Morphogenetic Studies of the Drosophila DA1 Ventral Olfactory Projection Neuron
Morphogenetic Studies of the Drosophila DA1 Ventral Olfactory Projection Neuron
Hung-Chang Shen 0 1
Jia-Yi Wei 0 1
Sao-Yu Chu 0 1
Pei-Chi Chung 0 1
Tsai-Chi Hsu 0 1
Hung-Hsiang Yu 0 1
0 Institute of Cellular and Organismic Biology, Academia Sinica , Taipei, Taiwan , 2 Graduate Institute of Life Sciences, National Defense Medical Center , Taipei , Taiwan
1 Editor: Hiromu Tanimoto, Tohoku University , JAPAN
In the Drosophila olfactory system, odorant information is sensed by olfactory sensory neurons and relayed from the primary olfactory center, the antennal lobe (AL), to higher olfactory centers via olfactory projection neurons (PNs). A major portion of the AL is constituted with dendrites of four groups of PNs, anterodorsal PNs (adPNs), lateral PNs (lPNs), lateroventral PNs (lvPNs) and ventral PNs (vPNs). Previous studies have been focused on the development and function of adPNs and lPNs, while the investigation on those of lvPNs and vPNs received less attention. Here, we study the molecular and cellular mechanisms underlying the morphogenesis of a putative male-pheromone responding vPN, the DA1 vPN. Using an intersection strategy to remove background neurons labeled within a DA1 vPNcontaining GAL4 line, we depicted morphological changes of the DA1 vPN that occurs at the pupal stage. We then conducted a pilot screen using RNA interference knock-down approach to identify cell surface molecules, including Down syndrome cell adhesion molecule 1 and Semaphorin-1a, that might play essential roles for the DA1 vPN morphogenesis. Taken together, by revealing molecular and cellular basis of the DA1 vPN morphogenesis, we should provide insights into future comprehension of how vPNs are assembled into the olfactory neural circuitry.
Data Availability Statement: All relevant data are
within the paper and its Supporting Information files.
Funding: This work was supported by Ministry of
Science and Technology, Taiwan grant#
MOST1032321-B-001-023 (https://www.most.gov.tw/?l=ch). The
funder had no role in study design, data collection
and analysis, decision to publish, or preparation of
Competing Interests: The authors have declared
that no competing interests exist.
Ensembles of neurons are linked into complex neural circuits in the nervous system in which
animals use them to process environmental information, e.g., light, sound and odors, etc., for
survival and offspring reproduction. A series of steps must occur in the formation of the neural
circuitry: stringent regulation of generation and survival of the number of neurons, cell fate
specification after the birth of neurons, accurate navigation of neuronal axons and dendrites to
their targets, appropriate patterns of axonal branches and dendritic arborizations within
neurons and correct connections among neurons for assembling into functional neural circuits [
Comprehensive identification of genes and molecules that regulate each cellular step described
above should provide insights into how the complex neural circuits are formed in the brain
throughout the animal kingdom.
In the Drosophila olfactory system, odors are detected by ~50 classes of ~1,300 olfactory
sensory neurons (OSNs) in the antennae and maxillary palps [
]. Odorant inputs are then
delivered by OSN axons to the primary olfactory center, the antennal lobe (AL), where OSN
axons make connections with neurites of ~250 projection neurons (PNs) and local
interneurons (LNs) [
]. Odorant signals are modulated by LNs, and then relayed by PNs to higher
olfactory centers, e.g., mushroom body and lateral horn (LH), for further decoding of the
olfactory information . These PNs and LNs are derived from at least five neural stem cells; the
derived neurons include three clusters of PNs (anterodorsal PNs (adPNs in the ALad1 lineage),
ventral PNs (vPNs in the ALv1 lineage), and lateroventral PNs (in the ALlv1 lineage)), a set of
ventral LNs (in the ALv2 lineage), and a lateral population of mixed PNs and LNs (in the ALl1
]. Of these four PN groups, adPNs and PNs from the ALl1 lineage (lPNs) have been
extensively explored at the molecular, cellular and functional levels, while the investigation of
those of vPNs and lvPNs received less attention [
]. Revealing the molecular and cellular
mechanisms underlying the morphogenesis of vPNs and lvPNs will advance our understanding of
how multiple populations of neurons are integrated into the functional olfactory system during
Here, we use the DA1 vPN, a neuron anatomically receiving the input from a class of
malepheromone responding OSNs (Or67d OSNs), as an entry to study the molecular and cellular
mechanisms underlying the morphogenesis of vPNs [
]. We identified R95B09-GAL4 which
labels the DA1 vPN within the AL together with background neurons outside of the AL from
the JFRC GAL4 collection . We then obtained a pure DA1 vPN labeling pattern by an
intersection strategy to remove background neurons from R95B09-GAL4, allowing us not only to
depict DA1 vPN morphogenesis but also to utilize it for the DA1 vPN phenotypic analysis in
loss-of-function study. Finally, to prove in principle that R95B09-GAL4 is a good reagent for
investigating the DA1 vPN development at the molecular level, we conducted a pilot RNA
interference (RNAi) knock-down screen and identified cell surface molecules, including Down
syndrome cell adhesion molecule 1 (Dscam1) and Semaphorin-1a (Sema-1a), that might
participate in the DA1 vPN morphogenesis [
]. Taken together, using R95B09-GAL4 we revealed
molecular and cellular basis of the DA1 vPN morphogenesis, which sets up a foundation for
future comprehensive understanding of DA1 vPN-mediated biological processes and neural
circuitry assembly in the olfactory system.
