The phosphatase SHP2 regulates the spacing effect for long-term memory induction

Abstract

A property of long-term memory (LTM) induction is the requirement for repeated training sessions spaced over time. This augmentation of memory formation with spaced resting intervals is called the spacing effect. We now show that in Drosophila, the duration of resting intervals required for inducing LTM is regulated by activity levels of the protein tyrosine phosphatase corkscrew (CSW). Overexpression of wild-type CSW in mushroom body neurons shortens the inter-trial interval required for LTM induction, whereas overexpression of constitutively active CSW proteins prolongs these resting intervals. These gain-of-function csw mutations are associated with a clinical condition of mental retardation. Biochemical analysis reveals that LTM-inducing training regimens generate repetitive waves of CSW-dependent MAPK activation, the length of which appears to define the duration of the resting interval. Constitutively active CSW proteins prolong the resting interval by altering the MAPK inactivation cycle. We thus provide insight into the molecular basis of the spacing effect.

Figure 1. Expression of Gain-of-Function (GOF) csw Mutants in Mushroom Body (MB) Neurons Impairs 24-Hour Memory

(A) Schematic representation of the SHP2 protein mapping point mutations detected in Noonan syndrome (top) and in leukemia (bottom) studied here. (B-C) Memory performance was determined in fly lines only carrying the transgene (UAS-csw;+) or expressing csw transgenes (OK107;UAS-csw). Transgene expression of csw mutant (from D61Y to N308D) and csw wild-type (WT) was targeted to the mushroom body using OK107-GAL4 (OK107;UAS-csw). (B) Immediate memory after a single trial was unaffected by the expression of the csw transgenes. (C) 24-hr memory after spaced training was reduced in all fly lines expressing mutants alleles (OK107;UAS-csw) compared with control (+/+ or OK107;+) flies, but not in lines only carrying those csw transgenes (UAS-csw;+). (D) 24-hr memory after spaced training in fly lines expressing a loss-of-function csw mutant (R465M) in the mushroom body neurons. Bars, mean ± SEM (n=8). Asterisks indicate p<0.05.

Figure 2. Acute Genetic or Pharmacological Interference of CSW Function Impairs 24-Hour Memory in Adult Fruit Flies

(A-D) Heat shock or temperature protocol to induce transgene expression in adult flies (top) and effect on 24-hr memory after spaced training (bottom). HS+ represents heat shock and HS− represents no heat shock. (A) Effect of acute expression of mutant and wild-type csw induced with a heat shock-GAL4 driver and a single heat shock. The memory performance in hsGAL4;UAS-D61Y after heat shock (HS+) was reduced compared with the same line without heat shock (HS−) or control (+/+, UAS-D61Y, UAS-WT, hsGAL4;UAS-WT) groups in both conditions (HS+ or HS−). (B) Effect of acute expression of csw-RNAi (hsGAL4;csw-RNAi) induced with three heat shocks (HS+). Memory was reduced in hsGAL4;csw-RNAi after heat shock (HS+) compared with the same line without heat shock (HS−) or control (+/+, UAS-csw-RNAi, UAS-ctrl-RNAi or hsGAL4;+) groups in both conditions (HS+ or HS−). (C and D) Effect on memory performance of inducible RNAi expression targeted to mushroom bodies using the GAL80^ts repressor of GAL4-mediated UAS-csw-RNAi expression, (C) at permissive temperature and (D) at restrictive temperature. Memory was reduced in the GAL80^ts;csw-RNAi;OK107 group at restrictive temperature compared with control (+/+, UAS-csw-RNAi, GAL80^ts;UAS-ctrl-RANi;OK107), but not at permissive temperature. (E) Effect of the SHP2 inhibitor, NSC-87877 (Inhib [+]) or vehicle (Inhib [−]) on 24-hr memory in flies expressing mutant or wild-type csw in mushroom body neurons. Memory was reduced in the control group (OK107;+) and in flies overexpressing the wild-type csw (WT) by the inhibitor (inhib [+]) compared with both groups fed with vehicle (Inhib [−]), but not in csw gof mutants. Bars, mean ± SEM (A, B and E n=8, and C and D n=5). Asterisks indicate p<0.05.

