Fig. 4: Transmembrane ___domain length of de novo designed proteins affects cellular trafficking and extracellular vesicle loading.
a Structure of synthetic, de novo designed transmembrane ___domain (TMD) proteins assessed. Two families of proteins were assessed, either containing four or twelve transmembrane domains. Different lengths of transmembrane domains in each family were assessed. b Pearson’s coefficient (R) from confocal microscopy images evaluating the subcellular localization of mRFP1-tagged de novo designed proteins with a plasma membrane-staining dye, Cell Mask. Data are reported in box and whisker plots collected from 30 cells from two independent experiments; each symbol is a single cell. To compare plasma membrane localization, a one-way ANOVA with Tukey’s multiple comparisons was performed on the dimeric 4TMD designs and an unpaired, two-tailed T-test was performed on the hexameric 12TMD designs (**p < 0.01; ****p < 0.0001). c Partitioning of de novo designed proteins with liquid-ordered lipid raft regions (LO) of GPMVs as determined by confocal microscopy analysis. Raft association did not significantly differ among de novo designed proteins (one-way ANOVA, Tukey’s multiple comparisons correction). Both 24 Å constructs did not associate with GPMV membranes. Data reported are the average and standard deviation of at least 20 GPMVs from three independent experiments (n = 20). For (b, c), the upper and lower bounds represent the minima and maxima of each data measurement, while the box plot marks the lower and upper quartile, as well as the median. d, e Protein loading and partitioning into EVs as calculated by western blots probing the 3x FLAG tag (Supplementary Fig. 16). Each lane of the protein gel received equal numbers of vesicles. e EV partitioning was calculated by normalizing EV protein loading by expression of each construct in EV producer cells as determined by western blot. EV partitioning values were normalized by 24 Å 4TMD HS-EV partitioning. A two-way ANOVA (main effects) was performed to compare proteins per vesicle and EV partitioning within each protein family (dimeric 4TMD and hexameric 12TMD designs), and comparisons were evaluated using Tukey’s multiple comparisons correction (*p < 0.05; ****p < 0.0001). Samples were compared to the respective 24 Å long transmembrane ___domain for their protein family. Comparisons not shown were not significant. n = 2 independent sample preparations and western blots for (d, e). Error bars represent the SEM. Source data are provided as a Source Data file.
