Skip to main content
Wiley Open Access Collection logoLink to Wiley Open Access Collection
. 2026 Oct 9;65(47):e2166480. doi: 10.1002/anie.2166480

Switching ATP‐Forming Phosphoryl‐Transfer Chemistry With a Restricted Set of Amino Acids

Sota Yagi 1,2, Subrata Dasgupta 2, Shunsuke Tagami 2,3, Satoshi Akanuma 1,✉
PMCID: PMC13772221  PMID: 42853966

ABSTRACT

Restricting amino acid diversity is expected to impair enzyme catalysis, but whether such constraints simply diminish catalytic activity or generate distinct catalytic chemistry remains unclear. We reconstructed an ancestral nucleoside diphosphate kinase (NDK) scaffold using 10 prebiotically plausible residues plus lysine and arginine. The resulting variant, Arc1‐12KR, retained the canonical NDK fold in its crystal structure despite lacking the catalytic histidine. Chromatographic and mass spectrometric analyses showed ADP disproportionation to ATP and AMP, an activity undetectable in the NDK scaffold prior to amino acid reduction. Kinetic analysis showed nonlinear dependence on ADP concentration, and docking/molecular dynamics simulations indicated that the putative active‐site region can accommodate two ADP molecules. These structural, computational, and mutational analyses suggest that compositional constraints reshape ATP‐forming phosphoryl‐transfer chemistry through a reorganized active site in which aspartate and arginine, together with coordinated Mg2+, may enable phosphoryl transfer via a noncanonical, histidine‐independent mechanism. Despite modest catalytic efficiency, this activity represents a viable catalytic solution under compositional constraints, demonstrating that ATP formation can be supported by minimal amino acid repertoires. Thus, restricting amino acid diversity can generate distinct yet convergent ATP‐forming catalytic solutions, which may represent chemically plausible early routes to ATP formation under primitive conditions.

Keywords: amino acids, ATP formation, enzymes, phosphoryl transfer, prebiotic chemistry


Constraints on amino acid composition relevant to prebiotic chemistry reshape enzyme chemistry by replacing the canonical mechanism within a conserved protein fold with a distinct ATP‐forming reaction, ADP disproportionation. Such constraints generate alternative catalytic solutions not observed in modern enzymes and point to chemically plausible early routes to ATP formation.

graphic file with name ANIE-65-e2166480-g006.webp

1. Introduction

Understanding how proteins first emerged is a central issue in elucidating the transition from prebiotic chemistry to biological evolution [1, 2, 3]. However, the influence of amino acid compositional constraints on the underlying chemistry of catalysis remains poorly understood. A key challenge in this context is the initial acquisition of amino acid building blocks [4, 5]. Modern proteins are generally composed of 20 L‐amino acids, and the last universal common ancestor (LUCA) is thought to have already used a complete amino acid repertoire [6, 7, 8]. The amino acid composition of proteins prior to LUCA, however, is not well defined [9, 10, 11]. Indeed, to our knowledge, there is no experimental evidence indicating that all 20 of the current proteinogenic amino acids were available during the earliest stages of life. More likely, these amino acids were instead gradually incorporated through the evolution of biosynthetic pathways [12, 13, 14]. This proposal is supported by several lines of evidence. For example, ten amino acids were synthesized in Miller's spark discharge experiment [15, 16, 17], eight were identified in the Murchison meteorite [18, 19], and recent analyses of samples from asteroid Bennu detected 14 proteinogenic amino acids, ten of which overlapped with those from the Miller experiment [20]. These findings suggest that primitive proteins were synthesized using only a limited subset of amino acids that were available on the early Earth, prior to the emergence of endogenous biosynthetic pathways. Specifically, the set of ten amino acids synthesized in Miller's spark discharge experiment are often referred to as “prebiotic amino acids,” because they plausibly represent the amino acid repertoire available on early Earth [12, 21, 22]. However, modern enzymes rely heavily on the chemical diversity of side chains for substrate binding, transition‐state stabilization, and catalysis. This raises a fundamental question as to whether polypeptides composed of a limited set of amino acids could support catalytic mechanisms and reaction specificity. Resolving this question is essential for understanding the emergence of functional enzymatic chemistry from prebiotic systems.

The plausibility of primitive protein synthesis using the prebiotic set of amino acids has been supported by multiple lines of evidence. The set of prebiotic amino acids have been shown to associate with self‐assembled fatty acid membranes, which enhances membrane stability. The localized concentration of these amino acids on the membrane surfaces could have facilitated peptide synthesis [21]. Computational and theoretical studies have further predicted that hypothetical proteins composed of the ten prebiotic amino acids would exhibit folding preferences and structural characteristics similar to those of modern proteins [23]. Moreover, another study suggested that among all possible combinations of ten amino acids drawn from the twenty used in modern proteins, the prebiotic set is particularly favorable for forming robust protein backbones [14]. Collectively, these findings indicate that the prebiotic amino acid subset was not only abundant in the early Earth environment but also inherently suited to forming stable, protein‐like structures.

Nucleoside diphosphate kinase (NDK) catalyzes the transfer of the γ‐phosphate from a nucleoside triphosphate to a nucleoside diphosphate, thereby maintaining the intracellular balance of nucleoside triphosphates [24]. Previously, we performed ancestral sequence reconstruction on NDK to produce several ancestral NDKs that might have existed in the last common ancestors of Archaea and Bacteria [25, 26]. Next, we systematically restricted the amino acid composition of Arc1, one of the archaeal ancestral NDKs, while preserving its catalytic function [27]. This approach yielded simplified proteins composed of only 13 amino acid types that retained detectable enzymatic activity and were thermostable above 70°C. Notably, the majority of amino acids in these simplified proteins overlapped with the aforementioned prebiotic amino acids. Furthermore, when histidine, asparagine, and tyrosine, which are not listed as prebiotic amino acids, were removed from a 13‐amino acid variant, its high thermal stability was retained, although its catalytic activity was completely lost.

In this study, we designed variants with distinct amino acid compositions to probe the effects of compositional constraints on catalytic function. A variant constructed from a 12‐amino‐acid set comprising prebiotic residues supplemented with lysine and arginine exhibited a noncanonical activity that disproportionates ADP yielding ATP and AMP. This activity is not observed in the original NDK prior to amino acid reduction, even as a promiscuous activity. These results indicate that restricting amino acid composition can alter catalytic chemistry and generate an alternative enzymatic activity. ATP synthesis is thought to have been central to early bioenergetics [28], suggesting that this activity could have contributed to primitive energy metabolism. These findings provide experimental evidence that reduced chemical complexity can reshape catalytic chemistry in primitive proteins during the origin and early evolution of life.

2. Results and Discussion

2.1. Emergence of Noncanonical ATP‐Forming Activity in a Reduced Amino Acid Variant

To investigate the relationship between amino acid composition and catalytic function in ancestral proteins, we first constructed two new variants of the reconstructed ancestral NDK Arc1. One of these variants, designated Arc1‐16, was designed with a unique 16‐amino acid alphabet that excluded the residues histidine, asparagine, tyrosine, and cysteine. Histidine, asparagine and tyrosine residues are thought to play important roles in the catalytic activity of Arc1 [27], and cysteine is a residue not present in original Arc1 (Figure S1; Table S1; Data S1). In addition, we constructed a 12‐amino acid variant, Arc1‐12KR, which was composed of the ten prebiotic amino acids supplemented with two basic residues, lysine and arginine (Figure S1; Table S1; Data S1). Note, the N‐terminal Met was excluded from the compositional definition.

Unless otherwise noted, all protein variants were expressed with an N‐terminal His6‐tag. For Arc1‐12KR, we confirmed that the His‐tag did not affect folding or enzymatic activity by comparing the tagged and tag‐free forms using circular dichroism (CD) spectroscopy and ATP‐synthesizing assays (Figure S2), respectively. In addition, the constructs lacking aromatic amino acids were appended with a Gly‐Gly‐Trp tag at the C‐terminus to enable spectroscopic detection of these proteins.

