Unexpectedly, a post-translational modification of DNA-binding proteins, initiating the cell response to single-strand DNA damage, was also required for long-term memory acquisition in a variety of learning paradigms. Our findings disclose a molecular mechanism based on PARP1-Erk synergism, which may underlie this phenomenon. A stimulation induced PARP1 binding to phosphorylated Erk2 in the chromatin of cerebral neurons caused Erk-induced PARP1 activation, rendering transcription factors and promoters of immediate early genes (IEG) accessible to PARP1-bound phosphorylated Erk2. Thus, Erk-induced PARP1 activation mediated IEG expression implicated in long-term memory. PARP1 inhibition, silencing, or genetic deletion abrogated stimulation-induced Erk-recruitment to IEG promoters, gene expression and LTP generation in hippocampal CA3-CA1-connections. Moreover, a predominant binding of PARP1 to single-strand DNA breaks, occluding its Erk binding sites, suppressed IEG expression and prevented the generation of LTP. These findings outline a PARP1-dependent mechanism required for LTP generation, which may be implicated in long-term memory acquisition and in its deterioration in senescence.
PolyADP-ribose polymerases (PARPs) catalyze an abundant post-translational modification of nuclear proteins by polyADP-ribosylation. In this modification, NAD (Nicotinamide adenine dinucleotide) derived ADP-ribosyl moieties form ADP-ribose polymers on glutamate, lysine and aspartate residues of PARPs and their substrates1,2. Binding of the most abundant nuclear polyADP-ribose polymerase PARP1 to DNA single-strand breaks activates the protein and thereby triggers DNA base-excision repair1,2.
Recent findings implicated PARP1 in additional processes in the chromatin, including gene expression regulated by chromatin remodeling, DNA methylation or recruitment of transcription factors2,3,4,5,6. Moreover, alternative mechanisms of PARP1 activation in the absence of DNA damage were identified in a variety of cell types and cell-free systems. They include PARP1 activation by a variety of signal transduction mechanisms inducing intracellular Ca2+ release and activation of phosphorylation cascades2,7,8,9.
Numerous findings implicated the phosphorylation of extracellular signal regulated kinase-2 (Erk2) in synaptic plasticity and long-term memory10,11,12. Interestingly, recent in vivo experiments also revealed a pivotal role of PARP1 activation in long-term memory acquisition during learning13,14,15,16,17,18, but the explicit molecular mechanism underlying this un-expected role of PARP1 has not been identified.
Here, we disclose a molecular mechanism in the chromatin of cerebral neurons, which is activated by stimulation-induced Erk-PARP1 binding and synergistic activity required for immediate early genes (IEG) expression implicated in long-term memory. Furthermore, identified intra-molecular re-arrangements in DNA-bound PARP1 preventing its binding to phosphorylated Erk2, interfered with stimulation-induced IEG expression and LTP generation in the presence of DNA single-strand breaks, usually accumulated in aged irreplaceable cerebral neurons19,20.
PARP1-dependent long-term potentiation in the hippocampal CA3-CA1 connections
Long-term potentiation (LTP) in the hippocampal CA3-CA1 connections is currently used as a model for long-term memory21,22,23. In our experiments, field excitatory postsynaptic potentials (fEPSPs) were recorded from hippocampal slices of mice. Long-term potentiation in the hippocampal CA3-CA1 connections was induced by a brief high frequency stimulation of the Schaffer collaterals using two sets of bipolar electrodes placed on both sides and equidistant from the recording pipette, such that two independent stimulation channels were used for each slice (Methods).
To examine a possible effect of PARP1 on LTP, hippocampal slices were prepared from WT and PARP1 KO mice (Methods). LTP was generated in response to high frequency (100 Hz, 1 sec) tetanic stimulation in hippocampal slices of WT mice. However, there was a striking attenuation of the potential in the potentiated pathway in hippocampal slices of PARP1 KO mice. LTP was not generated in the hippocampal CA3-CA1 connections of PARP1-KO mice (Fig. 1a–c).
To examine a possible effect of PARP1 activity on LTP generation, PARP1 activity was blocked by the potent PARP inhibitors PJ-34 and ABT-888 (Fig. 1e,f, n = 7 and n = 5 slices, respectively). PJ-34 and ABT-888 were added at concentrations that inhibited polyADP-ribosylation of PARP1 in the cortex and hippocampus of rats15. PJ-34 and ABT-888 were added to the recording medium 5 min after tetanic stimulation to one pathway, and 30 minutes before similarly stimulating the second pathway (Methods; Fig. 1a,e,f). The tetanic stimulations produced a pathway-selective LTP before the application of PARP inhibitors (Methods). The first tetanic stimulation caused LTP, maintained for 70 minutes at 1.48 ± 0.004 and 1.54 ± 0.01 above baseline (average values calculated before application of PJ-34 and ABT-888, respectively). LTP induced in the first pathway was maintained stable even after application of PARP inhibitors. In contrast, tetanic stimulation delivered to the second pathway after 30 min perfusion of each PARP1 inhibitor failed to produce LTP (average values measured at the end of experiments with PJ-34 and ABT-888, 1.04 ± 0.01 and 1.07 ± 0.01 above baseline, respectively). Thus, each of the PARP inhibitors applied 30 min before stimulation completely prevented LTP development without affecting the already developed LTP or baseline responses. These results implicated PARP1 in the generation of LTP by tetanic stimulation of the Schaffer collaterals. The PARP1 inhibitors did not block nor attenuated excitatory postsynaptic NMDA current, which evokes LTP in the hippocampal CA3-CA1 connections23 (Fig. S1).
In view of a similar effect of MEK inhibitors on LTP generation in the hippocampal CA3-CA1 connections24 (Fig. S2), and accumulating findings implicating Erk-induced IEG expression in LTP and long-term memory aquisition25,26,27,28, we examined possible role of PARP1 in Erk-induced IEG expression.
A PARP1-dependent immediate early gene expression in response to high frequency stimulation
Stimulation inducing LTP is restricted to a small subset of afferents in the hippocampus21,23. It was impossible to isolate the stimulated neurons for examining biochemical signals associated with LTP. To overcome this difficulty, we used a model system of cultured cerebral neurons stimulated by electrical stimulation (Methods). High frequency (tetanic) stimulation (3 repeats of a 100 Hz, 1 sec duration pulse, followed by a 10 sec pause) applied to cultured cerebral neurons caused synaptic potentiation, indicated by pre-synaptic vesicles recycling (Fig. S3), which induces post-synaptic excitatory currents and synaptic long-term potentiation29,30.
Stimulation-induced expression of the immediate early genes c-fos, zif268 and arc that are implicated in LTP and long-term memory acquisition25,26,27,28, was measured by RT-PCR in stimulated cultured cerebral neurons, 8–10 days after plating (Fig. 2).
We found that only high frequency stimulation (3 repeats of a 100 Hz, 1 sec duration pulse, followed by a 10 sec pause) induced expression of c-fos, zif268 and arc in the cultured cerebral neurons within minutes after stimulation (Fig. 2a). The expression of arc lagged after zif268 expression, probably due to Zif268 (Egr1) acting as one of arc transcription factors28. Notably, the high frequency stimulation did not induce a non-specific Erk-dependent gene expression (eg., cJun31 was not expressed; Fig. S4).
