Quantitative description of the interactions among kinase cascades underlying long-term plasticity of Aplysia sensory neurons

Abstract

Kinases play critical roles in synaptic and neuronal changes involved in the formation of memory. However, significant gaps exist in the understanding of how interactions among kinase pathways contribute to the mechanistically distinct temporal domains of memory ranging from short-term memory to long-term memory (LTM). Activation of protein kinase A (PKA) and mitogen-activated protein kinase (MAPK)-ribosomal S6 kinase (RSK) pathways are critical for long-term enhancement of neuronal excitability (LTEE) and long-term synaptic facilitation (LTF), essential processes in memory formation. This study provides new insights into how these pathways contribute to the temporal domains of memory, using empirical and computational approaches. Empirical studies of Aplysia sensory neurons identified a positive feedforward loop in which the PKA and ERK pathways converge to regulate RSK, and a negative feedback loop in which p38 MAPK inhibits the activation of ERK and RSK. A computational model incorporated these findings to simulate the dynamics of kinase activity produced by different stimulus protocols and predict the critical roles of kinase interactions in the dynamics of these pathways. These findings may provide insights into the mechanisms underlying aberrant synaptic plasticity observed in genetic disorders such as RASopathies and Coffin-Lowry syndrome.

Figure 1

Dynamic regulation of pRSK (A) and PKA catalytic subunits (PKAc) (B) by a brief (5 min) pulse of 50 μM 5-HT. A1, Representative confocal images of pRSK immunofluorescence in SNs at different times post-onset of 5-HT. A2, Summary data. The percent change was calculated as the change of pRSK level after 5-HT compared to time-matched control levels. pRSK increased immediately after 5-HT (n = 8), returned to basal level at 15 min (n = 9), followed by a delayed increase at about 45 min (n = 13), and then a return to basal level at 60 min (n = 8). B1, Representative confocal images of PKAc at different times post-onset of 5-HT. B2, Summary data. The percent change was calculated as the change of PKAc level after 5-HT compared to time-matched control levels. PKAc increased immediately after 5-HT (n = 10), returned to basal level at 15 min (n = 7), and remained at basal level at 45 min (n = 8). Data are represented as mean ± SEM. All scale bars are 40 μm. *p < 0.05.

Figure 2

The interactions between PKA and ERK (A,B), PKA and RSK (C,D), RSK and p38 MAPK (E). A1, Protocol for applying the PKA inhibitor KT5720 with 5 min 5-HT. A2, Representative confocal images of pERK in SNs at 45 min post-onset of 5-HT, in the absence or presence of KT5720. A3, Summary data. KT5720 significantly decreased pERK induced by 5-HT (n = 10). B1, Protocol for applying the PKA inhibitor Rp-cAMP with 5-HT. B2, Representative confocal images of pERK at 45 min post-onset, in the absence or presence of Rp-cAMP. B3, Summary data. Rp-cAMP significantly decreased pERK induced by 5-HT (n = 6). C1, Protocol. C2, Representative confocal images of pRSK immediately after 5-HT, in the absence or presence of KT5720. C3, Summary data. KT5720 significantly decreased pRSK induced by 5-HT (n = 7). D1, Protocol. D2, Representative confocal images of pRSK immediately after 5-HT, in the absence or presence of Rp-cAMP. D3, Summary data. Rp-cAMP significantly decreased pRSK induced by 5-HT (n = 9). E1, Protocol for applying the MEK inhibitor U0126 with 5-HT. E2, Representative confocal images of pRSK immediately after 5-HT, in the absence or presence of U0126. E3, Summary data. U0126 did not significantly attenuate the increase of pRSK immediately after 5-HT (n = 6). F1, Protocol for applying the RSK inhibitor BI-D1870 (BID) with 5-HT. F2, Representative confocal images of p-p38 MAPK at 45 min post-onset of 5-HT, in the absence or presence of BID. F3, Summary data. BID attenuated the induction of p-p38 MAPK 45 min post-onset of 5-HT (n = 9). Data are represented as mean ± SEM. All scale bars are 40 μm. *p < 0.05.

