Figure 4. Designed linear polyUb chains with either short or long Ub-Ub linkers are less effective at enhancing phase separation of UBQLN2.

(A) Experimental X-ray scattering curves (black) for designed M1-Ub4 constructs overlaid with scattering curve predicted (red) from single best structure (shown) derived from SASSIE conformational ensembles. Hydrophobic patch residues L8, I44, and V70 are represented as yellow spheres. (B) P(r) profiles of designed M1-Ub4 constructs with shorter Ub-Ub linkers. (C) Temperature-concentration phase diagrams of 450C with designed M1-Ub4 constructs. Phase separation of 450C and M1-Ub4 constructs occurs at temperatures above (but not below) phase boundaries. (D) Theory-derived curves overlaid on experimental data. (E) P(r) profiles of designed M1-Ub4 constructs with longer Ub-Ub linkers. (F) Temperature-concentration phase diagrams of 450C (fixed concentration of 50 µM) with designed M1-Ub4 constructs. (G) Theory-derived curves overlaid on experimental data. (H) Theory-derived hub inclusion energy for M1-Ub4 constructs of various Ub-Ub linker lengths (as color-coded in panels D and G). Label is added for M1-Ub4 (1–73) as this construct binds less effectively to 450C. (I) For 450C phase separation, theory-derived hub inclusion energy for all designed polyUb hubs in this study. Data points are color-coded as in panels D and G, and from Fig. 3F. Inclusion energies for M1-Ub4 and HT6-Ub constructs are represented as squares and circles, respectively. (J) For full-length UBQLN2 phase separation, theory-derived hub inclusion energy for naturally-occurring Ub4 polyUb chains, two designed M1-Ub4 constructs of longer Ub-Ub linker lengths, HT6-(GS)_x-Ub hubs (filled circles), and HT6-(PA)_x-Ub hubs (open circles). Parabolic trendlines (dotted lines) in panels I and J are drawn to guide the eye. R_g values in panels H, I, and J are of the different ligand hubs used.