Prebiotic synthesis of phosphoenol pyruvate by α-phosphorylation-controlled triose glycolysis

Article abstract of DOI:10.1038/nchem.2624

Phosphoenol pyruvate is the highest-energy phosphate found in living organisms and is one of the most versatile molecules in metabolism. Consequently, it is an essential intermediate in a wide variety of biochemical pathways, including carbon fixation, the shikimate pathway, substrate-level phosphorylation, gluconeogenesis and glycolysis. Triose glycolysis (generation of ATP from glyceraldehyde 3-phosphate via phosphoenol pyruvate) is among the most central and highly conserved pathways in metabolism. Here, we demonstrate the efficient and robust synthesis of phosphoenol pyruvate from prebiotic nucleotide precursors, glycolaldehyde and glyceraldehyde. Furthermore, phosphoenol pyruvate is derived within an α-phosphorylation controlled reaction network that gives access to glyceric acid 2-phosphate, glyceric acid 3-phosphate, phosphoserine and pyruvate. Our results demonstrate that the key components of a core metabolic pathway central to energy transduction and amino acid, sugar, nucleotide and lipid biosyntheses can be reconstituted in high yield under mild, prebiotically plausible conditions.

Full text of DOI:10.1038/nchem.2624

ARTICLES
PUBLISHED ONLINE: 10 OCTOBER 2016 | DOI: 10.1038/NCHEM.2624
Prebiotic synthesis of phosphoenol pyruvate by
α-phosphorylation-controlled triose glycolysis
Adam J. Coggins and Matthew W. Powner
*
Phosphoenol pyruvate is the highest-energy phosphate found in living organisms and is one of the most versatile
molecules in metabolism. Consequently, it is an essential intermediate in a wide variety of biochemical pathways, including
carbon fixation, the shikimate pathway, substrate-level phosphorylation, gluconeogenesis and glycolysis. Triose glycolysis
(generation of ATP from glyceraldehyde 3-phosphate via phosphoenol pyruvate) is among the most central and highly
conserved pathways in metabolism. Here, we demonstrate the efficient and robust synthesis of phosphoenol pyruvate
from prebiotic nucleotide precursors, glycolaldehyde and glyceraldehyde. Furthermore, phosphoenol pyruvate is derived
within an α-phosphorylation controlled reaction network that gives access to glyceric acid 2-phosphate, glyceric acid
3-phosphate, phosphoserine and pyruvate. Our results demonstrate that the key components of a core metabolic pathway
central to energy transduction and amino acid, sugar, nucleotide and lipid biosyntheses can be reconstituted in high yield
under mild, prebiotically plausible conditions.
he integrated pathways of metabolism are orchestrated by an late-stage oxidation. Amidotriphosphate (7) and diamidophosphate
array of enzymes and chaperones that are ultimately depen- (8)—the ammonolysis products of cyclotrimetaphosphate23–32
—
T
dent on a highly evolved genetic coding system essential to effect a remarkably selective phosphorylation of α-hydroxyaldehydes
all life on Earth. However, at the origins of life, before the advent and sugars in aqueous solution25,32, and are commonly regarded to
of sophisticated enzymatic catalysis and selection, metabolism was be prebiotically plausible24,25,28–32. It was of particular interest to us
required to provide the fundamental building blocks of life1–14
.
that these phosphorylations block entry of the key aldehyde com-
Although only an intriguingly small constellation of molecules ponents 2 and 3 into ribonucleotide syntheses1,18, which redirects
produce the diversity of contemporary life, the function of the indi- both 2 and 3 to non-natural phosphorylation patterns and pyranosyl
vidual biological components cannot be appreciated in isolation1,2. (rather than natural furanosyl) sugar and nucleotide structures33.
Therefore, it must follow that the generation and function of bio- In a seminal contribution to prebiotic chemistry, Eschenmoser and
chemical species at the origins of life requires a concerted and, co-workers outlined the possibility that this switch in reactivity
ideally, unified investigation14. Seeking to further establish the prebiotic could have predisposed the prebiotic synthesis to non-canonical
generational relationship between the essential molecular components nucleotides25,32–34. However, in contrast, we hypothesized that
of biology, we were struck by the constitutional relationship between α-phosphorylation would redirect 2 and 3 into another metabolic
nucleotide precursors glycolaldehyde (2) and glyceraldehyde (3) pathway, and provide a generational link between two distinct path-
and the carbon framework of phosphoenol pyruvate (1), biology’s ways within (proto)metabolism. Specifically, under prebiotically
highest-energy phosphate (free energy at pH 7, 25 °C (kJ mol−1): plausible conditions18, rapid dehydration of glyceraldehyde 2-phos-
1 = −61.9; creatine phosphate = −43.0; adenosine triphosphate = −30.5; phate (3-2P) was anticipated to yield phosphoenol pyruvaldehyde
pyrophosphate = −19.2; glucose-6-phosphate = −13.8)15–18
.
(9) and, upon oxidation, 1. Interestingly, given the central role of gly-
Pyruvate (6) synthesis has been reported at elevated temperatures colysis in shaping biochemical amino acid, nucleotide and lipid
(25 °C) and pressures (50–200 MPa) in the presence of iron sulfides syntheses and the metabolic relationship between 3, 4-2P, 4-3P, 1
(0.07% yield)19, during the abiotic degradation of sugars and sugar- and 6, the efficient generation of each species is observed after
phosphates20,21, and in good yield by lactic acid oxidation (70%)22. subtle variations of mild aqueous reaction conditions to establish a
However, access to the high-energy phosphate 1 required by metab- single chemical network that can assemble each of the components
olism has not been addressed. Although phosphorylation of glyceric of the triose glycolysis pathway of contemporary metabolism (Fig. 1).
acid (4) by cyclotrimetaphosphate (which, notwithstanding contrary
opinion30, is considered a promising prebiotic phosphorylating Results
agent23–29) yields a 1:1.2 mixture of glyceric acid 2- and 3-phos- Phosphoenol pyruvaldehyde (9). Eschenmoser and co-workers
phates (4-2P/4-3P; 38%)23 and amide derivatives can be accessed have reported the phosphorylation of tetrose and pentose sugars
by Passerini reactions31, the elimination of 4-2P (or its amide with 8 (ref. 32) and magnesium-dependent phosphorylation of 2
derivatives) to give 1 has not been demonstrated. Indeed, even and 3 with 7 (ref. 25). We have found that the pH-stabilization
under strongly acidic (pH 2), alkaline (pH 10) or general acid– provided by phosphate buffer (pH 7), as anticipated, removes
base (1 M phosphate, pH 7) conditions, 4-2P (60 mM) is remark- the reported32 requirement for periodic addition of both 8
ably stable for prolonged periods of incubation (>15 h at 60 °C; and a proton-exchange resin during diamidophosphate-mediated
Supplementary Fig. 1). To overcome the shortcomings of both phosphorylations (to counteract the stoichiometric liberation of
regioselective phosphorylation and elimination, we reasoned that ammonia and ensure that the phosphorylation reaction proceeds).
