Cell, Vol. 116, 869–881, March 19, 2004, Copyright 2004 by Cell Press
Synaptic Specificity Is Generated
by the Synaptic Guidepost Protein SYG-2
and Its Receptor, SYG-1
mation are thought to be encoded by specific molecular
cues, but the nature of these cues is largely unknown.
by activity-dependent synaptic refinement (Katz and
Like vertebrate synapses, synapses in the nematode
Caenorhabditis elegans are reliably formed between
specific cell types at specific subcellular regions (White
et al., 1976, 1986). In the nervous system of the adult
hermaphrodite, 302 neurons form about 5000 reproduc-
ible chemical synapses, 2000 neuromuscular junctions,
and 600 electrical synapses. A simple array of synapses
associated with egg laying develops in the late larval
stages of the hermaphrodite. The two HSN egg-laying
motor neurons, HSNL and HSNR, synapse onto vulval
muscle cells and onto the VC neurons VC4 and VC5;
VC4 and VC5 also synapse onto the vulval muscle cells.
The HSN and VC processes are in direct contact for
about 20 ?m, but synapses between these cells form
only at regions immediately adjacent to the vulva (White
et al., 1986). Synaptic guidepost cells in the vulval epi-
thelium initiate the formation of these synapses, driving
the accumulation of presynaptic proteins in the HSNL
neuron to the site of contact between the epithelial
guidepost cells and HSNL (Shen and Bargmann, 2003).
HSNL presynaptic differentiation appears normal as
long as the guidepost cell is present, even in the com-
plete absence of the normal postsynaptic targets, the
vulval muscles, and VC neurons. In the absence of the
guidepost cells, synaptic vesicle markers in HSNL fail
to accumulate at normal synaptic locations, and instead
form ectopic aggregates in anterior locations. Duplica-
tion of guidepost cells leads to the appearance of dupli-
cate synapses in HSNL.
Animals mutant for the syg-1 gene have HSNL synap-
tic defectssimilar to thoseobserved afterguidepost cell
(Shen and Bargmann, 2003). syg-1 encodes a trans-
membrane immunoglobulin superfamily protein that
acts in the presynaptic HSN axon, where it localizes to
synaptic sites prior to synapse formation in response
to the guidepost cell. These results suggest that SYG-1
might be the HSNL receptor for the vulval epithelial-
derived guidepost signal. SYG-1 is homologous to the
Drosophila proteins IrreC/roughest and Kirre/DUF, as
well as three mammalian proteins (Shen and Bargmann,
2003). During fly muscle development, Kirre/DUF, which
another immunoglobulin superfamily protein that is ex-
pressed in the muscle fusion-competent cells. Kirre/
DUF binding to SNS mediates the recognition between
founder cells and fusion-competent cells that initiates
muscle fusion (Dworak and Sink, 2002). In human and
mouse kidney development, a SYG-1-like protein,
NEPH1, and an SNS-like protein, nephrin, are required
for the formation of the podocyte slit membrane, a tight
intercellular adhesion within the kidney glomerulus (Do-
noviel et al., 2001; Tryggvason, 1999).
Here, we describe syg-2, a gene that appears to en-
code the guidepost signal in the vulval epithelial cells.
Kang Shen,1,2Richard D. Fetter,1
and Cornelia I. Bargmann1,*
1Howard Hughes Medical Institute
Department of Anatomy and
Department of Biochemistry and Biophysics
The University of California, San Francisco
San Francisco, California 94143
Synaptic connections in the nervous system are di-
rected onto specific cellular and subcellular targets.
Synaptic guidepost cells in the C. elegans vulval epi-
thelium drive synapses from the HSNL motor neuron
onto adjacent target neurons and muscles. Here, we
show that the transmembrane immunoglobulin super-
family protein SYG-2 is a central component of the
synaptic guidepost signal. SYG-2 is expressed tran-
formation. SYG-2 binds SYG-1, the receptor on HSNL,
and directs SYG-1 accumulation and synapse forma-
tion to adjacent regions of HSNL. syg-1 and syg-2
mutants havedefects in synaptic specificity;the HSNL
neuron forms fewer synapses onto its normal targets
and forms ectopic synapses onto inappropriate tar-
gets. Misexpression of SYG-2 in secondary epithelial
cells causes aberrant accumulation of SYG-1 and syn-
ing cells. Our results indicate that local interactions
termine specificity during synapse formation.
The process of synapse formation is characterized by
to their target regions by long-range and short-range
guidance cues, they are confronted with a variety of
potential target cells and target cell types. In both verte-
made by a neuron far exceeds its synaptic connections,
suggesting that neurons distinguish among potential
2001). For example, in the cat visual cortex, GABAergic
basket cell interneurons and chandelier cell interneu-
rons form synapses onto glutamatergic pyramidal cells,
but the different classes of GABAergic neurons do not
synapse onto each other (Somogyi et al., 1998). The
basket cells contact both the dendrites and the cell
bodies of pyramidal neurons, but only form synapses
onto the cell bodies, whereas the chandelier cells only
synapse onto the axon initial segment of the pyramidal
neurons (Somogyi et al., 1998). The selection of the
target cell type and subcellular region for synapse for-
2Present address: Department of Biological Sciences, Stanford Uni-
versity, Stanford, California 94305.
the absence of a signal from the vulval epithelial cells,
tions (Figures 1I and 1J).
From a direct visual screen for mutants with altered
SNB-1::YFP localization in HSNL, we identified two al-
Procedures). Like animals lacking the vulval epithelial
of syg-2(ky671) and syg-2(ky673) mutants had anterior
ectopic SNB-1::YFP vesicle clusters in HSNL. The ec-
topic SNB-1 clusters in syg-2 mutant animals appeared
during the L3 stage, when HSNL synapses normally
form, and persisted in adults. In addition to these ante-
rior clusters, 58% of syg-2(ky671) mutants had a re-
duced accumulation of SNB-1::YFP at the normal loca-
tion near the vulva. The vulval epithelium of syg-2
mutants was examined by differential interference con-
trast (DIC) microscopy at several developmental stages.
A normal complement of primary and secondary vulval
cells was induced during vulval development (3.0 ? 0.2
vulval precursor cells induced in syg-2(ky671), n ? 10).
Vulval cell divisions (scored in the L3 stage) and subse-
quent invagination during morphogenesis appeared to
be normal by light microscopy (data not shown). These
phenotypes suggest that syg-2 mutants have a defect
in signaling between the guidepost cells in the vulval
epithelium and the HSNL neuron.
