Skip to main content
. 2020 Apr 16;9:e55015. doi: 10.7554/eLife.55015

Figure 4. Oligomerization of Axin1 DAX and effects of oligomerization on signaling.

(A) Comparison of TEV-cleaved Axin1 DAX and Dvl2 DIX variants run on a Superose six size exclusion column (NG = E815N/E816G, NQ/NG = D793N/E794Q/E815N/E816G). Proteins were injected at 180 μM, and diluted to approximately 15 μM on the column. (B) SEC-MALS analysis of Axin1 DAX size as a function of concentration. (C) Superposition of the rat Axin1 crystal structure (PDB 1WSP; magenta) on one of the Dvl2 DIX domains (blue), showing potential clashes with another DIX protomer (green) across the inter-strand interface. Also shown is the substitution of N82 in DIX with E815 in DAX, which would eliminate the hydrogen bond with D63 of DIX and introduce electrostatic repulsion. Axin residue numbers are from the human Axin1 sequence. (D) Cryo-electron micrograph of DAX NQ/NG filaments.

Figure 4.

Figure 4—figure supplement 1. Sequence alignment of selected Dvl and Axin DIX domains.

Figure 4—figure supplement 1.

Dvl DIX domains are shown in the top block; sequence numbers correspond to the mouse Dvl2 domain used in this paper, which is identical to the human sequence shown. Axin DAX domains are shown in the lower block; sequence numbers correspond to human Axin1 sequence used in this study. Secondary structure elements are indicated. Arrowheads indicate sites of mutations used in this study. Black shading indicates 100% similarity; dark grey, 80–99% similarity; light grey, 60–79% similarity; and no shading,<60% similarity. Alignment was carried out in Geneious version 10.2.6 (Kearse et al., 2012) using the ClustalOmega plugin (Madeira et al., 2019). Similarity was calculated using the PAM250 algorithm (Schwartz and Dayhoff, 1978) with a threshold of 1.