This article appeared in a journal published by Elsevier.

The attached
copy is furnished to the author for internal non-commercial research
and education use, including for instruction at the authors institution
and sharing with colleagues.
Other uses, including reproduction and distribution, or selling or
licensing copies, or posting to personal, institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of the
article (e.g. in Word or Tex form) to their personal website or
institutional repository. Authors requiring further information
regarding Elsevier’s archiving and manuscript policies are
encouraged to visit:
http://www.elsevier.com/copyright
 Author's personal copy

Neuroscience Letters 450 (2009) 163–166

Contents lists available at ScienceDirect

Neuroscience Letters
journal homepage: www.elsevier.com/locate/neulet

TLS interaction with NMDA R1 splice variant in retinal ganglion cell line RGC-5
Widyawilis Selamat a , Ildasolha Jamari a , Yu Wang a , Toru Takumi b , Fulton Wong c , Ristuko Fujii a,∗
a
Waseda-Olympus Bioscience Research Institute, Waseda University Cooperative Bioscience Research Institute, 11 Biopolis Way, Helios #05-01/02, Singapore 138667, Singapore
b
Osaka Bioscience Institute, 6-2-4 Furuedai, Suita, Osaka 565-0874, Japan
c
Duke-NUS Graduate Medical School Singapore, Jalan Bukit Merah, Singapore 169547, Singapore

a r t i c l e i n f o a b s t r a c t

Article history: Translocated in liposarcoma (TLS or FUS) is a multifunctional protein component of the heterogenous
Received 26 September 2008 ribonuclear complex involved in the splicing of pre-mRNA and the export of fully processed mRNA
Received in revised form 9 November 2008 from the nucleus to the cytoplasm. As we determined that TLS was substantially expressed in the adult
Accepted 7 December 2008
retina, we investigated the functions of TLS in a rat retinal ganglion cell (RGC) line RGC-5. TLS was
found to be associated with N-methyl-d-aspartate (NMDA) receptor 1 (NR1) and myosinVa (MyoVa)
Keywords:
in a calcium-dependent manner. We demonstrated that TLS-associated NR1 could be one of the NR1
RNA binding protein
alternative splice variants, NR1-4, which was predominantly expressed in RGC-5. The degree of colocal-
NMDA R1 splice variants
Retinal ganglion cell
ization between TLS and NR1 was significantly decreased by depolarization of RGC-5 cells, indicating that
RGC-5 the depolarization-induced Ca2+ -influx triggered a redistribution of NR1 from the TLS–protein complex.
TLS (translocated in liposarcoma) These results suggested that TLS might be involved in a calcium-dependent trafficking of specific NR1
splice variants in RGCs.
© 2008 Elsevier Ireland Ltd. All rights reserved.

