Presynaptic Rac1 in the hippocampus selectively regulates working memory

Abstract

One of the most extensively studied members of the Ras superfamily of small GTPases, Rac1 is an intracellular signal transducer that remodels actin and phosphorylation signaling networks. Previous studies have shown that Rac1-mediated signaling is associated with hippocampal-dependent working memory and longer-term forms of learning and memory and that Rac1 can modulate forms of both pre- and postsynaptic plasticity. How these different cognitive functions and forms of plasticity mediated by Rac1 are linked, however, is unclear. Here, we show that spatial working memory is selectively impaired following the expression of a genetically encoded Rac1-inhibitor at presynaptic terminals, while longer-term cognitive processes are affected by Rac1 inhibition at postsynaptic sites. To investigate the regulatory mechanisms of this presynaptic process, we leveraged new advances in mass spectrometry to identify the proteomic and post-translational landscape of presynaptic Rac1 signaling. We identified serine/threonine kinases and phosphorylated cytoskeletal signaling and synaptic vesicle proteins enriched with active Rac1. The phosphorylated sites in these proteins are at positions likely to have regulatory effects on synaptic vesicles. Consistent with this, we also report changes in the distribution and morphology of synaptic vesicles and in postsynaptic ultrastructure following presynaptic Rac1 inhibition. Overall, this study reveals a previously unrecognized presynaptic role of Rac1 signaling in cognitive processes and provides insights into its potential regulatory mechanisms.

Figure 1.

Localization of the presynaptic Rac1 inhibitor construct. (A) Illustration of the presynaptic Rac1 inhibitor construct. (B) C57BL/6J primary neuron cultures expressing GFP, the negative control construct, or the presynaptic Rac1 inhibitor construct. (C) Fluorescence intensity profiles validated presynaptic localization of the constructs. (D) Timeline of AAV injection (n = 16 mice per construct) and behavioral experiments. (E) Representative immunohistochemistry images of brain slices expressing W56-GFP-Syn1a. Dotted lines indicate the outlines of the brain slices. Data are expressed as mean ± SEM with ns. not significant, ** p < 0.01 and *** p < 0.001.

Figure 2.

Hippocampal presynaptic Rac1 inhibition impairs spatial working memory. (A) Schematic of the eight-arm radial-arm maze test. (B) The total number of arm entries (Scr: 12.86 ± 1.123, W56: 14.29 ± 1.051, P = 0.3615) and (C) latency to the first arm entry (Scr: 23.93 ± 4.877 s, W56: 26.07 ± 8.106 s, P = 0.5020) were not significantly different, whereas (D) the number of arm entries before the first error (Scr: 5.286 ± 0.5074, W56: 3.714 ± 0.3544, P = 0.0174) and (E) latency to the first error (Scr: 115.3 ± 17.11 s, W56: 68.36 ± 11.01 s, P = 0.0248) were lower in the W56 group. (F) Schematic of the spontaneous alternation Y-maze test. (G) The total number of entries (Scr: 29.63 ± 2.185, W56: 38.20 ± 2.599, P = 0.0168) was higher in the W56 group, while (H) latency to the first arm entry (Scr: 2.875 ± 0.5719 s, W56: 2.811 ± 0.4422 s, P = 0.7039) was similar. (I) The percentage of alternations (Scr: 55.58 ± 3.34%, W56: 47.51 ± 2.12%, P = 0.030) was significantly lower in the W56 group. (J) Time spent in the decision area (Scr: 89.85 ± 6.555 s, W56: 86.03 ± 4.873 s, P = 0.6467) was not influenced by presynaptic Rac1 inhibition. (K) Schematic of the DNMTP T-maze test. (L) Percentage of correct choice trials during the DNMTP task. (M) Schematic of the light/dark transition test. (N) The latency to the first transition (Scr: 8.263 ± 1.313 s, W56: 9.827 ± 1.571, P = 0.2130) and (O) percentage of time spent in the light box (Scr: 47.22 ± 2.750%, W56: 45.42 ± 2.008%, P = 0.6054) were not significantly different. (P) Schematic of the open field test. No significant change was observed in (Q) the distance traveled (Scr: 2,135 ± 117.6 cm, W56: 2,030 ± 195.9 cm, P = 0.1417), (R) velocity (Scr: 8.665 ± 0.1518 cm/s, W56: 8.262 ± 0.2067 cm/s, P = 0.1242), (S) percentage of time in movement (Scr: 76.91 ± 2.035%, W56: 76.67 ± 2.103%, P = 0.9347), (T) percentage of time spent in the center (Scr: 43.20 ± 2.980%, W56: 46.84 ± 2.822%, P = 0.3842), and (U) number of stereotypic activities (Scr: 6,912 ± 445.5, W56: 6,470 ± 430.4, P = 0.4827). Data are expressed as mean ± SEM with ns. not significant, and * p < 0.05.

Figure 3.

Hippocampal presynaptic Rac1 inhibition does not affect other types of learning and memory. (A) Schematic and timeline of the Morris water maze test (Scr: n = 16 mice; W56: n = 15 mice). During the MWM acquisition trials, (B) average swimming speed and (C) escape latency showed no significant differences across treatment groups (two-way repeated measures ANOVA; Swimming speed: F(1,29) = 0.253, P = 0.619; Escape latency: F(1,29) = 1.40, P = 0.246). (D) Percentage of time spent in the platform quadrant during the probe trials was also not significantly different across treatment groups (F(1,29) = 0.024, P = 0.878). (E) Timeline of the fear conditioning test. (F) Schematic of fear conditioning sessions. (G) Percentage of time spent freezing. Data are expressed as mean ± SEM.

