Current Biology 16, 12–23, January 10, 2006 ª2006 Elsevier Ltd All rights reserved DOI 10.1016/j.cub.2005.11.062
A Conserved Role for Drosophila Neuroglian
and Human L1-CAM in Central-Synapse Formation
Tanja A. Godenschwege,1,* Lars V. Kristiansen,2
Smitha B. Uthaman,1Michael Hortsch,2
and Rodney K. Murphey1
1Department of Biology
Morrill Science Center
University of Massachusetts, Amherst
Amherst, Massachusetts 01003
2Department of Cell and Developmental Biology
University of Michigan
Ann Arbor, Michigan 48109
Background: Drosophila Neuroglian (Nrg) and its verte-
brate homolog L1-CAM are cell-adhesion molecules
(CAM) that have been well studied in early develop-
mental processes. Mutations in the human gene result
in a broad spectrum of phenotypes (the CRASH-
syndrome) that include devastating neurological dis-
orders such as spasticity and mental retardation.
Although the role of L1-CAMs in neurite extension
and axon pathfinding has been extensively studied,
much less is known about their role in synapse for-
Results: We found that a single extracellular missense
mutation in nrg849mutants disrupted the physiologi-
cal function of a central synapse in Drosophila. The
identified giant neuron in nrg849mutants made a syn-
aptic terminal on the appropriate target, but ultrastruc-
tural analysis revealed in the synaptic terminal a dra-
matic microtubule reduction, which was likely to be
the cause for disrupted active zones. Our results re-
veal that tyrosine phosphorylation of the intracellular
ankyrin binding motif was reduced in mutants, and
cell-autonomous rescue experiments demonstrated the
indispensability of this tyrosine in giant-synapse forma-
tion. We also show that this function in giant-synapse
formation was conserved in human L1-CAM but neither
in human L1-CAM with a pathological missense muta-
tion nor in two isoforms of the paralogs NrCAM and
Conclusions: We conclude that Nrg has a function in
synapse formation by organizing microtubules in the
synaptic terminal. This novel synaptic function is con-
served in human L1-CAM but is not common to all L1-
type proteins. Finally, our findings suggest that some
mans may result from a disruption in synapse formation
rather than in axon pathfinding.
The L1-type cell-adhesion molecules (CAMs) are prime
examples of multifunctional molecules with a multiplic-
ity of binding partners [1, 2]. In vertebrate genomes, typ-
ically four L1-type genes are found (L1-CAM, Neuro-
fascin, Nr-CAM, and CHL1), whereas the Drosophila
genome contains only a single L1-type family member,
called Neuroglian (Nrg) . In vertebrates, L1-type pro-
teins have been implicated in cell migration, neurite
extension, axon pathfinding, myelination, fasciculation,
synapse targeting, and LTP [4–8]. Similarly, in Drosoph-
ila, Nrg has been shown to be important in neurite
outgrowth, axon guidance, and sensory-neuron migra-
tion [7, 9–12]. The overall protein domain structures of
L1-type proteins are similar (Figure 1A), containing
a large extracellular domain consisting of six immuno-
globulin (Ig)-like repeats and up to five type-3 fibronec-
mains of most L1-type proteins contain ankyrin (FIGQY)
and ezrin-radixin-moesin (ERM) binding motifs, which
provide links to the spectrin and the actin cytoskeleton
[13–15]. Interestingly, phosphorylation of the ankyrin
phorylated motif becomes a target for other interaction
partners [16–18]. One such partner for the phospho-
FIGQY motif is doublecortin, a microtubule stabilizing
In humans, more than 140 mutations at different sites
throughout the L1-CAM gene have been identified as
the cause of a variety of neurological disorders that
are associated with mental retardation, hydrochepha-
lus, and spasticity of the lower limbs (CRASH syn-
drome) [20–23]. One of these, the pathological human
L1H210Qmutation, is similar to the Drosophila nrg849
allele, which contains a single point mutation (NrgS213L).
In the human, the mutation affects a surface residue
that is in the second Ig domain and disrupts homophilic
binding in trans [23, 24]. The fly mutation is in a similar
position, but the homophilic binding has not been
tested directly. The nrg849allele was originally discov-
ered in a histological screen for structural brain defects
and has been found to affect olfactory learning as well
as locomotor behavior in adult flies [25, 26]. Neuroglian
null mutants in flies are embryonic lethal. This is in con-
trast to the vertebrate L1-CAM null mutants, which are
viable probably as a result of redundancy of paralogous
genes. In both vertebrates and invertebrates, a com-
plete loss of function of Nrg/L1-CAM disrupts neurite
outgrowth and axonal guidance, making it difficult to
assess a potential function of this protein in later devel-
opmental processes such as synapse formation. In
contrast, nrg849flies are viable, with the mutation likely
to disrupt only a subset of Nrg functions. We used the
Drosophila giant fiber (GF) system to study the effects
of the missense mutation on neuronal development.
The GF system is a simple neuronal circuit that medi-
ates the escape response of the fly and allows one to
study the assembly of a central synapse anatomically
and physiologically at the single-cell level (Figure 1B)
nrg849Mutants Exhibit a Disrupted Giant Synapse
The function of the giant synapse in nrg849flies was
tested with standard electrophysiological methods (Fig-
ures 1B and 1C) [28, 30]. Whereas we focused on the
monosynaptic connection of GF-TTMn (tergo-trochan-
teral motor neuron)—which is a mixed chemical and
electrical synapse—in our study,we used thedisynaptic
connection from the GF to the flight motor neuron
of the presence of the GF in the target area [29–32]. Two
criteria, response latency and the ability to follow stimuli
strength and reliability of the synapse, and it was found
that the GF-TTMn connection was disrupted in virtually
all nrg849flies (Table 1).
We observed three mutant phenotypes. In most of the
specimens, the TTM response was absent (23%) or
severely weakened (61%), but the DLM response was
always present though occasionally weakened (Fig-
ure 1C, Table 1). Interestingly, the strength of the phys-
was weaker than in specimens that lack only the gap-
junctional connections [31, 33] or lack only the chemical
transmission , suggesting that both electrical and
chemical transmission of the giant synapse was af-
fected. Finally, we could not record a response from ei-
ther the TTM or the DLM in about 14% of nrg849mutant
animals, indicating that the GF is not in the target area,
probably as a result of a pathfinding error (Figure 1C).
ruption of the neuromuscular junction (NMJ) because
the thoracic stimulation revealed the presence of the
TTMn and DLMn NMJs in all genotypes (Figure 1C,
below dashed line).
With respect to morphology, in most cases (83%) the
mutant axons reached the target area in the second tho-
(Figures 2B2and 2B4). Although the GF-TTMn connec-
tion was disrupted in almost all nrg849flies, in about
one-third the presynaptic terminals appeared to be
Figure 1. Neuroglian and the Giant-Fiber
(A) Schematic of Neuroglian derivatives and
L1-CAM. Asterisks indicate a serine-to-leu-
cine missense mutation in the second Ig
domain in nrg849allele and the corresponding
histidine-to-glutamine human pathological
missense mutation in L1-CAM.
(B) Schematic of the giant-fiber system. For
simplicity, only the left GF circuitry has been
highlighted in color. The giant fiber (GF,
shown in orange) soma is in the brain, and
the axon makes a synaptic connection to
the peripheral synapsing interneuron (PSI, in
ron (TTMn, in blue) in the second thoracic
neuromere. ThePSI synapsesonto the dorsal
longitudinal motor neurons (DLMn, in green).
