(C) PLOS One
This story was originally published by PLOS One and is unaltered.
 . . . . . . . . . .



Aberrant dopaminergic activity during consolidation causes age-related memory generalization in Drosophila [1]

['Motomi Matsuno', 'Metropolitan Institute Of Medical Science', 'Learning', 'Memory Project', 'Setagaya', 'Tokyo', 'Nozomi Uemura', 'Department Of Computational Biology', 'Medical Sciences', 'Graduate School Of Frontier Sciences']

Date: 2026-04 

Overactivation of memory networks and pathways can induce post-traumatic stress disorders and memory generalization, where memories are recalled in inappropriate situations. Here, we demonstrate that age-related defects in long-term memories in Drosophila are also caused by memory generalization. Aversive memory engram cells are formed in both young and old flies trained in an odor avoidance task. However, while engrams in young flies are activated specifically by odors previously paired with electrical shocks, engrams in old flies are activated by shock-paired, unpaired, and novel odors. This enhancement of engram cell activation occurs because of increased activity of dopaminergic neurons during memory consolidation in old flies. Increased dopamine signaling results from an inability of old flies to inhibit glutamatergic activation and leads to increased activation of dopamine D2 receptors on engram cells. Our data suggest that increased dopaminergic activity after training generalizes the responsiveness of engram cells to disrupt appropriate memory recall.

Funding: This work was supported by Japan Society for Promotion of Science (JSPS) KAKENHI Grants (20K06753 to MM; 19H01013 and 21K18238 to MS; and 21K06403 and 18K06496 to JH). JSPS: https://www.jsps.go.jp/english/e-grants/grants01.html . The funder had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Although the formation and activity of neuronal engram cells is critical for LTM, the relationship between engram cells, glial transcription, aging, and memory generalization has been unclear. Here, we determine that altered glial regulation of glutamate signaling at old ages increases activity of dopaminergic neurons that regulate engram cell connectivity. This results in memory generalization. We further find that inhibiting dopamine D2 receptor activity prevents memory generalization.

Besides neuronal transcription, LTM also requires increases in glial transcription [ 15 ]. Spaced training activates the glial transcription factor, Repo, which increases the expression of the glutamate transporter, Eaat1 [ 16 ]. Increases in EAAT1 reduce synaptic glutamate by importing glutamate into glia. Eaat1 expression increases shortly after spaced training, indicating that glutamate activity is normally inhibited during memory consolidation. The homophilic cell adhesion molecule, Klingon (Klg), is expressed in both neurons and glia and is necessary for spaced training-dependent activation of Repo, indicating that Klg plays an important role in transmitting neuronal activity signals to glia [ 17 ]. Interestingly, aging causes decreased expression of Klg, Repo, and Eaat1, suggesting that defects in glial transcription and repression of glutamate signaling may be a cause of age-related defects in LTM. Supporting this idea, increased expression of Klg, Repo, or Eaat1 rescues age-related LTM defects [ 16 ].

In flies, age-related memory impairment (AMI) affects specific types of memory, including middle-term memory (MTM) and long-term memory (LTM) [ 2 , 7 – 9 ]. LTM, which lasts for over 7 days, requires increased protein synthesis and activation of transcription factors, including CREB and c-Fos, in specific neurons known as engram cells [ 10 – 12 ]. LTM engram cells in Drosophila can be identified as neurons that show training-dependent increases in Kayak, the fly homolog of c-Fos [ 12 ]. Engram cells are formed by multiple associative training trials with interspersed rest intervals (spaced training) and are reactivated during memory recall. Furthermore, the importance of c-Fos + engram cells in flies has been demonstrated by showing that their inactivation prevents memory recall [ 12 ]. While some studies have examined engram cells in old animals [ 13 , 14 ], the relationship between memory generalization and engram cell reactivation has been unclear.