NDK catalyzes the transfer of the γ‐phosphate from a nucleoside triphosphate to a nucleoside diphosphate (Figure 1A). To assess whether the new variants retained this canonical activity, enzyme assays were performed in which ATP production was monitored using GTP and ADP as substrates. Because both variants lacked histidine, a residue shown to be essential for catalysis, none were expected to retain the original enzymatic activity. Consistent with this expectation, Arc1‐16 did not display detectable catalytic activity under these conditions (Figure 1C; Data S2). Intriguingly, however, ATP synthesis activity was detected for Arc1‐12KR (Figure 1C inset). More surprisingly, ATP formation was also observed when GTP, one of the substrates, was omitted from the reaction mixture (Figure 1D; Table S1; Data S2). This result prompted us to hypothesize that Arc1‐12KR was catalyzing an alternative reaction, namely, the formation of ATP and AMP from two molecules of ADP (Figure 1B). To test this hypothesis, we analyzed the reaction products using high‐performance liquid chromatography (HPLC). Prior to incubation, only a single peak corresponding to ADP was observed in the reaction mixture (Figures 1E and S3). After incubation with Arc1‐12KR at 60°C for 10 min, two additional peaks emerged, corresponding to ATP and AMP (Figures 1E and S3). The presence of AMP and ATP together with residual ADP in the reaction mixture after incubation was further demonstrated by ion chromatography–mass spectrometry (IC–MS) analysis (Figure S4; Data S3). These results confirmed that Arc1‐12KR catalyzes the formation of ATP and AMP from two molecules of ADP.

FIGURE 1.

FIGURE 1

Emergence of alternative ATP‐forming activity under conditions of amino acid constraints. (A) Canonical NDK‐catalyzing reaction for the transfer of γ‐phosphate of a nucleoside triphosphate (e.g., GTP) to a nucleoside diphosphate (e.g., ADP). (B) ADP disproportionation reaction catalyzed by Arc1‐12KR and its variants, producing ATP and AMP from two molecules of ADP. (C) Time‐course of ATP formation catalyzed by Arc1, Arc1‐16, and Arc1‐12KR using ADP and GTP as substrates, corresponding to the canonical NDK‐catalyzing reaction shown in panel (A). The inset shows an expanded view highlighting the activities of Arc1‐16 and Arc1‐12KR. (D) Time‐course of ATP production from ADP as the sole substrate by Arc1 and its reduced amino acid set variants. Measurements were performed in triplicate (technical replicates), and data are presented as mean ± SE. (E) HPLC analysis of reaction products obtained after incubating ADP with Arc1‐12KR at 60°C for 10 min, demonstrating the formation of ATP and AMP as predicted by the disproportionation reaction shown in panel (B). A265, absorbance at 265 nm. (F) Initial reaction velocity (v) of ATP formation as a function of ADP concentration for the ADP disproportionation reaction catalyzed by Arc1‐12KR (0.3–5.0 mM ADP). Measurements were performed in triplicate (technical replicates), and data are presented as mean ± SE. Fitting of the data to the Hill equation is shown in Figure S6. (G) Expanded view of the low ADP concentration range (0.3–1.0 mM) shown in panel F. Power‐law fitting yielded an exponent of approximately 1.96. This result is consistent with an apparent second‐order dependence on ADP concentration.

To examine whether this unexpected activity was specific to Arc1‐12KR, we tested the original Arc1 and two 13‐amino acid‐alphabet variants, Arc1‐13FKR and Arc1‐13KMR, in which all phenylalanine residues or all methionine residues present in the original Arc1 sequence were restored to the Arc1‐12KR scaffold, respectively (Figure S1; Table S1; Data S1). When the activity was measured with ADP as the only substrate, no detectable ATP production was observed above background levels for these constructs (Figure 1D; Table S1; Data S2), indicating that the ability to catalyze ADP disproportionation is unique to Arc1‐12KR under the conditions used in this study.

To assess the oligomeric states of Arc1‐12KR and its related variants, we performed analytical gel filtration (Figure S5). Arc1‐13FKR and Arc1‐13KMR eluted at volumes corresponding to their respective hexameric assemblies, similar to the oligomeric state of Arc1 and many extant NDKs [29, 30, 31]. In contrast, Arc1‐12KR eluted as a single peak corresponding to a dimer. Notably, Arc1‐12KR, which uniquely exhibited the altered catalytic activity, existed predominantly as a dimer, whereas the hexameric variants did not display this activity. Structural analysis indicates that the interfaces involved in hexamer formation do not directly contribute to the active site. Therefore, while the change in oligomeric state correlates with the altered activity, a causal relationship between oligomerization and functional divergence cannot be established.

2.2. Noncanonical ATP‐Forming ADP Disproportionation in an NDK Scaffold

The reaction converting two molecules of ADP into ATP and AMP is, among known enzymes, catalyzed by adenylate kinases [32], but it has not been observed in extant NDKs or in the reconstructed ancestral NDK Arc1, even as a detectable promiscuous activity. Thus, within the NDK family, this ATP‐synthesizing activity represents a highly unusual and genuinely novel function. More generally, this finding shows that a conserved protein fold can catalyze a different reaction, while the same reaction can be catalyzed by distinct protein folds. From an evolutionary perspective, this activity may reflect a primitive mode of phosphoryl transfer that predates the appearance of histidine‐containing active sites. A restricted amino acid alphabet lacking histidine, asparagine, tyrosine, and other late‐emerging residues [33] would have constrained conventional catalytic strategies and forced a repurposing of residue side chains such as arginine and aspartate for substrate binding, charge compensation and possibly phosphate transfer. Our results therefore indicate that changes in biochemical function can arise not only from increasing molecular complexity but also from the adaptive exploitation of physicochemical constraints that reprogram catalytic mechanisms and enable the emergence of alternative chemical solutions.

Next, we examined the rate of ATP formation as a function of ADP concentration (Figure 1F; Data S2). The ADP disproportionation reaction approached saturation at ∼1.6 mM ADP. Given that nucleotide substrates are likely to enter the active site as Mg2 +–ADP complexes, depletion of free Mg2 + at high ADP concentrations could account for the apparent saturation behavior. The data could also be fitted to the Hill equation, giving a Hill coefficient of 2.7 (Figure S6). This Hill coefficient suggests that more than one ADP molecule is involved in the ATP‐forming reaction. In the low ADP concentration range (0.3–1.0 mM), the reaction rate determined from the linear phase of ATP formation exhibited an approximately second‐order dependence on ADP concentration (Figure 1G). These kinetic features are consistent with a model in which more than one ADP molecule is involved in a local catalytic event. However, the Hill coefficient should be interpreted with caution because Arc1‐12KR is dimeric and the apparent cooperativity may also reflect interactions between the two protomers or their active sites. Modern NDKs are oligomeric enzymes whose monomers are thought to be functionally independent within the complex. Indeed, differences in the oligomeric state of modern NDKs appear to have little effect on catalytic efficiency [34]. However, this behavior cannot be directly extrapolated to Arc1‐12KR. Therefore, the kinetic data are consistent with noncanonical ADP binding and/or catalytic organization, including a possible direct ADP to ADP phosphoryl transfer mechanism. Nonetheless, the data do not completely rule out other processes such as a ping‐pong‐like mechanism.

We also examined whether Arc1‐12KR is able to catalyze the reverse reaction. Upon incubating the enzyme with ATP and AMP under the standard assay conditions (2 mM Mg2 +), no detectable formation of ADP was observed (Figure S7A). However, increasing the Mg2 + concentration to 10 mM resulted in detectable ADP formation from ATP and AMP (Figure S7B), suggesting that the reverse reaction depends on Mg2 + availability relative to nucleotide concentration. Preferential coordination of Mg2 + by ATP may reduce formation of Mg–AMP complexes under Mg2 +‐limiting conditions, thereby suppressing the reverse reaction. Furthermore, when GDP was used in place of ADP, no reaction products were detected (Figure S7C), indicating that the observed activity is specific to adenylates and does not reflect a generalized NDK activity.

