(Received for publication, July 29, 1996, and in revised form, November 1, 1996)
From the Molecular Neurobiology Unit, Gerontology Research Center, NIA, National Institutes of Health, Baltimore, Maryland 21224
The N-methyl-D-aspartate
(NMDA) subtype of glutamate receptor plays important roles in synaptic
plasticity, the induction of long term potentiation, and
excitotoxicity. Mechanisms governing the regulation of expression of
its subunit genes remain largely unknown. The promoter of the essential
subunit of the NMDA receptor heteromer, NMDAR1, contains DNA binding
elements recognized by the nerve growth factor-inducible/early growth
reaction factor (NGFI/Egr) family of transcription factors that are
rapidly induced by neurotrophins, such as nerve growth factor (NGF).
This study examined the effect of NGF on the activity of the
N-methyl-D-aspartate receptor subunit 1 (NMDAR1) promoter/luciferase reporter constructs in PC12 cells, which
contain the high affinity TrkA receptor for NGF and the low affinity
p75NTR receptor for neurotrophins. NGF up-regulated the
activity of the NMDAR1 promoter by 3-4-fold in a time- and
dose-dependent manner. 5 deletional analysis of the
promoter indicated that the responsive element(s) resides in the
proximal region containing GSG and Sp1 sites. Mutational analysis of
these sites revealed that both were important for NGF regulation.
Transient expression of Egr-1 increased activity of the wild type
promoter but failed to increase activity of a GSG mutant promoter.
Other neurotrophins did not activate the promoter, while K-252a
inhibited the action of NGF. These results suggest that the NGF effect
is mediated by the high affinity NGF receptor, Trk A and that
neurotrophin binding to the low affinity neurotrophin receptor,
p75NTR, alone does not affect the promoter activity. Our
results suggest that NGF is able to up-regulate the activity of the
NMDAR1 promoter and may play a role in controlling the expression
levels of NMDA receptors.
The NMDA1 subtype of glutamate receptor is a unique ligand-gated calcium channel exhibiting voltage-dependent magnesium block, which suggests important roles for this receptor in developmental neuronal survival, synaptic plasticity, and long term potentiation (1-3). Overactivation of this receptor may be responsible for the glutamate-induced neuronal excitotoxicity in the postischemic brain, trauma-induced brain injury, and possibly chronic neurodegenerative disorders of aging such as Alzheimer's and Huntington's diseases (2).
Recent evidence suggests that functional NMDA receptors in neurons are heteromeric molecules assembled from an essential subunit, NMDAR1, and one or more of a second family of subunits termed NMDAR2-A-D (4-6). The NMDAR1 gene transcript undergoes alternative splicing to produce eight variants (7, 8). Different variants show distinct physiological and pharmacologic properties (9). Numerous studies have shown that the level of total NMDAR1 mRNA changes during development and maturation of the central nervous system (10-12), that there are region-specific changes in expression after experimentally induced ischemia (13-15), and that mRNA levels change after prolonged antagonist treatment (16). It also has been reported that message levels change due to 6-hydroxydopamine lesions (17), chronic dopamine antagonist treatment (18), estrogen treatment (19), epileptic seizures (20), and a diurnal rhythm (21). These changes in message levels may potentially influence the amount of the NMDAR1 subunit protein and result in the differential assembly of heteromeric NMDA receptors or a change in the number of functional channels. However, the mechanisms governing these changes in message levels remain largely unknown. Only recently have insights been obtained into the transcriptional regulation of the NMDAR1 gene (22).
Neurotrophins, including nerve growth factor (NGF), brain-derived neurotrophic factor (BDNF), neurotrophin 3 (NT-3), and neurotrophin 4/5 (NT-4/5), classically were found to support neuronal survival, growth, and differentiation (23-26). Recent studies revealed that they also are involved in other neuronal activities such as synaptic plasticity (27, 28) and possibly neurotoxicity (29). Interestingly, NMDA receptors also are involved in these processes (2). Activation of the NMDA receptor is able to up-regulate BDNF and NGF levels in hippocampal neurons (30), and most recently BDNF was shown to be required for the formation of LTP in hippocampal brain slices (26) and for the maturation of NMDA receptors in cultured cerebellar granule cells (31). Little evidence is available on the effect of neurotrophins on the expression of NMDA receptor genes.
