Autonomously Active Protein Kinase C in the Maintenance Phase of N -Methyl- D -aspartate Receptor-independent Long Term Potentiation*

In area CA1 of the hippocampus, the induction of long term potentiation (LTP) requires activation of either N -methyl- D -aspartate receptors (NMDA receptor-dependent LTP) or voltage-gated Ca 2+ channels (NMDA receptor-independent LTP). We have investigated biochemical sequelae of NMDA receptor-independent LTP induction. We find that a persistent increase in second messenger-independent protein kinase C activity is associated with the maintenance phase of NMDA receptor-independent LTP. This increase in protein kinase C activity is prevented by blocking LTP with nifedipine, a Ca 2+ channel antagonist, or kynurenic acid, a nonselective glutamate receptor antagonist. Additionally, we find an increase in the catalytic fragment of protein kinase C (PKM) in the maintenance phase of NMDA receptor-independent LTP, indicating that proteolytic activation of protein kinase C may account for its autonomous activation. This increase in the catalytic fragment of protein kinase C is also prevented by blocking LTP induction. These results are the first to demonstrate that persistent protein kinase C activation is a possible mechanism for the maintenance of NMDA receptor-independent LTP. Long term potentiation (LTP) 1 in area CA1 of the hippocampus has been categorized into two general types based on the pharmacology of induction: NMDA receptor-dependent LTP (NMDA LTP) and NMDA receptor-independent LTP (non-NMDA LTP) (1–4). NMDA LTP is blocked by NMDA receptor antagonists (5), and a variety of evidence suggests that Ca 2+ influx through the NMDA receptor ionophore (6–9) is required for NMDA LTP induction (4, 5, 10–12). On the other hand, non-NMDA LTP is not blocked by NMDA receptor antagonists (1–3, 13–17). Furthermore, induction of non-NMDA LTP can be blocked in LTP using affinity-purified antibodies against PKC ζ (41). The antibody used in the present studies, however, does not appear to recognize PKC ζ in the hippocampus in a side by side comparison with the PKC ζ antibody used in the study of Sacktor et al. (41) (data not shown, PKC ζ antibody generously provided by Dr. Todd Sacktor). These findings suggest differences in the posttranslational modifications of PKC observed in NMDA LTP and LTP K .

biochemical studies have shown directly that a lasting increase in both protein kinase C (PKC) activity (27,28) and Ca 2+ /calmodulin-dependent protein kinase activity (29) is associated with the maintenance phase of NMDA LTP. The recent demonstration that LTP can be induced without NMDA receptor activation in area CA1 affords us the opportunity to compare the biochemical and physiological mechanisms of the maintenance of these two different forms of LTP in the same population of synapses. If non-NMDA LTP shares common expression mechanisms with NMDA LTP, then they should converge on common biochemical pathways. This idea gains support from studies of the interaction between NMDA LTP and non-NMDA LTP in which establishment of one form of LTP occludes to some extent the subsequent induction of the other (15,17).
In the present study we tested the hypothesis that a persistent increase in second messengerindependent or autonomous PKC activity is associated with the maintenance phase of non-NMDA LTP. We have demonstrated that there is an increase in autonomous PKC activity associated with non-NMDA LTP maintenance. This effect was observed using a potent and selective PKC substrate in assays of kinase activity in vitro, and expression of the increased activity was blocked by the relatively selective PKC inhibitor peptide PKC- (19)(20)(21)(22)(23)(24)(25)(26)(27)(28)(29)(30)(31)(32)(33)(34)(35)(36). We also observed on Western blots an increase in the catalytic fragment of PKC (PKM) to be associated with non-NMDA LTP maintenance, providing additional evidence for a persistent increase in autonomous PKC activity and implicating proteolytic activation as a possible mechanism.

