Abstract
Key points
Signalling mechanisms for coincidence detection of paired stimuli during classical conditioning are fundamental for understanding the mechanisms of associative learning.
Bidirectional 3-phosphoinositide-dependent kinase-1 (PDK1) activity is signalled by TrkB neurotrophin receptors for paired stimuli and p75NTR for unpaired stimuli.
Adenosine 2A receptors modulate PDK1 responses directly as G proteins and by transactivation of TrkB.
Convergence of protein kinase A and PDK1 activity initiates signalling of paired stimuli during classical conditioning.
Abstract
How the neural substrates for detection of paired stimuli are distinct from unpaired stimuli is poorly understood and a fundamental question for understanding the signalling mechanisms for coincidence detection during associative learning. To address this question, we used a neural correlate of eyeblink classical conditioning in an isolated brainstem from the turtle, in which the cranial nerves are directly stimulated in place of using a tone or airpuff. A bidirectional response is activated in <5 min of training, in which phosphorylated 3-phosphoinositide-dependent kinase-1 (p-PDK1) is increased in response to paired and decreased in response to unpaired nerve stimulation and is mediated by the opposing actions of neurotrophin receptors TrkB and p75NTR. Surprisingly, blockade of adenosine 2A (A2A) receptors inhibits both of these responses. Pairing also induces substantially increased surface expression of TrkB that is inhibited by Src family tyrosine kinase and A2A receptor antagonists. Finally, the acquisition of conditioning is blocked by a PDK1 inhibitor. The unique action of A2A receptors to function directly as G proteins and in receptor transactivation to control distinct TrkB and p75NTR signalling pathways allows for convergent activation of PDK1 and protein kinase A during paired stimulation to initiate classical conditioning.
Introduction
During classical conditioning, temporally paired events lead to cellular changes underlying the formation of associative learning. When and how pairing of a neutral conditioned stimulus (CS) with a behaviour-evoking unconditioned stimulus (US) is first detected during learning is poorly understood. Current theories for coincidence detection are based largely on the NMDA receptor (NMDAR; Tsien, 2000), which is widely considered to be a molecular coincidence detector. However, studies on operant conditioning in Aplysia (Lorenzetti et al. 2008) and olfactory learning in Drosophila (Gervasi et al. 2010) indicate that coincidence detection may occur without any NMDAR-mediated signalling. These mechanisms instead use the adenylate cyclase (AC) complex, which serves as a point of molecular convergence for paired inputs that raises cAMP/protein kinase A (PKA) levels past a threshold required for downstream signalling.
The 3-phosphoinositide-dependent kinase-1 (PDK1) pathway is also likely to play an important role in coincidence detection by priming PKA for downstream substrate phosphorylation (Newton, 2003; Pearce et al. 2010). 3-Phosphoinositide-dependent kinase-1 localizes to the plasma membrane, where its activity is regulated by phosphoinositide 3-kinase (PI3K). 3-Phosphoinositide-dependent kinase-1 is constitutively active (Casamayor et al. 1999) but can be facilitated by PI3K and may provide an important mechanism for prolonged activation of multiple substrate targets required for learning (Mullasseril et al. 2007). While the PDK1 signalling cascade is not well characterized, it has been shown to be involved in induction of CA1 hippocampal long-term potentiation (LTP; Opazo et al. 2003) and long-term depression (LTD; Hou & Klann, 2004) and is activated during fear conditioning (Sui et al. 2008). Growth factor receptors, such as the tropomyosin-related kinases (Trks), stimulate PI3K activity (Reichardt, 2006; Duman & Voleti, 2012). One major neurotrophin, brain-derived neurotrophic factor (BDNF), is known to have a critical role in plasticity and learning by binding to its TrkB receptor to trigger positive cellular functions such as growth and survival (Lu et al. 2005; Gruart et al. 2007; Li & Keifer, 2008, 2009; Minichiello, 2009). Another class of receptor, the p75 neurotrophin receptor (p75NTR), is particularly sensitive to the proneurotrophin precursor proteins and generally results in negative cellular effects such as cell death. Evidence indicates that LTP is regulated by BDNF and TrkB, while LTD is signalled by its precursor proBDNF and p75NTR, suggesting a reciprocal functional relationship between the two neurotrophin receptors.
The classically conditioned eyeblink response is a well-studied model for investigating mechanisms underlying associative learning. In most vertebrates, an eyeblink response to a tone can be learned when the tone is repeatedly paired with an airpuff to the cornea that produces a blink. A neural analogue of eyeblink conditioning can be generated from an isolated brainstem preparation from the turtle to analyse the regulatory mechanisms critical for acquisition of learned responses (Keifer & Houk, 2011; Zheng et al. 2012). In place of using a tone or airpuff, paired stimulation of the auditory nerve (the ‘tone’ CS) with the trigeminal nerve (the ‘airpuff’ US) results in the acquisition of abducens nerve discharge that represents a neural correlate of blink conditioned responses (CRs).
Previously, we showed that PKA activation is critical for initiation of conditioning (Zheng & Keifer, 2009). Here, the signal transduction processes involved in detection of the earliest stages of paired nerve stimulation were explored. The results show that while unpaired stimuli reduce the level of p-PDK1, paired stimuli significantly increase it, resulting in downstream induction of PKA phosphorylation to initiate conditioning. This bidirectional regulation of PDK1 is signalled by TrkB and p75NTR complexed with adenosine 2A (A2A) receptors. The unique function of A2A receptors in signalling directly as G proteins and through transactivation of target receptors provides a basis for control of distinct signalling pathways that converge on activation of PDK1 and PKA to initiate classical conditioning.
Methods
Training procedures
Freshwater pond turtles, Trachemys scripta elegans, purchased from commercial suppliers were anaesthetized by hypothermia until torpid and decapitated. All experiments involving the use of animals were performed in accordance with the guidelines of the National Institutes of Health and were approved by the University of South Dakota Institutional Animal Care and Use Committee. The brainstem was transected at the levels of trochlear and glossopharyngeal nerves and the cerebellum was removed as described previously (Zheng et al. 2012). The preparation did not contain the red nucleus and consisted only of the pons, and was continuously bathed (2–4 ml min−1) with physiological saline containing (mm): 100 NaCl, 6 KCl, 40 NaHCO3, 2.6 CaCl2, 1.6 MgCl2 and 20 glucose, which was oxygenated with 95% O2–5% CO2 and maintained at room temperature (22–24°C) at pH 7.6. Suction electrodes were used for stimulation and recording of cranial nerves (see Fig.2A).
Figure 2.
Protein kinase A is phosphorylated within minutes of pairing by PDK1
A, experimental set-up. In vitro preparation of the pons is classically conditioned by paired stimulation of the auditory nerve [the ‘tone’ conditioned stimulus (CS)] with the trigeminal nerve [the ‘airpuff’ unconditioned stimulus (US)] while recording discharge in the ipsilateral abducens nerve that is representative of a neural correlate of a blink response. Representative physiological traces of an abducens nerve discharge before and after conditioning (CR) are shown. Pairing consists of a 100 Hz, 1 s duration CS to the auditory nerve that precedes a single shock US to the trigeminal nerve. One session of conditioning is comprised of 50 paired stimuli separated by 30 s and lasts a total duration of 25 min. B, Western blots show that protein kinase A (PKA) is phosphorylated during conditioning after 15 min of pairing (P) but not earlier. Unpaired stimuli, either US (U) or CS (C) alone, fail to result in PKA phosphorylation at any of the time points examined [P < 0.0001 compared with naive (N), n = 4 per group, one-way ANOVA followed by Fisher's and Bonferroni's post hoc tests]. Values in this and all subsequent figures represent means ± SEM. C, levels of p-PDK1 (P = 0.01) and p-PKA (P = 0.002) are significantly increased above naive after one session of conditioning (C1). Bath application of the PDK1 inhibitor BX during pairing reduced p-PDK1 to naive levels (P = 0.82) and inhibited conditioning-induced phosphorylation of PKA relative to naive (P = 0.62, n = 3 per group). *Represents significant differences from naïve throughout.