Exploration of the use of R95B09-GAL4 as a tool for studying DA1 vPN
To identify useful GAL4 lines that label vPNs, we sought to search for the Janelia GAL4
collection by viewing through imaging files in the website of Bloomington Drosophila stock center
(http://flystocks.bio.indiana.edu/Browse/gal4/gal4_Janelia.php) that portray expression
patterns of many available GAL4 lines. Fortunately, we identified R95B09- GAL4 (Bloomington
stock number (BL) 47276) which labels a single DA1 vPN (arrow and double-arrow in Fig 1A)
together with background neurons in brain regions outside of the AL (asterisks in Fig 1A) [
For this DA1 vPN, the dense dendritic innervation was clearly seen in the DA1 glomerulus at
the AL (double-arrow in Fig 1A), while some loose dendritic arborizations were also detected
in the ventral AL (arrow in the Fig 1A). In contrast, axonal elaboration pattern of the DA1 vPN
cannot be unambiguously identified in the LH due to the presence of background neurons in
the proximity of the same LH region (dash-circle for DA1 vPN axonal branches and three
2 / 12
Fig 1. Expression pattern of R95B09-GAL4 during development. (A-H) The expression pattern of
R95B09-GAL4 during development was revealed by using it to drive the expression of the mCD8::GFP
reporter (green). The genotypes of Figs 1–3 and S1 Fig are summarized in S2 Table. Brain neuropiles shown
in blue were stained with antibodies against Bruchpilot (Brp; A-C and I) or DN-cadherin (D-H). Staining
against Fasciclin II (FasII) and/or Neurotactin (Nrt) shown in magenta was performed to reveal landmarks for
comparison of the relative position of the developing DA1 vPN in D-H. Ventral and dorsal dendrites of the
DA1 vPN (arrow and double-arrow in A) were clearly observed in the AL, while the axonal pattern was
partially blocked by background neurons in the LH (three asterisks within dash-circle in A). The developing
DA1 vPN was detected after puparium formation (APF; arrows and double-arrows in D-I). Background
neurons were indicated with asterisks in the adult brain (A) and they were also seen in the brains of different
developmental stages with a transient and yet complex pattern (B-I). Two groups of background neurons
were seen from early larval to pupal stages (B-I): the first group is dorsal neurons that project their neurites
ventrally to the middle of the central brain and then bifurcate their neurites medially and laterally toward the
midline and optic lobes (arrowheads in Fig 1B–1F); the other one is a group of neurons in which their cell
bodies locate between the central brain and the ventral nerve cord and their neuritis project medially to the
midline (double-arrowheads in Fig 1B–1E). First instar larval (L1), second instar larval (L2) and white pupal
(WP). Scale bar: 10 μm for A-I, except for D as 14 μm.
asterisks in the dash-circle for background neurons in Fig 1A). We sought to identify whether
R95B09-GAL4 labels the DA1 vPN throughout morphogenesis since GAL4 lines may label
neurons in a transient and yet complex pattern over the course of development that makes
them unsuitable for examining the effects of genetic perturbation on the development of
neurons of interest (e.g., a group of dorsal neurons that project their neurites ventrally to the
middle of the central brain and bifurcate their neurites medially and laterally to the midline and
optic lobes can be seen from early larval to early pupal stages (pointed by arrowheads in Fig
1B–1F), but they gradually diminished after mid-late pupal stage (Fig 1G–1I)). We then
expressed mCD8::GFP under control of R95B09-GAL4 and examined the expression pattern
of this GAL4 from the early larval stage onwards through adulthood (Fig 1A–1I). Besides those
dorsal neurons, neuronal expression was also detected in a group of medially projected neurons
3 / 12
whose cell bodies locate between the central brain and the ventral nerve cord in R95B09-GAL4
started from the early larval stage (double-arrowheads in Fig 1B–1E). In contrast, neuronal
expression in the AL was not apparent in R95B09-GAL4 until after puparium formation (APF;
arrow in Fig 1F–1I). Interestingly, a single neuron (most likely the DA1 vPN) innervating the
AL was clearly observed in R95B09-GAL4 from 24-hour APF to the adult (Fig 1A and 1F–1I).
This result suggested that R95B09-GAL4 may be ideally utilized for perturbing gene function
and then examining the causal phenotype in the DA1 vPN.
Prior to using R95B09-GAL4 as a manipulation tool, we wondered whether the background
neurons labeled by R95B09-GAL4 could be removed through an intersection strategy, in which
flippase (FLP) was used to remove a FLP-recognition-target (FRT)>stop>FRT cassette that
blocks the expression of reporters (Fig 2A for the schematic illustration of this intersection
]. Notably, a pure DA1 vPN labeling pattern was obtained by intersecting
R95B09-GAL4 with a PN-expressing FLP line, GH146-FLP (Fig 2B), in which not only
dendritic innervations were clearly visualized in the AL (arrow and double-arrow for the ventral
AL and the DA1 glomerulus in Fig 2B) but also axonal arborizations were apparently detected
as dorsal and ventral groups of axonal branches in the LH (arrowhead and double-arrowhead
in Fig 2B) [
]. This clean DA1-vPN labeling pattern provides single cell resolution that can be
used as a tool to directly characterize the morphological progression of the developing DA1
vPN and to examine the effects of manipulating gene expression in the DA1 vPN
morphogenesis (both of them will be described below).