1 would be directly accessible with high efficiency and complete Phosphate-buffered phosphorylation of 2 (25 mM) by 8 (100 mM)
regiochemical control by mild, aldehyde-mediated reactivity and at ambient temperature gave clean phosphorylation after 22 h.
Department of Chemistry, University College London, 20 Gordon Street, London WC1H 0AJ, UK. e-mail: matthew.powner@ucl.ac.uk
*
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
1
© 2016 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

NATURE CHEMISTRY DOI: 10.1038/NCHEM.2624
ARTICLES
a
b
O
Y
O
C6 metabolites
HCN
–
7 or 8
O
H2O
H2O
2–
2–
OH
O
OPO3
OPO3
2-P
[ + NH3]
2
4-3P Y=OH
5-3P Y=NH2
CH2O
H2O
2–
OH
OPO3
O
O
OH
O
OH
O
OH
[O]
7 or 8
21-P
–
O
H2O
H2O
2–
2–
OH
OPO3
3-2P
OPO3
4-2P
3
2–
O
OPO3
H2O
OH
O
3-3P
O
P
O
[O]
–
O
H2N
X
–
H2O
2–
O
OPO3
2–
OPO3
1
3-
9
7 X=OP2O6
8 X=NH2
H2O
Acetyl CoA
O
Y
[O]
–
O
O
H2O
O
Citric acid
cycle
20 Y=OH
Acetyl CoA Y=CoA
6
Figure 1 | Reconstitution of the triose glycolysis pathway (blue nodes) in a mild prebiotically plausible aqueous reaction network controlled by
α-phosphorylation. a, Blue arrows: prebiotic pathway to phosphoenol pyruvate (1) and pyruvate (6) from glycolaldehyde (2) and glyceraldehyde (3). Green
arrow: divergent prebiotic synthesis of glyceric acid 2-phosphate (4-3P; Y = O) and phosphoserine (5-3P; Y = NH) from glycolaldehyde-2-phosphate (2-P),
which provides a generational link between glycolysis (blue) and serine metabolism (green). Red arrow: oxidative decarboxylation of pyruvate (6) to yield
acetate (20). Black arrow: quantitative oxidation of glyceraldehyde 2-phosphate (3-2P) to glyceric acid 2-phosphate (4-2P). Structures boxed in blue
highlight the alignment between the prebiotic network and the specific node on the metabolic triose glycolysis pathway. b, Enzymatically controlled pathway
from C6-metabolites to citric acid cycle via glyceraldehyde-3-phosphate (3-3P) and dihydroacetone phosphate (21-P).
However, as both imine formation and amidophosphate protonation
Having demonstrated the synthesis of 9 from 3, we next incubated
appear essential to the mechanism of phosphorylation, we suspected glycolaldehyde phosphate (2-P) with formaldehyde. At pH 10.7, 2-P
that the reaction could be promoted at lower pH. Accordingly, we is converted to 3-2P (66%) over the course of six days33; further incu-
observed an optimal rate of conversion at pH 4 that gave 90–96% bation (60 °C, pH 10, three days) led to elimination to furnish 9 (61%;
phosphorylation of 2 or 3 (25 mM) at ambient temperature in 2 h Supplementary Fig. 22). Alternatively, incubation in phosphate buffer
(Supplementary Fig. 4).
On incubating 3-2P (80 mM) in phosphate (0.5 M) we ob- Supplementary Fig. 23) from formaldehyde and 2-P. Demonstrating
served facile dehydration to (74%, one day, 60 °C; the phosphorylation of both prebiotic aldehyde precursors of RNA
(0.5 M, 60 °C, pH 7, 30 h) led to the direct synthesis of 9 (40%;
9
Supplementary Fig. 21). Furthermore, direct synthesis of 9 (50%) can lead, even at neutral pH, to high-energy phosphoenol ether 9.
was observed during the stoichiometric phosphorylation of 3 in
phosphate buffer (pH 4, 0.5 M, 60 °C, five days) to directly estab- Phosphoenol pyruvate (1). Struck by the simplicity and efficiency
lish the high-energy phosphoenol moiety of 1 under mild of this synthesis and the structural relationship between 9 and 1,
aqueous conditions.
we next turned our attention to the oxidation of 9. The range of
2
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
© 2016 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

NATURE CHEMISTRY DOI: 10.1038/NCHEM.2624
ARTICLES
and todorokite43,44. Additionally, phylogenetic analysis suggests that
manganese may have been the active centre of the earliest
enzymes38, and manganese concentrations are thought to have
been as high as 50 mM in the oceans of the early Earth44.
Although the global availability of redox-sensitive Mn(IV) before
atmosphere oxygenation has been questioned45–47, the deviation of
local environments and near-surface oxygen fugacity from global
values48 (for example, owing to photo-oxidation, ionization, volca-
nic outgassing, meteorite impacts and atmospheric water dis-
sociation and hydrogen escape) remains opaque49. However,
stimulated by the recent elucidation of facile prebiotically plausible
sugar, amino acid and lipid syntheses mediated by photo-oxidation
of transition metals in aqueous solution12,39,41, and recognizing the
prebiotically plausible (photo-)oxidation of Mn(II) to Mn(III) and
Mn(IV)22,50, and disproportionation of Mn(III) to Mn(II) and
MnO2 in the absence of complexing ligands, manganese dioxide
(MnO2) was investigated as the most constitutionally simple Mn(IV)
oxidant. Interestingly, we observed near-quantitative conversion of
9 (100 mM) to 1 (93%) on incubation with potassium cyanide
(500 mM) and manganese dioxide (20 equiv.) after 2 h, which
demonstrates the facile synthesis of 1 from both 2 and 3 (Fig. 2
and Supplementary Fig. 37).
We next investigated the direct chemical oxidation of 9 to 1. The
relatively low bond-dissociation energy of an aldehyde hydrogen
atom (360–370 kJ mol−1)51, the probably high prebiotic availability
of iron21 and prebiotic plausibility of a hydroxyl-radical-mediated
oxidation24,52 suggested the investigation of Fenton oxidation.
Interestingly, incubation of 9 (25 mM) with ferrous chloride (25 mM)
and hydrogen peroxide (100 mM) at pH 7–11 furnished 1 (30%,
25 °C, pH 10; Fig. 2 and Table 1). However, these aggressive
oxidation conditions also led to the concomitant degradation of 1
to formate and glycolate (12).
