Cleavage of Neuroligin-1
Kunimichi Suzuki,1Yukari Hayashi,1Soichiro Nakahara,2Hiroshi Kumazaki,3Johannes Prox,5Keisuke Horiuchi,6
Mingshuo Zeng,3Shun Tanimura,3Yoshitake Nishiyama,3Satoko Osawa,1Atsuko Sehara-Fujisawa,7Paul Saftig,5
Satoshi Yokoshima,3Tohru Fukuyama,3Norio Matsuki,2Ryuta Koyama,2Taisuke Tomita,1,8,* and Takeshi Iwatsubo1,4,8
1Department of Neuropathology and Neuroscience
2Laboratory of Chemical Pharmacology
3Laboratory of Synthetic Natural Products Chemistry, Graduate School of Pharmaceutical Sciences
4Department of Neuropathology, Graduate School of Medicine
The University of Tokyo, Hongo 7-3-1, Bunkyo, Tokyo 113-0033, Japan
5Institut fu ¨r Biochemie, Christian-Albrechts-Universita ¨t zu Kiel, D-24098 Kiel, Germany
6Department of Orthopedic Surgery, School of Medicine, Keio University, Shinjuku, Tokyo 160-8582, Japan
7Department of Growth Regulation, Institute for Frontier Medical Sciences, Kyoto University, Sakyo-ku, Kyoto 606-8507, Japan
8Core Research for Evolutional Science and Technology, Japan Science and Technology Agency
Neuroligin (NLG), a postsynaptic adhesion molecule,
is involved in the formation of synapses by binding to
a cognate presynaptic ligand, neurexin. Here we
report that neuroligin-1 (NLG1) undergoes ectodo-
main shedding at the juxtamembrane stalk region
to generate a secreted form of NLG1 and a
membrane-tethered C-terminal fragment (CTF) in
adult rat brains in vivo as well as in neuronal cultures.
Pharmacological and genetic studies identified
ADAM10 as the major protease responsible for
NLG1 shedding, the latter being augmented by
synaptic NMDA receptor activation or interaction
with solubleneurexin ligands. NLG1-CTF was subse-
quently cleaved by presenilin/g-secretase. Secretion
of soluble NLG1 was significantly upregulated under
a prolonged epileptic seizure condition, and inhibi-
tion of NLG1 shedding led to an increase in numbers
of dendritic spines in neuronal cultures. Collectively,
neuronal activity-dependent proteolytic processing
of NLG1 may negatively regulate the remodeling of
spines at excitatory synapses.
Formation and maintenance of the synaptic structure is a dy-
namic process that requires bidirectional interactions between
pre- and postsynaptic components. A diverse assortment of
cell adhesion molecules is present at the synapse and organizes
the synaptic specializations of both excitatory and inhibitory
central synapses (Dalva et al., 2007; Siddiqui and Craig, 2011).
Neuroligin (NLG) is one of the potent synaptogenic adhesion
proteins located at the postsynapse, which transsynaptically
binds to a presynaptic ligand, neurexin (NRX) (Ichtchenko
et al., 1995; Irie et al., 1997; Scheiffele et al., 2000; Su ¨dhof,
2008; Bottos et al., 2011). Mammals express four NLG genes
(i.e., NLG1 to NLG4). NLG polypeptides are type 1 transmem-
brane proteins with a large extracellular domain with homology
to acetylcholinesterases but lack critical residues in the active
site and interact with NRXs at the synaptic membrane surface
(Su ¨dhof, 2008). Notably, NLG1 is localized at glutamatergic
postsynapse, and overexpression of NLG1 induces the accumu-
lation of glutamatergic presynapse and postsynapse molecules
in vitro (Song et al., 1999; Scheiffele et al., 2000; Budreck and
Scheiffele, 2007). In contrast, NLG2 triggers the maturation of
GABAergic synapses, implicating specific functions of different
NLGs in theformation and maturation of different chemical types
of synapses in vitro and in vivo (Graf et al., 2004; Varoqueaux
et al., 2004, 2006).
Recent studies revealed that copy number variation or point
mutation in NLG genes are linked to autism spectrum disorder
(ASD), schizophrenia, or mental retardation (reviewed in Su ¨dhof,
2008). Notably, ASD-linked mutations in NLG genes have been
shown to affect the expression, folding, or dimerization of NLG
proteins to compromise their surface expression and binding
to NRXs (Comoletti et al., 2004; Levinson and El-Husseini,
2007; Zhang et al., 2009). Moreover, copy number variations
that are associated with an increased risk of ASD were identified
in NLG1 locus (Glessner et al., 2009). NLG1 knockout (KO) or
transgenic mice showed synaptic dysfunctions and ASD-like
behaviors (Varoqueaux et al., 2006; Chubykin et al., 2007; Blun-
dell et al., 2010; Dahlhaus et al., 2010). Thus, the levels of NLGs
within the synaptic membranes are presumed to directly modu-
late the synaptic functions in vivo. Although several reports indi-
cated that the surface levels of NLG1 are regulated by synaptic
activities through membrane trafficking (Schapitz et al., 2010;
Thyagarajan and Ting, 2010), the regulatory mechanisms to
control protein levels of NLG remains unclear. Here, we show
that NLG1 is sequentially cleaved by ADAM10 and g-secretase
to release its extra- and intracellular domain fragments, respec-
tively. Proteolytic processing of NLG1 resulted in the elimination
of NLG1 on the cell surface, thereby causing a decrease in the
410 Neuron 76, 410–422, October 18, 2012 ª2012 Elsevier Inc.
synaptogenic activity of NLG1. We further show that ADAM10-
mediatedshedding isregulated inanactivity-dependent manner
through NMDA receptor (NMDAR) activation or by binding to
secreted forms of NRXs. Our present results suggest that
neuronal activity and interaction with NRXs regulate the levels
of NLG1 via proteolytic processing to modulate the adhesion
system as well as the functions of synapses.
Proteolytic Processing of NLGs in Brains and Neuronal
NLGs are synaptogenic type 1 transmembrane proteins that
harbor large extracellular domains (Ichtchenko et al., 1995).
Whilethe levels of NLGsarepresumed tobecorrelated with their
physiological and pathological functions (Varoqueaux et al.,
2006; Chubykin et al., 2007; Glessner et al., 2009; Blundell
et al., 2010; Dahlhaus et al., 2010), little information is available
on the proteolytic mechanism of NLGs. Several lines of evidence
have indicated that a subset of type 1 transmembrane proteins
areprocessed by sequential cleavages by ectodomain shedding
and intramembrane cleavage, the latter being executed by
g-secretase (Beel and Sanders, 2008; Bai and Pfaff, 2011). To
test whether the levels of NLGs are regulated by proteolytic
processing, weanalyzed endogenous NLGpolypeptides in adult
rat brains (Figure 1A). Immunoblot analysis using antibodies that
specifically recognize the cytoplasmic region of NLG1 and NLG2
(see Figure S1 available online) revealed immunopositive bands
at ?20–25 kDa, in addition to full-length (FL) protein that
migrated at ?120 kDa. Because the predicted sizes of the cyto-
the ?20–25 kDapolypeptidesrepresent the membrane-tethered
C-terminal fragment (CTF) of endogenous NLGs. Multiple bands
corresponding to CTFs may represent different posttransla-
tional modifications (e.g., glycosylation, see below). To examine
whether these CTFs are processed by the g-secretase activity,
we incubated the membrane fractions of rat brains at 37?C and
detected the appearance of additional bands that migrate faster
than the CTFs with each NLG. Moreover, addition of DAPT,
a specific g-secretase inhibitor, abolished the generation of
the smaller CTFs, in a similar manner to that observed with a
well-known g-secretase substrate, amyloid precursor protein
(APP). These data suggested that NLG-CTFs are cleaved by
the g-secretase activity to release the intracellular domain
(ICD) (Figure 1B). In parallel with the generation of ICDs, we
observed a significant reduction in NLG1-FL upon incubation,
concomitant with the generation of a smaller NLG1 fragment,
which was detected by an antibody against the extracellular
region of NLG1 (Figure 1C). Generation of this extracellular frag-
mentof NLG1was decreasedbytreatmentwith metalloprotease
inhibitors (i.e., EDTA, TAPI2), supporting the notion that the
extracellular domain of NLG1 is processed by ectodomain
shedding. To test whether this processing occurs at synapses
under a physiological condition, we incubated synaptoneuro-
ulation of purified presynaptic boutons attached to postsynaptic
processes (Villasana et al., 2006; Kim et al., 2010) (Figures 1D
and 1E). After ultracentrifugation after incubation, soluble
NLG1 (sNLG1) as well as NLG1-ICD was detected in the soluble
fraction, which was abolished by coincubation with TAPI2 and
DAPT, respectively. To ascertain that these cleavages occur
in situ in neuronal cultures, we analyzed cell lysates and condi-
tioned media (CM) from mouse cortical primary neuronal
cultures obtained from embryonic day (E) 18 pups by immuno-
blotting and detected the secretion of an ?98 kDa single poly-
peptide in the conditioned media, which migrated at an identical
position to that generated upon incubation of the membrane
fractions, by an antibody against the extracellular domain of
NLG1 (Figures 1F and 1G). This band disappeared by treatment
with metalloprotease inhibitors (i.e., GM6001, TAPI2). These
data suggest that the extracellular domain of NLG1 is shed
by the metalloprotease activity to release sNLG1 into the
conditioned media. Furthermore, DAPT treatment caused the
accumulation of CTFs of NLG1 as well as of NLG2. Notably,
simultaneous administration of DAPT and metalloprotease
inhibitors decreased the accumulation of the CTFs. However,
endogenous NLG-ICD, which was observed upon incubation
of microsomes from brain lysates, was hardly detectable in cell
lysates from cultured primary neurons. This suggests that
NLG-ICD is a highly labile endoproteolytic product. These find-
ings led us to speculate thatNLGs areinitially processed bymet-
alloprotease at the extracellular region to generate sNLG and
membrane-tethered NLG-CTF, the latter being further cleaved
by the g-secretase activity (Figure 1H).