While animals need accurate memories, they also need to be able to generalize memories in order to respond properly to novel experiences that are similar but not identical to prior experiences [ 1 ]. Thus, a balance between memory specificity and generalization is required, with a shift in balance in either direction resulting in memory and behavioral impairments. Memory decreases with advancing age in many animals, including Drosophila [ 2 ]. However, it has been unclear whether altered memory generalization is a major cause of age-related memory impairments. Increases in generalization have been implicated in diseases associated with memory dysfunction, including post-traumatic stress disorder (PTSD), in which memories of past traumas can be triggered by inappropriate cues, severe depression, and anxiety [ 3 – 6 ].

Results

Memory engrams are formed in aged flies, but are reactivated by inappropriate stimuli To determine whether age-related impairments in LTM are caused by a reduction in the formation of memory engram cells, we first examined engram cells in young (raised for 3–5 days at 18 °C) and aged flies (raised for 40 days at 18 °C) after spaced training in an odor/fear association paradigm. As previously described [12], we identified potential engram cells as GFP-positive cells located dorsally to the mushroom body calyces 24 hours after spaced training of UAS-mCD8-GFP; kayak-Gal4/Tub p -Gal80ts flies (Fig 1A–1C). In these flies, GFP is expressed under control of a kayak (c-Fos) promoter. Flies were raised and aged at 18 °C to allow GAL80ts to inhibit background GFP expression, and trained at the Gal80ts restrictive temperature of 32 °C (Fig 1A). As seen in Fig 1C and 1D, we observed a significant increase in c-Fos-positive neurons after spaced training in both young and aged flies, suggesting that engram cells are formed to a similar extent in old flies. When we examined the reactivation of engram cells during recall using anti-pERK antibodies to detect neuronal activation upon odor exposure, we found that exposure to the shock-paired odor (CS+) reactivated a significantly larger percentage of potential engram cells compared to exposure to an unpaired odor (CS−) in young flies (Fig 1C and 1E). However, in aged flies, both the CS+ and CS− odors reactivated engram cells similarly (Fig 1C and 1E), suggesting that fear engrams are reactivated by both appropriate and inappropriate odors in aged flies. Thus, one defect of LTM in aged flies may be a recall of memories in inappropriate situations. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 1. Functional memory engrams are formed in aged flies but are activated by inappropriate odors. A) mCD8-GFP was expressed under control of the Drosophila c-Fos (kayak) promoter to visualize putative engram cells. GFP signals from background kayak expression were inhibited using GAL80ts and raising flies at permissive temperature (18 °C). 10× spaced training was performed at 32 °C to release GAL80ts repression and allow training-dependent kayak to induce GFP expression. B) Putative engram cells were identified as GFP (cFos) signals in the indicated mushroom body cell body region dorsal to the calyx 24 hours after spaced training of UAS-mCD8-GFP; kayak-Gal4/Tubp-Gal80ts flies. pERK signals observed after odor exposure 24 hours after training were used to identify cells activated during memory recall tests. C) Putative engram cells (mCD8-GFP signals) were observed after spaced training in young and aged flies (compare panels a and j to panels d, m, g, and p). pERK signals were observed after exposing flies to odors during memory recall tests 24 hours after training. In young flies, pERK signals after CS+ exposure overlapped significantly with GFP signals when compared to signals after CS− exposure. In aged flies, the overlap between pERK and GFP signals was similar after CS+ and CS− exposure. D) Quantification of putative engram cells in naïve and trained, young and aged flies. N = 7–18, *** indicates P < 0.001. E) Quantification of engram cell reactivation data. N = 6–10, *** indicates P < 0.001. F) Schematic for optogenetic reactivation of engram cells. Channelrhodopsin-YFP (ChR2-YFP) was expressed in putative engram cells using the same strategy of expressing GFP shown in (A). Twenty-four hours after spaced training, flies were tested for their preference of blue light (which activates channelrhodopsin) vs. green light. G) Color preferences of indicated flies. Positive color preference scores indicate that flies preferentially chose blue light over green light, while negative scores indicate that flies preferentially chose green light. N = 11–14, **** indicates P < 0.0001. Data in all bar graphs in this manuscript represent means ± SEMs. The data underlying this figure are available in S1 Data. https://doi.org/10.1371/journal.pbio.3003752.g001 To verify that the c-Fos-positive cells that we identified in aged flies encode functional aversive memory engrams, we next examined whether artificial activation of these cells induces avoidance behaviors. As described previously [12], we expressed blue-light inducible channelrhodopsin in putative engram cells by training UAS-ChR2::YFP/+; Kayak-Gal4/Tub p -Gal80ts flies at GAL80ts restrictive temperature. This allowed us to activate kayak (c-Fos)-positive engram cells by exposing flies to blue light (Fig 1F). Young untrained flies and young flies raised in the absence of all-trans retinal (ATR), which is required for producing functional channelrhodopsin, showed a strong preference for blue light when given a choice between blue and green light (Fig 1G). However, young flies fed ATR and trained in an odor avoidance task at the GAL80ts restrictive temperature lost this preference and instead showed a slight preference for green light (Fig 1G). This suggests that activation of putative aversive memory engram cells by blue light produces avoidance behaviors that are sufficient to eclipse natural light preferences. Similar to results in young flies, we found that aged trained and untrained flies raised in the absence of ATR showed a preference for blue light, while aged trained flies fed ATR showed a preference for green light. Thus, training flies in an aversive association task produces engram cells in both young and aged flies, and subsequent activation of these engram cells induces appropriate memory-associated avoidance behaviors.