To make further improvements to the direct oxidation, guided by
the conditions of nucleotide synthesis14,18, the prevalence of dry
heating and wet–dry cycles in prebiotic synthesis12,14,18,29,53, and
the simple photochemical synthesis and high abiotic abundance
of the chlorate family of molecules54–57, we were drawn to investigate
the role of these abiotic oxidants. Chlorates are ubiquitously formed
by stratospheric oxidation of chlorides on Earth and are widely
distributed by both wet and dry deposition57. Although ozone probably
plays a key role on Earth today, chlorates are also widely distributed
throughout the solar system (on Mars, the Moon and carbonaceous
chondrites including Murchison and Fayetteville). Little is yet
known regarding the origin of chlorates on Mars, but the known ter-
restrial syntheses do not account for their abundance or distribution
throughout the solar system54,55,57. Therefore, although the arid
environments implicated in chlorate synthesis are debated and the
delivery of chlorates to prebiotic chemistry remains unknown,
there are clearly abiotic mechanisms for the accumulation of chlor-
ates, which, if followed by wetting, could either mobilize deposits or
deliver materials to chlorate-rich locations. Moreover, the dispro-
portionation of atmospheric ClO2 in aqueous solutions would
directly deliver chlorite56,58, which is an excellent mild oxidant for
aldehydes in water, with a pH reactivity profile that is an ideal
match for 8-mediated phosphorylation59. Therefore, despite the
uncertain provenance of chlorates in prebiotic chemistry, the abun-
dance of chlorates in the solar system and their capability to act as
mild and highly selective nucleophilic oxidants in water meant that
we judged chlorates to be an interesting model oxidant for 9.
Chlorite oxidation of aldehydes can be mediated at low pH
(pH 3–6) in the presence of a phosphate buffer, which allows a
transition away from high-pH cyanide-catalysed oxidations. On
incubation of 9 (70 mM) with chlorite (100 mM) and a hypochlor-
ite scavenger such as methionine (140 mM) or dimethylsulfoxide
OH
O
O
[O]
H2N
+
NC
NC
H2O
2–
2–
2–
OPO3
10
OPO3
O
OPO3
11
13
HCN H2O
H2O
O
O
+
O
[O]
H
–
H2O
–
H2O
O
O
2–
2–
O
OPO3
OPO3
9
1
6
NaClO2
H2O
NaOCl H2O
NaOCl H2O
Cl
O
Cl
–
–
O
O
NaOCl
H2O
–
O
O
O
O
19
18
20
Figure 2 | Oxidation of phosphoenol pyruvaldehyde (9) and oxidative
decarboxylation of pyruvates (6 and 19). Oxidant ([O]) and oxidation
conditions are provided in Table 1. Blue arrows: aqueous umpolung oxidation
of phosphoenol pyruvaldehyde (9) to phosphoenol pyruvate (1) via
cyanohydrin (10). Red arrow: direct oxidation of phosphoenol pyruvaldehyde
(9) to phosphoenol pyruvate (1). Black arrow: quantitative phosphoenol
pyruvate (1) hydrolysis. Green arrows: oxidative decarboxylation of
α-ketoacids (6 and 19) yielding acetates (20 and 18), respectively.
potential conditions and oxidants available under prebiotic
constraints is unknown and probably highly varied. Therefore,
rather than develop a model geochemical scenario for aldehyde
oxidation, we sought to demonstrate a range of low-temperature,
aqueous model oxidations that operate through significantly
varied mechanistic pathways across a broad pH range (pH 3–13).
Activation towards oxidation through cyanohydrin formation35
initially appealed to us because high concentrations of hydrogen
cyanide are implicated in nucleotide, sugar and amino acid
abiogenesis1,5,8,12 and are observed in abiotic high abundance36,37
.
Incubation of 9 (33–200 mM) with sodium cyanide (167 mM to
1 M) led to rapid (<15 min) and quantitative conversion to cyano-
hydrin 10 in water (Supplementary Fig. 33). Because iron is likely to
be prebiotically abundant21 and clearly has biological importance38,
we first investigated ferricyanide ([Fe(CN)6]3−) oxidation of 10.
Ferricyanide is a prebiotically plausible model oxidant12,39–42 that
contains low-spin iron, which is easily reduced to the ferrocyanide
ion ([Fe(CN)6]4−) in a reversible redox couple and can be photo-
chemically regenerated from ferrocyanide12,39,41. We were pleased
to observe oxidation of 10 by ferricyanide between pH 11 and 13
(Fig. 2 and Supplementary Fig. 36). However, expedient oxidation
required significantly elevated pH and became limited by alkaline
nitrile hydrolysis.
To further improve the oxidation reaction and expand the scope
of potential model oxidants, we next investigated manganese, which
is the most common heavy metal in the Earth’s crust (0.1 wt%) after
iron43, and readily adopts +2, +3, +4, +6 and +7 oxidation states,
although naturally occurring systems are dominated by the +2, +3
and +4 oxidation states. Manganese is abundant in most geological
systems, covers 10–30% of the deep Pacific Ocean floor44, and is
readily found in minerals such as birnessite, pyrolusite, romanechite
(DMSO; 140 mM),
9
is quantitatively oxidized to
1
(Supplementary Figs 39 and 40).
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
3
© 2016 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

NATURE CHEMISTRY DOI: 10.1038/NCHEM.2624
ARTICLES
Table 1 | Aldehyde oxidation conditions and product yields after 2 h incubation of the substrate and specified oxidant ([O])
in water or phosphate buffer.
‡
*
†
–
O
O
O
O
O
[O]
[O]
[O]
O
R2
R2
–
–
–
H2O
H2O
H2O
H
O
O
H
O
O
2–
2–
OPO3
OPO3
R1
R1
O
2/3
12/4
6
20
9
1
Substrate
R1
R2
pH
[O] (equiv.)
Additive (equiv.)
Product
Yield (%)
2−
9
9
9
9
OPO32−
OPO32−
OPO32−
OPO32−
OPO32−
OPO3
OH
OH
H
H
–
=CH2
=CH2
=CH2
=CH2
=CH2
CH2OH
OPO3
CH2OH
OPO3
13
9
10
4
4
4
4
4
4
4
K3Fe(CN)6 (10)
MnO2 (20)
FeCl2 (1)
HCN (5)
HCN (5)
H2O2 (4)
DMSO (2)
–
DMSO (2)
DMSO (2)
DMSO (2)
DMSO (2)
DMSO (2)
–
1 (13)
1
1 (12)
1
50 (13)
93
30 (16)
Quant.
Quant.
Quant.
94
NaClO2 (1.4)
NaClO2 (2)
NaClO2 (1.4)
NaClO2 (1.4)
NaClO2 (1.4)
NaClO2 (1.4)
NaClO2 (1.4)
NaOCl (2)
9
18
3-2P
3-3P
3
2-P
2
4-2P
4-3P
4
12-P
12
2−
93
2−
Quant.