The expression of c-fos, zif268 and arc in response to the high frequency stimulation was suppressed in cerebral neurons treated with each of the PARP inhibitors PJ-34 (10 μM) and Tiq-A (50 μM). In addition, their expression was similarly suppressed after PARP1 silencing (by siRNA, 150 nM, 72 hours; Fig. 2a,b), or PARP1 genetic deletion in cerebral neurons of PARP1-KO mice (Fig. 2c). These results supported a possible implication of PARP1 in the stimulation-induced expression of c-fos, zif268 and arc in the cerebral neurons. So, PARP inhibition, PARP1 silencing or its genetic deletion similarly interfered with stimulation-induced IEG expression in cerebral neurons and LTP induction in hippocampal CA3-CA1 connections (Figs 1 and 2).
A possible role of PARP1 activation in the recruitment of RNA-polII or transcription factors to the IEG promoters32 seemed unlikely in-view of recent evidence for RNA-polII poised in the promoter of arc33, and IEG transcription factors bound to CBP (CREB binding protein), with its HAT (histone acetyl-transferase) activity induced by their phosphorylation34. Instead, we examined a possible role of PARP1 in the phosphorylation of poised transcription factors, initiating the expression of c-fos, zif268 and arc in response to stimulation.
PARP1 binding to phosphorylated Erk2 and its activation in response to high frequency stimulation
Transcription factors of c-fos, zif268 and arc are activated by Erk-induced phosphorylation34,35,36,37,38. We therefore examined the effect of high frequency electrical stimulation on Erk phosphorylation. Rat brain cerebral neurons in primary cultures were stimulated by a variety of electrical stimulations (8–10 days after plating; Methods). Erk was phosphorylated in nuclei of cerebral neurons stimulated by a high frequency stimulation (3 repeats of a 100 Hz, 1 sec duration pulse, followed by a 10 sec pause), and phosphorylated Erk2 co-immunoprecipitated with PARP1 in nuclear protein extracts of the stimulated cerebral neurons (Fig. 3a). In addition, PARP1 and its prominent substrate linker histone H1, were highly polyADP-ribosylated (Fig. 3b). This finding was consistent with PARP1 activation in response to the high frequency electrical stimulation. PARP1 activation was identified by its immunolabeled ADP-ribose residues, and it was quantified by the shift in the isoelectric point (pI) of PARP1 and its substrate H1 towards lower pH concomitantly with its polyADP-ribosylation (Fig. 3b and S5; Methods). In un-stimulated neurons and in neurons stimulated by low frequency stimulations, Erk1/2 were hardly phosphorylated in the nuclear protein extracts, Erk2 did not co-immunoprecipitate with PARP1, and PARP1 and H1 were not polyADP-ribosylated (Fig. 3a,b). Furthermore, PARP1 activation by high-frequency stimulation was prevented in cerebral neurons treated with the specific MEK inhibitor U0126 (10 μM; Fig. 3b), similarly to PARP1 inhibition by its inhibitor PJ-34 (10 μM). These results support a linkage between PARP1 binding to phosphorylated Erk2 in the nuclear extracts and PARP1 activation in response to the high-frequency stimulation (further examined by bioinformatics methods), reminiscent of recombinant PARP1 activation by recombinant phosphorylated Erk2 in a cell-free system9.
Next, we examined possible effects of Erk-PARP1 binding on IEG expression in the stimulated cerebral neurons.
Phosphorylated Erk2 bound to activated PARP1 was recruited to promoters of immediate early genes c-fos and zif268 in response to high frequency stimulation
We used the ChIP assay to identify recruited proteins to promoters of the immediate early genes cfos and zif268 in response to stimulation. Chromatin cross-linking following stimulation (3 repeats of a 100 Hz, 1 sec pulse, followed by a 10 sec pause) revealed phosphorylated Erk2 and acetylated histone H4 co-immunoprecipitated with DNA segments in the promoters of c-fos and zif268 in cerebral neurons of WT mice (Fig. 4a). In addition, PARP1 was bound to phosphorylated Erk2 in the chromatin segments, and PARP1 inhibition did not impair their binding (Fig. 4b). However, phosphorylated Erk2 and acetylated H4 hardly co-immunoprecipitated with the promoters of cfos and zif268 after PARP1 inhibition, or PARP1 genetic deletion in stimulated cerebral neurons of PARP1-KO mice (Fig. 4a), indicating that binding of phosphorylated Erk2 to PARP1 was required for phosphorylated Erk2 access to the promoters of cfos and zif268 in the stimulated cerebral neurons.
These results suggest that PARP1 binding to phosphorylated Erk2 inducing PARP1 activation9 (Fig. 3, and the effect of PARP1-Erk2 binding on PARP1 activation further examined by bioinformatics methods) and polyADP-ribosylation of the PARP1 substrate linker histone H1, may facilitate H1 release from the DNA2, rendering IEG promoters accessible to PARP1-bound phosphorylated Erk2 (Fig. 4b,c). In support, PARP1-bound to phosphorylated Erk2 did not co-immunoprecipitate with its substrate H1, unless polyADP-ribosylation was inhibited (Fig. 4b).
This outlines a possible synergism between Erk-induced PARP1 activation and polyADP-ribosylation of linker histone H1 facilitating recruitment of phosphorylated Erk2 to transcription factors of cfos and zif268 (Fig. 4a). A PARP1-mediated phosphorylation of their transcription factors, inducing the HAT activity of CBP, and their binding to specific elements in the IEG promoters34,35 complies with co-immunoprecipitation of PARP-bound phosphorylated Erk2 and acetylated histone with DNA segments in the IEG promoters and with transcription factors Elk1 and CERB34,35,36,37,38 (Fig. 4a,b). Co-immunoprecipitation of phosphorylated Erk2 or acetylated H4 with DNA segments in the IEG promoters was prevented by PARP1 inhibition or its genetic deletion in cerebral neurons of PARP1-KO mice (Fig. 4a). In accordance, stimulation-induced expression of cfos and zif268 was prevented by PARP1 inhibition or its genetic deletion (Fig. 2).
Notably, phosphorylated Erk co-immunoprecipitated with its cytoplasmic/nuclear substrate, Rsk (ribosomal S6 kinase)36 in the chromatin of both WT and PARP1 KO mice (Fig. 4b), suggesting a possible PARP1-independent Erk-induced gene expression via Rsk phosphorylation36 in PARP1 KO mice.
Identified docking sites of phosphorylated Erk in PARP1
We searched PARP1 domains for binding sites of Erk. Dot-blot analysis and co-immunoprecipitation of recombinant domains of PARP1 with recombinant phosphorylated Erk2 disclosed an exclusive binding of recombinant phosphorylated Erk2 to the F-domain of PARP1 (aa556-1014), which contains its WGR, helical (HD), and catalytic (CAT) ADP-ribosyl transferase domains39 (Figs 5a,b and 6a). Recombinant phosphorylated Erk2 did not bind to PARP1 domains containing its DNA binding sites (Zn1-Zn2), nor to the auto-modification domain of PARP1 (aa1-494) (Fig. 5a,b). Phosphorylated Erk2 did not bind to [32P]ADP-ribose polymers, and polyADP-ribosylation did not prevent the binding of recombinant PARP1 to recombinant phosphorylated Erk2 (Fig. 5a).