TTMn and DLMn drive the jump (TTM) and
flight muscles (DLM), respectively. The sche-
maticalsodepicts the stimulus andrecording
arrangement. Recording from the left GF cir-
cuitry is depicted, but we recorded from
both muscles on both sides. Two methods
of stimulation are illustrated: Brain stimula-
tion was used to activate the GF, which then
activates the motor neurons (TTMn and
DLMn), and thoracic stimulation was used
to bypass the GF and excite the motor neu-
(C)Recordings from wild-type andnrg849mu-
tant adult flies. In the wild-type, the TTM re-
sponse latency was 0.8 ms, and the DLM
response latencyis 1.4 ms. The TTM pathway
was able to follow stimuli at 100 Hz, but only
nine out of ten possible responses were
seen at the DLM (asterisk indicates the missing synaptic potential). In most nrg849mutants (61%), a weak TTM but a normal DLM response
was observed in response to brain stimulation. The TTM response latency was increased (1.6 ms) and/or had defects in following stimuli given
at 100 Hz. In this example, only the first synaptic potential was recorded in a series of ten stimuli. In 23%, we found no TTMn response but the
DLM response was present although sometimes weakened. Finally, in 14% of the nrg849specimens, no responses (asterisks) could be recorded
from either the TTM or the DLM when stimulated in the brain, but occasional spontaneous responses (#) indicated the presence of a NMJ. The
defects in the TTM pathway can be attributed to a disruption in the giant synapse because thoracic stimulation revealed that the NMJ was pres-
ent and was able to follow stimuli given at 100 Hz (shown below the dashed line). However, it should be noted that in many nrg849specimens, the
amplitude of the responses was smaller in comparison to wild-type controls, suggesting that the nrg849mutation may also affect the adult TTMn
Neuroglian/L1-CAM in Central-Synapse Formation
morphologically normal at the light-microscope level
(Figure 2B4, white bracket, Table 1). However, some
synaptic terminals did not grow to a normal diameter
(43%, Table 1, Figure 2B2, white bracket), and a few
failed to extend their presynaptic terminal laterally (ap-
proximately 10%, Table 1). Finally, a small number never
reached the target because of pathfinding errors in the
brain (17%, Table 1, Figure 2B1, white arrowhead). With
respect to these pathfinding errors in the brain, it should
be noted that all GFs stalled in similar position in the
suboesophangeal ganglion. Once the GFs had exited
the brain, we did not observe any guidance defects in
get area in the second thoracic neuromere. In contrast,
we did not observe any guidance defects in the medial
TTMn dendrite, which is the postsynaptic target of the
GF (Figure 2B3, black arrow).
Although 30% of the giant-fiber terminals in nrg849
mutant animals appeared anatomically normal at the
light-microscope level, the giant synapses of virtually
all specimens were functionally disrupted (Table 1). In
order to further characterize the structure–function cor-
relation as well as to test for the presence of a gap-
junctional connection, we obtained electrophysiological
sequently injected Lucifer Yellow into the giant axon of
the same specimens. This allowed for the detection of
defects in GF anatomy as well as in dye coupling be-
tween the GF and the TTMn. Overall, only 49% of the
GFsin nrg849specimens revealeddyecoupling between
the GF and the TTMn in comparison to controls (95%
specimens with trans-synaptic fills). Though the TTMn
response was abnormal in all specimens, we observed
trans-synaptic fills in some normal appearing as well
as in abnormal-sized giant nrg849terminals. Finally, it
should be noted that in those specimens with dye
coupling and a physiologically mutant TTM response
weaker in comparison to those in adults with a wild-type
ence of gap junctions in half of the specimens but sug-
gest that the electrical synapse is weakened.
Ultrastructural Defects in nrg849Specimens
To further characterize the synaptic defects in nrg849
mutants, we analyzed the ultrastructure of the axon as
well as the contact region between GF and TTMn (Fig-
ure 3A). We found multiple ultrastructural defects in
wild-type specimens were crowded with synaptic ves-
icles, pleomorphic vesicles, and tubulo-membranous
structures (Figures 3B2–4). In contrast, the presynaptic
terminal in nrg849mutant specimens appeared rather
‘‘empty’’ (Figures 3C2–4), with very few vesicles or vac-
uoles seen. We quantified the average amount of syn-
aptic and pleomorphic vesicles in a 0.25 mm2circle
around the T bars and found a highly significant differ-
ence (Figure 3G) between the wild-type (19 vesicles/
0.25 mm2) and the mutant (3 vesicles/0.25 mm2). How-
ever, it should be noted that the reduction of various
membranous vesicles and organelles is not only re-
stricted to the active zones surrounding the T bars but
is observed throughout the GF-TTMn contact region
(compare Figures 3B2–4and Figures 3C2–4). Additionally,
in wild-type specimens, many mitochondria (Figure 3E,
0.65 mitochondria/mm2) can be found in the synaptic
terminal, often closely associated with the synaptic
contact region (Figures 3B2–4), but significantly fewer
mitochondria (Figure 3E, 0.16 mitochondria/mm2) were
and they appeared to be much farther away from the
Table 1. GF-TTMn Synapse Physiology and Morphology under nrg Loss- and Gain-of-Function Conditions
in %Anatomy in %
+/y or +/+
nrg849/y or nrg849/nrg849
aUAS-constructs were expressed either presynaptically (pre) in the GF or postsynaptically (post) in the TTMn or both.
bA wild-type giant synapse is defined as a TTM with response latency % 1 ms and ability to follow stimuli one-to-one at 100 Hz when stimulated
in the brain.
cPercent of GFs that exhibited a synaptic terminal with normal diameter in the second thoracic neuromere.
dPercent of GFs that exhibited abnormally thin synaptic terminals in the second thoracic neuromere.
ePercent of GFs that exhibited abnormally short or no synaptic terminals although the GF have reached the target area in the second thoracic
fPercent of GFs terminals that were found in the brain as seen in Figure 3B1.
imens, many microtubules (Figure 3B3, black arrow-
heads) were observed in the synaptic terminal (Figure
3D, 4.59 microtubles/mm2) but were almost completely
absent in nrg849mutants (Figure 3D, 0.14 microtubles/
mm2, Figures 3C2–4). Strikingly, despite these apparent
differences, the comparison of wild-type and nrg849
specimens exhibited no differences in some of the
other inherent chemical and electrical synaptic compo-
nents. For example, presynaptic T bars (Figures 3B2–4
and2C2–4,blackarrows, insets)andpostsynaptic densi-
ties were seen in wild-type and mutant individuals.
the wild-type (0.43 T bars/mm2), the difference is not
statistically significant (Figure 3F, p = 0.44, two-tailed
Student’s t test). Similarly, the close approximation of
pre- and postsynaptic membrane, (intercellular space
1–5 nm, Figure 3, white arrowheads) often associated
with a single layer of presynaptic vesicles characteristic
was observed in mutant and wild-type synapses. This
nrg849mutants. This suggests that Nrg is not important
for the assembly ofT bars, but isessential for the normal
formation of other active-zone components, such as
synaptic vesicles and mitochondria, that are required
for proper synaptic function.
In order to determine whether the reduction of
vesicles and mitochondria in the nrg849mutant was
due to a defect in the axon, we compared wild-type
and nrg849axons ultrastructurally (Figures 3A, 3B1, and
3C1). In contrast to the synaptic terminal, the axon con-
tains the same average amount of microtubules/mm2
(Figures 3D). Recently, it has been demonstrated that
mutants with transport defects for mitochondria show
a difference in the abundance of mitochondria in the
axon [35, 36], but we saw no such differences between
wild-type and nrg849mutants in the axon (Figures 3E,
3B1, and 3C1). In conclusion, although we did not assay
the axonal transport directly, our ultrastructural analysis
suggest that the phenotypes found in the synaptic ter-
minal are unlikely to be a result of a defect of axon trans-
port or morphogenesis.