Quant.
Quant.
OH
–
6
4
20
*Oxidation of phosphoenol pyruvaldehyde (9) to phosphoenol pyruvate (1); †oxidation of glycolaldehyde (2) and glyceraldehyde (3) derivatives to glycolic acid (12) and glyceric acid (4) derivatives, respectively;
‡oxidative decarboxylation of pyruvate (6) to acetate (20). An expanded table of conditions and yields is provided in Supplementary Table 1. Quant., quantitative.
Glyceric acid 2-phosphate (4-2P) and glycolic acid phosphate Pyruvate (6). Compound (1) is highly stable at high pH (6%
(12-P). Gratifyingly, the oxidation by chlorite is general and hydrolysis to 6 observed over 2 days at pH 10 and 60 °C;
provides quantitative conversion of 2 and 2-P to glycolic acid Supplementary Fig. 93). However, 1 is readily hydrolysed to 6 at
(12) and glycolic acid phosphate (12-P), respectively (Table 1, lower pH in quantitative yield (pH 4–7, 2 days, 60 °C; Supplementary
Supplementary Figs 41 and 42). Phosphate 12-P is a key Fig. 94). Hydrolysis thus provides access to a second linchpin of
intermediate of photorespiration in plants15. Consequently, contemporary metabolism with excellent control and efficiency in
chlorite is also an ideal oxidant to convert 3-2P to 4-2P. water. Accessing 6 from 1 delivers improved yields (up to 69%
Following reaction of 3-2P with stoichiometric chlorite, we from 3, over four steps) relative to previously reported multiple-step
observed quantitative oxidation to 4-2P. Furthermore, upon syntheses (984-fold and 7-fold)19,20 and a comparable yield to the
phosphorylation of 3 with 8 followed by incubation at 60 °C in single-step oxidation of lactic acid (70%)22. More importantly, the
phosphate buffer (0.5 M) for 5 h and addition of chlorite/DMSO pathway mimics a metabolic trajectory with absolute specificity under
(1:1.4), we observed quantitative conversion of 3 to 1 and 4-2P mild aqueous conditions from simple prebiotic aldehydes.
(1:1) and thus realized the direct divergent synthesis of
two key intermediates of the universally conserved glycolysis Acetate (20). It is of significant interest that during our investigation
pathway (Fig. 3)17.
of phosphoenol pyruvaldehyde (9) oxidation with chlorite that in
the absence of a hypochlorite scavenger we observed over-
Glyceric acid 3-phosphate (4-3P) and phosphoserine (5-3P). oxidation to give chloroacetate (18; >98% yield with 2 equiv.
Interestingly, the hydrolysis that limited the ferricyanide oxidation chlorite; Supplementary Fig. 95). It is likely that 18 was accessed
of cyanohydrin 10 suggested direct prebiotic access to 4-3P and by electrophilic chlorination and oxidative decarboxylation of α-
phosphoserine (5-3P) through cyanohydrin and aminonitrile ketoacid 19 (Fig. 2), which is supported by the observed
hydrolysis, which provides access to a fourth component of the quantitative oxidative decarboxylation of 6 to yield 20 (ref. 61).
glycolysis pathway and a direct link to serine metabolism, Pyruvates 1 and 6 are both key nodes of central metabolism, but
respectively. On incubation of 2-P with sodium cyanide, we one of the most important metabolic connections is achieved by
observed smooth and quantitative conversion to cyanohydrin 14 oxidative decarboxylation of 6 to access acetyl coenzyme A (acetyl
(Supplementary Fig. 46), which is readily hydrolysed to 15 (27%, CoA)62, ketogenesis and the citric acid cycle (Fig. 1)6–9. The direct
750 mM phosphate, pH 7, three days, 75 °C; Supplementary Fig. 54) liberation of 20 by oxidation of 6 suggests an essential prebiotic
and 4-3P (31%, 750 mM phosphate, pH 12, 75 °C, five days; link between C3 and C2 metabolism within our system; an
Supplementary Fig. 59) with absolute control over the phosphate investigation of (prebiotic) pathways to acetyl transfer reagents39,41
regiochemistry (Fig. 4).
and C2 metabolism is currently under way in our laboratory.
Furthermore, incubation of 2-P with ammonium cyanide led to
the near-quantitative synthesis of glycolaldehyde phosphate amino- Discussion
nitrile (16; Supplementary Fig. 66). We found that the hydrolysis of Glycolysis is one of the most highly conserved metabolic pathways
aminonitrile 16 to amide 17 was aldehyde-catalysed; the cyanide of life and it is instrumental to the biosynthesis of amino acids,
carbon atom was mildly and efficiently locked into the carbon frame sugars, nucleotides and lipids, phosphorylation and the entry to
work of serine by a small excess of 2-P (<5%)60. This reactivity is the Krebs (citric acid) cycle2,16,17. The interrelated syntheses of 1,
only observed for aminonitrile hydrolysis, not cyanohydrin hydrolysis. 4, 4-2P, 4-3P, 5-3P, 6, 12, 12-P and 20 demonstrated herein
The aldehyde-catalysed hydrolysis occurs via 5-exo-dig hemiaminal suggest that prebiotic availability, as well as metabolic versatility,
cyclization and is quenched by excess cyanide. It is therefore of note may have positioned glycolysis at the heart of modern cellular
that we observed no evidence of intramolecular phosphate-assisted metabolism. The generational simplicity of this network also
hydrolysis, despite the ostensive potential for 6-exo-dig addition to suggests that α-phosphorylation and oxidation warrant further
the nitrile, which would stall 5-3P synthesis at glycolic acid 4-3P investigation with respect to the essential roles they may have
and, as a result, we observed efficient conversion of 2-P to 5-3P played in coordinating the first steps of life on Earth. Specifically,
(36%; Supplementary Fig. 83).