We have previously isolated and analyzed a 3-kb promoter region of the NMDAR1 gene (22, 32). A DNase I footprinting experiment showed protection of a proximal GC-rich sequence including a consensus GSG motif (GCG5GC) recognized by an immediate early gene family of transcription factors, i.e. nerve growth factor-inducible factors (NGFIs) or early growth reaction (Egr) factors (33-36). Neurotrophins induce a rapid increase in this family of factors both in vitro and in vivo (37). Several reports documented little or no change in NMDAR1 message after ischemic insult, while others showed marked increases (13-15). In this experimental paradigm, elevated levels of neurotrophins, NGFI-A, NMDA receptor activity, and subsequent cell loss also were documented (2, 30, 38-41). These findings prompted us to examine the effects of neurotrophin treatment on activity of the 3-kb promoter region of the NMDAR1 gene in hopes of better understanding the interrelationship between neurotrophic factor activity and NMDA gene regulation.
Plasmid Construction
Site-directed MutagenesisA megaprimer PCR method was used
to create point mutations of GSG and Sp1 elements in the proximal
356-bp promoter of the NMDAR1 gene (42). A 356/
1 fragment of the
NMDAR1 gene promoter was retrieved from pG3029 (22) by digestion with
BsmI and removal of overhang ends and then digestion with
SacI. The isolated fragment was subcloned into
HincII/SacI-digested pGEM3Zf(+) to form pG356. A
NdeI-linearized pG356 served as a template in all of the
following PCR reactions. An upstream primer, GB 46 (5
-atctagtcgacCATTCACGCCAACGCAG), tailed with a HincII
site, was used to generate the megaprimers with the following mutated
primers in antisense sequence: GB57 for Sp1a site,
5
-CGTTCTACataTgCCACCGCCCCCGCTTCCG; GB41 for Sp1b site,
5
-CGTTCTACCCCTCCCAtatgCCCCGCTTCCG; GB58 for GSG site,
5
-CGTTCTACCCCTCCCACCGCCCaaGCTTCCTG; GB72 for GSG/Sp1 × 2 sites,
5
-CGTTCTACataTgCCAtatgCCCCGCTTCCTG. The megaprimers were fractionated
by electrophoresis on an agarose gel and purified from the gel with
Glassmilk (Geneclean II kit, Bio 101, Inc., La Jolla, CA). In the
secondary PCR, mutated 356-bp promoters were generated using
megaprimers with a downstream primer, GB40
5
-gatatagtgtatgaGAGCTCCGGGCACAGCGG. All secondary PCR products were
cloned into a pGEM-T vector (Promega, Madison, WI) and sequenced to
confirm the mutations and the fidelity of the remaining sequences using
either the Sequenase II kit (U.S. Biochemical Corp./Amersham Corp.) or
a fluorescent Taq dye deoxy terminator cycle sequencing kit
(ABI, Foster City, CA). All oligonucleotides were synthesized with an
ABI 392 DNA/RNA synthesizer (Foster City, CA) and purified with G25
columns.
The mutated 356-bp promoter constructs were
excised from pGEM-T vector with HincII/SacI and
ligated into SmaI/SacI-digested pGL2-Basic, a
luciferase reporter gene vector to form pNRL356mSp1a, pNRL356mSp1b,
pNRL356mGSG, and pNRL356m GSG/Sp1X2 for transient expression. NMDAR1
promoter/reporter constructs for stable expression were prepared by
subcloning the promoter/luciferase gene fragments from the pGL2-Basic
constructs into pCDNAI/neo (Invitrogen, San Diego, CA). All mutated
pNRL356, pNRL356, and pNRL3029 (22) constructs were digested with
PvuI and blunted with T4 DNA polymerase. The constructs were
further digested with BamHI. The
PvuI/BamHI fragments containing a SV40
polyadenylation signal/NMDAR1 promoter/luciferase gene cDNA/SV40
small intron/SV40 polyadenylation signal were isolated. These fragments
were then cloned into EcoRV/BamHI-digested
pCDNAI/neo vector to yield pNRL356mSp1a/neo, pNRL356mSp1b/neo,
pNRL356m GSG/Sp1X2/neo, pNRL356mGSG/neo, pNRL356/neo, and
pNRL3029/neo. All DNA plasmids were purified from host
Escherichia coli using a Qiagen (Chatsworth, CA)
maxipreparation kit. Plasmids pCBEgr-1, containing full-length Egr-1
cDNA driven by the CMV promoter, and pCBEgr1331-374, lacking the DNA binding domain of Egr-1, were kindly supplied by Dr. Vakas P. Sukhatme.
