Regulation of synapse density by Pumilio RNA-binding proteins

Abstract

The formation, stabilization, and elimination of synapses are tightly regulated during neural development and into adulthood. Pumilio RNA-binding proteins regulate the translation and localization of many synaptic mRNAs and are developmentally downregulated in the brain. We found that simultaneous downregulation of Pumilio 1 and 2 increases both excitatory and inhibitory synapse density in primary hippocampal neurons and promotes synapse maturation. Loss of Pum1 and Pum2 in the mouse brain was associated with an increase in mRNAs involved in mitochondrial function and synaptic translation. These findings reveal a role for developmental Pumilio downregulation as a permissive step in the maturation of synapses and suggest that modulation of Pumilio levels is a cell-intrinsic mechanism by which neurons tune their capacity for synapse stabilization.

Keywords: CP: Molecular biology; CP: Neuroscience; Pumilio; RNA-binding proteins; local translation; mRNA regulation; synapse.

Figure 1.. Pumilios regulate synapse density

(A) Developmental mRNA expression of Pum1 and Pum2 (left) and their shared mRNA targets (right) in the mouse cortex. (B) Immunoblotting of Pum1 and Pum2 protein levels in primary rat hippocampal neurons at DIV6–DIV21. Mean ± SEM of n = 4 biological replicates. Mixed-effects analysis with Geisser-Greenhouse correction; significance indicator on the protein name indicates an overall fixed effect of DIV on protein level. (C) Neurons infected on DIV2 with shControl, shPum1, or shPum2 lentiviruses were immunostained at DIV14–DIV15 for Syn-I and PSD-95. (D) Quantification of (C). n = 80 fields per condition from 4 biological replicates, Kruskal-Wallis test with uncorrected Dunn’s test. (E) Neurons infected on DIV2 with shControl or shPum lentiviruses were immunostained at DIV14–DIV15 for Syn-I and PSD-95. (F) Quantification of (E). n = 59–60 fields per condition from 3 biological replicates, Mann-Whitney test. (G) Neurons infected on DIV1 with shControl or shPum lentiviruses and lentiviruses expressing shRNA-resistant Pum1 and Pum2 CDS (Pum1^R and Pum2^R) were immunostained at DIV21 for Syn-I and PSD-95. (H) Quantification of (G). n = 80 fields per condition from 4 biological replicates, Kruskal-Wallis test with uncorrected Dunn’s test. (A–H) The total number of co-localized puncta in a field was normalized to the area covered by the cell in the cell marker channel (EGFP or ZsGreen1). Insets, 2× magnification. Arrows indicate co-localized Syn-I+/PSD-95+ puncta. Scale bars, 10 μm. *p < 0.05, **p < 0.01, ****p < 0.0001, and ns, not significant.

Figure 2.. Dual Pumilio downregulation increases spine density but not dendrite complexity

(A) Neurons infected on DIV1 with shControl, shPum1, or shPum2 lentiviruses were immunostained for MAP2 at DIV14. Sholl profiles (left) were segmented into bins of 50 μm for analysis of area under the curve (right). Mean ± SEM of n = 67–78 neurons per condition from 4 biological replicates. Two-way ANOVA with Šídák’s multiple comparisons test. Scale bars, 50 μm. (B) Neurons infected on DIV1 with shControl or shPum lentiviruses were immunostained for MAP2 at DIV14. Sholl analysis was performed as in (A). Mean ± SEM of n = 70–77 neurons per condition from 4 biological replicates. Scale bars, 50 μm. (C) Neurons infected on DIV1 with shControl or shPum lentiviruses were immunostained for MAP2 and phalloidin at DIV19, and the MAP2 channel was used for quantification. The density and length of protrusions were measured from regions of interest (ROIs) of single dendrites as pictured. Quantification of n = 60 ROIs per condition from 3 biological replicates, Mann-Whitney test. Scale bars, 10 μm in full image and 2 μm in inset. (A–C) ****p < 0.0001 and ns, not significant.

Figure 3.. Pumilios regulate synapse density at late stages of maturation by restricting synapse stability

(A) Neurons infected on DIV1 with shControl or shPum lentiviruses were immunostained at DIV6–DIV21 for Syn-I and PSD-95. (B) Quantification of (A). Two-way ANOVA with Šidák’s multiple comparisons test. Mean ± SEM of 56–70 fields per condition/time point from 5 biological replicates. (C) Neurons infected on DIV15 with shControl, shPum1, or shPum2 lentiviruses were immunostained at DIV21 for Syn-I and PSD-95. (D) Quantification of (C). n = 100 optical fields from 5 biological replicates, Kruskal-Wallis test with uncorrected Dunn’s test. (E) Neurons infected on DIV15 with shControl or shPum1-shPum2 lentiviruses were immunostained at DIV21 for Syn-I and PSD-95. (F) Quantification of (E). n = 80 fields from 4 biological replicates, Mann-Whitney test. (G) Neurons infected on DIV1 with shControl or shPum lentiviruses were treated on DIV18 with 2.5 μM latrunculin A or DMSO and immunostained at DIV19 for Syn-I. (H) Quantification of (G). n = 59–60 fields from 3 biological replicates, two-way ANOVA with Šidák’s multiple comparisons test. (A–H) The total number of co-localized puncta in a field was normalized to the area covered by the cell in the cell marker channel (EGFP or β-III-tubulin). Inset, 2× magnification. Arrows indicate co-localized Syn-I+/PSD-95+ puncta. Scale bars, 10 μm. *p < 0.05, **p < 0.01, ****p < 0.0001, and ns, not significant.

Figure 4.. Pumilios regulate both excitatory and inhibitory synapses and control the expression of many synaptic genes

(A) Neurons infected on DIV1 with shControl or shPum lentiviruses were immunostained at DIV21 for VGAT and gephyrin. Arrows indicate co-localized VGAT+/gephyrin+ puncta (left). Quantification of inhibitory synapse count (center) and gephyrin puncta size (right) from n = 80 fields from 4 biological replicates, Mann-Whitney test. (B) Neurons infected on DIV1 with shControl or shPum lentiviruses were immunostained at DIV21 for VGLUT1 and VGAT (left). Quantification of puncta count (center) and VGLUT1/VGAT ratio (right) from n = 90–99 fields per condition from 5 biological replicates, Mann-Whitney test. (C) Gene Ontology analysis on neonatal mouse cortical brain iCLIP targets of Pum1, Pum2, or their intersection. (D) SynGO analysis of the shared Pum1/2 targets from (C). Major gene categories are shown in bold, and subcategories are shown in regular type. (E) Gene set enrichment analysis of RNA-seq data from conditional neural double knockout (Nestin-Cre-mediated, Ndcko) of Pum1 and Pum2 in mouse brain showing suppression or activation of biological processes related to synaptic development, translation, and mitochondrial function. (A and B) The total number of co-localized or individual puncta in a field was normalized to the area covered by the cell in the EGFP channel. Insets, 2× magnification. Scale bars, 10 μm. *p < 0.05, **p < 0.01, and ****p < 0.0001.