Figure 3. Altered LTM Formation via Genetic Manipulation of CSW

(A) Effect of the protein synthesis inhibitor, cycloheximide (CXM+), or vehicle (CXM−) on 24-hr memory after spaced training in flies expressing mutant or wild-type csw in the mushroom body neurons. Memory was reduced in the control group (OK107;+) and in flies overexpressing the wild-type csw (WT) by the inhibitor (CXM +) compared with both groups fed with vehicle (CXM −), but not in csw gof mutants. (B) 24-hr memory after spaced or massed training in control (OK107;+) and transgenic flies expressing mutant and wild-type csw in the mushroom body neurons. Spaced training promoted a higher performance in control (OK107;+) flies compared with the massed training in the same line or the mutant lines (D61Y and E76K). Note that massed training in flies overexpressing wild-type csw produced a similar memory score as spaced training. (C) 24-hr memory after one trial (1 T), or 5 and 10 trials of massed training (5 m and 10 m, respectively) in control (OK107;+) and flies overexpressing wild-type csw in mushroom body neurons (OK107;UAS-csw⁺), which showed memory enhancement after 10 m compared with all the control bars (OK107;+) or the same lines trained with 1 or 5 trials. The effect of the cold shock (D) or cycloheximide (CXM) (E) on 24-hr memory after 10 trials of massed training in flies overexpressing wild-type csw in mushroom body neurons. Memory enhancement was detected in OK107;UAS-csw⁺ in the absence of cold shock or CXM compared with the same line untreated or with the control (OK107;+) line with or without treatment. (F) 24-hr memory after 10 trials of massed training in flies expressing wild-type or mutant csw in alpha/beta neurons of the mushroom bodies using the c739-GAL4 driver (c739;UAS-csw) compared with control (c739;+). Memory enhancement was detected in flies expressing wild-type csw compared with the control group or mutant lines. (G) 24-hr memory after 10 trials of massed training compared between flies overexpressing wild-type and phosphatase-dead mutant (R465M) csw in the mushroom body neurons. * indicates WT vs. R465M, Student's t test. Bars, mean ± SEM (N=6). Asterisks indicate p<0.05.

Figure 4. Overexpression of the Normal csw⁺ Transgene In MB Neurons Shortens Resting Intervals in LTM Induction

(A) Schematic representation of the training protocols used in Panel B. These show the two initial trials of the protocols, which actually were composed of 10 trials. From top to bottom, first, symbols representing the conditioned stimuli (odors in blue or green) and the unconditioned stimuli (electric shock in red); and below training protocols. Spaced: training trials with a resting interval of 900 sec. Massed (standard): trials with a 45 sec of resting between trials. Massed (150 sec): trials separated by 150 sec of interval between electric shocks as in the standard massed protocol, but without control odor. Massed (45 sec): trials with the intervals between electric shocks reduced to 45 seconds. Odor control: control stimulation with odor. (B) 24-hr memory performance after 10 trials of training (schematized in Panel A) in flies overexpressing wild-type csw in the mushroom body neurons. * indicates memory reduction by shortening the inter-trial interval. Note: In these schematic representations, only the odor “OCT” is shown to be associated with the electric shock (US), whereas each performance index (PI) is the average of two experiments in parallel where the odor “OCT” and “MCH” were paired with electric shock. Bars, mean ± SEM (N=8). Asterisks indicate p<0.05.