A non-specific binding of Erk2 to recombinant PARP1 and its recombinant domains was excluded, as well as a possible binding of PARP1 and its F-domain to GST (glutathione S-transferase) attached as a fusion protein to recombinant phosphorylated Erk2 (Fig. 5a).
Notably, recombinant PARP1 did not co-immunoprecipitate with recombinant phosphorylated Erk2 in the presence of DNA damaged by single strand breaks (ssDNA; Methods) (Fig. 5c,d). However, their binding was restored after applying the recombinant DNA-binding domain of PARP1 (recombinant A-B domain containing its zinc fingers Zn1, Zn2; aa1-201) (Fig. 5d), possibly due to PARP1 displacement from its binding sites in nicked DNA40 (Fig. 5d). Furthermore, the recombinant F-domain of PARP1 interfered with the binding of recombinant PARP1 to recombinant phosphorylated Erk2 even in the absence of ssDNA, possibly due to their competition for common binding sites in phosphorylated Erk2 (Fig. 5a,b,d).
Next, residues in the F-domain of PARP1 (aa556-1014) were searched for known MAP kinases docking motifs41,42,43. Four sites on PARP1 partially match the known docking sites of MAP kinases in various proteins: 633KYPKK637, 683KK684, 747KKPPLL752 and 1007FNF1009 (Fig. 6a). In the helical domain of PARP1, 683 KK 684 partially matches docking motif for Erk41, 747 KKPPLL 752 matches docking motif DEJL for Erk42. In the catalytic domain of PARP1, 1007 FNF 1009 partially matches DEF docking motif (FXFP) for Erk43, and 633 KYPKK 637 in the WGR domain of PARP1 matches docking motif DEJL for Erk42. The domains in positively charged patches in the F-domain of PARP1 (aa633-637 and aa747-752; Fig. 6a) may bind to Erk via its negatively charged protein binding domain (CRS/CD region)41,44 (Fig. 6a). Indeed, mutations in the CRS/CD region of recombinant phosphorylated Erk2 interfered with the activation of recombinant PARP1 by recombinant phosphorylated Erk2 in a cell-free system9.
Notably, the recently disclosed PARP1 structural re-arrangements accompanying its binding to DNA39 occlude the indicated consensus docking sites of phosphorylated Erk in the F-domain of DNA-bound PARP1 (Fig. 6a). This may explain the failure of DNA-bound PARP1 to bind phosphorylated Erk2 (Fig. 5d), while its binding to histone H1 and other PARP1 substrates is not affected39,45 (Fig. 6c).
Intra-molecular dynamics in PARP1 bound to phosphorylated Erk2 can induce its activation
Intra-molecular dynamics in PARP1-bound to phosphorylated Erk2 homodimer was compared to intra-molecular dynamics in DNA-bound PARP1 by using the anisotropic network model (ANM)46 (http://ignmtest.ccbb.pitt.edu/cgi-bin/anm/anm1.cgi). This analysis was based on the potential Erk docking sites in the helical, catalytic and WGR domains of PARP1 (Fig. 6a), and on its predicted binding to homodimers of phosphorylated Erk in the nucleus9,47.
The resulting computed intra-molecular directions of motion in the combined complex of PARP1 bound to phosphorylated Erk2 expose the NAD binding site in the catalytic domain of PARP1 (Fig. 6b and S6). Exposure of its NAD binding site complies with the identified activation of Erk-bound PARP1 in stimulated cerebral neurons and in cell-free systems9 (Figs 3 and 6c, respectively).
The computed intra-molecular dynamics of PARP1 bound to phosphorylated Erk, exposing its NAD binding site anticipate polyADP-ribosylation of Erk-bound PARP1 (Fig. S6). This prediction is in consistence with the higher Erk-induced [32P]polyADP-ribosylation of recombinant PARP1 as compared to its DNA-induced [32P]polyADP-ribosylation at low [32P]NAD concentrations9 (Fig. 6c).
Thus, high frequency stimulation of cerebral neurons, inducing Erk phosphorylation and translocation to the nucleus47 may also induce PARP1 activation and PARP1-mediated IEG expression (Figs 2, 3, 4), unless the DNA is damaged by single strand breaks (Fig. 6). This notion was examined in stimulated cerebral neurons.
PARP1 binding to single-strand DNA breaks interfered with IEG expression
The expression of cfos and zif268 was measured by RT-PCR in cerebral neurons of PARP1- KO mice that were stimulated (100 Hz, 1 sec, 3 repeats, 10 sec pause) 72 hours after transfection with GFP-fusion vectors with constructs encoding full length PARP1 or PARP1 lacking its DNA binding domain (aa1-221; Methods). Expression of c-fos and zif268 was measured in the re-plated GFP-labeled transfected cerebral neurons (Methods). Cerebral neurons of PARP1-KO mice hardly expressed c-fos and zif268 (Fig. 7a). However, these genes were expressed in stimulated cerebral neurons of PARP1-KO mice transfected with either full length PARP1 or PARP1 lacking its DNA binding domain, evidence that Erk binding domains in PARP1 (but not its DNA binding domain) were necessary for stimulation induced cfos and zif268 expression (Fig. 7a).
Some of the transfected PARP1-KO cerebral neurons were treated before stimulation with H2O2 (1 mM, 10 min) causing single strand DNA breaks (Figs 7c and 8b). As a consequence, the expression of cfos and zif268 was very low in response to stimulation in PARP1-KO neurons transfected with full length PARP1, but was not impaired in PARP1-KO neurons transfected with PARP1 lacking its DNA binding domain (Fig. 7a). Thus, PARP1 binding to nicked DNA was required for preventing cfos and zif268 expression in the presence of single-strand DNA breaks.
In compliance, a brief pre-incubation of cultured rat cerebral neurons with H2O2 (1 mM; 10 min), or their exposure (60 min) to hypoxia causing DNA single-strand breaks (Fig. 7c,e) down-regulated cfos and zif268 expression and the synthesis of proteins/ transcription factors c-Fos, Zif268 and Arc following high-frequency stimulation (Fig. 7b,e).
Protein synthesis was monitored in stimulated cerebral neurons without or following treatment with H2O2. These neurons were stimulated by electrical stimulation (3 repeats of 100 Hz, 1 sec duration, each followed by 10 sec pause), without or following treatment with the nerve growth factor NGF ( 60 ng/ml, 5 min) (Fig. 7b). The effect of NGF, also inducing cfos and zif268 expression17, was examined because electrical stimulation was technically impossible under hypoxia (Fig. 7b,e).
Treatment with H2O2 causing single strand breaks, extensively attenuated the stimulation-induced synthesis of proteins c-Fos, Zif268 and Arc, unless cerebral neurons were pre-treated with the PARG (polyADP-ribose glycohydrolase) inhibitor gallotannin (100 μM, 60 min)48 (Fig. 7b).
PARG cleaves the negatively charged ADP-ribose polymers of PARP1, enabling its recurrent binding to the negatively charged DNA2,48. Thus, PARG inhibition interferes with PARP1 binding to nicked DNA2,48.