Inappropriate reactivation of engrams in aged flies is associated with memory generalization Our results suggest that a major cause of LTM defects in aged flies may be that aversive memories are recalled in inappropriate situations. In standard aversive odor testing experiments, flies are given a choice between two odors, the shock-paired (CS+) odor, and a non-shock-paired (CS−) odor. In young flies, aversive memory engrams are reactivated by the CS+, causing flies to avoid the CS+ and choose the CS− when given a choice between the two odors. However, if aversive engrams are reactivated by both CS+ and CS− odors in aged flies, flies should try to avoid both odors, leading to a reduction in behavioral scores. To determine whether aged flies show increased avoidance of both appropriate and inappropriate odors after training, we examined whether aged flies displayed an increased avoidance towards the CS+, CS−, and a novel odor after training when given a choice between the testing odor and air instead of a control odor. As seen in the left sections of Figs 2A and S1A, 24 hours after spaced training (paired training), young flies show a strong avoidance of the shock-paired odor, an attraction to the unpaired (CS−) odor [18,19], and a neutral response to novel odors. Control training protocols in which flies were mock trained by exposing them to odors in the absence of electrical shocks (odor-alone training) or where odors and shocks were unpaired resulted in neutral responses. Similar to young flies, old flies showed a strong avoidance of the CS+ odor after paired training. However, instead of an attraction to the CS− odor and a neutral response to a novel odor, they showed an avoidance of these odors (Figs 2A and S1A, right sections). Thus, compared to mock or control training methods, old flies subjected to paired training showed a significant increase in avoidance of both a novel odor and the CS+ odor, and a shift from attraction to avoidance of the CS− odor, when compared to young paired-trained flies. Altogether our results suggest that reactivation of aversive LTM engram cells by both appropriate and inappropriate odors in aged flies is accompanied by increased avoidance of these odors. This indicates that one component of age-related impairments in LTM is memory generalization, where aged flies lose the specificity of a memory. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 2. Long-term memory impairments in aged flies are associated with memory generalization and non-specific activation of engram cells. A) Avoidance of the CS+, CS−, and a novel odor, IAA, 24 hours after spaced training in young and aged flies. Positive avoidance scores indicate that more flies chose air over the odor while negative avoidance scores indicate that flies chose the odor over air. Spaced training (paired training) induced significant avoidance of the CS+ odor, approach toward the CS− odor, and no change in the response to a novel odor in young flies (N = 6–15). Control training included training in the absence of shocks (odor only, N = 6–8) and training where shocks were unpaired with either odor (unpaired training, N = 8–10). In aged animals, paired training resulted in a significant avoidance of the CS+, but also a significant avoidance of the novel odor compared to odor only and unpaired controls. Aged animals also had a tendency to avoid rather than approach the CS−. *, **, ***, and **** indicate P < 0.05, 0.01, 0.001, and 0.0001. B) Acute inhibition of Eaat1 induces memory generalization behaviors in young flies. Twenty-four hours after spaced training, control flies (−RU UAS-eaat1 RNAi/+, +RU UAS-eaat1 RNAi/+, and −RU UAS-eaat1 RNAi/+; glialGS/+) avoided the CS+ odor, but not the CS− odor. Inhibition of Eaat1 expression (+RU UAS-eaat1 RNAi/+; glialGS/+) caused flies to avoid both the CS+ and CS− odors. N = 6 for each condition. C, D) Memantine (C) or MK801 (D) feeding after spaced training reduces avoidance of the CS− odor and memory generalization in aged flies. N = 6–7. E, F) Feeding aged flies memantine (E) or MK801 (F) after spaced training reduces reactivation of engram cells by the CS− odor, resulting in significant differences between CS+ and CS− dependent engram reactivation. N = 5–6. The data underlying this figure are available in S1 Data. https://doi.org/10.1371/journal.pbio.3003752.g002 Flies were aged for 40 days at 18 °C to generate old flies for engram imaging experiments in order to inhibit background, non-training-dependent c-Fos expression, while they were aged for 23–25 days at 25 °C for behavioral experiments. To confirm that both methods of aging induce generalization, we verified that memory generalization occurs in flies aged at 18 °C similarly to those aged at 25 °C (S1B Fig). Paired training consists of exposing flies to two odors, the CS + , which is paired temporally with electrical shocks, and the CS−, which is presented 45 s after shock cessation. While the CS− may have previously been thought of as a control odor to the CS + , more current studies have determined that the CS− can become associated with learned safety or relief from electric shocks and becomes attractive to flies [18,19] (see also CS− effects in Figs 2A and S1A). The engram cells we identified can be classified as aversive since they are preferentially reactivated by the CS+ odor in young flies and induce aversive behaviors when activated. However, to remove possible confounding effects of CS− memory, we also performed CS+ training in which the CS + was paired to electrical shocks and no CS− was presented (S1C and S1D Fig). CS+ training produced similar numbers of GFP+ engram cells as paired training and these engram cells were reactivated preferentially by the CS+ odor compared to a novel odor in young flies, while they were reactivated similarly by CS+ and the novel odor in old flies. Furthermore, as a control, we performed mock odor-only training. In this situation, engram cells were formed, likely because of the altered context during training [20], but these engram cells were reactivated by odors at significantly lower numbers compared to paired and CS+ training in both young and old flies. Overall, our results indicate that odor-specific engram cells are formed by both paired and CS+ training in young flies, but these engram cells lose their odor specificity upon aging resulting in memory generalization. One possible explanation for apparent memory generalization in old flies is if spaced-trained old flies are unable to distinguish between different odors. However, we found that old flies are able to form short-term memories, which require the ability to distinguish odors, 24 hours after spaced-training (S2A Fig). We further found that a second type of consolidated memory, anesthesia-resistant memory (ARM), which does not require c-Fos positive engram cells, does not induce memory generalization in old flies (S2B Fig). In addition, generalization occurs after spaced training of old radish (rsh) mutants, which are specifically impaired for ARM (S2C Fig). Thus, memory generalization seems to be a specific phenomenon affecting the reactivation of LTM engram cells by multiple distinguishable odors.