ADAM10 and g-Secretase Are Responsible
for the Proteolytic Processing of NLG1
Next we analyzed the metabolism of NLGs in mouse embryonic
fibroblasts from Psen1?/?/Psen2?/?double knockout mice
(DKO cells), which completely lacks the g-secretase activity
(Herreman et al., 2000). Accumulation of NLG-CTFs was
observed upon the overexpression of hemagglutinin (HA)-
tagged NLGs in DKO cells (Figure 2A). However, the levels of
the accumulated NLG-CTFs were significantly reduced by the
coexpression of human PS1, indicating that g-secretase activity
is responsible for the processing of NLG-CTFs. ADAM10 is
known as a responsible enzyme for ectodomain shedding of
a subset of g-secretase substrates (e.g., Notch, APP, cadherin,
and CD44) at the membrane-proximal region of ectodomain
(Saftig and Reiss, 2011). To test whether ADAM10 is involved
in the processing of NLGs, we overexpressed HA-tagged
NLG1 or NLG2 in murine embryonic fibroblasts (MEFs)
obtained from ADAM10 knockout (Adam10?/?) or heterozygous
(Adam10+/?) mice (Figure 2B) (Hartmann et al., 2002). In
Adam10?/?MEF, the generation of sNLG1 was significantly
reduced. In contrast, no change in NLG1 processing was
observed in MEFs obtained from knockout mice of other
ADAMs (i.e., Adam8?/?, Adam17?/?, Adam19?/?, Adam9?/?;
Adam12?/?;Adam15?/?[TKO]) (Zhou et al., 2004; Weskamp
et al., 2006; Kawaguchi et al., 2007; Horiuchi et al., 2007). These
data strongly suggest that ADAM10 is a responsible enzyme for
the shedding of NLG1. Intriguingly, the level of soluble NLG2
secreted fromAdam10?/?MEFwasalmost comparable tothose
from other ADAM knockout MEFs, suggesting that ADAM10
specifically cleaves NLG1 but not NLG2. These data suggest
Proteolytic Regulation of Neuroligin-1
Neuron 76, 410–422, October 18, 2012 ª2012 Elsevier Inc. 411
that ADAM10 and g-secretase areresponsible for the proteolytic
processing of NLG1 in transfected fibroblasts.
To further examine the role of ADAM10 in the processing of
endogenous NLG1, we treated rat primary neurons obtained
from E18 pups with INCB3619, a known ADAM10/17 inhibitor
(Witters et al., 2008). INCB3619 abolished the secretion of
sNLG1 in a similar manner to that of sAPPa, the latter being
generated by ADAM10 (IC50: 1.6 mM) (Figures 3A and 3C). In
contrast, treatment with INCB3420, a derivative of INCB3619
that harbors a moderate ADAM10/17 inhibitory activity but
potently inhibits matrix metalloproteases (MMPs) (i.e., MMP2,
MMP9, MMP12, and MMP15) (Zhou et al., 2006), decreased
the NLG1 cleavage only at high concentrations (IC50: >10 mM)
(Figures 3B and 3C). In addition, INCB3420 did not affect the
sNLG1 production by incubation of synaptoneurosome fraction
of rat adult brain (Figure S2A). Moreover, other MMP-specific
inhibitors with different chemical structures (MMP2, MMP3,
MMP9, and MMP13 inhibitors) did not affect the sNLG1
Figure 1. Proteolytic Processing of Endogenous NLGs in Rat Brains
(A) Immunoblot analysis of membrane fraction from adult rat cortex using an antibody against the intracellular domain of NLGs. Molecular weight was indicated
on the left.
(B) Processing of NLG-CTFs in the cell-free assay using membrane fractions from adult rat cortex. Used antibodies are indicated below the panel. Arrows, open
arrows, and asterisks represent CTFs, ICDs, and nonspecific bands, respectively. IN, input; 4, incubation at 4?C; 37, incubation at 37?C; D, incubation at 37?C
(C)SheddingofNLG1-FLinthecell-freeassay.Arrowhead andanopenarrowheaddenoteNLG1-FL andsecretedNLG1,respectively.E,samplesincubatedwith
EDTA; T, samples incubated with TAPI2.
(D) Incubation of synaptoneurosome fraction from adult mouse brain. Open arrowhead, open arrow, arrowhead, and arrow denote sNLG1, NLG1-ICD, NLG1-FL,
and NLG1-CTF, respectively. 4, incubation at 4?C; 37, incubation at 37?C; D, incubation at 37?C with DAPT; T, incubation at 37?C with TAPI2.
(E) Densitometric analysis of sNLG1 (black bar) and NLG1-FL (gray bar) in (D) (n = 3, mean ± SEM; *p < 0.05; **p < 0.01; ***p < 0.001 versus 4 degrees by
Student’s t test).
(F) Primary neurons from E16 mouse at DIV10 were treated with indicated inhibitors for 48 hr. Conditioned medium (CM) and cell lysates were analyzed by
immunoblot analysis with antibodies indicated below the panels.
(G) Densitometric analysis of sNLG1 (black bar) and NLG1-ICD (white bar) in (F) (n = 4–6, mean ± SEM; ***p < 0.001 versus DMSO;###p < 0.001 versus DAPT by
Student’s t test).
(H) Schematic depiction of NLG processing.
Proteolytic Regulation of Neuroligin-1
412 Neuron 76, 410–422, October 18, 2012 ª2012 Elsevier Inc.
production or decreased only at high concentrations from rat
in NLG1 shedding in primary neurons (Figures 3B and 3D). We
neuroblastoma neuro2a cells (Figures 3E and 3F) as well as in
primary neurons from P1 Adam10flox/floxmice (Yoda et al.,
2011) (Figures 3G and 3H) by siRNA transfection and overex-
pression of Cre recombinase, respectively. Inhibition of NLG1
shedding, along with impairment of sAPPa generation as previ-
ously described (Jorissen et al., 2010; Kuhn et al., 2010), and
accumulation of NLG1-FL were observed both in Adam10
knockdown neuro2a cells and Adam10 knockout neurons.
Moreover, overexpression of ADAM10 in mouse primary
neurons increased the production of sNLG1 (Figure S2B). Lastly,
to demonstrate the physiological significance of ADAM10 in
NLG1 processing in vivo, we incubated the brain microsome
from postnatal day (P) 18 neuron-specific conditional ADAM10
knockout mice (Adam10flox/flox; CamKII-Cre) (Figure 3I). Notably,
thelevels of NLG1-FL wereincreased inthe microsome fractions
from brains of Adam10 conditional knockout mice, whereas
sNLG1 production was significantly decreased (Figures 3J and
3K). Taken together, we concluded that the major physiological
sheddase of NLG1 in brain is ADAM10.
We next examined the proteolytic processing of NLG1 at the
ectodomain in more detail. According to the molecular weight
cleavage site for shedding. To obtain further precise information
on the location and characteristics of the cleavage site, we
Figure 2. Proteolytic Processing of NLG1 in
Presenilin or ADAM Knockout Cells
(A and B) Immunoblot analysis of HA-NLG1 or HA-
NLG2 overexpressed in MEFs obtained from PS1/
PS2 DKO (A) or ADAM KO (B) mice. Twenty-four
hours after treatment of indicated compounds
(i.e., 10 mM of DAPT in A, 10 mM of INCB3619 in B),
CM and cell lysates were analyzed using anti-
bodies indicated below the panels. Genotypes of
ADAM knockout cells were indicated above the
lanes. TKO denotes cells derived from Adam9?/?;
Adam12?/?;Adam15?/?triple knockout mice.