Assuming that polyADP-ribosylation does not prevent the binding of PARP1 to phosphorylated Erk2 (Figs 3,4 and 5a), application of PARG inhibitors might preserve the binding of PARP1 to phosphorylated Erk2 by preventing PARP1 binding to single-strand DNA breaks (Fig. 7b). This assumption was examined in a cell-free system by measuring the dose-dependent effect of recombinant PARP1 polyADP-ribosylation on its binding to recombinant phosphorylated Erk2 in the presence of nicked DNA and βNAD (Fig. 7d).
The results indicated that binding of recombinant PARP1 to recombinant phosphorylated Erk2 in the presence of nicked DNA (ssDNA) was dependent on the intensity of PARP1 polyADP-ribosylation (Fig. 7d). The more intensely was PARP1 polyADP-ribosylated, the better it co-immunoprecipitated with phosphorylated Erk2 in the presence of ssDNA (Fig. 7d). This result complied with the preserved expression of cfos, zif268 and arc in the presence of nicked DNA in stimulated cerebral neurons treated with gallotannin (Fig. 7b,e).
Single-strand DNA breaks prevented LTP generation
Treatment causing DNA single strand breaks exclusively prevented LTP generation in stimulated hippocampal CA3-CA1 connections (Methods; Fig. 8). Notably, the baseline response was not affected, nor already generated LTP (Fig. 8a).
LTP generation in response to a brief high frequency stimulation of the Schaffer collaterals (100 Hz, 1 sec) was exclusively prevented by treatment inducing single-strand breaks (Fig. 8a).
The binding of phosphorylated Erk2 to PARP1 in cell nuclei prepared from the hippocampal slices was also prevented under these conditions (Fig. 8b). PARP1-Erk2 binding was examined in hippocampal slices briefly stimulated by high K+ induced depolarization49,50. This stimulation was used to enhance biochemical processes induced by the tetanic stimulation of the Schaffer collaterals21,22,23. PARP1-Erk2 co-immunoprecipitation was measured in cell nuclei prepared from hippocampal slices briefly exposed to high K+ (1 min wash with 50 mM K+ ACSF50), before and following treatment with H2O2 (1 mM; 15 min) (Fig. 8b). PARP1 co-immunoprecipitated with phosphorylated Erk2 only in the chromatin of stimulated hippocampal slices that were not treated with H2O2 (Fig. 8b). This result is consistant with depolarization-induced PARP1-Erk2 binding prevented by treatment causing DNA single strand breaks, which also prevented LTP generation in response to high frequency stimulation (Fig. 8a).
Ex vivo and in vivo experiments implicated Erk2-induced expression of specific immediate early genes in synaptic plasticity and long-term memory25,26,27,28,36,51. Our results suggest that Erk2-induced PARP1 activation mediates this activity of Erk. These results comply with the dependence of long-term memory acquisition during training on PARP1 activation13,14,15,16,17,18.
PARP inhibition did not affect excitatory post-synaptic NMDA currents (Fig. S1), inducing LTP in the hippocampal CA3-CA1 connections21,23. However, PARP1 was implicated in nuclear processes immediately following high-frequency stimulation inducing synaptic potentiation. These processes were examined in a model system of electrically stimulated cultured cerebral neurons29,30 (Fig. S3). High frequency stimulation induced binding of phosphrylated Erk2 to PARP1 in the chromatin of cerebral neurons (Figs 3 and 4), concomitantly with Erk-induced PARP1 activation9 (Figs 3b and 6, S6), polyADP-ribosylated linker histone H1 (Fig. 4b), and facilitated access of PARP1-bound phosphorylated Erk2 and acetylated histone H4 to promoters of immediate early genes cfos and zif26825,26,27,28 (Fig. 4a). PARP1-dependent access of phosphorylated Erk2 and acetylated H4 to the promoters of c-fos and zif268 and to their transcription factors Elk1 and CREB were identified by ChIP assay (Fig. 4). These results complied with PARP1-dependent IEG expression34,35,36,37,38 (Figs 2 and 7a).
In accordance, cfos, zif268 and arc expression was suppressed after PARP1 inhibition or its genetic deletion in cerebral neurons of PARP1-KO mice (Figs 2,4 and 7a). Furthermore, LTP was not generated after PARP1 or MEK inhibition (Fig. 1 and S2), or after PARP1 genetic deletion in the hippocampal CA3-CA1 connections of PARP1-KO mice (Fig. 1c), which indeed do not acquire long-term memory during training14.
Thus, PARP1-Erk binding promoting IEG expression (Figs 2, 3, 4 and 7a,d), promoted the forthcoming protein synthesis implicated in synaptic plasticity or long-lasting synaptic potentiation21,23,51. Recent findings indicating protein synthesis during the early phase of high-frequency induced LTP52 may support gene expression during early LTP. However, these results do not exclude extra-nuclear processes implicated in early LTP23.
Notably, LTP failed to develop in response to high frequency stimulation after treatment causing single strand DNA breaks preventing PARP1-Erk2 binding (Fig. 8), and abrogating the expression of c-fos and zif268 due to a predominant binding of PARP1 to single strand DNA breaks occluding its potential Erk binding sites (Figs 5, 6a and 7a,d).
Furthermore, LTP generated in response to high frequency stimulation 5–10 min before application of MEK or PARP1 inhibitors, was maintained intact (Figs 1 and S2). Moreover, generated LTP remained intact despite producing DNA single-strand breaks (Fig. 8). These results suggest that the brief effects of PARP1-Erk2 binding and their synergistic activity in the chromatin were required for the forthcoming LTP generation (Fig. S7).
The interference of DNA single strand breaks with IEG expression (Fig. 7) may attribute the previously observed aging-induced attenuation in gene expression53 to the accumulation of DNA single strand breaks in aged irreplaceable neurons19,20,53,54.
The DNA of mammalian cerebral neurons is constantly exposed to damaging processes, mostly by reactive oxygen species (ROS), which are normally produced in their mitochondria due to high-energy demands55. ROS cause single strand DNA breaks by oxidative reactions with the nucleic acids55. Thus, age-induced decline of antioxidant defensive mechanisms, the inability to replace aged neurons and the constant exposure of their DNA to oxidative stress during their life span, cause accumulation of single strand breaks in the DNA of cerebral neurons in senescence, despite the existing DNA repair mechanisms19,20.
Furthermore, failure to generate LTP due to accumulating DNA single-strand breaks in aged cerebral neurons (Fig. 8) could be implicated in the deterioration of memory acquisition and learning abilities, frequently experienced in senescence19,20,53. Thus, deterioration in learning abilities might not necessarily reflect death of cerebral neurons. It could result from the accumulation of amendable single-strand DNA breaks in aged irreplaceable cerebral neurons interfering with LTP generation (Figs 6, 7, 8). In this case, memory acquisition could be improved by attenuating the binding of PARP1 to nicked DNA (Figs 7 and 8). In support, recent evidence indicated an improved long-term memory acquisition of aged mice treated with the PARG inhibitor gallotannin48,56.
Notably, PARP1 inhibitors impaired long-term memory acquisition of trained animals only when administered at least 30 min before training13,15. Their application after training did not affect the already acquired memory of the trained animals13. Similarly, PARP inhibitors prevented the induction of LTP only when applied before stimulation (Fig. 1e,f). These findings may suggest a possible use of PARP1 inhibitors for erasing a specific memory without affecting past memories or learning abilities.