Figure 5. Training Induced CSW-Dependent Transient MAPK Activity

Training-induced MAPK activation in heads of control and transgenic fruit flies overexpressing csw in the mushroom bodies. (A) Representative western blot showing activated MAPK (p-MAPK) and total MAPK (T-MAPK) after spaced training. (B) MAPK activation in non-transgenic flies after training (Trained) or just manipulated (Naive). * Indicates trained vs. naïve at 20 min. (C) MAPK activation in naive flies (left side) and at different times after spaced training (right side) in mutant csw transgenic (D61Y and E76K), wild-type csw transgenic (WT) or control (OK107;+) lines. * indicates differences between lines overexpressing csw compared with the control (OK107;+) at the same time point. For D, E and F, training paradigms are represented with a red arrow indicating the time of sampling (top) and MAPK activation (bottom) in naïve flies (left side) and at the end of the inter-trial interval after 1, 2, 3, 9 and at three time points after the 10th training trial. MAPK activation in the control line (OK107;+) during spaced training (inter-trial interval, 900 sec) (D), and during massed training (inter-trial interval, 45 sec) (E). (F) MAPK activation in fruit flies overexpressing wild-type csw (OK107;UAS-csw⁺) during massed training (inter-trial interval, 45 sec). For D, E and F, * indicates an increase in MAPK activation compared with the basal state (naïve). Data points, mean + SEM (n=3 in B) and (n= 4 in C, D, E and F). (G-K) Effect of a trial of training on training-induced MAPK activation produced by a previous trial of training. (G) Protocols of training represented by a line with a first trial of training and a second trial (2^nd trial). Black arrows indicate the time at which the flies were harvested (45 sec or 930 sec) and processed for western blot. (H-K) MAPK activation in naïve (violet), at 45 sec (green) or 930’ sec (blue) after one trial; and at 930 sec after one trial plus a test trial of 30 sec (red) in the control group (OK107;+) (H); in fruit flies overexpressing wild-type CSW (OK107;UAS-csw⁺) (I); and flies overexpressing GOF mutants CSW D61Y (OK107;UAS-csw^D61Y) (J) or E76K (OK107;UAS-csw^E76K) (K). * indicates increased levels of MAPK compared with the basal state (naïve) and ** a statistical reduction of p-MAPK at 930 sec after the test trial compared with the same time point without a test trial (blue vs. red bar). Of note, in H and I, the level of MAPK activation after the test trial (red bar) is significantly smaller than the levels at the time of the peak of MAPK activation (blue and green bars, respectively; indicated by ***) and is not different from the naïve state. Data points, mean ± SEM (n=3). Asterisks indicate p<0.05.

Figure 6. Rescue of the GOF Mutations-Induced LTM Phenotype

(A) Dose response to the SHP2 inhibitor on 24-hr memory in non-transgenic flies (2022U). Flies were fed vehicle (0 μM) or varying doses of drug (25, 50, 75 and 100 μM) for eight hours before training, or for eight hours after training (100 μM). * indicates reduced memory performance compared with flies fed with vehicle (0 μM). (B) Effects of eight hours of SHP2 inhibitor 50 μM (inhib [+]) or vehicle (inhib [−]) before training on 24-hr memory in control (OK107;+) and flies overexpressing GOF or wild-type csw. * indicates memory rescue in mutants (D61Y and E76K) after drug feeding (inhib [+]) compared with the same line after vehicle (inhib [−]). (C) MAPK activation in non-transgenic flies treated with SHP2 inhibitor (100 μM) (Inhib [+]) or vehicle (Inhib [−]) after spaced training. * Indicates (Inhib [+]) vs. (Inhib [−]) at 20 min. Bars and data points, mean ± SEM (N=6 in A and B) and (n=3 in C). (D and E) 24-hr memory after 10 trials of spaced training with increasing resting intervals in controls (OK107;+) and flies overexpressing gain-of-function mutant (D61Y and E76K) or wild-type (WT) (D), or expressing a csw mutant with weaker gain-of function effect (I282V) (E). In D and E * indicates higher performance compared with the same fly line when the resting interval of 15 min was used. Bars, means ± SEM (N=8). Asterisks indicate p<0.05.

Figure 7. Schematic Representations of Training-Regulated MAPK Activity Correlated with Training Protocol and Genotype

Arrows indicate individual training trails. The curves are schematic representation of MAPK activity (the vertical for amplitude of MAPK activity and horizontal for time). The duration for the resting intervals for each training paradigm is indicated.