Several observations suggest that the ectopic aggre-
gates of SNB-1::YFP in syg-2 mutants represent dis-
placed clusters of synaptic vesicles. A second marker
for HSNL synapses, the vesicular monoamine trans-
ters, suggesting a general defect in the accumulation
of synaptic vesicle proteins (Nurrish et al., 1999; data
not shown). In addition, in unc-104; syg-2(ky671) and
absent from both the normal and the ectopic locations
(data not shown), suggesting that the ectopic clusters,
like normal synaptic vesicles, are transported by the
synaptic vesicle kinesin UNC-104 (Hall and Hedge-
The HSNL axon in syg-2(ky673) mutants exhibited an
axon defasciculation defect. In wild-type animals, the
HSNL axon initially lies along the ventral midline in mid
L3 stage, but an internal section of HSNL axon migrates
away from the ventral midline in the late L3 stage and
shifts laterally and dorsally in association with the vulval
epithelial cells as they undergo morphogenesis (Figures
4B? and 4F?). In syg-2(ky673), this lateral/dorsal shift
failed to occur and the axon remained near the midline
(Figure 5C). The defasciculation defect is not likely to
be the cause of the synaptic phenotype, since abnormal
synaptic vesicle accumulation was apparent at the L3
stage before HSNL axon detachment. syg-2(ky671) ani-
mals did not exhibit the defasciculation defect (Figure
5E); it is possible that this defect results from the com-
2(ky671) animals were viable, fertile, egg-laying proficient,
and coordinated. syg-2(ky673) animals were viable, fer-
tile but egg-laying defective, sluggish, and exhibited
uncoordinated kinking movement.
Figure 1. syg-2 Mutants Exhibit Synaptic Vesicle Defects Similar to
Those of syg-1 and Vulval Mutants
Paired fluorescence (left) and differential interference contrast (DIC)
(right) images of the vulval region. HSNL::SNB-1::YFP fluorescence
from kyIs235 is shown; arrows indicate synaptic vesicles, asterisks
indicate the vulva. All images are lateral views with anterior to the
left and ventral down. Scale bar is equal to 5?m.
(A and B) Wild-type L4 kyIs235 animal showing normal synaptic
vesicles positioned adjacent to the developing vulva.
(C and D) L4 syg-1(ky652); kyIs235 animal.
(E and F) L4 syg-2(ky673); kyIs235 animal.
(G and H) L4 syg-2(ky671); kyIs235 animal.
(I and J) L4 lin-11(n566); kyIs235 animal.
syg-2 mutants lack synapses at the normal location and
instead form synapses onto inappropriate target cells at
is expressed in the guidepost cells and exhibits hetero-
philic binding to SYG-1. Ourresults suggest that a SYG-1/
SYG-2 interaction stimulates the formation of appropriate
synapses while suppressing inappropriate synapses,
Mutations in syg-2 Disrupt Synaptic Vesicle
Clustering in HSNL
Synapses in the HSNL neuron were labeled with the
tagged synaptic vesicle protein SNB-1::YFP under the
control of the unc-86 promoter (Nonet, 1999; Shen and
Bargmann, 2003). SNB-1::YFP accumulates at either
side of the vulva in wild-type animals, at the sites of
the HSNL synapses defined by serial section electron
microscopy (White et al., 1986, Figures 1A and 1B). In
syg-1 and syg-2 Are Synaptic Specificity Mutants
synaptic markers, the details of synaptic structure are
SYG-2, a Synaptic Guidepost Signal
not visible by light microscopy. To further characterize
the HSNL synapses in syg-1 and syg-2 mutants, we
examined a 150 ?m region centered on the vulva of a
wild-type L4 animal, a syg-1 L4 animal, and three syg-2
L4animals (twoky673, oneky671) byserial sectionelec-
tron microscopy. HSNL neurons were identified based
on their unique ventrolateral cell body position, and
HSNL axons were traced from the cell body through
active zone structures were counted in each serial sec-
tion of each HSNL neuron (Figure 2). Postsynaptic tar-
gets were assigned by identifying the cell or cells
adjacent to the active zone, as C. elegans does not
have well-developed postsynaptic densities (White et
In the wild-type animal, clusters of HSNL synaptic
vesicles were observed in association with electron-
dense active zones, as seen previously (White et al.,
1986). A total of 20 active zones were found in HSNL
in the reconstructed region, of which 18 were located
between ?10 to 0 ?m. 11/18 active zones in this central
region were directed solely toward vulval muscles and
VC neurons. Seven synapses were oriented toward vul-
val epithelial cells and body wall muscles as potential
targets; three of these also included VC neurons and
vulval muscles as potential targets. In more anterior
regions, the wild-type HSNL neuron formed two syn-
apsesonto vulvalmuscle andthe VC4and VC5neurons.
Synapses of HSNL onto vulval epithelial cells and body
wall muscles were not observed by White et al. (1986),
perhaps because they sectioned older animals (see Dis-
active zone structures, the typical structure of a C. ele-
gans synapse (Figures 2B, 2D, and 2F). However, the
distribution of these synapses was different from those
of the wild-type animal (Figures 2A, 2C, 2E, and 2G).
observed atlocations well anteriorof the vulva(5–30 ?m
distant). Second, the accumulation of synaptic vesicles
and active zones at the vulva was less robust than that
observed in the wild-type animal. A smaller absolute
number of active zones (5 in syg-1(ky652), 0 in syg-
2(ky673), and 0 in syg-2(ky671)) as well as a smaller
percentage of the total number of the active zones were
within the 10 ?m region centered on the vulva. The total
number of active zones was reduced to a lesser extent
results match the mutant phenotypes observed with
SNB-1::YFP and other synaptic markers, and demon-
strate that the anterior clusters of synaptic markers in
the mutants correspond to morphological synapses.
Presynaptic densities (active zones) were similar in
size in wild-type, syg-1, and syg-2 animals, but the aver-
age number of vesicles per synapse was slightly re-
duced in the mutants (51%–68% of wild-type average,
Table 1). The spatial distribution of synaptic vesicles
within 180 nm of the membrane of presynaptic densities
was examined in wild-type and mutant animals (Figure
2H). A substantial fraction of synaptic vesicles (6% in
N2, 11%–17% in syg-1 and syg-2 mutants) were within
30 nm (one synaptic vesicle diameter) of the plasma
membraneof thepresynaptic density.Of thesevesicles,
47% in N2, 33% in syg-1(ky652), 41% in syg-2(ky671),
and 41% in syg-2(ky673) appeared to be in direct con-
tact with the plasma membrane or morphologically
(Reist et al., 1998).
mal in their choice of synaptic targets. Most of the ante-
rior synapses in mutants had vulval epithelial cells or
synapses were also made upon body epidermal (hypo-
dermal) cells (Figure 2). Even at the vulval region where
made onto inappropriate cell types. These results indi-
cate that the HSNL neurons in syg-1 and syg-2 mutants
attempt to form synapses, but fail to recognize their
appropriate synaptic targets, resulting in defects in syn-
SYG-2 Encodes an Immunoglobulin
The syg-2 gene was identified as the predicted coding
region C26G2.1 by genetic mapping and transformation
rescue of the synaptic phenotypes of syg-2(ky673) mu-
tants (Figure 3A and Experimental Procedures). cDNA
clones of syg-2 were obtained from the C. elegans EST
project (a generous gift from Dr. Yuji Kohara) or isolated
by RT-PCR (see Experimental Procedures). syg-2 en-
codes a protein in the immunoglobulin superfamily that
is predicted to contain a signal sequence, an extra-
cellular domain with seven immunoglobulin domains
and one fibronectin type III domain, a transmembrane
domain and a cytosolic domain that ends with consen-
sus binding sequence for type II PDZ domains (Figure
3B). SYG-2 shares its overall domain structure and sub-
stantial sequence similarity with the Drosophila protein
Sticks and Stones (SNS) and Hibris, as well as the verte-
brate nephrin protein (Figure 3C).
