,
93
and grows increasing larger through nymphal and adult
development due to continued production of these Kenyon
cells by the MBNBs (Malaterre et al., 2002).
4.4. Metamorphic plasticity of clawed Kenyon cells
Massive reorganization of the nervous system is
characteristic of metamorphosis, and has been extensively
described in the ventral nerve cord. As in other regions of
the nervous system, metamorphosis of the mushroom bodies
consists of restructuring of larval processes combined with
the generation of adult-specific neurons (reviewed in
Consoulas et al., 2000). Metamorphic restructuring of the
mushroom bodies of the dipteran brain was noted in P.
regina (Gundersen and Larsen, 1978), but was first
described in detail in D. melanogaster (Technau and
Heisenberg, 1982). By counting axon profiles in crosssections of the pedunculus, the authors observed a 40%
decrease in the number of axons within 12 h after puparium
formation, followed by an increase in axon number beyond
that observed prior to metamorphosis (Fig. 9). A bundle of
extremely thin fibers forming a tract through the center of
the pedunculus was observed to persist throughout the
period of degeneration. This tract was frequently absent in
mushroom bodies deranged (mbd) and mushroom body
Fig. 9. Selective labeling of clawed Kenyon cells reveals reorganization
during metamorphosis in D. melanogaster (from Lee et al. 1999). (A)
Evidence of axonal degradation is visible in the lobes by 12 h after
puparium formation (APF). The lobes appear thinner due to loss of axons
(arrows) and cellular debris is visible at the former tip of the medial lobe
(arrowhead). (B) Degradation of the lobes is complete by 18 h APF (arrow),
leaving the pedunculus thinner but relatively intact. (C) At 24 h APF,
clawed Kenyon cells have re-extended their dendrites into the medial lobe
only (arrow). (D) The medial lobe at 24 h APF contains axons of variable
length (arrows), indicating that process outgrowth is not yet complete.
Dotted line, brain midline. Scale bars ¼ 50 mm.
94
S.M. Farris, I. Sinakevitch / Arthropod Structure & Development 32 (2003) 79–101
defect (mud) mutants, in which the lobes were malformed
only in the adult. It was therefore proposed that the thin fiber
tract could provide guidance cues for the fibers that
repopulate the pedunculus and lobes in the adult.
The identity of the reorganizing and persisting Kenyon
cell populations proved somewhat elusive. The first hint as
to the result of axon reorganization was provided when
Yang et al. (1995) determined that clawed Kenyon cells
produced unbranched axons in the medial g lobe of the
adult. Armstrong et al. (1998) subsequently discovered that
these neurons are produced early in development, and
affirmed that the larval mushroom bodies contain only a
single vertical/medial lobe pair. Further evidence from
enhancer trap lines that labeled both of the larval lobes and
the adult g lobe led to the conclusion that these structures are
composed of the same population of Kenyon cells
(Armstrong et al., 1998). After axon degeneration, the
adult g lobe was seen to arise first. The a/b lobes were the
last to appear, indicating that their constituent Kenyon cells
were adult-specific (Armstrong et al., 1998; Lee et al.,
1999). Using the MARCM technique to label small
subpopulations of cells born at specific times in development (Lee and Luo, 1999), the degeneration and reorganization of larval clawed Kenyon cells into the adult g lobe
and the pupal origin of the neurons making up the adult a/b
lobes was confirmed (Lee et al., 1999). The a0 /b0 lobes first
identified by Crittenden et al. (1998) are made up of neurons
that are born after the clawed Kenyon cells but before
puparium formation and do not undergo reorganization (Lee
et al., 1999). The persistent thin fiber tract described by
Technau and Heisenberg (1982) is therefore likely composed of the axons of a0 /b0 neurons (Kurusu et al., 2002).
Reorganization of clawed Kenyon cells in the mushroom
bodies of D. melanogaster occurs during a period of
increasing ecdysteroid titers in the pupa (Kraft et al., 1998).
The ecdysone receptor (EcR), particularly the EcR-B1
isoform, is highly expressed in the clawed Kenyon cells of
the larval and early pupal mushroom bodies (Truman et al.,
1994; Lee et al., 2000a). Cultured Kenyon cells isolated at
this time respond to 20-hydroxyecdysone treatment with
increased neurite outgrowth and branching (Kraft et al.,
1998). The EcR protein forms a heterodimer with the
ultraspiracle (usp) gene product in order to regulate gene
expression (Yao et al., 1992, 1993). Clawed Kenyon cells of
EcR and usp mutants fail to degenerate and retain the
branched larval morphology into adulthood (Lee et al.,
2000a). Interestingly, usp mutations generated in g neurons
in vivo appear to disrupt neither initial axonal outgrowth nor
continued growth throughout development. These experiments indicate that ecdysone likely plays an important role
in the regulation of degeneration of clawed Kenyon cell
axons, but its role in process outgrowth in vivo is less clear.
Reorganization of clawed Kenyon cells has thus far been
reported only in D. melanogaster, so the question is whether
metamorphic restructuring of the mushroom bodies is a
common occurrence in holometabolous insects, or whether
it is a unique characteristic of a highly derived species.
Evidence for clawed Kenyon cell reorganization in other
holometabolous insects is at the present indirect and
relatively weak. The mushroom bodies of the moth Sphinx
ligustri L. (Lepidoptera, Sphingidae) contains clawed
Kenyon cells with unbranched axons that project medially
as in D. melanogaster (Pearson, 1971). Both a medial and a
vertical lobe is observed in larval Lepidoptera (Hanström,
1925; Nordlander and Edwards, 1968, 1970). Given the
early origin of clawed Kenyon cells in all insects, it is
perhaps possible that these neurons bifurcate in the larva
and reorganize during metamorphosis into the unbranched
adult form. In A. mellifera the picture is more complicated
as only some clawed Kenyon cells are unbranched, the
axons projecting into the vertical rather than the medial
component of the g layer (Strausfeld, 2002). The developmental history of this subset of clawed Kenyon cells is at
present unclear. The remaining branched clawed Kenyon
cells of the adult A. mellifera have tiny medial axons
forming a short medial g layer in the adult that appears to
have persisted unchanged from the larval stage (Farris,
Abrams and Strausfeld, unpublished observations). As these
examples indicate, definitive evidence for metamorphic
restructuring of clawed Kenyon cells exists only for D.
melanogaster at present. A great deal of further study will be
necessary before any conclusions on the extent of mushroom body reorganization in non-dipteran holometabolous
insects can be made.