In summary, the presented findings disclose a molecular mechanism in the chromatin of cerebral neurons, which is necessary for LTP generation, and can be manipulated by pharmacological interventions.
Antibodies and recombinant proteins used in the presented experiments
PARP1 and its recombinant domains were immunolabeled by the monoclonal antibody (Serotec, Cat # MCA1522; Oxford, UK) and the polyclonal antibody (Alexis, Cat # ALX210-302). Erk2 was immunolabeled by antibody directed against the c-terminal of Erk2 (#sc-154; Santa Cruz Biotechnology, CA, USA), phosphorylated Erk1/2 (Sigma), Elk1 and phosphorylated Elk1 (Cell Signaling Technologies, MA, USA). Antibodies directed against acetylated histones H3 and H4 were from Upstate Biotechnology (Millipore) CA, USA. Antibodies directed against transcription factors c-Fos (Cell Signaling Technologies), Egr1 (Zif268; Cell Signaling Technologies), phosphorylated CREB (phosphorylation of serine-133; Cell Signaling Technologies) and Arc (Novous Biologicals, Cambridge, UK). For cytochemistry, first antibodies were labeled by fluorescent secondary antibody: CyTM2 (green) or fluorescent CyTM3 (red) conjugated affinity pure goat-anti-rabbit or goat anti-mouse secondary antibodies (Jackson ImmunoResearch). Recombinant proteins: Elk1 (Elk1 residues 307–428 coupled to GST; Cell Signaling Technologies), recombinant human PARP1 was commercial (Alexis, Enzo Life sciences, NY,USA) or prepared by Dr John Pascal, Thomas Jefferson University, Philadelphia, USA, recombinant PARP1 domains were prepared in the lab of Dr Francoise Dantzer (Strasbourg, France) recombinant PARP1 (1-494aa) was prepared in the lab of Dr. John Pascal, as well as constructs in a GFP fusion vector of full-length PARP-1 and PARP-1 residues 201- 1014aa for expression in cultured neurons of PARP1 KO mice. Recombinant phosphorylated Erk2 was prepared in the lab of Prof Seger, Weizmann Institute of Science, Rehovot.
Primary cell cultures were prepared from brain cortex and hippocampus (cerebral neurons) of 18 to 19 day rat or mice embryos, as described before57. Experiments were conducted according to rules and regulations of Institutional Animal Care and Use Committee.
Nuclear protein extracts
Cell nuclei were isolated from cultured cerebral neurons as described before57. Nuclear proteins were extracted after incubation (30 min on ice) in a high salt concentration buffer, containing 0.42 M NaCl, 1.5 mM MgCl2, 0.2 mM EDTA, 25% glycerol, 20 mM Tris-HCl pH 8.0, protease and phosphatase inhibitors. Supernatants obtained after centrifugation (15,000 rpm 4 °C, 15 min) contained extracted nuclear proteins.
Electrophysiology in hippocampal slices
The methods of recording from hippocampal slices were described before23,58. Briefly, male 129/Sv mice (2–2.5 month-old) were rapidly decapitated and their brains were removed and placed in ice cold ACSF containing (mM) 124 NaCl, 2 KCl, 26 NaHCO3, 1.24 KH2PO4, 2.5 CaCl2, 2 MgSO4 and 10 glucose, at pH7.4. The hippocampi were cut into 350–400 μm transverse slices using a McIlwain tissue chopper. Slices were incubated for 1.5 h in carbogenated (5% CO2 and 95% O2) ACSF at room temperature in a holding chamber. Recording was made from slices that are slightly submerged in a standard chamber at 33.8–34.0 °C with a flow rate of 2.5 ml ACSF/min. Field excitatory postsynaptic potentials (fEPSPs) were recorded in stratum radiatum of the CA1 region of hippocampal slices through a glass pipette containing 0.75 M NaCl (4 MΩ). Synaptic responses were evoked by stimulation of the Schaffer collaterals using two sets of bipolar electrodes placed on both sides and equidistant from the recording pipette, such that two independent stimulation channels were used for each slice23,58 (Fig. 1a). LTP was induced by high-frequency stimulation (100 Hz, 1 sec). Before applying the stimulation, evoked fEPSPs (50% of maximum amplitude) were recorded for a stable baseline period of at least 10 min. Stimulation of one pathway did not cause any noticeable change in response to stimulation of the second pathway, verifying their independence58. Data acquisition and off-line analysis were performed using pCLAMP 9.2 (Axon Instruments, Inc). All numerical data are expressed as mean ± SEM, and fEPSP slope changes after stimulation and drug application were calculated with respect to baseline. PARP1(−/+) 129/Sv mice were donated by Dr Dantzer (Strasbourg) and bred for PARP1 (−/−) mice in Cohen-Armon’s lab under the rules and regulations of the Institutional Animal Care and Use Committee.
The effect of H2O2 on LTP was examined as follows: After 20 minutes of baseline recording, first tetanic stimulation (100 Hz, 1 sec) was applied to pathway 1, which resulted in LTP of a magnitude of 1.64 ± 0.004. H2O2 (at final concentration in ACSF 1 mM, Sigma Aldrich) was added for 15 minutes after potentiated pathway has stabilized. Its application did not affect either magnitude of already established LTP or baseline responses of pathway 1. Following 5 minutes of washout of H2O2, a tetanus was delivered to pathway 2 to investigate H2O2 impact on LTP induction. Post-tetanic long-term potentiation failed to develop (LTP 1.12 ± 0.01, p < 0.001 was measured). After 1 hour of recording and adjustment of stimulation intensity of both pathways to baseline level, a second tetanic stimulation was applied to each pathway. LTP did not developed in both pathways (LTP 1.19 ± 0.01 ;p < 0.001 was measured in pathway 1 and 1.05 ± 0.01 p < 0.05 in pathway 2, correspondingly).
Electrical stimulation (‘bath stimulation’) of cerebral neurons in primary culture was applied by a pulse generator controlled by pCLAMP 6.0 (Axon Instruments, Inc.) and a digital to analogue conversion (D/A 1200 Digidata), as described before7. Usually, train of pulses (0.5 msec) of 1 sec duration, 1–100 Hz frequency, was repeated 3 times, each followed by 10 sec pause intervals. Pulse amplitude was the minimal voltage required for a break of action potential in randomly chosen neurons in the culture, as described before7. About 30 Volt was applied to the bath solution (5 ml growth medium containing MEM eagle enriched by 2 mM Glutamax, 0.6% glucose, 5% Horse Serum, 20 μg/ml gentamycin). This stimulation caused pre-synaptic vesicles recycling (Fig. S3).
Culturing neurons on glia cells
Mouse postnatal cultures were platted on glia cells prepared from rat E19 embryos, as detailed before59. Glia cells proliferated for 10 days before plating the mouse culture. This procedure was used for re-plating of transfected neurons and for plating cerebral neurons of PARP1 KO mice.
PARP1 activation in cerebral neurons
Two-dimensional (2-D) gel electrophoresis was used to identify stimulation-induced activation of the positively charged DNA-binding protein PARP1 by the shift in its isoelectric point (pI) towards lower pH, due to polyADP-ribosylation adding negatively charged phosphates to PARP1. This method was used to estimate PARP1 activation in situ13,15,57. In support, PARP1 activation was measured by the shift in the pI of [32P]polyADP-ribosylated PARP1 in isolated nuclei of stimulated cerebral neurons incubated with [32P]NAD (1 μCi/sample; 1000 mCi/mmol; Amersham, UK) (Fig. S4).