To confirm that this open reading frame represents
the syg-2 gene, we identified the molecular lesions in
the syg-2 alleles. syg-2(ky671) is associated with a G to
A mutation at the SL1 splice acceptor site at the first
exon of the syg-2 gene (Figure 3A). RT-PCR analysis
demonstrated that the mutation results in missplicing
of the SL1 splice leader to the sequence CAG at nucleo-
tide 122 of syg-2, downstream of the initiator ATG and
the signal sequence. The closest ATG to this alternative
splice site is in a different reading frame from the start
codon of the wild-type protein. syg-2(ky671) is likely to
represent a loss-of-function mutation in the syg-2 gene.
rrf-3 animals in which RNAi of C26C2.1 was induced by
feeding with double stranded RNA-expressing bacteria
n ? 200), consistent with the interpretation that syg-2
corresponds to C26G2.1.
syg-2(ky673) is associated with a 32 kb deletion that
begins at nucleotide 6027 of cosmid C26G2 and ends
at nucleotide 27146 of cosmid F40E10. The region cov-
frames, syg-2(C26G2.1), C26G2.2, slt-1(F40E10.4), and
calsequestrin (F40E10.3). The synaptic defects in syg-
2(ky673) were entirely rescued by PCR fragments that
contained only C26G2.1, and were not rescued by a
regions of slt-1 and calsequestrin. Moreover, null muta-
Figure 2. HSNL Forms Synapses onto Abnormal Target Cells in syg-1 and syg-2 Mutants
Serial electron microscopic reconstruction of HSNL axons near the vulval region. Each plot on the left (A, C, E, and G) represents reconstruction
of one L4 animal of the given genotype. The X axis labels the position of the EM section. Minus value is posterior, plus value is toward anterior.
0 ?m marks the anterior boundary of the developing vulva; the vulval region extends from about ?10 ?m to 0 ?m in each animal. Blue
diamonds on the Y axis represent the number of synaptic vesicles found in HSNL in each EM section. Each triangle represents an active
zone. The color of the triangle indicates the apparent postsynaptic targets of that active zone, including all potential corecipients when the
active zone is oriented toward several cells. Images on the right are representative images of active zones in different genotypes. Scale bar
equals 200 nm in (B), 150 nm in (D) and (F). Asterisks in (B), (D), and (F) identify the HSNL process.
(A) HSNL EM profile of a mid-L4 wild-type N2 animal. Most synaptic vesicles are found within the boundary of the vulva (?10 ?m ?0 ?m).
Active zones are exclusively localized to the same region. Most synapses are formed onto VCs and vm2 vulval muscles.
SYG-2, a Synaptic Guidepost Signal
Table 1. Summary of Synapse Morphometry by EM Analysis
within Avg. PSD
Vesicles per Synapse
27.8 ? 2.3 (n ? 10)
14.3 ? 2.0 (n ? 10)
15.3 ? 2.5 (n ? 12)
18.8 ? 1.9 (n ? 17)
66.6 ? 3.1
72.0 ? 3.5
70.9 ? 4.1
78.1 ? 3.3
46-88 (n ? 16)
57-95 (n ? 10)
46-99 (n ? 11)
54-97 (n ? 15)
PSD, presynaptic density. Size was measured as length of the cell membrane underneath the PSD. All averages are mean ? SEM.
aTotal number of vesicles in membrane contact, or vesicles within 30 nm, are given in parentheses following the average number per synapse.
tions in the slt-1 gene (Hao et al., 2001) did not cause
HSNL synaptic defects (data not shown). Thus, the mu-
tant phenotypes in syg-2(ky673) are caused at least in
part by the deletion of C26C2.1. Even the more severe
phenotypes of syg-2(ky673) that were not shared by
syg-2(ky671), including its HSNL axon defasciculation
phenotype, its egg-laying defect, and its uncoordina-
tion, were rescued by C26C2.1. It is possible that some
of these severe defects represent synthetic phenotypes
of several genes in the deletion, including syg-2, or that
they represent the null phenotype of syg-2.
the primary vulval epithelial cells leads to clustering of
SYG-1 and synaptic vesicles. In the mid-L3 stage, the
growth cone of HSNL has just migrated past the center
of the developing vulva (Figures 4C and 4D). On the
HSNL axon, SYG-1::GFP becomes clustered adjacent
to the developing vulva at the same time in L3 that SYG-
2::GFP expression began (Shen and Bargmann, 2003;
Figures 4C, 4D, 4G, and 4H). SNB-1::YFP, the synaptic
vesicle marker, clusters at the same location several
hours later (Shen and Bargmann, 2003). The later disap-
pearance of SYG-2::GFP correlates with a more diffuse
distribution of SYG-1::GFP in adult animals (Figures 4K
and 4L), which suggests that syg-2 might be required
for the early specification of synapses, but not required
for the maintenance of synaptic function.
During embryogenesis, SYG-2::GFP is expressed in
(data not shown). In the L1 and L2 larval stages, when
many motor neuron synapses develop, SYG-2::GFP is
expressed in head and body wall muscles and in punc-
tate structures near the ventral nerve cord that may
correspond to body wall muscle arms. We have not
determined whether syg-2 affects synapse formation in
other cell types.
syg-2 Is Expressed in the Primary Vulval Epithelial
Cells during HSNL Synapse Formation
In situ hybridization of C26G2.1 completed as part of
NEXTDB, the Nematode Expression Pattern Database,
revealed expression of C26G2.1 mRNA exclusively near
the vulval region from L3 to adult stages (http://
in embryos. The expression of syg-2 was examined fur-
ther by generating a reporter fusion in which the GFP
coding region was fused to a 16 kb genomic clone with
8 kb of upstream sequence as well as the entire coding
region of syg-2. This fusion gene rescued the HSNL
In the vulval region, SYG-2::GFP was expressed in
beginning at the mid to late L3 stage (Figures 4A and
4B). The expression of SYG-2::GFP in the primary vulval
epithelial cells increased as animals enter L4 stage (Fig-
ures 4E and 4F), then disappeared within the first 6 hr
matched the expression of syg-2 mRNA observed by
in situ hybridization. The dynamic expression of SYG-
2::GFP corresponds to the time at which the signal from
Expression of syg-2 in Vulval Epithelial Cells Is
Necessary to Localize SYG-1
A signal from the primary vulval epithelial cells is suffi-
cient to localize SYG-1 protein, SNB-1 clusters, and
other synaptic markers to adjacent regions of the HSNL
axon (Shen and Bargmann, 2003). To ask if the SYG-2
protein corresponds to this vulval signal, we examined
the synaptic localization of SYG-1 in syg-2 mutants. In
wild-type animals, SYG-1 was exclusively localized to
the segment of the axon contacting the primary vulval
epithelial cells (Figures 5A and 5B). In syg-2(ky671) and
(B) Electron micrograph of an HSNL synapse formed onto a VC process in the N2 animal shown in (A). The arrow points to the electron-dense
presynaptic active zone.
(C) HSNL EM profile of a mid-L4 syg-1(ky652) animal. Many synaptic vesicles are found anterior to the boundary of the vulva. Active zones
are also found anterior to the vulva. The majority of the synapses are formed onto body wall muscles and hypodermal cells.
