Exploring the evolution and function of Canoe’s intrinsically disordered region in linking cell-cell junctions to the cytoskeleton during embryonic morphogenesis

Abstract

One central question for cell and developmental biologists is defining how epithelial cells can change shape and move during embryonic development without tearing tissues apart. This requires robust yet dynamic connections of cells to one another, via the cell-cell adherens junction, and of junctions to the actin and myosin cytoskeleton, which generates force. The last decade revealed that these connections involve a multivalent network of proteins, rather than a simple linear pathway. We focus on Drosophila Canoe, homolog of mammalian Afadin, as a model for defining the underlying mechanisms. Canoe and Afadin are complex, multidomain proteins that share multiple domains with defined and undefined binding partners. Both also share a long carboxy-terminal intrinsically disordered region (IDR), whose function is less well defined. IDRs are found in many proteins assembled into large multiprotein complexes. We have combined bioinformatic analysis and the use of a series of canoe mutants with early stop codons to explore the evolution and function of the IDR. Our bioinformatic analysis reveals that the IDRs of Canoe and Afadin differ dramatically in sequence and sequence properties. When we looked over shorter evolutionary time scales, we identified multiple conserved motifs. Some of these are predicted by AlphaFold to be alpha-helical, and two correspond to known protein interaction sites for alpha-catenin and F-actin. We next identified the lesions in a series of eighteen canoe mutants, which have early stop codons across the entire protein coding sequence. Analysis of their phenotypes are consistent with the idea that the IDR, including the conserved motifs in the IDR, are critical for protein function. These data provide the foundation for further analysis of IDR function.

The Dilute domain in Canoe is not essential for linking cell junctions to the cytoskeleton but supports morphogenesis robustness

“…The DIL domains of Cno and afadin are well conserved in sequence and length, with 47% of the 374 fly amino acids identical to the corresponding rat 393 amino acids ( Gurley et al, 2023 ; Fig. 1B ).…”

“…cno R2 is zygotically embryonic lethal, so in a cross of +/cno R2 parents, the 25% of embryos who are cno R2 /cno R2 die as embryos, whereas the +/cno R2 and +/+ progeny from this cross are fully viable. However, due to the strong maternal contribution of wild-type Cno, although cno maternal and zygotic mutants have defects in most morphogenetic movements, most cno R2 homozygous null zygotic mutants only have mild defects in head involution but few defects in other morphogenetic events like dorsal closure ( Gurley et al, 2023 ; Sawyer et al, 2009 ). This creates a sensitized situation where small reductions in maternal Cno function can enhance these defects.…”

“…In most cno R2 zygotic mutants derived from +/cno R2 parents, germband extension, retraction and dorsal closure go to completion, such that the only defects in most embryos are in head involution, leading to a disrupted head skeleton ( Fig. 8E,F ; Gurley et al, 2023 ) – 89% of embryos were in these categories, whereas only 11% had more severe phenotypes ( Fig. 8F–H ; quantified in Fig.…”

“…Cno and Afadin share five predicted folded protein domains, followed by a long intrinsically disordered region (IDR) ( Fig. 1A ; Gurley et al, 2023 ). At their N-terminus are two Ras-association (RA) domains, which bind the small GTPase Rap1 ( Boettner et al, 2000 ).…”

“…The only other proteins known to have DIL domains are the Cno/afadin family, found in all animals, and their distantly related and less studied vertebrate paralogs RADIL and Rasip. The DIL domains of human afadin and Drosophila Cno share 48% amino acid sequence identity ( Gurley et al, 2023 ). Conservation in the vertebrate lineage is even stronger, with the DIL domains of human and zebrafish afadin sharing 89% identity.…”