4.5. Conserved and divergent aspects of process outgrowth
in the mushroom bodies
As with patterns of neurogenesis, there is much common
ground in the developmental processes of neuropil formation across the insects. Although the dendritic neuropil of
the calyces is not strictly arranged by birthdate, the neurites
providing the arborization are, and this age-dependent
ordering is maintained into the pedunculus and lobes. The
calyces are established by the dendrites of clawed Kenyon
cells, which represent the entire neuropil, and are subsequently filled in with the processes of later born spiny
neurons. Whether concentric or laminar, axons of newborn
Kenyon cells enter the lobes via a discrete tract, such that
axons located progressively further from this core have
correspondingly earlier birthdates. Lobes and layers made
up of the early-born clawed Kenyon cells (g) are constructed
first. Newborn Kenyon cells with axons in the core display
unusual phenotypes, such as the production of collaterals
and glutamate immunoreactivity, indicating that they are
functioning in some unknown, transient role prior to
differentiating into mature Kenyon cells. A role for Kenyon
cells as guidance cues for extrinsic and intrinsic neurons
during development has been proposed, but remains as yet
unproven. Finally, the presence of unbranched clawed
Kenyon cells in the adult stage of holometabolous insects
S.M. Farris, I. Sinakevitch / Arthropod Structure & Development 32 (2003) 79–101
hint at a widely occurring process of metamorphic
reorganization of these neurons.
No evidence of massive clawed Kenyon cell reorganization is observed in hemimetabolous insects, perhaps due to
the absence of complete metamorphosis. In P. americana
and A. domestica, clawed Kenyon cells maintain their
medial and vertical branches throughout life (Schürmann,
1973; Strausfeld and Li, 1999). The drastic reorganization
of the axonal projection pattern of clawed Kenyon cells in
the holometabolous insects indicates that they likely have
different targets and perform different functions in the larval
and adult mushroom bodies (Technau and Heisenberg,
1982; Armstrong et al., 1998; Lee et al., 1999). Perhaps such
reorganization is not necessary in the hemimetabolous
insects, in which juvenile and adult behaviors are not as
vastly divergent as those of holometabolous larvae and
adults.
5. Genetic control of mushroom body development
Genetic studies in D. melanogaster have led to the
discovery of a number of genes that guide mushroom body
development. The first mushroom body structural mutants
proved to be deficient in learning and memory, thus
providing some of the first solid evidence for mushroom
body functioning in these processes. Subsequent screens
have isolated a variety of mutants in all aspects of
mushroom body development. The majority of mutants
described exhibit some manner of disruption of process
outgrowth and branching, perhaps because the mutant
phenotypes tend to be particularly severe and are therefore
easily identified.
5.1. Genes involved in neurogenesis
Two of the earliest described mushroom body structural
mutants have since proved to have defects in neurogenesis.
The mushroom body defect (mud) and mushroom bodies
miniature (mbm) mutants produce too many and too few
Kenyon cells, respectively. In mud mutants, up to 25
MBNBs are found in each brain hemisphere, rather than
four in wild type brains, leading to a large increase in
Kenyon cell number (Technau and Heisenberg, 1982;
Prokop and Technau, 1994). The mud gene encodes a
coiled-coil protein containing conserved actin-binding
motifs (Guan et al., 2000). The mbm phenotype is sexspecific, with adult females having a tiny portion of the wild
type number of Kenyon cells, while males have normal or
even larger mushroom bodies (Heisenberg et al., 1985). The
female phenotype first manifests in the late larval stage.
Another structural mutant that was identified in these initial
screens, mushroom bodies deranged (mbd), has been
reported to have decreased Kenyon cell number in the
adult (Tettamanti et al., 1997), although no neurogenesis
95
defect was noted in the original accounts (Technau and
Heisenberg, 1982).
Other neurogenesis mutants typically exhibit smaller
mushroom bodies resulting from a reduction in Kenyon cell
number. Aside from being critically important for development of the compound eye, the ‘master gene’ eyeless (ey) is
also expressed in mushroom body primordia and intrinsic
neurons. Both hypomorphic and overexpression mutants
display a decrease in Kenyon cell number (Callaerts et al.,
2000; Noveen et al., 2000).
The enok and rho A genes, when mutant, produce
immediate and severe defects in neurogenesis that arrest
mushroom body development entirely. Only the g lobe is
produced in enok mutants due to MBNB arrest shortly after
larval hatching (Scott et al., 2001). Induction of mutant
clones later in development, however, leads to a similar
gradual cessation of MBNB proliferation, so this gene is not
specifically involved in clawed Kenyon cell production.
Enok has been shown to encode a histone acetyltransferase;
thus the observed mutant phenotype observed in the rapidly
dividing MBNBs. The rhoA gene is involved in cytokinesis,
and causes a similar rapid arrest of MBNB activity after
induction of mutant clones (Lee et al., 2000b).
5.2. Genes involved in process outgrowth and branching
The early neuroanatomical mutant screens turned up a
number of genes that appeared to affect axon pathfinding,
particularly during metamorphic reorganization. Along with
increased neurogenesis in the larval stage, the mushroom
bodies of mud mutant flies exhibit a complete disruption of
the adult lobes (Heisenberg, 1980). Degenerating axons in
the early pupa are apparently unable to reenter the
protocerebral neuropil to form the adult lobes, and instead
form a large disorganized mass around the calyx. The
mushroom bodies deranged (mbd) mutant displays a similar
phenotype (Heisenberg et al., 1985; Heisenberg, 1989). The
a0 /b0 core fibers are often missing in the early pupae of both
of these mutants, providing additional evidence that these
Kenyon cells guide the g and a/b axons into the lobes during
metamorphosis (Technau and Heisenberg, 1982).
Ey mutants display a wide variety of neuropil defects,
which are highly variable depending on the particular
mutant allele and genetic background. The ey JD and ey D1DA
hypomorphic mutant alleles lead to a gross reduction in
neuropil size in homozygous adults (Callaerts et al., 2000),
while the ey 2 and ey R hypomorphic alleles cause few
apparent adult defects but generate fused, reduced or absent
lobes in larvae (Noveen et al., 2000). Interestingly, overexpression of ey in these flies also causes lobe fusion and
reduction. Null ey mutant alleles ey J5.71 and ey C7.20 cause
minor defects in homozygous larvae, but Kenyon cell
development appears to be arrested during the pupal
reorganization so that adult neuropils are completely ablated
(Kurusu et al., 2000).