EXPERIMENTAL PROCEDURES
Conventional hippocampal slices were prepared from 4-8-week-old, male, albino, Sprague-Dawley rats as described (30). Slices were perfused with physiological saline in an interface chamber at 1-2 ml/min at 33-34 °C. Physiological saline contained (in mM) NaCl (123), KCl (3.5), CaCl 2 (2), MgCl 2 (1.2), NaHCO 3 (25), dextrose (10), pH 7.4, when saturated with 95% O 2 and 5% CO 2 . Extracellular field recordings were made in the stratum radiatum of area CA1 using 1-10 megohm electrodes filled with 0.5 M NaCl. Slices were stimulated continuously at 0.05 Hz with a Teflon-insulated, bipolar, platinum electrode. Unless noted otherwise, 50 µM DL-2-amino-5-phosphonovaleric acid was present in all experiments for at least 1 h prior to base-line synaptic recording. LTP K was induced with a 10-min application of TEA (25 mM). After recording from the potentiated slice for the indicated time, a control slice, which was not exposed to TEA, was recorded from briefly (<2 min) to ensure viability. Both potentiated and control slices were taken from the same slice preparation. Slices were then frozen on glass at dry ice temperature, and the region of area CA1 between the stimulating and recording electrodes was dissected from the slices under 10-20× magnification. All further manipulations to control and potentiated slices were performed simultaneously under the same conditions so that the control and potentiated slices could be directly compared.
It is important to note that control and experimental slices for one experiment were incubated simultaneously in physiological saline for the same period of time. Also, control and experimental slices from the same slice preparation were frozen, homogenized, and assayed simultaneously so that a direct comparison could be made between control and potentiated slices within the same experiment.
Western blot procedures were performed essentially as described (28,32). Proteins from control and LTP K , homogenates (5-10 µg) from the same slice preparation were separated on 10% polyacrylamide gels. Proteins were transferred onto Immobilon-P transfer membranes. Blots were incubated with a polyclonal PKC antibody (generously provided by Freesia Huang, National Institutes of Health) diluted 1:2000, exposed to a secondary antibody diluted 1:2000, exposed to 20 µCi of 125 I-labeled protein A (1 Ci = 37 GBq), and then placed on film at −70 °C. This antibody is known to recognize native PKC and the catalytic fragment of PKC (PKM) in both tissue homogenates (not shown) (33)(34)(35) and purified preparations (28,31,33). Moreover, brief incubation of hippocampal homogenates with Ca 2+ results in the generation of a proteolytic fragment of PKC that elutes from ionexchange and gel filtration columns as expected for PKM and that is recognized by this antibody (35).

Autonomous PKC Activity in the Maintenance Phase of LTP K
LTP K is the most robust form of non-NMDA LTP that can be induced in a relatively homogeneous, large population of synapses in the hippocampus. We have taken advantage of this robust form of non-NMDA LTP to test the hypothesis that a persistent increase in PKC activity is associated with the maintenance phase of non-NMDA LTP. LTP K was induced in the presence of 50 µM DL-2-amino-5-phosphonovaleric acid, and pEPSPs were monitored for 45 min following TEA washout (Fig. 1, A and B). For the control we recorded briefly (<2 min) from another slice from the same preparation that was not exposed to TEA. Isolated CA1 regions from the control and potentiated slices were then assayed for PKC activity using the selective PKC substrate peptide NG-(28-43) (31,36). Autonomous, Ca 2+independent PKC activity was significantly increased in potentiated CA1 regions 45 min following TEA washout (Fig. 1C, solid bars; control, 0.59 ± 0.12 pmol/min/µg protein; LTP K , 0.98 ± 0.13 pmol/min/µg; n = 14; p = 0.038, unpaired Student's t test).
We confirmed that the increase in NG-(28-43) phosphorylation associated with LTP K is due to the activity of PKC by including a relatively selective PKC inhibitor peptide PKC-(19-36) (19,27,28) in assays in vitro. The increase in NG-(28-43) phosphorylation was blocked by 5 µM PKC-(19-36) (Fig. 1C, hatched bars). Therefore, as in NMDA LTP, a persistent increase in autonomous PKC activity is associated with the maintenance phase of LTP K .