To generate riboprobes for NMDAR1
mRNA, plasmid pG3.7K/S was linearized with PstI for
transcribing riboprobe 1 (32). A SacI/BsmI sequence with nucleotides 2328-2660 (6) complementary to a common
region of all NMDAR1 mRNA splicing variants was subcloned from pN60
(43) into pGEM3Zf(+) to form pNR1-4, which was linearized with
EcoRI for riboprobe 4. To transcribe an antisense riboprobe of luciferase mRNA, pGEM-luc (Promega) was linearized
with EcoRV. For in vitro
transcription/translation, Egr-1 or Egr-1331-374 cDNA including
5
- and 3
-untranslated region was retrieved with EcoRI from
pCB plasmid and subcloned into pGEM3Zf(+) to generate pGEgr-1 and
pGEgr-1
331-374 with inserts at 3
5
orientation in the vector
multiple cloning sites. A SmaI digestion of pGEgr-1 and
pGEgr-1
331-374 and religation of vector/cDNA containing
fragments removed most of the 5
-untranslated region and an outside ATG and formed pGEgr-1/-sm and pGEgr-1
331-374/-sm.
Transfections of PC12 Cells and Reporter Gene Assay
PC12 cells were cultured and maintained as described previously
(22). PC12 cells were transfected with constructs using Lipofectin
(Life Technologies, Inc.). For transient expression, the constructs
were co-transfected with pCMV (Clontech, Palo Alto, CA). Twenty
hours after transfection, NGF (2.5 S; Promega) was added at 100 ng/ml.
The cells were harvested at different times after adding NGF, and both
luciferase and
-galactosidase activities in the cell lysates were
measured (22). In these experiments, the DNA content in the cell
lysates was also measured with Hoechest Dye 33258 (Polyscience,
Warrington, PA) to correlate the relative reporter activity to cell
number in individual samples (44). In some experiments, pNRL356 or
pNRL356mGSG was co-transfected into PC12 cells with pCBEgr-1 or
pCBEgr-1/
331-374 in addition to pCMV
. Two days after
transfection, activities of luciferase and
-galactosidase in cell
lysates were measured. To establish stable transfectants, only
luciferase-containing constructs were used, and 2 days after
transfection the cells were cultured with Geneticin (G-418, Life
Technologies) at 300 µg/ml. Each construct was transfected in
duplicate transfections in 100-mm culture dishes, and all transfectants
for the same constructs were pooled together after a 3-week selection.
The level of luciferase in the transfectants was measured as described
before (22). The total genomic DNA in the lysate was quantitated with
Hoechest Dye 33258, and the integrated NMDAR1 promoter/luciferase genes
were measured as described previously.2 The
stable transfectants were subcultured on 60-mm dishes coated with
collagen and treated with NGF. At various times, cells were collected
for a luciferase assay. In some experiments, stable transfectants were
pretreated with either actinomycin D at 1 µg/ml or K-252a at 50 nM for 15 min and treated with or without NGF. Some stable
transfectants or wild type PC12 cells were cultured and treated with
NGF for RNA extraction.
Measurement of mRNA Level
Total RNA in stable transfectants or wild type PC12 cells was
extracted with an RNAzol kit (Cinna/Biotecx, Friendswood, TX) and
quantitated by measuring the absorbance at 260 nm. A RNase protection
assay (Ambion, Inc., Austin, TX) was used to measure the level of
endogenous NMDAR1 mRNA and the expressed luciferase mRNA (32).
Briefly, two riboprobes for NMDAR1 mRNA were used: riboprobe 1 for
the 5-untranslated region was described before (32). Riboprobe 4 was
transcribed from EcoRI-linearized pGNR1-4 by Sp6 RNA
polymerase and has a 330-nucleotide sequence complementary to NMDAR1
mRNA. For the luciferase mRNA assay, an antisense probe was
synthesized from EcoRV-linearized pGEM-luc by T7
RNA polymerase to form a 413-nucleotide probe containing 356 nucleotides complementary to luciferase mRNA. Ten µg of total RNA
were hybridized with riboprobes overnight at 42 °C and digested with
RNase ONE (Promega) for 30 min for NMDAR1 and 20 min for luciferase
mRNA. Protected RNA probes were resolved on 7 M
urea/polyacrylamide gel and exposed to x-ray film as well as to a
PhosphorImager plate. The amount of mRNA was analyzed on a
PhosphorImager SI system (Molecular Dynamics, Sunnyvale, CA).
