Journal: Science Advances

Article Title: Sequence-dependent scale for translocon-mediated insertion of interfacial helices in membranes

doi: 10.1126/sciadv.ads6804

Figure Lengend Snippet: The experimentally tested sequences are embedded as large points with black outlines. ( A ) UMAP embeddings with UniProt-derived sequences colored light blue if found in soluble domains and beige if found in TM domains. The UMAP embeddings cluster into two large lobes, delineating helices within soluble proteins from TM helices. All sequences tested with the LepB glycosylation assay (table S2) were visually clustered based on their embeddings and the predicted statistics over the three dispositions [aqueous ( P Wat ), interfacial ( P Int ), and TM ( P TM )] are shown as histograms. Each of the three helical dispositions rapidly becomes more dominant along a particular direction in the UMAP space, shown by the three vectors using the same color scheme for the three dispositions as in (green, yellow, and red for inserted, interfacial, and translocated location, respectively). ( B ) The same UMAP embeddings, colored according to average hydrophobicity computed using the Δ G PDB scale, show a clear trend toward higher hydrophobicity (i.e., lower average Δ G PDB ) for helices from soluble proteins from left to right. ( C ) UMAP embeddings of helical peptides with known function appear in distinct regions of the UMAP space. AMPs, which tend to be soluble and negatively charged, cluster in the upper half of the left lobe. Lytic peptides and the C-terminal segments of viral fusion proteins occupy the boundary region where helices from soluble proteins and TM proteins can be found. TM viral fusion domains, which must insert into a TM state to function, can be found at the rightmost edge of the right lobe, implying a strong preference for the TM disposition. Points labeled with circular markers represent individual sequences, while diamonds represent means in the UMAP space over a functional family.

Article Snippet: In particular, the interfacial helical state is not well represented in our PDB files, which are dominated by multispan TM helices.

Techniques: Derivative Assay, Glycoproteomics, Labeling, Functional Assay