2.3. Dissecting the Roles of Arginine and Lysine in Noncanonical ATP Synthesis

To further dissect the contribution of basic amino acids in the newly observed catalytic activity of Arc1‐12KR, we designed three additional variants: Arc1‐11K, in which all arginine residues were replaced with lysine (Figure S1; Table S1; Data S1); Arc1‐11R, in which all lysine residues were replaced with arginine (Figure S1; Data S1); and Arc1‐10, in which both lysine and arginine residues were replaced with prebiotic amino acids selected from the remaining ten used in Arc1‐12KR (Figure S1; Data S1). The corresponding genes were codon‐optimized for expression in Escherichia coli and artificially synthesized. Despite codon optimization, no expression was detected for either Arc1‐11R or Arc1‐10 in E. coli under our standard expression conditions, suggesting that these variants did not fold into an adequate tertiary structure and were readily degraded. Due to this poor expression, Arc1‐11R and Arc1‐10 could not be purified. In contrast, Arc1‐11K was expressed at high levels and the recombinant protein successfully purified for further analysis.

Gel filtration chromatography revealed that Arc1‐11K, like Arc1‐12KR, exists predominantly as a dimer (Figure S5; Table S1). Far‐UV circular dichroism spectroscopy demonstrated that Arc1‐11K exhibits a secondary structure content highly reminiscent of Arc1‐12KR (Figure S8). These findings indicated that the arginine residues could be substituted for lysine without disrupting the overall protein fold. Catalytic activity assays using ADP as the sole substrate showed that Arc1‐11K retained the ability to catalyze the conversion of two ADP molecules into ATP and AMP (Figure 1D; Table S1; Data S2). However, the rate of reaction was approximately half that of Arc1‐12KR, suggesting arginine contributes more effectively than lysine to this noncanonical catalytic function.

The variant completely lacking basic amino acids (Arc1‐10) failed to fold into a stable structure, highlighting a paradox; namely, although basic residues were probably rare under prebiotic conditions [35], they seem to be essential for proper protein folding and activity. Previous work has shown that primitive proteins can be stabilized in high‐salt environments [22], and that proteins devoid of basic residues can interact with RNA via metal‐ion mediation [36]. Future experiments should clarify how environmental factors, particularly the availability of mono and divalent cations, might have compensated for the absence of basic residues and enabled enzymatic activity in the early Earth environment.

However, we cannot rule out the possibility that lysine, arginine, or both these amino acids were in fact synthesized abiotically and already present in early protein synthesis [37, 38]. Tawfik and coworkers proposed that the first nucleic acid‐binding proteins may have arisen from short, simple sequences containing ornithine, which is not used in extant proteins [39]. By contrast, Thoma and Powner showed that ornithine and diaminobutyric acid rapidly cyclize intramolecularly, blocking further peptide extension, whereas the shortest basic α‐amino acid diaminopropionic acid supports peptide formation and may therefore have substituted for or complemented lysine and arginine during early evolution [40]. This idea is also supported by a previous experimental study comparing proteinogenic and nonproteinogenic cationic amino acids [41]. The study showed that Lys, Arg, and His formed larger oligomers more efficiently than nonproteinogenic analogues such as ornithine and diaminobutyric acid. These proteinogenic amino acids primarily generated linear peptides through the α‐amine, whereas the nonproteinogenic analogues were more prone to intramolecular lactam formation. Thus, the selection of Lys and Arg may have reflected both their functional roles in RNA binding and catalysis as well as their chemical compatibility with stable linear peptide formation.

2.4. Structural Basis of Histidine‐Independent ATP Synthesis Mediated by Arc1‐12KR

To gain insight into the structural basis of the unexpected catalytic activity observed in Arc1‐12KR, we determined the crystal structure of the apo form of this variant at 3.3 Å resolution by x‐ray crystallography (Table S2). The asymmetric unit contained six independent dimeric assemblies (Figure S9A), consistent with the gel filtration data, indicating that the protein predominantly exists as a dimer in solution. Superposition of the six dimers revealed notable conformational heterogeneity within a loop region (residues 49–59) located adjacent to the putative active site (Figure S9B,C). This heterogeneity suggests inherent flexibility at the loop region, which may be relevant to the catalytic mechanism. Except for this flexible motif, the six protomers exhibited highly conserved backbone conformations, indicating that the overall fold is well maintained despite the reduced amino acid alphabet. Next, the crystal structure of Arc1‐12KR was compared with that of the ancestral Arc1 protein (Figure 2A). The two structures aligned closely, with the majority of backbone atoms showing excellent overlap. However significant deviation was observed in the loop formed by residues 49–59, which is located near the active site and likely involved in substrate recognition and positioning (Figure S10).

FIGURE 2.

FIGURE 2

Crystal structure of Arc1‐12KR and its MD‐generated ADP–Mg2+ complex models. (A) Superposition of the Arc1‐12KR dimer (yellow) with the crystal structure of Arc1 (black and gray, PDB code; 3VVT). (B) Model of the Arc1‐12KR–ADP–Mg2+ complex generated by MD simulations. (C) Close‐up view of the ADP‐ and Mg2+‐binding site in the Arc1‐12KR–ADP–Mg2+ complex model. Direct and water‐mediated hydrogen bonds involving W1 and W2 stabilize ADP binding in the catalytic pocket. The Mg2+ ion is directly coordinated by the α‐ and β‐phosphate groups of ADP and the side chain carboxylate group of Asp118, whereas Asp115 is connected to Mg2+ through water‐mediated interactions involving W3 and W4. Additional views focusing on ADP binding and Mg2+ coordination are shown in Figure S12. (D) Overall view of the Arc1‐12KR complex with two ADP molecules and Mg2+ after MD simulations. (E) Close‐up view of the two‐ADP model after MD simulations. ADP_1 occupies the putative ADP‐binding pocket, whereas ADP_2 is located closer to the protein surface near this pocket. Additional views focusing on ADP_1 and ADP_2 are shown in Figure S14A,B, respectively. Dashed lines indicate possible hydrogen bonding, electrostatic, or metal‐coordination interactions. The model suggests that the putative active site can spatially accommodate two ADP molecules.

To directly visualize substrate binding, we attempted cocrystallization and soaking experiments of Arc1‐12KR with ADP and a nonhydrolyzable nucleotide analog, adenosine 5′‐(α,β‐methylene)diphosphate. However, electron density maps revealed no interpretable density corresponding to ADP or its analog in any of the datasets obtained. Instead, a structural model of the Arc1‐12KR–Mg2+–ADP complex was generated using molecular dynamics (MD) simulations (Figure S11; Data S4). Starting from the most “opened” apo conformation observed among the six crystallographically independent dimers, a single ADP molecule was docked into the putative active site. To examine whether a second ADP molecule can be accommodated near this region, a two‐ADP model was also generated and subjected to MD simulations (Data S5). In addition, MD simulations were performed for the apo form of Arc1‐12KR to provide a reference for comparison with the ADP‐bound models (Data S6).

The one‐ADP‐bound Arc1‐12KR model revealed that ADP is located in the canonical binding pocket, with its phosphate groups coordinated by a central Mg2 + ion (Figures 2B,C and S12A). In the modeled active site, the side chain guanidinium groups of Arg85 and Arg102 directly interact with the β‐phosphate group of ADP through hydrogen bonds, indicating that these two arginine residues play important roles in catalysis. Two water molecules (W1 and W2) contribute to ADP binding through a series of hydrogen bonds. The side chain amino group of Lys55 forms a hydrogen bond with the α‐phosphate of ADP via W1. In addition, the hydroxyl group of Thr91 forms a hydrogen bond with ADP via W2. The Mg2 + ion is coordinated by the side chain carboxylate group of Asp118 as well as the α‐ and β‐phosphate groups of ADP. In addition, Asp115 is connected to Mg2 + through water‐mediated interactions involving water molecules W3 and W4 (Figure S12B).