In summary, we found dramatic differences in the
mutant synaptic terminal but not inthe axon, suggesting
that the nrg849phenotypes are due to a disruption in the
formation of a mature, fully functional synapse.
Pre- and Postsynaptic Nrg Function Requires
the Intracellular Domain
The synaptic defects observed in nrg849specimens are
likely to be caused by the nrg mutation because the
Figure 2. GF and TTMn Morphology in Wild-Type and nrg849
(A1) depicts the GF cell bodies and dendrites in the brain of an adult
(A2) The GF axons extended their presynaptic terminals laterally
(black brackets) in the second thoracic neuromere (T2, higher mag-
nification in the inset).
(A3) The medial dendrite of TTMn (black arrow) is the postsynaptic
target of the GF (the white arrow is the lateral dendrite of TTMn).
(A4) shows a trans-synaptic fill (black arrowheads) from the presyn-
aptic GF terminal (black bracket) to the TTMn of a physiologically
wild-type giant synapse (0.8 ms response latency; 100% responses
to 100 Hz stimuli).
(B1) Pathfinding defects were found with low penetrance in the brain
of nrg849specimens. In this adult, the left GF terminal has stalled in
the suboesophangeal ganglion (white arrowhead).
(B2) In this adult nrg849individual, both GFs reached T2, but had
abnormally thin terminals (white brackets) as seen in higher magni-
fication in the inset.
(B3) The medial dendrite (black arrow) and the lateral dendrite (white
arrow) of the TTMn in nrg849mutant flies are indistinguishable from
(B4)A weak trans-synaptic fill (black arrowheads) across the presyn-
aptic GF terminal (white bracket) was observed in this nrg849speci-
men, which had a weakened TTM response (1.2 ms response la-
tency, 50% response to 100 Hz stimuli) but the synaptic terminal
appeared normal at light-microscopic level.
Neuroglian/L1-CAM in Central-Synapse Formation
Figure 3. Ultrastructure of the Presynaptic Terminals
(A) Schematic of the GF system showing the three ultrastructural regions examined (Axon1, Axon2, and Synapse).
(B1) Thin section through a wild-type axon.
(B2–4) Ultrastructure of the presynaptic terminal in the wild-type. The electron micrograph of the contact region of the GF and the TTMn in wild-
type specimens suggests the presence of chemical and electrical synapses indicated by T bars (black arrows, inset in B4) and the 1–5 nm
intercellular space often associated with a single layer of vesicles (white arrowheads), respectively.
(C1) Thin section through an axon in a mutant specimen.
(C2–4). Ultrastructure of nrg849mutant synapses. There was a dramatic reduction of synaptic and pleomorphic vesicles, as well as of tubular
structures in mutant presynaptic terminals. In addition, the number of mitochondria (M) and microtubules (black arrowheads) in nrg849mutants
appeared to be reduced, and they were not as closely associated with the synaptic contact region (C3) as in wild-type preparations.
(D–G) Quantification of phenotypes in wild-type and nrg849mutants; for each genotype, eight giant fibers in five different specimens were ana-
lyzed with two-tailed Student’s t test. No ultrastructural differences were detected between the axonal regions (axon1 and axon2) of wild-type
and nrg849mutant flies. (D) shows the average amount of microtubules per square micrometer in the wild-type and the nrg849. A significant dif-
ference in microtubule density was seen between wild-type and mutant synapses but not in axons. In the wild-type, a total of 222 mm2in region
Axon1, 291 mm2inregion Axon2, and 456mm2inregion Synapse wereanalyzed.Inthe nrg849mutant, atotal of 96mm2inregion Axon1, 291mm2in
region Axon2, and 594 mm2in region Synapse were analyzed. (E) shows the average amount of mitochondria per square micrometer in the wild-
type and the nrg849. A significant difference in the density of mitochondria was seen between the wild-typeand thenrg849synaptic terminals, but
no difference was seen for the axons. The total area analyzed was the same in wild-type and nrg849specimens. (F) shows the average number of
T bars per mm length of contact with the TTMn. No significant difference between wild-type and nrg849mutants was seen with respect to density
of T bars at the GF-TTMn contact region. A total length of 364 mm and 384 mm was analyzed in the wild-type and the nrg849, respectively. (G)
are found around T bars in comparison to the wild-type. A total of 74 T bars in wild-type and 100 T bars in nrg849mutants were analyzed. GF
denotes giant fiber, and TTMn denotes tergo-trochanteral motor neuron. The scale bar is 1 mm in (B1–3) and 0.5 mm in (B4).
GF-TTMn connection was also disrupted in trans-heter-
ozygotes of nrg849and the null mutant nrg14(Table 1).
In order to prove that the phenotypes are due to a dis-
ruption in Nrg function and therefore indicative of an en-
dogenous function in giant-synapse formation, we tried
to rescue the phenotypes by cell-autonomous expres-
sion of wild-type Nrg in the nrg849genetic background.
We used two different UAS-nrg constructs to character-
ize Nrg’s functional and spatial requirements in giant-
synapse formation: full-length neuronal Nrg180(UAS-
nrg180) and an artificial Nrg isoform (UAS-nrgGPI), in
which the extracellular Nrg domain is attached to the
membrane surface by a GPI-moiety (Figure 1A). It has
normal homophilic binding properties, and the expres-
sion of the transgene (UAS-nrgGPI) has been demon-
strated to efficiently activate EGF and Echinoid recep-
tors [11, 37, 38].
Simultaneous pre- and postynaptic expression of the
wild-type (UAS-nrg180) in the GF system of nrg849mu-
tants completely restored the function of the giant syn-
apse in all specimens (Table 2). Expressing exclusively
either pre- or postsynaptically had a slightly lower ca-
pacity to rescue the synaptic defects (70%–77%). This
is involved in giant-synapse formation. In contrast to
full-length Nrg180, no rescue was observed when the
NrgGPIwas expressed pre- or postsynaptically or both
Though NrgGPIhad no rescue capacity, the transgene
does get expressed because its expression in wild-type
background had a dominant-negative effect and disrup-
ted the giant synapse. Overexpression of UAS-nrgGPIon
either side of the synapse had a mildly disruptive effect
on synapse function and morphology but was synergis-
tically enhanced by simultaneous expression on both
sides of the synapse (Table 1). Interestingly, expression
of NrgGPIin the GFS throughout development mimicked
the nrg849phenotypes (Table 1, Figure S1 in the Supple-
mental Data available online); although essentially all
GF-TTMn connections were physiologically disrupted,
21% of the GF terminals appeared to be anatomically
normal at the light-microscopic level whereas 45% had
thin synaptic terminals (Figure S1A, white brackets,
Table 1). A failure to extend a presynaptic terminal later-
ally in the second thoracic neuromere occurred in 31%
ofthe cases (Table 1, Figure S1B, asterisks). Pathfinding
defects in the brain were only rarely observed (Table 1).
