As detailed above, we found that Prosapip1 neuronal knockout mice have deficits in LTP in the dHP, and LTP is the hallmark of learning and memory (Bliss and Collingridge, 1993; Frey and Morris, 1998; Collingridge et al., 2004). We also found that global neuronal knockout and dHP-specific knockout of Prosapip1 results in deficits in spatial learning and memory. Specifically, we showed that Prosapip1 in the dHP is required for normal performance in the novel object recognition, novelty T-maze, 3-chamber social interaction, and Barnes maze tests, which all require a unique form of learning and memory (Davis et al., 1992; Broadbent et al., 2010; Barker and Warburton, 2011; Rosenfeld and Ferguson, 2014; Lueptow, 2017; Pitts, 2018; Sánchez-Rodríguez et al., 2022). The novel object recognition test is primarily examining recognition memory, and with an inter-trial interval of 24 hr, specifically long-term memory (Broadbent et al., 2010; Barker and Warburton, 2011; Antunes and Biala, 2012; Lueptow, 2017; Cinalli et al., 2020). The novelty T-maze focuses on both spatial learning and memory, and with its short, 1 min inter-trial interval, this procedure is examining spatial working memory (Sharma et al., 2010; d’d’Isa et al., 2021). The 3-chamber social interaction test is assessing baseline sociability and also social recognition memory (Meira et al., 2018; Tzakis and Holahan, 2019; Wang and Zhan, 2022; Cope et al., 2023; Wei et al., 2024). Finally, the Barnes maze tests both spatial learning and working memory in intra-day trials, while simultaneously testing long-term contextual and spatial memory in inter-day trials and the probe test (Bach et al., 1995; Sharma et al., 2010; Rosenfeld and Ferguson, 2014; Pitts, 2018). Interestingly, loss of hippocampal LTP has been shown to impair spatial, but not contextual memory in the Barnes maze (Bach et al., 1995). As presented here, Prosapip1 knockout mice significantly reduced distance traveled to exit but did not switch to spatial searching. In this example, Prosapip1 knockout mice are retaining the contextual understanding of escaping the platform but do not recall the spatial location of the exit. Additionally, the link between NMDAR function and learning and memory is well established (Newcomer et al., 2000; Li and Tsien, 2009). For example, blockage of hippocampal NMDA receptors impairs spatial learning in rats (Davis et al., 1992; Bye and McDonald, 2019). Prosapip1 is therefore likely controlling the reinforcement of learning and memory by PSD scaffolding, stabilization, and GluN2B synaptic localization, leading to LTP.
We observed that Prosapip1 knockout specifically in the dHP replicated the recognition, social, and spatial learning and memory deficits exhibited by the global neuronal knockout mice, suggesting that Prosapip1 is controlling these learning and memory processes specifically in the dHP. The dHP primarily controls memory formation and recall (Eichenbaum, 1997; Broadbent et al., 2004; Squire et al., 2004; Pilly and Grossberg, 2012). We found that LTP in the CA1 subregion of the dHP was reliant on Prosapip1. The CA1 subregion is critically involved in contextual memory, object recognition memory, and spatial memory (Tsien et al., 1996; Lee and Kesner, 2004; Daumas et al., 2005; Sanderson et al., 2009; Sharma et al., 2010; Stevenson et al., 2018; Bye and McDonald, 2019; Cinalli et al., 2020; Jeong and Singer, 2022). Loss of Prosapip1 in CA1 is likely leading to decreased performance in the novel object recognition, novelty T-maze, and Barnes maze tests. The lack of social recognition displayed by Prosapip1(fl/fl);Syn1-Cre(+) mice and AAV-Cre-infected mice is likely attributed to the loss of Prosapip1 in the CA2 subregion of the dHP, which is the primary subregion controlling social recognition memory (Meira et al., 2018; Tzakis and Holahan, 2019; Wang and Zhan, 2022; Cope et al., 2023; Wei et al., 2024). Specifically, silencing the CA2 subregion of the dHP impairs social memory formation and consolidation (Meira et al., 2018). However, the CA3 and DG subregions of the dHP are also involved in spatial and contextual memory (Broadbent et al., 2004; Lee and Kesner, 2004; Daumas et al., 2005). As our conditional knockout strategy resulted in Prosapip1 deletion from the whole dHP, further studies are required to dissect the subregion specificity of the contribution of Prosapip1 to recognition, social, and spatial learning and memory processes.
Memory consists of three primary processes: encoding, consolidation, and retrieval (Straube, 2012). In this study, the defect in memory function is likely due to a failure to encode new information (Bye and McDonald, 2019) or consolidate this ‘short-term’ into ‘long-term’ memory (Yang et al., 2022). The spatial T-maze experiment utilized a short inter-trial interval of 1 min, which requires working spatial memory (Sharma et al., 2010), and Prosapip1 knockout mice exhibited a failure to encode new information. Similarly, the Barnes maze training trials were separated by an inter-trial interval of 30 min, but Prosapip1 knockout mice did not acquire spatial memory between training trials, nor during the longer consolidation periods between days, again implying a failure to encode spatial information or consolidate this information. The lack of synaptic localization of GluN2B is likely underlying the loss of memory encoding or consolidation (Nachtigall et al., 2024). It is unlikely that NMDAR dysfunction is affecting the retrieval of memory, as studies have exhibited rats’ ability to use previously acquired spatial information during NMDAR blockage (Bast et al., 2005; Mackes and Willner, 2006; Bye and McDonald, 2019).
Prosapip1 belongs to the Fezzin family of proteins (Wendholt et al., 2006). It is important to note that other Fezzins do not compensate for the loss of Prosapip1 in the dHP. Knockout of other Fezzins, like PSD-Zip70, also leads to cognitive deficits (Mayanagi et al., 2015). However, these deficits were attributed to the action of PSD-Zip70 in the PFC. Therefore, one could hypothesize that proteins in this family enact their function in specific brain subregions.
In summary, Prosapip1 in the dorsal hippocampus is integral to the synaptic localization of SPAR, PSD-95, and GluN2B, which are required for the formation of LTP and subsequent spatial learning and memory behavior. Abnormalities with PSD proteins are associated with neuropsychiatric disorders (Kaizuka and Takumi, 2018), and further unraveling of the physiological role of Prosapip1 may unlock insights into normal and abnormal mechanisms of learning and memory.