(GQR93) Buffered by speed, unmasked by stress: how a photosynthetic antenna hides mutational variation

Abstract

Cryptic genetic variation—differences that do nothing in one environment and matter in another—is a central idea in evolutionary genetics, but it is rarely traced to a physical mechanism. We traceone in the Fenna–Matthews–Olson (FMO) protein, the antenna that passes excitation energy to thereaction centre in green sulfur bacteria. The modelling is anchored in experiment: crystal and cryo-EMstructures, Hamiltonians fitted to optical spectra, the spectral shifts of seven site-directed mutants, andmeasured excited-state lifetimes, with transfer and quench rates calibrated to published values. Eachprediction was written down before the calculation that tested it. Three results follow. First, the highest-energy exciton of FMO is almost invisible to the exit pigment (weight 3.7×10−5), because two routesinto the exit cancel. Second, this geometric darkness offers no protection: the protein bath empties thestate in 0.25 ps, five times faster than classical noise of the same strength, and isolating it by energyslows the escape only as the 2.5th power of the gap. Third, under reducing conditions every mutantdelivers within 0.5 percentage points of wild type at all temperatures from −5 to 48 ◦C, because transfer(picoseconds) outruns loss (nanoseconds). Oxidative quenching, which cuts the excited-state lifetime to65 ps, removes that margin: the spread across mutants grows 16- to 35-fold, to 8–17 points dependingon the exit model, and their rank order changes. The direction of the largest effect, a 9–11-point gainfor Y16F in the two structure-based exit models, depends on how the exit is shared between pigments3 and 4. Point-dipole couplings from the FMO–reaction-centre structure put most of it on pigment4, which disagrees with the published charge-based calculation, so we report this sign as unresolved.Across 158 natural FMO sequences, the mutated sites are highly conserved, but conservation does nottrack their effect on energy delivery. The protein therefore buffers variation by a timescale margin, andan environmental change that shrinks the margin exposes it. The reducing-condition differences arebuffered, not neutral: if fitness tracks delivered energy, bacterial population sizes make them likely tobe visible to selection. This is a candidate physical mechanism for gene-by-environment interaction;whether it produces cryptic variation in natural populations, or changes fitness, is untested.