Inhibition of glutamate activity during memory consolidation prevents memory generalization and inappropriate reactivation of engrams How does aging cause an increase in memory generalization? LTM formation requires an increase in amounts of the glial glutamate transporter, EAAT1, which functions to clear synaptic glutamate during memory consolidation [16]. Aged flies are unable to increase Eaat1 expression after training, while artificial expression of Eaat1 in glia is able to rescue LTM defects in aged flies [16]. This suggests that inappropriate glutamate activity during memory consolidation may induce generalization, leading to reduced LTM. To test this idea, we examined whether artificial inhibition of Eaat1 induces memory generalization in young flies. We used a glial geneswitch (GS) driver to acutely knock down Eaat1 expression in glia in young adult flies and observed memory generalization, similar to what we observe in aged wild-type flies (Fig 2B). In a complementary approach, we also inhibited NMDA-type glutamate receptors after spaced training in aged flies using the pharmacological agents, memantine and MK801, and observed an increase in memory specificity (Fig 2C and 2D). When we examined odor-dependent reactivation of engram cells in aged flies, we found that memantine and MK801 feeding after spaced training reduced activation of engram cells by the CS− odor, restoring the preferential activation of engram cells by the CS+ odor (Fig 2E and 2F). These results suggest that activation of NMDA receptors by glutamate during consolidation causes engram cells to be reactivated by inappropriate odors, leading to memory generalization and LTM defects in aged flies.