The active sites of the modeled Arc1‐12KR and the Pyrococcus horikoshii OT3 NDK‐ADP‐Mg2 + complex (PDB ID: 2DY9) are compared in Figure 3 and superimposed in Figure S13. Several critical differences between the two active sites were observed. The bound ADP adopts different positions in the active site and, therefore, the phosphorus atom of the β‐phosphate is shifted by ∼1.3 Å, and the position of the coordinated Mg2 + ion differs by ∼2.3 Å between the two structures. Residue 115, which is a catalytically essential histidine in all known NDKs, is replaced by an aspartate in Arc1‐12KR. In canonical NDKs, including that from P. horikoshii, the conserved histidine residue functions as a transient phosphoryl acceptor during catalysis. In Arc1‐12KR, Asp115 is located near the Mg2 +‐binding site and may contribute to the ADP disproportionation reaction. The carboxylate moiety of Asp118 in Arc1‐12KR is also positioned close to the Mg2 + ion, whereas the corresponding aspartate residue in P. horikoshii NDK is oriented away from the active site. In contrast, Arg85 and Arg102 in Arc1‐12KR and the corresponding arginine residues in P. horikoshii NDK are similarly positioned and oriented in the active site. This conservation of relative spatial positioning and orientation suggests that these arginine side chains play comparable roles in phosphate binding. The structural analyses suggest that Arc1‐12KR adopts a distinct His‐independent catalytic strategy primarily derived from alterations to acidic residues and metal coordination, while preserving the canonical phosphate‐binding arginine residues. The active site architecture in Arc1‐12KR reflects adaptation resulting from a restricted amino acid alphabet. These findings demonstrate how simplified sequence information can generate functional enzymatic folds with altered catalytic features.

FIGURE 3.

FIGURE 3

Comparison of the ADP‐binding pockets of the Arc1‐12KR‐ADP‐Mg2+ complex model (left) and the crystal structure of P. horikoshii NDK bound to ADP and Mg2+ (right). The side chains of the four key residues in each structure are also shown.

To further examine whether a second ADP molecule (ADP_2) can be accommodated near the putative active site, we used the one‐ADP‐bound model as a template and docked a second ADP molecule to generate a docked two‐ADP starting structure in which the second ADP was positioned near this region. MD simulations were then performed using this starting structure (Figures 2D,E, S11C, and S14; Data S5). In the resulting two‐ADP model, the binding mode of the first ADP molecule (ADP_1) was largely unchanged. The β‐phosphate group of ADP_1 and the carboxylate group of Asp118 coordinated with the Mg2 + ion, and Arg85 and Arg102 interacted with the β‐phosphate group of ADP_1 (Figure S14A). In addition, Arg85 was also positioned to interact with the β‐phosphate group of ADP_2. The β‐phosphate group of ADP_2 was further bound via hydrogen bonds with the hydroxyl group of Thr91 and the guanidinium group of Arg53 (Figure S14B). These results suggest that the putative active site of Arc1‐12KR can readily accommodate a second ADP molecule.

2.5. Mutational Analysis Identifies Residues Critical for Noncanonical ATP Synthesis

Next, we aimed to identify crucial residues for the newly acquired ATP synthesis activity. Mutants of four candidate residues suggested from analysis of the crystal structure of Arc1‐12KR and its MD‐generated ADP‐bound models were generated. Asp115 and Asp118, both oriented toward the catalytic Mg2 + ion, were individually substituted with alanine. The resultant D115A and D118A mutants retained 60% and 68% of the original ATP synthesizing activity, respectively (Table 1; Figure S15; Data S2). However, simultaneous substitution of both residues drastically reduced this activity to ∼6% (Table 1; Figure S15; Data S2), suggesting that the presence of at least one negatively charged side chain is important for coordinating Mg2 + and for electrostatic contributions to ADP binding.

TABLE 1.

Specific ATP‐synthesizing activity via ADP disproportionation for Arc1‐12KR and its variants at 60°C.

Mutants Specific activity (U/mg) a
Arc1‐12KR 5.0 ± 0.3
D115A 3.0 ± 0.2
D118A 3.4 ± 0.4
D115A/D118A 0.30 ± 0.01
R85M 0.67 ± 0.04
R102M 2.4 ± 0.1
R85M/R102M 0.21 ± 0.01
R53M 0.21 ± 0.01
a

Specific activities are expressed as mean ± SE (n = 3).

Arg85 and Arg102, predicted to interact with the phosphate moieties of ADP, were individually replaced with methionine. R102M maintained approximately 50% of the original ATP synthesizing activity, but R85M resulted in a significant reduction in activity to ∼13% (Table 1; Figure S15; Data S2). The double mutant lacking both arginine residues displayed only ∼4% activity (Table 1; Figure S15; Data S2), indicating the cooperative role of Arg85 and Arg102 in positioning the substrate and stabilizing the transition state. Notably, Arg85 appears to play a more central role than Arg102 in this variant, despite the fact that, in extant NDKs, both Arg85 and Arg102 are identified as catalytically important residues [42].

The two‐ADP model suggested that the guanidinium group of Arg53 interacts with the β‐phosphate group of the second ADP molecule (Figures 2E and S14B). To examine the importance of this residue, we constructed the R53M variant of Arc1‐12KR. This substitution markedly decreased ADP disproportionation activity to approximately 4% that of Arc1‐12KR (Table 1; Figure S16A). In contrast, introduction of the same substitution into Arc1 did not significantly reduce the canonical phosphate transfer activity from GTP to ADP. The specific activities of Arc1 and its R53M variant were 2100 and 1800 U/mg, respectively (Figure S16B). These results suggest that Arg53 is particularly important for the ADP disproportionation reaction of Arc1‐12KR, rather than being generally required for the canonical NDK reaction. This finding also supports the idea that accommodation of a second ADP molecule near the putative active site is important for ATP formation by Arc1‐12KR.

Intriguingly, the variant Arc1‐11K, in which all arginine residues including Arg85 and Arg102 were replaced with lysine, retained 50% of the ATP‐synthesizing activity of Arc1‐12KR (Figure 1D; Table S1; Data S2). Therefore, the catalytic roles of Arg85 and Arg102 can be compensated by lysine, although with reduced efficiency. These findings likely reflect a compensatory redistribution of catalytic roles that emerged under the constraints of being restricted to a prebiotically plausible set of amino acids.

2.6. Mechanistic Basis of Histidine‐Independent Phosphoryl Transfer

Canonical NDKs catalyze γ‐phosphate transfer from NTP to NDP via a ping‐pong mechanism involving a transient phosphohistidine intermediate (Figure 4, left) [43]. This mechanism depends on a conserved histidine residue (His115 in Arc1), which serves as the nucleophile and is transiently phosphorylated during catalysis. By contrast, Arc1‐12KR entirely lacks histidine and catalyzes disproportionation between two ADP molecules to yield ATP and AMP. Although aspartate residues have been reported to form transient phosphoaspartate intermediates in other enzymatic contexts [44, 45], mutational analyses indicate that neither Asp115 nor the nearby Asp118 is individually essential for ATP synthesis in Arc1‐12KR, as alanine substitutions at these positions retain substantial activity (Table 1; Figure S15A). Simultaneous substitution of both residues strongly reduced enzyme activity, but ATP formation remained detectable above the background level (Figure S15B). Therefore, ATP formation in Arc1‐12KR does not appear to strictly require an Asp‐dependent ping‐pong‐like mechanism. However, the activity was markedly decreased by the double Asp substitution, showing that these residues are important for efficient ATP formation. Thus, an Asp‐dependent phosphorylated enzyme intermediate cannot be excluded as one possible part of the catalytic mechanism in Arc1‐12KR. Kinetic analyses also showed that, at low ADP concentrations, the rate of ATP formation has an approximately second‐order dependence on ADP concentration (Figure 1G). This result is consistent with the involvement of two ADP molecules in the ATP‐forming reaction and supports a possible direct ADP to ADP phosphoryl‐transfer model (Figure 4, right). In addition, the R53M substitution in Arc1‐12KR, which was designed based on the two‐ADP model, reduced the ADP disproportionation activity to approximately 4% that of Arc1‐12KR (Table 1; Figure S16A; Data S2), whereas the same substitution in Arc1 did not markedly reduce the canonical phosphate transfer activity from GTP to ADP (Figure S16B; Data S2). This result suggests that Arg53 is very important for ATP formation by Arc1‐12KR and supports the idea that accommodation of the second ADP near the putative active site is crucial for the enzymatic reaction. However, these kinetic and mutational results do not rule out other mechanisms, including an Asp‐dependent ping‐pong‐like mechanism. In conclusion, the direct transfer model remains a possible explanation for the His‐independent ATP‐forming activity of Arc1‐12KR.