For RT-PCR profiling
we used RNeasy Plus mini kit (Qiagene, CA, USA) for RNA preparation, and we used RevertAid First Strand cDNA Synthesis Kit #K1622 (Thermo scientific) for cDNA preparation. Primers that initiated amplification of the indicated cDNA segments in the rat genes cfos, Zif268 and arc (forward and reverse) were: for c-fos, 5′GTTCCTGGCAATAGTGTGTTC3′ and 5′GCTGAAGAGCTACAGTACGTG3′, for arc, 5′TGGAGTCTTCAGACCAGGTG3′ and 5′GCTGGCTTGTCTTCACCTTC3′, for zif268, 5′CAGGAGTGATGAACGCAAGA3′ and 5′AGCCCGGAGAGGAGTAAGTG3′, for c-jun1, 5′TGAGAACTTGACTGGTTGCG3′ and 5′CAGGTGGCACAGCTTAAACA3′. For the control gene GAPDH; 5′CTGGAAAGCTGTGGCGTGATGG3′ and 5′TCCTCAGTGTAGCCCAG GATGC3′ and for the control gene β-actin, 5′AGAGCTATGAGCTGCCTGAC3′ and 5′AATTGAATGTAGTTTCATGGATG3′.
Primers that initiated amplification of the indicated cDNA segments in the genes c-fos, zif268 and arc (forward and reverse) in mice were: c-fos forward 5′TCCGGGCTGCACTACTTA3′ and reverse 5′TGTTTCACGAACAGGTAAGGT3′, respectively. For zif268, 5′GTGTGGTGGCC TCCCCGGCT3′ and 5′CACTGACGGCGACGGGAAGCC3′, respectively. For arc, 5′GCCACAAATGCAGCTGAAGCAG3′ and 5′GTGGTGTGGTGATGCCCTTTCC3′, respectively. For the control gene β-actin forward 5′GGGCTGTATTCCCCTCCAT3′ and reverse 5′GCGTGAGGGAGAGCATAGC3′, respectively.
Chromatin immunoprecipitation (ChIP) assay
We used the ChIP assay protocol5 to identify binding of phosphorylated Erk2 and Acetylated H4 (AcH4) to promoters of c-fos, and zif268. DNA-bound proteins were crosslinked to the DNA of stimulated mice cerebral neurons (by formaldehyde 1%) at different intervals after stimulation. The crosslinked chromatin was cleaved into segments of approximately 1000-bp by sonication on ice (Probe Sonicator; Heat Systems Inc., Farmingdale, USA). Promoters and transcription factors were co-immunoprecipitated by antibody directed against AcH4 (#06-866 anti-acetyl-H4 antibody directed against epitope aa2-19 in H4 acetylated on lysines 5,8,12,16; Upstate Biotechnology (Millipore) CA, USA), or by antibody directed against the c-terminal of Erk2 (#sc-154; Santa Cruz Biotechnology, CA, USA). Both DNA and proteins were recovered from the crosslinked chromatin segments after co-immunoprecipitation as described before5. For DNA isolation, formaldehyde cross-linking was reversed by heating (65 °C, 2 h). After protein digestion, DNA was purified on Zymo-SpinTM columns (ChIP DNA Clean & Concentrator kit, Zymo research corp.). DNA segments in the promoters were amplified by RT-PCR by using the following primers: for the promoter of c-fos, primers 5′GTGCTGCCGTCCTTTAAAAC-3′ and 5′-GAGAGAGGGGCTGAGAAGCT-3′ (amplified segment 60–204) and primers 5′CTGCACTGATTTGGGATGGG-3′ and 5′-TAGGAGAAGCAAGTACGCAGC3′ (amplified segment 98–150). For the promoter of Zif268 5′-TGGGGCTCCCGAAATACAAC-3′ and 5′-AAGAGGGGGACTTGGCTTTG-3′ (amplified segment 382–395), and primers 5′-AGGACGGAGGGAATAGCCTT-3′ and 5′-ACTGGTTC TTGGGACACTGC-3′ (amplified segment 659–787). Proteins were recovered from the croslinked chromatin after 15 boiling in sample buffer.
Expression of PARP1 in PARP1-KO cerebral neurons
Cerebral neurons of PARP1-KO mice were transfected 24 hours after plating with two plasmids (in mammalian expression vector (pEGFP-N1) encoding GFP-fusion full-length PARP1 or GFP-fusion PARP1 lacking residues aa1-201 (lacking the DNA binding zinc fingers domain), which were prepared in Dr John Pascal Lab (Jefferson University, Philadelphia). The expression of c-fos and zif268 was measured by RT-PCR in transfected GFP-labeled KO cerebral neurons sorted by FACS, 60–72 hours after transfection. The transfected neurons were re-plated on cultured rat glia cells and stimulated by bath stimulation, without or after treatment with H2O2 (1 mM, 10 min).
DNA isolation and detection of DNA breaks
DNA was isolated from the nuclei of cultured neurons using the PureLink genomic DNA kit (Invitrogen, Cat # K1820-01). Single strand DNA breaks were identified on alkali agarose gels containing 1% agarose, 50 mM NaCl, 1 mM EDTA, soaked for 60 min with 30 mM NaOH and 1 mM EDTA, as described before7. Double strand DNA breaks were detected in 1% agarose gel at pH 7.4. The migration of DNA in 1% agarose gel was detected by staining under UV illumination. In cell-free experiments we used commercial ssDNA (salmon sperm DNA) carrying numerous single strand breaks (Sigma).
PARP1 silencing by siRNA
This method was described before5,9. Two sequences, aa800-807 and aa890-897, in the PARP1 catalytic domain were targeted for PARP1 silencing. PARP1 targeted siRNA was prepared by Darmacon (Lafayette CO, USA). For control we used the non specific siRNA#2 (non-spec. rat siRNA; Darmacon). Cerebral neurons were transfected by XtremeGENE siRNA transfection reagent (Cat no. 04476093001, Roche Diagnostic, GmbH Mannheim, Germany). PARP1 silencing was achieved 72 hours after transfection with 100–200 nM siRNA.
Bioinformatic analysis of PARP1 binding to phosphorylated Erk2
Identified docking sites of Erk on the F-domain of PARP141,42,43,44: Phosphorylated Erk2 homodimer47 was reconstructed from the crystal contact interface in PDB (Protein Data Bank). Phosphorylated Erk2 homodimer was docked on the helical, catalytic and WGR domains of PARP1 (PDB 4DQY). Details are included in Supplementary Methods.
Treatment with H2O2
Cerebral neurons in cell culture (10 days after plating), and hippocampal slices were exposed to H2O2 (1 mM, 10–15 min), and then thoroughly washed, as described before7.
Cerebral neurons under hypoxia conditions
Cultured rat cerebral neurons were exposed after over-night starvation (MEM-Eagle growth medium containing 0.5% Horse serum instead of 5% in normal growth medium, 0.6% glucose, 2 mM Glutamax and 20 μg/ml Gentamycin) to hypoxia at 37 °C for 60 min. Hypoxia was imposed after replacing the normal atmosphere in a close chamber with 100% Argon. A similar procedure was described before60.
was used to identify bound recombinant proteins or nuclear proteins as described before5,9. Binding to specific antibodies trapped the proteins on Protein A/G Agarose Beads (1 h, 4 °C). The bound proteins were recovered (1–2 min, boiling in sample buffer) separated on polyacrylamide SDS gel and immunodetected on Western blots.