The US was a twofold threshold single shock applied to the trigeminal nerve and the CS was a 1 s, 100 Hz, train stimulus applied to the ipsilateral auditory nerve that was below the threshold amplitude required to produce activity in the abducens nerve. Neural responses were recorded from the ipsilateral abducens nerve that innervates the extraocular muscles controlling movements of the eye, nictitating membrane and eyelid. The abducens nerve projects to the retractor bulbi and lateral rectus muscles that retract and abduct the eye and serves to raise the nictitating membrane passively, while the pyramidalis muscle actively moves the nictititating membrane and eyelid (Keifer, 1993). As turtles lack an orbicularis muscle, the facial nerve is unlikely to have a role in blinking (Zhu & Keifer, 2005).
For paired stimuli, the CS–US interval was 20 ms, which was defined as the time between the CS offset and the onset of the US, and the intertrial interval between paired stimuli was 30 s. A pairing session was composed of 50 CS–US presentations lasting 25 min for one complete session (C1). For multiple pairing sessions, there was a 30 min rest period between pairing sessions, during which there was no stimulation. Conditioned responses were defined as abducens nerve activity that occurred during the CS and exceeded an amplitude of twofold above the baseline recording level. Unpaired stimulation consisted of application of either the CS or US alone using the same procedures described above. Naive preparations were not stimulated but remained in the bath for an equivalent time period to the experimental preparations. Backward pairing consisted of US–CS pairing, in which the onset of the US was simultaneous with the onset of the CS (Keifer et al. 1995).
Pharmacological agents
The following compounds were bath applied to brainstem preparations prior to and during nerve stimulation or to unstimulated preparations for the same time period: the competitive PDK1 antagonist BX-912 or BX-795 (0.3 μm for a total of 2 h; Axon Medchem, Reston, VA, USA; no differences were observed between the two compounds); the p75NTR signalling inhibitor TAT-Pep5 (1 μm for 1.5 h; Calbiochem, Billerica, MA, USA); anti-rat p75NTR extracellular domain (REX) antibody (a kind gift of Louis Reichardt at University of California, San Francisco, USA; 50 μg ml−1 for 2 h); antibody to TrkB (5 μg ml−1 incubated overnight; 20542; Santa Cruz Biotechnology, Dallas, TX, USA; Li & Keifer, 2008); the selective adenosine A2A receptor antagonist ZM 241385 (50 nm for 1 h; Tocris, Minneapolis, MN, USA); the A2A receptor agonist CGS 21680 (20 nm for 1 h; Tocris); adenosine (0.2 mm for 15 min; Tocris); cleavage-resistant human proBDNF (100 ng ml−1 for 45 min; Alomone Labs, Jerusalem, Israel); the selective TrkB agonist 7,8-dihydroxyflavone (DHF; 10 μm for 45 min; Tocris); the cell-permeable PI3K inhibitor wortmannin (200 nm for 1 h; Sigma, St. Louis, MO, USA); the inhibitor of sphingomyelinase sphingolactone-24 (25 μm for 1 h; Santa Cruz Biotechnology; Luther et al. 2013); the phospholipase C (PLC) inhibitor U73122 (4 μm for 1 h; Tocris; Bleasdale et al. 1990); and the Src family tyrosine kinase inhibitor PP2 or its negative control PP3 (0.5 μm for 1 h; Tocris; Bain et al. 2003).
Enzyme-linked immunosorbent assay
Levels of IP3 were determined with a commercial ELISA kit (catalogue no. MBS701131; MyBiosource, San Diego, CA, USA), which has high sensitivity and specificity for detection of rat IP3. The limit of sensitivity was fixed at 2.5 pg ml−1. The measurements were performed according to the manufacturer's instructions with modification. After a period of equilibration, brainstems received unpaired or paired nerve stimulation for 5 or 25 min and then were homogenized in cold PBS [1 mg (5 μl PBS)−1] and stored overnight at −80°C. After three freeze–thaw cycles were performed to break the cell membranes, homogenates were centrifuged for 20 min at 9000 g at 4°C. Standards or samples were added to plates that were precoated with antibody provided by the kit. There were no differences in brainstem IP3 levels between the 5 and 25 min stimulation groups and so these data were combined and are expressed in picograms per millilitre.
Immunoprecipitation and Western blotting
Immediately after the physiological experiments were performed, brainstem samples for protein analysis were obtained by dissecting a portion of tissue from the pons containing the abducens nuclei only on the stimulated side, frozen in liquid nitrogen and stored at −80°C. Brainstems were homogenized in lysis buffer with a protease and phosphatase inhibitor cocktail. Protein samples were precleared with protein A/G–agarose, and supernatants were incubated with the primary antibodies or with non-specific rabbit or mouse IgG as a control at 4°C for 2 h. Protein A/G–agarose was added to the protein samples and incubated at 4°C overnight. Immunoprecipitated samples or IgG control samples were washed with ice-cold lysis buffer and dissociated by heating for 5 min in the loading buffer and then subjected to SDS-PAGE. For all Western blots and co-immunoprecipitation experiments, both input protein and IgG controls were loaded at the same time. The following primary antibodies were used for co-immunoprecipitation and/or Western blotting: t-PDK1 (3062; Cell Signaling, Danvers, MA, USA); p-PDK1 (S241; 3061; Cell Signaling); t-PKA (06-903; Millipore, Danvers, MA, USA); p-PKA (T197; 4781; Cell Signaling); TrkB (20542; Santa Cruz Biotechnology); p75NTR (G323A; Promega, Madison, WI, USA); A2A (13937; Santa Cruz Biotechnology); and actin for loading controls (1501R; Millipore). The specificity of the p-PDK1 antibody in turtle brain compared with rodent is shown in Fig.1. The specificity of the PKA antibody was shown by Zheng & Keifer (2009). Proteins were detected by the ECL Plus chemiluminescence system (Amersham, Pittsburgh, PA, USA) or the Odyssey infrared imaging system (Li-Cor Biosciences, Lincoln, NE, USA). All Western blots were run with three naive lanes for comparison with the experimental lanes and quantified by computer-assisted densitometry relative to naive.
Figure 1.
Specificity of the phosphorylated 3-phosphoinositide-dependent kinase-1 (p-PDK1) antibody in turtle brain
Immunoblots showing bands labelled with the p-PDK1 antibody (which recognizes the autophosphorylation site at Ser241 on its activation loop) from turtle brain lysates and brain lysates from mouse and rat. Two bands are prominent for each species; the turtle bands are shifted slightly higher relative to the mouse and rat. 3-Phosphoinositide-dependent kinase-1 has several different isoforms identified in humans. Isoform 1 is 556 amino acids and has a molecular weight of 63 kDa (accession NM_002613), whereas isoforms 2 and 3 are 506 (58 kDa) and 530 amino acids (60 kDa), respectively. It is likely that these isoforms are visible as the upper and lower bands in the blots for the turtle, rat and mouse. Turtle PDK1 protein (Chelonia mydas; GenBank accession EMP32206) has been identified at 554 amino acids, with a moleular wieght of 63 kDa and a putative sequence identified. This sequence shares 89% sequence identity with chickens, 80% with mice and 79% with human. Additionally, three variants in genomic DNA are predicted for PDK1 in the turtle Chrysemys picta bellii (accession XM_005288887.1). The figures for turtle PDK1 here show the most prominent lower band. The upper band undergoes similar conditioning-related changes to those illustrated for the lower band.