the remarkable chemical control that α-phosphorylation achieves
4
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
© 2016 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

NATURE CHEMISTRY DOI: 10.1038/NCHEM.2624
ARTICLES
OH
OH
a
HO
2–
OPO3
3-2P
O
b
2–
OPO3
9
O
O
OH
HO
HO
2–
2–
OPO3
OPO3
4-2P
1
c
9.0
6.0
5.5
5.0
4.5
4.0
3.5
Chemical shift (ppm)
Figure 3 | 1H NMR spectra (600 MHz, H2O/D2O 9:1, 25 °C) showing quantitative divergent synthesis of phosphoenol pyruvate (1)/glyceric acid
2-phosphate (4-2P) in 0.5 M phosphate buffer. a, Spectrum showing the incubation of glyceraldehyde (3; 25 mM) and diamidophosphate (8; 50 mM) at
initial pH 4 and 25 °C for 1 h in phosphate buffer, which yields glyceraldehyde-2-phosphate (3-2P; blue). b, Spectrum to show incubation of glyceraldehyde-
2-phosphate (3-2P) in pH 7 phosphate buffer at 60 °C for 5 h, yielding a 1:1 mixture of glyceraldehyde-2-phosphate (3-2P; blue) and phosphoenol
pyruvaldehyde (9; red). c, Spectrum showing the quantitative concomitant oxidation of glyceraldehyde-2-phosphate (3-2P) and phosphoenol pyruvaldehyde
(1) to glyceric acid 2-phosphate (4-2P) and phosphoenol pyruvate (9) by chlorite/DMSO (1:1.4) in pH 4 phosphate buffer at 25 °C after 1 h.
in reconstituting the pathway of glycolysis indicates that a transition the early Earth is warranted before unification of prebiotic and
from α-phosphorylation to terminal phosphorylation may greatly geochemical redox strategies49.
simplify the transition from abiotic reactions to contemporary
It is of note that chlorite oxidation is not only quantitative but that
metabolism. The essential and ubiquitous roles of 4-2P and 1 in its reactivity profile is also pH matched with the maximal reactivity of
modern biochemistry may be vestiges of an earlier reactivity upon the prebiotic phosphorylating agent 8 (optimal α-phosphorylation at
which modern glycolysis/gluconeogenesis are built from the low pH). At the concentrations investigated (25–500 mM), back-
‘bottom up’17. Although modern metabolism enters triose glycolysis ground hydrolysis does not compete with Schiff base-mediated
through glyceraldehyde 3-phosphate (3-3P), 3-3P is unstable with phosphoryl transfer; however, the increased efficiency of Schiff base
respect to its ketose isomer (21-P) and is predisposed to undergo formation and increased proportion of protonated amidophosphate
elimination to methyl glyoxal with concomitant loss of phosphate leads to a significantly improved rate of phosphorylation. This pH
activation. Biochemically, the isomerization of triose sugar phos- profile also matches the pH requirements (pH < 7) for controlled
phates is mediated to the limit of diffusion by triose-phosphate iso- ribonucleotide assembly by a leading model for prebiotic synthesis18.
merase, but at equilibrium, the formation of 21-P is still favoured Therefore, it is also interesting to note that an acidic drying phase is
20:1 (ref. 63). It is of note, however, that we have demonstrated required during ribonucleotide synthesis to induce phosphorylation
that glyceraldehyde 2-phosphate (3-2P) blocks these deleterious and stereocontrolled rearrangement14,18, and these arid, low-pH
reaction pathways without the requirement for enzymatic control. conditions have remarkable parity with the proposed environmental
Furthermore, by retaining aldehyde activation to E1cb-type elimin- pathways to chlorine oxides54–57. However, the relevance of oxidized
ation, 3-2P readily undergoes a predisposed elimination to access the chlorine reagents in prebiotic chemistry can certainly be questioned,
high-energy phosphoenol moiety, which is kinetically prohibited for and further investigation of the potential to accumulate and transport
4-2P and its amide derivatives.
chlorates55,57 and chlorine dioxide56,58 would be needed to verify facile
Aldehyde oxidations are commonplace in chemistry, and a prebiotic access to 4-2P, which we only found to be efficient by
variety of oxidants can be used in the facile oxidation of 9 to 1. chlorite oxidation. It is also of note that although 4-2P is a biological
The conditions on the early Earth are far from certain, and we intermediate of glycolysis, it is not an intermediate of our prebiotic
have not sought to specify geochemical constraints for aldehyde network; rather, it is a terminal node (Fig. 1; blue arrows), and
oxidation. We have, however, demonstrated umpolung and conven- without the aldehyde activation exploited to yield 9, it is difficult to
tional polarity oxidation, nucleophilic, electrophilic and radical envisage elimination to deliver 1 (Supplementary Fig. 1).
oxidation, heterogeneous and homogenous oxidation, and both
It is striking that the prebiotically plausible nucleotide synthesis
outer- and inner-sphere oxidation of 9 to give 1, and consequently reported by Powner et al.18, is achieved, like prebiotic entry to glycoly-
suspect that the facile oxidation of 9 in water is extremely general sis, through a regiochemical phosphorylation isomer of the canonical
and thus prebiotically plausible. Nonetheless, further investigation form. Ribonucleotide synthesis18 occurs via 2′,3′-cyclic phosphates
of redox-sensitive elements (including Mn, Fe, Co, Ni and Cu) on rather than terminal 5′-phosphates, and triose glycolysis is accessed
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
© 2016 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
5

NATURE CHEMISTRY DOI: 10.1038/NCHEM.2624
ARTICLES
Aldehyde oxidation. Method A: Aldehyde (30–100 mM), sodium cyanide (5 equiv.
or none) and an internal standard (DSS, pentaerythritol, or acetate (20)) were
dissolved in H2O/D2O (9:1) or D2O and adjusted to the specified pH/pD with 1 M
HCl/NaOH. Cyanohydrin formation was monitored by 1H NMR spectroscopy and,
after equilibration (0.1–1 h), the oxidant (K3Fe(CN)6, or MnO2) was added. The
solution was incubated at ambient temperature and the specified pH for 1–2 h. The
solution was filtered and NMR spectra were acquired. The carboxylic acid product
was confirmed by sample spiking and the yield (Table 1 and Supplementary Table 1)
was quantified with respect to the internal standard.
X
X
O
HCN
H2O
Y
NC
/– NH
+
3
2–
2–
2–
O
OPO3
OPO3
OPO3
2-P
14 X=OH
16 X=NH2
15 X=OH Y=NH2
17 X=NH2 Y=NH2
4-3P X=OH Y=OH
5-3P X=NH2 Y=OH
Method B: Phosphoenol pyruvaldehyde (9; 25 mM) and ferrous chloride
(0–2 equiv.) were dissolved in phosphate buffer (0.75 M, H2O, pH 10) and stirred
rapidly at 0 °C. Hydrogen peroxide (30% wt/wt, 1–4 equiv.) was added and the
reaction was stirred at 0–15 °C for 30 min. The reaction was warmed to ambient
temperature over 30 min and the solids removed by centrifugation. The supernatant
was either combined with D2O (10% vol/vol) or lyophilized and re-dissolved in D2O
(1 ml). NMR spectra were acquired, and the formation of 1 was confirmed by sample
spiking. The yield (Table 1 and Supplementary Table 1) was quantified with respect
to an internal standard (20) and titration of an external standard (DSS).