Gel Mobility Shift
Transcription factor Egr-1 or Egr1-331-374 was synthesized
in vitro from pGEgr-1/-sm or pGEgr-1
331-374/-sm using
the TNT Sp6 coupled recticulocyte lysate system (Promega) following the manufacturer's instructions. A luciferase plasmid was included in some
reactions to monitor the synthesis efficiency and possible inhibitory
contamination from templates. The gel mobility shift assay was
performed as described previously (22). The probe used was a
double-stranded oligonucleotide GB63/67 encompassing the GSG element in
the NMDAR1 promoter, and its sequences are as follows: GB63 for sense
strand, 5
-CAGGAAGCGGGGGCGGTGGGAGGGGTAGAAC; GB67 for antisense strand,
5
-GTTCTACCCCTCCCACCGCCCCCGCTTCCTG. The probe was labeled at the 5
-end
with 32P by T4 DNA polynucleotide kinase. For
competition, in addition to nonlabeled GB63/67, oligonucleotide GB27/28
containing a GSG core sequence and GB58/62 carrying a mutated GSG
element were synthesized. GB58 is the same as the one used in
site-directed mutagenesis, and the others have sequences as follows:
GB62 for mutated GSG sense strand, 5
-CAGGAAGCttGGGCGGTGGGAGGGGTAGAACG; GB27 for sense strand, 5
-TCAGGAAGCGGGGGCGGTGGGA; GB28 for antisense strand, 5
-TCCCACCGCCCCCGCTTCCTGA. In some experiments, a GSG consensus (Santa Cruz Biotechnology, Santa Cruz, CA) was labeled at the
5
-end with 32P by T4 DNA polynucleotide kinase
and used as a probe to examine the activity of synthesized Egr-1
protein. Polyclonal antibodies against Egr-1 or Sp1 protein (Santa Cruz
Biotechnology) were used to confirm the specificity of Egr-1 binding.
In a typical reaction, a slowly migrating band was formed after 1 µl
of synthesized Egr-1 protein reacted with 25 fmol of labeled GSG
consensus. This band was abolished by preincubation with a 20-fold
excess of nonlabeled probe but not by an 80-fold excess of Sp1
consensus (Promega). The same band was supershifted only by Egr-1
antibody and not by Sp1 antibody. Synthesized Egr-1
331-374 failed
to form this band with labeled GSG consensus.
We previously cloned and characterized a 3-kb promoter of the
NMDAR1 gene. In luciferase reporter assays, this promoter exhibited high level expression in PC12 cells. This activity resided in a
proximal, GC-rich region of the promoter near multiple transcriptional start sites (22). To test whether this promoter was sensitive to
neurotrophic factor regulation, we initially transfected the full-length NMDAR1 promoter, fused to a luciferase gene (pNRL3029), into PC12 cells and added NGF to the cells 20 h after
transfection. Transfection efficiency was monitored by co-transfecting
a lacZ gene driven by a CMV promoter. We also measured the
DNA content of each transfected plate of cells to control for
differences in proliferation (slowed) and differentiation (enhanced) of
the NGF-treated PC12 cells versus untreated cells. NGF
treatment increased both the luciferase activity and the
-galactosidase activity of transfected cells. The luciferase
activity increased with time and reached a maximum 48 h after the
addition of NGF (Fig. 1A), while
-galactosidase activity was elevated at the earliest time point and
remained above untreated cells throughout the experiment (Fig.
1B). These results suggested that NGF may up-regulate the NMDAR1 promoter and that the CMV promoter may have NGF-responsive elements. This result precluded the use of this lacZ
construct as a control for transfection efficiency. We then tested
the effect of NGF on PC12 cell lines stably transfected with either a
3029- or a 356-bp proximal NMDAR1 promoter fused to a luciferase gene. NGF caused a time- and dose-dependent increase in
luciferase activity (Fig. 1, C and D). Maximal
activity, ~3.5-fold over untreated cells, was reached in 4 h,
and NGF was effective at concentrations of 5 ng/ml with a maximum
effect at about 50 ng/ml. This is in the range of its differentiating
effect on PC12 cells (45). Similar NGF effects were seen in cell lines
expressing either promoter fragment. The decline in luciferase activity
after peak activity may be due to exhaustion of NGF from the culture
medium. To confirm that the increase was due to transcriptional
up-regulation and not translational modification of luciferase
reporter, we measured luciferase mRNA levels by an RNase protection
assay. Luciferase message levels increased as early as 30 min after NGF treatment and reached a maximal elevation of ~3.2-fold above
untreated cells about 4 h after treatment (Fig.