Also similar to nrg849, trans-synaptic fills between the
GF and the TTMn were seen with less penetrance (only
15% of the dye-injected GFs, Figure S1D). Although
the overall dendritic structure of TTMn was not affected
by the expression of UAS-nrgGPI, large vesicles were
occasionally observed in the medial TTMn dendrites
(Figure S1C, black arrowheads). Finally, overexpression
of one or multiple copies of UAS-nrg180in wild-type
background had no disruptive effect on the giant syn-
apse anatomically or physiologically (Table 1), sug-
gesting that the NrgGPIacts as a dominant-negative
modulator of synapse formation. It presumably does
so by competing with endogenous Nrg for the same
extracellular interaction partner(s) and thereby uncou-
pling an extracellular signal from an intracellular output.
In summary, our results indicate that Nrg has an en-
dogenous pre- and postsynaptic function in giant-
synapse formation, which is disrupted in the nrg849
allele. We also demonstrate that the intracellular domain
is indispensable for giant-synapse formation, suggest-
ing that the extracellular mutation affects intracellular
Phosphorylation of an Ankyrin Binding Site
Is Essential for Giant Synapse-Formation
In order to address the question of whether intracellular
signaling via the ankyrin binding motif is important for
giant-synapse assembly, we used three assays. First,
we determined whether the phosphorylation of the
ankyrin binding motif (FIGQY) was affected in nrg849
flies. We found that the expression of total amount of
endogenous neuronal Nrg180was not different between
wild-type controls and nrg849flies (Figure 4A). In con-
trast, when we used an antibody that is specific for
the tyrosine-phosphorylated ankyrin binding motif, we
found that there was approximately a 50% reduction
of the phospho-FIGQY Nrg form in the nrg849mutants
in comparison with wild-type flies (Figure 4B) . This
implies that the extracellular mutation in nrg849disrupts
tyrosine-phosphorylation of the ankyrin binding motif in
the intracellular domain.
Second, to further investigate the involvement of the
highly conserved tyrosine in the ankyrin binding motif
in giant-synapse formation, we generated a transgene
sine is changed to a phenylalanine . This mutation
prevents phosphorylation of the ankyrin binding motif
and has been shown in vertebrates to completely dis-
rupt binding to phospho-FIGQY-specific proteins such
as doublecortin . However, ankyrin binding affinity
to the NrgY1234Fprotein is not abolished but rather re-
duced by approximately half, and extracellular homo-
philic adhesion is still present . When we expressed
Table 2. Rescue of the Physiological nrg849Phenotype
aUAS-constructs were expressed either presynaptically (pre) in the
GF or postsynaptically (post) in the TTMn or both.
bA wild-type synapse is defined as a TTM with response latency
% 1 ms and the ability to follow stimuli one-to-one at 100 Hz when
stimulated in the brain.
Neuroglian/L1-CAM in Central-Synapse Formation
NrgY1234Finannrg849mutantbackground, wefound that
cue synaptic phenotypes when compared to control
flies (Table 2). This suggests that the signaling pathway
disrupted by the extracellular missense mutation in
nrg849involves the tyrosine in the FIGQY-motif and
that this signaling pathway is probably not dependent
In our final assay, we then expressed NrgY1234Fin
wild-type background and found that one copy of
UAS-nrgY1234Fhad a disruptive effect on giant-synapse
formation (Table 1). This indicates that NrgY1234Falso
works as a dominant negative and disrupts wild-type
function during giant-synapse formation.
In conclusion, our data suggest a signaling mecha-
nism via phosphophorylation of the FIGQY motif that is
important for giant-synapse formation and that this in-
tracellular output is disrupted by the mutation in the ex-
tracellular domain of Nrg849.
Critical Periods for Nrg in Giant-Synapse Assembly
The data suggest that the phenotypes in nrg849are de-
rived from a defect in synapse formation. In order to fur-
ther test the hypothesis that Nrg has an endogenous
function during synapse assembly, we used two ap-
proaches to determine the temporal requirements: We
perature-sensitive allele (nrg3) to temporally affect Nrg
function. We used the TARGET system , which al-
lows temporal control of the expression of UAS-nrgGPI,
to determine the critical period for the disruptive effect
pressed NrgGPIpre- and postsynaptically for 24 hr (ap-
proximately 16%) ofpupal development (PD) atdifferent
developmental stages in a wild-type background and
assayed the function of the GF-TTMn synapse in adults
(Figures 5A and 5B). Expression of NrgGPIduring path-
finding had no effect on the giant-synapse physiology
in adults (Figure 5B). The same treatment in a stabilized,
mature synapse had only a mild effect on synapse
function (Figure 5B, 70%–80% wild-type responses).
However, the presence of NrgGPIduring nascent syn-
apse formation (Figures 5A and 5B, yellow bar, 35%–
50% wild-type responses) and synapse maturation (Fig-
ure 5B, 25% wild-type responses) severely disrupted
the GF-TTMn connection.
Second, we used a temperature-sensitive allele (nrg3)
to generate a temporal loss-of-function background at
different times and asked when disruption of endoge-
nous Nrg affected assembly and function of the giant
synapse . A temporal loss offunction of endogenous
Nrg for 12.5% or 25% of pupal development during syn-
during synapse maturation was able to disrupt the giant
synapse anatomically (Figures S2B–S2D) and physio-
logically (Figure 5C), with phenotypes similar to nrg849
mutants. However, it should be noted that temporal
loss of function of Nrg during synapse formation and
maturation resulted in much more anatomically pro-
nounced phenotypes in comparison to nrg849mutants.
For example, we see with high penetrance (50% in com-
parison to 10% in nrg849, Table 1) that the complete
synaptic terminal is lacking when the temporal loss of
function was induced during synapse formation (Fig-
ure S2B). This suggests that the missense mutation in
nrg849does not abolish all functions of the protein in
These results show a critical period, beginning at ap-
proximately 25% of pupal development, for Nrg in syn-
apse assembly. Because pathfinding and targeting is
complete by this time, these results suggest that Nrg
has a pathfinding-independent function in the assembly
of the GF-TTMn synapse. In conclusion, our findings
clearly demonstrate that Nrg has an endogenous func-
tion during synapse formation and synapse maturation.
Human L1-CAM Is Able to Rescue the Synaptic
Defects in nrg849
A previous study demonstrated that Nrg180, human L1-
CAM, rat-NrCAM, human-NCAM, and Drosophila FasII
Figure 4. Expression of Wild-Type and Mutant Nrg and Human L1 in the Drosophila
(A) Expression of neuronal Nrg in wild-type and nrg849mutants. The extracellular missense mutation does not affect the expression levels of the
protein; similar amounts are detected in wild-type and nrg849mutants. Immunostaining against Tango, a transcription factor, serves as a loading
(B) Reduction of theFIGQY-phosphorylated Nrg forminnrg849mutants. Less immunoreactivity is foundwith an anti-phospho-FIGQY antibody in
nrg849mutants when compared to wild-type. Different amounts of homogenized CNS were loaded as indicated by brackets, and anti-Tango
staining serves as a loading control.
the fly, but only L1 not L1H210Qwas able to rescue the phenotypes in nrg849mutants. GF-driver flies (A307, no expression) served as a negative
control and anti-Tango immunostaining as a loading control.
are able to rescue axon pathfinding defects of specific
sensory neurons under nrg loss-of-function conditions,
suggesting a functional overlap between these proteins
in this developmental process . We used the same
transgenic lines as well as the expression of a chicken-
Neurofascin isoform in order to determine whether Nrg
cell-adhesion molecules or L1-type proteins.