(D) Electron micrograph of an HSNL synapse formed onto body wall muscle in the syg-1(ky652) animal shown in (C). The arrow points to the
electron-dense presynaptic active zone.
(E) HSNL EM profile of a mid-L4 syg-2(ky673) animal. Synaptic vesicles and active zones are exclusively found anterior to the boundary of
the vulva. Most synapses are formed onto body wall muscles and hypodermal cells. A second animal of this genotype had similar phenotypes
(data not shown).
(F) Electron micrograph of an HSNL synapse formed onto body wall muscle in the syg-2(ky673) animal shown in (C). The arrow points to the
electron-dense presynaptic active zone.
(G) HSNL EM profile of a mid-L4 syg-2(ky671) animal. Synaptic vesicles and active zones are exclusively found anterior to the boundary of
the vulva. Most synapses are formed onto body wall muscles and hypodermal cells.
(H) Distribution of synaptic vesicles within 180 nm of presynaptic active zone membrane in wild-type, syg-1, and syg-2 animals.
Figure 3. syg-2 Encodes a Transmembrane Immunoglobulin Superfamily Protein
(A) syg-2 maps to C26G2.1 on the right arm of the X chromosome. syg-2(ky671) is associated with a G to A mutation at the SL1 splice acceptor
site of C26G2.1. syg-2(ky673) is a deletion of ?32 kb that affects four predicted open reading frames, C26G2.1, C26G2.2, F40E10.4, and
F40E10.3. A subclone that only contains the genomic region of C26G2.1 fully rescued the synaptic phenotype of syg-2(ky673) and syg-2(ky671)
[see Experimental Procedures]
(B) Predicted domain structure of SYG-2. Seven Ig domains and a fibronectin type III repeat were identified by the SMART program http://
(C) Phylogenetic analysis of syg-2. SYG-2 is closely related to SNS and hibris in Drosophila and human nephrin, and more distantly related
to SYG-1 in C. elegans as well as NEPH1, RST/IrreC, and Kirre/DUF.
(D) Amino acid alignment of SYG-2, Drosophila SNS, and human nephrin. Dark underlines denote Ig domains and dashed underline denotes
fibronectin type III domains.
syg-2(ky673) mutants, SYG-1 was diffusely distributed
on the HSNL axon, suggesting that SYG-2 is essential
for SYG-1 localization at synaptic sites (Figures 5C–5F).
This pattern of SYG-1 mislocalization is identical to the
are absent (Shen and Bargmann, 2003).
The LIM homeodomain mutant lin-11(n566) is associ-
ated with cell lineage defects in the secondary vulval
cell lineage as well as unknown defects in the primary
vulval cell lineage. Primary vulval cells are generated in
this mutant but they fail to function normally (Ferguson
et al., 1987; Freyd et al., 1990; Gupta et al., 2003). lin-
11 mutants have defects in synaptic vesicle localization
(Figures 1I and 1J). In lin-11 mutants, SYG-2::GFP ex-
pression was absent from the primary vulval epithelial
SYG-2, a Synaptic Guidepost Signal
Figure 4. syg-2 Is Expressed in the Primary Vulval Epithelial Cells during L3 and L4 Stages
SYG-2::GFP is a biologically active GFP fusion at the C terminus of a 16 kb genomic syg-2 clone. Images are taken from animals carrying
the extrachromosomal array kyEx684 (A, B, E, F, I, J, M, and N) or animals of equivalent age expressing SYG-1::GFP (kyIs288) (C, D, G, H, K,
(A and B) Fluorescence and DIC images of a mid-L3 stage kyEx684 animal. Arrows point to the four primary vulval epithelial cells.
(C and D) SYG-1 GFP and DIC image of an animal at the same stage as (A) and (B).
(E and F) Fluorescence and DIC images of a mid-L4 stage kyEx684 animal. Arrows point to the four primary vulval epithelial cells on the left
side of the animal. Expression of GFP in the primary vulval epithelial cells is higher than in (A). Low expression can be seen in body wall muscles.
(G and H). SYG-1 GFP and DIC image of an animal at the same stage as (E) and (F).
(I and J) Fluorescence and DIC images of a young adult kyEx684 animal. Note absence of expression in the vulval cells, and faint fluorescence
in vulval muscles.
(K and L). SYG-1 GFP and DIC image of an animal at the same stage as (I) and (J). Arrows in (K) point to diffusely localized SYG-1 on HSNL
axon outside of synaptic regions.
(M and N) Fluorescence and DIC images of a mid-L4 stage lin-11(n566); kyEx684 animal. Note absence of expression in the primary vulval
epithelial cells. Low expression of GFP can be seen in body wall muscles, similar to that of the wild-type animal in (E).
cells, whereas SYG-2 expression was maintained in
other regions such as body wall muscle (Figures 4M and
4N). This result suggests that syg-2 may represent a
target for lin-11-regulated gene expression within the
primary vulval cells. SYG-1 was diffusely localized on
the HSNL axon in lin-11(n566) mutant animals (Figures
5G and 5H), providing further evidence that the expres-
sion of SYG-2 in the primary vulval epithelial cells corre-
lates with the clustering of SYG-1 and synaptic vesicles
in the adjacent HSNL axon.
Expression of syg-2 in Secondary Vulval
Epithelial Cells Relocalizes SYG-1 Protein
and SNB-1-Labeled Vesicles
To ask whether SYG-2 directly specifies the location of
SYG-1 and synaptic vesicles, we expressed SYG-2 in
vulval cells (Figures 6D, 6E, and 6F). These results sug-
gest that SYG-2 expression inthe secondary vulval cells
is sufficient to attract SYG-1 protein from the HSNL
SYG-2 expression was also sufficient to cluster SNB-
1-expressing synaptic vesicles along secondary vulval
cells. In wild-type L4 animals, SNB-1 vesicles in HSNL
were clustered at the same segment of the HSNL axon
as SYG-1 (Figures 6G, 6H, and 6I). In syg-2(ky673);
Ex(egl-17::syg-2) animals, synaptic vesicles were clus-
tered at the segment of the HSNL axon that contacted
secondary vulval cells (Figures 6J, 6K, and 6L). This
result suggests that the interaction between SYG-1 and
These genetic experiments strongly suggest that
SYG-1 and SYG-2 interact with each other in vivo, and
that this interaction play an essential role in determining
the specificity of HSNL synaptic connections. To ask if
SYG-1 and SYG-2 could interact directly, we performed
Drosophila S2 cell aggregation assays with SYG-1 and
SYG-2. S2 cells were cotransfected with a SYG-1 cDNA
and GFP, or a SYG-2 cDNA and dsRED, and transfected
fected cells aggregated with themselves, suggesting
fected cells specifically aggregated with SYG-2-trans-
fected cells, suggesting that SYG-1 has a direct hetero-
philic interaction with SYG-2 (Figure 7C). Aggregation
was complete within 60 min of incubation (Figure 7D).