(C) Densitometric analysis of HA-sNLG1 and
HA-sNLG2 in Adam10+/?and Adam10?/?cells (B)
(n = 4, mean ± SEM; ***p < 0.001 versus hetero-
zygous cells [black bar] by Student’s t test).
analyzed mutant forms of NLG1 overex-
pressed in COS-1 cells (Figure 4A). Re-
combinant NLG1-FL was sequentially
cleaved similarly to the endogenous
NLG1-FL in primary neurons, whereas
NLG1-ICD was not observed (Figures
4B and 4C). To identify the extracellular
cleavage site of NLG1, we have system-
atically generated a series of alanine
substitutions around the juxtamembrane
stalk region, at sites consistent with the
molecular weight of sNLG1 and NLG1-
CTF: K674QDD/AAAA, P678KQQ/AAAA,
P682SPF/AAAA, S686VDQ/AAAA, R690DYS/AAAA, and T694E/AA
(Figure 4A). Among these mutants, generation of sNLG1 was
significantly decreased in the PKQQ/AAAA mutant, whereas
the level of mutant FL protein was increased (Figures 4D and
4E). The observation that overexpressed PKQQ/AAAA mutant
NLG1 accumulated at dendritic spines in rat primary neurons
similarly to wild-type (WT) NLG1 and showed spinogenic effects
supported the view that introduction of PKQQ/AAAA mutation
did not affect the trafficking of NLG1 (see Figure 8C). In contrast,
alanine substitutions at regions proximal to the membrane
caused an augmentation of NLG1 shedding. Intriguingly, the
levels of sNLG1 were significantly increased in the PSPF/AAAA
and SVDQ/AAAA mutants, whereas the corresponding FL
proteins were decreased, suggesting that the region from
Pro682to Gln689negatively regulates the NLG1 shedding. These
data suggest that the region between Pro678to Gln681is critical
for the sNLG1 production.
We thenoverexpressed NLG1-DE, a recombinant polypeptide
corresponding to NLG-CTF starting from residue Val682fused
with a signal peptide (Figure 4A). The levels of NLG1-DE were
increased by DAPT treatment, indicating that NLG1-DE was
processed by g-secretase (Figure 4F). However, NLG1-ICD
was again undetectable in lysates of transfected cells. Indeed,
recombinant NLG1-ICD, which corresponds to the predicted
g-secretase product starting at an intramembranous residue
Val717, was detected only after proteasome inhibitor treatment
(Figure 4G), suggesting that NLG1-ICD is highly labile to protea-
somal degradation. In fact, the immunostaining of NLG1-ICD
Proteolytic Regulation of Neuroligin-1
Neuron 76, 410–422, October 18, 2012 ª2012 Elsevier Inc. 413
was relatively weak and predominantly detected in neuronal
nuclei and somata, whereas NLG1-FL, as well as the other
mutants, was localized to the somatodendritic compartments
in transfected primary neurons (Figure S3). We further examined
the significance of the PDZ-binding motif located at the C
terminus of NLG1 and found that deletion of this motif did not
impact on the shedding as well as the g-secretase-mediated
cleavage (Figures 4A, 4B, 4C, and 4F). Collectively, these data
Figure 3. Pharmacological and Genetic Analyses of NLG1 Shedding in Neurons
(A and B) Immunoblot analysis of CM of rat primary neurons obtained from E17–E18 pups treated with ADAM (A) or MMP (B) inhibitors at DIV4. CM was collected
at DIV6. Used compounds and concentration are shown above the lanes.
(C and D) Densitometric analysis of sNLG1 from neurons treated with ADAM (C) or MMP (D) inhibitors (n = 4–8, mean ± SEM;###p < 0.001 versus mock by
Student’s t test; *p < 0.05 versus mock by one-way ANOVA followed by Dunnett’s post test).
(E) Immunoblot analysis of overexpressed HA-NLG1 in neuro2a cells cotransfected with siRNA duplex for nontarget or Adam10. CM and cell lysates were
analyzed by immunoblotting using anti-HA for HA-NLG1, APP597 for endogenous APP, and anti-ADAM10 and DM1A for a-tubulin.
(F) Densitometric analysis of HA-NLG1 and HA-sNLG1 in (E) (n = 4, mean ± SEM; *p < 0.05, **p < 0.01 versus nontarget siRNA by Student’s t test).
(G) Immunoblot analysis of mouse primary neurons infected with recombinant adenoviruses. Neurons were obtained from wild-type or Adam10flox/floxmouse at
P1 and infected at DIV4 and replaced with fresh medium at DIV6. Samples were obtained at DIV8 and subjected to immunoblot analysis.
(H) Densitometric analysis of sNLG1 and sNLG2 in (G) (n = 3–6, mean ± SEM; ***p < 0.001 versus lacZ-infected cells by Student’s t test).
(I) Immunoblot analysis of brain microsomes obtained from Adam10flox/flox; CamKII-Cre (cKO) and Adam10flox/flox(WT) mice at P18.
(J) Cell-free assayof brain microsomes from cKOand WT mice.After incubation,sampleswere centrifuged toseparatesolubleand membranefractionsand then
subjected to immunoblot analysis.
(K) Relative changes in the levels of NLG1-FL and sNLG1 in (J) compared with that at 4 degrees (n = 3, mean ± SEM; *p < 0.05 versus WT by Student’s t test).
Proteolytic Regulation of Neuroligin-1
414 Neuron 76, 410–422, October 18, 2012 ª2012 Elsevier Inc.
and g-secretase to release sNLG1 and a highly labile NLG1-ICD.
Neuronal Activity and Ligand Binding Upregulate NLG1
It has been shown that some g-secretase substrates (e.g., APP,
N-cadherin, and EphA4) undergo cleavage in an activity-depen-
dent manner in neurons (Kamenetz et al., 2003; Marambaud
et al., 2003; Reiss et al., 2005; Inoue et al., 2009). To investigate
theeffect ofsynapticactivityonNLG1processing,wetreated rat
primary neuronal culture at day in vitro (DIV) 11 with a set of
compounds. Fifteen minute treatments with glutamate or
NMDA significantly increased the sNLG1 level in the conditioned
media, which was abolished by addition of NMDA receptor
antagonists (i.e., D-AP5 and MK-801) (Figures 5A and 5B).
Intriguingly, pretreatment with MK-801 (Figures 5C and 5D), an
open-channel blocker of NMDA receptor (Huettner and Bean,
1988), completely inhibited the NLG1 shedding induced by
glutamate, suggesting that the physiological activation of
functional NMDA receptors is sufficient for the generation of
sNLG1 at the glutamatergic synapses. To examine whether the
shedding regulates the cell surface level of NLG1, we performed
Figure 4. Proteolytic Processing of Mutant Forms of NLG1 at the Predicted Cleavage Site in COS-1 Cells
(A) Schematic depiction of mutant NLG1 constructs used in this study. Numbers represent corresponding amino acid residues in mouse NLG1. Transmembrane
and PDZ-binding domains are indicated as gray and black boxes, respectively. Boxes with SS and HA represent signal sequence and inserted HA tag,
respectively. Primary amino acid sequences of the stalk region of NLG1 and NLG2 are shown by single letter code. Locations of alanine substitution mutations in
NLG1 were indicated by underbar. Asterisks represent O-glycosylation sites in the stalk region of NLG1.
(B–G) Immunoblot analysis of CM as well as lysates of COS cells expressing NLG1DPDZ (B), NLG1 with amino acid substitutions (D), NLG1-DE (F), or NLG1-ICD
(G) treated with indicated compounds. Note that NLG1-ICD was detected only in epoxomicin (epox)-treated COS cells. Densitometric analysis of secreted HA-
sNLG1 standardizedby HA-NLG1 expression of lysates in(E) and (D) wereshown in(C) and (E), respectively (n = 3,mean ± SEM; ***p <0.001 versus HA-NLG1 by
Student’s t test).
Proteolytic Regulation of Neuroligin-1
Neuron 76, 410–422, October 18, 2012 ª2012 Elsevier Inc. 415
a cell surface biotinylation experiment in rat primary neurons
(Figure 5F). Treatment with TAPI2 or GM6001 significantly
increased the surface levels of NLG1 (Figures 5G and 5H). More-
over, secretion of biotinylated sNLG1 was detected in the condi-
tioned media of labeled primary neurons (Figures 5I, 5J, and 5K).
