Conversion of pyridoxal to pyridoxamine with NH ₃ and H ₂ on nickel generates a protometabolic nitrogen shuttle under serpentinizing conditions

Abstract

Serpentinizing hydrothermal vents are likely sites for the origin of metabolism because they produce H₂ as a source of electrons for CO₂ reduction while depositing zero‐valent iron, cobalt, and nickel as catalysts for organic reactions. Recent work has shown that solid‐state nickel can catalyze the H₂‐dependent reduction of CO₂ to various organic acids and their reductive amination with H₂ and NH₃ to biological amino acids under the conditions of H₂‐producing hydrothermal vents and that amino acid synthesis from NH₃, H₂, and 2‐oxoacids is facile in the presence of Ni⁰. Such reactions suggest a metallic origin of metabolism during early biochemical evolution because single metals replace the function of over 130 enzymatic reactions at the core of metabolism in microbes that use the acetyl‐CoA pathway of CO₂ fixation. Yet solid‐state catalysts tether primordial amino synthesis to a mineral surface. Many studies have shown that pyridoxal catalyzes transamination reactions without enzymes. Here we show that pyridoxamine, the NH₂‐transferring intermediate in pyridoxal‐dependent transamination reactions, is generated from pyridoxal by reaction with NH₃ (as little as 5 mm) and H₂ (5 bar) on Ni⁰ as catalyst at pH 11 and 80 °C within hours. These conditions correspond to those in hydrothermal vents undergoing active serpentinization. The results indicate that at the origin of metabolism, pyridoxamine provided a soluble, organic amino donor for aqueous amino acid synthesis, mediating an evolutionary transition from NH₃‐dependent amino acid synthesis on inorganic surfaces to pyridoxamine‐dependent organic reactions in the aqueous phase.

Recent studies show that Ni⁰ catalyzes amino acid synthesis from 2‐oxoacids via reductive amination. We show that the cofactor pyridoxal is reductively aminated to pyridoxamine, the NH₂‐transferring intermediate of transaminations, with NH₃ (~ 5 mm) and H₂ (5 bar) on Ni⁰ catalysts within hours in water under serpentinizing hydrothermal vent conditions (pH 11 and 80 °C). At metabolic origin, this hybrid catalytic shuttle connects solid‐state catalysts to aqueous phase pyridoxamine‐dependent transaminations.

Abbreviations

PL

pyridoxal

PLP

pyridoxal 5′‐phosphate

PM

pyridoxamine

PMP

pyridoxamine 5′‐phosphate

Introduction

Nitrogen is essential to life. By dry weight, modern cells consist of ~ 55% protein and ~ 20% nucleic acids at the level of polymers [1], with ~ 50% carbon and ~ 10% nitrogen by elements [2]. At life's origin, amino acids and bases had to be supplied by prebiotic reactions without the help of enzymes, but how? The two main competing theories for how C and N entered prebiotic chemistry are genetics first and metabolism first. Genetics‐first theories posit that highly reactive cyanide or alkyl cyanides (nitriles) generated from meteorite impacts [3] were the source of nucleic acid bases [4, 5], amino acids [6, 7], and more recently some cofactors [8], with both C and N entering prebiotic chemistry via nitrile bonds. Nitrile moieties do not, however, occur in any substrate or product among the 400 reactions that generate amino acids, bases, and cofactors in the biochemistry of cells [9], such that genetics‐first theories do not speak to the origin of reactants, catalysts, or compounds in metabolism. Metabolism first theories hold that C and N entered prebiotic chemistry along reaction routes that strongly resemble some, but not all, modern enzymatic pathways, that inorganic surfaces and cofactors served as catalysts at origins [10], and that the first organisms were autotrophs that obtained their carbon from CO₂ [11, 12, 13, 14].

