The GTPase Rem2 regulates synapse development and dendritic morphology

Abstract

Rem2 is a member of the Rad/Rem/Rem2/Gem/Kir subfamily of small Ras-like GTPases that was identified as an important mediator of synapse development. We performed a comprehensive, loss- of-function analysis of Rem2 function in cultured hippocampal neurons using RNAi to substantially decrease Rem2 protein levels. We found that knockdown of Rem2 decreases the density and maturity of dendritic spines, the primary site of excitatory synapses onto pyramidal neurons in the hippocampus. Knockdown of Rem2 also alters the gross morphology of dendritic arborizations, increasing the number of dendritic branches without altering total neurite length. Thus, Rem2 functions to inhibit dendritic branching and promote the development of dendritic spines and excitatory synapses. Interestingly, binding to the calcium-binding protein calmodulin is required for the Rem2 regulation of dendritic branching. However, this interaction is completely dispensable for synapse development. Overall, our results suggest that Rem2 regulates dendritic branching and synapse development via distinct and overlapping signal transduction pathways.

Figure 1

Rem2 is widely expressed in primary hippocampal neurons. (a) Representative images of a stretch of dendrite immunostained for Rem2 (red) and MAP2 (blue) from a neuron transfected with either a GFP-expressing plasmid plus: (top rows) a control plasmid (“Control”), or a Rem2 RNAi-resistant cDNA (“OE”); (bottom rows) a control plasmid (“Control”), or a shRNA plasmid targeting Rem2 (“RNAi”). (b) Quantification of Rem2 and MAP2 staining intensity in Control and overexpressed conditions. (c) Representative images of a Control neuron immunostained for Rem2 and NeuN. Asterisks indicate p<0.01 compared to control conditions by two-sample t-test. Scale bars indicate (a) 5 μm and (b) 50 μm. n > 50 neurons/condition for all experiments. (See Supplementary Table 1 for detailed statistical information for the data presented in this and subsequent figures.)

Figure 2

Rem2 regulates the development and morphological maturation of dendritic spines. (a) Representative images of a stretch of dendrite from a neuron that has been transfected with either a GFP-expressing plasmid plus (clockwise from top left) a control plasmid (“Control”), a shRNA plasmid targeting Rem2 (“RNAi”), a shRNA plasmid targeting Rem2 plus a Rem2 RNAi-resistant cDNA (“Rescue”), or a Rem2 RNAi-resistant cDNA alone (“OE”). (b) Quantification of dendritic spine density for each condition. n > 20 neurons/condition. (c) Quantification of dendritic spine length and (d) dendritic spine head width for each condition. Asterisks indicate p<0.001 compared to control condition by two-way between-subjects ANOVA with Tukey posthoc test. n > 1000 spines/condition. Scale bar indicates 5 μm.

Figure 3

Rem2 inhibits branching but maintains the total length of the dendritic arborization. (a) Representative images of the morphology of a (clockwise from top left) Control, RNAi, Rescue, and OE neuron. Total neurite lengths are as follows (in μm): Control= 2997.16, RNAi= 3057.23, OE= 2882.6, Rescue= 2776.21. (b) Quantification of dendritic branching for each condition via Sholl analysis. Asterisks indicate p<0.001 compared to control condition by two-way repeated measures ANOVA with Tukey posthoc test. n > 40 neurons/condition. (c) Quantification of the number of neurites for each condition. Asterisk indicates p<0.001 compared to control condition by two-way between-subjects ANOVA with Tukey posthoc test. (d) Quantification of the mean and (e) total neurite length for each condition. Asterisks indicate p<0.05 compared to control condition by two-way between-subjects ANOVA with Tukey posthoc test. n > 800 neurites/condition. Scale bar indicates 50 μm.

