N-cadherin and neuroligins cooperate to regulate synapse formation in hippocampal cultures

Abstract

Cadherins and neuroligins (NLs) represent two families of cell adhesion proteins that are essential for the establishment of synaptic connections in vitro; however, it remains unclear whether these proteins act in concert to regulate synapse density. Using a combination of overexpression and knockdown analyses in primary hippocampal neurons, we demonstrate that NL1 and N-cadherin promote the formation of glutamatergic synapses through a common functional pathway. Analysis of the spatial relationship between N-cadherin and NL1 indicates that in 14-day in vitro cultures, almost half of glutamatergic synapses are associated with both proteins, whereas only a subset of these synapses are associated with N-cadherin or NL1 alone. This suggests that NL1 and N-cadherin are spatially distributed in a manner that enables cooperation at synapses. In young cultures, N-cadherin clustering and its association with synaptic markers precede the clustering of NL1. Overexpression of N-cadherin at this time point enhances NL1 clustering and increases synapse density. Although N-cadherin is not sufficient to enhance NL1 clustering and synapse density in more mature cultures, knockdown of N-cadherin at later time points significantly attenuates the density of NL1 clusters and synapses. N-cadherin overexpression can partially rescue synapse loss in NL1 knockdown cells, possibly due to the ability of N-cadherin to recruit NL2 to glutamatergic synapses in these cells. We demonstrate that cadherins and NLs can act in concert to regulate synapse formation.

FIGURE 1.

Spatial distribution of N-cadherin and NL1 at glutamatergic synapses. A, confocal images of 6- and 14-DIV neurons transfected at 2 and 10 DIV with GFP-NL1 and immunolabeled with VGlut-1 and N-cadherin. The low magnification image (far right) illustrates a neuron transfected with GFP-NL1 with a mask outlining the area of the cell chosen for analysis. Higher magnification images depict representative raw images of GFP-NL1 clusters and GFP-NL1 clusters following thresholding, as well as immunostaining for VGlut-1 and N-cadherin. Open arrows denote VGlut-1 clusters with no colocalization; closed arrows show colocalization between VGlut-1 and GFP-NL1; open arrowheads show colocalization between VGlut-1 and N-cadherin; closed arrowheads show triple colocalization between VGlut-1, GFP-NL1, and N-cadherin. B, quantification of the proportion of VGlut-1 puncta within the mask that is associated with GFP-NL1, N-cadherin (N-cad), or both. n = an average of 857.6 μm of neurite/cell from 23–27 cells per condition from three separate cultures. Scale bars = 10 μm.

FIGURE 2.

N-cadherin overexpression enhances the density of NL1 clusters in young but not mature neurons. A, confocal images of 6-DIV neurons transfected at 2 DIV with GFP-NL1 or GFP-NL1 + N-cadherin (Ncad)-CFP and immunolabeled with VGlut-1 and PSD-95. Raw images of GFP-NL1 are shown on the far left, and thresholded images are shown in the second column. B, quantification of 6-DIV neurons demonstrates an increase in NL1 density, synapse density, and density of synaptically localized NL1 in N-cadherin-overexpressing cells. C, quantification of 14-DIV neurons demonstrates no change in NL1 density, synapse density, or density of synaptically localized NL1 in N-cadherin-overexpressing cells. n = 20 cells per condition from three separate cultures. *, p < 0.05; **, p < 0.01 (Student's t test). Scale bar = 10 μm.

FIGURE 3.

siRNA-mediated knockdown of N-cadherin and NL1. A, HEK293 cells transfected with N-cadherin siRNA plus wild-type N-cadherin-CFP display a significant decrease in N-cadherin-CFP levels compared with cells transfected with control nonspecific siRNA plus wild-type N-cadherin-CFP. N-cadherin siRNA did not attenuate levels of siRNA-resistant (res) N-cadherin-CFP. B, HEK293 cells transfected with NL1 siRNA plus wild-type HA-NL1 exhibit a significant decrease in HA-NL1 levels compared with cells transfected with control nonspecific siRNA plus wild-type HA-NL1. Conversely, NL1 siRNA did not attenuate levels of siRNA-resistant HA-NL1. n = three blots with three independent cultures. WB, Western blot.

FIGURE 4.

N-cadherin and NL1 cooperate to regulate synapse formation. A–C, confocal images of 14-DIV neurons transfected at 10 DIV with the indicated constructs and immunolabeled with VGlut-1 and PSD-95. D, quantification of the density of synapses on transfected neurons normalized to cells transfected with control siRNA. Asterisks along the x axis denote significant difference from cells transfected with control siRNA. *, p < 0.05; **, p < 0.01; ***, p < 0.001 (one-way analysis of variance with Tukey's post hoc test). n = an average of 640 μm of neurite/cell from 21–50 cells per condition from four separate cultures. E, the density of NL1 clusters is decreased in N-cadherin (N-cad) siRNA-expressing cells. n = 28–29 cells/condition from three separate cultures. ***, p < 0.001 (Student's t test). Scale bar = 10 μm.

FIGURE 5.

The proportion of NL2 at glutamatergic synapses is increased in NL1 knockdown cells overexpressing N-cadherin. A–C, confocal images of 14-DIV neurons transfected at 10 DIV with the indicated constructs and immunolabeled with VGlut-1 (blue) and NL2 (red). There is no change in the proportion of NL2 associated with VGlut-1 in N-cadherin-overexpressing cells (A, arrowheads; C). There is a significant increase in the proportion of NL2 associated with VGlut-1 in NL1 knockdown cells overexpressing N-cadherin (N-cad) (B, arrowheads; D). n = 9–15 cells/condition from two separate cultures. **, p < 0.01 (Student's t test). Scale bars = 10 μm. E, model for the functional interaction between N-cadherin and NLs. N-cadherin colocalizes with VGlut-1 clusters prior to the recruitment of NL1, and overexpression of N-cadherin enhances the recruitment of NL1 to these sites (i). Recruitment of NL1 to nascent synapses enhances the further development of synapses through NL1 interaction with neurexin (ii). In the absence of NL1, N-cadherin recruits NL2 to nascent synapses partially rescuing the effect of NL1 knockdown on synapse density (iii).