For carbon, a coherent picture of primordial CO₂ assimilation in hydrothermal vents is emerging in metabolism first theories. Among the known pathways of CO₂ fixation [15, 16], the acetyl‐CoA pathway is the oldest [12, 15]. It is the only pathway of CO₂ fixation that occurs in bacteria and archaea [17], and the only route that simultaneously supports CO₂ fixation and ATP synthesis [18, 19]. Its reactions from H₂ and CO₂ to pyruvate proceed without enzymes [20, 21, 22, 23, 24] under the conditions of serpentinizing hydrothermal vents [25] using only solid‐state transition metals Fe⁰, Co⁰, Ni⁰ [20], and their alloys as catalysts. These native metals are naturally deposited [26] by the highly reducing (H₂‐rich) conditions of serpentinizing hydrothermal systems [27, 28]. The same transition metals are coordinated in the active sites of the enzymes [29, 30] and cofactors [31, 32, 33] of the acetyl‐CoA pathway, providing continuity of substrates, catalysts, and products between its metal‐catalyzed and enzymatic versions [34].

For nitrogen, a coherent picture of primordial N incorporation into protometabolism under hydrothermal conditions is also coming into focus. In laboratory serpentinization reactions, NH₄ ⁺ is readily generated from N₂ and H₂ in water reactions with peridotite, the host rock of serpentinizing systems [35]. As a hydrothermal route of NH₃ incorporation into metabolism, Kaur et al. [36] recently showed that in the presence of H₂ and as little as 6 mm NH₄ ⁺, Ni⁰ will catalyze reductive aminations of 2‐oxoacids to form amino acids. These reductive aminations produced glycine, alanine, aspartate, glutamate, valine, leucine, and isoleucine with yields of up to approximately 50% over 72 h at room temperature and a pH of 7–8. Both native Ni and the amount of H₂ that Kaur et al. [36] employed (5 bar) are observed in actively serpentinizing hydrothermal vents [25, 27, 28]. Abiotic glycine synthesis in a serpentinizing system has also been reported [37] (Fig. 1B). These reactions generate an ample supply of amino acids under primordial hydrothermal vent conditions. Yet they tether amino acid synthesis to solid‐state metal (mineral) surfaces before the origin of enzymes.

Fig. 1.

Biological and prebiotic N incorporation. (A) In organisms that live from the reduction of CO₂ with H₂, carbon backbones for amino acid synthesis are provided by the acetyl‐CoA pathway and the incomplete reverse citric acid cycle [12, 15, 34]. Nitrogen is assimilated as ammonium via N activation by glutamine synthase (GS), and reductive amination by glutamate synthase (GluS). Glutamate donates the amino group to various a‐keto acids via PLP‐dependent transaminases. In some organisms, the aminotransferase activity of GS is circumvented by glutamate dehydrogenase (GDH), which generates Glu from α‐ketoglutarate, ammonium, NAD⁺, and NAD(P)H. Another route of N incorporation entails the synthesis of carbamoyl phosphate (required in de novo pyrimidine synthesis) from NH₃ and CO₂ [46]. (B) Pyruvate is synthesized from H₂ and CO₂ under hydrothermal vent conditions using various native metals including Ni (shown), Fe, Co, and their alloys as catalysts [20, 21, 22, 23, 24]. Native Ni catalyzes reductive amination of various a‐keto acids with H₂ including pyruvate [36]. Metal ions (Fe²⁺) will convert pyruvate and glyoxylate to citric cycle intermediates including α‐ketoglutarate [47], but Ni–H₂ will catalyze the same sets of TCA cycle reactions [36]. Transamination reactions can be catalyzed by PLP or PL alone or by metal ions alone, (including Ni²⁺), or by combining PL with metal ions [40]. PL, pyridoxal; PLP, pyridoxal phosphate. (C) General mechanism of a PLP‐dependent transamination. In transaminases, PLP is enzyme‐bound as an aldimine to a lysyl side chain amino group [39] but is shown here as the free aldehyde for simplicity.

In metabolism, nitrogen incorporated as amino acids is distributed into other amino acid biosynthesis via transaminases, enzymes that require the cofactor pyridoxal phosphate (PLP). PLP interconverts 2‐oxoacids and amino acids (Fig. 1A). In a primordial protometabolism, transaminations can be catalyzed without enzymes by PLP alone [38] or even by metal ions in solution alone, without PLP, albeit at reduced rates [39, 40]. Recent work on PLP‐dependent transaminations in the context of early evolution has uncovered specific reaction mechanisms [39, 40, 41, 42] and shown that soluble metal ions can significantly impact the PLP‐dependent nonenzymatic reactions [39, 40, 43, 44], as can the inclusion of small peptides instead of metal ions [45]. While pyridoxal‐dependent transaminations efficiently interconvert 2‐oxoacids and amino acids [39, 40, 43, 44], they are dependent upon reductive aminations for net N incorporation. The mechanisms of PLP‐dependent transaminations always generate pyridoxamine phosphate (PMP) as the intermediate N‐donating species [41] (Fig. 1C).