Figure 4

Rem2 affects excitatory synapse development and dendritic branching via distinct pathways. L317G= an RNAi-resistant cDNA encoding a Rem2 protein that does not bind calmodulin. (a) Quantification of excitatory synapse density as PSD-95/synapsin overlapping puncta for Control, RNAi, OE, Rescue, L317G OE, and L317G Rescue conditions. Asterisk indicates p<0.001 compared to control condition by two-way between-subjects ANOVA with Tukey posthoc test. (b) Quantification of dendritic branching for all conditions via Sholl analysis. Asterisks indicate p<0.05 compared to control condition by two-way repeated measures ANOVA with Tukey posthoc test. (c) Representative images of neurons from Control, RNAi, and L317G Rescue conditions. Top image is GFP signal to depict the dendritic arbor. Bottom images are a stretch of dendrite from above neuron to depict synaptic staining (see Supplementary Figure 3 for separated fluorescence channels for all experiments with synaptic staining). Scale bars indicate 50 μm (neuron images) or 5 μm (stretch of dendrite images). n > 40 neurons/condition for all experiments.

Figure 5

The ability of Rem2 to bind GTP is essential for excitatory synapse development and proper dendritic arborization. S129N= an RNAi-resistant cDNA encoding a Rem2 protein that does not bind GTP. (a) Quantification of excitatory synapse density as PSD-95/synapsin overlapping puncta for Control, RNAi, OE, Rescue, S129N OE, and S129N Rescue conditions. Asterisks indicate p<0.001 compared to control condition by two-way between-subjects ANOVA with Tukey posthoc test. (b) Quantification of dendritic branching for all conditions via Sholl analysis. Asterisks indicate p<0.05 compared to control condition by two-way repeated measures ANOVA with Tukey posthoc test. (c) Representative images of neurons from Control, RNAi, and S129N Rescue conditions. Top image is GFP signal to depict the dendritic arbor. Bottom images are a stretch of dendrite from above neuron to depict synaptic staining. Scale bars indicate 50 μm (neuron images) or 5 μm (stretch of dendrite images). n > 40 neurons/condition for all experiments.

Figure 6

The ability of Rem2 to bind 14-3-3 proteins is essential for excitatory synapse development and proper dendritic arborization. S69A/S334A= an RNAi-resistant cDNA encoding a Rem2 protein that does not bind 14-3-3 proteins. (a) Quantification of excitatory synapse density as PSD-95/synapsin overlapping puncta for Control, RNAi, OE, Rescue, S69A/S334A OE, and S69A/S334A Rescue conditions. Asterisks indicate p<0.05 compared to control condition by two-way between-subjects ANOVA with Tukey posthoc test. (b) Quantification of dendritic branching for all conditions via Sholl analysis. Asterisks indicate p<0.05 compared to control condition by two-way repeated measures ANOVA with Tukey posthoc test. (c) Representative images of neurons from Control, RNAi, and S69A/S334A Rescue conditions. Top image is GFP signal to depict the dendritic arbor. Bottom images are a stretch of dendrite from above neuron to depict synaptic staining. Scale bars indicate 50 μm (neuron images) or 5 μm (stretch of dendrite images). n > 40 neurons/condition for all experiments.

Figure 7

The roles of Rem2 in excitatory synapse development and dendritic branching are independent of Rem2 binding to calcium channels. R236A/L263A= an RNAi-resistant cDNA encoding a Rem2 protein that does not bind the β subunit of calcium channels. (a) Quantification of excitatory synapse density as PSD-95/synapsin overlapping puncta for Control, RNAi, OE, Rescue, R236A/L263A OE, and R236A/L263A Rescue conditions. Asterisks indicate p<0.05 compared to control condition by two-way between-subjects ANOVA with Tukey posthoc test. (b) Quantification of dendritic branching for all conditions via Sholl analysis. Asterisks indicate p<0.01 compared to control condition by two-way repeated measures ANOVA with Tukey posthoc test. (c) Representative images of neurons from Control, RNAi, and R236A/L263A Rescue conditions. Top image is GFP signal to depict the dendritic arbor. Bottom images are a stretch of dendrite from above neuron to depict synaptic staining. Scale bars indicate 50 μm (neuron images) or 5 μm (stretch of dendrite images). n > 40 neurons/condition for all experiments.

Figure 8

Model depicting a putative interaction between Rem2 and CaMK signaling pathways. Rem2 limits dendritic complexity by sequestering calcium-bound CaM from CaMKs, thus limiting CaMK-dependent dendritic outgrowth.