Quantitative analysis confirmed that SYG-1 transfected
cells only aggregate with SYG-2 transfected cells (Fig-
Figure 5. syg-2 Expression in the Primary Vulval Epithelial Cells Is
Essential for the Synaptic Localization of SYG-1
(A and B) Fluorescence and overlaid DIC images of a mid-L4 stage
kyIs288 animal. SYG-1 is localized to a restricted segment of HSNL
axon within the boundary of the developing vulva. kyIs288 ? unc-
(C and D) Fluorescence and overlaid DIC images of a mid-L4 stage
syg-2(ky673); kyIs288 animal. SYG-1 is diffusely localized across
the axon of HSNL. HSNL also fails to defasciculate dorsally near
the vulval region in this mutant.
(E and F) Fluorescence and overlaid DIC images of a mid-L4 syg-
2(ky671); kyIs288 animal. SYG-1 is diffusely localized across the
axon of HSNL. HSNL shows normal dorsal defasciculation near the
(G and H) Fluorescence and overlaid DIC images of a mid-L4 lin-
11(n566); kyIs288 animal. SYG-1 is diffusely localized across the
axon of HSNL. HSNR is also visible in this animal due to midline
crossover defect in lin-11 mutants. HSNL also fails to defasciculate
dorsally near the vulval region.
Asterisks mark the center of the developing vulva.
Synaptic connections form between particular classes
of neurons at particular subcellular locations. Molecu-
lar cues trigger the initial formation of the connections;
or eliminate these connections. Synapses made by the
C. elegans HSNL neuron are initiated by the guidepost
role of the primary vulval epithelial cells, acting through
superfamily protein SYG-2 as the molecular guidepost
to specify the location of synapses, probably through a
direct binding interaction. In addition, they drive the
selection of the appropriate postsynaptic targets and
exclusion of inappropriate targets by HSNL, thereby
contributing to synaptic specificity.
secondary vulval epithelial cells using an egl-17 pro-
moter in syg-2(ky673) mutants (Burdine et al., 1998).
Secondary vulval cells have a more ventral and lateral
position than primary vulval cells, which directly flank
the developing vulva (Figure 6C).
In wild-type early L4 stage animals, the primary vulval
epithelial cells expressed SYG-2::GFP, the HSN axon
contacted the ventral edge of the two most ventral pri-
mary vulval epithelial cells, and SYG-1 in HSNL was
localized to the contact sites between the HSNL axon
and the primary vulval cells (Figures 6A, 6B, and 6C). In
syg-2 mutants, SYG-1 was evenly distributed on the
HSNL axon (Figures 5C–5F). In syg-2(ky673); Ex(egl-
17::syg-2) animals, syg-2 should be expressed in the
secondary vulval cells but not the primary vulval cells.
In these animals, SYG-1 in HSNL was localized to the
contact sites between HSNL and the ventral secondary
syg-1 and syg-2 Are Synaptic Specificity Mutants
Both syg-1 and syg-2 mutants show a displacement of
synapses from the normal location near the vulva to
more anterior locations. The displaced synapses are
apparently normal, with typical active zones and vesicle
clusters at the electron microscope level. However, the
synapses in syg-1 and syg-2 mutants are made onto
abnormal postsynaptic cells, including body wall mus-
SYG-2, a Synaptic Guidepost Signal
Figure 6. Misexpression of SYG-2 in Secondary Vulval Epithelial Cells Is Sufficient to Target SYG-1 and Synaptic Vesicle Clusters to Adjacent
(A, B, and C) Fluorescence, overlaid DIC images, and schematic drawing of a mid-L4 kyIs288 animal. SYG-1 is localized to the axon region
that contacts the primary vulval epithelial cells. kyIs288 ? unc-86::syg-1::gfp
(D, E, and F) Fluorescence, overlaid DIC images, and schematic drawing of a mid-L4 syg-2(ky673); kyEx672; kyIs288 animal. SYG-1 is localized
to the axon region that contacts the secondary vulval epithelial cells. kyEx672 ? egl-17::syg-2. Diffuse fluorescence represents the secondary
cells, which express GFP faintly in this strain.
(G, H, and I) Fluorescence, overlaid DIC images, and schematic drawing of a mid-L4 kyIs235 animal. Synaptic vesicles are localized to the
axon region that contacts the primary vulval epithelial cells. kyIs235 ? unc-86::snb-1::yfp.
(J, K, and L) Fluorescence, overlaid DIC images and schematic drawing of a mid L4 syg-2(ky673); kyEx672; kyIs235 animal. Synaptic vesicles
are localized to the axon region that contacts the secondary vulval epithelial cells. kyEx672 ? egl-17::syg-2. Diffuse fluorescence represents
the secondary cells, which express GFP faintly in this strain.
cles and hypodermal cells, which normally do not re-
ceive extensive synaptic input from HSNL. Even in re-
gions near the vulva where the normal HSNL targets
are present, those targets are no longer preferentially
are necessary for HSNL to distinguish its normal synap-
tic targets from abnormal synaptic targets.
We do not know whether the synapses made onto
inappropriate targets are active. The HSNL releases
serotonin and acetylcholine neurotransmitters (Wein-
shenker et al., 1995, Desai et al., 1988), and body wall
muscles have receptors for acetylcholine, so communi-
cation could take place at these connections if the vesi-
cles are released at the ectopic active zones.
onto vulval epithelial cells in the posterior region of the
muscles. Although they represented a small fraction of
the total synapses in wild-type, the existence of syn-
post cells may serve as transient targets for HSNL syn-
in the adult stage may promote the transfer of synapses
from the guidepost cells to the mature postsynaptic
gous to subplate neurons, which serve as transient tar-
of the cortical targets (Allendoerfer and Shatz, 1994).
SYG-2, an Immunoglobulin Superfamily Protein,
Belongs to a Family of Homophilic
and Heterophilic IgSF Proteins
syg-2 encodes a transmembrane immunoglobulin su-
perfamily (IgSF) protein that is similar to two Drosophila
Figure 7. SYG-1 Interacts with SYG-2
Drosophila S2 cells cotransfected with SYG-1 and GFP, or SYG-2 and dsRED, subjected to self- and cross-aggregation assays.
(A) S2 cells cotransfected with SYG-1 and GFP do not aggregate.
(B) S2 cells cotransfected with SYG-2 and dsRED do not aggregate.
(C) S2 cells cotransfected with SYG-1 and GFP aggregate with S2 cells cotransfected with SYG-2 and dsRED.
(D) Time course of S2 cell aggregation. Cells cotransfected with GFP and SYG-1 were mixed with cells cotransfected with dsRED and SYG-2.
The formation of three types of cell aggregates is plotted against time. To ? total number of cells counted.
Blue trace represents the formation of cell aggregates containing both red and green cells Green circles represents cell aggregates containing
only green cells. Red triangles represents cell clusters containing only red cells. In separate experiments, cells cotransfected with GFP and
SYG-1 were tested for self-aggregation (green diamonds) and cells cotransfected with dsRED and SYG-2 were tested for self-aggregation
proteins, sticks and stones (SNS) and hibris (HIB). Its
putative receptor on HSNL, SYG-1, is an IgSF protein
closely related to two other Drosophila proteins, RST/
IrreC and Kirre/DUF. The four IgSF proteins RST/IrreC,
ila muscle fusion (Ruiz-Gomez et al., 2000, Strunkeln-
berg et al., 2001). During muscle development, muscle
founder cells attract and fuse with fusion-competent
cells to form muscles of different size and shapes. Rec-
ognition between the muscle founder cells and the fu-
sion-competent cells is mediated by RST/IrreC and
Kirre/DUF on the founder cells, and SNS and HIB in the
fusion-competent cells. Kirre/DUF binds to SNS in a S2
cell aggregation assay, suggesting that direct interac-
tion between membrane immunoglobulin proteins can
mediate specific adhesion during muscle fusion. Rst/
IrreC also functions in axon targeting in the fly central
nervous system (Schneider et al., 1995).