FIGURE 4.

FIGURE 4

Possible histidine‐independent phosphoryl transfer model for Arc1‐12KR. In canonical NDKs (left), phosphoryl transfer proceeds through a phosphohistidine intermediate [43]. In the reduced amino acid variant Arc1‐12KR (right), the deprotonated sidechain of Asp118 coordinates with Mg2 +, while Arg85 and Arg102 together with Mg2 + may help position the β‐phosphate of one ADP molecule (ADP_1). Arg53 and Arg85 may additionally contribute to binding of the second ADP molecule (ADP_2). This arrangement represents one possible model for direct phosphoryl transfer between two ADP molecules in the absence of a catalytic histidine residue. However, alternative mechanisms, including formation of a phosphorylated enzyme intermediate, cannot be excluded.

Our structural and mutational analyses suggest that the noncanonical activity relies on cooperative interactions among Mg2 +‐coordinating aspartate and phosphate‐binding arginine residues. Together, these residues may partially substitute for the catalytic role normally played by histidine in canonical NDKs. This occurs by organizing the Mg2 +‐ADP complex, stabilizing the negative charge on the phosphate groups, and facilitating phosphoryl transfer, which does not involve a covalent intermediate. These observations indicate that, when the amino acid alphabet is constrained, alternative catalytic mechanisms can emerge that diverge from those employed by modern enzymes.

2.7. Implications for Early ATP‐Forming Chemistry

The ability of Arc1‐12KR to catalyze a reaction that is absent from its modern homologs suggests that structural plasticity and chemical adaptability may have played key roles in the functional diversification of primordial enzymes. Rather than representing an entirely new catalytic activity, the observed enzymatic function may instead reveal a latent capacity that was accessible on early protein scaffolds but was subsequently lost as more efficient and specialized mechanisms evolved. Our findings demonstrate that simplification of amino acid composition can reveal alternative activities that are no longer retained by modern, fully optimized enzymes.

Consistent with this notion, Arc1‐12KR and Arc1‐11K catalyze the same overall reaction as modern adenylate kinases despite having no detectable homology and entirely distinct tertiary structures. This functional convergence implies that phosphoryl transfer reactions may have arisen independently multiple times during early evolution from nonhomologous protein scaffolds. The canonical NDK reaction, which involves transfer of the γ‐phosphate from NTP to NDP, is nearly universal across extant life. By contrast, the ADP disproportionation reaction characterized here may reflect an older stage of molecular evolution. In environments where NTPs were scarce or absent, a mechanism that uses ADP as both phosphate donor and acceptor could have provided a plausible route for transient ATP formation and energy recycling in prebiotic systems. As illustrated in Figure 5, we envisage a prebiotic world in which kinase‐like proteins resembling Arc1‐12KR/Arc1‐11K with a limited amino acid repertoire generated ATP via ADP disproportionation. With the subsequent expansion of the amino acid alphabet, this chemistry was later superseded by modern NDKs, adenylate kinases, and rotary ATPases.

FIGURE 5.

FIGURE 5

Conceptual model illustrating possible relationships among ATP‐generating enzyme systems. Under prebiotic conditions with a limited amino acid repertoire, kinase‐like proteins resembling Arc1‐12KR or Arc1‐11K may have generated ATP through ADP disproportionation (2ADP → ATP + AMP). As the amino acid alphabet expanded toward the LUCA era, this chemistry was likely superseded by the emergence of modern NDKs, which catalyze phosphoryl transfer between NTPs and NDPs, and by adenylate kinases, which catalyze the reversible reaction 2ADP ⇌ ATP + AMP. Ultimately, the evolution of rotary ATP synthases enabled efficient ATP production from ADP and inorganic phosphate.

The ATP‐forming turnover rate (1.5 s– 1) of Arc1‐12KR was substantially lower than that (550 s– 1) of the original Arc1 enzyme. Thus, the activity of Arc1‐12KR is unlikely to be sufficient for a physiological function comparable to that of modern NDKs or adenylate kinases. Modern adenylate kinases catalyze the same net reaction using a distinct protein fold. In adenylate kinases, phosphoryl transfer proceeds directly between two bound nucleotides and does not involve a covalent phosphorylated‐enzyme intermediate. A phosphoryl transfer step of >500 s– 1 has been reported, and Mg2 + and active site Arg residues have been shown to be important for organizing the phosphate groups and promoting phosphoryl transfer [46]. Arc1‐12KR is different from adenylate kinases in terms of its overall fold and catalytic efficiency. However, Arc1‐12KR also catalyzes ATP formation from two ADP molecules without the catalytic histidine present in canonical NDKs. Mutational analyses of Arc1‐12KR indicate that Arg residues, together with Mg2 +, are important for efficient ATP formation. Thus, although details of the reaction mechanism mediated by Arc1‐12KR remain unresolved, comparison with adenylate kinase suggests that Mg2 + and basic residues may provide common catalytic elements for His‐independent phosphoryl transfer. To place this activity in the context of early ATP‐forming chemistry, it is also useful to compare Arc1‐12KR with simpler phosphoryl transfer polypeptides [47]. Short P‐loop prototype polypeptides reported to mediate the same ADP disproportionation reaction show apparent turnover numbers of approximately 2 × 10−4 s−1. Thus, although the activity of Arc1‐12KR is much lower than that of optimized modern enzymes, it is substantially higher than that of previously reported rudimentary ATP‐forming polypeptides. Arc1‐12KR is therefore best interpreted not as a physiologically optimized enzyme, but as a weak His‐independent ATP‐forming catalyst that may have played a crucial role in early ATP‐forming chemistry.

Although the ADP disproportionation reaction is thermodynamically close to neutral, its evolutionary and metabolic implications could be substantial. Local or transient increases in ATP concentration would have facilitated downstream biosynthetic reactions and early nucleotide‐dependent processes [28]. At the same time, the coproduced AMP could have contributed to nucleotide signaling functions in primitive biochemical networks [32]. In proto‐cellular environments, such reactions, when combined with compartmentalization or ion gradients, may have helped generate and sustain nonequilibrium states, thereby supporting the emergence of proto‐metabolic organization before dedicated ATP synthase systems evolved [48, 49].

3. Conclusion

In this study, we demonstrate that ATP‐forming activity emerges in the reconstructed kinase‐like enzymes Arc1‐12KR and Arc1‐11K even though they are built only from prebiotic and basic amino acids. Despite its modest catalytic efficiency, this activity represents a viable catalytic solution under severe compositional constraints. These results indicate that restricting the amino acid repertoire reshapes phosphoryl transfer chemistry rather than merely reducing canonical activity, thereby uncovering catalytic strategies that are masked in modern enzymes. The emergence of noncanonical ATP‐forming activity demonstrates that such constraints can generate chemically viable solutions for energy conversion even under severely restricted amino acid compositions, recapitulating conditions relevant to early evolution. Our results reveal a many‐to‐many relationship between protein structure and catalytic function, whereby a conserved fold can catalyze different reactions while the same reaction can be catalyzed by distinct folds. These findings suggest that early protein evolution was shaped by physicochemical constraints that defined a minimal chemical space for catalytic function, in which functional innovation and potentially primitive routes to ATP formation could emerge prior to the establishment of fully optimized enzymatic mechanisms.