Strikingly, we found that the human neuronal L1-CAM
function (Table 2). In contrast, four related genes were
tested and failed to rescue the phenotype (rat-NrCAM,
chicken-Neurofascin, human-NCAM, and its Drosophila
homolog Fasciclin II). Hence, it seems that the synaptic
function disrupted in nrg849flies is not purely adhesion
dependent. Although we cannot exclude that alternative
splice variants of NrCAM and Neurofascin other than the
tested ones may have rescue capacity in nrg849flies, our
results nevertheless show that the synaptic function is
ture of all L1-type isoforms.
The single point mutation (NrgS213L, Figure 1A) in the
Drosophila nrg849allele is predicted to affect the surface
residue in the second Ig domain homologous to the
pathological human L1H210Qmutation . In contrast
to wild-type L1, the expression of L1H210Qin the nrg849
mutant background was not able to rescue the function
the UAS-L1H210Qtransgene gets expressed with similar
efficiency as human UAS-L1 (Figure 4C1). These results
suggest that Drosophila Nrg and vertebrate L1-CAM
function that involves the second Ig domain.
The function of L1-type proteins in early neuronal devel-
opment has been intensively studied [3, 6], but less is
known about their contribution to synapse formation.
Recently, it has been shown that the L1-type protein
Neurofascin (but not L1-CAM) is important in direct-
ing GABAergic innervation of the Purkinje axon initial
segment and that this involves ankyrin . Here, we
provide evidence that the ankyrin-independent form of
formation and that this novel function in synapse forma-
common to all L1-type isoforms.
A Conserved Function of Nrg/L1-CAM
during Synapse Formation Is Distinct
from Its Pathfinding Function
L1-type proteins are multifunctional proteins with many
interaction partners, and the complete loss of function
of Nrg protein results in defects in neurite outgrowth
Figure 5. The Timing of Events during Pupal
(A) The timeline illustrates the definition of the
various phases of the giant-synapse assem-
bly during pupal development (PD), which
has been described previously . The yel-
low highlight indicates the critical period for
(B) Temporal effects of NrgGPIat different de-
velopmental stages with the TARGET system
. NrgGPIwas expressed for approximately
16% of pupal development (red bars, ‘‘on’’) at
different stages during the assembly of the
GF-TTMn connection. The expression of
NrgGPIthroughout development resulted in
only 2% of individuals with a wild-type TTM
response (right red bar). In control flies lack-
ing expression of NrgGPI, there was no effect
on the GF-TTMn connection (100% wild-type
responses, rightmost blue bar).
giant-fiber development. Temperature-sensi-
tive nrg3mutants were raised at permissive
temperature (blue line) and exposed for
12.5% (top) or 25% (bottom) of pupal devel-
opment to nonpermissive temperature (red
bars) at different time points during GF devel-
opment. The treated specimens were ana-
lyzed electrophysiologically as adults. In (B)
and (C), the percent at the top of the bar indi-
cates the number of wild-type specimens
obtained after treatment at the correspond-
ing developmental stage. nrg3mutants con-
stantly raised at nonpermissive (non-perm.)
were lethal (y) but when raised at constant
permissive temperature were viable, and
80% of the adults had a wild-type TTM re-
Neuroglian/L1-CAM in Central-Synapse Formation
and axonal guidance as well as in lethality in Drosophila
mutation in the second Ig domain are viable, but syn-
apse formation was disrupted in virtually all animals
and pathfinding was rarely affected. This suggests that
the nrg849missense mutation does not affect the overall
function of Nrg but rather disrupts a subset of functions
that is important for giant-synapse formation, but non-
essential for the outgrowth or guidance of the GF
axon. This implies that the functions of Nrg during
axon pathfinding and synapse formation are distinct
Earlier studies revealed that the extracellular domain
of Nrg and L1-CAM functioning as a ligand is sufficient
for many developmental processes including the stimu-
lation of neurite outgrowth, sensory axon pathfinding,
and eye development [11, 37–39, 41, 42]. For example,
expression of NrgGPIis able to rescue certain aspects
of pathfinding in a nrg loss-of-function background
and is able to activate Echinoid, epidermal growth fac-
tor, (EGF), and fibroblast growth factor (FGF) receptor
signaling [11, 38, 41, 43]. In addition, NrgGPIhas normal
homophilic binding properties and can induce ankyrin
recruitment when binding to full-length Nrg [37, 39].
More recently, it has been shown that Nrg180, human
L1-CAM, rat-NrCAM, human-NCAM, and Drosophila
FasII are able to rescue axon pathfinding defects of
specific sensory neurons under nrg loss-of-function
teins in axonal guidance . However, although neuro-
nal Nrg180and its human homolog L1-CAM were able to
rescue the synaptic dysfunction in the nrg849missense
mutants, we did not find functional overlap with NrgGPI,
rat-NrCAM, NCAM, or FasII. This supports our hypothe-
from its well-known function during neurite extension
and axonal guidance and shows that the intracellular
domain of Nrg is indispensable for synapse formation.
Mechanism of Nrg Function in Synapse Formation
We show that the highly conserved tyrosine of the an-
kyrin binding site (FIGQY) in the intracellular domain is
essential for Nrg function in giant-synapse formation,
and the ultrastructural phenotypes suggest that one of
its functions may be to stabilize microtubules in the syn-
aptic terminal. L1-type proteins have been shown to
generate different microdomains that are either ankyrin
free or ankyrin containing, as determined by the phos-
phorylation status of the tyrosine in the ankyrin binding
motif [16,17]. Thehomophilic interaction ofL1-type pro-
teins in trans or heterophilic interaction in cis induces
the recruitment of ankyrin to the unphosphorylated
FIGQY, which in turn interacts with the spectrin cyto-
skeleton [44–46]. Recently, it has been shown that the
spectrin cytoskeleton is important for stabilization of
the neuromuscular junction and that the loss of spectrin
can induce synapse disassembly and retraction .
Interestingly, synaptic boutons in the NMJ lacking
spectrin exhibit ultrastructural phenotypes similar to
the giant synapse of nrg849mutants, they are devoid of
microtubules, and the synaptic vesicle density is se-
verely reduced. Hence, FIGQY-unphosphorylated Nrg
may have an effect on the microtubule cytoskeleton by
its connection to the spectrin cytoskeleton via ankyrin.
Although we cannot generally exclude a function for
ankyrin binding Nrg in giant-synapse formation, a dis-
ruption in this pathway seems not to be the primary
cause for the phenotypes in nrg849mutants. We have
strong evidence that in nrg849mutants, signaling via
ankyrin-independent Nrg is disrupted and this affects
giant-synapse formation. We found that in nrg849flies,
tyrosine phosphorylation of the FIGQY motif is re-
duced. Furthermore, NrgY1234F, which still has residual
binding affinity for ankyrin, had no capacity to rescue
mutant synapses but rather had a dominant-negative
effect on synapse formation when expressed in a wild-
type background . Interestingly, Nrg/L1 with a phos-
phorylated FIGQY motif is found at cell-cell contact
sites in the nervous system and does not recruit an-
kyrin but has been proposed to bind to phospho-
FIGQY-specific proteins [16, 17]. In vertebrates, one
of these phospho-FIGQY-specific proteins has been
recently identified as doublecortin, a neurogenesis-
specific protein that stabilizes microtubules [16, 17,
19, 48]. Hence, we propose that the phosphorylated
FIGQY Nrg has a function in giant-synapse formation
possibly by anchoring of microtubules in the synaptic
terminal via a protein similar to doublecortin.