Biotinylation assays
Brainstem preparations were incubated in physiological saline containing 1 mg ml−1 EZ-link sulfo-NHS-LC Biotin (Pierce, Rockford, IL, USA) during the conditioning procedure or drug treatment for the required time period. The total elapsed time of incubation in the biotin did not exceed 2 h. After the experiment, brainstems were washed in ice-cold physiological saline and frozen in liquid nitrogen. Tissue samples were homogenized in lysis buffer (20 mm Tris pH 8.0, 1 mm EDTA, 1% NP-40, 0.15 m NaCl, 10 mm Na4P2O7 and 5% glycine) with a protease (Roche, Madison, WI, USA) and phosphatase inhibitor cocktail (Sigma), rotated at 4°C for 2 h, centrifuged at 14,000 g for 20 min at 4°C, and the supernatants were stored at −80°C. Biotinylated proteins (200–300 μg) were incubated with UltraLink Immobilized Streptavidin (Pierce) at 4°C overnight. Streptavidin–protein complexes were washed with ice-cold washing buffer and pellets resuspended in 2× SDS/β-mercaptoethanol and boiled for 5 min before separation by 8% SDS–PAGE followed by Western blotting. To confirm that the assays only pulled down surface protein, blots were incubated with β-actin, and only the whole-cell lysis input lane showed a band, while the lanes containing biotinylated cell membranes were blank.
Statistical analysis
All data were analysed using a one-way ANOVA followed by post hoc analysis using Fisher's and Bonferroni's tests and are presented as means ± SEM. The P and n values are given in the text, where n values represent the number of brainstem preparations.
Results
Timing of PKA phosphorylation in conditioning and requirement for PDK1 activation
The experimental set-up for studies on in vitro classical conditioning is illustrated in Fig. 2A. Suction electrodes were used for stimulation of the trigeminal and auditory nerves and to record neural activity characteristic of eyeblink responses from the ipsilateral abducens nerve using an isolated brainstem preparation (Keifer, 1993; Keifer et al. 1995). Paired stimuli consisted of a 1 s, 100 Hz CS to the auditory nerve followed almost immediately by a single shock trigeminal nerve US. Further details are given in the Methods. Our previous studies demonstrated that phosphorylation of PKA has a critical role in initiating the signalling cascades required for driving AMPARs into synapses during conditioning (Zheng & Keifer, 2009). Protein kinase A undergoes dramatically increased phosphorylation 15 min after the onset of paired stimuli (P) but not at earlier time points compared with unstimulated naive (N) preparations (Fig.2B; P < 0.0001 vs. N). Unpaired stimuli, such as the US (U) or CS (C) alone or randomly presented pseudoconditioning trials, produced no change in p-PKA levels. In order to examine further the mechanisms responsible for PKA activation shortly after pairing begins, the PDK1 inhibitors BX-795 or BX-912 (BX), which block the S241 autophosphorylation site, were used. Following conditioning for one pairing session (C1, which lasted 25 min in duration), p-PDK1 was markedly increased above naive levels (Fig.2C; P = 0.01 vs. N), as was p-PKA (P = 0.002). Bath application of BX (0.3 μm) followed by conditioning resulted in a reduction in p-PDK1 to levels similar to naive values (P = 0.82 vs. N). This treatment also resulted in significantly reduced p-PKA compared with normal conditioning (Fig.2C; P = 0.62 vs. N). These data suggest that conditioning-dependent activation of PKA requires phosphorylation by PDK1 to prime its activation loop residue.
Bidirectional regulation of PDK1 during paired or unpaired stimulation is mediated by TrkB and p75NTR
How detection of paired stimuli is distinct from unpaired stimuli to result in the formation of learned associations is a fundamental question for understanding the mechanisms underlying learning. The signal transduction events evoked by paired and unpaired stimuli during in vitro conditioning diverge in less than 5 min of stimulation. As shown in Fig 3A, the level of p-PDK1 was significantly reduced following unpaired stimulation of either the US or the CS alone (both, P < 0.0001 vs. N), while it was increased in response to paired stimulation (P = 0.003 vs. N) compared with naive preparations. Closer scrutiny of the timing of these responses shows that after only 5 min of pairing (or application of 10 paired stimuli) p-PDK1 levels were significantly increased (Fig.3B; P = 0.01, paired groups vs. N), while they were reduced after unpaired stimulation (P < 0.0002, unpaired groups vs. N). This surprising bidirectional response in p-PDK1 was maintained for at least one session, or after 25 min, of training. In fact, this response was first detectable after 3 min of stimulation, or after only six paired or unpaired stimuli, but not after 1 min or two stimulus presentations. Significantly, backward pairing (BP), in which the stimuli remain paired but are reversed so that the single shock US precedes the CS, failed to result in any changes in p-PDK1 (Fig.3C; P = 0.70). Backward pairing is a more difficult learning protocol and results in no conditioned responding or inconsistent effects in behaving animals and in this preparation (Keifer et al. 1995) because a CS presented after the US has little predictive value for signalling the occurrence of the US.
Figure 3.
Bidirectional responses of p-PDK1 to paired and unpaired stimuli are mediated by p75NTR and TrkB neurotrophin receptors
A, unpaired stimulation of the US (U) or CS (C) results in significantly decreased levels of p-PDK1 (U and C, P < 0.0001), while paired stimulation results in increased levels (P = 0.003; n = 16 per group) compared with unstimulated naive preparations. B, bidirectional responses of p-PDK1 to unpaired and paired stimuli occur within 5 min after stimulation onset and are maintained throughout 25 min (P < 0.0002 for unpaired groups vs. N, P = 0.01 for paired groups vs. N, n = 4 per group). C, backward pairing (BP), in which the US precedes the CS, results in no changes in p-PDK1 (P = 0.70; n = 3 per group). D, bath application of the peptide inhibitor of p75NTR, TAT-Pep5, inhibits or reverses p-PDK1 levels to unpaired stimulation for 5 and 25 min compared with naive (P = 0.14, unpaired groups vs. N, n = 4 per group). Responses to paired stimulation are not altered by the treatment (P = 0.001, paired groups vs. N, n = 4 per group). E, bath application of the REX antibody, which suppresses p75NTR function, reverses the decreased response of p-PDK1 to unpaired stimulation, while the increased response to pairing is unaffected (P = 0.003, unpaired groups vs. N; P = 0.008, paired group vs. N; n = 3 per group). F, application of a TrkB antibody that inhibits TrkB function results in suppression of the elevated responses of p-PDK1 to paired stimulation but does not affect decreased responses to unpaired stimulation (P = 0.004, unpaired groups vs. N; P = 0.93, paired groups vs. N; n = 4 per group). *Represents significant differences from naïve throughout.
The bidirectional responses observed for p-PDK1 are reminiscent of LTD- and LTP-like responses that have been attributed to p75NTR and TrkB neurotrophin receptors through the action of proBDNF and BDNF, respectively (Lu et al. 2005; Reichardt, 2006). Brain-derived neurotrophic factor is a key signal transduction element in in vitro classical conditioning and is required for synaptic AMPAR delivery during CR acquisition (Li & Keifer, 2008, 2009; Keifer et al. 2009). Therefore, we tested whether p75NTR or TrkB receptors were involved in the p-PDK1 responses using selective inhibitors. Bath application of TAT-Pep5 (1 μm), a peptide inhibitor of p75NTR, blocked or reversed the reduction in p-PDK1 to unpaired stimulation (Fig.3D; P = 0.14, unpaired groups vs. N) but had no effect on the increase in p-PDK1 after paired stimulation (P = 0.001, paired groups vs. N). Blockade of p75NTR function by the REX antibody during conditioning resulted in similar findings for p-PDK1 during unpaired and paired stimulation, namely the normally reduced p-PDK1 response to unpaired stimuli was significantly reversed, while the increased p-PDK1 response to paired stimuli remained intact (Fig.3E; P = 0.003, unpaired groups vs. N; P = 0.008, paired group vs. N). In contrast, application of a TrkB antibody that inhibits its function blocked the pairing-evoked increase in p-PDK1 but had no effect on p-PDK1 suppression induced by unpaired stimuli (Fig.3F; P = 0.004, unpaired groups vs. N; P = 0.93, paired groups vs. N). The results are consistent with the conclusion that the p75NTR receptor mediates the reduction in p-PDK1 in response to unpaired stimuli, while TrkB mediates the increase produced by pairing.