Activation of PPL1 dopaminergic neurons by glutamate during consolidation inhibits LTM Where does glutamate signaling need to be inhibited to prevent memory generalization? Upregulation of Eaat1 in glial cells after spaced training requires neuron/glia interactions mediated by the homophilic cell adhesion molecule, Klg [16]. Thus, to identify candidate neurons and glia involved in LTM and memory generalization, we knocked down klg in various cell types and searched for reductions in LTM (Fig 3A). We found that knocking down klg in the ellipsoid bodies and mushroom bodies had no significant effects on memory. In contrast, knocking down klg in astrocyte-glia, glutamatergic neurons, and dopaminergic neurons significantly reduced LTM. This suggested that EAAT1 needs to inhibit glutamate signaling between presynaptic glutamate neurons and postsynaptic dopaminergic neurons during LTM consolidation to prevent memory generalization. We further found that knocking down klg specifically in the PPL1 subset of dopaminergic neurons using the MB060B driver inhibited LTM, while knocking down klg in the PAM cluster dopaminergic neurons did not (Fig 3A). PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 3. Suppression of Klg in astrocytes, dopaminergic neurons, and glutamatergic neurons inhibits long-term memory. A) klg was knocked down (klgRNAi) in indicated cell types using indicated drivers. Drivers/cell types where klg knockdown significantly inhibited LTM are shown in blue. N = 5–19. *, **, ***, and **** indicate P < 0.05, 0.01, 0.001, and 0.0001. B) Representative images of glutamate sensor (iGlusnFR) fluorescence in young and old flies under different conditions. The glutamate sensor, iGlusnFR, was expressed in PPL1 neurons from the MB060B driver and ROIs (green and red circles) were chosen at the α2α′2 compartments of the mushroom bodies. In young flies, glutamate signals were reduced 3 hours after spaced training compared to naïve animals. This reduction was not observed when young flies were fed the EAAT1 blocker, UCPH-101. Reductions in glutamate concentrations 3 hours after spaced training were not observed in aged flies. C) Quantification of the data shown in B). N = 20–42 for young flies, and N = 15 for old flies. D) Pharmacological inhibition of EAAT1 activity impairs LTM and induces memory generalization in young flies. Young flies were fed UCPH-101 after spaced training and tested for LTM (right panel) or tested for avoidance of the CS+ and CS− odors at 24 hours (left panel). In the left panel, no significant differences in avoidance of the CS+ and CS− odors were found in flies fed UCPH-101. Right panel, N = 8, Left panel, N = 5–6. The data underlying this figure are available in S1 Data. https://doi.org/10.1371/journal.pbio.3003752.g003 The MB060B driver labels a subset of PPL1 neurons that are proposed to regulate consolidation and forgetting of LTM [21,22]. To verify that glutamate input to these neurons is inhibited during memory consolidation, we expressed the fluorescent glutamate sensor, iGlusnFR, in these neurons to measure extracellular glutamate after spaced training [23,24]. Compared to naïve flies, we measured a significant reduction in glutamate 3 hours after spaced training in young flies (Fig 3B and 3C), which correlates with an increase in Eaat1 expression after training (S3A Fig). This result is consistent with previous studies [16] and suggests that increased Eaat1 expression during consolidation decreases glutamatergic activation of PPL1 neurons. When we fed young flies a glutamate transporter blocker, UCPH-101, after training, we found that the reduction in glutamate at 3 hours was abolished, LTM was impaired, and avoidance of the CS− odor was increased (Fig 3B–3D). This indicates that glutamate transporter activity is required to reduce glutamate during consolidation, and a failure to do this results in LTM impairment and memory generalization. In contrast to young flies, we did not observe a decrease in extracellular glutamate after spaced training in aged flies. This result is consistent with the inability of aged flies to increase EAAT1 expression after spaced training [16] and prevent memory generalization.