Nrg links the plasma membrane to the cytoskeleton,
where it organizes and stabilizes the synaptic terminal,
and when this function is disrupted, it may lead to the
synaptic phenotypes seen in nrg849animals. Hence,
Nrg may beimportant for the stability oftheactive zones
by providing a scaffolding function at the synapse that
affects local signaling. Alternatively, nrg849synaptic ter-
minals may represent nascent synapses that have never
matured because the dramatic cytoskeletal defects in
Finally, it should be noted that Nrg may not only have a
scaffolding function that affects signaling important for
synapse formation but may also signal itself. For exam-
ple, vertebrate L1-CAM can activate mitogen-activated
regulated kinases 1 and 2, which are known to play a
Unfortunately, our results do not allow us to unequiv-
ocally distinguish which interaction on the extracellular
side is affected by the mutation in nrg849flies. Missense
mutations of a surface residue in the second Ig domain
in L1-type proteins have been shown to affect homo-
philic interactions, heterophilic interactions, or both si-
multaneously [1, 52]. The finding that expression of
Nrg180on bothsides ofthesynapse had rescuecapacity
could be simply due to the fact that heterophilic interac-
tion on both sides of the synapse may contribute to syn-
apse formation. However, because NrgS213Lprotein has
some residual function, it is also possible that homo-
philic interaction between wild-type and mutant Nrg
protein could be the reason for a rescue capacity of
Nrg180on either side of the synapse. Hence, we suggest
that the missense mutation in nrg849flies probably dis-
ing giant-synapse formation and thatthe combination of
these interactions results in signaling via the phosphor-
ylated, ankyrin binding motif that is important for syn-
The GFSas aModelto Study L1-CAM Function under
Normal and Pathological Conditions
The finding that vertebrate L1-CAM, but not the tested
paralog isoforms of NrCAM and Neurofascin, rescued
the synaptic defects in the Drosophila nrg849mutants
suggests that the synaptic function is conserved but is
not a feature of all L1-type proteins and therefore is
ical mutations in L1 have been studied in cell culture,
and the results suggest that some pathological defects
may be the result of disrupted neurite outgrowth, exten-
sion,andbranching [23,53].However, untilnow,noneof
the pathological mutations have been characterized
with respect to their affect on synapse formation. We
show that human L1 with a corresponding pathological
mutation (L1H210Q) in the second Ig domain  was
not able not rescue the synaptic defects in the Drosoph-
ila nrg849allele. Hence, it is possible that some of the
clinical L1 phenotypes in humans may also be attributed
to synaptic defects rather than axonal growth and path-
endogenous role of Nrg and L1 in synapse formation but
may also help to understand the pathology of L1-related
All stocks were grown at 22ºC or 25ºC on standard medium. The
following fly stocks were used: UAS-nrgGPI(2ndchromosome),
UAS-nrgY1234F(2ndas well as 2ndand 3rdchromosome), UAS-fasII
(2ndchromosome), UAS-humanL1-CAM (RSLE isoform, 3rdchromo-
some), UAS-hL1H210Q-CAM, UAS-rat-NrCAM (on 2ndor 3rdchro-
mosome), UAS-chicken-Neurofascin (isoform NF17, 2ndand 3rd
chromosome), UAS-humanNCAM (2ndor 3rdchromosome), UAS-
nrg180(2ndchromosome), nrg849, nrg14, and nrg3[9–12, 25, 38, 54].
Three P[GAL]4 lines that express in the GF system were used:
A307, which shows strong expression in the GF and its postsynaptic
targets , c17, which drives expression in the GF but not in its
postsynaptic targets , and shakB-Gal4, which drives expression
only postsynaptically but not presynaptically .
Dye Injection and Immunocytochemistry
For dye injection, the dissected CNS was mounted dorsal side up on
poly-lysine-coated slides. The preparation was covered with saline
and viewed with a 403 water-immersion lens. The GF was identified
in the connective by using DIC optics. A glass electrode (20–80 mU)
filled with 1% aqueous Lucifer Yellow was used to impale the GF in
the connective, and the dye was injected into the GF axon by pass-
ing hyperpolarizing current (3–5 nA) with a Getting 5A amplifier (Get-
immunohistochemistry with rabbit anti-Lucifer Yellow (1:2000,
Molecular Probes) was performed as previously described .
In order to reveal the lac-Z expression, we dissected the CNS of
adults and pupae expressing UAS-lacZ in 100 mM phosphate buffer
(PB) and performed immuncytochemistry with polyclonal rabbit anti
b-galactosidase antibody (1:6000, Cappel, Tunhout, Belgium) as
previously described . For immunoblotting, the CNS of adults
were homogenized in 23 Laemmli probe buffer. The probes were
run on a 4%–20% gradient SDS-page gel (Bio-Rad) and blotted
onto Immobilon-P membrane (Millipore) with the Protean II ready-
gel system (Biorad). Antibodies were used at the following concen-
trations: rabbit anti-phospho-FIGQY (0.8 mg/ml) , mouse mono-
Bank, Iowa, 1:200), goat L1-CAM antisera (sc-1508, 1:200, Santa
Cruz Biotechnology), goat anti-mouse IgG peroxidase-conjuguated
antibody(Sigma),donkey anti-goatIgGHRP (sc-2020,1:2000,Santa
Cruz Biotechnology) and goat anti-rabbit IgG peroxidase-conjug-
uated antibody (Jackson ImmunoResearch Laboratories). For signal
visualization we used Western Lightning Chemiluminescence Re-
agent Plus Kit (PerkinElmer Life Science).
Intracellular recordings from TTM and DLM muscles were obtained
from adult flies in a method similar to that described earlier . The
have been previously described .
The adult CNS was dissected from nrg849and wild-type controls
(Biocore), prefixed in 2.5% glutaraldehyde for 24 hr, and osmicated
(1% osmium tetroxide) for 1 hr. After dehydration, the nervous sys-
tems were embedded in epon-araldite. Serial ultrathin sections (50–
60 nm) were counterstained with 1% aqueous uranyl acetate and
lead citrate and examined with a JEOL100s electron microscope.
The GF-TTMn contact region was photographed, and the negatives
were scanned with a high-resolution flatbed scanner.
Temporal Expression of UAS-nrgGPIand Temporal Loss
of Function with nrg3
We crossedUAS-nrgGPIfliestoA307/Cyo;tub-Gal80ts/TM6. Theoff-
spring were raised at the permissive temperature (we found that
room temperature, 22ºC, was sufficient to suppress NrgGPIdisrup-
tive effects) to prevent UAS-nrgGPIexpression. NrgGPIexpression
was induced for 24 hr (approx. 16% of PD) by a temperature shift
to 30ºC at P0, P0 + 24 hr, P0 + 48 hr, P0 + 72 hr, P0 + 120 hr, and
in adults. The TTM response of adult UAS-nrgGPI/A307;Gal80ts/+
specimens was determined and compared to the following control
specimens: UAS-nrgGPI/A307;TM6/+ (continuous NrgGPIexpres-
Gal80ts/+ (continuously inhibited NrgGPIexpression).
We raised temperature-sensitive nrg3 mutants at permissive tem-
perature (18ºC). As described above, we exposed the specimens to
30ºC for 24 hr (approximately 12.5% of PD) or 48 hr (approximately
25% of PD) at different developmental stages. After the tempera-
tures shift, we continued to raise the nrg3 mutants at 20ºC and as-
sessed the adults electrophysiologically and anatomically with dye
injection into the GFs and compared them to non-temperature-
shifted nrg3 flies, which were continuously raised at 20ºC.
Supplemental Data include two figures and are available with this
article online at: http://www.current-biology.com/cgi/content/full/
We thank X. Shan-Crofts and P. Caruccio for excellent technical as-
sistance, D. A. Callaham for his help and advice with the ultrastruc-
ces. We also thank Drs. P. Callaerts and Y.Y. Kang for sharing
information on unpublished data and fly stocks, as well as Drs. L.