Activation of PDK1 is regulated by adenosine A2A receptors complexed with p75NTR and TrkB
The action of TrkB receptors in the earliest stages of conditioning was curious because mature BDNF is minimally expressed until one session of conditioning (or after 25 min), when it is substantially increased by protein synthesis (Keifer et al. 2009; Ambigapathy et al. 2013). It was unclear what might activate TrkB after 5 min of nerve stimulation if not BDNF itself. An earlier study showed that adenosine or its agonists may activate Trk receptors through an interaction with the A2A receptor (Lee & Chao, 2001). We tested the hypothesis that increased p-PDK1 mediated by TrkB was activated by adenosine during pairing by using application of the selective A2A receptor antagonist, ZM 241385. Not only did administration of ZM 241385 (50 nm) block the increase in p-PDK1 after 5 min of paired stimulation, but surprisingly, it also inhibited the reduction in p-PDK1 in response to unpaired stimulation (Fig.4A; P = 0.20, all groups vs. N). The specificity of ZM 241385 for A2A receptors in our preparation was confirmed by incubating preparations with ZM 241385 and the A2A agonist CGS 21680 simultaneously, which resulted in blockade of reduced p-PDK1 responses normally produced by CGS alone (not shown).
Figure 4.
Activation of PDK1 is regulated by adenosine 2A (A2A) receptors
A, the selective A2A antagonist ZM 241385 inhibits responses of p-PDK1 to 5 min of unpaired (U, US; C, CS alone) and paired nerve stimulation compared with naive (P = 0.20; n = 3 per group). B and C, analysis of interactions of A2A, p75NTR and TrkB receptors during early stages of pairing using co-immunoprecipitation shows that there is a strong association of p75NTR with A2A receptors in naive preparations that is relatively unchanged after conditioning (B; P = 0.90, ANOVA all groups, n = 3 per group). The interaction of TrkB with A2A and p75NTR is weak in naive preparations but is substantially increased after conditioning (C; TrkB-A2A, P = 0.0008, conditioned groups vs. N, n = 3 per group; TrkB-p75NTR, P = 0.03, C5 vs. N, n = 3 per group). Abbreviations: C5, paired stimulation for 5 min; C15, paired stimulation for 15 min; IP, immunoprecipitate; IB, immunoblot. Input (whole-cell lysates from naive) and IgG lanes are also shown. D, application of the adenosine receptor agonist adenosine (Ad; 0.2 mm, 15 min; P = 0.0001, n = 8) or the A2A receptor agonist CGS 21680 alone (CGS; 20 nm, 45 min; P < 0.0001, n = 4) results in significantly reduced levels of p-PDK1 compared with naive. Cleavage-resistant proBDNF that stimulates p75NTR also reduces p-PDK1 (pro; 100 ng ml−1, 45 min; P < 0.0001, n = 4). In contrast, the TrkB receptor agonist DHF results in significantly increased p-PDK1 (10 μm, 45 min; P < 0.0001, n = 5). E, application of adenosine or the TrkB receptor agonist DHF results in significant activation of p-PKA above naive levels (P < 0.0001, Ad and DHF vs. N, n = 5 per group). *Represents significant differences from naïve throughout.
Given that adenosine receptors interact with and modu-late other types of receptors (Dias et al. 2013) and considering the results using ZM 241385, we wondered whether A2A receptors associate directly with p75NTR and TrkB in this preparation. Co-immunoprecipitation studies showed that A2A and p75NTR receptors interacted strongly, not only in naive preparations but also in those that were conditioned for 5 or 15 min (Fig.4B). TrkB receptors also immunoprecipitated with A2A, but this association was weak in naive preparations and strengthened during conditioning (Fig.4C; P = 0.0008, conditioned groups vs. N). As previous evidence suggested that p75NTR and Trk receptors might interact physically (Bibel et al. 1999), we also performed co-immunoprecipitation on samples from our preparation to determine whether this was the case here. The findings show that TrkB was again only weakly associated with p75NTR in naive preparations but interacted strongly with p75NTR after 5 min of pairing, while this interaction declined after 15 min (Fig.4C; P = 0.03, C5 vs. N). From these data it can be inferred that TrkB is recruited to A2A-p75NTR within 5 min of conditioning onset to form an A2A–p75NTR–TrkB receptor complex that regulates p-PDK1.
To obtain further support for the findings that A2A, p75NTR and TrkB receptors mediate bidirectional changes in p-PDK1 during conditioning, application of selective agonists was used to assess their effect on p-PDK1 when applied alone (Fig.4D). Application of adenosine (0.2 mm) or an A2A receptor agonist, CGS 21680 (20 nm), resulted in significantly decreased levels of p-PDK1 compared with naive preparations incubated in normal saline (Fig 4D; P = 0.0001 and P < 0.0001, respectively). Application of cleavage-resistant proBDNF (100 ng ml−1), which is not converted by proteolytic cleavage to mature BDNF and activates p75NTR, also reduced p-PDK1, similar to adenosine and its agonist (P < 0.0001). In contrast, a selective agonist of TrkB, DHF (10 μm), resulted in significantly increased p-PDK1 to 117% above naive levels (P < 0.0001). As anticipated, the TrkB agonist also significantly stimulated levels of p-PKA to 146% above naive values (Fig.4E; P < 0.0001), while application of adenosine also unexpectedly elevated p-PKA to similar values (P < 0.0001).
These pharmacological data are consistent with the interpretation that A2A and p75NTR receptors generate the reduction in p-PDK1 observed during unpaired stimulation, while activation of TrkB receptors results in increased p-PDK1 during paired stimulation. These data also suggest, however, that adenosine modulates both the positive and the negative changes in p-PDK1 to nerve stimulation, because the A2A receptor antagonist ZM 241385 inhibits both responses.
Signalling by PDK1 involves sphingolipids and Src family kinases
As its name suggests, PDK1 is activated by the PI3K lipid signalling cascade that generates phosphatidylinositol-3,4,5-triphosphate and anchors it to the membrane. To assess whether the bidirectional changes in p-PDK1 in response to nerve stimulation are signalled by corresponding changes in IP3, ELISA was performed on brainstem homogenates after paired or unpaired stimulation (Fig.5A). Consistent with our findings on p-PDK1, values for IP3 after unpaired stimulation were significantly reduced relative to naive preparations (C, CS alone; P = 0.02), while they were elevated after paired stimulation (P; P < 0.0001). To confirm these findings, bath application of the PI3K inhibitor wortmannin (200 nm) blocked the stimulus-induced changes in p-PDK1 to both unpaired (US or CS alone) and paired stimulation, indicating that phosphatidylinositol lipids underlie these responses (Fig.5B; P = 0.39, all groups).
Figure 5.
Signalling by PDK1 during nerve stimulation involves sphingolipids and Src family tyrosine kinases
A, ELISA for IP3 showed significantly reduced levels in preparations that received unpaired stimulation (C, CS alone; P = 0.02 vs. N, n = 4 per group) and increased levels after paired stimulation (P; P < 0.0001) compared with naive. B, bath application of the PI3K inhibitor wortmannin (200 nm) did not affect basal levels of p-PDK1 when applied to naive preparations (Nw, naive in wortmannin) but inhibited the stimulus-induced changes to unpaired (U, US; C, CS alone) and paired stimulation (P) for 5 min (P = 0.39, ANOVA all groups, n = 3 per group). C, the sphingomyelinase inhibitor sphingolactone-24 (Sphingo; 25 μm) blocked the negative response of p-PDK1 to unpaired nerve stimulation for 5 min, but the positive response to paired stimuli was unaffected (P = 0.59, unpaired groups vs. N, P = 0.006 paired group vs. N, n = 6 per group). D, application of the Src family tyrosine kinase inhibitor PP2 (0.5 μm) suppressed the increase in p-PDK1 that was normally observed after 5 min of paired nerve stimulation (P = 0.58 vs. N, n = 5 per group), while its negative control PP3 did not (P = 0.04 vs. N). The PLC inhibitor U73122 (U7; 4 μm) also failed to affect increased p-PDK1 after pairing (P = 0.002). *Represents significant differences from naïve throughout.