Dachshund (dac), another regulatory gene primarily
96
S.M. Farris, I. Sinakevitch / Arthropod Structure & Development 32 (2003) 79–101
known for its role in eye development, is expressed in
mature Kenyon cells throughout development (Kurusu et al.,
2000; Martini et al., 2000). Null mutant homozygotes have
misdirected or missing vertical lobes in the larval mushroom
bodies and disorganized medial lobes in the pupa and adult,
with individual stray fibers observed throughout.
Although ey and dac act synergistically during eye
development, studies using hypomorphic ey mutants
initially failed to support such a relationship in the
mushroom bodies. Experiments using ey null mutants,
however, indicate that ey and dac work together in the
mushroom bodies as they do in the eye. Dac heterozygotes
were shown to enhance the abnormal phenotype of ey
homozygous null mutants, and null double mutants showed
a strong degeneration of the lobe neuropil in the adult
(Kurusu et al., 2000).
Fasciclin II ( fasII), a cell adhesion molecule best studied
for its important role in axon fasciculation in the embryonic
nerve cord, is expressed in g and a/b Kenyon cells of the D.
melanogaster mushroom bodies (Cheng et al., 2001). As
with ey mutants, phenotypes vary between the two mutant
studies published to date. Fas II null and hypomorph
mutants display no obvious adult phenotype (Cheng et al.,
2001), while hypomorph mutant larvae often have reduced
vertical lobes and fused medial lobes (Kurusu et al., 2002).
In those larvae in which both lobes are reduced the calyces
are often enlarged, indicating pathfinding defects that cause
axons to bunch up around the calyx as in mud and mbd
mutants. As might be expected in neurons defective for
proteins involved in fasciculation, the axons of fasII null
mushroom body clones do not respect the concentric
arrangement of the pedunculus and lobes. Rather than
entering the lobes through the core region, fasII mutant
axons apparently penetrate the lobes at random and are thus
irregularly distributed with no relation to birthdate (Kurusu
et al., 2002). Overexpression also leads to disruption of the
core in the developing mushroom bodies.
Other genes that appear to play roles in axon guidance in
the mushroom bodies, particularly during reorganization,
are members of the Rho and Rac GTPase cascades (Awasaki
et al., 2000; Ng et al., 2002) and the D. melanogaster
homologue of the human gene for Down syndrome cell
adhesion molecule, DSCAM (Wang et al., 2002). Induction
of trio (a member of the Rho GTPase cascade) and Rac
GTPase mutant clones causes disruption of later born lobe
systems in the adult, indicating non-cell autonomous
guidance effects on wild-type neurons. Trio mutants have
shortened lobes indicating axon stalling, as well as
individual misguided fibers leaving the calyx at random
trajectories (Awasaki et al., 2000). Interestingly, clones
containing mutant Rho GTPase, which is part of the same
chemical cascade as trio, do not have lobe defects but rather
overgrown dendrites resulting in large calyces (Lee et al.,
2000b). The successive mutation of three Rac GTPases
(Rac1, Rac2 and Mtl) leads to defects in branching (loss of a
single lobe), guidance (axons form a ball around calyx as in
mud and mbd), and finally growth (axons do not extend past
pedunculus) (Ng et al., 2002). DSCAM mutant g Kenyon
cells have no noticeable defects, but a0 /b0 and a/b mutant
cells have supernumerary branches that are often directed in
only the vertical or medial direction (Wang et al., 2002). In
another instance of apparent guidance roles for a0 /b0 axons,
mutant clones induced immediately before the production of
a0 /b0 neurons often exhibit mutant projection patterns (both
axon branches extending into a single lobe) of wild-type a/b
neurons as well. In mutants in which only a non-doubled
lobe is present wild-type neurons extend a single
unbranched axon into this lobe, indicating that DSCAM
also regulates axon branching.
As mentioned previously, ecdysteroid hormones play an
important role in g Kenyon cell reorganization during
metamorphosis. This is underscored by the dramatic
phenotype of adult flies mutant for the ecdysone receptor
isoform B1 (EcR) or ultraspiracle (usp), the gene products of
which act as a heterodimer to effect ecdysteroid-regulated
gene transcription (Lee et al., 2000a). Adult mutant flies
show no reorganization of g, with the constituent clawed
Kenyon cells branching into a medial and a vertical
component as in the larva. EcR-B1 is expressed only in g
Kenyon cells and its effects appear to be completely cellautonomous, indicating direct effects of ecdysteroids on
clawed Kenyon cells. Interestingly, several known EcR/
USP gene targets had no noticeable effects on reorganization when mutant; perhaps another pathway is involved in
mushroom body neuronal remodeling.
A large class of mushroom body structural mutants also
have more widespread effects on the development of
midline neuropils such as the central complex. These
mutants manifest phenotypically as a fusion of the medial
lobes of the two mushroom bodies across the midline.
Medial lobe fusion phenotypes typically involve the b and b0
lobes, and in severe cases the g lobe as well. One such gene,
beta lobes fused (bef), was described in the early mutant
screen that turned up mud, mbm and mbd (Heisenberg,
1980); since that time many more medial lobe fusion
mutants have been added to the list. The genes ciboulot,
brainstorming, split central complex, alpha lobes absent,
brainwashing, fused lobes and castor all display central
complex disruption and variable degrees of medial lobe
fusion when mutant (Boquet et al., 2000a,b; Hitier et al.,
2001). Missing vertical lobes are also seen in more severe
cases, and the mutant alpha lobes absent has been of special
use in functional studies, as mutants are shown to be
deficient in long-term memory (Pascual and Préat, 2001).
These genes are not always expressed in the mushroom
bodies, but rather in midline cells such as the glia of the
transient interhemispheric fiber ring (TIFR) (Simon et al.,
1998; Boquet et al., 2000b), indicating that the functional
gene products may be part of a pathway generating
repulsive signals preventing Kenyon cells from crossing
the midline.