Notably, increased sNLG1 by NMDA treatment also was bio-
tinylated, suggesting that the proteolytic processing of NLG1
occurs at the cell surface and regulates the levels of cell surface
Figure 5. Activity-Dependent Shedding of NLG1 in Primary Neurons
(A) Primary neurons from E18 rat cortex were treated with indicated compounds at DIV11 for 15 min. CM as well as cell lysates were analyzed by immunoblotting
using antibodies indicated below the panels.
(B) Densitometric analysis of the amounts of sNLG1 secreted from rat primary neurons treated as in (A). Statistical analysis was carried out by Student’s t test
(n = 5–9, mean ± SEM, ***p < 0.001 versus mock;###p < 0.001 versus Glu).
(C–E) Production of sNLG1 from rat primary neurons from E17–E18 pups pretreated with NMDAR antagonists at DIV11 (C). Representative result of immunoblot
analysis of CM is shown in (D). Noncompetitive NMDAR antagonist, MK-801, totally abolished the production of sNLG1. NBM indicates Neurobasal medium.
Densitometric analysis of sNLG1 is shown in (E) (n = 3, mean ± SEM; ***p < 0.001 versus (?) by Student’s t test).
(F–H) Surface protein levels were examined by the surface biotinylation technique in rat primary neurons obtained from E17–E18 pups at DIV6 (F). Representative
result of immunoblot analysis is shown in (G). TAPI2 or GM6001 treatment from DIV4 to DIV6 caused a significant accumulation of total as well as cell surface
NLG1. Densitometric analysis of sNLG1 is shown in (H) (n = 4, mean ± SEM; *p < 0.05; **p < 0.01 versus by Student’s t test).
(I–K) Detection of de novo secretion of biotinylated sNLG1 from DIV9 rat primary neurons from E17–E18 pups (I). Representative result of immunoblot analysis is
shown in (J). Densitometric analysis of sNLG1 is shown in (K) (n = 3–5, mean ± SEM; **p < 0.01; ***p < 0.001 versus mock by Student’s t test).
Proteolytic Regulation of Neuroligin-1
416 Neuron 76, 410–422, October 18, 2012 ª2012 Elsevier Inc.
NLG1. Taken together with the results of synaptoneurosome
incubation (Figure 1D), these data indicate that glutamatergic
synaptic transmission through NMDA receptor activation modu-
lates the levels of NLG1 at the synaptic membrane.
It has been shown that several g-secretase substrates are
cleaved upon binding with cognate membrane-tethered or
soluble ligands, e.g., Delta/Jagged for Notch (Mumm et al.,
2000), Hyaluronan for CD44 (Sugahara et al., 2003), BDNF for
p75 (Kenchappa et al., 2006), and VEGF-A for VEGF receptor
(Swendeman et al., 2008). Recently, it was reported that NRXs
undergo proteolytic processing, which is augmented by gluta-
mate treatment (Bot et al., 2011; Saura et al., 2011). We also
observed the metalloprotease-dependent production of soluble
forms of endogenous NRXs in rat primary neurons (Figure 6A).
Treatment with TAPI2 or GM6001 caused the accumulation of
NRX-FL and inhibited the accumulation of NRX-CTF, which
was detected upon DAPT treatment (Figure 6B).These data indi-
cate that NRXs also are sequentially cleaved by metalloprotease
and g-secretase in primary neurons. To investigate whether the
binding of NRX regulates the production of sNLG1, we coincu-
bated primary neurons with the conditioned media of HEK293T
Figure 6. Increased Shedding of NLG1 by
Soluble NRXs Derived by Proteolytic Pro-
well as cell lysates of primary neurons with indi-
cated compounds. Antibody against extracellular
(A) or intracellular (B) domain of NRX was used.
(C) Metalloprotease-dependent secretion of re-
combinant soluble NRXs from HEK293T stably
expressing NRX1a-dsRed or NRX1b-dsRed.
(D) DIV11 primary neurons from E18 rat were
cultured for 24 hr in the presence of conditioned
media derived from HEK293T stably expressing
of sNLG1 (white arrowhead) and NLG1-FL (black
arrowhead) were analyzed by immunoblotting.
(E) Densitometric analysis of sNLG1 in (D) (n = 5–6,
mean ± SEM; **p < 0.01; ***p < 0.001 versus EGFP
by Student’s t test).
cells expressing NRX1a or NRX1b, which
contained soluble forms of the NRXs(Fig-
ure 6C). Accumulation of NRX1b immu-
noreactivity at endogenous NLG1 puncta
was observed in rat primary neurons
treated with the HEK293T conditioned
media containing sNRX1b, suggesting
that recombinant sNRX1b is capable of
interacting with NLG1 at synapses (Fig-
ure S4). Intriguingly, release of sNLG1
from neurons was significantly increased
byaddition ofthesoluble NRX-containing
media (Figures 6D and 6E). This result
indicates that ligand binding at the cell
surface regulates the shedding of NLG1.
We also analyzed the activity-depen-
dent NLG1 processing in vivo. Pilocar-
pine treatment induces glutamate-mediated synaptic activation,
resulting in status epilepticus associated with synapse remodel-
ing (Isokawa, 1998; Kurz et al., 2008). In agreement with the
previous reports (Kamenetz et al., 2003), APP processing was
promoted in the brains of 8-week-old epileptic mice (Figure 7A).
Moreover, the level of sNLG1 was significantly increased,
whereas that of the membrane-associated NLG1-FL was
decreased, suggesting that NLG1 shedding was augmented in
brains by pilocarpine-induced seizures (Figure 7B). Taken
by the excitatory activity in vivo as well as in vitro.
The Effect of NLG1 Processing on the Spine Density
To analyze the functional impact of NLG1 processing on its
spinogenic activity, we overexpressed NLG1 and its derivatives
in dentate granule cells of the organotypic hippocampal slice
culture obtained from P6 rat, in which local-circuit synaptic
interactions are preserved. Overexpression of NLG1-FL signifi-
cantly increased the spine density at the apical dendrites of
granule cells. However, NLG2-FL failed to induce spines, sug-
gesting that NLG1 specifically increased the spine density at
Proteolytic Regulation of Neuroligin-1
Neuron 76, 410–422, October 18, 2012 ª2012 Elsevier Inc. 417
glutamatergic synapses as previously described (Figure 8A)
(Scheiffele et al., 2000; Graf et al., 2004). Overexpression of
NLG1DPDZ that lacks the C terminus failed to increase the spine
number,suggesting thatthespinogenic effectofNLG1isdepen-
dent on the PDZ-binding motif in rat dentate granule cells.
Reduction in the amount of transfected NLG1 cDNA led to loss
of the spinogenic effect of NLG1 (see 0.1 mg HA-NLG1, Figures
8A and 8B), indicating that the protein level of NLG1 is critical
to the de novo formation of the dendritic spine (Figure 8B).
Notably, TAPI2 treatment of cultures transfected with reduced
effect on spine density to a level comparable to that in cells
transfected with 1.0 mg/ml of HA-NLG1. Thus, we reasoned that
ectodomain shedding negatively regulates the spinogenic effect
of NLG1 in hippocampal granule cells. Next, we analyzed the
effects of fragment forms of NLG1 corresponding to its proteo-
lytic products (i.e., NLG1-DE and NLG1-ICD) on the spine
density at a similar level to NLG1-FL, suggesting that the NLG1-
CTF lacking the ectodomain retains the spinogenic effect.
However, NLG1-ICD failed to increase the spine density. Thus,
the function of membrane-tethered form of NLG1-ICD (aka,
Figure 7. Status Epilepticus Induced by Pilocarpine Promoted the
Shedding of NLG1 In Vivo
(A) Immunoblot analysis of TS soluble (soluble fraction, left) and insoluble
(membrane fraction, right) of cortices from 8-week-old mice with status epi-
lepticus (SE) 1 hr after the injection of pilocarpine. Mice injected with saline
were used as control (ctrl). NLG1, APP, and their derivatives were probed with
relevant antibodies. bIII-tubulin, calnexin, and nicastrin were used as loading
controls for each fraction, respectively. A representative set of immunoblot
data is shown.
(B) Densitometric analysis of immunoblots. Protein levels in SE brain were
Figure 8. Effect of Truncated NLG1 on Spine Formation
(A) Representative confocal images of apical dendrites visualizedby GFP inrat
hippocampal slice cultures prepared from P6 rats transfected with GFP with
NLG1 constructs. We used 1 mg of plasmids for recombinant proteins unless
the amount is indicated. White bar represents 1 mm.
(B) Quantification of the spine density of neurons in the indicated conditions.