We previously showed that bath application of wortmannin inhibits acquisition and expression of CRs (Zheng & Keifer, 2008). Activation of p75NTR has been shown to result in the generation of the sphingolipid ceramide derived from membrane-bound sphingomyelin through the activity of sphingomyelinase (Reichardt, 2006; Luther et al. 2013). Ceramide has complex cellular effects and has been shown to inhibit or stimulate the PI3K pathway. Therefore, we asked whether ceramide might mediate the activity of p-PDK1 during nerve stimulation. Interestingly, application of the sphingomyelinase inhibitor sphingolactone-24 (25 μm) blocked the reduction of p-PDK1 in response to unpaired stimulation (Fig.5C; P = 0.59, unpaired groups vs. N) but failed to have an effect on the pairing-evoked increase in p-PDK1, which was elevated 115% above naive (P = 0.006). These results suggest that the negative effect of unpaired stimulation on p-PDK1 is mediated by metabolism of the sphingolipid ceramide, while the positive p-PDK1 response is not.
In order to define more specifically which signalling pathways are activated by pairing to result in increased p-PDK1, preparations were treated with the Src family tyrosine kinase inhibitor PP2 (0.5 μm) during 5 min of paired nerve stimulation. The results show that this compound inhibited the normally increased p-PDK1 response (Fig.5D; P = 0.58, PP2 vs. N). The negative control for PP2, PP3 (0.5 μm), did not suppress the increase in p-PDK1 in response to paired stimulation (Fig.5D; P < 0.05, PP3 vs. N). The PLC inhibitor U73122 (4 μm) also failed to affect the increased p-PDK1 response to paired stimulation (Fig.5D; P = 0.002, U73122 vs. N). These findings are consistent with the interpretation that TrkB receptors stimulate increased p-PDK1 in response to paired stimulation primarily through the action of the Src kinases.
Surface expression of TrkB is recruited within minutes of pairing and is mediated by A2A receptors and Src family kinases
It has been shown by others, using cultured neurons, that TrkB translocates to microdomains rich in cholesterol and sphingolipids in the plasma membrane (lipid rafts) in response to stimulation (Nagappan & Lu, 2005). In light of these findings, we examined whether TrkB was translocated to the membrane surface during conditioning using biotinylation assays. Paired stimulation of only 5 min resulted in significantly increased surface expression of biotinylated TrkB to 146% above naive levels that was further enhanced after 15 min of conditioning (Fig.6A; P = 0.0002 and P < 0.0001, respectively). Moreover, application of adenosine (0.2 mm; P = 0.0003) or the TrkB agonist DHF alone (10 μm; P = 0.001) also induced TrkB surface expression (Fig.6A). Unpaired stimulation not only failed to recruit surface TrkB but resulted in significantly less biotinylated TrkB, at 79% of naive values (Fig.6B, UP; P = 0.02). To determine whether TrkB recruitment to the membrane was dependent on phosphorylation by Src family tyrosine kinases, application of the inhibitor PP2 (0.5 μm) not only resulted in suppression of surface TrkB during pairing but also reduced TrkB in response to adenosine or DHF application (Fig.6B; P < 0.001, PP2-treated groups vs. paired group). Finally, the A2A receptor antagonist ZM 241385 (50 nm) applied to the bath during 5 min of paired stimulation resulted in inhibition of TrkB surface expression (Fig.6B; P = 0.13 compared with naive, P = 0.0005 vs. paired group). Taken together, these data suggest that TrkB is recruited to the membrane, where it interacts with A2A–p75NTR receptor complexes within 5 min of paired stimulation, and that this is mimicked by application of adenosine or DHF. The data showing blockade by ZM 241385 also support the interpretation that activation of A2A receptors during pairing results in transactivation of TrkB through phosphorylation by Src kinases that recruit it to the membrane.
Figure 6.
Surface TrkB expression is induced by pairing or A2A and TrkB receptor agonists and is dependent on Src family kinases
A, paired nerve stimulation for 5 or 15 min significantly increases surface expression of TrkB compared with naive (P = 0.0002 and P < 0.0001, respectively, n = 5 per group), as does application of adenosine (0.2 mm for 15 min; P = 0.0003, n = 5) or DHF (10 μm for 45 min; P = 0.001, n = 5). B, in a different set of experiments, paired nerve stimulation for 5 min increased biotinylated TrkB (P = 0.003, n = 4), while unpaired stimuli (UP) for 15 min significantly reduced surface TrkB compared with naive (P = 0.02, n = 5). Treatment of preparations with the Src kinase inhibitor PP2 (0.5 μm) suppressed surface TrkB in response to pairing and agonist application compared with normal pairing (P5min, P = 0.001; Ad, P = 0.0008; DHF, P = 0.0002; n = 5 per group). Treatment with the A2A antagonist ZM 241385 (ZM; 50 nm) during 5 min of pairing also inhibited expression of surface TrkB (P = 0.13 compared with naive, P = 0.0005 compared with normal P5min, n = 5). Blots for surface actin, total TrkB and total actin are also shown. Input lanes are whole-cell lysates from naive. Control lanes (Ctl) are whole-cell lysis samples without biotin treatment. *Represents significant differences from naïve throughout.
Effects of PDK1 inhibition on conditioned responding
Finally, to confirm that PDK1 is critical for coincidence detection and acquisition of CRs, conditioning was examined in the presence of the PDK1 inhibitor, BX (0.3 μm; Fig.7). When BX was applied to the bath prior to paired stimulation followed by training, the acquisition of CRs, which normally occurs by the second pairing session, was completely blocked (mean 0% CRs; Fig.7A). However, when BX was applied after CR acquisition had already occurred and was in the expression phase, there was no inhibition of conditioned responding (mean 100% CRs during pairing sessions 3–6; P = 0.87, sessions 3–6 vs. session 2; Fig.7B). Representative physiological recordings from the abducens nerve during the experiments summarized in Fig.7B show an unconditioned response (UR) during the first session in normal saline and a CR followed by the UR during the session 6 in BX (Fig.7C). Western blots from these conditioning experiments indicate that p-PDK1 was inhibited in the presence of BX regardless of whether it was applied at the beginning of training, when CRs are normally acquired, or later, after CRs have been expressed (Fig.7D). Importantly, however, p-PKA was inhibited only after BX treatment during acquisition resulting in blockade of CRs, while it was increased relative to naive and was similar to normal conditioning when acquisition proceeded in normal saline and BX was applied thereafter. These data suggest that once PKA is activated by cAMP and p-PDK1 during the acquisition phase of conditioning, it remains activated and supports CR expression.
Figure 7.
Acquisition of conditioning is blocked in the presence of the PDK1 inhibitor BX
A, paired stimulation for two pairing sessions during bath application of BX (0.3 μm) prevented acquisition of conditioned responses (CRs) recorded in the abducens nerve (mean 0% CRs for sessions 1 and 2; n = 6). B, acquisition of CRs was obtained after the second pairing session in normal saline. BX was applied to the bath after the second session and remained in the bath while CRs continued to be expressed for four additional pairing sessions (mean 100% CRs for sessions 3–6; n = 6). C, representative physiological recordings from the abducens nerve during the experiments in B during the first session in normal saline, showing the UR alone, and during session 6 in BX, showing a CR (arrow) followed by the UR. There was no change in physiological responsiveness recorded in the abducens nerve to stimulation in BX, as illustrated in the records. The CS–US stimuli are also indicated. D, Western blots for phosphorylated and total PDK1 and PKA taken from naive preparations (N), those conditioned for two sessions in normal saline (C2) and preparations conditioned for two sessions in BX during acquisition (shown in A; BX-C2) or expression of CRs in session 6 (shown in B; BX-C6). Levels of p-PDK1 are similar to naive after BX treatment in both training groups, but p-PKA is elevated similar to normal conditioning when BX is applied after CR acquisition (BX-C6).