The best-characterized medial lobe fusion gene is linotte
S.M. Farris, I. Sinakevitch / Arthropod Structure & Development 32 (2003) 79–101
(lio, also known as derailed), which encodes a receptor
tyrosine kinase and was initially isolated in a screen for
learning and memory mutants (Dura et al., 1993, 1995). The
phenotype is particularly severe, with the a and sometimes
the a0 lobes reduced or absent and the b and b0 and perhaps g
lobes fused across the midline. Lio was initially thought to
97
be expressed only in glial cells of the TIFR, but a lio-lacZ
reporter gene produces Kenyon cell staining as well
(Moreau-Fauvarque et al., 1998; Simon et al., 1998). In
either case, expression of lio is not observed in the adult,
indicating that lio plays a role only in the development of
mushroom body neurons.
Fig. 10. Schematic diagram of a developing mushroom body of P. americana, sagittal view, illustrating both general and specific aspects of insect mushroom
body development. Neuroblasts (MbNb) are located centrally with respect to each calyx (C) of the insect mushroom bodies. Cell bodies within and around the
calyx are arranged in an age gradient. Class I and class II Kenyon cells, with their characteristic locations and process organizations, appear to be common to
the mushroom bodies of the majority of insect taxa. In all cases, class II cells are among the first born intrinsic neurons and their cell bodies are located at the
outer periphery of the cell body region. Class I cells make up the remainder of the Kenyon cell population and their cell bodies fill the calyx cup. Kenyon cell
axons travel through the pedunculus (P) and bifurcate into the medial (M) and vertical (V) lobes, where they are also arranged by age. Class II axons form the
anterior g layer (or the separate g lobe in D. melanogaster) while class I axons occupy progressively more posterior laminae, layers or lobes as neurogenesis
proceeds. Newborn class I cells form with their axons the core of the pedunculus and lobes and are distinct in both their morphology (often producing
collaterals) and gene expression patterns. The existence of class III cells outside of the Dictyoptera has not been confirmed thus far. In P. americana these cells
are born even before the class II cells and form a separate accessory calyx. Their axons occupy the most anterior laminae of the cockroach lobes, and in the
vertical lobe form a looping structure termed the lobelet. Kenyon cells homologous to dictyopteran class III cells may form the posterior/accessory calyx
system in the Orthoptera.
98
S.M. Farris, I. Sinakevitch / Arthropod Structure & Development 32 (2003) 79–101
5.3. Genes and mushroom body development
The study of the molecular control of mushroom body
development is in its infancy, with a few well-described
genes and little knowledge of how they function to effect
proper axon organization. Further comparisons with wellstudied pathways from other systems, such as those
regulated by ecdysteroids or ey and dac, should fill in
some of these gaps in time. In-depth investigation of
individual aspects of mushroom body development, such as
axon branching or reorganization, will be helpful in the
search for suites of genes that are involved in the regulation
of these specific processes. Such widely-conserved developmental events are likely to share homologous genes with
other insects and arthropods, and perhaps even vertebrates.
6. Conclusions
Comparative and molecular analyses of mushroom body
development have provided important information about the
formation of this complex brain region. Although speciesspecific differences exist, a clear, coherent picture of the
basic events in insect mushroom body development and the
subsequent organization of these brain structures is beginning to be realized (Fig. 10). Fundamental principles of
neurogenesis and cellular organization are highly conserved, and may be compared to the birthdate-dependent
ordering of layered brain structures in the mammalian brain
such as the cerebral cortex. Continued investigation of these
conserved events at the cellular and molecular level are
likely to uncover evolutionarily ancient genes and pathways
that are fundamentally important for the construction of
complex nervous systems. Comparative studies of less
universal aspects of development, such as Kenyon cell
reorganization during metamorphosis, will give insight into
the evolution of the insect mushroom bodies. Finally, the
diversity of behavioral ecologies within closely related
insect groups provides an excellent basis for studies of
correlations between life history and brain development,
and how these factors interact to effect observed trends in
brain evolution.
Acknowledgements
The authors wish to thank Dr Nicholas Strausfeld, whose
insightful comments greatly improved this manuscript. We
would also like to thank Dr Birgit Ehmer for her assistance
in translating the German literature and Dr Samuel Beshers
for providing Atta cephalotes pupae. The authors are also
grateful for the comments and suggestions of two
anonymous reviewers, that enhanced the clarity and
completeness of this manuscript. SMF was supported by a
National Institutes of Health Program Project Grant NIH2
P01 NS28495. IS received support from a Human Frontiers
award HFSP RG0143/2000-B.
References
Afify, A., 1960. Über die postembryonale Entwicklung des Zentralnervensystems (ZNS) bei der Wanderheuschrecke Locusta migratoria
migratorioides (R.U.F.) (Orthoptera-Acrididae). Zoologisches Jahrblatt
Abteilung für Anatomie 78, 1–38.
Armstrong, J.D., deBelle, J.S., Wang, Z., Kaiser, K., 1998. Metamorphosis
of the mushroom bodies; large-scale rearrangements of the neural
substrates for associative learning and memory in Drosophila. Learning
and Memory 5, 102–114.
Awasaki, T., Saito, M., Sone, M., Suzuki, E., Sakai, R., Ito, K., Hama, C.,
2000. The Drosophila trio plays an essential role in patterning axons by
regulating their directional extension. Neuron 26, 119–131.
Bieber, M., Fuldner, D., 1979. Brain growth during the adult stage of a
holometabolous insect. Naturwissenschaften 66, 426.
Boquet, I., Boujemaa, R., Carlier, M.-F., Préat, T., 2000a. Ciboulot
regulates actin assembly during Drosophila brain metamorphosis. Cell
102, 797–808.
Boquet, I., Hitier, R., Dumas, M., Chaminade, M., Préat, T., 2000b. Central
brain postembryonic development in Drosophila: implication of genes
expressed at the interhemispheric junction. Journal of Neurobiology 42,
33 –48.
Brandon, J., Coss, R., 1982. Rapid dendritic spine shortening during onetrial learning: the honeybee’s first orientation flight. Brain Research
252, 51–61.
Bretschneider, F., 1914. Über die Gehirne des Goldkäfers und des
Lederlaufkäfers. Zoologischer Anzeiger 43, 490–497.
Callaerts, P., Leng, S., Clements, J., Benassayag, C., Cribbs, D., Kang, Y.,
Walldorf, U., Fischbach, K.-F., Strauss, R., 2000. Drosophila Pax-6/
eyeless is essential for normal adult brain structure and function. Journal
of Neurobiology 46, 73–88.
Cayre, M., Strambi, C., Strambi, A., 1994. Neurogenesis in an adult insect
brain and its hormonal control. Nature 368, 57–59.