Method C: Aldehyde (70–350 mM), an internal standard (DSS or 20) and
the specified additive (none, NH4Cl, H2O2, DMS, DMSO, sulfamic acid or
L-methinone) were dissolved in phosphate buffer (60 mM, D2O, pH 4–7). Sodium
chlorite (5 M, 1.4 equiv.) was added in five portions over 1 h at 0 °C, then the
solution was warmed to ambient temperature and NMR spectra were acquired.
The carboxylic acid product was confirmed by sample spiking, and the yield
Figure 4 | Divergent synthesis of glyceric acid 3-phosphate (4-3P) and
phosphoserine (5-3P) from glycolaldehyde-2-phosphate (2-P). By
incubating 2-P with hydrogen cyanide, cyanohydrin (14) or aminonitrile (16)
(in the presence of ammonia) can be formed quantitatively. Subsequent
nitrile hydrolysis furnishes 15 and 4-3P with absolute control over the
phosphate regiochemistry.
via 3-2P rather than terminal 3′-phosphate 3-3P. The parity of these
model reaction conditions and the ostensive regiochemical switch
from α-phosphorylation both suggest that an integrated and unifying
system to concurrently establish nucleotide synthesis and triose
metabolism may have been fundamentally important in constructing (Table 1 and Supplementary Table 1) was quantified with respect to the
internal standard.
the first steps towards contemporary metabolism.
Phosphoenolpyruvaldehyde cyanohydrin (10). 1H NMR (400 MHz, D2O, pD 9.5)
Methods
δH 5.10 (1H, s, CH), 4.95 (1H, t, J = 1.8 Hz, CH2), 4.83 (1H, t, J = 1.8 Hz, CH2).
Phosphorylation. Aldehyde (25–500 mM) and sodium 3-(trimethylsilyl)-1-
propanesulfonate (DSS, 2.5 mM) were dissolved in phosphate buffer (500 mM,
H2O/D2O 9:1, pH 4–7). Diamidophosphate (8; 1–4 equiv.) was added, and the
solution was incubated at ambient temperature. 1H and 31P NMR spectra
were periodically acquired, and phosphorylation was quantified with respect to the
internal DSS standard. Aldehyde phosphates were isolated on a preparative scale by
ion-exchange chromatography (Dowex-1 × 8, HCO3− form, eluent 0 → 0.5 M
Et3NH·HCO3) and barium (Ba(OAc)2, 1 equiv.) or calcium (Ca(OAc)2, 1 equiv.)
salt precipitation from aqueous ethanol (EtOH/H2O 4:1) or aqueous acetone
(Me2CO/H2O 1.2:1). Sodium salts were acquired by ion-exchange chromatography
(Dowex-50 W × 8, Na+ form) followed by lyophilization.
31P NMR (161 MHz, D2O, pH 9.5) δP 0.2 (aps).
Phosphoenol pyruvate (1). 1H NMR (400 MHz, D2O, pD 4) δH 5.78 (1H, t, J = 2.2 Hz,
(C3)-Ha), 5.44 (1H, t, J = 2.2 Hz, (C3)-Hb). 13C NMR (151 MHz, D2O) δC 168.1
(d, J = 7.2 Hz, C1), 146.2 (d, J = 7.2 Hz, C2), 108.6 (d, J = 3.9 Hz, C3). 31P NMR
(161 MHz, D2O) δP −4.0 (aps).
Glyceric acid (4). 1H NMR (400 MHz, D2O, pD 4) δH 4.28 (1H, dd, J = 4.5,
3.8 Hz, CH), 3.81 (1H, ABX, J = 12.1, 3.8 Hz, CH2), 3.78 (1H, ABX, J = 12.1,
4.5 Hz, CH2). 13C NMR (151 MHz, D2O) δC 176.8 (C1), 72.3 (C2), 64.0 (C3).
Glyceric acid 2-phosphate (4-2P). 1H NMR (400 MHz, D2O, pD 4) δH 4.61 (1H, dt,
J = 9.5, 3.6 Hz, CH), 3.89 (2H, d, J = 3.6 Hz, CH2). 13C NMR (151 MHz, D2O)
δC 175.0 (d, J = 3.9 Hz, C1), 75.9 (d, J = 5.0 Hz, C2), 63.7 (d, J = 5.0 Hz, C3).
31P NMR (161 MHz, D2O) δP 0.1 (d, J = 9.5 Hz).
Glycolaldehyde 2-phosphate (2-P). Yield: 90% (4 equiv. 8, pH 4, 25 mM, 4 h, room
temperature (RT)). M.p. 135–145 °C (dec). 1H NMR (600 MHz, D2O) δH 5.11
(1H, t, J = 4.5 Hz, (C1)-H), 3.75 (2H, dd, J = 7.4, 4.5 Hz, (C2)-H2). 13C NMR
(151 MHz, D2O) δC 89.9 (d, J = 7.2 Hz, C1), 67.4 (d, J = 4.9 Hz, C2). 31P NMR
(162 MHz, D2O) δP 4.47 (t, J = 7.4 Hz). m/z (ESI−): 139 (100%, [M–H]+).
Glycolic acid (12). 1H NMR (400 MHz, D2O, pD 4) δH 4.13 (2H, s, CH2). 13C NMR
(151 MHz, D2O) δC 177.4 (C1), 60.2 (C2).
Glyceraldehyde 2-phosphate (3-2P). Yield: 96% (4 equiv. 8, pH 4, 25 mM, 4 h, RT).
M.p. 110–120 °C (dec). 1H NMR (600 MHz, D2O) δH 5.05 (1H, d, J = 3.8 Hz,
(C1)-H), 4.09 (2H, dddd, J = 8.8, 5.7, 4.5, 3.8 Hz, (C2)-H), 3.78 (1H, ABX, J = 12.1,
4.5 Hz, (C3)-Ha), 3.75 (1H, ABX, J = 12.1, 5.7 Hz, (C3)-Hb). 13C NMR (151 MHz,
D2O) δC 90.1 (d, J = 3.6 Hz, C1), 77.5 (d, J = 5.5 Hz, C2), 62.3 (d, J = 4.4 Hz, C3).
31P NMR (161 MHz, D2O) δP 3.6 (d, J = 8.8 Hz). m/z (ESI+): 215 (100%, [M-H+]).
High-resolution mass spectrometry (HRMS) ([C3H5O6PNa2+H]+) calcd 214.9692,
found 214.9690.
Glycolic acid phosphate (12-P). 1H NMR (400 MHz, D2O, pD 4) δH 4.37 (2H, d,
J = 8.0 Hz, CH2). 13C NMR (151 MHz, D2O) δC 175.1 (C1), 63.2 (d, J = 5.0 Hz, C2).
31P NMR (161 MHz, D2O) δP 0.0 (aps).
