Fig. 5: Selection of mutations for the Aes^N engineering strategy.

a Model of Aes^N aggregation. Unsaturated aggregation-prone regions (APRs) in disordered regions of Aes^N serve as nuclei for soluble aggregate formation. b Sequence-based consensus prediction of amyloidogenic pattern formation in the Aes^N fragment using the web tool AMYLPRED2 including 10 different prediction algorithms⁴⁰. The upper graph indicates the number of positive hits of the different algorithms plotted against the native Aes^N fragment sequence. APRs labeled in red are defined by ≥5 positive hits (dashed line indicates threshold). The heat map below shows the effect of the selected single mutations on the same aggregation-prone site prediction. c X-ray structure of the Aes intein, obtained with the fused protein MYIDTD-Aes^N(S1A)-GSH-Aes^C(N159A)-SVYLN (9P). Shown is an overlay of chain A and B in the unit cell (for details see Supplementary Table 4). The three strongest APRs are marked in the structure (red). The artificial GSH inker is shown in dark gray. Residues T69, F75 and M118 are indicated with red arrows. d Primary sequence of the Aes^N (aa1−120) and the Aes^C fragment (aa121-139), each with 3 extein residues. Indicated are positively (red) and negatively (blue) charged residues, conserved motifs (underlined), key catalytic amino acids (green) and the location of secondary structure elements. For panel b, source data are provided as a Source Data file.