Though Ni⁰ and H₂ can incorporate NH₄ ⁺ into protometabolism as a primordial source of organic nitrogen [36], the catalyst physically ties both NH₄ ⁺ incorporation and amino acid synthesis to solid‐state metal surfaces in or on the Earth's crust. For microbial cells to emerge, the reactions of metabolism had to become soluble, and the catalysts (enzymes) had to become independent of solid‐state catalysts. To explore possible intermediate steps in the transition from metal‐catalyzed reductive amination to pyridoxal‐catalyzed transamination, we investigated the ability of Ni⁰ to catalyze H₂‐dependent reductive amination of pyruvate to alanine in the presence or absence of pyridoxal to determine if pyridoxal impacts reductive amination under conditions of active serpentinization. We also tested whether H₂ and NH₄ ⁺ over Ni⁰ will reductively aminate pyridoxal to pyridoxamine as a freely diffusible, active amino donor for soluble transamination reactions.

Results and Discussion

Pyridoxal promotes reductive amination over nickel

Actively serpentinizing hydrothermal systems can harbor alkaline effluent with a pH of 8–10 or, in hyperalkaline systems, up to pH 12 or greater [25]. We first examined the effect of pH on reductive amination of pyruvate to alanine in the presence or absence of pyridoxal with all reactants at 20 mm and H₂ at 5 bar (or ~ 3.9 mm using Henry's law) in the presence of commercial nickel powder (Ni⁰) as the catalyst. After 18 h at 80 °C, we observed a 7.6% conversion of pyruvate to alanine at pH 11 and a 34% conversion in the presence of pyridoxal (Fig. 2A) using ¹H NMR to assay products (see Materials and methods). This indicates that at pH 11, alkalinity observed in actively serpentinizing hydrothermal vents [25], pyridoxal is compatible with Ni⁰, that is, it is not sequestered by the metal, it is stable during the reaction, and it does not inactivate the catalyst. Under these conditions, pyridoxal (PL) increases the alanine yield by a factor of 4.4 (Fig. 2A). This increase is due in part to the effect of pyridoxal on the competing lactate synthesis reaction. Without PL, 87% of pyruvate is converted to lactate, but only 7.5% conversion to lactate is observed in the presence of PL at 80 °C (Fig. 3). We also monitored pyridoxamine (PM) accumulation (Fig. 2B) via ¹H NMR. At pH 11, under the addition of PL, only 3.1% of PL accumulates as PM while 34% of pyruvate is converted to alanine. An increase in PL‐dependent alanine accumulation at a low steady‐state PM concentration suggests that PM is serving as a shuttle, undergoing cycles of reductive amination via the nickel catalyst, converting pyruvate to alanine, and regenerating PL in the process. In the absence of nickel catalyst, no alanine accumulates (Fig. 4).

Fig. 2.

Effect of pH and addition of pyridoxal on the nickel‐catalyzed reductive amination of pyruvate at 5 bar H₂ and 80 °C over 18 h (Table S1). Pyruvate concentration was set to 20 mm and the ratio between pyruvate, ammonium chloride, and pyridoxal was 1 : 1 : 1. The amount of nickel nanopowder was 1.5 mmol in a total volume of 1.5 mL. The reaction was performed at 80 °C for 18 h under a 5 bar hydrogen atmosphere. Error bars in the figure represent the standard deviation (SD). Each reaction was performed in triplicate. (A) Concentration of alanine in presence and absence of pyridoxal. (B) Concentration of pyridoxamine correlating with the concentration of alanine (A) in presence of pyridoxal.

Fig. 3.