Chloroacetic acid (18). 1H NMR (400 MHz, D2O) δH 4.02 (2H, s, CH2).
Acetic acid (20). 1H NMR (400 MHz, D2O) δH 1.96 (3H, s, CH3).
Phosphoenol pyruvaldehyde (9). Method A: Glyceraldehyde 2-phosphate (3-2P;
70 mM) and pentaerythritol (10 mM) were dissolved in phosphate buffer (500 mM,
H2O/D2O 9:1, pH 7) and incubated at 60 °C. NMR spectra were periodically
acquired and dehydration was quantified with respect to the internal pentaerythritol
standard. A maximum yield of 9 (74%) with residual 3 (12%) was observed at 23 h.
Method B: Glycolaldehyde 2-phosphate (2-P; 80 mM), formaldehyde (800 mM)
and DSS (15 mM) were dissolved in phosphate buffer (500 mM, H2O/D2O 9:1,
pH 7) and incubated at 60 °C. NMR spectra were periodically acquired and the yield
of 9 was quantified with respect to the internal DSS standard. A maximum yield of
9 (40%) was observed at 30 h.
Compound 9 was isolated on preparative scale by barium (Ba(OAc)2, 1 equiv.)
salt precipitation from aqueous ethanol (EtOH/H2O 4:1). Sodium salts were acquired
by ion-exchange chromatography (Dowex-50W × 8, Na+ form) followed by
lyophilization. An analytical sample was prepared by reversed-phase HPLC (isocratic
gradient of 0.1% NEt3H·HCO3 in water, retention time = 2.68 min). M.p. 100–115 °C
(decomposed). Infrared (IR, cm−1, powder) 1,686, 1,618, 1,562. 1H NMR (600 MHz,
D2O) δH 9.25 (1H, d, J = 3.0 Hz, (C1)-H), 6.17 (1H, dd, J = 2.8, 1.8 Hz, (C3)-Ha),
5.84 (1H, dd, J = 2.8, 2.0 Hz, (C3)-Hb), 5.29 (0.04H, s, (C1)-Hhydrate), 4.87 (0.04H, apt,
J = 2.1 Hz, (C3)-H2hydrate), 4.85 (1H, apt, J = 2.1 Hz, (C3)-Hh2ydrate), 3.18 (7H, q,
J = 7.3 Hz, NCH2CH3), 1.26 (10.5H, t, J = 7.4 Hz, NCH2CH3). 13C NMR (151 MHz,
D2O) δC 191.8 (d, J = 6.1 Hz, C1), 152.9 (d, J = 7.2 Hz, C2), 152.7 (C2hydrate),
119.5 (d, J = 3.9 Hz, C3), 93.6 (C3hydrate), 87.3 (C1hydrate), 47.3 (NCH2CH3),
8.9 (NCH2CH3). 31P NMR (161 MHz, D2O)δP −4.1 (aps, Phydrate), −4.2 (aps, P). m/z (ESI
+):197(100%, [M + H]+). HRMS ([C3H3O5PNa2+H]+)calcd196.9586, found196.9583.
Pyruvate (6). Phosphoenol pyruvate (1; 63 mM) and pentaerythritol (16 mM) were
dissolved in H2O/D2O (9:1, pH 4–10), or phosphate buffer (0.5 M, H2O/D2O, 9:1,
pH 7) and incubated at 60 °C for 50 h. 1H NMR spectra were acquired periodically
and the yield was quantified with respect to the internal pentaerythritol standard.
Yield 95–97% (pH 4 or phosphate buffer, 50 h, 60 °C). 1H NMR (600 MHz, H2O/
D2O 9:1, pH 4) δH 2.27 (3H, s, CH3keto), 1.44 (0.7H, s, CH3hydrate). 13C NMR
(151 MHz, H2O/D2O 9:1, pH 4) δC 205.2 (C1keto), 177.2 (C1hydrate), 170.5 (C2keto),
93.9 (C2hydrate), 27.1 (C3keto), 25.9 (C3hydrate).
Cyanohydrin/aminonitrile hydrolysis. Aldehyde (200 mM), sodium cyanide
(0.95–5 equiv.), ammonium chloride (0–5 equiv.) and DSS (20–25 mM) were
dissolved in H2O/D2O (9:1; adjusted to pH/pD 2, 7 or 9.5 with 4 M HCl or 4 M
NaOH) or in phosphate buffer (750 mM, H2O/D2O 9:1, pH 7). Cyanohydrin and
aminonitrile formation were monitored by NMR spectroscopy. Upon equilibration
(5–100 h), the solution was incubated at 25, 60 or 75 °C (3–6 days) and NMR
spectra were periodically acquired to monitor nitrile/amide hydrolysis. The
solution was adjusted to pH/pD 12 with 4 M NaOH/NaOD and incubation was
continued (5–30 h). NMR spectra were periodically acquired to monitor
amide hydrolysis.
Glycolaldehyde phosphate cyanohydrin (14). Yield 94–98% (5 h, pH 2–9.5, RT).
1H NMR (400 MHz, H2O/D2O 9:1, pH 7) δH 4.75 (obs., CH), 3.95 (2H, m, CH2).
13C NMR (151 MHz, H2O/D2O 9:1, pH 7) δC 120.1 (C1), 65.7 (d, J = 4.4 Hz, C3),
62.3 (d, J = 6.6 Hz, C2). 31P NMR (161 MHz, H2O/D2O 9:1, pH 7) δP 3.8 (aps).
6
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
© 2016 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

NATURE CHEMISTRY DOI: 10.1038/NCHEM.2624
ARTICLES
Glyceramide 3-phosphate (15). Yield 27% (phosphate buffer, 3 d, 75 °C). 1H NMR
(600 MHz, H2O/D2O 9:1, pH 7) δH 4.25 (1H, dd, J = 5.2, 3.2 Hz, CH), 4.00
(1H, ABXY, J = 11.6, 7.6, 3.2 Hz, CH2), 3.92 (1H, ABXY, J = 11.6, 7.6, 5.2 Hz, CH2).
13C NMR (151 MHz, H2O/D2O 9:1, pH 7) δC 179.4 (C1), 73.5 (d, J = 7.2 Hz,
C2), 67.5 (d, J = 4.4 Hz, C3). 31P NMR (161 MHz, H2O/D2O 9:1, pH 7) δP 4.4
(t, J = 7.6 Hz).
19. Cody, G. D. et al. Primordial carbonylated iron–sulfur compounds and the
synthesis of pyruvate. Science 289, 1337–1340 (2000).
20. Weber, A. The sugar model: catalysis by amines and amino acid products. Orig.
Life Evol. Biosph. 31, 71–86 (2001).
21. Keller, M. A., Turchyn, A. V. & Ralser, M. Non-enzymatic glycolysis and pentose
phosphate pathway-like reactions in a plausible Archean ocean. Mol. Syst. Biol.