Cayre, M., Strambi, C., Tirard, A., Renucci, M., Charpin, P., Augier, R.,
Strambi, A., 1995. Effects of juvenile hormone on polyamines of the fat
body and neural tissue of the cricket Acheta domesticus. Comparative
Biochemistry and Physiology 222, 241–250.
Cayre, M., Strambi, C., Charpin, P., Augier, R., Meyer, M.R., Edwards,
J.S., Strambi, A., 1996. Neurogenesis in adult insect mushroom bodies.
The Journal of Comparative Neurology 371, 300– 310.
Cayre, M., Strambi, C., Charpin, P., Augier, R., Strambi, A., 1997a.
Inhibitory role of ecdysone on neurogenesis and polyamine metabolism
in the adult cricket brain. Archives of Insect Biochemistry and
Physiology 35, 85–97.
Cayre, M., Strambi, C., Charpin, P., Augier, R., Strambi, A., 1997b.
Specific requirement of putrescine for the mitogenic action of juvenile
hormone on adult insect neuroblasts. Proceedings of the National
Academy of Sciences of the United States of America 94, 8238– 8242.
Cayre, M., Malaterre, J., Charpin, P., Strambi, C., Strambi, A., 2000a. Fate
of neuroblast progeny during postembryonic development of mushroom
bodies in the house cricket, Acheta domesticus. Journal of Insect
Physiology 46, 313–319.
Cayre, M., Strambi, C., Strambi, A., Charpin, P., Ternaux, J.-P., 2000b.
Dual effect of ecdysone on adult cricket mushroom bodies. European
Journal of Neurosciences 12, 633–642.
Cayre, M., Malaterre, J., Strambi, C., Charpin, P., Ternaux, J.-P., Strambi,
A., 2001. Short- and long-chain natural polyamines play specific roles
in adult cricket neuroblast proliferation and neuron differentiation in
vitro. Journal of Neurobiology 48, 315– 324.
Chanas-Sacre, G., Rogister, B., Moonen, G., Leprince, P., 2000. Radial glia
S.M. Farris, I. Sinakevitch / Arthropod Structure & Development 32 (2003) 79–101
phenotype: origin, regulation, and transdifferentiation. Journal of
Neuroscience Research 61, 357 –363.
Cheng, Y., Endo, K., Wu, K., Rodan, A., Heberlein, U., Davis, R., 2001.
Drosophila fasciclin II is required for the formation of odor memories
and for normal sensitivity to alcohol. Cell 105, 757 –768.
Consoulas, C., Duch, C., Bayline, R.J., Levine, R.B., 2000. Behavioral
transformations during metamorphosis: remodeling of neural and motor
systems. Brain Research Bulletin 53, 571–583.
Coss, R.G., Brandon, J.G., Globus, A., 1980. Changes in morphology of
dendritic spines on honeybee calycal interneurons associated with
cumulative nursing and foraging experience. Brain Research 192,
49–59.
Crittenden, J.R., Skoulakis, E.M.C., Han, K.-A., Kalderon, D., Davis, R.L.,
1998. Tripartite mushroom body architecture revealed by antigenic
markers. Learning and Memory 5, 38–51.
Datta, S., 1995. Control of proliferation activation in quiescent neuroblasts
of the Drosophila central nervous system. Development 121,
1173–1182.
deBelle, J.S., Heisenberg, M., 1994. Associative odor learning in
Drosophila abolished by chemical ablation of mushroom bodies.
Science 263, 692–695.
Dujardin, F., 1850. Mémoire sur le système nerveux des insectes. Annales
des Sciences Naturelles, Zoologie 14, 195 –206.
Dura, J., Préat, T., Tully, T., 1993. Identification of linotte, a new gene
affecting learning and memory in Drosophila melanogaster. Journal of
Neurogenetics 9, 1– 14.
Dura, J., Taillebourg, E., Préat, T., 1995. The Drosophila learning and
memory gene linotte encodes a putative receptor tyrosine kinase
homologous to the human RYK gene product. FEBS Letters 21,
250–254.
Durst, C., Eichmuller, S., Menzel, R., 1994. Development and experience
lead to increased volume of subcompartments of the honeybee
mushroom body. Behavioral and Neural Biology 62, 259–263.
Edwards, J., 1969. Postembryonic development and regeneration of the
insect nervous system. Advances in Insect Physiology 6, 97 –137.
Ehmer, B., Hoy, R., 2000. Mushroom bodies of vespid wasps. The Journal
of Comparative Neurology 416, 93–100.
Fahrbach, S.E., Strande, J.L., Robinson, G.E., 1995. Neurogenesis is absent
in the brains of adult honey bees and does not explain behavioural
neuroplasticity. Neuroscience Letters 197, 145– 148.
Farris, S.M., Strausfeld, N.J., 2000. Development of a learning and memory
neuropil: neurogenesis and lamina formation in the mushroom bodies of
the cockroach nymph. Society for Neuroscience Abstracts 26.
Farris, S.M., Strausfeld, N.J., 2001. Development of laminar organization
in the mushroom bodies of the cockroach: Kenyon cell proliferation,
outgrowth, and maturation. The Journal of Comparative Neurology
439, 331 –351.
Farris, S.M., Strausfeld, N.J., 2003. A unique mushroom body substructure
common to basal cockroaches and to termites. The Journal of
Comparative Neurology 456, 305– 320.
Farris, S.M., Robinson, G.E., Davis, R.L., Fahrbach, S.E., 1999. Larval and
pupal development of the mushroom bodies in the honey bee, Apis
mellifera. The Journal of Comparative Neurology 414, 97–113.
Farris, S.M., Robinson, G.E., Fahrbach, S.E., 2001. Experience- and agerelated outgrowth of intrinsic neurons in the mushroom bodies of the
adult worker honey bee. Journal of Neuroscience 21, 6395–6404.
Ganeshina, O., Schafer, S., Malun, D., 2000. Proliferation and programmed
cell death of neuronal precursors in the mushroom bodies of the
honeybee. The Journal of Comparative Neurology 417, 349–365.
Greenough, W.T., Volkmar, F.R., 1973. Pattern of dendritic branching in
rat occipital cortex after rearing in complex environments. Experimental Neurology 40, 491 –504.
Gronenberg, W., 2001. Subdivisions of hymenopteran mushroom body
calyces by their afferent supply. The Journal of Comparative Neurology
436, 474 –489.