Depiction of DA1 vPN morphogenesis by intersecting R95B09-GAL4
Since the presence of background neurons in R95B09-GAL4 hinders a clear depiction of how
the DA1 vPN develops, we then used the above FLP-out strategy with the intersection of
GH146-FLP and R95B09-GAL4 to examine the morphogenesis of the DA1 vPN (Fig 2C–2H).
Axonal branches were initially observed to occupy at the ventral part of the developing LH
(double-arrowheads in the Fig 2C and 2D) and these ventral axonal branches became more
elaborated after 24-hour APF (double-arrowheads in Fig 2B and 2E–2H). Intriguingly,
characteristic dorsal axonal branches did not appear until 24-hour APF (Fig 2C–2E; arrowhead
pointed at dorsal axonal branches in Fig 2E) and subsequently became more prominent
(arrowheads in Fig 2B and 2F–2H). On the other hand, dendrites of the DA1 vPN were initially
seen to distribute ventrally in the developing AL within the first 24 hours APF (arrows in Fig
2C–2E). The typical dorsal dendrites of the DA1 vPN were not visible in the developing AL
until around 36-hour APF (double-arrow in Fig 2F). Dorsal dendrites of the DA1 vPN
gradually accumulated and became denser from 36-hour to 60-hour APF (double-arrows in Fig 2F–
2H), which eventually led to restriction of most of the dendrites to the DA1 glomerulus at the
adult stage (double-arrow in Fig 2B). In contrast, the ventral dendrites of the DA1 vPN were
gradually diminished during the pupal stage, becoming minimal processes by the adult stage
(arrows in Fig 2B and 2F–2H). We should note that animals homozygous for GH146-FLP,
R95B09-GAL4, and flip-out-cassette GFP reporter occasionally contained visible developing
DA1 vPN between the white pupa stage (WP) and 12-hour APF (~10% and ~30%, respectively;
n>20; Fig 2C and 2D), while the developing DA1 vPN could be detected in nearly 100% of
heterozygous animals of the same genotype from 24-hour APF to the adulthood (n>20; Fig 2B
and 2E–2H). Taken together, the above results suggested that the GAL4 of R95B09-GAL4
should be stably expressed in the DA1 vPN after 24-hour APF during DA1 vPN
morphogenesis, which makes R95B09-GAL4 an ideal tool to examine the effect of perturbing gene
expression on the development of the DA1 vPN.
4 / 12
Fig 2. Use of an intersection strategy to visualize the developing DA1 vPN. (A) A schematic illustration
of how to obtain a pure DA1 vPN expression pattern by intersecting the expression of a membrane bound
GFP reporter, myristoyl-anchored GFP (myr-GFP) reporter, with R95B09-GAL4 and GH146-FLP. All neurons
contain the genetic components of GH146-FLP, R95B09-GAL4 and UAS-FRT<stop<FRT-myr-GFP in the
genome. However, the expression of myr-GFP is initially blocked by the presence of a FRT<stop<FRT
cassette. GH146-FLP is expressed to remove this stop cassette in neurons of panels a and b but not in those
of panel c. On the other hand, R95B09-GAL4 is expressed in neurons of panel b and c but not in those of
panel a. Therefore, myr-GFP can be only expressed in neurons of panel b in which GH146-FLP and
R95B09-GAL4 are co-expressed. (B) A flattened confocal image revealed the pure DA1 vPN expression
pattern that was obtained by intersecting R95B09-GAL4 and GH146-FLP. Dense and loose dendritic
innervations in the DA1 glomerulus and the ventral portion of the AL were pointed by double-arrow and arrow,
respectively. Characteristic dorsal and ventral axonal branches were indicated by arrowhead and
doublearrowhead, respectively. (C-H) The morphogenesis of the developing DA1 vPN can be observed in confocal
images taken from the brains of Drosophila in which R95B09-GAL4 was intersected with GH146-FLP; brains
were processed at the following stages: white pupa (WP) (C), 12-hour (D), 24-hour (E), 36-hour (F), 48-hour
(G), and 60-hour (H) after puparium formation (APF). In WP, axons and dendrites of the DA1 vPN were
present at the ventral portions of the developing LH and AL, respectively. The characteristic dorsal axonal
branches of the DA1 vPN started to appear around 24-hour APF (arrowhead in E) and became increasingly
prominent afterward (arrowheads in F-H). Ventral axonal braches were also observed to elaborate after
24-hour APF (double-arrowheads in C-H). In contrast, the DA1 vPN dendrites remained at the ventrolateral
5 / 12
developing AL prior to 36-hour APF (arrows in C-H) and started to elaborate dorsally in the developing AL at
around 36-hour AFP (double-arrow in F). Gradually, the DA1 vPN dendrites accumulated at the dorsolateral
part of the developing AL, with a resulting reduction of dendrites at the ventrolateral developing AL from
36-hour to 60-hour APF (double-arrows in F-H). The morphology of the developing DA1 vPN at 60-hour APF
(H) was quite similar to that of the adult DA1 vPN (B). Brain neuropiles were stained with antibodies against
Brp (A and B) or DN-cadherin (C-H) (shown in blue). Staining against FasII and Nrt (shown in magenta) was
performed to reveal landmarks for comparison of the relative position of the developing DA1 vPN in C-H.
