Reduced neuronal self-avoidance in mouse starburst amacrine cells with only one Pcdhg isoform

Abstract

The clustered protocadherins (cPcdhs) are a family of ~60 homophilic cell adhesion molecules expressed across three gene clusters (Pcdha, Pcdhb, and Pcdhg) with a variety of essential roles in the developing nervous system. Some of these roles rely on specific isoforms, while others are more consistent with a model of isoform redundancy or a requirement for diversity. The γ-Pcdhs (expressed from the Pcdhg gene cluster) are particularly important for neuronal self-avoidance in starburst amacrine cells in the mouse retina. Here, we used mouse mutants to test two of the C-type isoforms - γC4 and γC5 - and found that neither was required for normal self-avoidance. Conversely, when we analyzed a mutant with only γC4 intact, we found significant failures in self-avoidance that could not be completely rescued by overexpression of this isoform from a transgene. We have recently found that this isoform is essential for normal neuronal survival during development, and our new findings here support the hypothesis that γC4 is specialized for the survival function at the expense of a significant role in self-avoidance.

Figure 1:. Individual γC-type isoforms are not required for ON-SAC self-avoidance.

A) A schematic of the Pcdhg locus with γA and γB type variable exons at the 5’ end and γC type isoforms at the 3’ end adjacent to the three constant exons. The Pcdhg^C4KO allele harbors a frameshifting 13-bp deletion disrupting the γC4 isoform. To circumvent the neonatal lethality of homozygous mutants, Pcdhg^C4KO is made compound heterozygous with Pcdhg^fcon3 and crossed with Chat^Cre to limit the loss of function to starburst amacrine cells (SACs). Pcdhg^C5KO includes a 7-bp frameshifting deletion in the Pcdhgc5 exon. B) RNAscope in adult retina verified that, along with many other retinal cell types, SACs expressed Pcdhgc5. Projections of ON SACs labeled with Brainbow in Chat^Cre from (C-E) Pcdhg^+/+, (F-H) Pcdhg^C4KO/fcon3, and (I-K) Pcdhg^C5KO/C5KO retinas show a range of normal, radially symmetrical morphologies. Individual SAC morphologies were quantified by (L) radial asymmetry and (M) Elo score. Scale bar is 50 μm in B, 100 μm in C-K.

Figure 2:. Disrupted self-avoidance in Pcdhg^1R1/1R1 mutant retinas.

A) A schematic representation of the Pcdhg^1R1 allele indicates a large deletion from Pcdhga1 to Pcdhgc3 and a frameshifting mutation in Pcdhgc5, leaving only Pcdhgc4 intact. Projections through the ON SAC plexus stained for Chat in (B-D) Pcdhg^+/+ and (E-G) Pcdhg^1R1/1R1 retinas. Plexus in Pcdhg^1R1/1R1 is characterized by large gaps between dendritic fascicles, as quantified in H and I. J-L) Individual Pcdhg^1R1/1R1 SACs labeled with Brainbow in Chat^Cre show self-avoidance failures with a variety of severity, quantified in M and N. Scale bar is 100 μm. **** is P < 0.0001, ** is P < 0.01 by Tukey (H,M) or Dunn’s (I,N) multiple comparison test.

Figure 3:. Transgene expression from C4-GFP partially rescues self-avoidance in Pcdhg^1R1/1R1.

A) Schematic representation of the C4-GFP allele knocked in at the Rosa locus. Expression from the transgene occurs after Cre-mediated recombination of a stop cassette. B-D) ON SAC plexus images from Pcdhg^1R1/1R1::C4-GFP retinas demonstrates interdendritic gaps and dendritic fasciculation intermediate between wild type controls and Pcdhg^1R1/1R1 mutants, as quantified in E and F. G-I) Individual SACs labeled with Brainbow in Chat^Cre also showed incomplete rescue after transgene expression. Quantified in J-K. Scale bar is 100 μm. **** is P < 0.0001, *** is P < 0.001, ** is P < 0.01, * is P < 0.05, ns is not significant by Tukey (E,J) or Dunn’s (F,K) multiple comparison test.

Figure 4:. Transgene expression from C4-GFP is unable to rescue self-avoidance in SACs with both Pcdhg and Pcdha disrupted.

A) Brainbow vectors were modified to include a cassette for shRNA knockdown of Pcdha by introducing a human U6 promoter and hairpin RNA sequence antisense to the EF-1α promoter regulating fluorescent protein expression. shα::Brainbow AAV vectors were injected into (B-D) Pcdhg^+/+, (E-G) Pcdhg^1R1/1R1, and (H-J) Pcdhg^1R1/1R1::C4-GFP retinas. Self-avoidance defects quantified in K and L were severe, and the presence of the C4-GFP transgene was unable to provide significant rescue. Scale bar is 100 μm. **** is P < 0.0001, *** is P < 0.001, ns is not significant by Tukey (K) or Dunn’s (L) multiple comparison test.

Figure 5:. Localization of γC4-GFP does not depend on other γ-Pcdh isoforms.

The distribution of GFP-tagged protein in ON SACs was compared between Pcdhg^1R1/+ (with one copy of the other 21 Pcdhg isoforms, A-B) and Pcdhg^1R1/1R1mutants (without any of the other 21 Pcdhg isoforms, C-D, F). The ratio of fluorescent signal between somas and dendrites was measured in E, with no significant difference between the two genotypes. When SAC somas were adjacent to each other (F) there was increased GFP labeling at cell-cell junctions, consistent with γC4 successfully trafficking to the cell membrane and engaging in trans interactions in the absence of other γ-Pcdh isoforms.