Selective targeting and local translation of mRNAs in the dendritic
spines have been implicated in synapse remodeling and synaptic
plasticity [18]. The mechanism of mRNA transport in the neuronal
dendrites has been well established [8,12,17]. The dendritic mRNAs
are transported as a large RNA–protein complex termed “RNA gran-
ule”, which contains mRNAs, ribosomal RNA and proteins such as
RNA binding proteins or translational factors [8,10,12]. Our recent
study has revealed that TLS [4], an RNA binding protein compo-
nent of the RNA granule, associates with an actin-dependent motor
protein myosinVa (MyoVa) in mouse cortical and hippocampal neu-
rons [7,21]. In addition to the RNA binding ability, TLS is implicated
in NMDA receptor (NR)-multi-protein complex formation [9]. In
the pyramidal neurons of CA1 layer of adult rat hippocampus, a
subset of TLS punctuate clusters has actually been shown to colo-
calize with NMDA receptor 1 (NR1) [1]. However, a physiological
significance of the interaction between TLS and NR1 has remained
unclear.
To date, eight splice variants for NR1 have been reported [15,19].
They are synthesized by alternative NR1 splicing processes the dele-
tion of exon 21 in the C-terminus, the use of alternative acceptor
site resulting in the exclusion of exon 22, the insertion of exon 5 in
the N-terminus domain, or all possible combinations of these splic-
ing events [19,23]. Exons 21 and 22 encode the C-terminus splice
cassettes termed C1 and C2, respectively. The exclusion of exon 22
∗ Corresponding author. Tel.: +65 6478 9721; fax: +65 6478 9416.
E-mail address: fujiir@waseda.jp (R. Fujii).
introduces a new coding region termed C2 . The presence or absence
of the C1 and C2 cassettes, influences membrane trafficking of NR1
[3].
NR1 has been found to be expressed homogenously in the gan-
glion cell layer (GCL) of rodent retina [2]. It has been reported that
NR1 splicing variant NR1-2, which lacks the C1 and contains the
C2, is most abundantly expressed in rat retina GCL whereas NR1-4,
which lacks the C1 and contains the C2 , and NR1 pan type are found
at moderate levels [12]. As we found that TLS was substantially
expressed in mouse retina and RGC-5 cells, here we specifically
address the question of whether TLS is involved in regulation of
the alternative NR1 splicing or trafficking of the NR1 splice variants
at both mRNA and protein levels in RGC-5 cells.
Rat retinal ganglion cell (RGC) lines RGC-5 cells were provided
by Dr. Agarwal (UNT Health Science Center, TX) and cultured as
described previously [13]. Differentiation of RGC-5 was induced as
reported by Frassetto et al. [5].
In this study, the following primary antibodies were used:
mouse anti-TLS monoclonal antibody (N-terminus 1–117 antigen,
BD Biosciences), and rabbit polyclonal antibodies raised against
MyoVa (C-terminus 1782–1799 antigen, Sigma–Aldrich), NR1 (N-
terminus 918–138 antigen, Sigma–Aldrich), and the C-terminus of
rat NR1 C2 splice variant (Millipore). Phosphorylation of proteins
was detected using OMNI-PhosTM Blend rabbit polyclonal antibody
(Chemicon).
Colocalization between TLS and NR1 in RGC-5 cells was assessed
by immunocytochemistry after fixation of the cells in methanol at
−20 ◦ C for 10 min. The immunofluorescent images were acquired by