Gu, S.-H., Tsia, W.-H., Chiang, A.-S., Chow, Y.-S., 1999. Mitogenic effects
of 20-hydroxyecdysone on neurogenesis in adult mushroom bodies of
99
the cockroach, Diploptera punctata. Journal of Neurobiology 39,
264– 274.
Guan, Z., Prado, A., Melzig, J., Heisenberg, M., Nash, H., Raabe, T., 2000.
Mushroom body defect, a gene involved in the control of neuroblast
proliferation in Drosophila, encodes a coiled-coil protein. Proceedings
of the National Academy of Sciences of the United States of America
97, 8122–8127.
Gundersen, R., Larsen, J., 1978. Postembryonic development of the lateral
protocerebral lobes, corpora pedunculata, deutocerebrum, tritocerebrum of Phormia regina Meigen (Diptera: Calliphoridae). Journal of
Insect Morphology and Embryology 7, 467–477.
Hanström, B., 1925. Comparison between the brains of the newly hatched
larva and the imago of Pieris brassicae. Entomologisk Tidskrift 46,
43– 52.
Heisenberg, M., 1980. Mutants of brain structure and function: what is the
significance of the mushroom bodies for behavior. In: O Siddiq, P.B.,
Hall, L.M., Hall, J.C. (Eds.), Development and Biology of Drosophila,
Plenum, New York, pp. 373–390.
Heisenberg, M., 1989. Genetic approach to learning and memory
(mnemogenetics) in Drosophila melanogaster. In: Rahmann, M.,
(Ed.), Fortschritte der Zoologie: Fundamentals of Memory Formation:
Neuronal Plasticity and Brain Function, vol. 37. Gustav Fischer Verlag,
Stuttgart, New York, pp. 1–45.
Heisenberg, M., Borst, A., Wagner, S., Byers, D., 1985. Drosophila
mushroom body mutants are deficient in olfactory learning. Journal of
Neurogenetics 2, 1–30.
Hitier, R., Chaminade, M., Préat, T., 2001. The Drosophila castor gene is
involved in postembryonic brain development. Mechanisms of Development 103, 3–11.
Holmgren, N., 1916. Zur vergleichenden Anatomie des Gehirns von
Polychaeten, Onychophoren, Xiphosuren, Arachniden, Drustaceen,
Myriapoden, und Insekten. Kongliga Svenska Vetenskaps Akademiens
Handlingar 56, 1–303.
Ito, K., Hotta, Y., 1992. Proliferation pattern of postembryonic neuroblasts
in the brain of Drosophila melanogaster. Developmental Biology 149,
134– 148.
Ito, K., Awano, W., Suzuki, K., Hiromi, Y., Yamamoto, D., 1997. The
Drosophila mushroom body is a quadruple structure of clonal units each
of which contains a virtually identical set of neurones and glial cells.
Development 124, 761–771.
Kraft, R., Levine, R.B., Restifo, L.L., 1998. The steroid hormone 20hydroxyecdysone enhances neurite growth of Drosophila mushroom
body neurons isolated during metamorphosis. Journal of Neuroscience
18, 8886–8899.
Kucharski, R., Maleszka, R., Hayward, D.C., Ball, E.E., 1998. A royal jelly
protein is expressed in a subset of Kenyon cells in the mushroom bodies
of the honey bee brain. Naturwissenschaften 85, 343–346.
Kurusu, M., Nagao, T., Walldorf, U., Flister, S., Gehring, W.J., FurukuoTokunaga, K., 2000. Genetic control of development of the mushroom
bodies, the associative learning centers in the Drosophila brain, by the
eyeless, twin of eyeless, and dachshund genes. Proceedings of the
National Academy of Sciences of the United States of America 97,
2140–2144.
Kurusu, M., Awasaki, T., Masuda-Nakagawa, L.M., Kawauchi, H., Ito, K.,
Furukubo-Tokunaga, K., 2002. Embryonic and larval development of
the Drosophila mushroom bodies: concentric layer subdivisions and the
role of fasciclin II. Development 129, 409–419.
Larsson, M., Hansson, B., Strausfeld, N., 2002. The mushroom bodies of
the scarab beetle Pachnoda marginata. Society for Neuroscience
Abstracts, 28.
Lee, T., Luo, L., 1999. Mosaic analysis with a repressible cell marker for
studies of gene function in neuronal morphogenesis. Neuron 22,
451– 461.
Lee, T., Lee, A., Luo, L., 1999. Development of the Drosophila mushroom
bodies: sequential generation of three distinct types of neurons from a
neuroblast. Development 126, 4065–4076.
Lee, T., Marticke, S., Sung, C., Robinow, S., Luo, L., 2000a. Cell-
100
S.M. Farris, I. Sinakevitch / Arthropod Structure & Development 32 (2003) 79–101
autonomous requirement of the USP/EcR-B ecdysone receptor for
mushroom body neuronal remodeling in Drosophila. Neuron 28,
807– 818.
Lee, T., Winter, C., Marticke, S.S., Lee, A., Luo, L., 2000b. Essential roles
of Drosophila RhoA in the regulation of neuroblast proliferation and
dendritic but not axonal morphogenesis. Neuron 25, 307 –317.
Li, Y., Strausfeld, N.J., 1997. Morphology and sensory modality of
mushroom body extrinsic neurons in the brain of the cockroach,
Periplaneta americana. The Journal of Comparative Neurology 387,
631– 650.
Li, Y.-S., Strausfeld, N.J., 1999. Multimodal efferent and recurrent neurons
in the medial lobes of cockroach mushroom bodies. The Journal of
Comparative Neurology 409, 647 –663.
Malaterre, J., Strambi, C., Chiang, A.-S., Aouane, A., Strambi, A., Cayre,
M., 2002. Development of cricket mushroom bodies. The Journal of
Comparative Neurology 452, 215 –227.
Malun, D., 1998. Early development of mushroom bodies in the brain of the
honeybee Apis mellifera as revealed by BrdU incorporation and
ablation experiments. Learning and Memory 5, 90–101.
Malzacher, P., 1968. Die Embryogenese des Gehirns paurometaboler
Insekten Untersuchungen an Carausius morosus und Periplaneta
americana. Zeitschrift für Morphologie der Tiere 62, 103–161.
Martini, S.R., Roman, G., Meuser, S., Mardon, G., Davis, R.L., 2000. The
retinal determination gene, dachshund, is required for mushroom body
cell differentiation. Development 127, 2663–2672.