Scale bar: 10 μm.
Identification of cell surface molecules that participate in DA1 vPN
Since cell surface molecules have been shown to participate in axonal and dendritic
], we next sought to take RNAi knock-down approach to inhibit the expression of
cell surface molecules that may affect the morphogenesis of the DA1 vPN (Fig 3A for a
wildtype DA1 vPN). Just to prove in principle that R95B09-GAL4 can be served as an ideal tool for
identifying essential cell surface molecules in the DA1 vPN morphogenesis, we utilized
R95B09-GAL4 to drive the expression of a Dscam1-RNAi transgene, Dscam1RNAi18i, which
Fig 3. Using R95B09-GAL4 as a tool to identify Dscam1 and Sema-1a essential for the DA1 vPN
morphogenesis. Images of DA1 vPN phenotypes with gene perturbation were shown in green: wild-type
(A), RNAi knock-down of Dscam1 (B-D) and Sema-1a (E and F). Brain neuropiles were stained with an
antibody against Brp (shown in blue in A-F). Dense and loose dendritic innervations in the DA1 glomerulus
and the ventral portion of the AL were pointed by double-arrows and arrows in A-D. Dorsal and ventral
axonal branches were indicated by arrowheads and double-arrowheads in A, C, E and F. (B and D) In
Dscam1RNAi18i and Dscam1RNAi17.2i knock-down animals, ventral axons of the DA1 vPN appeared to
aggregate in the LH (yellow double-arrowheads) that might result in the failure of branching out dorsal
axons (yellow arrowheads). (C) In Dscam1RNAi17.1i knock-down animals, no axonal phenotype was
observed as compared to the Dscam1RNAi18i and Dscam1RNAi17.2i knock-down animals
(doublearrowhead and arrowhead). (E and F) Axonal and dendritic phenotypes were observed in the DA1 vPNs
expressing two independent Sema-1a RNAi transgenes, Sema-1aRNAiTRiP-V20 (BL34320) and
Sema1aRNAisy (home-made in the current study): mis-projected axons (yellow triple-arrowhead) were seen to
channel through the optic tubercle (OT) to reach the superior medial protocerebrum (SMP); dendrites
were also found to mis-target to the DA3 glomerulus (yellow triple-arrow). Scale bar: 10 μm.
6 / 12
potentially knocks down all Dscam1 isoforms by specifically targeting the common exon 18 of
Dscam1 that has previously shown to affect the axonal and dendritic development of vPNs
. Notably, a highly penetrant phenotype was observed in the axonal morphogenesis of the
Dscam1-deficient DA1 vPN, in which the axons were unable to elaborate at the LH, resulting
in crumpling up ventral axonal branches and failing to sprout out most of dorsal axonal
branches (yellow double-arrowhead and yellow arrowhead in Fig 3B; phenotypic percentage of
all RNAi knock-down results can be seen in S1A Fig). We then used two additional
Dscam1RNAi transgenic lines to confirm the finding of Dscam1RNAi18i-derived axonal defects:
Dscam1RNAi17.1i and Dscam1RNAi17.2i can specifically silence dendritic- and axonal-expressed
Dscam1 isoforms by targeting exons 17.1 and 17.2 of Dscam1, respectively [
the axonal phenotypes of crumpling up ventral axonal branches and failing to sprout out most
of dorsal axonal branches were recapitulated by the expression of Dscam1RNAi17.2i but not
Dscam1RNAi17.1i (double-arrowhead and arrowhead in Fig 3C for Dscam1RNAi17.1i
knockdown samples; yellow double- arrowhead and yellow arrowhead in Fig 3D for Dscam1RNAi17.2i
knock-down samples). We should note that no dendritic phenotype was observed by the
expression of any of three Dscam1-RNAi transgenes (double-arrows and arrows in Fig 3B–
3D). Taken together, these Dscam1-RNAi knock-down results suggested that R95B09-GAL4 is
indeed an excellent reagent for the molecular study of the DA1 vPN development.