Effect of temperature, time, and addition of pyridoxal on the nickel‐catalyzed accumulation of lactate (Table S2). Pyruvate, ammonium, and pyridoxal concentrations were set to 20 mm. The catalyst (nickel nanopowder) was added as 1.5 mmol of undissolved solid phase powder in a total reaction volume of 1.5 mL. The reaction was performed under a 5 bar hydrogen atmosphere at (A) 40 °C, (B) 60 °C, (C) 80 °C, and (D) 100 °C. The pH was set to 11 with KOH. Error bars in the figure represent the standard deviation (SD). Each reaction was performed in triplicate. The corresponding alanine concentrations are given in Fig. 5.

Fig. 4.

Alanine synthesis requires the presence of nickel catalyst. Pyruvate, pyridoxal, and ammonium were reacted in the presence of nickel catalyst (Table S3). Product accumulation after the removal of nickel nanopowder, hydrogen, and ammonium chloride is shown. Pyruvate, ammonium, and pyridoxal concentrations were set to 20 mm (or 0 in the corresponding control). The catalyst (nickel nanopowder) was added as 1.5 mmol of undissolved solid phase powder in a total reaction volume of 1.5 mL. The reaction was performed under a 5 bar hydrogen atmosphere or 5 bar argon in the control, pH was set to 11 with KOH, the temperature was set to 100 °C, and the reaction time was 18 h. Error bars in the figure represent the standard deviation (SD). Each reaction was performed in triplicate. Accumulation of alanine and lactate in the absence of hydrogen stems from nickel‐dependent reduction, the midpoint potential of the pyruvate to lactate reduction (E ₀′) is −190 mV, the midpoint potential (E ₀′) of Ni⁰ to Ni²⁺ oxidation is −260 mV [67]. Pyridoxamine concentrations are shown on the right‐hand y‐axis.

Pyridoxal promotes amination at higher temperature

Serpentinizing hydrothermal vents exhibit temperature gradients as effluent interfaces with seawater [48, 49, 50]. In the range of 0–100 °C, most uncatalyzed biological reactions take place more rapidly with increasing temperature, such that catalysts can increase rate more effectively at lower temperatures than at higher temperatures, itself an argument in favor of an origin of metabolism at high temperature [51]. At 40 °C and pH 11, the addition of PL apparently inhibited alanine accumulation as we could only detect alanine accumulation in the absence of PL (Fig. 5). The enhancing effect of PL on alanine accumulation increased with temperature, at 80 °C and 100 °C up to 48% of pyruvate was converted to alanine, while temperature had little effect on alanine accumulation in the absence of PL at pH 11.

Fig. 5.

Effect of temperature, time, and addition on the formation of pyridoxamine in nickel‐catalyzed reductive amination of pyruvate (Table S4). Pyruvate concentration was set to 20 mm and the ratio between pyruvate, ammonium chloride, and pyridoxal was 1 : 1 : 1. Pyridoxamine concentrations are given on the right axis. The amount of nickel nanopowder was 1.5 mmol in a total volume of 1.5 mL. The reaction was performed under a 5 bar hydrogen atmosphere at (A) 40 °C, (B) 60 °C, (C) 80 °C and (D) 100 °C. The pH was set to 11 with KOH. Error bars in the figure represent the standard deviation (SD). Each reaction was performed in triplicate. We noticed that after 18 h at 80 °C, the PM concentration is three times lower than in Fig. 2, likely due to the use of different batches of nickel. No PM was detected in reactions performed at 40 °C.

Pyridoxamine accumulation was also temperature dependent (Fig. 5), with alanine accumulating after 18 h only in reactions that led to PM accumulation of 1.2%. This is consistent with the role of PM as the aminating agent in PL‐dependent pyruvate conversion to alanine in the present samples. A main effect of PL was to favor reductive amination of pyruvate over pyruvate reduction to lactate. After 72 h, 9.4% pyruvate conversion to alanine by H₂, NH₃, and Ni⁰ alone was observed without PL, while under the addition of PL alanine and lactate accumulated in equal proportions (Fig. 6).

Fig. 6.