Masson, C., 1970. Mise en évidence, au cours de l’ontogénèse d’une fourmi
primitive (Mesoponera caffraria F. Smith), d’une prolifération tardive
au niveau des cellules globuleuses (kGlobuli cellsl) des corps
pedonculés. Zeitschrift für Zellforschung 106, 220–231.
Menzel, R., 2001. Searching for the memory trace in a mini-brain, the
honeybee. Learning and Memory 8, 53 –62.
Millikin, C., Stark, R., 1990. Growth of the corpora pedunculata during
development of the cockroach brain. American Zoologist 30, 121A.
Mizunami, M., Iwasaki, M., Okada, R., Nishikawa, M., 1998a. Topography
of four classes of Kenyon cells in the mushroom bodies of the
cockroach. The Journal of Comparative Neurology 399, 162–175.
Mizunami, M., Okada, R., Li, Y.-S., Strausfeld, N., 1998b. Mushroom
bodies of the cockroach: activity and identities of neurons recorded in
freely moving animals. The Journal of Comparative Neurology 402,
501– 519.
Mobbs, P.G., 1982. The brain of the honeybee Apis mellifera L. The
connections and spatial organization of the mushroom bodies.
Philosophical Transactions of the Royal Society of London B 298,
309– 354.
Moreau-Fauvarque, C., Taillebourg, E., Boissoneau, E., Mesnard, J., Dura,
J.-M., 1998. The receptor tyrosine kinase gene linotte is required for
neuronal pathway selection in the Drosophila mushroom bodies.
Mechanisms of Development 78, 47 –61.
Ng, J., Nardine, T., Harms, M., Tzu, J., Goldstein, A., Sun, Y., Dietzl, G.,
Dickson, B., Luo, L., 2002. Rac GTPases control axon growth,
guidance and branching. Nature 416, 442 –447.
Nordlander, R.H., Edwards, J.S., 1968. Morphology of the larval and adult
brains of the monarch butterfly, Danaus plexippus plexippus L. Journal
of Morphology 126, 67–94.
Nordlander, R.H., Edwards, J.S., 1970. Postembryonic brain development
in the monarch butterfly, Danaus plexippus plexippus L. III. Morphogenesis of centers other than the optic lobes. Wilhelm Roux’ Archive
164, 247–260.
Noveen, A., Daniel, A., Hartenstein, V., 2000. Early development of the
Drosophila mushroom body: the roles of eyeless and dachshund.
Development 127, 3475–3488.
O’Shea, M., Colbert, M., Williams, L., Dunn, S., 1998. Nitric oxide
compartments in the mushroom bodies of the locust brain. Neuroreport
9, 333 –336.
Okada, R., Ikeda, J., Mizunami, M., 1999. Sensory responses and
movement-related activities in extrinsic neurons of the cockroach
mushroom bodies. Journal of Comparative Physiology A 185,
115–129.
Panov, A.A., 1957. The structure of the brain in insects in successive stages
of postembryonic development. Revue d’Entomologie de URSS 36,
269–284.
Panov, A., 1962. The nature of cell reproduction in the central nervous
system of the nymph of the house cricket (Gryllus domesticus L. Orth.,
Insecta). Doklady Akademii Nauk SSSR 143, 471–474.
Panov, A., 1966. The correlation of ontogenetical development of the
central nervous system of Gryllus domesticus L. and Gryllotalpa
gryllotalpa L. (Orthoptera, Grylloidea). Revue d’Entomologie de URSS
45, 326–340.
Pascual, A., Préat, T., 2001. Localization of long-term memory within the
Drosophila mushroom body. Science 294, 1115–1117.
Pearson, L., 1971. The corpora pedunculata of Sphinx ligustri L. and other
Lepidoptera: an anatomical study. Philosophical Transactions of the
Royal Society of London B 259, 477–516.
Prokop, A., Technau, G.M., 1994. Normal function of the mushroom body
defect gene of Drosophila is required for the regulation of the number
and proliferation of neuroblasts. Developmental Biology 161,
321–337.
Salecker, I., Boeckh, J., 1995. Embryonic development of the antennal
lobes of a hemimetabolous insect, the cockroach Periplaneta
americana: light and electron microscopic observations. The Journal
of Comparative Neurology 352, 33–54.
Schildberger, K., 1984. Multimodal interneurons in the cricket brain:
properties of identified extrinsic mushroom body cells. Journal of
Comparative Physiology A 154, 71– 79.
Schrader, K., 1938. Untersuchungen über die Normalentwicklung des
Gehirns und Gehirntransplantationen bei der Mehlmotte Ephestia
kühniella Zeller nebst einigen Bemerkungen über das Corpus Allatum.
Biologisches Zentralblatt 58, 52 –90.
Schröter, U., Malun, D., 2000. Formation of antennal lobe and mushroom
body neuropils during metamorphosis in the honeybee, Apis mellifera.
The Journal of Comparative Neurology 422, 229– 245.
Schürmann, F., 1973. Über die Struktur der Pilzkörper des Insektengehirns.
III Die Anatomie der Nervenfasern in den Corpora pedunculata bei
Acheta domesticus L., (Orthopt.): eine Golgi-Studie. Zeitschrift für
Zellforschung und Mikroskopische Anatomie 145, 247– 285.
Schürmann, F., Ottersen, O., Honegger, H., 2000. Glutamate-like
immunoreactivity marks compartments of the mushroom bodies in
the brain of the cricket. The Journal of Comparative Neurology 418,
227–239.
Scott, E., Lee, T., Luo, L., 2001. enok encodes a Drosophila putative
histone acetyltransferase required for mushroom body neuroblast
proliferation. Current Biology 11, 99–104.
Scotto, L.S., Strambi, C., Strambi, A., Charpin, P., Augier, R., Aouane, A.,
Cayre, M., 2000. Influence of environmental simulation on neurogenesis in the adult insect brain. Journal of Neurobiology 45, 162–171.
Scotto-Lomassese, S., Strambi, C., Strambi, A., Cayre, M., 2002. Sensory
inputs stimulate progenitor cell proliferation in an adult insect brain.
Current Biology 12, 1001– 1005.
Simon, A., Boquet, I., Synguélakis, M., Préat, T., 1998. The Drosophila
putative kinase linotte (derailed) prevents central brain axons from
converging on a newly described interhemispheric ring. Mechanisms of
Development 76, 45–55.