Blocking Synaptic Potentiation Prevents the Increase in Autonomous PKC Activity
To control for possible effects of low frequency test stimulation or the solution change protocol on PKC activity, slices were stimulated at the test frequency (0.2 Hz, low frequency stimulation) for the duration of a typical LTP experiment, and a "sham" solution change protocol was performed with normal recording saline. Low frequency stimulation without TEA application caused no synaptic potentiation and no significant change in PKC activity (control, 0.656 ± 0.09 pmol/min/µg; low frequency stimulation, 0.688 ± 0.12 pmol/min/µg; n = 5; not shown).
To strengthen the association between LTP K and the increased PKC activity and to rule out a direct effect of TEA application on PKC activity, we blocked LTP K with the nonselective glutamate receptor antagonist kynurenic acid and assayed PKC activity. Kynurenic acid (10 mM) blocked the induction of LTP K (Fig. 2A). This result is consistent with a previous report that LTP K is blocked by the glutamate receptor antagonist CNQX (3). Kynurenic acid was used because its effects washed out within the time course of our experiments (Fig. 2B). Blocking LTP K prevented the increase in PKC activity ( Fig. 2C; control, 0.55 ± 0.07 pmol/ min/µg; TEA and kynurenic acid, 0.73 ± 0.21; n = 6), suggesting that the requirements for LTP K induction and for the persistent increase in PKC activity are similar.
As an additional control, we blocked LTP K with the voltage-gated calcium channel antagonist nifedipine. In a series of five experiments, nifedipine (10 µM) significantly reduced LTP K (Fig. 3A; pEPSP slope, 108 ± 1% of control, n = 5), and this also prevented the increase in PKC activity ( Fig. 3B; control, 0.369 ± 0.03 pmol/min/µg; TEA and nifedipine, 0.289 ± 0.06 pmol/min/µg; n = 5). The effect of nifedipine on the magnitude of LTP K , however, proved to be variable in subsequent experiments in which little or no effect of nifedipine on LTP K was observed (n = 4, not shown). The variable effects of nifedipine on LTP K indicate that another source of Ca 2+ entry may also play a role in LTP K . Nonetheless, the associated increase in PKC activity was not observed when LTP K induction was blocked by either kynurenic acid or nifedipine. These findings support the hypothesis that the increase in PKC activity plays a role in LTP K maintenance.

Time Course of the Increase in Autonomous PKC Activity
If the increase in PKC activity is involved in the maintenance of LTP K then the increase should persist during LTP K maintenance. To test this prediction we monitored LTP K for 3 h following TEA washout (Fig. 4A) and assayed PKC activity. A significant increase in basal PKC activity was observed 3 h into LTP K maintenance ( Fig. 4B; control, 0.243 ± 0.07 pmol/ min/µg; LTP K , 0.686 ± 0.17 pmol/min/µg; n = 7; p < 0.05). These results further indicate that the increase in PKC activity is not due to an effect of residual TEA because TEA should be completely washed out after 3 h (see Fig. 2B). Control PKC activity was lower in the 3-h LTP experiments than in the 45-min LTP experiments (Fig. 4B). The decrease in autonomous PKC activity with longer incubation times has been observed previously 2 and may be due to gradual deterioration of slices.
The onset of the increase in PKC activity was also examined. No significant increase in basal PKC activity was observed 5 or 20 min following the washout of TEA (Fig. 4B). Therefore, the increase in PKC activity is gradual or delayed in onset relative to the presumed induction phase of LTP K .
No significant change in total PKC activity stimulated with Ca 2+ , phosphatidylserine and 1,2-dioctanoyl-sn-glycerol was observed at any time point tested (5 min, 20 min, 45 min, 3 h; data not shown).

Posttranslational Modification of PKC Is Associated with LTP K Maintenance
To provide an additional line of evidence for an increase in autonomous PKC activity that does not rely on substrate selectivity or inhibitor selectivity, we tested for posttranslational modification of PKC in the maintenance phase of LTP K . The best characterized mechanism for generating an increase in autonomous PKC activity is proteolytic activation (37)(38)(39)(40). Proteolytic activation of PKC occurs when the regulatory domain is removed, leaving the 45-50-kDa catalytic domain (PKM), which is active in the absence of second messenger activators (37)(38)(39)(40). To test the hypothesis that the persistent increase in PKC activity associated with LTP K is due to proteolytic activation of PKC, a Western blot analysis was performed using a polyclonal antibody that recognizes multiple PKC isoforms as well as proteolytically activated PKC (33)(34)(35) and preferentially binds to unphosphorylated PKC (28). A significant increase in a 45-kDa immunoreactive PKC fragment was observed ( Fig.  5A; LTP K , 119 ± 6% of control; n = 10; p = 0.009, unpaired Student's t test). Previous characterization of this fragment indicates that its immunoreactivity, molecular weight, and charge character are consistent with that of proteolytically activated PKC (PKM) (31,(33)(34)(35)(37)(38)(39)(40). To demonstrate that the increase in PKM is specifically associated with LTP K , we blocked LTP K with kynurenic acid (see Fig. 2A). Blocking LTP K completely eliminated the increase in PKM (Fig. 6). Thus, we have provided additional evidence for an increase in autonomous PKC activity associated with LTP K , and this increase in PKC activity may involve proteolytic activation of PKC.
In NMDA LTP, Klann et al. (28) observed a decrease in native PKC immunoreactivity that was reversed by treatment of homogenates with phosphatases, indicating an increase in phosphorylation of PKC. Using the same antibody and blotting conditions, we saw no evidence for a change in phosphorylation of PKC as no change in native PKC immunoreactivity was observed in LTP K (Fig. 5B; LTP K , 125 ± 15% of control, n = 10). This result complements our finding of no change in total PKC activity in LTP K . Also, in contrast to Klann et al. (28), we observed an increase in the catalytic fragment of PKC. An increase in a catalytic fragment of PKC isoform ζ has been previously described in NMDA LTP using affinity-purified antibodies against PKC ζ (41). The antibody used in the present studies, however, does not appear to recognize PKC ζ in the hippocampus in a side by side comparison with the PKC ζ antibody used in the study of Sacktor et al. (41) (data not shown, PKC ζ antibody generously provided by Dr. Todd Sacktor). These findings suggest differences in the posttranslational modifications of PKC observed in NMDA LTP and LTP K .