2A). Actinomycin D totally abolished the NGF
effect while having no effect on activity in untreated cells (Fig.
2B). These results suggest that the NGF effect was at the
NMDAR1 promoter and was due to enhanced transcription of the reporter
gene.
Since the 356-bp proximal promoter construct showed an NGF response
similar to that of the 3-kb promoter (Fig. 1, C and
D), we mutagenized the GSG and the two tandem Sp1 elements
on this shorter fragment to test the importance of these sites in
mediating the NGF effect. This fragment has basal activity similar to
the 3-kb promoter in unstimulated cells (22). In transient
transfections, mutagenesis of any one of the sites reduced basal
activity to about 50% that of wild type promoter (Fig.
3A). In PC12 cell lines stably expressing
these same mutant constructs, single mutations led to similar
reductions in activity (Fig. 3B). However, mutations in both
Sp1 sites (one of which overlaps the GSG site) reduced activity to
about 10% of wild type in both transfection assays. We then examined
the NGF responsiveness of these cell lines stably expressing the
various mutations. Mutagenesis at each site alone reduced, but did not
abolish, NGF responsiveness (Fig. 3B). However, mutagenesis
at both Sp1 sites greatly reduced the NGF responsiveness.
The role of the GSG site in the NGF response was tested by transiently co-transfecting Egr-1/NGFI-A protein with the NMDAR1 promoter/luciferase reporter gene in PC12 cells. In these cells, overexpression of Egr-1/NGFI-A increased basal activity by about 2-fold. Transfection of a deletional mutant of Egr-1/NGFI-A, lacking the DNA binding domain failed to enhance activity. In another experiment, co-transfecting Egr-1/NGFI-A did not increase the activity of the GSG mutant luciferase construct (Fig. 3C). Also, we noticed that increasing the amount of co-transfected Egr-1 construct resulted in a reduced reporter activity. This reduction also was seen with increased deletional mutant of Egr-1/NGFI-A. Therefore, we believe that this reduction was due to a nonspecific inhibition due to excess DNA in the transfection. To correlate the transfection results seen in Fig. 3C with a molecular mechanism, we examined the interaction of synthesized Egr-1 protein with the GSG element in the NR1 promoter. In a gel shift experiment (Fig. 3D), an NR1 promoter oligonucleotide probe containing the GSG element formed a specific band with synthesized Egr-1 protein. The specificity of this band was confirmed with a GSG consensus probe, competitors, and antibody supershift (see "Experimental Procedures"). In Fig. 3D, mutated GSG oligonucleotide failed to compete the binding compared with strong competition by both nonlabeled probe and a short oligonucleotide containing the GSG core sequence. These experiments demonstrate that both the GSG and Sp1 sites play important roles in both basal and NGF-regulated activity of the NMDAR1 promoter.
To examine how NGF mediates its effect, we examined the role of low
affinity neurotrophin receptor, p75NTR, and high affinity
Trk A receptor. PC12 cells express both of these receptors but lack
high affinity Trk B and C receptors for BDNF and NT-3 (46). Treatment
of cells with BDNF, NT-3, and NT-4 at concentrations (100 ng/ml)
sufficient to activate the p75NTR failed to increase
luciferase activity in cell lines stably expressing the 356-bp proximal
NMDAR1 promoter construct (Fig. 4A).
Furthermore, the NGF effect was completely blocked by preincubation
with K-252a (Fig. 4B), a potent inhibitor of Trk receptor
kinase activity (47). These results suggest that NGF was having its
effect by activation of Trk A receptors and not via the
p75NTR low affinity receptors.
PC12 cells endogenously express low levels of NMDAR1 message in one
predominant transcript and a very low level of receptor subunit protein
(48). We examined whether NGF could regulate the expression of the
endogenous gene by RNase protection assays with two probes to the
NMDAR1 transcript. Even after 4 days of treatment with NGF, when
neurite extension had occurred, PC12 cells did not show any changes in
the levels of endogenous NMDAR1 message (Fig. 5). This
experiment was repeated on several different PC12 cell lines with
similar results. The half-life of the endogenous message was not
different in cells treated with and without NGF (data not shown). Also,
we did not detect any differences in the sequence of the PC12 cell
NMDAR1 gene promoter relative to the rat sequence that we originally
characterized and used in the reporter constructs described here.