0304-3940/$ – see front matter © 2008 Elsevier Ireland Ltd. All rights reserved.
doi:10.1016/j.neulet.2008.12.014
 Author's personal copy
164 W. Selamat et al. / Neuroscience Letters 450 (2009) 163–166
Fig. 1. TLS expression in mouse retina and RGC-5 cells: (A) equal amounts (20 ␮g) of
protein lysates prepared from adult mouse retina (retina), RGC-5 and adult mouse
cortex were separated by SDS-PAGE and immunoblotted with anti-TLS monoclonal
(TLS) and anti-actin monoclonal (actin) antibodies. (B) Cryosection of adult mouse
retina was immunostained with anti-TLS antibody. TLS (green) was localized in gan-
glion cell layer (GCL) and inner nuclear layer (INL). Inset: higher magnification view
of GCL. Bar: 50 ␮m, 15 ␮m for inset.
a confocal laser scan microscopy using FV1000 equipped with IX81
microscope and UPlanApo 100× Oil Iris/1.35 objective (Olympus).
Pearson’s correlation coefficient was calculated using FV10-ASW
software (Olympus) to determine the degree of colocalization
between the two fluorescent signals.
C57BL/6 adult mice retinas and cortices, and RGC-5 cells were
homogenized in lysis buffer [1% TritonX-100, 50 mM Tris–HCl
(pH 7.5)] supplemented with protease inhibitors and phos-
phatase inhibitor cocktail (Nacalai Tesque). TLS–protein complex
in the lysate was immunoprecipitated by anti-TLS mouse mono-
clonal antibody (BD Biosciences). Alternatively, MyoVa-associated
proteins were immunoprecipitated using anti-MyoVa antibody
(Sigma–Aldrich). Proteins in the immunocomplexes were eluted in
SDS-PAGE sample buffer and resolved on 10% SDS-PAGE according
to standard protocols. The proteins were analyzed by immunoblot-
ting using the Western LightningTM enhanced chemiluminescent
(ECL) detection system (Perkin-Elmer).
The PFA-fixed mouse eyes were sectioned at 8 ␮m using Cryostat
CM1850 (Leica). The sections were immunostained with anti-TLS
monoclonal antibody (BD Biosciences followed by the indirect
detection of immunoreactive signals using anti-mouse Alexa 488-
conjugated goat IgG (Invitrogen). Immunofluorescent images were
acquired by using FV1000 (Olympus).
Total RNA from mouse retina and cortex, and RGC-5 were
prepared using Sepasol II (Nacalai Tesque) and subjected to RT-
PCR for the detection of cDNA fragments of NR1 splice variants.
Immunoprecipitation/RT-PCR (IP/RT-PCR) was carried out essen-
tially as described previously [6]. Briefly, mouse adult cortex or
RGC-5 cells were homogenized in RNase-free SDS-lysis buffer
Fig. 2. TLS interacts with myosinVa and NMDAR1 in differentiated RGC-5 cells:
(A) the expression of the different glutamate receptors inclusive of the different
NMDAR subtypes (NR1, NR2A, NR2B), mGluR5 and TLS in RGC-5 was detected
by immunoblotting. The same amounts of protein extracts (20 ␮g) from undiffer-
entiated control RGC-5 cells (ctrl) and RGC-5 differentiated by staurosporin (diff,
SS) were resolved by SDS-PAGE. Actin was probed as an indication for equal pro-
tein loading in both lanes. (B) RGC-5 cell lysate was immunoprecipitated with
anti-TLS antibody after treatment with 200 nM staurosporin (SS) or left as con-
trol. The immunoprecipitated proteins were separated in SDS-PAGE followed by
the detection of phosphorylated-tyrosine/serine/threonine residues in the proteins
(anti-phospho-Y/S/T). The immunoblot was reprobed with anti-TLS antibody (TLS).
(C) RGC-5 was lysed in 0.2% SDS-containing lysis buffer. The cell lysate was incubated
with an anti-TLS mouse monoclonal antibody (aTLS) or mouse IgG and control (IgG).
The immunoprecipitated materials (IP) and whole cell lysate (20 ␮g) (WL) were
analyzed by immunoblotting using anti-NR1, anti-MyoVa and anti-TLS antibodies.
[6,21]. Then one-third of the volume was used for normal IP, and
the rest was used for RT-PCR.
All animal experiments in this study were regulated by the
National Advisory Committee on Laboratory Animal Research in
Singapore, and Committee for Animal Experimentation in the
School of Science and Technology at Waseda University in Japan.
TLS was substantially expressed in the retinal GCL of the mouse
retina and in a rat retinal ganglion cell line RGC-5 [13] (Fig. 1A). To
localize TLS in mouse retinas, the retina sections were immunos-
tained with anti-TLS monoclonal antibody. The microscopic image
showed that TLS was expressed and localized distinctly in the GCL
and moderately in the inner nuclear layer (INL) (Fig. 1B).
NR1, NR2A, NR2B and mGluR5 were expressed in both undif-
ferentiated control and differentiated RGC-5 cells (Fig. 2A). A
non-specific protein kinase inhibitor staurosporin (SS) was used
 Author's personal copy

W. Selamat et al. / Neuroscience Letters 450 (2009) 163–166 165

Fig. 3. TLS protein does not associate with NR1 mRNA in RGC-5 cells: (A) expression of NR1-1, NR1-2, and NR1-4 in RGC-5 and mouse cortex was determined by RT-PCR.
(B) Schematic representation of IP/RT-PCR. (C) TLS protein in the IP samples was detected by immunoblotting. (D) IP/RT-PCR was performed to detect any NR1 splice variant
mRNAs in the immunoprecipitates in RGC-5 cells (IP, RGC-5) or in mouse cortex (IP, cortex). IgG control did not immunoprecipitate NR1 splice variant mRNAs (IgG). The same
amount of cDNA (250 ng) from RGC-5 and mouse cortex was subjected to PCR as a positive control (ctrl).

