81-10 The Mineral within the Protein: Predicting Structural and Functional Properties of Proteins from Mineralogical Scaffolds

Abstract:

Metals within protein active sites can be compared, in geometric terms, to how those same metals are coordinated within a mineral. Ligands around the metal in a protein often reproduce the coordination found in a natural mineral phase. Our previous work on microbial recovery of critical minerals, which pairs computational genomics with an Eh-pH framework built on Pourbaix and Baas Becking relations, led us to ask whether that mineral can serve as a template to predict a protein’s binding affinity, selectivity, and function.

Terrahedral extracts a metal’s donor shell, superimposes it on real mineral coordinates by Kabsch alignment, and scores the geometric strain against the closest mineral. So far, we have applied it across ten metals (Ca, V, Mn, Fe, Co, Ni, Cu, Zn, Mo, W) and the lanthanides, 50+ metalloprotein superfamilies, and 11,832 metal sites, spanning six main roles: electron transfer, hydrolytic and Lewis-acid chemistry, redox and oxygen catalysis, calcium signaling, lanthanide binding, and structural support. A consistent pattern emerged where (1) rigid, catalytically inert sites match their mineral analogs closely, while (2) catalytic and selective sites deviate from them. Because the size and direction of that deviation can track how the metal is used, the strain itself carries information about a site’s function.

We demonstrate the usefulness of this alignment through three examples. In rare-earth systems, we track binding affinity across natural proteins and synthetic peptides, including lanmodulin, whose sites bind lanthanides at picomolar affinity. In manganese systems, we explore how early research into photosystem II could have anticipated its Mn₄CaO₅ cluster from a mineral template. In acireductone dioxygenase, one cupin scaffold cleaves a different carbon-carbon bond depending on whether it holds iron or nickel, cobalt, or manganese, and the closest-mineral match shifts with the metal, showing that metal identity, not the fold, establishes the reaction.

Coordination geometry and the Eh-pH framework together give a route to predict which metals a protein binds, how tightly, and under what conditions, treating a protein metal site as a mineral. These predictions can support selective metal recovery and catalysis, under which designing proteins and peptides to match chosen geometric strain or congruence to a mineral may increase metal specificity and reaction efficiency. 

© Copyright 2026 The Geological Society of America (GSA), all rights reserved.