Garcia-Alonso, Vann Bennett, S. DeBelle, and R. Strauss for provid-
ingflystrains and/orantibodies. L.Kristiansen is gratefulforhissup-
port by Drs. E. Bock and V. Berezin. This work was supported by
a grant from the National Sciences Foundation (IBN-0132819) to
M.H. and by the National Institutes of Health grant R01-NS044609
Received: October 18, 2005
Revised: November 21, 2005
Accepted: November 22, 2005
Published: January 9, 2006
1. Haspel, J., and Grumet, M. (2003). The L1CAM extracellular re-
gion: A multi-domain protein with modular and cooperative
binding modes. Front. Biosci. 8, s1210–s1225.
2. Hortsch, M. (1996). The L1 family of neural cell adhesion mole-
cules: Old proteins performing new tricks. Neuron 17, 587–593.
Neuroglian/L1-CAM in Central-Synapse Formation
3. Hortsch, M. (2000). Structural and functional evolution of the L1
family: Are four adhesion molecules better than one? Mol. Cell.
Neurosci. 15, 1–10.
4. Lemmon, V., Farr, K.L., and Lagenaur, C. (1989). L1-mediated
axon outgrowth occurs via a homophilic binding mechanism.
Neuron 2, 1597–1603.
5. Matsumoto-Miyai, K., Ninomiya, A., Yamasaki, H., Tamura, H.,
Nakamura,Y.,and Shiosaka, S.(2003).NMDA-dependent prote-
olysis of presynaptic adhesion molecule L1 in the hippocampus
by neuropsin. J. Neurosci. 23, 7727–7736.
ecule L1: Signaling pathways and growth cone motility. J. Neu-
rosci. Res. 49, 1–8.
7. Itoh, K., Cheng, L., Kamei, Y., Fushiki, S., Kamiguchi, H., Gut-
wein, P., Stoeck, A., Arnold, B., Altevogt, P., and Lemmon, V.
(2004). Brain development in mice lacking L1–L1 homophilic
adhesion. J. Cell Biol. 165, 145–154.
8. Ango, F., di Cristo, G., Higashiyama, H., Bennett, V., Wu, P., and
Huang, Z.J. (2004). Ankyrin-based subcellular gradient of neuro-
fascin, an immunoglobulin family protein, directs GABAergic in-
nervation at purkinje axon initial segment. Cell 119, 257–272.
9. Bieber, A.J., Snow, P.M., Hortsch, M., Patel, N.H., Jacobs, J.R.,
Traquina, Z.R., Schilling, J., and Goodman, C.S. (1989). Dro-
sophila neuroglian: A member of the immunoglobulin superfam-
ily with extensive homology to the vertebrate neural adhesion
molecule L1. Cell 59, 447–460.
10. Hall, S.G., and Bieber, A.J. (1997). Mutations in the Drosophila
neuroglian celladhesionmoleculeaffectmotor neuronpathfind-
ing and peripheral nervous system patterning. J. Neurobiol. 32,
11. Islam, R., Kristiansen, L.V., Romani, S., Garcia-Alonso, L., and
Hortsch, M. (2004). Activation of EGF receptor kinase by
L1-mediated homophilic cell interactions. Mol. Biol. Cell 15,
12. Kristiansen, L.V., Velasquez, E., Romani, S., Baars, S., Berezin,
V., Bock, E., Hortsch, M., and Garcia-Alonso, L. (2005). Genetic
analysis of an overlapping functional requirement for L1- and
NCAM-type proteins during sensory axon guidance in Drosoph-
ila. Mol. Cell. Neurosci. 28, 141–152.
13. Davis, J.Q., and Bennett, V. (1994). Ankyrin binding activity
shared by the neurofascin/L1/NrCAM family of nervous system
cell adhesion molecules. J. Biol. Chem. 269, 27163–27166.
14. Cheng, L., Itoh, K.,and Lemmon,V. (2005).L1-mediated branch-
ing is regulated by two ezrin-radixin-moesin (ERM)-binding
sites, the RSLE region and a novel juxtamembrane ERM-binding
region. J. Neurosci. 25, 395–403.
15. Dickson, T.C., Mintz, C.D., Benson, D.L., and Salton, S.R. (2002).
Functional binding interaction identified between the axonal
CAM L1 and members of the ERM family. J. Cell Biol. 157,
16. Chen, L.,Ong, B., and Bennett, V.(2001).LAD-1, the Caenorhab-
ditis elegans L1CAM homologue, participates in embryonic and
gonadal morphogenesis and is a substrate for fibroblast growth
factor receptor pathway-dependent phosphotyrosine-based
signaling. J. Cell Biol. 154, 841–855.
17. Jenkins, S.M., Kizhatil, K., Kramarcy, N.R., Sen, A., Sealock, R.,
and Bennett, V. (2001). FIGQY phosphorylation defines discrete
populations of L1 cell adhesion molecules at sites of cell-cell
contact and in migrating neurons. J. Cell Sci. 114, 3823–3835.
18. Garver, T.D., Ren, Q., Tuvia, S., and Bennett, V. (1997). Tyrosine
phosphorylation at a site highly conservedin the L1 family of cell
adhesion molecules abolishes ankyrin binding and increases
lateral mobility of neurofascin. J. Cell Biol. 137, 703–714.
19. Kizhatil, K., Wu, Y.X., Sen, A., and Bennett, V. (2002). A new
activity of doublecortin in recognition of the phospho-FIGQY
tyrosine in the cytoplasmic domain of neurofascin. J. Neurosci.
20. Kenwrick, S., Watkins, A., and De Angelis, E. (2000). Neural cell
recognition molecule L1: Relating biological complexity to
human disease mutations. Hum. Mol. Genet. 9, 879–886.
21. Bateman, A., Jouet, M., MacFarlane, J., Du, J.S., Kenwrick, S.,
and Chothia, C. (1996). Outline structure of the human L1 cell
adhesion molecule and the sites where mutations cause neuro-
logical disorders. EMBO J. 15, 6050–6059.
22. Fransen, E., Van Camp, G., D’Hooge, R., Vits, L., and Willems,
P.J. (1998). Genotype-phenotype correlation in L1 associated
diseases. J. Med. Genet. 35, 399–404.
23. De Angelis, E., Watkins, A., Schafer, M., Brummendorf, T., and
Kenwrick, S. (2002). Disease-associated mutations in L1 CAM
interfere with ligand interactions and cell-surface expression.
Hum. Mol. Genet. 11, 1–12.
24. Kang, Y.Y. (2004). The role of the Drosophila L1 CAM Neuroglian
in postembryonic brain development: From cell adhesion to sig-
naling. PhD thesis, Department of Biology and Biochemistry,
University of Houston.
locomotor behavior in the Drosophila brain. J. Neurosci. 13,
26. de Belle, J.S., and Heisenberg, M. (1996). Expression of Dro-
sophila mushroom body mutations in alternative genetic back-
grounds: A case study of the mushroom body miniature gene
(mbm). Proc. Natl. Acad. Sci. USA 93, 9875–9880.
27. King, D.G., and Wyman, R.J. (1980). Anatomy of the giant fibre
pathway in Drosophila. I. Three thoracic components of the
pathway. J. Neurocytol. 9, 753–770.
28. Tanouye, M.A., and Wyman, R.J. (1980). Motor outputs of giant
nerve fiber in Drosophila. J. Neurophysiol. 44, 405–421.