We then performed a pilot screen using R95B09-GAL4 to express RNAi transgenes carried
by fly stocks available in our hand at the time to knock down the expression of 34 genes (all of
them are available at the Bloomington Drosophila stock center; S1 Table): we found that
lossof-function (LOF) of Sema-1a displayed a highly penetrant phenotype on the axonal and
dendritic morphogenesis of the DA1 vPN. By crossing the R95B09-GAL4 line with the
UAS-Sema1aRNAiTRiP-V20 transgenic line (which expresses Sema-1a-short hairpin RNA (shRNA) using
the Valium 20 backbone vector [
]; BL34320), we observed that their progeny exhibited both
axonal and dendritic phenotypes: (1) axons mis-projected away from the LH to the lateral edge
of the superior medial protocerebrum (SMP) via the dorsal bottle neck of the optic tubercle
(OT; yellow triple-arrowhead in Fig 3E); (2) dendrites were no longer limited to the DA1
glomerulus and mis-targeted into the DA3 glomerulus (yellow triple-arrow in Fig 3E). We tried,
but failed, to use the other Sema-1a-RNAi line, UAS-Sema-1aRNAiTRiP-V10 (which expresses
Sema-1a-long double strand RNA (dsRNA) using the Valium 10 backbone vector [
BL29554) to verify the Sema-1aRNAiTRiP-V20-derived DA1 vPN morphological defects (data
not shown and S1A Fig). However, this might not be a surprising result since it has been
reported that Valium 10-based RNAi transgenes generally exhibit lower gene knock-down
efficiency than Valium 20-based RNAi transgenes as described in the Bloomington RNAi stock
webpage (http://flystocks.bio.indiana.edu/Browse/RNAi/RNAi_all.php) [
]. To confirm the
authenticity of Sema-1aRNAiTRiP-V20-derived axonal and dendritic defects in the DA1 vPN, we
have made the third line carrying another Sema-1a-RNAi transgene, UAS-Sema-1aRNAisy,
whose targeting sequence is distinct from the UAS-Sema-1aRNAiTRiP-V20 transgene.
Intriguingly, the Sema-1aRNAisy- expressing DA1 vPN did recapitulate the phenotypes of axonal
misprojection out of the LH and dendritic mis-targeting to the DA3 glomerulus (yellow
triplearrowhead and yellow triple-arrow Fig 3F). These Sema-1a-RNAi knock-down results again
demonstrated the power of using R95B09-GAL4 as a manipulating reagent for the
morphogenetic study of the DA1 vPN.
Besides Dscam1 and Sema-1a, knock-down of derailed (drl), Derailed 2 (Drl-2), Eph receptor
tyrosine kinase (Eph), Leukocyte-antigen-related-like (Lar) and Protein tyrosine phosphatase 4E
(Ptp4E) with RNAi transgenic lines (BL39002, BL25961, BL39066, BL34956, and BL 38369,
respectively) resulted in a low penetration of dendritic phenotypes in the DA1 vPN (S1A–S1C
Fig). LOF of drl (6%) and Lar (9%) exhibited aberrant dendritic phenotypes with a slightly
7 / 12
higher percentage compared to the rest of LOF of three genes (~3%; S1A Fig). The aberrant
dendritic phenotype observed in the Lar-RNAi knock-down samples included partial dendritic
innervation in the DA1 glomerulus, dendritic mis-targeting to the DA4l glomerulus and
excessive dendritic arborizations distributed in the VL1 glomerulus and the posterior ventrolateral
AL (yellow double-arrows, yellow triple-arrows and yellow arrows, respectively, in S1B Fig).
Intriguingly, LOF of drl and Ptp4E displayed a similar but less severe Lar-deficient dendritic
defect (yellow double-arrows, yellow triple-arrows and yellow arrows, respectively, in S1C and
S1D Fig). On the other hand, samples with LOF of Drl-2 and Eph displayed dendritic shifting
phenotypes from the DA1 glomerulus to ventroposterior and anteromiddle portions of the AL,
respectively (yellow arrow, yellow double-arrows and yellow triple-arrows in S1E and S1F Fig).
Interestingly, the dendrites of Eph-deficient DA1 vPN distributed roughly in three clusters
along the anteromiddle portion of the AL, including the D, DA3 and DA4l glomeruli within
the dorsal cluster, the VA1d and DC3 glomeruli within the middle cluster and the VA5 and
VL1 glomeruli within the ventral cluster (yellow triple-arrows in S1F Fig). However, none of
above phenotypes resulting from the loss of Lar, Ptp4E, Drl-2 and Eph can be recapitulated by
other independent RNAi transgenic lines (BL40938, BL60008, BL55893 and BL60006,
respectively; S1A Fig), which kind of undermined the above finding for the roles of these genes in the
DA1 vPN morphogenesis. Nonetheless, the RNAi knock-down results on Dscam1 and
Sema1a together with the above results of the depiction of the developing DA1 vPN suggested that
R95B09-GAL4 could serve as an ideal tool for future systematically studying the molecular and
cellular basis of the DA1 vPN morphogenesis, which could lead to a better understanding of
how vPNs are assembled into the olfactory system.
In the Drosophila olfactory neural circuit, neurites of OSNs, LNs and PNs are assembled
together to form the primary olfactory center, the AL. Of the four PN groups, the development
of adPNs and lPNs has been extensively explored, while the development of lvPNs and vPNs
remains understudied [
]. In the current study, we investigate the molecular and cellular basis
of the morphogenesis of a putative male-pheromone responding vPN, the DA1 vPN. At the
cellular level, we depicted when and where stereotyped patterns of axonal branches and
dendritic arborizations are established in the DA1 vPN. At the molecular level, we identified genes
of cell surface molecules, including Dscam1 and Sema-1a, which might be essential for forming
the characteristic axonal and dendritic patterns of the DA1 vPN. By revealing the molecular
and cellular mechanisms underlying the DA1 vPN morphogenesis, we should provide insights
into how vPNs are assembled in the olfactory neural circuit.