Figure 4.

Hippocampal postsynaptic Rac1 inhibition does not affect hippocampal-dependent learning and memory. (A) Illustration of the postsynaptic Rac1 inhibitor constructs. (B,C) Fluorescence intensity profiles validated postsynaptic localization of the constructs. The radial-arm maze revealed no significant differences in (D) the total number of arm entries (Scr: 32.25 ± 1.896, W56: 30.56 ± 1.245, P = 0.4627), (E) number of arm entries before making the first error (Scr: 5.563 ± 0.3532, W56: 5.563 ± 0.3287, P = 0.8199), and (F) latency to the first error (Scr: 37.19 ± 3.500 s, W56: 34.56 ± 2.630 s, P = 0.9919). The spontaneous alternation Y-maze indicated no significant difference in (G) the total number of arm entries (Scr: 38.75 ± 2.452, W56: 38.63 ± 1.798, P = 0.9675), (H) latency to the first arm entry (Scr: 12.68 ± 1.819 s, W56: 10.07 ± 2.240 s, P = 0.3226), and (I) percentage of alternations (Scr: 45.61 ± 2.254%, W56: 42.56 ± 2.723%, P = 0.3953). (J) The percentage of correct choice trials in the DNMTP T-maze was comparable for both groups throughout the test. In the Morris water maze, (K) the escape latency from acquisition sessions (F(1,29) = 0.04019, P = 0.8425) and (L) the time spent in the platform quadrant (F(1,29) = 0.05692, P = 0.8131) also revealed no significant differences. (M) Percentage of time spent freezing. The difference between treatment groups was significant during the contextual trial 10 days after training (two-way repeated measures ANOVA; Treatment: F(1,27) = 6.169, P = 0.0195), but not in other trials. Data are expressed as mean ± SEM with ns. not significant, * p < 0.05, ** p < 0.01, *** p < 0.001, and **** p < 0.0001.

Figure 5.

EM analysis of presynaptic Rac1 inhibition in hippocampal cells. (A,B) Representative EM images of the dorsal CA1 stratum radiatum transduced with AAV-hSyn1-Scr-Syn1a (A) and AAV-hSyn1-W56-Syn1a (B). (C) Diameter of axonal boutons (Scr: 493.8 ± 28.38 nm, W56: 421.5 ± 41.47 nm; P = 0.1437). (D) Number of synaptic vesicles per bouton (Scr: 17.28 ± 1.552, W56: 20.00 ± 3.486; P = 0.9877). (E) Number of dendritic spines per 100 μm² (Scr: 51.64 ± 1.927 spines/100 μm², W56: 52.46 ± 2.068 spines/100 μm²; P = 0.7730). (F) Bar graph and frequency distribution of diameter of synaptic vesicles show that presynaptic Rac1 inhibition leads to larger vesicles (Scr: 30.02 ± 0.3845 nm, W56: 36.35 ± 0.4993 nm; P < 0.0001). (G) Synaptic vesicles were located relatively further from the active zone in the W56 group (Scr: 0.3193 ± 0.008885, W56: 0.3869 ± 0.01032; P ≈ 0.0001). The position of individual vesicles within a presynaptic terminal was normalized as a value ranging from 0 to 1, where 0 is closest to the active zone. (H) Area of dendritic spines was larger following presynaptic Rac1 inhibition (Scr: 124,400 ± 4,342 nm², W56: 144,700 ± 5,676 nm²; P = 0.0276). Kolmogorov-Smirnov test was performed on data that did not pass normality test. Data are expressed as mean ± SEM with ns. not significant, * p < 0.05, *** p < 0.001, and **** p < 0.0001.

Figure 6.

Functional protein interaction network of presynaptic Rac1. (A) Schematic of the presynaptic Rac1 BioID construct. (B,C) Presynaptic Rac1 BioID constructs are co-localized with a presynaptic marker Bassoon at synaptic boutons in C57BL/6J primary neuron cultures. (D) Timeline of AAV injection and protein sample collection. (E,F) Expression of presynaptic Rac1 BioID constructs (E) and biotinylation (F) were verified using Western blot. (G) Volcano plot of the biotinylated proteins. Red dots represent the 19 proteins that were significantly enriched in the Rac1 CA proteome. Dashed lines correspond to p-value = 0.05 (horizontal) or FC = ±1.5 (vertical). (H) Network of biotinylated proteins that were significantly enriched (FC > 1.5 and P < 0.05) in the CA proteome. Edges represent protein-protein interactions identified from our proteomic data (red) or reported from STRING (blue). (I) Phosphorylated peptides that were significantly enriched in the CA proteome. Numbers in gray boxes represent phosphorylation sites. Arrows represent predicted kinase–substrate pairs with high percentile scores (>90). (J) Schematic of Syntaxin-1 and Munc18a (PDB ID: 3C98). A red residue represents the phosphorylation site (Ser 109). (K) Schematic of Synaptotagmin-1 and SNARE complex (PDB ID: 5CCG). Colors are assigned randomly for different proteins/domains. Data are expressed as mean ± SEM with ** p < 0.01.

Figure 7.

Model of presynaptic Rac1 regulation and its signaling cascades in spatial working memory. This study suggests this process is driven by kinases-mediated phosphorylation of GEF/GAPs and vesicle proteins, which regulate vesicle cycle and actin dynamics.