Discussion
Coincidence detection prior to NMDAR activation
Many ideas for coincidence detection underlying synaptic modifications during learning have been founded on Donald Hebb's hypothesis that synaptic strengthening occurs in response to simultaneous pre- and postsynaptic activity (Hebb, 1949). The NMDAR uniquely fulfils Hebb's postulate because it requires both presynaptic release of glutamate and postsynaptic depolarization to relieve the Mg2+ block of the channel pore in order to function and is therefore widely considered to be a molecular coincidence detector (Tsien, 2000). While studies have demonstrated the unequivocal role of NMDARs in associative learning and memory, whether they are involved specifically in coincidence detection is questionable. While not excluding NMDAR function directly, the studies by Lorenzetti et al. (2008) in Aplysia and Gervasi et al. (2010) in Drosophila indicate that the AC–cAMP complex underlies the synergistic responses of PKA associated with paired stimulation and conditioned responding. More recently, olfactory conditioning using transgenic flies in which the Mg2+ block site of dNMDARs was mutated resulted in intact associative learning and short-term memory evaluated immediately and 1 h after training. However, CREB-dependent gene expression and long-term memory formation assayed 1 day after training was inhibited (Miyashita et al. 2012). While these data implicate NMDARs in long-term memory, the observation that short-term associative learning was retained in mutant flies suggests that NMDARs are not involved or only minimally involved in early stages of this learning. Rather, AC–cAMP and PKA may be used to initiate learning-related kinase signalling, which is upstream from CREB activity and protein synthesis.
In our model of in vitro eyeblink classical conditioning, coincidence detection occurs well before activity of NMDARs. This conclusion comes from several lines of evidence. First, the conditioning-induced increase in p-PKA is not affected by AP-5 after 15 min or one pairing session (25 min) of conditioning (Zheng & Keifer, 2009). Second, the synaptic insertion of GluA1-containing AMPARs during early conditioning is not sensitive to application of the NMDAR antagonist AP-5, but the later synaptic delivery of GluA4 subunits that follows is inhibited (Li & Keifer, 2009). Moreover, while the PKA agonist Sp-cAMPs induces synaptic delivery of both GluA1- and GluA4-containing AMPARs, only delivery of GluA4 subunits is suppressed by co-application of Sp-cAMPs and AP-5 (Zheng & Keifer, 2009). Together, these data support the conclusion that coincidence detection in this system precedes NMDAR function and, furthermore, our two-stage model for classical conditioning (Zheng et al. 2012; Zheng & Keifer, 2014), in which the earliest stage of AMPAR trafficking uses NMDAR-independent mechanisms for synaptic delivery of GluA1-containing AMPARs.
Signalling mechanisms for coincidence detection during in vitro classical conditioning
Within 5 min of paired stimulation, we observed substantially increased TrkB membrane surface expression, enhanced TrkB interaction with A2A and p75NTR receptors and increased formation of IP3 and p-PDK1 activation. This is followed by phosphorylation of PKA at Thr197 by PDK1 that initiates downstream signalling cascades for conditioning (Zheng & Keifer, 2009). The initiation of this process is dependent on the Src family tyrosine kinases and activity of the A2A receptor.
In light of these data, we constructed a working model to summarize the signal transduction process associated with the detection of paired stimuli that initiates classical conditioning (Fig.8). First, we hypothesize that adenosine is co-released with the neurotransmitter glutamate from presynaptic nerve terminals during stimulation, but the exact source and nature of this release is currently uncertain. However, both the auditory (CS) and trigeminal nerves (US) are immunopositive for adenosine receptors (not shown). In support of a role for adenosine in conditioning, endogenous adenosine and activation of A2A receptors has a critical role in acquisition of trace eyeblink conditioning in behaving mice (Fontinha et al. 2009). Here, adenosine stimulation of the G protein-coupled A2A receptor results in two distinct processes (Fig.8), namely the direct activation of AC–cAMP (Dias et al. 2013) and the phosphorylation of TrkB through transactivation, an indirect process, in which one receptor is activated by another receptor that responds to a different ligand (Lee & Chao, 2001; Rajagopal et al. 2004). Evidence for transactivation of TrkB by A2A is supported by the observation that the A2A inhibitor ZM 241385 blocks TrkB-mediated increases in p-PDK1 and TrkB surface expression in response to paired stimulation.
Figure 8.
Model for coincidence detection during classical conditioning
The signalling pathways shown are simplified to focus on the coincidence detection function of the receptors described here. Pairing of the CS and US during conditioning induces two distinct postsynaptic processes. First, there is direct activation of the AC–cAMP cascade through the G protein-coupled adenosine 2A receptor; and secondly, these receptors phosphorylate TrkB through transactivation by using the Src family tyrosine kinases. Phosphorylation of TrkB recruits it to the membrane within 5 min of paired nerve stimulation, where it stimulates the activity of PI3K and PDK1 in a conditioning-dependent manner. The accumulation of cAMP through direct activation of A2A acts to release the catalytic subunits of the PKA holoenzyme, allowing it to become phosphorylated by the convergent activity of PDK1 that is required for downstream substrate phosphorylation by PKA to initiate the signal transduction cascade that underlies conditioning. Alternatively, unpaired nerve stimulation activates p75NTR receptors and the sphingomyelin signalling pathway that results in a reduction in PI3K and PDK1 activity. Interestingly, this pathway is also controlled by A2A receptors, which immunoprecipitate with p75NTR, possibly through a unique form of transactivation (i.e. transinhibition).
The Src family tyrosine kinases play an important role in the transactivation of TrkB. This was demonstrated by using the antagonist PP2, which inhibited TrkB surface expression as well as suppressing p-PDK1 in response to paired stimuli. Failure of the PLC inhibitor U73122 to suppress p-PDK1 to pairing suggests that this TrkB signalling pathway has minimal involvement in coincidence detection. However, this negative finding cannot rule out a role for PLC activation entirely, because evidence suggests it is important for both LTP and trace eyeblink conditioning in mice (Gruart et al. 2007). Activation of TrkB leads recruitment of several adaptor proteins, including Src homology 2 domain-containing (Shc) adaptor protein and Grb2-associated binder-1 (GAB1) to activate the PI3K/PDK1 pathway (Reichardt, 2006; Duman & Voleti, 2012).
This signalling process takes place while cAMP is generated from AC directly by A2A receptors. The cAMP acts to release the catalytic subunits of the PKA holoenzyme through a conformational change that allows the priming phosphorylation of its activation loop (Thr197) by p-PDK1 required for downstream substrate phosphorylation to initiate conditioning. While the requirement for PDK1 in PKA activation in vivo is not universal (Newton, 2003; Pearce et al. 2010), data here show that PKA phosphorylation during conditioning is suppressed by the PDK1 blocker BX. 3-Phosphoinositide-dependent kinase-1 binding and phosphorylation of its targets is regulated not only by the conformational state of the substrate (to expose the activation loop) but also by the subcellular localization of its binding partners (Newton, 2003). Therefore, the activities of PDK1 and PKA are required to be spatially and temporally convergent to signal an association between the paired inputs.
Evidence using TAT-Pep5 or the REX antibody suggests that the reduction in p-PDK1 observed after unpaired CS or US alone nerve stimulation is mediated by p75NTR and sphingomyelin signalling pathways (Fig.8). Notably, the decrease in p-PDK1 is mimicked by application of adenosine or its agonists. This finding raises a perplexing question, i.e. why does adenosine application result in decreased levels of p-PDK1 and increased p-PKA? Bath application of adenosine is clearly unlike nerve stimulation. The decrease in p-PDK1 produced by adenosine application is not blocked by the p75NTR antagonist TAT-Pep5 (data not shown), which is different from observations using unpaired nerve stimulation. However, this response to adenosine is inhibited by sphingolactone-24, which inhibits the downstream p75NTR pathway described here for unpaired nerve stimulation. The finding that ZM 241385 inhibits both the positive and negative p-PDK1 responses to paired and unpaired stimuli, respectively, indicates that A2A receptors are involved in lowering p-PDK1 levels as well as enhancing them. Therefore, there is an unspecified modulatory control of the p75NTR pathway by A2A receptors, perhaps through a unique form of transactivation (i.e. transinhibition). This suggests that A2A receptors exert a push–pull control over PI3K/PDK1 levels and may explain why p-PDK1 responses to unpaired nerve stimulation are reversed when p75NTR is blocked.