DISCUSSION
We have demonstrated that a persistent increase in protein kinase activity is associated with the early maintenance phase of LTP K . The evidence that this increase in kinase activity is due to protein kinase C is compelling.
Finally, we observed an increase in an approximately 45-kDa fragment of PKC associated with the maintenance phase of LTP K . A variety of evidence suggests that this 45-kDa band is the constitutively active, catalytic fragment of PKC (PKM). A persistent increase in autonomous PKC activity is associated with the increase in the 45-kDa fragment following LTP K , as expected for an increase in PKM. The PKC antibody we used has been well characterized and is known to recognize native PKC and the catalytic fragment of PKC (PKM) in both tissue homogenates, including hippocampus (33-35) (not shown) and purified preparations (28,31,33). Also, brief incubation of hippocampal homogenates with Ca 2+ results in the generation of a 45-50-kDa proteolytic fragment of PKC recognized by the same antibody (35) (not shown) and a large increase in basal PKC activity (35) (not shown). Both the immunoreactivity and the enzyme activity elute from ion-exchange chromatography as expected for PKM (35). These data strongly suggest that the 45-kDa PKC fragment is the constitutively active, catalytic fragment of PKC, PKM. An increase in a PKC fragment of the appropriate molecular weight for proteolytically activated PKC (PKM) alone provides evidence for an increase in basal PKC activity.
These results provide strong evidence that PKC is persistently activated in the maintenance phase of LTP K . Also, the present studies provide the first direct evidence suggesting persistent PKC activation as a possible mechanism for the maintenance of a form of NMDA receptor-independent LTP.
The increase in basal PKC activity, associated with non-NMDA LTP, like PKM, is independent of typical second messenger activators of PKC. The increased PKC activity is Ca 2+ -independent because basal PKC activity is measured in the presence of an excess of the Ca 2+ chelator EGTA. Also, it is unlikely that a persistent increase in a second messenger is responsible for the increase in basal PKC activity because of the manner in which our samples are prepared. A fraction of the CA1 region (20-30 µg of protein; volume, ≤1 µl) from LTP or control slices is homogenized in 100 µl of buffer and often diluted further. Only 5 µl of this homogenate is added to a 50-µl kinase assay reaction. Therefore, our sample is diluted at least 1:1000 before each kinase assay reaction is performed. Increased kinase activity under such dilute conditions suggests an intrinsic modification of the enzyme such as proteolytic activation. Thus, the increase in PKC activity is due to an autonomously active form of the enzyme, and the most parsimonious explanation is that the increased PKC activity is a result of the increase in PKM.
Considering this finding with previous results, NMDA LTP and LTP K have in common persistent activation of PKC in their maintenance phase. The increase in basal PKC activity is of similar magnitude in both forms of LTP (Fig. 1B) (27,28) and is prevented by blocking LTP (Figs. 2 and 4) (27,28). If we assume that the effector substrates of PKC are similar in the two forms of LTP, then these results complement the finding that NMDA LTP can partially occlude subsequent LTP K expression and vice versa (17). It is interesting that two forms of LTP with distinct induction mechanisms may converge on a common effector mechanism, persistent activation of PKC, in the maintenance phase.
Previous studies, using the same antibody and protocols used in the present study, indicate that the mechanism of PKC activation in NMDA LTP is increased phosphorylation of PKC (28) and not proteolytic activation because no change in the PKM fragment was observed on Western blots in NMDA LTP (28) (but see Sacktor et al. (41)). In the present studies, no increase in PKC phosphorylation was observed while an increase in PKM was observed. Thus, the two distinct means of Ca 2+ entry in NMDA LTP and LTP K appear to cause different posttranslational modifications of PKC. Differences in subcellular localization of Ca 2+ influx or in intracellular Ca 2+ concentrations between the two forms of LTP are possible explanations for this observation. Also, the activation of two distinct signaling pathways by Ca 2+ influx through NMDA receptors or through voltage-gated Ca 2+ channels in hippocampal neurons has been recently described (42).
An increase in a proteolytic fragment of PKC has been previously observed in NMDA LTP using an antibody specific for the PKC ζ isoform (41). The antibody used in the present study and the study of Klann et al. (28) does not appear to recognize this PKC ζ fragment or native PKC ζ in direct comparisons using Western blots of hippocampal homogenates with both antibodies (data not shown). Thus, the increase in PKM that we observe in LTP K is different from that observed in NMDA LTP.
In conclusion, we have found that, as in NMDA LTP, a persistent increase in basal PKC activity is associated with LTP K , a form of NMDA receptor-independent LTP. The mechanism of persistent PKC activation in LTP K may involve generation of the catalytically active fragment of PKC, PKM. Although differences in posttranslational modification of PKC occur in the two forms of LTP, these results indicate that persistent activation of PKC may play a role in the expression of both NMDA receptor-dependent and NMDA receptorindependent forms of LTP, suggesting a general role for PKC activity in LTP. It will be interesting to examine whether PKC activity is strictly required for the expression of LTP K and other forms of NMDA receptor-independent LTP.  12%, n = 14). B, example pEPSPs taken before (a), during (b), and 45 min after (c) TEA washout. C, increase in basal, Ca 2+ -independent PKC activity 45 min into LTP K (solid bars). Addition of PKC-(19-36) (5 µM) to kinase assays in vitro blocks the increase in PKC activity associated with LIT K (hatched bars).