Effect of addition of pyridoxal on the product ratio between alanine and lactate in nickel‐catalyzed reductive amination of pyruvate (Table S5). Pyruvate concentration was set to 20 mm and the ratio between pyruvate, ammonium chloride, and pyridoxal was 1 : 1 : 1. The amount of nickel nanopowder was 1.5 mmol in a total volume of 1.5 mL. The reaction was performed at 100 °C for 72 h under a 5 bar hydrogen atmosphere, and pH was set to 11. Error bars in the figure represent the standard deviation (SD). Each reaction was performed in triplicate.

The addition of PL had a pronounced effect at low ammonium concentrations. At 5 mm ammonium, we detected 11% pyruvate conversion after 18 h at 100 °C but only 1.9% of pyruvate converted to alanine without PL using commercial nano‐nickel powder (Fig. 7A). For comparison, Kaur et al. [36] detected 16.7% pyruvate conversion to alanine at 6 mm NH₄ ⁺ and 25 °C after 72 h using silicate‐supported nano‐nickel catalysts. Above 100 mm NH₄ ⁺, alanine accumulation showed no significant difference with or without the addition of PL. However, below 100 mm NH₄ ⁺, the addition of PL significantly enhanced alanine formation compared to the H₂ and NH₃‐driven reaction, with the effect becoming more pronounced as NH₄ ⁺ concentration decreased (Fig. 7A). This is also consistent with a role for PM in PL‐enhanced reductive amination over nickel.

Fig. 7.

Effect of ammonium chloride concentration and addition of pyridoxal on the nickel‐catalyzed reductive amination of pyruvate (Tables S6 and S7). Pyruvate and pyridoxal concentrations were set to 20 mm. Pyridoxamine concentrations are given on the right axis. The amount of nickel nanopowder was 1.5 mmol in a total volume of 1.5 mL. The reaction was performed at 100 °C for 18 h under a 5 bar H₂ atmosphere, and pH was set to 11. Error bars in the figure represent the standard deviation (SD). Each reaction was performed in triplicate. (A) Concentration of alanine in presence and absence of pyridoxal; and concentration of pyridoxamine in presence of pyridoxal. (B) Concentration of lactate in presence and absence of pyridoxal.

In cells, PL is highly reactive and typically bound to the ε‐amino group of lysine in enzymes through a Schiff base linkage [52], such that the content of the free cofactor is very low. We tested pyridoxal concentrations ranging from 1 to 20 mm (Fig. 8). The yield of alanine generally decreased with decreasing PL concentrations. Up to 5 mm PL, there was an increased alanine yield compared to the control without PL. However, below 5 mm, no significant effect from PL addition could be detected.

Fig. 8.

Effect of the pyridoxal concentration on the Ni‐catalyzed reductive amination of pyruvate (Table S8). Pyruvate and ammonium concentrations were set to 20 mm. Pyridoxamine concentrations are given on the right axis. The amount of nickel nanopowder was 1.5 mmol in a total volume of 1.5 mL. Error bars in the figure represent the standard deviation (SD). Reactions were performed in triplicate at 100 °C for 18 h under a 5 bar H₂ atmosphere, and pH was set to 11.

Conclusion

Biochemical views on the origin of metabolism posit that inorganic surfaces and cofactors served as catalysts before the advent of enzymes [10]. Pyridoxal is a textbook case of cofactor‐only catalysis in that a pyridoxal‐containing transaminase will accelerate the rate of the uncatalyzed reaction by a factor of 10¹⁸, with PLP alone contributing a factor of 10¹⁰ to rate enhancement and the enzyme contributing the remaining 10⁸‐fold increase [53]. Iron sulfides were long thought to be the inorganic forerunners of enzymes [13, 54]. FeS minerals will catalyze some CO‐dependent reactions [55] as well as some reductive aminations in the laboratory, but the latter only at NH₄ ⁺ concentrations above 1 m [56]. Recent work indicates that for prebiotic CO₂ reduction and reductive amination, native Fe, Co, and Ni and their alloys are more effective and more versatile catalysts than FeS minerals and in many cases generate products that are identical to compounds of microbial metabolism [20, 21, 22, 23, 24].