Glutamate activation induces low-frequency activity in PPL1 neurons To next determine how reduced glutamate signaling affects PPL1 dopaminergic neuronal activity during memory consolidation, we examined the activity of PPL1 neurons using a bioluminescence-based neural activity reporter. Longitudinals lacking (Lola) is a transcription factor whose expression is regulated by neuronal activity, and Lola promoter luciferase constructs have been used to measure neuronal activity [25–27]. We used flies expressing Lola p -luciferase in PPL1 neurons to measure the activity of these neurons after spaced training. In young flies, PPL1 activity after training was similar to activity in naïve and unpaired control flies (Fig 4A and 4B). In contrast, in aged flies, PPL1 activity was significantly increased between 3 and 6 hours after spaced training (Fig 4A and 4B). Administration of riluzole, a glutamate receptor inhibitor [28], or memantine, an NMDA receptor antagonist (Fig 4C), abolished this increase, suggesting that it is caused by excess glutamate activity. Altogether, our data are consistent with a model where glutamate activity needs to be reduced during consolidation to prevent memory generalization. In young flies, this is accomplished by increasing glial EAAT1 expression. However, aged flies are unable to increase EAAT expression resulting in increased PPL1 activity during consolidation. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 4. Activity of PPL1 dopaminergic neurons increases after spaced training in aged flies. A) Activity of PPL1 DANs was measured in young and aged, naïve and trained MB060B>FLP, Lola(FRT)stop(FRT)Luc flies (N = 18 for each condition). No significant differences were observed in luciferase activity between young and aged naïve flies or between young naïve and young trained flies. Asterisks indicate significant differences between aged trained vs. young trained flies. *, **, ***, and **** indicate P < 0.05, 0.01, 0.001, and 0.0001. B) Luminescent activity in indicated flies 3–6 hours after spaced training. Activity was normalized to that of young naïve flies. N = 9–18. C) Effects of riluzole and memantine feeding on PPL1-DAN activity. N = 17–22 for riluzole experiments and N = 11–14 for memantine experiments. D) Power spectrum in PPL1 neurons of young and aged, naïve and spaced trained flies detected using GCaMP. E) Effects of the NMDAR inhibitor, MK801, and the EAAT1 inhibitor, UCPH-101 on power at frequencies <0.05 Hz. N = 9–14. The inset is a representative image of the α2α′2 compartment region of the mushroom bodies and ROIs where GCaMP fluorescence was measured in PPL1 neurons. The data underlying this figure are available in S1 Data. https://doi.org/10.1371/journal.pbio.3003752.g004 To verify that spaced training increases PPL1 activity after training in aged flies, we next examined PPL1 activity using calcium imaging. We observed a significant increase in Ca2+ oscillations at frequencies below 0.05 Hz 3 hours after spaced training in aged flies that was not present in young flies (Fig 4D and 4E). Feeding aged flies the NMDA receptor antagonist, MK801, reduced PPL1 activity to identical levels as young flies (Fig 4E), suggesting that increased glutamate signaling to NMDA receptors is responsible for increased PPL1 activity in aged flies. Furthermore, feeding the Glu transporter blocker UCPH-101 after training caused an increase in Ca2+ oscillations below 0.05 Hz in young flies, suggesting that impaired Glu uptake after training is responsible for the increase in Ca2+ oscillations in aged flies. UCPH-101 feeding might also be expected to increase Ca2+ oscillations in old flies, but instead we observed a decrease, possibly indicating that further increases in glutamate activity above the increase already found in old flies may induce glutamate toxicity [29]. Neuronal activity increases phosphorylation of ribosomal protein S6, and phospho-S6 can be used to identify active neurons [30]. We examined phosphorylated S6 in TH+ neurons and found that spaced training increases phospho-S6 signals in aged, but not young, flies (S3B and S3E Fig). This increase does not occur after massed training (S3C Fig), supporting the idea that spaced training specifically increases dopaminergic activity in aged flies. Ectopic expression of the glutamate transporter, eaat1, in glia reduced phospho-S6 signals in dopaminergic neurons after spaced training in aged flies (S3D Fig), suggesting that reductions in Eaat1 are responsible for increased dopaminergic activity and impaired LTM in aged flies.