Statistical analysis was carried out by Dunnett’s multiple comparison test
(n = 6–35, mean ± SEM, *p < 0.01 versus mock).
(C) Representative confocal images of immunostained dendrites in rat
hippocampal primary neurons obtained from E17–E18 pups transfected with
GFP and WT or PKQQ_AAAA mutant NLG1. Transfection into rat primary
neuronswasperformed atDIV6andfixed atDIV20.Whitebarrepresents5mm.
(D) Quantification of the spine density of neurons under the indicated condi-
tions. Statistical analysis was carried out by Dunnett’s multiple comparison
test (n = 14 [mock], 55 [WT], and 42 [PKQQ_AAAA]). Mean ± SEM, *p < 0.05
Proteolytic Regulation of Neuroligin-1
418 Neuron 76, 410–422, October 18, 2012 ª2012 Elsevier Inc.
NLG1-DE or NLG1-CTF) was abolished by liberation from the
membrane by the g-secretase cleavage and subsequent degra-
dation. Finally, to directly test whether NLG1 shedding modu-
lates the spinogenic function, we analyzed the dendritic spines
of transfected rat hippocampal primary neurons obtained from
E18 pups (Figure 8C). We transfected wild-type or PKQQ/
AAAA mutant NLG1 together with green fluorescent protein
(GFP) into primary neurons at DIV6 and fixed them at DIV20.
The numbers of spines in neurons expressing wild-type NLG1
showed an increased trend compared to those in mock-trans-
fected neurons, but not with a statistical significance. However,
the spine density was significantly increased in neurons trans-
fected with the mutant NLG1 (Figure 8D), suggesting that
cleavage-deficient mutation enhanced the NLG1 function in
primary neurons. Taken together, our results indicate that the
sequential processing of NLG1 negatively regulates the spino-
Sequential Proteolytic Processing of NLG1 by ADAM10
To date, all known g-secretase substrates are shown to be first
shed at the extracellular domain to generate a soluble ectodo-
main as well as a membrane-tethered CTF. ADAM10 is a well-
characterized physiological sheddase for a number of g-secre-
tase substrates (e.g., APP, cadherin, and Notch) (Reiss et al.,
2005; Jorissen et al., 2010; Kuhn et al., 2010). Both g-secretase
and ADAM10 have been implicated in the regulation of neural
stem cell number by modulation of Notch signaling in the devel-
oping CNS (Jorissen et al., 2010). Recently, it was shown that
metalloprotease and g-secretase-mediated cleavage in mature
neurons regulates the synaptic function (Rivera et al., 2010; Res-
tituito et al.,2011). Herewe systematically analyzed the process-
ing of NLG1 by pharmacological and genetic approaches. Using
ADAM10 is responsible for NLG1 shedding and that C-terminal
Notably, significant reduction in the sNLG1 production was simi-
larly observed in two distinct lines of Adam10flox/floxmice (i.e.,
exon 1 floxed mice for in vitro Cre-mediated gene excision in
that ADAM10 is a major sheddase for NLG1 in vivo in brains. In
addition, Adam10-dependent sNLG1 production and NLG1
accumulation were observed in primary neurons as well as in
adult mouse brains, suggesting that NLG1 is shed by ADAM10
at both developmental and mature stages in neurons. Our data
unequivocally indicate that the cell surface level of NLG1 is regu-
lated by ADAM10/g-secretase-mediated sequential processing,
which may in turn negatively modulate its spinogenic activity.
It is noteworthy that ADAM10 prefers Leu, Phe, Tyr, and Gln at
P10position for cleavage (Caescu et al., 2009), although no
consensus cleavage sequence has been reported. Our observa-
tion that shedding of NLG1 was inhibited in PKQQ/AAAA mutant
suggests that the Gln680or Gln681at the stalk region of NLG1 is
the candidate cleavage site for ADAM10-mediated shedding.
Unexpectedly, we found that NLG2 was not a suitable substrate
for ADAMs so far examined. This is consistent with the previous
results that ADAM10 is localized at the excitatory postsynapses
at which NLG1 is present (Marcello et al., 2007), whereas NLG2
resides in the GABAergic postsynapses (Graf et al., 2004).
Indeed, primary amino acid sequence of the stalk region of
NLG2 is totally different from that of NLG1 (Figure 3A). Thus,
othermetalloprotease(s) presentintheinhibitory synapseshould
be responsible for NLG2 shedding. Intriguingly, the expression
levels of NLG1, but not NLG2, was significantly increased in
the brains of ADAM10 transgenic mice, suggesting a specific
functional correlation between NLG1 and ADAM10 (Prinzen
et al., 2009). Identification of the responsible proteases and rele-
vant auxiliary components at different types of synapses would
provide important information on the proteolytic control of
neuronal adhesion molecules.
Physiological Significance of the Activity-Dependent
Processing of NLG1
The level of NLG1 in neurons has been shown to regulate the
number, ratio of NMDA/AMPA receptors, and electrophysiolog-
ical functionsof theexcitatory synapsesin vitro andin vivo(Song
et al., 1999; Chih et al., 2006; Varoqueaux et al., 2006; Chubykin
et al., 2007). Here, we show that NLG1 is cleaved in a neuronal
activity-dependent manner, resulting in a loss of its spinogenic
function. Moreover, pretreatment with MK-801 completely abol-
ished the processing of NLG1 induced by glutamate, suggesting
that the NLG1 level is homeostatically controlled by the excit-
atory synaptic, but not extrasynaptic, transmission. Increased
shedding of NLG1 was also observed in pilocarpine-treated
mice. Interestingly, profound decreases in the density, as well
as alterations in shape and size, of dendritic spines by aberrant
Ca2+signaling have been observed in epileptic mouse models
(Isokawa, 1998; Kochan et al., 2000; Kurz, et al., 2008). Aberrant
Ca2+signaling also affects ADAM10 activity via calmodulin
kinase as well as calcineurin (Nagano et al., 2004; Kohutek
et al., 2009). These results support the idea that NLG1 process-
ing is involved in the remodeling of dendritic spines at glutama-
tergic synapses in vivo. We also found that soluble NRX
treatment augments the NLG1 shedding. In this regard, the
activity-dependent proteolytic cleavage of NRX at the presy-
napse (Bot et al., 2011; Saura et al., 2011) may be functionally
linked to the processing of NLG. It has been reported that
Fc-fused recombinant NRX extracellular domain inhibited the
synaptogenic activity of NLG (Scheiffele et al., 2000; Levinson
et al., 2005); it is possible that soluble NRX functions as a nega-
tiveregulator of NLG1via induction of shedding.Ligand-induced
shedding has been reported in several g-secretase substrates
(Mumm et al., 2000; Sugahara et al., 2003; Kenchappa et al.,
2006; Findley et al., 2007) too, although the molecular mecha-
nisms whereby the ligands activate the processing remain
unknown. Ligand/receptor complex formation has been shown
to increase the ADAM activity in the ephrin/Eph receptor system
(Janes et al., 2009). In the case of Notch, ‘‘pulling’’ movement
induced by the endocytosis of bound ligand is thought to cause
a structural change leading to cleavage (Gordon et al., 2007).
Notably, mucin-like O-linked glycosylation, which might create
steric hindrance against ADAM10, was identified in the juxta-
membranous stalk region of NLG1 (i.e., Ser683and Ser686,
respectively; Hoffman et al., 2004). We also have found that
Proteolytic Regulation of Neuroligin-1
Neuron 76, 410–422, October 18, 2012 ª2012 Elsevier Inc. 419
amino acid substitutions including the O-glycosylation sites (i.e.,
PSPF/AAAA and SVDQ/AAAA mutants) increased the shedding
of NLG1 (Figure 4C). Thus, it is possible that the binding of
soluble NRX induces structural changes in the stalk region of
NLG1 in a way to expose the cleavage site and/or activate
the ADAM10 activity. Further studies on the mechanism of
site would clarify the mechanism of NLG1 shedding.
What, then, is the physiological function of the NLG1 frag-
ments? Extracellular domain of NLG1 is sufficient for binding its
ligands (Ichtchenko et al., 1995). Intriguingly, soluble form of
NLG1 has been shown to inhibit the synaptogenic effect of
TSP1 in immature neurons (Xu et al., 2010). Moreover, clustering
assembly of presynaptic terminals (Dean et al., 2003), raising the
possibility that sNLG1 may bind to soluble as well as membrane-
tethered forms of ligands and modulate their functions. Unex-
pectedly, however, overexpression of NLG1-DE that lacks the
extracellular domain retained the capacity to induce dendritic
spines in granule cells. Our observation is consistent with the
previous data showing that the conserved cytoplasmic domain,
2009; Shipman et al., 2011). Importantly, however, overexpres-
sion of NLG1-ICD, which retains the intact intracellular domain,
failed to increase the spine numbers. Moreover, NLG1-ICD was
highly labile and degraded by proteasomal activity. These data
indicate that the membrane tethering as well as the stability of
the cytoplasmic domain of NLG1 is critical to the spinogenic
activity. This is distinct from the activation of the ‘‘conventional’’
g-secretase substrates, e.g., Notch, by g-cleavage, although
tions of g-secretase substrates by cleavage, e.g., ephrin-B1 and
DCC (Tomita et al., 2006; Parent et al., 2005).
