Inhibitory to non-inhibitory evolution of the ζ subunit of the F1FO-ATPase of Paracoccus denitrificans and α-proteobacteria as related to mitochondrial endosymbiosis.

Introduction: The ζ subunit is a potent inhibitor of the F1FO-ATPase of Paracoccus denitrificans (PdF1FO-ATPase) and related α-proteobacteria different from the other two canonical inhibitors of bacterial (ε) and mitochondrial (IF1) F1FO-ATPases. ζ mimics mitochondrial IF1 in its inhibitory N-terminus, blocking the PdF1FO-ATPase activity as a unidirectional pawl-ratchet and allowing the PdF1FO-ATP synthase turnover. ζ is essential for the respiratory growth of P. denitrificans, as we showed by a Δζ knockout. Given the vital role of ζ in the physiology of P. denitrificans, here, we assessed the evolution of ζ across the α-proteobacteria class. Methods: Through bioinformatic, biochemical, molecular biology, functional, and structural analyses of several ζ subunits, we confirmed the conservation of the inhibitory N-terminus of ζ and its divergence toward its C-terminus. We reconstituted homologously or heterologously the recombinant ζ subunits from several α-proteobacteria into the respective F-ATPases, including free-living photosynthetic, facultative symbiont, and intracellular facultative or obligate parasitic α-proteobacteria. Results and discussion: The results show that ζ evolved, preserving its inhibitory function in free-living α-proteobacteria exposed to broad environmental changes that could compromise the cellular ATP pools. However, the ζ inhibitory function was diminished or lost in some symbiotic α-proteobacteria where ζ is non-essential given the possible exchange of nutrients and ATP from hosts. Accordingly, the ζ gene is absent in some strictly parasitic pathogenic Rickettsiales, which may obtain ATP from the parasitized hosts. We also resolved the NMR structure of the ζ subunit of Sinorhizobium meliloti (Sm-ζ) and compared it with its structure modeled in AlphaFold. We found a transition from a compact ordered non-inhibitory conformation into an extended α-helical inhibitory N-terminus conformation, thus explaining why the Sm-ζ cannot exert homologous inhibition. However, it is still able to inhibit the PdF1FO-ATPase heterologously. Together with the loss of the inhibitory function of α-proteobacterial ε, the data confirm that the primary inhibitory function of the α-proteobacterial F1FO-ATPase was transferred from ε to ζ and that ζ, ε, and IF1 evolved by convergent evolution. Some key evolutionary implications on the endosymbiotic origin of mitochondria, as most likely derived from α-proteobacteria, are also discussed.

Evolution of the Inhibitory and Non-Inhibitory ε, ζ, and IF1 Subunits of the F1FO-ATPase as Related to the Endosymbiotic Origin of Mitochondria.

The F1FO-ATP synthase nanomotor synthesizes >90% of the cellular ATP of almost all living beings by rotating in the "forward" direction, but it can also consume the same ATP pools by rotating in "reverse." To prevent futile F1FO-ATPase activity, several different inhibitory proteins or domains in bacteria (ε and ζ subunits), mitochondria (IF1), and chloroplasts (ε and γ disulfide) emerged to block the F1FO-ATPase activity selectively. In this study, we analyze how these F1FO-ATPase inhibitory proteins have evolved. The phylogeny of the α-proteobacterial ε showed that it diverged in its C-terminal side, thus losing both the inhibitory function and the ATP-binding/sensor motif that controls this inhibition. The losses of inhibitory function and the ATP-binding site correlate with an evolutionary divergence of non-inhibitory α-proteobacterial ε and mitochondrial δ subunits from inhibitory bacterial and chloroplastidic ε subunits. Here, we confirm the lack of inhibitory function of wild-type and C-terminal truncated ε subunits of P. denitrificans. Taken together, the data show that ζ evolved to replace ε as the primary inhibitor of the F1FO-ATPase of free-living α-proteobacteria. However, the ζ inhibitory function was also partially lost in some symbiotic α-proteobacteria and totally lost in some strictly parasitic α-proteobacteria such as the Rickettsiales order. Finally, we found that ζ and IF1 likely evolved independently via convergent evolution before and after the endosymbiotic origin mitochondria, respectively. This led us to propose the ε and ζ subunits as tracer genes of the pre-endosymbiont that evolved into the actual mitochondria.