NMDAR-dependent activation of PPL1 neurons regulates memory generalization Our results indicate that aberrant activation of PPL1 neurons during consolidation causes memory generalization in aged flies. If this is the case, artificial activation of PPL1 neurons in young flies should induce generalization, while inhibiting these neurons in aged flies should suppress generalization. Thus, we examined whether altering PPL1 activity after spaced training using the DREADD system [31,32] affected LTM and generalization. When we activated the inhibitory DREADD, hM4Di, in PPL1 neurons by feeding MB060B>hM4Di flies CNO after training, we observed no change in LTM or generalization in young flies, while we observed a restoration of LTM (Fig 5A, left panel) and a reduction of generalization in aged flies (Fig 5A, right panel). Conversely, when we activated the stimulatory hM1Dq DREADD in PPL1 neurons after training, we observed a significant decrease in LTM and an increase in generalization in young flies, but no change in memory or generalization in aged flies (Fig 5B). Thus, LTM impairment and increased generalization can be caused by aberrant activation of PPL1 neurons in young flies, while these defects can be rescued by inhibition of PPL1 in aged flies. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 5. PPL1 activity regulates memory generalization. A) Effects of inhibition of PPL1 neurons on LTM scores (left panel) and generalization (right panel) in young and aged flies. Inhibitory DREADDs were activated after spaced training by feeding flies CNO. N = 5–7. *, **, ***, and **** indicate P < 0.05, 0.01, 0.001, and 0.0001. B) Effects of stimulation of PPL1 neurons on LTM scores (left panel) and generalization (right panel) in young and aged flies. Stimulatory DREADDs were activated as described in A). N = 5–10. C, D) Knocking down NR1 in PPL1 neurons rescues LTM defects (C) and suppresses memory generalization (D) in aged flies. N = 8–14 for LTM experiments, and N = 5–7 for generalization experiments. E, F) Overexpression of Mg2+ block-deficient NR1 in PPL1 neurons inhibits LTM (E) and induces memory generalization (F) in young flies. N = 5–10 for each condition. The data underlying this figure are available in S1 Data. https://doi.org/10.1371/journal.pbio.3003752.g005 To verify that glutamate regulates PPL1 activity through NMDARs, we next manipulated NMDARs on PPL1 neurons and examined effects on LTM and generalization. We found that decreasing NMDAR expression using RNAi in PPL1 neurons rescued LTM defects and reduced generalization in aged flies without affecting memory or generalization in young flies (Figs 5C, 5D, and S4A). Notably, the converse experiment, overexpression of wild-type NMDARs did not inhibit LTM or induce generalization in young flies. Activation of NMDARs requires two steps, (1) binding of extracellular glutamate and (2) depolarization to remove Mg2+ ions that block channel conductance, suggesting that overexpression of NMDARs on its own is insufficient to increase glutamate-dependent activation of PPL1 neurons. When we overexpressed an NMDAR Mg2+ block mutant, which can be activated by glutamate in the absence of depolarization (NR1 N631Q), we observed a reduction in LTM and an increase in generalization in young flies (Figs 5E, 5F, and S4A). This suggests that age-related impairments in LTM occur through aberrant activation of NMDARs through a combination of two mechanisms, increased glutamate signaling and removal of Mg2+ block. One possible mechanism for removing Mg2+ block at old ages would be an age-dependent reduction in extracellular Mg2+ concentration. However, when we measured the extracellular concentration of various cations at young and old age, we did not observe any decrease in Mg2+ upon aging (S4B Fig). Thus, a different mechanism, such as increased depolarization of PPL1 neurons, occurs upon aging to contribute to memory reductions and increased generalization.