Taken together, our present results provide compelling evi-
dence that proteolytic processing is a molecular mechanism
regulating the NLG1 levels as well as its spinogenic function.
Further functional analysis would be required to determine
whether spines modulated by NLG1 shedding are functional.
However, previous results showing that changes in spines by
overexpression or knockdown of NLG1 correlated with synaptic
transmission (Chih et al., 2006; Levinson et al., 2005; Chubykin
et al., 2007) may support our view that the proteolytic cleavage
by ADAM10 and g-secretase downregulates the cell-surface
levels of NLG1, which in turn negatively affects the synaptogenic
function. Considering the recent implication of aberrant levels
that alterations in the proteolytic processing of NLG1 may
also be involved in the etiology of the neurodevelopmental
Chemicals, Immunological Methods, Animals, Plasmids, Cell
Culture, Transfection, Recombinant Virus Infection, and RNA
All experimental procedures were performed in accordance with the
guidelines for animal experiments of the University of Tokyo. Primary neuron
culture, immunoblot analyses, and immunocytochemistry experiments were
performed as previously described with some modifications (Tomita et al.,
1998; Fukumoto et al., 1999). For in vitro Cre-mediated Adam10 ablation,
primary cortical neurons were obtained from E16 pups of Adam10flox/flox
mice, in which the first exon was floxed (Yoda et al., 2011). For analysis of
neuron-specific conditional Adam10 knockout mice, brains of P18 exon 2
floxed Adam10flox/floxmice (Jorissen et al., 2010) crossed with CamKII-Cre
mice (J.P. and P.S., unpublished data) were homogenized to obtain micro-
some fractions. Other animals were obtained from Japan-SLC. See Supple-
mental Experimental Procedures for details.
Mouse Model of Status Epileptics by Pilocarpine Treatment
Male 8-week-old BALB/C mice were injected with scopolamine methylnitrate
(Tokyo Chemical Industry) (1 mg/kg, intraperitoneally [i.p.]) to protect against
peripheral autonomic effects caused by subsequent pilocarpine administra-
tion. Fifteen to thirty minutes later, mice were injected with pilocarpine-HCl
(SIGMA) (330–380 mg/kg, i.p.) or saline (Otsuka), and then we scored the
seizure intensity according to a previously described method (Patel et al.,
1988). We defined status epilepticus (status epilepsy) as a continuous seizure
lasting longer than 30 min. One hour after the injection of pilocarpine, mice
were sacrificed to isolate the cerebrums. See Supplemental Experimental
Procedures for details.
Analysis of Spinogenic Function of NLG1
Hippocampal slice cultures were prepared from P6 rats as previously
described (Koyama et al., 2007). Granule cells in the cultured slices at DIV5
were transfected with the plasmids encoding NLG1 and its derivatives
(1.0 mg/ml in HBSS) using the single-cell electroporation method (Nakahara
et al., 2009). Transfection of mutant NLG1 in rat hippocampal primary neurons
were performed at DIV6 and fixed at DIV20. See Supplemental Experimental
Supplemental Information includes four figures and Supplemental Experi-
mental Procedures and can be found with this article online at http://dx.doi.
We thank Drs. C. Blobel (Hospital for Special Surgery, New York), R. Balice-
Gordon (University of Pennsylvania), P. Scheiffele (University of Basel), B. De
Strooper (VIB Leuven), K. Hozumi (Tokai University), F. Fahrenholtz (Johannes
Gutenberg University Mainz), T. Kitamura (The University of Tokyo), and J. Ta-
kagi (Osaka University) for materials. We are also grateful to our laboratory
members for helpful discussions and technicalassistance. Thiswork was sup-
ported by Grants-in-Aid for Young Scientists (S) from Japan Society for the
Promotion of Science (JSPS) (for T.T.), Challenging Exploratory Research
from JSPS (for T.T.), Scientific Research on Innovative Areas ‘‘Foundation of
Synapse and Neurocircuit Pathology’’ from the Ministry of Education, Culture,
Sports, Science, and Technology (MEXT) (for T.T. and T.I.), the Cell Science
Research Foundation (for T.T.), Core Research for Evolutional Science and
Technology of the Japan Science and Technology Agency (for Y.H., T.T.,
and T.I.), Japan, and the Deutsche Forschungsgemeinschaft SFB877 TP:A3
(for P.S.). K.S. is a research fellow of JSPS.
Accepted: October 2, 2012
Published: October 18, 2012
Bai, G., and Pfaff, S.L. (2011). Protease regulation: the Yin and Yang of neural
development and disease. Neuron 72, 9–21.
Beel, A.J., and Sanders, C.R. (2008). Substrate specificity of g-secretase and
other intramembrane proteases. Cell. Mol. Life Sci. 65, 1311–1334.
Blundell, J., Blaiss, C.A., Etherton, M.R., Espinosa, F., Tabuchi, K., Walz, C.,
Bolliger, M.F., Su ¨dhof, T.C., and Powell, C.M. (2010). Neuroligin-1 deletion
Proteolytic Regulation of Neuroligin-1
420 Neuron 76, 410–422, October 18, 2012 ª2012 Elsevier Inc.
results in impaired spatial memory and increased repetitive behavior.
J. Neurosci. 30, 2115–2129.
Bot, N., Schweizer, C., Ben Halima, S., and Fraering, P.C. (2011). Processing
of the synaptic cell adhesion molecule neurexin-3b by Alzheimer disease a-
and g-secretases. J. Biol. Chem. 286, 2762–2773.
Bottos, A., Rissone, A., Bussolino, F., and Arese, M. (2011). Neurexins and
neuroligins: synapses look out of the nervous system. Cell. Mol. Life Sci. 68,
Budreck, E.C., and Scheiffele, P. (2007). Neuroligin-3 is a neuronal adhesion
protein at GABAergic and glutamatergic synapses. Eur. J. Neurosci. 26,
Caescu, C.I., Jeschke, G.R., and Turk, B.E. (2009). Active-site determinants of
substrate recognition by the metalloproteinases TACE and ADAM10.
Biochem. J. 424, 79–88.
Chih, B., Gollan, L., and Scheiffele, P. (2006). Alternative splicing controls
selective trans-synaptic interactions of the neuroligin-neurexin complex.
Neuron 51, 171–178.
Chubykin, A.A., Atasoy, D., Etherton, M.R., Brose, N., Kavalali, E.T., Gibson,
J.R., and Su ¨dhof, T.C. (2007). Activity-dependent validation of excitatory
versus inhibitory synapses by neuroligin-1 versus neuroligin-2. Neuron 54,
Comoletti, D., De Jaco, A., Jennings, L.L., Flynn, R.E., Gaietta, G., Tsigelny, I.,
Ellisman, M.H., and Taylor, P. (2004). The Arg451Cys-neuroligin-3 mutation
associated with autism reveals a defect in protein processing. J. Neurosci.
Dahlhaus, R., Hines, R.M., Eadie, B.D., Kannangara, T.S., Hines, D.J., Brown,
C.E., Christie, B.R., and El-Husseini, A. (2010). Overexpression of the cell
adhesion protein neuroligin-1 induces learning deficits and impairs synaptic
plasticity by altering the ratio of excitation to inhibition in the hippocampus.
Hippocampus 20, 305–322.
Dalva, M.B., McClelland, A.C., and Kayser, M.S. (2007). Cell adhesion mole-
cules: signalling functions at the synapse. Nat. Rev. Neurosci. 8, 206–220.
Dean, C., Scholl, F.G., Choih, J., DeMaria, S., Berger, J., Isacoff, E., and
Scheiffele, P. (2003). Neurexin mediates the assembly of presynaptic termi-
nals. Nat. Neurosci. 6, 708–716.
Findley, C.M., Cudmore, M.J., Ahmed, A., and Kontos, C.D. (2007). VEGF
induces Tie2 shedding via a phosphoinositide 3-kinase/Akt dependent
pathway to modulate Tie2 signaling. Arterioscler. Thromb. Vasc. Biol. 27,
Fukumoto, H., Tomita, T., Matsunaga, H., Ishibashi, Y., Saido, T.C., and
Iwatsubo, T. (1999). Primary cultures of neuronal and non-neuronal rat brain
cells secrete similar proportions of amyloid b peptides ending at Ab40 and
Ab42. Neuroreport 10, 2965–2969.