In the present study, we focused our attention on the signalling events during the earliest stages of pairing before CRs are recorded, but are they relevant to acquisition of conditioning? Data indicate that without these initial signalling cascades, conditioning fails to occur. Previously, we showed that inhibition of PKA activity by Rp-cAMPs blocked both the acquisition and the expression of CRs (Zheng & Keifer, 2009). Furthermore, application of the same TrkB antibody shown here to suppress enhanced p-PDK1 responses to paired nerve stimulation completely inhibited conditioning (Li & Keifer, 2008). Similar results on conditioning were also obtained for the PI3K antagonist, wortmannin (Zheng & Keifer, 2008). Findings here indicate that the PDK1 inhibitor BX, when applied at the beginning of pairing, inhibits not only p-PDK1 but also downstream p-PKA, leading to a failure in the acquisition of CRs. However, when BX was applied after acquisition had occurred in normal saline, p-PKA activity was maintained through its autoregulatory mechanisms and supported CR expression. Our recent data using chromatin immunoprecipitation assays also indicate that BX or Rp-cAMPs inhibit CREB binding to the turtle BDNF promoter regions, which is required for BDNF protein expression and conditioned responding (Li & Keifer, 2008; Keifer et al. 2009; G. Ambigapathy, Z. Zheng & J. Keifer, unpublished data). Taken together, activation of signalling pathways involving AC–cAMP, PI3K and PDK1 that lead to phosphorylation of PKA can be considered substrates for coincidence detection during the earliest stages of pairing and are fundamental to generate conditioned responding.
Delayed activation of PKA relative to PDK1 during pairing
The activation of PDK1 is substantially increased within 5 min of paired nerve stimulation during conditioning, but PKA does not undergo significant phosphorylation until 15 min of pairing. What underlies this delay in PKA activation? In previous studies, we (Zheng & Keifer, 2014) and others (Huganir & Nicoll, 2013) have shown that assembly of scaffolding proteins during learning is required for synaptic AMPAR delivery. In our model system, these include protein kinase A-anchoring protein and synapse-associated protein 97; interactions that are detected 15 min after pairing to deliver GluA1-containing AMPARs to the postsynaptic density (Zheng & Keifer, 2014). The assembly of protein complexes takes time to form and move into the postsynaptic density. For example, redistribution of protein kinase A-anchoring protein into synaptosomal fractions in response to chemical LTD was observed after 5 min of treatment but was not significantly increased until after 15 min (Smith et al. 2006). Given that PDK1 undergoes autophosphorylation (Casamayor et al. 1999), maintaining it in an ‘on’ state, such a time delay does not pose a problem for convergence of PKA with active PDK1. It may also take time for enough cAMP or activated PDK1 to accumulate locally to a functional concentration threshold, which results in a delay in production of phosphorylated PKA. Whatever the mechanism, the autophosphorylation of PDK1 is central to sustaining the prolonged kinase activation required for learning (see also Mullasseril et al. 2007).
What is the true nature of coincidence detection?
While we have described key signal transduction events that occur at the beginning of paired stimulation during classical conditioning, how paired CS–US presentations lead to signalling changes while backward pairing (US–CS) does not is key to a full understanding of the mechanisms underlying coincidence detection. One possibility is that transactivation is central to this process because it provides directionality to signalling processes; that is, CS–US presentations lead to conditioning, while US–CS training does not. For example, if A2A receptors are associated mainly with the CS inputs, then the CS will be required to precede the US if transactivation of nearby receptors is a necessary mechanism. Therefore, one critical issue is clarification of the arrangement of the presynaptic nerve terminals with the underlying postsynaptic membrane receptors. How coincident stimuli are detected is fundamental to understanding the mechanisms for associative learning and the learning disorders that characterize a number of neurological syndromes and disease states. In the context of the present study, it is intriguing that decreased plasma phospholipids, particularly the phosphatidylinositols, appear to provide promising biomarkers for preclinical Alzheimer's disease (Mapstone et al. 2014).
Acknowledgments
We thank Drs Louis Reichardt (University of California, San Francisco) for the generous gift of the REX antibody and Victor Huber (University of South Dakota) for help with the ELISA. We thank Dr Samuel Sathyanesan for critically reading the manuscript.
Glossary
- A2A
adenosine 2A receptor
- AC
adenylate cyclase
- BDNF
brain-derived neurotrophic factor
- CREB
cAMP response element-binding protein
- CR
conditioned response
- CS
conditioned stimulus
- IP3
inositol trisphosphate
- LTD
long-term depression
- LTP
long-term potentiation
- PDK1
3-phosphoinositide-dependent kinase-1
- PI3K
phosphoinositide 3-kinase
- PKA
protein kinase A
- PKC
protein kinase C
- PLC
phospholipase C
- Trk
tropomyosin-related kinase
- UR
unconditioned response
- US
unconditioned stimulus
Additional information
Competing interests
None declared.
Author contributions
Conception and design of the experiments: J.K. Collection, analysis and interpretation of data, drafting and revising the article: J.K. and Z.Z. Both authors approved the final version of the manuscript. All persons designated as authors qualify for authorship, and those who qualify for authorship are listed.
Funding
This work was supported by the National Institutes of Health grant NS051187 (J.K.).
References
- Ambigapathy G, Zheng Z, Li W. Keifer J. Identification of a functionally distinct truncated BDNF mRNA splice variant and protein in Trachemys scripta elegans. PLoS One. 2013;8:e67141. doi: 10.1371/journal.pone.0067141. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Bain J, McLauchlan H, Elliot M. Cohen P. The specificities of protein kinase inhibitors: an update. Biochem J. 2003;371:199–204. doi: 10.1042/BJ20021535. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Bibel M, Hoppe E. Barde Y-A. Biochemical and functional interactions between the neurotrophin receptors trk and p75NTR. EMBO J. 1999;18:616–622. doi: 10.1093/emboj/18.3.616. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Bleasdale JE, Thakur NR, Gremban RS, Bundy GL, Fitzpatrick FA, Smith RJ. Bunting J. Selective inhibition of receptor-coupled phospholipase C-dependent processes in human platelets and polymorphonuclear neutrophils. J Pharmacol Exp Ther. 1990;255:756–768. [PubMed] [Google Scholar]
- Casamayor A, Morrice NA. Alessi DR. Phosphorylation of Ser-241 is essential for the activity of 3-phosphoinositide-dependent protein kinase-1: identification of five sites of phosphorylation in vivo. Biochem J. 1999;342:287–292. [PMC free article] [PubMed] [Google Scholar]
- Dias RB, Rombo DM, Ribeiro JA, Henley JM. Sebastiao AM. Adenosine: setting the stage for plasticity. Trends Neurosci. 2013;36:248–257. doi: 10.1016/j.tins.2012.12.003. [DOI] [PubMed] [Google Scholar]
- Duman RS. Voleti B. Signaling pathways underlying the pathophysiology and treatment of depression: novel mechanisms for rapid-acting agents. Trends Neurosci. 2012;35:47–56. doi: 10.1016/j.tins.2011.11.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Fontinha BM, Delgado-García JM, Madroñal N, Ribeiro JA, Sebastião AM. Gruart A. Adenosine A2A receptor modulation of hippocampal CA3-CA1 synapse plasticity during associative learning in behaving mice. Neuropsychopharmacology. 2009;34:1865–1874. doi: 10.1038/npp.2009.8. [DOI] [PubMed] [Google Scholar]