Figure 3

(A) Schematic model of PKA and MAPK signaling pathways. 5-HT regulates the PKA and MAPK cascades via multiple pathways. Red, blue, green dashed lines represent newly added pathways. Blue denotes the ERK/RSK/p38 MAPK feedback loop, green denotes the NT-dependent pathways. Each number represents a signaling pathway (not equation numbers). Arrowheads indicate activation, circular ends indicate repression. (B–D) Simulated dynamics of kinases after one pulse of 5-HT. (B) Control simulations (black solid curves). Empirical data points (red circles) are from this study (PKAc and pRSK) and from Zhang et al. (pERK and p-p38 MAPK). Numbers “1”, “2” in B2 represent two waves of increases in pRSK. The “1” in B3 represents the transient decrease of p-p38 MAPK. The “2” in B3 represents the delayed increase of p-p38 MAPK. (C) Dynamics of kinases with a simulated inhibition of downstream effects of PKA, ended 45 min post-onset of 5-HT. The slow increases of pERK (C1), pRSK (C2) and p-p38 MAPK (C3) were blocked. Blue curves are simulations with inhibitors, black dashed curves are control simulation. This inhibition suppressed the delayed activation of pERK as observed empirically in Fig. 2A, and pRSK and p-p38 MAPK subsequently decreased (black arrows). Control simulation curves added in this and following panels and figures, are for the convenience of comparison with the curves in the presence of inhibitors. Dashed curves are used to make overlapped curves visible. (D) Dynamics of kinases with simulated inhibition of RSK ending 60 min post-onset of 5-HT. The delayed increase of p-p38 MAPK (D3) was blocked by RSK inhibition, but pERK (D1) and pRSK (D2) remained elevated for 60 min. Green curves are simulations with inhibitors, black dashed curves are control simulation. This RSK inhibition suppressed the delayed activation of p-p38 MAPK (D3, black arrow) as observed empirically (Fig. 2F). Reduced p-p38 MAPK activity disinhibited MEK. The disinhibition led to a sustained increase of pERK and pRSK at 1 h (D1-2, red arrows).

Figure 4

Schematic network including transcription factors CREB1/2, regulated by PKA and MAPK signaling pathways. (A) Complete pathways in the extended model. The pathways from Fig. 3A are represented as solid, thinner, lines, and the pathway numbers are indicated with unfilled circles. The newly added pathways are dashed, thicker lines, and pathway numbers are indicated with filled circles. The PKA and MAPK cascades interact to regulate the phosphorylation and activities of CREB1 and CREB2, which regulate the expression of TBL. TBL subsequently activates TGF-β. Violet denotes two ERK – > TBL pathways: ERK- > RSK- > CREB1- > TBL, pathway 7- > 16- > 19); ERK- > CREB2- > TBL, pathway 17- > 20). Arrowheads indicate activation, circular ends indicate repression. (B) Network highlighting the addition of new pathways, with the previous pathways made transparent.

Figure 5

(A,B) Simulated dynamics of pERK, pRSK, p-p38 MAPK, PKAc, pCREB1, pCREB2_p38, NT, TBL and TGF-β levels after one or two 5-min pulses of 5-HT with ISI of 45 min, without block of TrkB (A), or with block of TrkB applied during both pulses of 5-HT (B). (A) Black solid curves, simulations after two pulses of 5-HT. Green curves, simulations after one pulse of 5-HT. Red arrows in A1 and A4 represent the overlapped increases of pERK and PKAc after two pulses of 5-HT. (B) Blue curves, simulations after two pulses with inhibitors; black dashed curves, control simulations without inhibitors, same as black solid curves in A. Numbers “1”, “2”, and “3” in A2, A5, A8, B2, B5, and B8 represent the waves or late shoulders of increase. Numbers “1”, and “2” in A3, A6, B3, and B6 represent the waves of decrease. (C) Empirical validation that TrkB Fc applied during both pulses of 5-HT suppressed the increase of pERK at 1 h post-onset of second pulse of 5-HT (n = 7). Data are represented as mean ± SEM. All scale bars are 40 μm. * p < 0.05.

Figure 6

Simulated dynamics of pERK, TGF-β, and TBL after one pulse of 5-HT, with or without suppression of pathway 8 (purple ‘x’) or pathway 11 (green ‘x’) in (A). (B) Dynamics of pERK, TGF-β, and TBL after one pulse of 5-HT with blocking of pathway 8 (purple curves). Black dashed curves are control simulations with no suppression. (C–E) Dynamics of pERK, TGF-β, and TBL after two pulses of 5-HT (ISI = 45 min) with block of pathway 11 in the absence of block of the Trk pathway (C, green curve), or in the presence of Trk inhibitor during either the first (D, green curve) or second pulse (E, green curve) of 5-HT. Also shown are the effects of only blocking the Trk pathway during the first (D, blue curve) or second (E, red curve) pulse of 5-HT. Black dashed curves are control simulations. The histograms in boxes in D and E denote the integral values of pERK, TGF-β, and TBL between 50 and 280 min, with bar colors matching the corresponding curves.

Figure 7

Simulated dynamics of kinases, pCREB1, pCREB2_p38, TBL and inducer levels after five pulses of 5-HT using the Standard protocol (A, black curves), or Enhanced protocol (A, red curves); or two pulses of 5-HT with ISI of 15 min (B, blue curves), 45 min (B, black curves) or 60 min (B, red curves). Arrows represents peak inducer levels.