Sinakevitch, I., Farris, S.M., Strausfeld, N.J., 2001. Taurine-, aspartate- and
glutamate-like immunoreactivity identifies chemically distinct subdivisions of Kenyon cells in the cockroach mushroom body. The Journal of
Comparative Neurology 439, 352–367.
Sinakevitch, I., Strausfeld, N.J., 2002. Affinity of mushroom body modules
to antibodies raised against neuromodulators and transmitters. FENS
Forum July 2002, Paris. European Journal of Neuroscience Abstracts
228.13, p 445.
Strausfeld, N.J., 1976. Atlas of an Insect Brain, Springer-Verlag, Berlin.
Strausfeld, N.J., 2002. Organization of the honey bee mushroom body:
S.M. Farris, I. Sinakevitch / Arthropod Structure & Development 32 (2003) 79–101
representation of the calyx within the vertical and gamma lobes. The
Journal of Comparative Neurology 450, 4 –33.
Strausfeld, N.J., Li, Y.-S., 1999. Representation of the calyces in the medial
and vertical lobes of cockroach mushroom bodies. The Journal of
Comparative Neurology 409, 626– 646.
Strausfeld, N.J., Hansen, L., Li, Y., Gomez, R.S., Ito, K., 1998. Evolution,
discovery, and interpretations of arthropod mushroom bodies. Learning
and Memory 5, 11–37.
Strausfeld, N.J., Homberg, U., Kloppenburg, P., 2000. Parallel organization
in honey bee mushroom bodies by peptidergic Kenyon cells. The
Journal of Comparative Neurology 424, 179 –195.
Strausfeld, N.J., Vilinski, I., Sinakevitch, I., 2003. The mushroom bodies of
Drosophila melanogaster: an immunocytological and Golgi study of
Kenyon cell organization in the calyces and lobes. Microscopy
Research and Technique in press.
Takeuchi, H., Kage, E., Sawata, M., Kamikouchi, A., Ohashi, K., Ohara,
M., Fujiyuki, T., Kunieda, T., Sekimizu, K., Natori, S., Kubo, T., 2001.
Identification of a novel gene, Mblk-1, that encodes a putative
transcription factor expressed preferentially in the large-type Kenyon
cells of the honeybee brain. Insect Molecular Biology 10, 487–494.
Technau, G.M., 1984. Fiber number in the mushroom bodies of adult
Drosophila melanogaster depends on age, sex, and experience. Journal
of Neurogenetics 1, 113 –126.
Technau, G.M., Heisenberg, M., 1982. Neural reorganisation during
metamorphosis of the corpora pedunculata in Drosophila melanogaster.
Nature 295, 405– 407.
Tettamanti, M., Armstrong, J.D., Endo, K., Yang, M.Y., FurukuboTokunaga, K., Kaiser, K., Reichert, H., 1997. Early development of the
Drosophila mushroom bodies, brain centres for associative learning and
memory. Development Genes and Evolution 207, 242– 252.
Truman, J.W., Bate, M., 1988. Spatial and temporal patterns of
neurogenesis in the CNS of Drosophila melanogaster. Developmental
Biology 125, 146 –157.
Truman, J., Talbot, W., Fahrbach, S., Hogness, D., 1994. Ecdysone receptor
expression in the CNS correlates with stage-specific responses to
ecdysteroids during Drosophila and Manduca development. Development 120, 219 –234.
Verkhusha, V.V., Otsuna, H., Awasaki, T., Oda, H., Tsukita, S., Ito, K.,
2001. An enhanced method of red fluorescent protein DsRed for double
labeling and developmental timer of neural fiber bundle formation.
Journal of Biological Chemistry 276, 29621–29624.
101
Volkmar, F.R., Greenough, W.T., 1972. Rearing complexity affects
branching of dendrites in the visual cortex of the rat. Science 176,
1445–1447.
Vowles, D., 1964. Olfactory learning and brain lesions in the wood ant
(Formica rufa). Journal of Comparative Physiology and Psychology 58,
105– 111.
Wang, J., Zugates, C.T., Liang, I.H., Lee, C.-H.J., Lee, T., 2002.
Drosophila dscam is required for divergent segregation of sister
branches and suppresses ectopic bifurcation of axons. Neuron 33,
559– 571.
Weiss, M.J., 1974. Neuronal connections and the function of the corpora
pedunculata in the brain of the American cockroach, Periplaneta
americana (L.). Journal of Morphology 142, 21 –70.
Weiss, M., 1977. Evolutionary heritage of the corpora pedunculata of the
orthopteran brain. American Zoologist 17, 924.
Weiss, M., 1981. Structural patterns in the corpora pedunculata of
Orthoptera: a reduced silver analysis. The Journal of Comparative
Neurology 203, 515–553.
Withers, G.S., Fahrbach, S.E., Robinson, G.E., 1993. Selective neuroanatomical plasticity and division of labour in the honeybee. Nature 364,
238– 240.
Withers, G.S., Fahrbach, S.E., Robinson, G.E., 1995. Effects of experience
and juvenile hormone on the organization of the mushroom bodies of
honey bees. Journal of Neurobiology 26, 130– 144.
Witthöft, W., 1967. Absolute Anzahl und Verteilung der Zellen im Hirn der
Honigbiene. Zeitschrift für Morphologie der Tiere 61, 160–164.
Yang, M.Y., Armstrong, J.D., Vilinsky, I., Strausfeld, N.J., Kaiser, K.,
1995. Subdivision of the Drosophila mushroom bodies by enhancertrap expression patterns. Neuron 15, 45 –54.
Yao, T., Segraves, W., Oro, A., McKeown, M., Evans, R., 1992.
Drosophila ultraspiracle modulates ecdysone receptor function via
heterodimer formation. Cell 71, 63–72.
Yao, T., Forman, B., Jiang, Z., Cherbas, L., Chen, J., McKeown, M.,
Cherbas, P., Evans, R., 1993. Functional ecdysone receptor is the
product of EcR and ultraspiracle genes. Nature 366, 476– 479.
Younossi-Hartenstein, A., Nassif, C., Green, P., Hartenstein, V., 1996.
Early neurogenesis of the Drosophila brain. The Journal of Comparative Neurology 370, 313–329.
Zars, T., Fischer, M., Schulz, R., Heisenberg, M., 2000. Localization of a
short-term memory in Drosophila. Science 288, 672–675.