Ancient nitrogenases are ATP dependent

Abstract

Life depends on a conserved set of chemical energy currencies that are relics of early biochemistry. One of these is ATP, a molecule that, when paired with a divalent metal ion such as Mg²⁺, can be hydrolyzed to support numerous cellular and molecular processes. Despite its centrality to extant biochemistry, it is unclear whether ATP supported the function of ancient enzymes. We investigate the evolutionary necessity of ATP by experimentally reconstructing an ancestral variant of the N₂-reducing enzyme nitrogenase. The Proterozoic ancestor is predicted to be ~540-2,300 million years old, post-dating the Great Oxidation Event. Growth rates under nitrogen-fixing conditions are ~80% of those of wild type in Azotobacter vinelandii. In the extant enzyme, the hydrolysis of two MgATP is coupled to electron transfer to support substrate reduction. The ancestor has a strict requirement for ATP with no other nucleotide triphosphate analogs (GTP, ITP, and UTP) supporting activity. Alternative divalent metal ions (Fe²⁺, Co²⁺, and Mn²⁺) support activity with ATP but with diminished activities compared to Mg²⁺, similar to the extant enzyme. Additionally, it is shown that the ancestor has an identical efficiency in ATP hydrolyzed per electron transferred to the extant of two. Our results provide direct laboratory evidence of ATP usage by an ancient enzyme.IMPORTANCELife depends on energy-carrying molecules to power many sustaining processes. There is evidence that these molecules may predate the rise of life on Earth, but how and when these dependencies formed is unknown. The resurrection of ancient enzymes provides a unique tool to probe the enzyme's function and usage of energy-carrying molecules, shedding light on their biochemical origins. Through experimental reconstruction, this research investigates the ancestral dependence of a nitrogen-fixing enzyme on the energy carrier ATP, a requirement for function in the modern enzyme. We show that the resurrected ancestor does not have generalist nucleotide specificity. Rather, the ancestor has a strict requirement for ATP, like the modern enzyme, with similar function and efficiency. The findings elucidate the early-evolved necessity of energy-yielding molecules, delineating their role in ancient biochemical processes. Ultimately, these insights contribute to unraveling the intricate tapestry of evolutionary biology and the origins of life-sustaining dependencies.

Keywords: ancestral sequence reconstruction; energy; nitrogen fixation.

Fig 1

NifH₂ cycle of Mo-nitrogenase. The NifH₂ homodimer is shown in blue; it houses two nucleotide-binding sites and a [4Fe-4S] redox active cluster. One catalytic half (NifDK) of the NifD₂K₂ heterotetramer is shown in yellow (K) and green (D); each catalytic half houses one [8Fe-7S] (P-cluster) and a [7Fe-9S-Mo-C-homocitrate] (FeMo-co) active site metallocluster. When NifH₂ binds to NifDK, an electron is transferred from the [4Fe-4S] cluster to FeMo-co mediated by P-cluster. NifH₂ then hydrolyzes two ATP, dissociates with bound MgADP and an oxidized [4Fe-4S] cluster. MgADP is then exchanged for MgATP and the [4Fe-4S] cluster is reduced, preparing NifH₂ for binding and electron transfer.

Fig 2

Evolution of the ATP/ADP-binding site in nitrogenase NifH. (A) Nitrogenase NifHDK protein phylogeny. Phylogenetic positions of ancestral (Anc^AK029) and extant A. vinelandii (“Extant A. vin”) nitrogenase variants discussed in the main text are highlighted by colored circles. (B) Protein sequence alignment of NifH ATPase signature motifs (labeled) (54) site index from A. vinelandii NifH. (C) Residue-level intermolecular interactions in the ADP-binding site of A. vinelandii NifH (PDB 1FP6). All displayed residues are conserved between Anc^AK029 and WT. Curved lines delineate residues with nonspecific interactions that shape the binding site. (D) Estimated age range of nitrogenase ancestors mapped to Earth’s environmental history. Nitrogenases hosted by heterocystous cyanobacteria provide a minimum age constraint based on oldest fossil evidence (60). Atmospheric oxygenation plot and relative marine metal abundances are from Lyons et al. (61).

Fig 3

Diazotrophic growth of WT and Anc^AK029 A. vinelandii strains. (A) Diazotrophic growth curve of A. vinelandii expressing either the WT or Anc^AK029 nitrogenase proteins. Line plots represent the average ln OD₆₀₀ of three biological replicates per strain, and error bars indicate ±1 SD. (B) Doubling times of WT and Anc^AK029. (C) Midpoint times of WT and Anc^AK029. (B and C) Bars represent the average of three biological replicates per strain, and error bars indicate ±1 SD.

Fig 4

Ancestral nitrogenase behavior with alternative nucleotides and divalent metals. (A) Specific activity for N₂ reduction by extant and Anc^AK029 NifHDK with different nucleotides complexed with Mg²⁺. Shown are percent specific activities for NH₃ formation relative to the maximum activity seen with MgATP. n.d., not detectable (specific activity of Anc^AK029 is ~40% of extant; specific activities are listed in Table S1). (B) Specific activity for N₂ reduction by extant and Anc^AK029 NifHDK with ATP and different divalent metal ions. Shown are percent specific activities for NH₃ formation relative to the maximum activity seen with Mg²⁺ (specific activity of Anc^AK029 is ~40% of extant; specific activities are listed in Table S1). Bars represent the average of three replicates, and error bars represent ±1 standard deviation. (C) Inorganic phosphate released per electron transferred during N₂ reduction. Shown is inorganic phosphate measured normalized to total electrons in products H₂ and NH₃ for extant and Anc^AK029 nitrogenases when allowed to turn over for 2 min under 1 atm of N₂. Bars represent the average of three replicates, and error bars represent ±1 standard deviation. (D) Simplified nitrogenase phylogeny shown in Fig. 2A indicates evolutionary relationship between variants analyzed in panels A–C.