In this study, we utilized an FLP-out intersection strategy to obtain a pure DA1 vPN labeling
pattern, which allowed us to clearly depict the formation of axonal branches and dendritic
arborizations of the DA1 vPN from early pupal stage to the adulthood. Intriguingly, both dendrites
and axons of the DA1 vPN were initially accumulated at the ventral parts of the developing AL
and LH, respectively, and its characteristic dorsal dendritic arborizations and axonal branches
were not seen until the mid-early pupal stage (Fig 2). What is the molecular mechanism
underlying the cellular processes for generating appropriate dendritic and axonal patterns of the DA1
vPN? From RNAi knock-down screen we found two molecules, Dscam1 and Sema-1a, which
might play a crucial role in the DA1 vPN morphogenesis. Axonal aggregation and axonal
misprojection defects were observed in the DA1 vPN when Dscam1 and Sema-1a were knocked
down, which were the similar axonal phenotypes that exhibited in other types of neurons (Fig
]. In contrast, a novel phenotype of mis-targeted dendrites to the DA3 glomerulus
was observed when Sema-1a was deficient in the DA1 vPN (Fig 3A and 3C). Previous reports
8 / 12
have demonstrated that Sema-1a operates as a receptor for PNs to regulate the process of initial
dendritic targeting via the interaction with secreted semaphorins, Sema-2a and Sema-2b, at the
early pupal stage [
]. This initial dendritic targeting process was mediated via two opposite
gradients of Sema-1a and Sema-2a/2b in the developing AL: the dorsolateral (DL) expression of
Sema-1a in dendrites of PNs and the ventromedial (VM) expression of Sema-2a/2b in axons of
degenerating larval OSNs [
]. The dendrites of DL1 adPNs and DA1 lPNs tended to make a
DL-to-VM shift when Sema-1a was removed from these PNs or Sema-2a/2b was deficient from
the degenerating larval OSNs [
]. In our current study, however, we found that the dendrites
of the DA1 vPN were specifically mis-targeted to the DA3 glomerulus when Sema-1a was
deficient (Fig 3A and 3C). It is interesting, although perplexing at the same time, that the same
molecule, i.e., Sema-1a, plays distinct roles in forming appropriate dendritic patterns for different
types of PNs. Through loss-of-function and rescue studies, we have demonstrated that Sema-1a
plays a novel role in the prevention of aberrant dendritic invasion specifically into the DA3
glomerulus from surrounding adPNs (D, DA4m, DA4l, DC3 and VA1d adPNs) and vPNs (DA1
and diffuse vPNs) but not lPNs (DA1 and DL3 lPNs), which is different from the previously
described function of Sema-1a for its graded expression in guiding PN initial dendritic targeting
within the AL (Shen et al., manuscript in preparation).
Unlike most of adPNs and lPNs with uni-glomerular dendritic patterns, dendrites of many
types of vPNs innervate multiple glomeruli in the AL [
]. Interestingly, animals with LOF
of Lar, drl, Drl-2, Eph and Ptp4E, despite of occurring in a low frequency, all displayed a
tendency to switch their dendritic innervation from single DA1 glomerulus into multiple
glomeruli (Fig 3A and 3D–3H). This observation raises a reasonable speculation in which the initial
ventral localization of axons and dendrites of the DA1 vPN represents a default state for the
development of vPNs and perturbation of the gene expression of cell surface molecules like
Lar, drl, Drl-2, Eph and Ptp4E could lead to dendritic arborizations of vPNs from
uni-glomerulus into multiple glomeruli in the AL. Nonetheless, future investigation is required to validate
whether the dendritic morphogenesis of the DA1 vPN and various types of vPNs might be
sculpted by various combinations of cell surface molecules. Revealing the molecular
mechanism underlying the regulation of vPN morphogenesis should promise a comprehensive
understanding of how the complex olfactory system is formed in the brain to decode odorant
information from the external world.
Materials and Methods
Making the fly stock carrying a RNAi transgene for knock-down of
Standard molecular biological techniques were used to generate UAS-Sema-1aRNAisy which
encodes microRNA (miRNA) carrying the unique Sema-1a sequence, agagcaaggatcaggaa
ataat, for knocking down the expression of Sema-1a. The design of miRNA for knock-down
of Sema-1a followed the strategy described previously [
]. The cDNA containing the miRNA
backbone with tandem repeats of Sema-1a target sequence was subsequently cloned into
pJFRC7-20XUAS-IVS-mCD8::GFP in Xho 1 and Xba 1 sites [
]. The cDNA construct of
UASSema-1aRNAisy was injected into the fly stock carrying an attP docking site (VK00033; e.g.,
BL9750) to generate the transgenic fly via the service provided by Rainbow Transgenic Flies, Inc.
The fly strains used in this study were as follows: (1) R95B09-GAL4 (BL47267); (2)
]; (3) GH146-FLP [
]; (4) UAS-FRT<stop<FRT-myr-GFP [
9 / 12
UAS-Dscam1RNAi17.1i, UAS-Dscam1RNAi17.2i and UAS-Dscam1 RNAi18i; [
UAS-Sema1aRNAisy; (9–47) UAS-RNAiTRiP stocks from the TRiP collection (see the S1 table for the list of
these UAS-RNAiTRiP stocks).