- Gervasi N, Tchénio P. Preat T. PKA dynamics in a Drosophila learning center: coincidence detection by Rutabaga adenylyl cyclase and spatial regulation by Dunce phosphodiesterase. Neuron. 2010;65:516–529. doi: 10.1016/j.neuron.2010.01.014. [DOI] [PubMed] [Google Scholar]
- Gruart A, Sciarretta C, Valenzuela-Harrington M, Delgado-García JM. Minichiello L. Mutation at the TrkB PLCγ-docking site affects hippocampal LTP and associative learning in conscious mice. Learn Mem. 2007;14:54–62. doi: 10.1101/lm.428307. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Hebb DO. The Organization of Behavior: A Neuropsychological Theory. New York: Wiley; 1949. [Google Scholar]
- Hou L. Klann E. Activation of the phosphoinositide 3-kinase-Akt-mammalian target of rapamycin signaling pathway is required for metabotropic glutamate receptor-dependent long-term depression. J Neurosci. 2004;24:6352–6361. doi: 10.1523/JNEUROSCI.0995-04.2004. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Huganir RL. Nicoll RA. AMPARs and synaptic plasticity: the last 25 years. Neuron. 2013;80:704–717. doi: 10.1016/j.neuron.2013.10.025. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Keifer J. In vitro eye-blink reflex model: role of excitatory amino acid receptors and labeling of network activity with sulforhodamine. Exp Brain Res. 1993;97:239–253. doi: 10.1007/BF00228693. [DOI] [PubMed] [Google Scholar]
- Keifer J, Armstrong KE. Houk JC. In vitro classical conditioning of abducens nerve discharge in turtles. J Neurosci. 1995;15:5036–5048. doi: 10.1523/JNEUROSCI.15-07-05036.1995. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Keifer J. Houk JC. Modeling signal transduction in classical conditioning with network motifs. Front Mol Neurosci. 2011;4:1–8. doi: 10.3389/fnmol.2011.00009. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Keifer J, Sabirzhanov BE, Zheng Z, Li W. Clark TG. Cleavage of proBDNF to BDNF by a tolloid-like metalloproteinase is required for acquisition of in vitro eyeblink classical conditioning. J Neurosci. 2009;29:14956–14964. doi: 10.1523/JNEUROSCI.3649-09.2009. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Lee FS. Chao MV. Activation of Trk neurotrophin receptors in the absence of neurotrophins. Proc Natl Acad Sci USA. 2001;98:3555–3560. doi: 10.1073/pnas.061020198. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Li W. Keifer J. Coordinate action of pre- and postsynaptic brain-derived neurotrophic factor is required for AMPAR trafficking and acquisition of in vitro classical conditioning. Neuroscience. 2008;155:686–697. doi: 10.1016/j.neuroscience.2008.06.043. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Li W. Keifer J. BDNF-induced synaptic delivery of AMPAR subunits is differentially dependent on NMDA receptors and requires ERK. Neurobiol Learn Mem. 2009;91:243–249. doi: 10.1016/j.nlm.2008.10.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Lorenzetti FD, Baxter DA. Byrne JH. Molecular mechanisms underlying a cellular analog of operant reward learning. Neuron. 2008;59:815–828. doi: 10.1016/j.neuron.2008.07.019. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Lu B, Pang PT. Woo NH. The Yin and Yang of neurotrophin action. Nat Rev Neurosci. 2005;6:603–614. doi: 10.1038/nrn1726. [DOI] [PubMed] [Google Scholar]
- Luther JA, Enes J. Birren SJ. Neurotrophins regulate cholinergic synaptic transmission in cultured rat sympathetic neurons through a p75-dependent mechanism. J Neurophysiol. 2013;109:485–496. doi: 10.1152/jn.00076.2011. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Mapstone M, Cheema AK, Fiandaca MS, Zhong X, Mhyre TR, MacArthur LH, Hall WJ, Fisher SG, Peterson DR, Haley JM, Nazar MD, Rich SA, Berlau DJ, Peltz CB, Tan MT, Kawas CH. Federoff HJ. Plasma phospholipids identify antecedent memory impairment in older adults. Nat Med. 2014;20:415–418. doi: 10.1038/nm.3466. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Minichiello L. TrkB signaling pathways in LTP and learning. Nat Rev Neurosci. 2009;10:850–860. doi: 10.1038/nrn2738. [DOI] [PubMed] [Google Scholar]
- Miyashita T, Oda Y, Horiuchi J, Yin JCP, Morimoto T. Saitoe M. Mg2+ block of Drosophila NMDA receptors is required for long-term memory formation and CREB-dependent gene expression. Neuron. 2012;74:887–898. doi: 10.1016/j.neuron.2012.03.039. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Mullasseril P, Dosemeci A, Lisman JE. Griffith LC. A structural mechanism for maintaining the ‘on-state’ of the CaMKII memory switch in the post-synaptic density. J Neurochem. 2007;103:357–364. doi: 10.1111/j.1471-4159.2007.04744.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Nagappan G. Lu B. Activity-dependent modulation of the BDNF receptor TrkB: mechanisms and implications. Trends Neurosci. 2005;28:464–471. doi: 10.1016/j.tins.2005.07.003. [DOI] [PubMed] [Google Scholar]
- Newton AC. Regulation of the ABC kinases by phosphorylation: protein kinase C as a paradigm. Biochem J. 2003;370:361–371. doi: 10.1042/BJ20021626. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Opazo P, Watabe AM, Grant SG. O'Dell TJ. Phosphatidylinositol 3-kinase regulates the induction of long-term potentiation through extracellular signal-related kinase-independent mechanisms. J Neurosci. 2003;23:3679–3688. doi: 10.1523/JNEUROSCI.23-09-03679.2003. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Pearce LR, Komander D. Alessi DR. The nuts and bolts of AGC protein kinases. Nat Rev Mol Cell Biol. 2010;11:9–22. doi: 10.1038/nrm2822. [DOI] [PubMed] [Google Scholar]
- Rajagopal R, Chen Z-Y, Lee FS. Chao MV. Transactivation of Trk neurotrophin receptors by G-protein-coupled receptor ligands occurs on intracellular membranes. J Neurosci. 2004;24:6650–6658. doi: 10.1523/JNEUROSCI.0010-04.2004. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Reichardt LF. Neurotrophin-regulated signalling pathways. Phil Trans R Soc B Biol Sci. 2006;361:1545–1564. doi: 10.1098/rstb.2006.1894. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Smith KE, Gison ES. Dell'Acqua ML. cAMP-dependent protein kinase postsynaptic localization regulated by NMDA receptor activation through translocation of an A-kinase anchoring protein scaffold protein. J Neurosci. 2006;26:2391–2402. doi: 10.1523/JNEUROSCI.3092-05.2006. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Sui L, Wang J. Li B-M. Role of the phosphoinositide 3-kinase-Akt-mammalian target of the rapamycin signaling pathway in long-term potentiation and trace fear conditioning memory in rat medial prefrontal cortex. Learn Mem. 2008;15:762–776. doi: 10.1101/lm.1067808. [DOI] [PubMed] [Google Scholar]
- Tsien JZ. Linking Hebb's coincidence-detection to memory formation. Curr Opin Neurobiol. 2000;10:266–273. doi: 10.1016/s0959-4388(00)00070-2. [DOI] [PubMed] [Google Scholar]
- Zheng Z. Keifer J. Protein kinase C-dependent and independent signaling pathways regulate synaptic GluR1 and GluR4 AMPAR subunits during in vitro classical conditioning. Neuroscience. 2008;156:872–884. doi: 10.1016/j.neuroscience.2008.08.042. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Zheng Z. Keifer J. PKA has a critical role in synaptic delivery of GluR1- and GluR4-containing AMPARs during initial stages of acquisition of in vitro classical conditioning. J Neurophysiol. 2009;101:2539–2549. doi: 10.1152/jn.91282.2008. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Zheng Z. Keifer J. Sequential delivery of synaptic GluA1 and GluA4-containing AMPARs by SAP97 anchored protein complexes in classical conditioning. J Biol Chem. 2014;289:10540–10550. doi: 10.1074/jbc.M113.535179. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Zheng Z, Sabirzhanov B. Keifer J. Two-stage AMPA receptor trafficking in classical conditioning and selective role for glutamate receptor subunit 4 (tGluA4) flop splice variant. J Neurophysiol. 2012;108:101–111. doi: 10.1152/jn.01097.2011. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Zhu D. Keifer J. Distribution of facial motor neurons in the pond turtle Pseudemys scripta elegans. Neurosci Lett. 2005;373:134–137. doi: 10.1016/j.neulet.2004.10.022. [DOI] [PubMed] [Google Scholar]