Fig. 2. Blocking LTP K with kynurenic acid prevents the increase in PKC activity
A, LTP K was completely blocked when TEA was applied in the presence of 10 mM kynurenic acid (n = 5). B, physiological effects of kynurenic acid (10 mM) are completely reversible (n = 4). C, no significant increase in PKC activity was observed in slices exposed to TEA in the presence of kynurenic acid. Control slices were similarly perfused with 10 mM kynurenic acid.

Fig. 3. Blocking LTP K with 10 µM nifedipine prevents the increase in basal PKC activity
A, time course of pEPSP slope before and after TEA application (25 mM) in the presence of 10 µM nifedipine and 50 µM APV. Potentiation is significantly reduced to 108 ± 1% of base line 45 min after TEA washout (n = 5). B, no change in basal PKC activity occurs after TEA application in the presence of 10 µM nifedipine and 50 µM APV (n = 5). Control slices were similarly perfused with nifedipine.

Fig. 4. Time course of the increase in basal PKC activity associated with LTP K maintenance
A, 3-h time course of LTP K . pEPSP slope is plotted as a function of time. TEA (25 mM) application In the presence of 50 µM APV caused a lasting potentiation (144 ± 6%, n = 7). Inset numbers represent times in minutes where PKC activity was measured in B. B, the increase in PKC activity associated with LTP K persists for at least 3 h following TEA application. Basal PKC activity was measured in separate experiments at various time intervals following onset of TEA washout (5 min, n = 5; 20 min, n = 7; 45 min, n = 14; 180 min, n = 7). A significant increase in basal PKC activity was observed 45 and 180 min following TEA washout (* indicates statistical significance using an unpaired Student's t Left, example Western blot analysis using a polyclonal antibody raised against classical PKC isoforms demonstrates no significant increase in PKM when LTP K is blocked by kynurenic acid. Right, grouped densitometry data from kynurenic acid block experiments show no increase in PKM when LTP K is blocked by kynurenic acid (89 ± 8% of control, n = 6).