Fly brain preparation for molecular and cellular studies of the DA1 vPN
Dissection, immunostaining, and mounting of fly brains were performed as described in a
standard protocol [
]. Primary antibodies used in this study included rabbit antibody against GFP
(1:800, Invitrogen), rat monoclonal antibody against DN-cadherin (DN-Ex#8, 1:100, DSHB)
and mouse monoclonal antibodies against Bruchpilot (nc82, 1:100, DSHB), Fasciclin II (1D4,
1:50, DSHB) and Neurotactin (BP106, 1:50, DSHB). Secondary antibodies conjugated to
different fluorophores (Alexa 488, 546, and 647 (Invitrogen)) were used at a 1:800 dilution in this
study. Immunofluorescent images were collected by Zeiss LSM 700 or 780 confocal microscopy
and projected using LSM browser.
S1 Fig. A pilot RNAi screen to identify molecules potentially important for the DA1 vPN
morphogenesis. The result of the pilot RNAi knock-down screen for the DA1 vPN
morphogenesis was summarized in the panel A. Images of DA1 vPN phenotypes with gene
perturbation were shown in green: Lar (B), drl (C), Ptp4E (D), Drl-2 (E) and Eph (F). Brain neuropiles
were stained with an antibody against Brp (shown in blue in B-F). Characteristic dorsal and
ventral axonal branches were indicated by arrowheads and double-arrowheads in B-F. (A)
Percentage of DA1 vPN phenotypes in the bar graph was used to show the effect of RNAi
knockdown of 35 genes. The vertical axis indicated the RNAi lines used for knocking down 35 genes
with their Bloomington stock number (BL) or original sources and with their examined sample
sizes (n). (B-D) Similar DA1 vPN dendritic defects were observed in animals of LOF of Lar, drl
and Ptp4E: lacking a fully dendritic innervation in the DA1 glomerulus (yellow double-arrows),
mis-targeting dendrites to the DA4l glomerulus (yellow triple-arrows) and arborizing excessive
dendrites in the VL1 glomerulus and the posterior ventrolateral AL (yellow arrows). (E and F)
DA1 vPN dendritic shifting phenotypes were observed in Drl-2 and Eph RNAi knock-down
animals, in which the dendritic innervation was absent from the DA1 glomerulus (yellow
double-arrows) and found to distribute at ventroposterior and anteromiddle portions of the AL,
respectively (yellow arrows, yellow double-arrows and yellow triple-arrows). The dendrites of
Eph-deficient DA1 vPN were found as three clusters within the AL: (1) D, DA3 and DA4l
glomeruli, (2) VA1d and DC3 glomeruli and (3) VA5 and VL1 glomeruli (yellow triple-arrows).
Scale bar: 10 μm.
S1 Table. UAS-RNAiTRiP fly stocks used in the current study. The UAS-RNAiTRiP fly stocks
used in this study that were available from Bloomington Drosophila stock center were
summarized with their names of targeting genes, stock numbers (BL#), Valium vectors for
constructing RNAi transgenes, attP integration sites for generating RNAi transgenic fly stocks and the
targeting sequence information of RNAi transgenes.
S2 Table. Genotypes of flies in the experiments shown in the indicated figure panels.
Genotypes of flies in the Figs 1–3 and S1 Fig were summarized in this supplemental table.
10 / 12
We thank the TRiP at Harvard Medical School (NIH/NIGMS R01-GM084947) for providing
the transgenic RNAi fly stocks used in this study. This work was supported by the Ministry of
Science and Technology (MOST103-2321-B-001-023) and the Institute of Cellular and
Organismic Biology, Academia Sinica, Taiwan.
Conceived and designed the experiments: HCS JYW HHY. Performed the experiments: HCS
JYW SYC PCC TCH. Analyzed the data: HCS JYW SYC PCC. Wrote the paper: HHY.
11 / 12
1. Kandel ER . Principles of neural science . 5th ed. New York: McGraw-Hill ; 2013 .
2. Komiyama T , Luo L . Development of wiring specificity in the olfactory system . Curr Opin Neurobiol . 2006 ; 16 ( 1 ): 67 - 73 . doi: 10 .1016/j.conb. 2005 . 12 .002 PMID: 16377177 .
3. Yu HH , Awasaki T , Schroeder MD , Long F , Yang JS , He Y , et al. Clonal development and organization of the adult Drosophila central brain . Curr Biol . 2013 ; 23 ( 8 ): 633 - 43 . doi: 10 .1016/j.cub. 2013 . 02 .057 PMID: 23541733; PubMed Central PMCID : PMCPMC3637848 .
4. Hong W , Luo L. Genetic control of wiring specificity in the fly olfactory system . Genetics . 2014 ; 196 ( 1 ): 17 - 29 . doi: 10 .1534/genetics.113.154336 PMID: 24395823; PubMed Central PMCID : PMCPMC3872183 .
5. Couto A , Alenius M , Dickson BJ . Molecular, anatomical, and functional organization of the Drosophila olfactory system . Curr Biol . 2005 ; 15 ( 17 ): 1535 - 47 . doi: 10 .1016/j.cub. 2005 . 07 .034 PMID: 16139208 .