Author Contributions

Sota Yagi: investigation, methodology, data curation, visualization, writing – review and editing, writing – original draft. Subrata Dasgupta: investigation, methodology, data curation, writing – review and editing, visualization, writing – original draft. Shunsuke Tagami: supervision, writing – review and editing. Satoshi Akanuma: conceptualization, investigation, funding acquisition, writing – original draft, writing – review and editing, data curation, supervision, project administration, methodology, visualization.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Experimental Section, Figures S1−S16, Tables S1−S2, Data S1, Legends for Data S2 to S7 (PDF). Data S2: Data of activity measurements (Excel). Data S3: Numerical IC–MS intensity data before normalization. Data S4: Atomic coordinates files (Arc1‐12KR_1ADP_MD_model.pdb). Data S5: Atomic coordinates files (Arc1‐12KR_2ADP_MD_model.pdb). Data S6: Atomic coordinates files (Arc1‐12KR_apo_MD_model.pdb). Data S7: Data of position‐specific amino acid frequencies (Excel). Supporting File 1: anie75062‐sup‐0001‐SuppMat.pdf.

Supporting File 2: anie75062‐sup‐0002‐S2.xlsx.

Supporting File 3: anie75062‐sup‐0003‐S3.xlsx.

ANIE-65-e2166480-s004.xlsx (886.4KB, xlsx)

Supporting File 4: anie75062‐sup‐0004‐SuppMat.pdb.

Supporting File 5: anie75062‐sup‐0005‐SuppMat.pdb.

Supporting File 6: anie75062‐sup‐0006‐SuppMat.pdb.

Supporting File 7: anie75062‐sup‐0007‐S7.xlsx.

Acknowledgments

This work was supported by Japan Society for the Promotion of Science (JSPS) KAKENHI (Grant numbers 19K21903, 21H01200, 23K20903, and 26K00825) to SA; and by the Astrobiology Center Program of the National Institutes of Natural Sciences (Grant Numbers AB022003, AB031007 and AB0609) to SA.

Data Availability Statement

The data that supports the findings of this study are available in the Supporting Information of this article.