Knockdown of Dop2 receptors in engram cells reduces generalization in aged flies Altogether, our data suggest that dopamine released from PPL1 neurons during memory consolidation alters plasticity such that memory engram cells are activated by a variety of odor-responsive neurons. To next determine whether dopamine may be acting directly on engram cells to alter their activity, we knocked down dopamine receptor subtypes (Dop1R1, Dop1R2, Dop2R, and DopEcR) in kayak-positive engram cells at Gal80ts restrictive temperatures using tub-Gal80ts/UAS-Dop1R1RNAi;kayak-Gal4/+, tub-Gal80ts/+;kayak-Gal4/UAS-Dop1R2RNAi, tub-Gal80ts/+;kayak-Gal4/UAS-Dop2RRNAi and tub-Gal80ts/+;kayak-Gal4/UAS-DopEcRRNAi flies. We raised these flies at 18 ℃ where Gal80ts prevents knockdown, and then shifted them to 30 ℃ to knock down dopamine receptors during spaced training and memory consolidation. Knocking down Dop1R1, Dop1R2, and Dop2R receptors inhibited long-term memory in young flies, while knocking down DopEcR had no significant effects (left panels of Figs 6A, 6B, S5A, and S5B). Knocking down Dop1R1, Dop1R2, and DopEcR did not affect the already impaired LTM of old flies, while knocking down Dop2R improved LTM (left panels of Figs 6A, 6B, S5A, and S5B). Furthermore, while knocking down the other DopRs had no effects on generalization in either young or old flies, knocking down Dop2R prevented memory generalization in old flies without affecting memory specificity in young flies (right panels of Figs 6A, 6B, S5A, and S5B). These data are consistent with a model where dopamine release from PPL1 neurons acts on Dop2Rs on engram cells to induce memory generalization. Also consistent with this model, we found that feeding flies the Dop2R antagonist sulpiride during memory consolidation also prevented memory generalization in old flies (Fig 6C). PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 6. Inhibition of Dop2R reduces memory generalization. A) Left panel, LTM scores of young and old tub-GAL80ts/+; kayak-GAL4/UAS-Dop2RRNAi flies trained at GAL80ts permissive (-hs) or restrictive (+hs) temperatures. Temperature shifts for +hs flies are described in Fig 1A. Knocking down Dop2R improves LTM, specifically in old flies. N = 6–13. Right panel, avoidance of the CS+ and CS− odors in these flies. Knocking down Dop2R inhibits memory generalization in old flies. N = 5. *, **, ***, and **** indicate P < 0.05, 0.01, 0.001, and 0.0001. B) Left panel, LTM scores of young and old tub-GAL80ts/UAS-Dop1R1RNAi;kayak-GAL4/+ flies trained at GAL80ts permissive (−hs) or restrictive (+hs) temperatures. Knocking down Dop1R1 does not affect LTM in old flies. N = 6–14. Right panel, avoidance of the CS+ and CS− odors in these flies. Knocking down Dop1R1 does not affect memory generalization in old flies. N = 5. C) Sulpiride feeding restores memory specificity in old flies. 5 μM sulpiride or vehicle was fed to flies from the end of training until testing, and avoidance of the CS+ and CS− odors was measured. N = 5. The data underlying this figure are available in S1 Data. https://doi.org/10.1371/journal.pbio.3003752.g006

[END]
---
[1] Url: https://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.3003752

Published and (C) by PLOS One
Content appears here under this condition or license: Creative Commons - Attribution BY 4.0.

via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/