Like Kaur et al. [36], we found that inorganic surfaces of Ni⁰ can catalyze the reductive amination of pyruvate with H₂. The alanine yield increased under conditions that correspond closely to those of actively serpentinizing hydrothermal vents (pH 11, 80 °C) [28, 57, 58] (Fig. 2). Reductive amination of pyruvate with H₂ over Ni resulted in a 4.4‐fold increase in alanine yield in the presence of PL (Fig. 5). We also found that in the presence of H₂ and 20 mm NH₄ ⁺, Ni⁰ will reductively aminate PL to PM (Fig. 5), the amino‐donating intermediate of transamination reactions. The yield of the reductive amination of pyridoxal depends on the pH (Fig. 2), time, and temperature (Fig. 5). The variation in PM concentrations in both figures can be attributed to the impact of these factors on the rate of reductive amination [41]. The PM amounts at pH 11, 18 h, and 80 °C differ across different experiments (Figs 2 and 5), which are attributable to different nickel catalyst batches from the supplier (Sigma‐Aldrich, St. Louis, Missouri, USA). Pyruvate, the substrate converted to alanine in our experiments, is readily produced from H₂ and CO₂ over nickel and nickel‐iron alloys at 20–100 °C and pH 8–10 in laboratory simulations of serpentinizing conditions [20, 21, 22, 23, 24].

The addition of PL increased pyruvate conversion to alanine relative to the competing reaction to lactate at lower NH₄ ⁺ concentrations down to 5 mm NH₄ ⁺ over 18 h (Fig. 7A,B). Although abiogenic NH₄ ⁺ synthesis has not been reported in modern serpentinizing systems, laboratory simulations of rock‐water interactions during serpentinization show that peridotite (a rock substrate for serpentinization) will react with water, H₂, and N₂ (300 °C, 50 bar N₂) to generate NH₃ at amounts corresponding to 350 μm over 29 days [35]. An abiotic source of PL has not been reported, but compounds similar to PL are readily obtained from heating glycolaldehyde with ammonia, though other routes from sugars and ammonia are also possible [59]. The present results show that PL and PM are compatible with native Ni, as PL enhances the Ni‐catalyzed reaction (Fig. 1), and PM‐dependent pyruvate amination is not inhibited by nickel (Fig. 9). Cofactor compatibility with native Ni, Co, and Fe was also observed for NADH [60, 61], and compatibility with Fe⁰ was also observed for ferredoxin, an electron carrier protein with two redox‐active 4Fe4S clusters [62].

Fig. 9.

PM‐dependent pyruvate amination is not inhibited by nickel (Table S9). Pyruvate and PM concentrations were set to 20 mm. The reaction was performed under 5 bar argon atmosphere, pH was set to 11 with KOH, the temperature was set to 100 °C, and the reaction time was 18 h in the presence or absence of nickel catalyst (1.5 mmol of undissolved solid phase powder in a total reaction volume of 1.5 mL). Error bars in the figure represent the standard deviation (SD). Each reaction was performed in triplicate.

Pyridoxal is a widespread cofactor in metabolism. It is estimated that PLP is used by roughly 1% of all protein‐coding genes in prokaryotes [63]. Pyridoxal is clearly the most versatile of all cofactors in terms of reaction types. In enzymatic reactions involving amino acid, oxoacid, and amine substrates, PLP catalyzes transaminations, Claisen condensations, β‐ and γ‐eliminations, β‐ and γ‐substitutions, epimerizations, racemizations, decarboxylations, transaldolations, and S‐adenosyl methionine‐dependent radical reactions [52, 64]. Underlying its catalytic proficiency, many pyridoxal‐dependent enzymes sequester the cofactor via aldimine formation with the side chain amino group of lysine, which is then displaced by the substrate amino moiety via transamination [52, 65] (Fig. 10). It is possible that PL can catalyze a broader spectrum of reactions under protometabolic conditions than just transaminations.

Fig. 10.

Reactions of PL. (A) The intermediates in PL catalyzed enzymatic reactions with amino acids, redrawn from Ref. [52]. (B) Proposal for reductive amination of PL to PM with ammonium and H₂ over Ni⁰ (see text). (C) Proposal for reversible PM‐mediated transamination half reaction to alanine, with ketimine hydrolysis instead of reaction with a second 2‐oxoacid regenerating PL, analogous to the mechanism of PL‐premediated transaminations [41].