The mammalian orthologs of SYG-1 and SYG-2,
NEPH1and nephrin,playessentialroles inkidneydevel-
opment. Mice or humans lacking NEPH1 or nephrin die
perinatally from kidney failure due to the malformation
of the podocyte slit membrane, a tight epithelial cell
adhesion complex that is an essential component of the
glomerular filter (Khoshnoodi and Tryggvason, 2001).
Both NEPH1 and nephrin are expressed in a subset of
neurons in the central nervous system of mice, but their
et al., 2001, Putaala et al., 2001).
SYG-2 expressed on different cells, but no homophilic
binding of SYG-1 or SYG-2 in trans. This interaction
fits with the apparent heterophilic interaction of SYG-2-
expressing vulval epithelial cells and SYG-1-expressing
HSNL neurons. However, it is worth noting that many
different interactions have been observed between
members of the superfamily that includes SYG-1 and
SYG-2. In S2 aggregation assays, cells expressing RST/
IrreC exhibit homophilic binding (Schneider et al., 1995),
whereas cells expressing the similar protein Kirre/DUF
act as an inhibitor of the interaction between SNS and
SYG-2, a Synaptic Guidepost Signal
RST/IrreC and Kirre/DUF in vivo (Dworak et al., 2001).
Cells expressing mammalian nephrin exhibit homophilic
binding; in addition, nephrin and NEPH1 may bind in cis
on the same cell (Gerke et al., 2003, Liu et al., 2003). It
will be interesting to determine whether the nematode,
fly, and mammalian proteins are intrinsically different in
their binding properties, or whether cellular context or
regulation can shift their behavior between homophilic,
heterophilic, cis, and trans modes of binding.
ecules, such as Toll, Semaphorin, and Beat (Rose and
Chiba, 2000). Many of these are IgSF proteins. In gen-
eral, loss-of-function mutants of these genes have only
subtle defects in synaptic targeting, perhaps because
of functional overlap between multiple synaptogenic
ecules suggested to affect synapse formation include
cadherins, neurexins, and neuroligins (Benson et al.,
Hierarchy in Synaptic Choices
Evidence from both vertebrates and invertebrates indi-
cates that synaptic choice is not a hard-wired decision
made exclusively between two neurons, but rather a
flexible decision based on developmental context. If the
normal postsynaptic target of a neuron is absent, it can
synapse onto other target cells in vivo or in vitro. Cul-
tured hippocampal neurons can even form synapses
onto themselves in vitro, a situation that is rarely seen
in vivo (Bekkers and Stevens, 1991). Certain targets are
favored during synapse formation, but neurons appar-
ently have the capacity to form synapses onto multiple
potential target cells.
epithelial cells, secondary vulval epithelial cells, the
PVQL, PVPR, and AVKL axons, body wall muscles and
hypodermal cells in addition to the VC axons and vulval
randomly onto cells that come into contact with HSNL.
sive contact with HSNL along the ventral nerve cord in
both wild-type animals and syg mutants, but no syn-
apses onto PVQL were observed in any genetic back-
ground. These results suggest that the synaptic speci-
ficity of HSNL did not disappear entirely in the syg
mutants; rather, the specificity changed to favor syn-
apse formation onto a broader set of cells. The mole-
onto body wall muscles and hypodermal cells in syg
mutants are unknown. Our results suggest a hierarchy
of synaptic choices for HSNL, with synaptic specificity
representing the combined actionof SYG-2, SYG-1, and
other molecules that remain to be discovered.
SYG-2 Acts as a Vulval Guidepost Signal
that Interacts with SYG-1 to Specify
HSNL Synaptic Connections
The development of HSNL synapses is initiated by a
signalfrom theprimaryvulvalepithelial cells.Ourresults
argue that expression of SYG-2, driven by the LIN-11
transcription factor within primary vulval epithelial cells,
represents this guidepost signal. SYG-2 expression be-
gins immediately before SYG-1 clusters on the HSNL
axon, and SYG-2 is necessary for SYG-1 clustering to
occur. Synaptic vesicles accumulate in the same region
of HSNL in response to SYG-2, presumably indirectly
through the action of SYG-1.
Misexpression of SYG-2 in the secondary vulval epi-
mulation of synaptic vesicles to the region of HSNL that
SYG-2 is sufficient as well as necessary for localization
of HSNL synapses to adjacent regions. SYG-1 binds to
SYG-2 in cell aggregation assays, suggesting that
SYG-2 provides a direct signal to SYG-1 in HSNL. How-
ever, expression of SYG-2 in body wall muscles or PVQ
neurons did not relocalize HSNL synaptic vesicles (K.S.,
unpublished data), suggesting that other genes in vulval
epithelial cells may also contribute to guidepost
SYG-1 and SYG-2 are expressed in complex patterns
at earlier stages of C. elegans development. SYG-1 is
cord neurons, and SYG-2 is expressed in head muscles,
body wall muscles and a small number of head neurons.
It is possible that SYG-1/SYG-2 interactions contribute
to thespecificity ofsynaptic connectionsin severalneu-
ronal circuits in C. elegans. Although SYG-1 and SYG-2
have the potential to specify some synapses, their re-
stricted expression indicates that additional proteins
must contribute to synapse formation in C. elegans as
well as other animals. Several other immunoglobulin su-
perfamily proteins (IgSF) are implicated in synaptic de-
velopment. Sidekick proteins localize to synapses and
target axons to specific sublaminar structures in the
the number and strength of synaptic connections (Ben-
son et al., 2001). Mammalian SynCAM is widely ex-
pressed in the nervous system and drives the formation
of glutamatergic synapses in vitro (Biederer et al., 2002).
Sidekicks, ApCAM, fasciclin II and SynCAM all show
homophilic interactions, which are not observed for
SYG-1 and SYG-2. Studies on Drosophila neuromuscu-
lar junction formation have identified both stimulatory
synaptogenic molecules, such as NetrinB, CAP, Con-
nectin, and FasIII, and inhibitory anti-synaptogenic mol-
Strains and Genetics
Wild-type animals were C. elegans variety Bristol, Strain N2. Strains
were maintained using standard methods (Brenner, 1974). Animals
were grown at 20?C unless otherwise specified. Some strains were
provided by the Caenorhabditis Genetics Center.
LGI, lin-11(n566); LGII, unc-104(e1265); LGV, kyIs235 [unc-
86::snb-1::yfp, unc-4::lin-10::dsRED, odr-1::dsRED]; LGX, syg-
1(ky652), syg-2(ky673), syg-2(ky671); kyIs288 [unc-86::syg-1::gfp,
odr-1::rfp]; kyEx684[syg-2::gfp]; kyEx672[egl-17::syg-2, odr-1::gfp].