at 200 nM to induce the differentiation of RGC-5 cells. However,
the phosphorylation state of TLS in SS-treated cells was similar
to that in the intact control cells (pTLS, Fig. 2B). Therefore, we
decided to use the differentiated RGC-5 in this study. As a myosin
family is implicated in the trafficking of glutamate receptors [14],
MyoVa may be involved in the trafficking of NR complexes to post-
synaptic membrane or intracellular compartments [3]. Since the
previous proteomics of mouse brain has revealed that TLS is a
component of NR multi-complex containing NR1 and MyoVa [9],
we decided to confirm their interaction in RGC-5 by immuno-
precipitation using anti-TLS antibody. Immunblots showed that a
protein complex of MyoVa, TLS and NR1 was present in RGC-5 cells
(Fig. 2C).
Next we examined which NR1 splice variants were expressed
in RGC-5 cells. Expression of three major NR1 splice variants, C2
splice type (NR1-1 and NR1-2) and C2 type (NR1-4), was semi-
quantitatively assessed by RT-PCR. NR1-3 was not expressed in
the mouse cerebral cortex and adult mouse retina at a detectable
level (our unpublished data) and was thus not examined in greater
detail. We observed that RGC-5 cells predominantly expressed NR1-
4 mRNA at a significant level (Fig. 3A). There was no detectable
expression of the other splice variants in RGC-5 cells.
Known data of TLS binding to mRNA [6,22] led us to examine
whether TLS binds to the functional NR1 splice variant mRNAs in
RGC-5 cells. For this purpose, IP/RT-PCR was performed (Fig. 3B).
The immunoblot analysis demonstrated that TLS was immuno-
precipitated with the TLS antibody (Fig. 3C). The RT-PCR results,
however, showed that the TLS antibody did not immunoprecipi-
tate any NR1 splice variant mRNAs with TLS in RGC-5 cells (RGC-5,
Fig. 3D), although all forms of the NR1 splice variants tested were
present in the IP samples prepared from mouse cortex by the same
IP/RT-PCR (cortex, Fig. 3D). The results indicated that TLS did not