Though the reaction mechanism of PL‐promoted reductive amination is unknown, the reactions of carbonyls with ammonia and reactions of PL (Fig. 10A) are well‐studied and can provide a guide. Imine formation from aldehydes or ketones and ammonia is fast and freely reversible in water. Under alkaline conditions, the equilibrium in the reaction of NH₄ ⁺ and pyruvate lies far on the side of the oxoacid, with imine reduction being irreversible even with the mild reductant NaBH₃CN [41]. In the present study, at 100 °C pH 11 and 5 bar H₂, the midpoint potential of the H₂ oxidation reaction H₂ → 2e⁻ + 2H⁺ is ca. −820 mV, strongly reducing conditions, such that the reductive amination of pyridoxal over Ni (Fig. 5) might proceed as sketched in Fig. 10B, with PM condensing with pyruvate to form the aldimine [41], followed by hydrolysis releasing alanine and regenerating PL (Fig. 10C). This reaction sequence could account for the low PM concentration relative to alanine (Fig. 7A). Imine reduction (Fig. 10B) also takes place during the enzymatic reductive amination reaction of α‐ketoglutarate with NH₄ ⁺ and NADH in the glutamate dehydrogenase reaction mechanism [66].

Our findings are consistent with the proposal that core biochemistry arose from reactions of H₂, CO₂, and NH₃ with the help of solid‐state catalysts in serpentinizing hydrothermal systems [21, 46]. The path from protometabolic reactions to cells entails a transition from solid state to soluble catalysts, whereby the latter could be metal ions or cofactors at first, followed by enzymes [10]. Reductive amination of PL to PM represents such an intermediate state in the development of prebiotic N incorporation before the origins of enzymes (Fig. 11). While soluble metal ions alone can catalyze transamination reactions, as can PL alone [40], net N incorporation is required for the accumulation of nitrogenous compounds prior to the origin of enzymes. Pyridoxamine synthesized on solid‐state catalysts is a diffusible aminating agent. In primordial metabolic evolution, PM could have freed N‐incorporating reactions from physical contact with solid‐state catalysts, acting as a shuttle that allowed surface metabolism (reductive amination) to become soluble metabolism (transamination) using organic cofactors as catalysts. While amino acids synthesized by reductive amination on Ni⁰ [36] could serve a similar function in transferring syntheses to the aqueous phase, PM can participate in a broad spectrum of reactions [52], which could have accelerated early biochemical evolution.

Fig. 11.

An evolutionary intermediate. Reductive amination of PL on solid‐state catalyst surfaces generates a soluble aminating reagent that could participate in amino acid synthesis from α‐ketoacid or any number of PM‐catalyzed reactions [52], marking a transition from solid state [36] to soluble catalysis in early biochemical evolution.

Materials and methods

Reaction

Reactions contained 20 mm each of pyruvate, pyridoxal hydrochloride (Merck, Sigma‐Aldrich, Darmstadt, Germany) or pyridoxamine (Merck Millipore, Billerica, Massachusetts, USA), and ammonium chloride, dissolved in distilled water in Falcon tubes. We used the nonphosphorylated forms of the cofactors (PL and PM, respectively) because of their better solubility in water relative to PLP and PMP. The pH was adjusted to 11 through the addition of 1 m KOH. Nickel nanopowder (Sigma‐Aldrich, St. Louis, Missouri, USA) was weighed out in an anaerobic glovebox (GS 79821; GS Glovebox System, Malsch, Germany). Samples contained 1 mmol of nano nickel per mL reaction volume. Samples (3 mL glass vials) were prepared in the glove box, placed in glass sample holders, and closed with corresponding lids (VWR International, Darmstadt, Germany), which were punctured to permit gas exchange before placing in the reactor (Berghof BR‐300 with BTC‐3000 temperature controller). The reactor was filled with 5‐bar hydrogen (99.999%; Air Liquide, Paris, France). After completion, reactors were depressurized and glass vial contents (metal powder and supernatant), were transferred to 2 mL Eppendorf tubes and centrifuged for 20 min at 16 060 g (Biofuge Fresco, Heraeus, Thermo Fisher Scientific Inc., Waltham, Massachusetts, USA). Supernatants were analyzed by NMR.