Glessner, J.T., Wang, K., Cai, G., Korvatska, O., Kim, C.E., Wood, S., Zhang,
H., Estes, A., Brune, C.W., Bradfield, J.P., et al. (2009). Autism genome-wide
copy number variation reveals ubiquitin and neuronal genes. Nature 459,
Gordon, W.R., Vardar-Ulu, D., Histen, G., Sanchez-Irizarry, C., Aster, J.C., and
Blacklow, S.C. (2007). Structural basis for autoinhibition of Notch. Nat. Struct.
Mol. Biol. 14, 295–300.
Graf, E.R., Zhang, X., Jin, S.X., Linhoff, M.W., and Craig, A.M. (2004).
Neurexins induce differentiation of GABA and glutamate postsynaptic special-
izations via neuroligins. Cell 119, 1013–1026.
Hartmann, D., de Strooper, B., Serneels, L., Craessaerts, K., Herreman, A.,
Annaert, W., Umans, L., Lu ¨bke, T., Lena Illert, A., von Figura, K., and Saftig,
P. (2002). The disintegrin/metalloprotease ADAM 10 is essential for Notch
signalling but not for a-secretase activity in fibroblasts. Hum. Mol. Genet.
Herreman, A., Serneels, L., Annaert, W., Collen, D., Schoonjans, L., and De
Strooper, B. (2000). Total inactivation of g-secretase activity in presenilin-
deficient embryonic stem cells. Nat. Cell Biol. 2, 461–462.
Hoffman, R.C., Jennings, L.L., Tsigelny, I., Comoletti, D., Flynn, R.E., Sudhof,
T.C., and Taylor, P. (2004). Structural characterization of recombinant soluble
rat neuroligin 1: mapping of secondary structure and glycosylation by mass
spectrometry. Biochemistry 43, 1496–1506.
Horiuchi, K., Miyamoto, T., Takaishi, H., Hakozaki, A., Kosaki, N., Miyauchi, Y.,
Furukawa, M., Takito, J., Kaneko, H., Matsuzaki, K., et al. (2007). Cell surface
colony-stimulating factor 1 can be cleaved by TNF-a converting enzyme or
endocytosed in a clathrin-dependent manner. J. Immunol. 179, 6715–6724.
Huettner, J.E., and Bean, B.P. (1988). Block of N-methyl-D-aspartate-
activated current by the anticonvulsant MK-801: selective binding to open
channels. Proc. Natl. Acad. Sci. USA 85, 1307–1311.
Ichtchenko, K., Hata, Y., Nguyen, T., Ullrich, B., Missler, M., Moomaw, C., and
Su ¨dhof, T.C. (1995). Neuroligin 1: a splice site-specific ligand for b-neurexins.
Cell 81, 435–443.
Inoue, E., Deguchi-Tawarada, M., Togawa, A., Matsui, C., Arita, K., Katahira-
Tayama, S., Sato, T., Yamauchi, E., Oda, Y., and Takai, Y. (2009). Synaptic
activity prompts g-secretase-mediated cleavage of EphA4 and dendritic spine
formation. J. Cell Biol. 185, 551–564.
Irie, M., Hata, Y., Takeuchi, M., Ichtchenko, K., Toyoda, A., Hirao, K., Takai, Y.,
Rosahl, T.W., and Su ¨dhof, T.C. (1997). Binding of neuroligins to PSD-95.
Science 277, 1511–1515.
temporal lobe epilepsy. Neurosci. Lett. 258, 73–76.
Griesshaber, B., Sabet, O., Grabenbauer, M., Ting, A.Y., Saftig, P.,
Bastiaens, P.I., and Lackmann, M. (2009). Cytoplasmic relaxation of active
Eph controls ephrin shedding by ADAM10. PLoS Biol. 7, e1000215.
P.W.,Wimmer-Kleikamp, S.H., Frangakis,A.S., Treble,K.,
Jorissen, E., Prox, J., Bernreuther, C., Weber, S., Schwanbeck, R., Serneels,
L., Snellinx, A., Craessaerts, K., Thathiah, A., Tesseur, I., et al. (2010). The
disintegrin/metalloproteinase ADAM10 is essential for the establishment of
the brain cortex. J. Neurosci. 30, 4833–4844.
Kamenetz, F., Tomita, T., Hsieh, H., Seabrook, G., Borchelt, D., Iwatsubo, T.,
Sisodia, S., and Malinow, R. (2003). APP processing and synaptic function.
Neuron 37, 925–937.
Kawaguchi, N., Horiuchi, K., Becherer, J.D., Toyama, Y., Besmer, P., and
Blobel, C.P. (2007). Different ADAMs have distinct influences on Kit ligand
processing: phorbol-ester-stimulated ectodomain shedding of Kitl1 by
ADAM17 is reduced by ADAM19. J. Cell Sci. 120, 943–952.
Kenchappa, R.S., Zampieri, N., Chao, M.V., Barker, P.A., Teng, H.K.,
Hempstead, B.L., and Carter, B.D. (2006). Ligand-dependent cleavage of
the P75 neurotrophin receptor is necessary for NRIF nuclear translocation
and apoptosis in sympathetic neurons. Neuron 50, 219–232.
Kim, S.H., Fraser, P.E., Westaway, D., St George-Hyslop, P.H., Ehrlich, M.E.,
and Gandy, S. (2010). Group II metabotropic glutamate receptor stimulation
triggers production and release of Alzheimer’s amyloid(b)42 from isolated
intact nerve terminals. J. Neurosci. 30, 3870–3875.
Ko, J., Zhang, C., Arac, D., Boucard, A.A., Brunger, A.T., and Su ¨dhof, T.C.
(2009). Neuroligin-1 performs neurexin-dependent and neurexin-independent
functions in synapse validation. EMBO J. 28, 3244–3255.
Kochan, L.D., Churn, S.B., Omojokun, O., Rice, A., and DeLorenzo, R.J.
(2000). Status epilepticus results in an N-methyl-D-aspartate receptor-depen-
dent inhibition of Ca2+/calmodulin-dependent kinase II activity in the rat.
Neuroscience 95, 735–743.
Kohutek, Z.A., diPierro, C.G., Redpath, G.T., and Hussaini, I.M. (2009). ADAM-
10-mediated N-cadherin cleavage is protein kinase C-a dependent and
promotes glioblastoma cell migration. J. Neurosci. 29, 4605–4615.
Koyama, R., Muramatsu, R., Sasaki, T., Kimura, R., Ueyama, C., Tamura, M.,
Tamura, N., Ichikawa, J., Takahashi, N., Usami, A., et al. (2007). A low-cost
method for brain slice cultures. J. Pharmacol. Sci. 104, 191–194.
Kuhn, P.H., Wang, H., Dislich, B., Colombo, A., Zeitschel, U., Ellwart, J.W.,
Kremmer, E., Rossner, S., and Lichtenthaler, S.F. (2010). ADAM10 is the phys-
iologically relevant, constitutive a-secretase of the amyloid precursor protein
in primary neurons. EMBO J. 29, 3020–3032.
Proteolytic Regulation of Neuroligin-1
Neuron 76, 410–422, October 18, 2012 ª2012 Elsevier Inc. 421
Kurz, J.E., Moore, B.J., Henderson, S.C., Campbell, J.N., and Churn, S.B.
(2008). A cellular mechanism for dendritic spine loss in the pilocarpine model
of status epilepticus. Epilepsia 49, 1696–1710.
Levinson, J.N., and El-Husseini, A. (2007). A crystal-clear interaction: relating
neuroligin/neurexin complex structure to function at the synapse. Neuron 56,
Levinson, J.N., Che ´ry, N., Huang, K., Wong, T.P., Gerrow, K., Kang, R.,
Prange, O., Wang, Y.T., and El-Husseini, A. (2005). Neuroligins mediate excit-
atory and inhibitory synapse formation: involvement of PSD-95 and neurexin-
1b in neuroligin-induced synaptic specificity. J. Biol. Chem. 280, 17312–
Marambaud, P., Wen, P.H., Dutt, A., Shioi, J., Takashima, A., Siman, R., and
Robakis, N.K. (2003). A CBP binding transcriptional repressor produced by
the PS1/ε-cleavage of N-cadherin is inhibited by PS1 FAD mutations. Cell
Marcello, E., Gardoni, F., Mauceri, D., Romorini, S., Jeromin, A., Epis, R.,
Borroni, B., Cattabeni, F., Sala, C., Padovani, A., and Di Luca, M. (2007).