Isolation and Mapping of syg-2 Mutants
P0 kyIs235 animals were mutagenized with EMS using standard
procedures (Brenner, 1974) and screened semiclonally. Five inde-
pendent F1 animals were transferred onto a single plate, and HSNL
under the 100? objective of a Zeiss Axioplan II compound micro-
scope using fluorescent microscopy. Two alleles of syg-2, ky673,
and ky671, were isolated from a screen covering about 800 haploid
genomes. syg-2 is a large gene, which may account for the relatively
high frequency of alleles isolated in a small screen.
syg-2(ky673) was mapped between stP2 and stP72 Tc1 markers
on the right arm of chromosome X, and further mapped between
the left boundary of mnDf13 and the right boundary of mnDf18 using
deficiency mapping. syg-2(ky671) was mapped between stP2 and
stP72 Tc1 markers on the right arm of chromosome X.
grids and stained with uranyl acetate and Sato’s lead (Sato, 1968).
Sections were photographed with a JEOL 1200 EX/II TEM operated
at 80 kV. Every other section was photographed to follow HSNL
processes from the HSNL cell body to a position 35–40 ?m anterior
of the developing vulva. For detailed analysis of synapses, every
section was photographed for a distance of 5–6 sections on either
side of the presynaptic density.
To assess the distribution of clear synaptic vesicles with respect
to the presynaptic membrane, the negatives were digitized at 1600
dpi with and Epson 1680 scanner and distances measured in Pho-
toshop in a manner similar to that in Reist et al. (1998). Micrographs
of2160 lines/mmgratingreplicas takenatthe sametime andmagni-
fication as the HSNL processes (Ted Pella, Inc.) were measured to
convert distances in pixels to nanometers. Preliminary analyses
indicated that vesicles within 150–180 nm were associated with a
specific presynaptic density in syg-1 and syg-2 mutants. Therefore,
a distance of 180 nm surrounding a presynaptic density was used
for examining the spatial distribution of vesicles around a synapse.
Triangulation was used to determine the effective distance of vesi-
aptic density. The total number of vesicles counted for this study
was 464 in 16 synapses for N2, 142 in 11 synapses for syg-1(ky652),
318 in 17 synapses for syg-2(ky671), and 181 in 12 synapses for
Standard molecular techniques were used (details available upon
request). To identify themutations in syg-2(ky673) and syg-2(ky671),
open reading frame and splice junctions of the mutant alleles were
amplified by PCR and sequenced with an ABI sequencing machine.
The partial cDNA clone yk674g, corresponding to the predicted
open reading frame C26G2.1, was obtained from Yuji Kohara. The
5? end of the syg-2 cDNA was isolated by RT-PCR from wild-type
was subcloned under an egl-17 promoter (Burdine et al., 1998) in a
pSM vector (Shen and Bargmann, 2003). The plasmid was injected
into syg-2(ky673) at 20 ng/?l. For the S2 cell aggregation assays,
full-length cDNA of syg-1 and syg-2 were subcloned into a UAS
sequence-containing vector. UAS-GFP, UAS-GAL-4, UAS-DsRED,
and actin-GAL-4 were generous gifts from Tom Kornberg.
To analyze the expression pattern of syg-2, GFP was inserted at
the C terminus of a genomic clone that contained 10 kb of upstream
region and the entire coding region of the syg-2 gene. This clone
was injected at 20 ?g/?l into N2 animals. syg-2::gfp fully rescue the
synaptic vesicle phenotype of syg-2(ky673) and syg-2(ky671).
We thank Joe Hill and Hai Nguyen for excellent technical support;
Steve McCarroll for the pSM vector; Zemer Gitai for the unc-86
promoter; Paul Sternberg for the egl-17 promoter; Oliver Hobert for
the lin-11B promoter; Gretchen Ehrenkaufer and Tom Kornberg for
S2 cells and Drosophila expression vectors; Yuji Kohara for a partial
syg-2 cDNA and for the NEXTDB expression database. This work
was supported by the Howard Hughes Medical Institute. K.S. is a
Hughes Medical Institute.
S2 Cell Aggregation Assay
Drosophila S2 cells were transfected using Effectene (Gibco Life
Science) and S2 cell aggregation assays were performed as de-
scribed (Schneider et al., 1995). Briefly, S2 cells were cotransfected
with UAS-GFP, UAS-SYG-1 cDNA, and actin-GAL-4, or with UAS-
dsRED, UAS-SYG-2, and actin-GAL-4. Cells were allowed to grow
for 36 hr after transfection, collected and washed in PBS (phos-
phate-buffered saline), and resuspended in Schneider’s Drosophila
medium at 2 ? 106cells/ml. 1 ml of this suspension was placed in
2 ml tubes and rotated at 30 rpm for various amount of time at room
temperature. Aliquots were spotted onto slides and immediately
counted for aggregate formation under an Axioplan II microscope,
using DIC microscopy to examine aggregates and fluorescence mi-
croscopy to identify cotransfected cells.
Received: October 24, 2003
Revised: January 8, 2004
Accepted: January 15, 2004
Published: March 18, 2004
Allendoerfer, K.L., and Shatz, C.J. (1994). The subplate, a transient
neocortical structure: its role in the development of connections
between thalamus and cortex. Annu. Rev. Neurosci. 17, 185–218.
Bekkers, J.M., and Stevens, C.F. (1991). Excitatory and inhibitory
autaptic currents in isolated hippocampal neurons maintained in
cell culture. Proc. Natl. Acad. Sci. USA 88, 7834–7838.
Benson, D.L., Colman, D.R., and Huntley, G.W. (2001). Molecules,
maps and synapse specificity. Nat. Rev. Neurosci. 2, 899–909.
Biederer, T., Sara, Y., Mozhayeva, M., Atasoy, D., Liu, X., Kavalali,
E.T., and Su ¨dhof, T.C. (2002). SynCAM, a synaptic adhesion mole-
cule that drives synapse assembly. Science 297, 1525–1531.
Burdine, R.D., Branda, C.S., and Stern, M.J. (1998). EGL-17(FGF)
expression coordinates the attraction of the migrating sex myo-
blasts with vulval induction in C. elegans. Development 125, 1083–
Desai, C., Garriga, G., McIntire, S.L., and Horvitz, H.R. (1988). A
genetic pathway for the development of the Caenorhabditis elegans
HSN motor neurons. Nature 336, 638–646.
Donoviel, D.B., Freed, D.D., Vogel, H., Potter, D.G., Hawkins, E.,
Barrish, J.P., Mathur, B.N., Turner, C.A., Geske, R., Montgomery,
C.A., et al. (2001). Proteinuria and perinatal lethality in mice lacking
NEPH1, a novel protein with homology to NEPHRIN. Mol. Cell. Biol.
Dworak, H.A., and Sink, H. (2002). Myoblast fusion in Drosophila.
Bioessays 7, 591–601.