References

  • 1. Frenkel‐Pinter M., Samanta M., Ashkenasy G., and Leman L. J., “Prebiotic Peptides: Molecular Hubs in the Origin of Life,” Chemical Reviews 120 (2020): 4707–4765, 10.1021/acs.chemrev.9b00664. [DOI] [PubMed] [Google Scholar]
  • 2. Hlouchová K., “Peptides En Route From Prebiotic to Biotic Catalysis,” Accounts of Chemical Research 57 (2024): 2027–2037, 10.1021/acs.accounts.4c00137. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3. Ezerzer Y., Frenkel‐Pinter M., Kolodny R., and Ben‐Tal N., “A Building Blocks Perspective on Protein Emergence and Evolution,” Current Opinion in Structural Biology 91 (2025): 102996, 10.1016/j.sbi.2025.102996. [DOI] [PubMed] [Google Scholar]
  • 4. Kitadai N. and Maruyama S., “Origins of Building Blocks of Life: A Review,” Geoscience Frontiers 9 (2018): 1117–1153, 10.1016/j.gsf.2017.07.007. [DOI] [Google Scholar]
  • 5. Cleaves H. J. 2nd, “The Origin of the Biologically Coded Amino Acids,” Journal of Theoretical Biology 263 (2010): 490–498, 10.1016/j.jtbi.2009.12.014. [DOI] [PubMed] [Google Scholar]
  • 6. Mat W. K., Xue H., and Wong J. T., “The Genomics of LUCA,” Frontiers in Bioscience 13 (2008): 5605–5613, 10.2741/3103. [DOI] [PubMed] [Google Scholar]
  • 7. Hendrickson T. L., Wood W. N., and Rathnayake U. M., “Did Amino Acid Side Chain Reactivity Dictate the Composition and Timing of Aminoacyl‐tRNA Synthetase Evolution?,” Genes 12 (2021): 409, 10.3390/genes12030409. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8. Ilardo M., Bose R., Meringer M., et al., “Adaptive Properties of the Genetically Encoded Amino Acid Alphabet are Inherited From Its Subsets,” Scientific Reports 9 (2019): 12468, 10.1038/s41598-019-47574-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9. Wehbi S., Wheeler A., Morel B., et al., “Order of Amino Acid Recruitment Into the Genetic Code Resolved by Last Universal Common Ancestor's Protein Domains,” Proceedings of the National Academy of Sciences of the United States of America 121 (2024): e2410311121, 10.1073/pnas.2410311121. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10. Philip G. K. and Freeland S. J., “Did Evolution Select a Nonrandom “Alphabet” of Amino Acids?,” Astrobiology 11 (2011): 235–240, 10.1089/ast.2010.0567. [DOI] [PubMed] [Google Scholar]
  • 11. Kirschning A., “On the Evolutionary History of the Twenty Encoded Amino Acids,” Chemistry—A European Journal 28 (2022): e202201419, 10.1002/chem.202201419. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12. Higgs P. G. and Pudritz R. E., “A Thermodynamic Basis for Prebiotic Amino Acid Synthesis and the Nature of the First Genetic Code,” Astrobiology 9 (2009): 483–490, 10.1089/ast.2008.0280. [DOI] [PubMed] [Google Scholar]
  • 13. Tretyachenko V., Vymětal J., Neuwirthová T., Vondrášek D., Fujishima K., and Hlouchová K., “Modern and Prebiotic Amino Acids Support Distinct Structural Profiles in Proteins,” Open Biology 12 (2022): 220040, 10.1098/rsob.220040. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14. Solis A. D., “Reduced Alphabet of Prebiotic Amino Acids Optimally Encodes the Conformational Space of Diverse Extant Protein Folds,” BMC Evolutionary Biology 19 (2019): 158, 10.1186/s12862-019-1464-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15. Ferus M., Pietrucci F., Saitta A. M., et al., “Formation of Nucleobases in a Miller–Urey Reducing Atmosphere,” Proceedings of the National Academy of Sciences of the United States of America 114 (2017): 4306–4311, 10.1073/pnas.1700010114. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16. Weber A. L. and Miller S. L., “Reasons for the Occurrence of the Twenty Coded Protein Amino Acids,” Journal of Molecular Evolution 17 (1981): 273–284, 10.1007/BF01795749. [DOI] [PubMed] [Google Scholar]
  • 17. Miller S. L., “A Production of Amino Acids Under Possible Primitive Earth Conditions,” Science 117 (1953): 528–529, 10.1126/science.117.3046.528. [DOI] [PubMed] [Google Scholar]
  • 18. Cronin J. R., “Origin of Organic Compounds in Carbonaceous Chondrites,” Advances in Space Research 9 (1989): 59–64, 10.1016/0273-1177(89)90364-5. [DOI] [PubMed] [Google Scholar]
  • 19. Cronin J. R. and Pizzarello S., “Amino Acids in Meteorites,” Advances in Space Research 3 (1983): 5–18, 10.1016/0273-1177(83)90036-4. [DOI] [PubMed] [Google Scholar]
  • 20. Glavin D. P., Dworkin J. P., Alexander C. M. O. D., et al., “Abundant Ammonia and Nitrogen‐Rich Soluble Organic Matter in Samples From Asteroid (101955) Bennu,” Nature Astronomy 9 (2025): 199–210, 10.1038/s41550-024-02472-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21. Cornell C. E., Black R. A., Xue M., et al., “Prebiotic Amino Acids Bind to and Stabilize Prebiotic Fatty Acid Membranes,” Proceedings of the National Academy of Sciences of the United States of America 116 (2019): 17239–17244, 10.1073/pnas.1900275116. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22. Longo L. M., Lee J., and Blaber M., “Simplified Protein Design Biased for Prebiotic Amino Acids Yields a Foldable, Halophilic Protein,” Proceedings of the National Academy of Sciences of the United States of America 110 (2013): 2135–2139, 10.1073/pnas.1219530110. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23. Ángyán A. F., Ortutay C., and Gáspári Z., “Are Proposed Early Genetic Codes Capable of Encoding Viable Proteins?,” Journal of Molecular Evolution 78 (2014): 263–274, 10.1007/s00239-014-9622-3. [DOI] [PubMed] [Google Scholar]
  • 24. Deville‐Bonne D., Sellam O., Merola F., Lascu I., Desmadril M., and Véron M., “Phosphorylation of Nucleoside Diphosphate Kinase at the Active Site Studied by Steady‐State and Time‐Resolved Fluorescence,” Biochemistry 35 (1996): 14643–14650, 10.1021/bi960945m. [DOI] [PubMed] [Google Scholar]
  • 25. Akanuma S., Yokobori S., Nakajima Y., Bessho M., and Yamagishi A., “Robustness of Predictions of Extremely Thermally Stable Proteins in Ancient Organisms,” Evolution; Internation Journal of Organic Evolution 69 (2015): 2954–2962, 10.1111/evo.12779. [DOI] [PubMed] [Google Scholar]
  • 26. Akanuma S., Nakajima Y., Yokobori S.‐I., et al., “Experimental Evidence for the Thermophilicity of Ancestral Life,” Proceedings of the National Academy of Sciences of the United States of America 110 (2013): 11067–11072, 10.1073/pnas.1308215110. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27. Shibue R., Sasamoto T., Shimada M., Zhang B., Yamagishi A., and Akanuma S., “Comprehensive Reduction of Amino Acid Set in a Protein Suggests the Importance of Prebiotic Amino Acids for Stable Proteins,” Scientific Reports 8 (2018): 1227, 10.1038/s41598-018-19561-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28. Pinna S., Kunz C., Halpern A., et al., “A Prebiotic Basis for ATP as the Universal Energy Currency,” PLoS Biology 20 (2022): e3001437, 10.1371/journal.pbio.3001437. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29. Misra G., Aggarwal A., Dube D., Zaman M. S., Singh Y., and Ramachandran R., “Crystal Structure of the Bacillus anthracis Nucleoside Diphosphate Kinase and Its Characterization Reveals an Enzyme Adapted to Perform Under Stress Conditions,” Proteins: Structure, Function, and Bioinformatics 76 (2009): 496–506, 10.1002/prot.22364. [DOI] [PubMed] [Google Scholar]
  • 30. Moréra S., LeBras G., Lascu I., Lacornbe M. L., Véron M., and Janin J., “Refined X‐Ray Structure of Dictyostelium discoideum Nucleoside Diphosphate Kinase at 1.8 Å Resolution,” Journal of Molecular Biology 243 (1994): 873–890, 10.1006/jmbi.1994.1689. [DOI] [PubMed] [Google Scholar]
  • 31. Webb P. A., Perisic O., Mendola C. E., Backer J. M., and Williams R. L., “The Crystal Structure of Human Nucleoside Diphosphate Kinase, NM23‐H2,” Journal of Molecular Biology 251 (1995): 574–587, 10.1006/jmbi.1995.0457. [DOI] [PubMed] [Google Scholar]
  • 32. Dzeja P. and Terzic A., “Adenylate Kinase and AMP Signaling Networks: Metabolic Monitoring, Signal Communication and Body Energy Sensing,” International Journal of Molecular Sciences 10 (2009): 1729–1772, 10.3390/ijms10041729. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33. Trifonov E. N., “Consensus Temporal Order of Amino Acids and Evolution of the Triplet Code,” Gene 261 (2000): 139–151, 10.1016/S0378-1119(00)00476-5. [DOI] [PubMed] [Google Scholar]
  • 34. Georgescauld F., Song Y., and Dautant A., “Structure, Folding and Stability of Nucleoside Diphosphate Kinases,” International Journal of Molecular Sciences 21 (2020): 6779, 10.3390/ijms21186779. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35. McDonald G. D. and Storrie‐Lombardi M. C., “Biochemical Constraints in a Protobiotic Earth Devoid of Basic Amino Acids: The “BAA(‐) World”,” Astrobiology 10 (2010): 989–1000, 10.1089/ast.2010.0484. [DOI] [PubMed] [Google Scholar]
  • 36. Giacobelli V. G., Fujishima K., Lepšík M., et al., “In Vitro Evolution Reveals Noncationic Protein–RNA Interaction Mediated by Metal Ions,” Molecular Biology and Evolution 39 (2022): msac032, 10.1093/molbev/msac032. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37. Bizzarri B. M., Santalucia R., Mancin E., et al., “Synthesis of Peptides, Peptide Nucleic Acid Components, and Other Biogenic Compounds From a Three‐Component Prebiotic Mixture,” Chemistry—A European Journal 32 (2026): e02419, 10.1002/chem.202502419. [DOI] [PubMed] [Google Scholar]
