Resurrecting Full-length Ancestral Schizorhodopsins and Heliorhodopsins with Structure-guided, Indel-aware Sequence Reconstruction

Microbial rhodopsins exhibit diverse functions ranging from ion pumps and channels to light sensors, despite sharing a common seven-transmembrane (7TM) architecture. Understanding how this functional diversity evolved is a long-standing problem, and ancestral sequence reconstruction (ASR) offers a direct route to inferring and experimentally testing plausible ancestral rhodopsins. However, ASR of 7TM proteins is often limited by alignment ambiguity and insertion-deletion (indel) uncertainty, especially in extra-membrane (EM) loops and termini. As a result, many studies focus on trimmed transmembrane (TM) cores and treat EM regions by manual curation, leaving the evolutionary history of full-length architecture difficult to test experimentally. Here we reconstruct and resurrect full-length ancestral schizorhodopsins (Anc-SzR) and heliorhodopsins (Anc-HeR), two microbial rhodopsin families that share a retinal-binding 7TM core but differ in membrane topology and EM secondary-structure elements. Starting from untrimmed alignments, we combine structure-consistent multiple sequence alignments and profile-based evolutionary models with an explicit indel-aware refinement that merges amino-acid ancestral states with binary ancestral gap inference on a fixed topology. Indel-aware refinement prevents artificially overextended ancestors and yields compact full-length sequences. AlphaFold3 predictions for the indel-corrected ancestors support high-confidence 7TM folds and recover lineage-specific EM features, including characteristic {beta}-strands and short helices. Both Anc-SzR and Anc-HeR can be ex-pressed in Escherichia coli and recovered as stable, colored, retinal-binding holoproteins. In a whole-cell pH assay, Anc-SzR shows light-driven proton-transport activity, whereas Anc-HeR shows no detectable ion-pumping signal, consistent with extant heliorhodopsins. Together, these results show that full-length, indel-aware ASR can produce experimentally tractable ancestral microbial rhodopsins and enable direct tests of how EM architecture evolves alongside the 7TM core.