Fig. 4. TLS interacts with NR1 splice variant protein: (A) protein extracts of RGC-5 cells were immunoprecipitated with anti-MyoVa antibody in lysis buffer supplemented with
either 0.1 mM CaCl2 (Ca) or 1 mM EGTA (EGTA), or left as control (ctrl), followed by SDS-PAGE and immunoblot analyses using anti-NR1 (pan NR1), anti-NR1 C2 protein (C2 ),
anti-TLS (TLS) and anti-MyoVa (MyoVa) antibodies. IgG: rabbit IgG control. Representative data from four independent experiments are shown. (B) Immunocytochemistry
of RGC-5 using anti-NR1 (NR1, red) and anti-TLS (TLS, green) antibodies. Representative images are shown. Arrows: colocalization of TLS with NR1 in the neurites. (C)
Immunocytochemistry of RGC-5 after depolarization induced by 75 mM KCl (KCl) or left as control (control). Representative merged color images are shown (red: NR1, green:
TLS). Bars: 20 ␮m.
 Author's personal copy
166 W. Selamat et al. / Neuroscience Letters 450 (2009) 163–166
associate with any form of the functional NR1 splice variant mRNAs
in RGC-5 cells.
In RGC-5 cells, NR1 of approximately 116 kDa was coimmuno-
precipitated with TLS and MyoVa (Fig. 2C). As the molecular size
of NR1 observed seemed to be lower than its known molecu-
lar weight, we examined the types of NR1 splice variants in the
MyoVa-immunocomplex. The results showed that MyoVa and TLS
were coimmunoprecipitated with NR1 C2 splice variant, or most
likely NR1-4, in lysis buffer despite of its substantially low protein
level (C2 , ctrl, Fig. 4A). Although the association of NR1 C2 with
the TLS–MyoVa complex was enhanced in the presence of Ca2+ -
chelating reagent EGTA (1 mM) (EGTA, Fig. 4A), the C2 variant was
still coimmunoprecipitated with MyoVa in the presence of 0.1 mM
CaCl2 , but not with TLS (C2 , Ca, Fig. 4A).
Finally, we wanted to establish if TLS translocates with NR1
in RGC-5 cells as TLS formed a protein complex with NR1 and
MyoVa in calcium-dependent manner. To address this issue, colo-
calization between TLS and NR1 were examined by a double
fluorescent immunocytochemistry. Depolarization of RGC-5 cells
by KCl has been shown to evoke a calcium influx [20]. Therefore,
the effect of the depolarization on the TLS–NR1 complex was exam-
ined after treating the cells with 75 mM KCl for 2 min or left as
control, followed by methanol-fixation. The fixed cells were then
immunostained with mouse anti-TLS and rabbit anti-NR1 antibod-
ies. The dual-fluorescent images showed that both TLS and NR1
were distributed in the cell soma and the neurites (Fig. 4B and C).
Populations of TLS granules were colocalized with NR1 in the neu-
rites (arrows, Fig. 4B). Next, Pearson’s correlation coefficients were
calculated to quantitatively determine the degree of colocalization
between TLS and NR1 in the immunofluorescent images. Pearson’s
correlation coefficient was significantly higher in control intact cells
than in KCl-treated cells (0.487 ± 0.203 in control, 0.233 ± 0.136 in
KCl-treated cells; t-test, n = 20, P < 0.01; Fig. 4C). The result indi-
cated that the depolarization induced a redistribution of TLS and
NR1 from the protein complex.
Our results demonstrate that TLS is expressed in RGC-5 and asso-
ciates with the NR1 complex that is translocated in the neurites,
but is not bound to NR1 mRNAs. Thus, TLS may play an important
role in trafficking of NR1 rather than in the NR1 splicing regula-
tion in RGC-5. Supporting this view, most of the TLS was observed
in the cytoplasm (Fig. 1B, Fig. 4B and C) while some populations
of the TLS clusters were translocated to the neurites (Fig. 4B). Our
live cell monitoring of TLS–GFP in RGC-5 cells revealed that a sub-
set of TLS clusters repeatedly gathered and dispersed into smaller
particles, which then moved towards the distal ends of the neurites,
albeit at a slow rate (∼5 ␮m/min, maximum rate ∼12 ␮m/min) (our
unpublished data).
We found that RGC-5 cells predominantly expressed NR1-4
(Fig. 3A). Kreutz et al. reported that the expression level of NR1-
4 was relatively kept high and did not change over the time course
of 4 weeks after optic nerve injury, while the expression lev-
els of other splice variants were eventually decreased [11]. We
initially hypothesized that TLS might stabilize the NR1-4 mRNA.
However, our IP/RT-PCR analyses failed to show the association
of TLS protein with NR1-4 mRNA in RGC-5 (Fig. 3D) or in mouse
retina that had undergone optic nerve injury (our unpublished
data).
It has been reported that the different NR1 splice variants on the
cell surface regulate the surface expression of functional NR com-
plex [16]. The NR complex containing NR1-4 or C2 type is most
abundant on the cell surface and most permeable to extracellu-
lar Ca2+ [16]. It is possible that TLS may modulate the cell surface
expression of NR1-4 according to the magnitude of Ca2+ -influx
in RGC. It would be interesting to further investigate a spatial-