Dot Blot analysis searching PARP1 domains binding phosphorylated Erk2
Binding of recombinant phosphorylated Erk2 (1 μg) to recombinants of PARP1 and to recombinant domains of PARP1 was examined by dot-blot analysis. In addition, binding of phosphorylated Erk2 to polyADP-ribosylated recombinant human PARP1 and free [32P]-labeled poly(ADP-ribose were examined. The blots were blocked in ‘binding buffer’ (50 mM Tris-HCl pH7.5, 120 mM NaCl, 0.1% NP40, 0.5 mM PMSF, 20 mg/ml BSA) for 30 min at room temperature. These blots were incubated for 3 h at room temperature in ‘binding buffer’ containing 5 μg/ml of each of the purified recombinant: human PARP1, recombinant domains of human PARP1, polyADP-ribosylated human PARP1, or 2 μg/ml of free [32P]polyADP-ribose polymers. The blots were then washed with TBS-Tween-20 0.1%. Anchored proteins were detected on the nitrocellulose membrane with the appropriate antibodies. Binding to free [32P] poly(ADP-ribose) was measured by autoradiography.
All the methods were carried out in accordance with the approved guidelines. All experimental protocols were approved by the Institutional Animal Care and Use Committees of the Sheba Medical Center and the Tel-Aviv University.
How to cite this article: Visochek, L. et al. A PARP1-ERK2 synergism is required for the induction of LTP. Sci. Rep. 6, 24950; doi: 10.1038/srep24950 (2016).
Schreiber, V., Datzer, F., Amë, J.-C. & de Murcia, G. Novel functions for an old molecule. Nat. Rev. Mol. Cell Biol. 7, 517–528 (2006).
Gibson, B. A. & Kraus, W. L. New insights into the molecular and cellular functions of poly(ADP-ribose) and PARPs. Nat Rev Mol Cell Biol. 13, 411–424 (2012).
Caiafa, P., Guastafierro, T. & Zampieri, M. Epigenetics: poly(ADP-ribosyl)ation of PARP-1 regulates genomic methylation patterns FASEB J. 23, 672–678 (2009).
Ohlsson, R., Lobanenkov, V. & Klenova, E. Does CTCF mediate between nuclear organization and gene expression? Bioessays 32, 37–50 (2010).
Geistrikh, I. et al. Ca2+ induced PARP-1 activation and ANF expression are coupled events in cardiomyocytes Biochem J. 438, 337–347 (2011).
Matveeva, E. et al. Involvement of PARP1 in the regulation of alternative splicing. Cell Discovery 2, 15046, doi: 10.1038/celldisc.2015.46 (2015).
Homburg, S. et al. A fast signal- induced activation of poly(ADP-ribose) polymerase: A novel downstream target of phospholipase C. J. Cell Biol. 150, 293–308 (2000).
Ju, B. G. et al. Activating the PARP-1 sensor component of the groucho/ TLE1 corepressor complex mediates a CaM-Kinase Iidelta-dependent neurogenic gene activation pathway. Cell 119, 815–829 (2004).
Cohen-Armon, M. et al. DNA-Independent PARP-1 Activation by Phosphorylated ERK2 Increases Elk1 Activity: A Link to Histone Acetylation. Mol Cell 25, 297–308 (2007).
Sweatt, D. J. Mitogen activated protein kinases in synaptic plasticity and memory Cur Opin Neurobiol. 14, 311–317 (2004).
Samuels, I. S. et al. Deletion of Erk2-mitogen-activated protein kinase identifies its key roles in cortical neurogenesis and cognitive functions. J. Neurosci. 28, 6983–6995 (2008).
Maharana, C., Sharma, K. P. & Sharma, S. K. Feedback mechanism in depolarization-induced sustained activation of extracellular signal regulated kinase in the hippocampus. Scientific Rep. 3, 1103 (2013).
Cohen-Armon, M. et al. Long-term memory requires polyADPribosylation. Science 304, 1820–1823 (2004).
Piskunova, T. S. et al. Deficiency in Poly(ADP-ribose) Polymerase-1 (PARP-1) Accelerates Aging and Spontaneous Carcinogenesis in Mice. Curr Gerontol Geriatr Res. 754190, doi: 10.1155/2008/754190 (2008).
Goldberg, S., Visochek, L., Giladi, E., Gozes, I. & Cohen-Armon, M. PolyADP-ribosylation is required for long-term memory formation in mammals. J. Neurochem. 111, 72–79 (2009).
Hernandez, A. I. et al. Poly(ADPribose) polymerase-1 is necessary for Long-Term Facilitation in Aplysia . J. Neurosci. 29, 9553–9562 (2009).
Wang, S.-H. et al. NGF promotes long-term memory formation by activating poly(ADP-ribose)polymerase-1. Neuropharmacology 63, 1085–1092 (2012).
Fontán-Lozano, A. et al. Histone H1 polyADP-ribosylation regulates the chromatin alterations required for learning consolidation. J Neurosci. 30, 13305–13313 (2010).
Lu, T. et al. Gene regulation and DNA damage in the ageing human brain. Nature 429, 883–891 (2004).
Kumar, A. Long-term potentiation at CA3-CA1 hippocampal synapses with special emphasis on aging, disease and stress Frontiers. in Aging Neurosci. 3, 2–20 (2011).
Bliss, T. V. P. & Collingridge, G. L. A synaptic model of memory: long-term potentiation in the hippocampus Nature 361, 31–39 (1993).
Albensi, B. C., Oliver, D. R., Toupin, J. & Odero, G. Electric stimulation protocols for hippocampal synaptic plasticity and neuronal hyper-excitability - are they effective or relevant? Exp. Neurol. 204, 1–13 (2007).
Sala, C. & Segal, M. Dentritic spines: The locus of structural and functunal plasticity. Physiol Rev. 94, 141–188 (2014).
English, J. D. & Sweatt, D. J. A requirement for mitogen activated protein kinase cascade in hippocampal long-term potentiation. J. Biol. Chem. 272, 19103–19106 (1997).
Flavell, S. W. & Greenberg, M. E. Signaling mechanisms linking neuronal activity to gene expression and plasticity of the nervous system. Annu Rev Neurosci. 31, 563–590 (2008).
Loebrich, S. & Nedivi, E. The function of activity-regulated genes in the nervous system. Physiol. Rev. 89, 1079–1103 (2009).
Jones, M. W. et al. A requirement for the immediate early gene Zif268 in the expression of late LTP and long-term memories Nat. Neurosci. 4, 289–296 (2001).
Clark, P. J., Bhattacharya, T. K., Miller, D. S. & Rhodes, J. S. Induction of c-Fos, Zif268, and Arc from acute bouts of voluntary wheel running in new and pre-existing adult mouse hippocampal granule neurons. Neuroscience 184, 16–27 (2011).
Bi, G. & Poo, M. Synaptic Modifications in Cultured Hippocampal Neurons: Dependence on Spike Timing, Synaptic Strength, and Postsynaptic Cell Type. J. Neurosci. 18, 10464–10472 (1998).