Product identification

After centrifugation, 600 μL of each sample was transferred to an NMR tube (VWR International). DSS (2,2‐dimethyl‐2‐silapentane‐5‐sulfonate) was added to a final concentration of 1 mm as a reference for calibration. NMR spectra were measured on a Bruker Avance III – 600 MHz spectrometer (Bruker Corporation, Billerica, Massachusetts, USA) by the Center for Molecular and Structural Analytics at Heinrich Heine University Düsseldorf. Spectra were analyzed using chenomx nmr suite version 9.02 software (Chenomx Inc., Edmonton, Alberta, Canada). Accumulation of PM and alanine identified by NMR was confirmed in selected probes by mass spectrometry (Figs 12 and 13).

Fig. 12.

HPLC‐MS analysis of standard solutions of pyridoxamine and alanine. Base peak chromatogram (BPC) and extracted ion chromatogram (EIC) of 50 mmol standard solutions of pyridoxamine (A) and alanine (B). The insets show the mass spectra for pyridoxamine m/z = 169.0971 and alanine m/z = 90.0549, respectively.

Fig. 13.

HPLC‐MS analysis of the effect of pH and addition of pyridoxal on the nickel‐catalyzed reductive amination of pyruvate. Reaction was performed at 5 bar H₂ and 80 °C over 18 h. Searching for the EICs of 169.0971 m/z of pyridoxamine and 90.0549 of alanine in the base peak chromatogram (BPC) (A) revealed clear signals for both educts in the reaction solution. The insets show the mass spectra for pyridoxamine m/z = 169.0971 (B) and alanine m/z = 90.0549 (C), respectively.

LC–MS analysis of pyridoxamine

Pyridoxamine was identified using the Dionex UltiMate 3000 UPLC system (Thermo Scientific, Germering, Germany) coupled to a maXis 4G (Bruker Daltonics, Bremen, Germany) quadrupole‐time‐of‐flight (Q‐TOF) mass spectrometer connected to an electrospray (ESI) ion source. Sample volumes of 10 μL were applied to a 3 mm by 150 mm C18 XSelect® HSS T3 column (2.5 μm particle size, 100 Å pore diameter; Waters, Drinagh, Wexford, Ireland) and separated using a binary gradient with a flow rate of 0.4 mL·min⁻¹. Mobile phase A was water + 0.1% formic acid, and mobile phase B was methanol + 0.1% formic acid. Starting with 8% B, a linear gradient to 95% B was applied from 2.5 to 10 min, followed by 95% B for additional 5 min and return to 8% B within 1 min. The system was equilibrated with 8% B for another 2 min prior to the next injection. The MS (positive‐ion mode) was run at 3.5 kV capillary voltage, 1 bar nebulizer pressure, 8 L·min⁻¹ dry gas flow, and dry temperature of 200 °C. Data acquisition was performed with compass hystar software (version 5.5) (Bruker). Pyridoxamine was quantified from full‐scan MS data (mass range 50–1000 m/z) using the dataanalysis (version 4.2) software (Bruker).

Conflict of interest

The authors declare no conflict of interest.

Author contributions

MB, MLS, and WFM designed the research. MLS, MBu, CGG, and SM performed experiments. MB, MLS, LS, SM, JM, and WFM interpreted and analyzed the data. MB, MLS, and WFM visualized the data. MB, MLS, MBu, SM, LS, and WFM wrote the paper. MB, MLS, LS, CGG, JM, and WFM edited the paper.

Peer review

The peer review history for this article is available at https://www.webofscience.com/api/gateway/wos/peer‐review/10.1111/febs.17357.

Supporting information

Table S1. Raw data Fig. 2.

Table S2. Raw data Fig. 3.

Table S3. Raw data Fig. 4.

Table S4. Raw data Fig. 5.

Table S5. Raw data Fig. 6.

Table S6. Raw data Fig. 7A.

Table S7. Raw data Fig. 7B.

Table S8. Raw data Fig. 8.

Table S9. Raw data Fig. 9.

Data availability statement

The data that support the findings of this study are available in Figs 2, 3, 4, 5, 6 and the Supporting Information of this article.