Synapse-associated protein-97 mediates a-secretase ADAM10 trafficking
and promotes its activity. J. Neurosci. 27, 1682–1691.
Mumm, J.S., Schroeter, E.H., Saxena, M.T., Griesemer, A., Tian, X., Pan, D.J.,
Ray, W.J.,and Kopan, R.(2000). Aligand-inducedextracellularcleavageregu-
lates g-secretase-like proteolytic activation of Notch1. Mol. Cell 5, 197–206.
T., Nakajima, M., Shinohara, M., and Saya, H. (2004). Cell-matrix interaction
via CD44 is independently regulated by different metalloproteinases activated
in response to extracellular Ca(2+) influx and PKC activation. J. Cell Biol. 165,
Nakahara, S., Tamura, M., Matsuki, N., and Koyama, R. (2009). Neuronal
hyperactivity sustains the basal dendrites of immature dentate granule cells:
Hippocampus 19, 379–391.
Parent, A.T., Barnes, N.Y., Taniguchi, Y., Thinakaran, G., and Sisodia, S.S.
(2005). Presenilin attenuates receptor-mediated signaling and synaptic func-
tion. J. Neurosci. 25, 1540–1549.
Patel, S., Meldrum, B.S., and Fine, A. (1988). Susceptibility to pilocarpine-
induced seizures in rats increases with age. Behav. Brain Res. 31, 165–167.
Prinzen, C., Tru ¨mbach, D., Wurst, W., Endres, K., Postina, R., and Fahrenholz,
F. (2009). Differential gene expression in ADAM10 and mutant ADAM10 trans-
genic mice. BMC Genomics 10, 66.
Reiss, K., Maretzky, T., Ludwig, A., Tousseyn, T., de Strooper, B., Hartmann,
D., and Saftig, P. (2005). ADAM10 cleavage of N-cadherin and regulation of
cell-cell adhesion and b-catenin nuclear signalling. EMBO J. 24, 742–752.
Restituito, S., Khatri, L., Ninan, I., Mathews, P.M., Liu, X., Weinberg, R.J., and
Ziff, E.B. (2011). Synaptic autoregulation by metalloproteases and g-secre-
tase. J. Neurosci. 31, 12083–12093.
Rivera, S., Khrestchatisky, M., Kaczmarek, L., Rosenberg, G.A., and Jaworski,
D.M. (2010). Metzincin proteases and their inhibitors: foes or friends innervous
system physiology? J. Neurosci. 30, 15337–15357.
Saftig, P., and Reiss, K. (2011). The ‘‘A Disintegrin And Metalloproteases’’
ADAM10 and ADAM17: novel drug targets with therapeutic potential? Eur. J.
Cell Biol. 90, 527–535.
Saura, C.A., Servia ´n-Morilla, E., and Scholl, F.G. (2011). Presenilin/g-secre-
tase regulates neurexin processing at synapses. PLoS ONE 6, e19430.
Schapitz, I.U., Behrend, B., Pechmann, Y., Lappe-Siefke, C., Kneussel, S.J.,
Wallace, K.E., Stempel, A.V., Buck, F., Grant, S.G., Schweizer, M., et al.
(2010). Neuroligin 1 is dynamically exchanged at postsynaptic sites.
J. Neurosci. 30, 12733–12744.
Scheiffele, P., Fan, J., Choih, J., Fetter, R., and Serafini, T. (2000). Neuroligin
expressed in nonneuronal cells triggers presynaptic development in contact-
ing axons. Cell 101, 657–669.
Shipman, S.L., Schnell, E., Hirai, T., Chen, B.S., Roche, K.W., and Nicoll, R.A.
(2011). Functional dependence of neuroligin on a new non-PDZ intracellular
domain. Nat. Neurosci. 14, 718–726.
Siddiqui, T.J., and Craig, A.M. (2011). Synaptic organizing complexes. Curr.
Opin. Neurobiol. 21, 132–143.
Song, J.Y., Ichtchenko, K., Su ¨dhof, T.C., and Brose, N. (1999). Neuroligin 1 is
a postsynaptic cell-adhesion molecule of excitatory synapses. Proc. Natl.
Acad. Sci. USA 96, 1100–1105.
Su ¨dhof, T.C. (2008). Neuroligins and neurexins link synaptic function to cogni-
tive disease. Nature 455, 903–911.
Sugahara, K.N., Murai, T., Nishinakamura, H., Kawashima, H., Saya, H., and
Miyasaka, M. (2003). Hyaluronan oligosaccharides induce CD44 cleavage
and promote cell migration in CD44-expressing tumor cells. J. Biol. Chem.
Swendeman, S., Mendelson, K., Weskamp, G., Horiuchi, K., Deutsch, U.,
Scherle, P., Hooper, A., Rafii, S., and Blobel, C.P. (2008). VEGF-A stimulates
ADAM17-dependent shedding of VEGFR2 and crosstalk between VEGFR2
and ERK signaling. Circ. Res. 103, 916–918.
Thyagarajan, A., and Ting, A.Y. (2010). Imaging activity-dependent regulation
of neurexin-neuroligin interactions using trans-synaptic enzymatic bio-
tinylation. Cell 143, 456–469.
Tomita, T., Tokuhiro, S., Hashimoto, T., Aiba, K., Saido, T.C., Maruyama, K.,
and Iwatsubo, T. (1998). Molecular dissection of domains in mutant presenilin
2 that mediate overproduction of amyloidogenic forms of amyloid b peptides.
increase secretion of Ab42. J. Biol. Chem. 273, 21153–21160.
Tomita, T., Tanaka, S., Morohashi, Y., and Iwatsubo, T. (2006). Presenilin-
dependent intramembrane cleavage of ephrin-B1. Mol. Neurodegener. 1, 2.
Varoqueaux, F., Jamain, S., and Brose, N. (2004). Neuroligin 2 is exclusively
localized to inhibitory synapses. Eur. J. Cell Biol. 83, 449–456.
Varoqueaux, F., Aramuni, G., Rawson, R.L., Mohrmann, R., Missler, M.,
Gottmann, K., Zhang, W., Su ¨dhof, T.C., and Brose, N. (2006). Neuroligins
determine synapse maturation and function. Neuron 51, 741–754.
Villasana, L.E., Klann, E., and Tejada-Simon, M.V. (2006). Rapid isolation of
synaptoneurosomes and postsynaptic densities from adult mouse hippo-
campus. J. Neurosci. Methods 158, 30–36.
Weskamp, G., Ford, J.W., Sturgill, J., Martin, S., Docherty, A.J., Swendeman,
S., Broadway, N., Hartmann, D., Saftig, P., Umland, S., et al. (2006). ADAM10
is a principal ‘sheddase’ of the low-affinity immunoglobulin E receptor CD23.
Nat. Immunol. 7, 1293–1298.
Witters, L., Scherle, P., Friedman, S., Fridman, J., Caulder, E., Newton, R., and
Lipton, A. (2008). Synergistic inhibition with a dual epidermal growth factor
receptor/HER-2/neutyrosinekinase inhibitor andadisintegrin andmetallopro-
tease inhibitor. Cancer Res. 68, 7083–7089.
esis in hippocampal neurons through neuroligin 1. Nat. Neurosci. 13, 22–24.
Yoda, M., Kimura, T., Tohmonda, T., Uchikawa, S., Koba, T., Takito, J.,
Morioka, H., Matsumoto, M., Link, D.C., Chiba, K., et al. (2011). Dual functions
of cell-autonomous and non-cell-autonomous ADAM10 activity in granulopoi-
esis. Blood 118, 6939–6942.
Zhang, C., Milunsky, J.M., Newton, S., Ko, J., Zhao, G., Maher, T.A., Tager-
Flusberg, H., Bolliger, M.F., Carter, A.S., Boucard, A.A., et al. (2009). A neuro-
ligin-4 missense mutation associated with autism impairs neuroligin-4 folding
and endoplasmic reticulum export. J. Neurosci. 29, 10843–10854.
Chiusaroli, R., Hahn, R., Wilkes, D., Fisher, P., et al. (2004). Essential role for
ADAM19 in cardiovascular morphogenesis. Mol. Cell. Biol. 24, 96–104.
Zhou, B.B., Peyton, M., He, B., Liu, C., Girard, L., Caudler, E., Lo, Y., Baribaud,
F., Mikami, I., Reguart, N., et al. (2006). Targeting ADAM-mediated ligand
cleavage to inhibit HER3 and EGFR pathways in non-small cell lung cancer.
Cancer Cell 10, 39–50.
Proteolytic Regulation of Neuroligin-1
422 Neuron 76, 410–422, October 18, 2012 ª2012 Elsevier Inc.