Neurotrophins and glutamate are thought to be involved in neuronal survival, growth, and plasticity (1-3, 23-25, 49-51). We therefore examined the effects of NGF treatment on activity of the previously isolated 3-kb promoter region of the NMDAR1 gene in PC12 cells. In this study, we observed that NGF, the prototype of neurotrophins, up-regulated the activity of NMDAR1 promoter/luciferase reporter constructs in both transient and stable transfection experiments. The dose response for this effect paralleled the neurite outgrowth and differentiation effects of NGF on PC12 cells and, like the continuous requirement for NGF to maintain a viable differentiated state in these cells (52), NMDAR1 promoter activation also required the continual presence of NGF. These characteristics of the NGF effect on the NMDAR1 promoter are in agreement with the effects of neurotrophins in establishing and maintaining the neuronal phenotype (52, 53). The expression of neurotrophins and their receptors is elevated during central nervous system development and declines in adulthood (54). Interestingly, expression of the NMDAR1 gene follows neurotrophin expression and also is elevated during early development compared with adult levels (10, 11). The NMDAR1 gene and several neurotrophin genes have an overlapping regional expression; high levels of both types of genes are found in the hippocampus, striatum, and cortex (4, 5, 23). Therefore, the preexistence of neurotrophins in the developing brain may be, in part, responsible for the high level of NMDAR1 message in the late postnatal period. Our result showing a gradual decline in promoter activity, apparently paralleling a decline in NGF from the culture medium, suggests that the continual presence of NGF is required for stable NMDAR1 message levels. This result also agrees with the continual need for trophic factor to support neuronal growth and maintain a differentiated state in vivo and in vitro (24, 52, 53). However, in models of severe ischemia, when neurotrophic factor expression is greatly elevated, NMDAR1 gene expression also can be induced in the hippocampus (see below) (13, 38, 39).
The proximal region (see Fig. 3A) of the NMDAR1 promoter contains a consensus GSG site that overlaps with a consensus Sp1 site, which, in turn, is separated by a single nucleotide from a second Sp1 site (5/6 consensus). Using site-directed mutagenesis experiments, we found that both the GSG-overlapping Sp1 and the downstream Sp1 sites were required for the NGF up-regulation of activity (Fig. 3B). Mutation of each site individually reduced the NGF effect by about 50%, while mutation of both sites almost completely abolished the NGF effect. Our previous results showed that deletion of this region of the promoter abolished basal activity of the promoter (22); the double mutation reported here also dramatically reduced basal activity.
The GSG element is a consensus binding motif for a group of NGF-inducible transcription factors termed NGFI/Egr transcription factors (i.e. NGFI-A/Egr-1, NGFI-C, Egr-2, and Egr-3) that are immediate early genes (33-36). Their expression is rapidly and transiently induced by a number of factors including serum, growth factors, and glutamate (35, 55, 56). They are thought to be the first wave of genes induced after receptor activation (52, 53). In PC12 cells, NGF-induced expression of NGFI-A and -C reaches a peak at 30 min (33, 34). In the present study, NGF up-regulated NMDAR1 promoter activity with a maximum at 4 h, following the NGFI-A response and may represent an example of a second wave of gene expression initiated by NGF. Furthermore, co-transfection of a cDNA for Egr-1 enhanced activity about 2-fold. Also, in gel mobility shift experiments, an oligonucleotide probe containing the intact GSG element of the NMDAR1 promoter was able to form a shifted complex with in vitro synthesized Egr-1. These results suggest that NGF-inducible genes play a role in activation of NMDAR1 gene expression.
The double mutation of the GSG/Sp1 and Sp1 sites abolished the NGF effect, suggesting that in addition to the GSG binding factors, Sp1 protein may be important in the NGF response. Sp1 is a ubiquitous transcription factor that may recruit TATA-box binding proteins or initiation complexes to the initiation site in promoters lacking consensus TATA-boxes (57, 58). The possibility exists that Sp1 proteins are required for NMDAR1 promoter activity and may interact with NGFI proteins to maximally activate NMDAR1 transcription.