29. Godenschwege, T.A., Hu, H., Shan-Crofts, X., Goodman, C.S.,
and Murphey, R.K. (2002). Bi-directional signaling by Sema-
phorin 1a during central synapse formation in Drosophila. Nat.
Neurosci. 5, 1294–1301.
30. Godenschwege, T.A., Simpson, J.H., Shan, X., Bashaw, G.J.,
Goodman, C.S., and Murphey, R.K. (2002). Ectopic expression
in the giant fiber system of Drosophila reveals distinct roles for
roundabout (Robo), Robo2, and Robo3 in dendritic guidance
and synaptic connectivity. J. Neurosci. 22, 3117–3129.
31. Phelan, P., Nakagawa, M., Wilkin, M.B., Moffat, K.G., O’Kane,
C.J., Davies, J.A., and Bacon, J.P. (1996). Mutations in shak-
ing-B prevent electrical synapse formation in the Drosophila
giant fiber system. J. Neurosci. 16, 1101–1113.
32. Blagburn, J.M., Alexopoulos, H., Davies, J.A., and Bacon, J.P.
(1999). Null mutation in shaking-B eliminates electrical, but not
chemical, synapses in the Drosophila giant fiber system: A
structural study. J. Comp. Neurol. 404, 449–458.
33. Thomas, J.B., and Wyman, R.J. (1984). Mutations altering syn-
aptic connectivity between identified neurons in Drosophila.
J. Neurosci. 4, 530–538.
34. Allen, M.J., Shan, X., Caruccio, P., Froggett, S.J., Moffat, K.G.,
and Murphey, R.K. (1999). Targeted expression of truncated
glued disrupts giant fiber synapse formation in Drosophila.
J. Neurosci. 19, 9374–9384.
35. Guo, X., Macleod, G.T., Wellington, A., Hu, F., Panchumarthi, S.,
smaier, K.E. (2005). The GTPase dMiro Is Required for Axonal
Transport of Mitochondria to Drosophila Synapses. Neuron 47,
36. Stowers, R.S., Megeath, L.J., Gorska-Andrzejak, J., Meinertzha-
gen, I.A., and Schwarz, T.L. (2002). Axonal transport of mito-
chondria to synapses depends on milton, a novel Drosophila
protein. Neuron 36, 1063–1077.
37. Hortsch, M., Wang, Y.M., Marikar, Y., and Bieber, A.J. (1995).
The cytoplasmic domain of the Drosophila cell adhesion mole-
culeneuroglianis notessentialfor its homophilic adhesiveprop-
erties in S2 cells. J. Biol. Chem. 270, 18809–18817.
38. Islam, R., Wei, S.Y., Chiu, W.H., Hortsch, M., and Hsu, J.C.
(2003). Neuroglian activates Echinoid to antagonize the Dro-
sophila EGF receptor signaling pathway. Development 130,
39. Hortsch, M., Homer, D., Malhotra, J.D., Chang, S., Frankel, J.,
Jefford, G., and Dubreuil, R.R. (1998). Structural requirements
foroutside-inandinside-outsignaling byDrosophila neuroglian,
a member of the L1 family of cell adhesion molecules. J. Cell
Biol. 142, 251–261.
40. McGuire, S.E., Le, P.T., Osborn, A.J., Matsumoto, K., and Davis,
R.L. (2003). Spatiotemporal rescue of memory dysfunction in
Drosophila. Science 302, 1765–1768.
41. Forni, J.J., Romani, S., Doherty, P., and Tear, G. (2004). Neuro-
glian and FasciclinII can promote neurite outgrowth via the
FGF receptor Heartless. Mol. Cell. Neurosci. 26, 282–291.
42. Doherty, P., Williams, E., and Walsh, F.S. (1995). A soluble chi-
meric form of the L1 glycoprotein stimulates neurite outgrowth.
Neuron 14, 57–66.
43. Garcia-Alonso, L., Romani, S., and Jimenez, F. (2000). The EGF
and FGF receptors mediate neuroglian function to control
growth cone decisions during sensory axon guidance in Dro-
sophila. Neuron 28, 741–752.
the Na+/K+ ATPase are essential for septate junction function in
Drosophila. J. Cell Biol. 161, 979–989.
45. Jenkins, S.M., and Bennett, V. (2001). Ankyrin-G coordinates as-
sembly of the spectrin-based membrane skeleton, voltage-
gated sodium channels, and L1 CAMs at Purkinje neuron initial
segments. J. Cell Biol. 155, 739–746.
46. Malhotra, J.D., Tsiotra, P., Karagogeos, D., and Hortsch, M.
(1998). Cis-activation of L1-mediated ankyrin recruitment by
TAG-1 homophilic cell adhesion. J. Biol. Chem. 273, 33354–
47. Pielage, J., Fetter, R.D., and Davis, G.W. (2005). Presynaptic
spectrin is essential for synapse stabilization. Curr. Biol. 15,
48. Moores, C.A., Perderiset, M., Francis, F., Chelly, J., Houdusse,
A., and Milligan, R.A. (2004). Mechanism of microtubule stabili-
zation by doublecortin. Mol. Cell 14, 833–839.
49. Schmid, R.S., Pruitt, W.M., and Maness, P.F. (2000). A MAP
kinase-signaling pathway mediates neurite outgrowth on L1
and requires Src-dependent endocytosis. J. Neurosci. 20,
50. Schaefer, A.W., Kamiguchi, H., Wong, E.V., Beach, C.M., Land-
reth, G., and Lemmon, V. (1999). Activation of the MAPK signal
cascade by the neural cell adhesion molecule L1 requires L1 in-
ternalization. J. Biol. Chem. 274, 37965–37973.
51. Sweatt, J.D. (2004). Mitogen-activated protein kinases in synap-
tic plasticity and memory. Curr. Opin. Neurobiol. 14, 311–317.
52. De Angelis, E., MacFarlane, J., Du, J.S., Yeo, G., Hicks, R., Rath-
jen, F.G., Kenwrick, S., and Brummendorf, T. (1999). Pathologi-
cal missense mutations of neural cell adhesion molecule L1 af-
fect homophilic and heterophilic binding activities. EMBO J.
53. Cheng, L., and Lemmon, V. (2004). Pathological missense muta-
tions of neural cell adhesion molecule L1 affect neurite out-
growth and branching on an L1 substrate. Mol. Cell. Neurosci.
the neurofascin gene and analysis of developmentally regulated
alternative splicing. J. Biol. Chem. 272, 28742–28749.
55. Allen, M.J., Drummond, J.A., and Moffat, K.G. (1998). Develop-
ment of the giant fiber neuron of Drosophila melanogaster.
J. Comp. Neurol. 397, 519–531.
56. Jacobs, K., Todman, M.G., Allen, M.J., Davies, J.A., and Bacon,
J.P. (2000). Synaptogenesis in the giant-fibre system of Dro-
sophila: Interaction of the giant fibre and its major motorneuro-
nal target. Development 127, 5203–5212.
57. Hortsch, M., Bieber, A.J., Patel, N.H., and Goodman, C.S. (1990).
Drosophila neuroglian. Neuron 4, 697–709.
58. Murphey, R.K., Froggett, S.J., Caruccio, P., Shan-Crofts, X.,
Kitamoto, T., and Godenschwege, T.A. (2003). Targeted expres-
sion of shibire ts and semaphorin 1a reveals critical periods for
synapse formation in the giant fiber of Drosophila. Development
Neuroglian/L1-CAM in Central-Synapse Formation