Transgenic strains were created as previously described (Mello and
Fire, 1995). Genomic DNA covering the coding region of C26G2.1,
ping PCR reactions (high fidelity PCR kit, Roche Inc.). The PCR
products were injected as a mix at 20 ng/?l using odr-1::gfp as a
coinjection marker. Four out of five lines showed complete rescue
of the synaptic vesicle phenotype of syg-2(ky673). A genomic sub-
clone of cosmid C26G2 that covers the slt-1 region was injected
into syg-2(ky673). No rescue of the synaptic vesicle phenotype was
observed (n ? 6). The same slt-1 genomic clone fully rescued the
AVM axon guidance of slt-1(eh15) (Hao et al., 2001).
L4 hermaphrodites of syg-1(ky652), syg-2(ky673), syg-2(ky671), and
N2 were prepared for conventional transmission electron micros-
copy by fixation in 0.8% glutaraldehyde, 0.8% osmium tetroxide in
0.1 M Na-cacodylate buffer, [pH 7.4] for 2 hr at 4?C with rotation,
with a change of fixative after the first hour. The animals were rinsed
3 ? 15 min with 100 mM Na-cacodylate buffer at 4?C, and postfixed
overnight with 0.5% osmium tetroxide in 100 mM Na-cacodylate
buffer at 4?C with rotation. Following postfixation, the specimens
were rinsed 2 ? 15 min with Na-cacodylate buffer, 2 ? 15 min with
water, and then stained en bloc with 1% aqueous uranyl acetate
for 1 hr at room temperature. Specimens were rinsed with water,
dehydrated with ethanol, passed through propylene oxide, and em-
bedded in Eponate 12 resin. Serial 50–60 nm sections were cut with
a Leica Ultracut T microtome, collected on Formvar coated slot
SYG-2, a Synaptic Guidepost Signal Download full-text
Dworak, H.A., Charles, M.A., Pellerano, L.B., and Sink, H. (2001).
Characterization of Drosophila hibris, a gene related to human
nephrin. Development 128, 4265–4276.
Ferguson, E.L., Sternberg, P.W., and Horvitz, H.R. (1987). A genetic
pathway for the specification of the vulval cell lineages of Caeno-
rhabditis elegans. Nature 326, 259–267.
Freyd, G., Kim, S.K., and Horvitz, H.R. (1990). Novel cysteine-rich
motif and homeodomain in the product of the Caenorhabditis ele-
gans cell lineage gene lin-11. Nature 344, 876–879.
Gerke, P., Huber, T.B., Sellin, L., Benzing, T., and Walz, G. (2003).
Homodimerization and heterodimerization of the glomerular podo-
cyte proteins nephrin and NEPH1. J. Am. Soc. Nephrol. 14, 918–926.
Gupta, B.P., Wang, M., and Sternberg, P.W. (2003). The C. elegans
LIM homeobox gene lin-11 specifies multiple cell fates during vulval
development. Development 130, 2589–2601.
Hall, D.H., and Hedgecock, E.M. (1991). Kinesin-related gene unc-
104 is required for axonal transport of synaptic vesicles in C. ele-
gans. Cell 65, 837–847.
Hao, J.C., Yu, T.W., Fujisawa, K., Culotti, J.G., Gengyo-Ando, K.,
Mitani, S., Moulder, G., Barstead, R., Tessier-Lavigne, M., and Barg-
mann, C.I. (2001). C. elegans slit acts in midline, dorsal-ventral, and
anterior-posterior guidance via the SAX-3/Robo receptor. Neuron
tion of cortical circuits. Science 274, 1133–1138.
make-up of the glomerular podocyte slit diaphragm. Exp. Nephrol.
Liu, G., Kaw, B., Kurfis, J., Rahmanuddin, S., Kanwar, Y.S., and
Chugh, S.S. (2003). Neph1 and nephrin interaction in the slit dia-
phragm is an important determinant of glomerular permeability. J.
Clin. Invest. 112, 209–221.
Mello, C., and Fire, A. (1995). DNA transformation. Methods Cell
Biol. 48, 451–482.
Nonet, M.L. (1999). Visualization of synaptic specializations in live
C. elegans with synaptic vesicle protein-GFP fusions. J. Neurosci.
Methods 89, 33–40.
Nurrish, S., Segalat, L., and Kaplan, J.M. (1999). Serotonin inhibition
of synaptic transmission: Galpha(0) decreases the abundance of
UNC-13 at release sites. Neuron 24, 231–242.
Putaala, H., Soininen, R., Kilpelainen, P., Wartiovaara, J., and Trygg-
vason, K. (2001). The murine nephrin gene is specifically expressed
in kidney, brain and pancreas: inactivation of the gene leads to
massive proteinuria and neonatal death. Hum. Mol. Genet. 10, 1–8.
Reist, N.E., Buchanan, J., Li, J., DiAntonio, A., Buxton, E.M., and
Schwarz, T.L. (1998). Morphologically docked synaptic vesicles are
reduced in synaptotagmin mutants of Drosophila. J. Neurosci. 18,
Rose, D., and Chiba, A. (2000). Synaptic target recognition at Dro-
sophila neuromuscular junctions. Microsc. Res. Tech. 49, 3–13.
Ruiz-Gomez, M., Coutts, N., Price, A., Taylor, M.V., and Bate, M.
(2000). Drosophila dumbfounded: a myoblast attractant essential
for fusion. Cell 102, 189–198.
Sato, T. (1968). A modified method for lead staining of thin sections.
J. Electron Microsc. (Tokyo) 17, 158–159.
Schneider, T., Reiter, C., Eule, E., Bader, B., Lichte, B., Nie, Z.,
expression of the irreC-rst protein is required for normal axonal
projections of columnar visual neurons. Neuron 15, 259–271.
Shen, K., and Bargmann, C.I. (2003). The immunoglobulin superfam-
ily protein SYG-1 determines the location of specific synapses in
C. elegans. Cell 112, 619–630.
Somogyi, P., Tamas, G., Lujan, R., and Buhl, E.H. (1998). Salient
features of synaptic organisation in the cerebral cortex. Brain Res.
Brain Res. Rev. 26, 113–135.
Strunkelnberg, M., Bonengel, B., Moda, L.M., Hertenstein, A., de
Couet, H.G., Ramos, R.G., and Fischbach, K.F. (2001). rst and its
paralogue kirre act redundantly during embryonic muscle develop-
ment in Drosophila. Development 128, 4229–4239.
Tryggvason, K. (1999). Unraveling the mechanisms of glomerular
ultrafiltration: nephrin, a key component of the slit diaphragm. J.
Am. Soc. Nephrol. 11, 2440–2445.
Weinshenker, D., Garriga, G., and Thomas, J.H. (1995). Genetic and
pharmacological analysis of neurotransmitters controlling egg lay-
ing in C. elegans. J. Neurosci. 15, 6975–6985.
White, J.G., Southgate, E., Thomson, J.N., and Brenner, S. (1976).
The structure of the ventral nerve cord of Caenorhabditis elegans.
Philos. Trans. R. Soc. Lond. B Biol. Sci. 275, 327–348.
White, J.G., Southgate, E., Thomson, J.N., and Brenner, S. (1986).
elegans. Philos. Trans. R. Soc. Lond. B Biol. Sci. 314, 1–340.
Yamagata, M., Weiner, J., and Sanes, J. (2002). Sidekicks. Synaptic
adhesion molecules that promote lamina-specific connectivity in
the retina. Cell 110, 649–660.