  • 38. Patel B. H., Percivalle C., Ritson D. J., Duffy C. D., and Sutherland J. D., “Common Origins of RNA, Protein and Lipid Precursors in a Cyanosulfidic Protometabolism,” Nature Chemistry 7 (2015): 301–307, 10.1038/nchem.2202. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39. Longo L. M., Despotović D., Weil‐Ktorza O., et al., “Primordial Emergence of a Nucleic Acid‐Binding Protein via Phase Separation and Statistical Ornithine‐to‐Arginine Conversion,” Proceedings of the National Academy of Sciences of the United States of America 117 (2020): 15731–15739, 10.1073/pnas.2001989117. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40. Thoma B. and Powner M. W., “Selective Synthesis of Lysine Peptides and the Prebiotically Plausible Synthesis of Catalytically Active Diaminopropionic Acid Peptide Nitriles in Water,” Journal of the American Chemical Society 145 (2023): 3121–3130, 10.1021/jacs.2c12497. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41. Frenkel‐Pinter M., Haynes J. W., C M., et al., “Selective Incorporation of Proteinaceous Over Nonproteinaceous Cationic Amino Acids in Model Prebiotic Oligomerization Reactions,” Proceedings of the National Academy of Sciences of the United States of America 116 (2019): 16338–16346, 10.1073/pnas.1904849116. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42. Moréra S., Chiadmi M., LeBras G., Lascu I., and Janin J., “Mechanism of Phosphate Transfer by Nucleoside Diphosphate Kinase: X‐Ray Structures of the Phosphohistidine Intermediate of the Enzymes From Drosophila and Dictyostelium ,” Biochemistry 34 (1995): 11062–11070. [PubMed] [Google Scholar]
  • 43. Schaertl S., Konrad M., and Geeves M. A., “Substrate Specificity of Human Nucleoside‐Diphosphate Kinase Revealed by Transient Kinetic Analysis,” Journal of Biological Chemistry 273 (1998): 5662–5669, 10.1074/jbc.273.10.5662. [DOI] [PubMed] [Google Scholar]
  • 44. Piserchio A., Isiorho E. A., Abzalimov R., Dalby K. N., and Ghose R., “Phosphorylation of a Conserved Aspartate in the Catalytic Site of Eukaryotic Elongation Factor 2 Kinase,” Protein Science 35 (2026): e70442, 10.1002/pro.70442. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45. Shin D. H., Roberts A., Jancarik J., et al., “Crystal Structure of a Phosphatase With a Unique Substrate Binding Domain From Thermotoga maritima ,” Protein Science 12 (2003): 1464–1472, 10.1110/ps.0302703. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46. Kerns S. J., Agafonov R. V., Cho Y.‐J., et al., “The Energy Landscape of Adenylate Kinase During Catalysis,” Nature Structural & Molecular Biology 22 (2015): 124–131, 10.1038/nsmb.2941. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47. Vyas P., Malitsky S., Itkin M., and Tawfik D. S., “On the Origins of Enzymes: Phosphate‐Binding Polypeptides Mediate Phosphoryl Transfer to Synthesize Adenosine Triphosphate,” Journal of the American Chemical Society 145 (2023): 8344–8354, 10.1021/jacs.2c08636. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48. Lane N. and Martin W. F., “The Origin of Membrane Bioenergetics,” Cell 151 (2012): 1406–1416, 10.1016/j.cell.2012.11.050. [DOI] [PubMed] [Google Scholar]
  • 49. Lane N., “Proton Gradients at the Origin of Life,” BioEssays 39 (2017): 1600217, 10.1002/bies.201600217. [DOI] [PubMed] [Google Scholar]
  • 50. Pace C. N., Vajdos F., Fee L., Grimsley G., and Gray T., “How to Measure and Predict the Molar Absorption Coefficient of a Protein,” Protein Science 4 (1995): 2411–2423, 10.1002/pro.5560041120. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 51. Gill S. C. and von Hippel P. H., “Calculation of Protein Extinction Coefficients From Amino Acid Sequence Data,” Analytical Biochemistry 182 (1989): 319–326, 10.1016/0003-2697(89)90602-7. [DOI] [PubMed] [Google Scholar]
  • 52. Hirata K., Yamashita K., Ueno G., et al., “ ZOO: An Automatic Data‐Collection System for High‐Throughput Structure Analysis in Protein Microcrystallography,” Acta Crystallographica Section D: Structural Biology 75 (2019): 138–150, 10.1107/S2059798318017795. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 53. Ito S., Ueno G., and Yamamoto M., “DeepCentering: Fully Automated Crystal Centering Using Deep Learning for Macromolecular Crystallography,” Journal of Synchrotron Radiation 26 (2019): 1361–1366, 10.1107/S160057751900434X. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 54. Nakamura Y., Baba S., Mizuno N., et al., “Computer‐Controlled Liquid‐Nitrogen Drizzling Device for Removing Frost From Cryopreserved Crystals,” Acta Crystallographica Section F: Structural Biology Communications 76 (2020): 616–622. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 55. Okazaki N., Hasegawa K., Ueno G., Murakami H., Kumasaka T., and Yamamoto M., “Mail‐in Data Collection at SPring‐8 Protein Crystallography Beamlines,” Journal of Synchrotron Radiation 15 (2008): 288–291, 10.1107/S0909049507064679. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 56. Yamashita K., Hirata K., and Yamamoto M., “KAMO: Towards Automated Data Processing for Microcrystals,” Acta Crystallographica Section D: Structural Biology 74 (2018): 441–449, 10.1107/S2059798318004576. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 57. Kabsch W., “XDS,” Acta Crystallographica Section D: Biological Crystallography 66 (2010): 125–132, 10.1107/S0907444909047337. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 58. Adams P. D., Afonine P. V., Bunkóczi G., et al., “PHENIX: A Comprehensive Python‐Based System for Macromolecular Structure Solution,” Acta Crystallographica Section D: Biological Crystallography 66 (2010): 213–221, 10.1107/S0907444909052925. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 59. Mirdita M., Schütze K., Moriwaki Y., Heo H., Ovchinnikov S., and Steinegger M., “ColabFold: Making Protein Folding Accessible to All,” Nature Methods 19 (2022): 679–682, 10.1038/s41592-022-01488-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 60. Emsley P. and Cowtan K., “Coot: Model‐Building Tools for Molecular Graphics,” Acta Crystallographica Section D: Biological Crystallography 60 (2004): 2126–2132, 10.1107/S0907444904019158. [DOI] [PubMed] [Google Scholar]
  • 61. Snyder H. D. and Kucukkal T. G., “Computational Chemistry Activities With Avogadro and ORCA,” Journal of Chemical Education 98 (2021): 1335–1341, 10.1021/acs.jchemed.0c00959. [DOI] [Google Scholar]
  • 62. Neese F., Wennmohs F., Becker U., and Riplinger C., “The ORCA Quantum Chemistry Program Package,” Journal of Chemical Physics 152 (2020): 224108, 10.1063/5.0004608. [DOI] [PubMed] [Google Scholar]
  • 63. Eberhardt J., Santos‐Martins D., Tillack A. F., and Forli S., “AutoDock Vina 1.2.0: New Docking Methods, Expanded Force Field, and Python Bindings,” Journal of Chemical Information and Modeling 61 (2021): 3891–3898, 10.1021/acs.jcim.1c00203. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 64. Honorato R. V., Trellet M. E., Jiménez‐García B., et al., “The HADDOCK2.4 Web Server for Integrative Modeling of Biomolecular Complexes,” Nature Protocols 19 (2024): 3219–3241, 10.1038/s41596-024-01011-0. [DOI] [PubMed] [Google Scholar]
  • 65. Van Der Spoel D., Lindahl E., and Hess B., “GROMACS: Fast, Flexible, and Free,” Journal of Computational Chemistry 26 (2005): 1701–1718. [DOI] [PubMed] [Google Scholar]
  • 66. Brooks B. R., Bruccoleri R. E., Olafson B. D., States D. J., Swaminathan S., and Karplus M., “CHARMM: A Program for Macromolecular Energy, Minimization, and Dynamics Calculations,” Journal of Computational Chemistry 4 (1983): 187–217, 10.1002/jcc.540040211. [DOI] [Google Scholar]
  • 67. Hess B., Bekker H., Berendsen H. J. C., and Fraaije J. G. E. M., “LINCS: A Linear Constraint Solver for Molecular Simulations,” Journal of Computational Chemistry 18 (1997): 1463–1472, 10.1002/(SICI)1096-987X(199709)18:12<1463::AID-JCC4>3.0.CO;2-H. [DOI] [Google Scholar]
  • 68. Harvey M. J. and De Fabritiis G., “An Implementation of the Smooth Particle Mesh Ewald Method on GPU Hardware,” Journal of Chemical Theory and Computation 5 (2009): 2371–2377, 10.1021/ct900275y. [DOI] [PubMed] [Google Scholar]
  • 69. Dasgupta S. and Bhaumik P., “Unraveling the Molecular Architecture of Mosquito D1‐Like Dopamine Receptors: Insights Into Ligand Binding and Structural Dynamics for Insecticide Development,” Proteins: Structure, Function, and Bioinformatics 93 (2025): 1157–1170, 10.1002/prot.26783. [DOI] [PubMed] [Google Scholar]
  • 70. Mukherjee S., Dasgupta S., Panja S. S., and Adhikari U., “Structural Insight to Human Retinoid X Receptor Alpha‐Thyroid Hormone Receptor Beta Heterodimer by Molecular Modelling and MD‐Simulation Studies: Role of Conserved Water Molecules,” Journal of Biomolecular Structure and Dynamics 41 (2023): 9828–9839, 10.1080/07391102.2022.2147097. [DOI] [PubMed] [Google Scholar]
  • 71. Dasgupta S., Mukherjee S., Mukhopadhyay B. P., Banerjee A., and Mishra D. K., “Recognition Dynamics of Dopamine to Human Monoamine Oxidase B: Role of Leu171/Gln206 and Conserved Water Molecules in the Active Site Cavity,” Journal of Biomolecular Structure and Dynamics 36 (2018): 1439–1462, 10.1080/07391102.2017.1325405. [DOI] [PubMed] [Google Scholar]
  • 72. Humphrey W., Dalke A., and Schulten K., “VMD: Visual Molecular Dynamics,” Journal of Molecular Graphics 14 (1996): 33–38, 10.1016/0263-7855(96)00018-5, 27‐28. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Experimental Section, Figures S1−S16, Tables S1−S2, Data S1, Legends for Data S2 to S7 (PDF). Data S2: Data of activity measurements (Excel). Data S3: Numerical IC–MS intensity data before normalization. Data S4: Atomic coordinates files (Arc1‐12KR_1ADP_MD_model.pdb). Data S5: Atomic coordinates files (Arc1‐12KR_2ADP_MD_model.pdb). Data S6: Atomic coordinates files (Arc1‐12KR_apo_MD_model.pdb). Data S7: Data of position‐specific amino acid frequencies (Excel). Supporting File 1: anie75062‐sup‐0001‐SuppMat.pdf.

Supporting File 2: anie75062‐sup‐0002‐S2.xlsx.

Supporting File 3: anie75062‐sup‐0003‐S3.xlsx.

ANIE-65-e2166480-s004.xlsx (886.4KB, xlsx)

Supporting File 4: anie75062‐sup‐0004‐SuppMat.pdb.

Supporting File 5: anie75062‐sup‐0005‐SuppMat.pdb.

Supporting File 6: anie75062‐sup‐0006‐SuppMat.pdb.

Supporting File 7: anie75062‐sup‐0007‐S7.xlsx.

Data Availability Statement

The data that supports the findings of this study are available in the Supporting Information of this article.


Articles from Angewandte Chemie (International Ed. in English) are provided here courtesy of Wiley

RESOURCES