temporal trafficking of NR1 splice variants by TLS in the mouse
retina since TLS was localized not only in GCL but also in the INL.
Acknowledgements
This work was supported in part by grants from Ministry of
Education, Culture, Sports, Science and Technology (MEXT) and
Academic Frontier Project for Private Universities. F.W. is supported
by DUKE-NUS GMS Block Grant Funding. We thank N. Agarwal for
a kind contribution of RGC-5, S. Tong and N. Dendukuri for their
excellent technical supports.
References
[1] A. Belly, F. Moreau-Gachelin, R. Sadoul, Y. Goldberg, Delocalization of the mul-
tifunctional RNA splicing factor TLS/FUS in hippocampal neurones: exclusion
from the nucleus and accumulation in dendritic granules and spine heads,
Neurosci. Lett. 379 (2005) 152–157.
[2] J.H. Brandstätter, E. Hartveit, M. Sassoè-Pognetto, H. Wässle, Expression of
NMDA and high-affinity kainite receptor subunit mRNAs in the adult rat retina,
Eur. J. Neurosci. 6 (1994) 1100–1112.
[3] R.C. Carroll, R.S. Zukin, NMDA-receptor trafficking and targeting: implications
for synaptic transmission and plasticity, Trends Neurosci. 25 (2002) 571–577.
[4] A. Crozat, P. Aman, N. Mandahl, D. Ron, Fusion of CHOP to a novel RNA-binding
protein in human myxoid liposarcoma, Nature 363 (1993) 640–644.
[5] L.J. Frassetto, C.R. Schlieve, C.J. Lieven, A.A. Utter, M.V. Jones, N. Agarwal, L.A.
Levin, Kinase-dependent differentiation of a retinal ganglion cell precursor,
Invest. Ophthalmol. Vis. Sci. 47 (2006) 427–438.
[6] R. Fujii, T. Takumi, TLS facilitates transport of mRNA encoding an actin-
stabilizing protein to dendritic spines, J. Cell Sci. 118 (2005) 5755–5765.
[7] R. Fujii, S. Okabe, T. Urushido, K. Inoue, A. Yoshimura, T. Tachibana, T. Nishikawa,
G.G. Hicks, T. Takumi, The RNA binding protein TLS is translocated to dendritic
spines by mGluR5 activation and regulates spine morphology, Curr. Biol. 15
(2005) 587–593.
[8] N. Hirokawa, mRNA transport in dendrites: RNA granules, motors, and tracks,
J. Neurosci. 26 (2006) 7139–7142.
[9] H. Husi, M.A. Ward, J.S. Choundhary, W.P. Blackstock, S.G. Grant, Proteomics
analysis of NMDA receptor–adhesion protein signaling complexes, Nat. Neu-
rosci. 3 (2000) 661–669.
[10] Y. Kanai, N. Dohmae, N. Hirokawa, Kinesin transports RNA: isolation and char-
acterization of an RNA-transporting granule, Neuron 43 (2004) 513–525.
[11] M.R. Kreutz, T.M. Böckers, J. Bockmann, C.I. Seidenbecher, B. Kracht, C.K. Vorw-
erk, J. Weise, B.A. Sabel, Axonal injury alters alternative splicing of the retinal
NR1 receptor: the preferential expression of the NR1b isoforms is crucial for
retinal ganglion cell survival, J. Neurosci. 18 (1998) 8278–8291.
[12] A.M. Krichevsky, K.S. Kosik, Neuronal RNA granules: a link between RNA local-
ization and stimulation-dependent translation, Neuron 32 (2001) 683–696.
[13] R.R. Krishnamoorthy, P. Agarwal, G. Prasanna, K. Vopat, W. Lambert, H.J. Sheedlo,
I.-H. Pang, D. Shads, R.J. Wordinger, T. Yorio, A.F. Clark, N. Agarwal, Characteri-
zation of a transformed rat retinal ganglion cell line, Mol. Brain Res. 86 (2001)
1–12.
[14] M.-F. Lisè, T.P. Wong, A. Trinh, R.M. Hines, L. Liu, R. Kang, D.J. Hines, J. Lu, J.R.
Goldenring, Y.T. Wang, A. El-Husseini, Involvement of myosin Vb in glutamate
receptor trafficking, J. Biol. Chem. 281 (2006) 3669–3678.
[15] N. Nakanishi, R. Axel, N.A. Schneider, Alternative splicing generates functionally
distinct N-methyl-d-aspartate receptors, Proc. Natl. Acad. Sci. U.S.A. 89 (1992)
552–556.
[16] S. Okabe, A. Miwa, H. Okado, Alternative splicing of the C-terminal domain
regulates cell surface expression of the NMDA receptor NR1 subunit, J. Neurosci.
19 (1999) 7781–7792.
[17] E.M. Schuman, J.L. Dynes, O. Steward, Synaptic regulation of translation of den-
dritic mRNAs, J. Neurosci. 26 (2006) 7143–7146.
[18] O. Steward, E.M. Schuman, Compartmentalized synthesis and degradation of
proteins in neurons, Neuron 40 (2003) 347–359.
[19] H. Sugihara, K. Moriyoshi, T. Ishii, M. Masu, S. Nakanishi, Structures and proper-
ties of seven isoforms of the NMDA receptor generated by alternative splicing,
Biochem. Biophys. Res. Commun. 185 (1992) 826–832.
[20] K.T. Tchedre, R.Q. Huang, A. Dibas, R.R. Krishnamoorthy, D.H. Dillon, T. Yorio,
Sigma-1 receptor regulation of voltage-gated calcium channels involves a direct
interaction, Invest. Ophthalmol. Vis. Sci. 49 (2008) 4993–5002.
[21] A. Yoshimura, R. Fujii, Y. Watanabe, S. Okabe, K. Fukui, T. Takumi, Myosin-Va
facilitates the accumulation of mRNA/protein complex in dendritic spines, Curr.
Biol. 16 (2006) 2345–2351.
[22] H. Zinszner, J. Sok, D. Immauel, Y. Yin, D. Ron, TLS (FUS) binds RNA in vivo and
engages in nucleo-cytoplasmic shuttling, J. Cell Sci. 110 (1997) 1741–1750.
[23] R.S. Zukin, M.V. Bennett, Alternatively spliced isoforms of the NMDAR1 receptor
subunit, Trends Neurosci. 18 (1995) 306–313.