PC12 cells express the high affinity NGF receptor Trk A but not Trk B and Trk C (46). Our results suggest that the NGF effect is due to activation of the Trk A receptor rather than the low affinity p75NTR receptor. The NGF effect was blocked by K-252a, an inhibitor of Trk tyrosine kinase activity (47). BDNF, NT-3, and NT-4, all of which interact with the p75NTR (23, 25, 46), did not induce activity of the NMDAR1 promoter. Since the Trk receptors share similar signal transduction pathways (46), the results reported here in PC12 cells may be extrapolated to the central nervous system, where several other neurotrophic factors including BDNF in the cortex may regulate expression of NMDA receptor genes (23). In global ischemia models, BDNF expression is elevated in the penumbral region and is seen in hippocampal neurons (38, 39). Maximal expression of BDNF is seen in severe ischemia, a condition under which NMDAR1 gene expression also is elevated (13, 38, 39). These increases are seen in CA1 pyramidal neurons of the hippocampus, which are highly sensitive to glutamate-induced toxicity (2, 41). These results suggest that high levels of neurotrophins may be required to up-regulate the NMDAR1 gene. Interestingly, an in vitro study of cultured primary cortical neurons that contain Trk B receptors (23) showed that a high dose of BDNF potentiated NMDA-induced necrosis (29).
Neurotrophins are generally thought to protect neurons from various stimuli (23, 46). In fact, toxic activation of NMDA receptors is known to increase expression of NGF, BDNF, TrkB, and TrkC (30). Our results with NGF activation of NMDAR1 expression suggest a potential for a lethal cycle of trophic factor up-regulation increasing NMDA receptor expression. This overexpression of NMDA receptors may then be responsible for neurotoxicity. However, it is possible that the levels of neurotrophins elicited by various insults may be lower than that required to induce expression of NMDA receptor genes; alternatively, these neurotrophins may induce other gene cascades that are protective of neurons. Interestingly, cerebellar granule neurons require NMDA receptor activity in addition to trophic factor support for survival and differentiation (49-51). Granule cells express Trk B receptors (23). BDNF may accelerate the maturation (expression) of NMDA receptor genes normally seen in embryonic neurons in culture (31). It may be that low levels of a positive feedback loop are necessary for optimum survival of neurons, but a major insult that dramatically increases expression of one of these players may induce an accelerated neurotoxic feedback loop (13, 38, 39).
PC12 cells express one of eight possible isoforms of the NMDAR1 gene and low levels of NMDAR2-A, -C, and -D (48, 59, 60). However, several studies have shown that these cells have very little functional NMDA channel activity. Sucher et al. (48) reported barely detectable NMDAR1 protein levels relative to the level of transcripts. Interestingly, in several studies and in results reported here, NGF failed to increase the levels of endogenous NMDAR1 message (48). This resistance to NGF treatment was not due to a change in mRNA stability or mutations in the endogenous promoter sequence. As we have reported previously, the proximal region of the NMDAR1 promoter is GC-rich and has a high frequency of CpG islands, which are subject to DNA methylation (61, 62). Generally, methylation suppresses basal gene expression, and in our stable transfectants, where the NMDAR1 promoter/reporter gene has been integrated into the PC12 genome, the NMDAR1 promoter sequence should be subject to similar suppression. We observe no evidence of suppression of basal activity with our reporters. Interestingly, Macleod et al. (63) observed that the mouse aprt gene promoter is rich in CpG islands yet appears to be unmethylated. They propose that an Sp1 site in this promoter prevented methylation. The NMDAR1 promoter has two proximal Sp1 sites, but it is unlikely that DNA methylation is related to a lack of the NGF effect on the endogenous gene. Rather, in view of the positional uncertainty of regulatory elements in the genome (64), this blockade of the NGF effect on the endogenous gene may be caused by other trans-acting factors that act on uncharacterized cis-elements located outside the 3-kb NMDAR1 promoter. In view of the low levels of NMDAR1 mRNA in PC12 cells (22) and a lack of an NGF response of the endogenous gene (Fig. 5) (48), it will be important to determine if this phenomenon is PC12 cell-specific. This is especially critical in light of recent studies suggesting that various neurotrophic factors such as NGF and BDNF may regulate NMDA receptor gene expression.
We thank the Aluminum Association for support during the course of these studies and Drs. Walter Horton and Nikki Holbrook for helpful discussions.