6. Kurtovic A , Widmer A , Dickson BJ . A single class of olfactory neurons mediates behavioural responses to a Drosophila sex pheromone . Nature . 2007 ; 446 ( 7135 ): 542 - 6 . doi: 10 .1038/nature05672 PMID: 17392786 .
7. Pfeiffer BD , Jenett A , Hammonds AS , Ngo TT , Misra S , Murphy C , et al. Tools for neuroanatomy and neurogenetics in Drosophila . Proc Natl Acad Sci U S A . 2008 ; 105 ( 28 ): 9715 - 20 . doi: 10 .1073/pnas. 0803697105 PMID: 18621688; PubMed Central PMCID : PMCPMC2447866 .
8. Ni JQ , Zhou R , Czech B , Liu LP , Holderbaum L , Yang-Zhou D , et al. A genome-scale shRNA resource for transgenic RNAi in Drosophila . Nat Methods . 2011 ; 8 ( 5 ): 405 - 7 . doi: 10 .1038/nmeth.1592 PMID: 21460824; PubMed Central PMCID : PMCPMC3489273 .
9. Wong AM , Wang JW , Axel R . Spatial representation of the glomerular map in the Drosophila protocerebrum . Cell . 2002 ; 109 ( 2 ): 229 - 41 . PMID: 12007409 .
10. Hong W , Zhu H , Potter CJ , Barsh G , Kurusu M , Zinn K , et al. Leucine-rich repeat transmembrane proteins instruct discrete dendrite targeting in an olfactory map . Nat Neurosci . 2009 ; 12 ( 12 ): 1542 - 50 . doi: 10 .1038/nn.2442 PMID: 19915565; PubMed Central PMCID : PMCPMC2826190 .
11. Tessier-Lavigne M , Goodman CS . The molecular biology of axon guidance . Science . 1996 ; 274 ( 5290 ): 1123 - 33 . PMID: 8895455 .
12. Missaire M , Hindges R . The role of cell adhesion molecules in visual circuit formation: from neurite outgrowth to maps and synaptic specificity . Dev Neurobiol . 2015 ; 75 ( 6 ): 569 - 83 . doi: 10 .1002/dneu.22267 PMID: 25649254 .
13. Shi L , Yu HH , Yang JS , Lee T. Specific Drosophila Dscam juxtamembrane variants control dendritic elaboration and axonal arborization . J Neurosci . 2007 ; 27 ( 25 ): 6723 - 8 . doi: 10 .1523/JNEUROSCI. 1517- 07 . 2007 PMID: 17581959 .
14. Zhu H , Hummel T , Clemens JC , Berdnik D , Zipursky SL , Luo L . Dendritic patterning by Dscam and synaptic partner matching in the Drosophila antennal lobe . Nat Neurosci . 2006 ; 9 ( 3 ): 349 - 55 . doi: 10 .1038/ nn1652 PMID: 16474389 .
15. Komiyama T , Sweeney LB , Schuldiner O , Garcia KC , Luo L . Graded expression of semaphorin-1a cellautonomously directs dendritic targeting of olfactory projection neurons . Cell . 2007 ; 128 ( 2 ): 399 - 410 . doi: 10 .1016/j.cell. 2006 . 12 .028 PMID: 17254975 .
16. Sweeney LB , Chou YH , Wu Z , Joo W , Komiyama T , Potter CJ , et al. Secreted semaphorins from degenerating larval ORN axons direct adult projection neuron dendrite targeting . Neuron . 2011 ; 72 ( 5 ): 734 - 47 . doi: 10 .1016/j.neuron. 2011 . 09 .026 PMID: 22153371; PubMed Central PMCID : PMCPMC3365565 .
17. Lai SL , Awasaki T , Ito K , Lee T . Clonal analysis of Drosophila antennal lobe neurons: diverse neuronal architectures in the lateral neuroblast lineage . Development . 2008 ; 135 ( 17 ): 2883 - 93 . doi: 10 .1242/ dev.024380 PMID: 18653555 .
18. Tanaka NK , Endo K , Ito K. Organization of antennal lobe-associated neurons in adult Drosophila melanogaster brain . J Comp Neurol . 2012 ; 520 ( 18 ): 4067 - 130 . doi: 10 .1002/cne.23142 PMID: 22592945 .
19. Chen CH , Huang H , Ward CM , Su JT , Schaeffer LV , Guo M , et al. A synthetic maternal-effect selfish genetic element drives population replacement in Drosophila . Science . 2007 ; 316 ( 5824 ): 597 - 600 . doi: 10 .1126/science.1138595 PMID: 17395794 .
20. Pfeiffer BD , Ngo TT , Hibbard KL , Murphy C , Jenett A , Truman JW , et al. Refinement of tools for targeted gene expression in Drosophila . Genetics . 2010 ; 186 ( 2 ): 735 - 55 . doi: 10 .1534/genetics.110.119917 PMID: 20697123; PubMed Central PMCID : PMCPMC2942869 .
21. Lee T , Luo L . Mosaic analysis with a repressible cell marker for studies of gene function in neuronal morphogenesis . Neuron . 1999 ; 22 ( 3 ): 451 - 61 . PMID: 10197526 .