Tao, H.-Z. W., Zhang, L. I., Bi, G.-Q. & Poo, M. Selective presynaptic propagation of long-term potentiation in defined neural networks. J. Neurosci. 20, 3233–3243 (2000).
Leppa, S., Saffrich, R., Ansorge, W. & Bohmann, D. Differential regulation of c- Jun by ERK and JNK during PC12 cell differentiation. EMBO J. 17, 4404–4413 (1998).
Oei, Li S., Griesenbeek, J., Schweiger, M. & Ziegler, M. Regulation of RNA polymeraseII-dependent transcription by polyADP-ribosylation of transcription factors. J Biol Chem. 273, 31644–31647 (1998).
Saha, R. N. et al. Rapid activity-induced transcription of Arc and other IEGs relies on poised RNA polymerase-II. Nat. Neurosci. 14, 848–856 (2011).
Buchwalter, G., Gross, C. & Wasylyk, B. Ets ternary complex transcription factors. Gene 324, 1–14 (2004).
Besnard, A., Gala-Rodriguez, B., Vanhoutte, P. & Caboche, J. Elk1 a transcription factor with multiple facets in the brain. Frontiers in Neurosci. 5, 35, doi: 10.3389/fnins.00035 (2011).
Thomas, G. M. & Huganir, R. L. MAPK cascade signaling and synaptic plasticity. Nature 5, 173–183 (2004).
Li, Q. J. et al. MAP kinase phosphorylation-dependent activation of Elk-1 leads to activation of the co-activator p300. EMBO J. 15, 281–291 (2003).
Korzus, E., Rosenfeld, M. G. & Mayford, M. CBP histone acetyltransferase activity is a critical component of memory consolidation. Neuron 42, 961–972 (2004).
Langelier, M.-F., Planck, J.-L., Roy, S. & Pascal, J. M. Structural Basis for DNA Damage– Dependent Poly(ADP-ribosyl)ation by Human PARP-1. Science 336, 728–732 (2012).
Ali, A. A. et al. The zinc-finger domains of PARP1 cooperate to recognize DNA strand breaks. Nat. Struct. Mol. Biol. 19, 685–694 (2012).
Tanoue, T., Adachi, M., Moriguchi, T. & Nishida, E. A conserved docking motif in MAP kinases common to substrates activators and regulators. Nat. Cell Biol. 2, 110–116 (2000).
Jacobs, D., Glossip, D., Xing, H., Muslin, A. J. & Kornfeld, K. Multiple docking sites on substrate proteins form modular system that mediates recognition by Erk MAP kinase. Gene Dev. 13, 163–175 (1999).
Fantz, D. A., Jacobs, D., Glossip, D. & Kornfeld, K. Docking sites on substrate proteins direct extra-cellular signal regulated kinase to phophorylate specific residues J. Biol. Chem. 276, 27256–27265 (2001).
Tanoue, T. & Nishida, E. Docking interactions in the mitogen-activated protein kinase cascades. Pharmacology and Therapeutics 2–3, 193–202 (2002).
Langelier, M. F., Planck, J. L., Roy, S. & Pascal, J. M. Crystal structures of poly(ADP-ribose) polymerase1 (PARP1) zinc fingers bound to DNA: structural and functional insights into DNA-dependent PARP1 activity. J. Biol. Chem. 286, 10690–10701 (2011).
Atilgan, A. R., Durell, S. R., Jernigan, R. L., Demirel, M. C., Keskin, O. & Bahar, I. Anisotropy of fluctuation dynamics of proteins with an elastic network model. Biophys. J. 80, 505–515 (2001).
Khokhlatchev, A. V. et al. Phosphorylation of the MAP Kinase ERK2 promotes its homodimerization and nuclear translocation. Cell 93, 605–615 (1998).
Finch, K. E., Knezevic, C. E., Nottbohm, A. C., Partlow, K. C. & Hergenrother, P. J. Selective small molecule inhibition of poly(ADP-ribose) glycohydrolase (PARG) ACS Chem Biol. 7, 563–570 (2012).
Fleck, M. W., Palmer, A. M. & Barrionuevo, G. Potassium-induced long-term potentiation in rat hippocampal slices Brain Res 580, 100–105 (1992).
Feng, Z. & Durand, D. M. Effects of potassium concentration on firing patterns of low-calcium epileptiform activity in anesthetized rat hippocampus: Inducing of persistent spike activity. Epilepsia 47, 727–736 (2006).
Tabuchi, A. Synaptic plasticity-regulated gene expression: a key event in the long-lasting changes of neuronal function. Biol Pharm Bull 31, 327–335 (2007).
Fonseca, R., Nagerl, V. & Bonhoeffer, T. Neuronal activity determines the protein synthesis dependence of long-term potentiation. Nat. Neurosci. 9, 478–480 (2006).
Mattson, M. P. & Magnus, T. Ageing and neuronal vulnerability Nat. rev. Neurosci. 7, 278–294 (2006).
Kann, O. & Kovács, R. Mitochondria and neuronal activity. Am J. Physiol Cell Physiol. 292, C641–C657 (2007).
Evans, M. D., Dizdaroglu, M. & Cooke, M. S. Oxidative DNA damage and disease: Induction, repair and significance. Mutat. Res. 567, 1–61 (2004).
Tian, Y. et al. High molecular weight persimmon tannin ameliorates cognition deficits and attenuates oxidative damage in senescent mice induced by D-galactose. Food Chem Toxicol. 49, 1728–1736 (2011).
Visochek, L. et al. PolyADP-ribosylation is involved in neurotrophic activity. J. Neurosci. 25, 7420–7428 (2005).
Grigoryan, G. & Segal, M. Prenatal stress alters noradrenergic modulation of LTP in hippocampal slices. J Neurophysiol. 110, 279–85 (2013).
Ivenshitz, M. & Segal, M. Simultaneous NMDA-dependent long-term potentiation of EPSCs and long-term depression of IPSCs in cultured rat hippocampal neurons. J Neurosci. 26, 1199–1210 (2006).
Guo, S. et al. Neuroprotection via matrix-trophic coupling between cerebral endothelial cells and neurons Proc.Natl. Acad. Sci. USA 105, 7582–7587 (2008).
We thank Dr Liron Miller and Radka Holbova, Sheba Medical Center, for the maintenance of PARP1 KO mice and Efrat Biton in the lab of Prof Menahem Segal, for preparing neuronal cell cultures from PARP1 KO newborn mice. This work was supported by NIH grant 1R21DA027776 and the Israeli Ministry of Health grant (M. C.-A.).
The authors declare no competing financial interests.
About this article
Cite this article
Visochek, L., Grigoryan, G., Kalal, A. et al. A PARP1-ERK2 synergism is required for the induction of LTP. Sci Rep 6, 24950 (2016). https://doi.org/10.1038/srep24950
Presynaptic long-term potentiation requires extracellular signal-regulated kinases in the anterior cingulate cortex
Molecular Pain (2020)
Signal Transduction and Targeted Therapy (2019)
Neural Plasticity (2019)
Psoralidin Stimulates Expression of Immediate-Early Genes and Synapse Development in Primary Cortical Neurons
Neurochemical Research (2018)