10, 725–737 (2014).
Glyceric acid 3-phosphate (4-3P). Yield 31% (phosphate buffer, 3 days, 75 °C,
then pH 12, 5 h, 75 °C). 1H NMR (600 MHz, H2O/D2O 9:1, pH 7) δH 4.11 (1H, dd,
J = 6.1, 2.8 Hz, CH), 3.93 (1H, ABXY, J = 11.2, 5.8, 2.8 Hz, CH2), 3.80 (1H,
ABXY, J = 11.2, 6.1, 5.8 Hz, CH2). 13C NMR (151 MHz, H2O/D2O 9:1, pH 7)
δC 179.6 (C1), 73.5 (C2), 67.5 (d, J = 4.4 Hz, C3). 31P NMR (161 MHz, H2O/D2O
9:1, pH 7) δP 4.0 (t, J = 5.8 Hz).
22. Guzman, M. I. & Martin, S. T. Prebiotic metabolism: production by mineral
photoelectrochemistry of α-ketocarboxylic acids in the reductive tricarboxylic
acid cycle. Astrobiology 9, 833–842 (2009).
23. Kolb, V. & Orgel, L. E. Phosphorylation of glyceric acid in aqueous solution
using trimetaphosphate. Orig. Life Evol. Biosph. 26, 7–13 (1996).
24. Pasek, M. A., Kee, T. P., Bryant, D. E., Pavlov, A. A. & Lunine, J. I. Production of
potentially prebiotic condensed phosphates by phosphorus redox chemistry.
Angew. Chem. Int. Ed. 47, 7918–7920 (2008).
25. Krishnamurthy, R., Arrhenius, G. & Eschenmoser, A. Formation of
glycolaldehyde phosphate from glycolaldehyde in aqueous solution. Orig. Life
Evol. Biosph. 29, 333–354 (1999).
26. Rabinowitz, J., Lores, J., Krebsbach, R. & Rogers, G. Peptide formation in the
presence of linear or cyclic polyphosphates. Nature 224, 795–796 (1969).
27. Saffhill, R. Selective phosphorylation of the cis-2′,3′-diol of unprotected
ribonucleosides with trimetaphosphate in aqueous solution. J. Org. Chem. 36,
2881–2883 (1970).
Glycolaldehyde phosphate aminonitrile (16). Yield 90% (pH 9.5, 100 h, RT).
1H NMR (600 MHz, H2O/D2O 9:1, pH 9.5) δH 4.09 (1H, t, J = 5.3 Hz, CH), 3.91
(1H, ABXY, J = 10.3, 5.9, 5.3 Hz, CH2), 3.85 (1H, ABXY, J = 10.3, 5.9, 5.3 Hz, CH2).
13C NMR (151 MHz, H2O/D2O 9:1, pH 9.5) δC 122.0 (C1), 65.4 (d, J = 4.0 Hz,
C3), 44.5 (d, J = 8.4 Hz, C2). 31P NMR (161 MHz, H2O/D2O 9:1, pH 9.5)
δP 3.8 (t, J = 5.9 Hz).
Phosphoserinamide (17). Yield 65% (pH 9.5, 3 d, RT). 1H NMR (400 MHz,
H2O/D2O 9:1, pH 9.5) δH 3.86 (1H, ABXY, J = 11.6, 5.4, 5.2 Hz, CH2), 3.84
(1H, ABXY, J = 11.6, 5.4, 5.2 Hz, CH2), 3.65 (1H, t, J = 5.2 Hz, CH). 13C NMR
(151 MHz, H2O/D2O 9:1, pH 9.5) δC 178.5 (C1), 66.5 (d, J = 4.4 Hz, C3), 55.4
(d, J = 7.7 Hz, C2). 31P NMR (161 MHz, H2O/D2O 9:1, pH 7) δP 4.1 (t, J = 5.4 Hz).
28. Mullen, L. B. & Sutherland, J. D. Formation of potentially prebiotic amphiphiles
by reaction of β-hydroxy-n-alkylamines with cyclotriphosphate. Angew. Chem.
Int. Ed. 46, 4166–4168 (2007).
29. Orgel, L. E. Prebiotic chemistry and the origin of the RNA world. Crit. Rev.
Biochem. Mol. Biol. 39, 99–123 (2004).
30. Keefe, A. D. & Miller, S. L. Are polyphosphates or phosphate esters prebiotic
reagents? J. Mol. Evol. 41, 693–702 (1995).
Phosphoserine (5-3P). Yield 36% (pH 9.5, 3 days, RT, then pH 12, 3 days, 75 °C).
1H NMR (400 MHz, H2O/D2O 9:1, pH 12) δH 3.93 (1H, ABXY, J = 10.1, 4.5, 4.5 Hz,
CH2), 3.81 (1H, ABXY, J = 10.1, 6.2, 4.5 Hz, CH2), 3.46 (1H, dd, J = 6.2, 4.5 Hz,
CH). 13C NMR (151 MHz, H2O/D2O 9:1, pH 12) δC 181.3 (C1), 67.9 (d, J = 5.0 Hz,
C3), 57.6 (d, J = 7.7 Hz, C2). 31P NMR (161 MHz, H2O/D2O 9:1, pH 12) δP 4.4
(t, J = 4.5 Hz).
31. Sutherland, J. D., Mullen, L. B. & Buchet, F. F. Potentially prebiotic Passerini-
type reactions of phosphates. Synlett 14, 2161–2163 (2008).
32. Krishnamurthy, R., Guntha, S. & Eschenmoser, A. Regioselective α-
phosphorylation of aldoses in aqueous solution. Angew. Chem. Int. Ed. 39,
2281–2285 (2000).
33. Muller, D. et al. Chemie von α-Aminonitrilen. Aldomerisierung von
Glycolaldehyd-phosphat zu racemischen Hexose-2,4,6-triphosphaten und
(in Gegenwart von Formaldehyd) racemischen Pentose-2,4-diphosphaten:
rac-Allose-2,4,6-triphosphat und rac-Ribose-2,4-diphosphat sind die
Reaktionshauptprodukte. Helv. Chim. Acta 73, 1410–1468 (1990).
34. Eschenmoser, A. Chemical etiology of nucleic acid structure. Science 284,
2118–2124 (1999).
35. Corey, E. J., Gilman, N. W. & Ganem, B. E. New methods for the oxidation
of aldehydes to carboxylic acids and esters. J. Am. Chem. Soc.
90, 5616–5617 (1968).
36. Goldman, N., Reed, E. J., Fried, L. E., William Kuo, I.-F. & Maiti, A. Synthesis of
glycine-containing complexes in impacts of comets on early Earth. Nat. Chem. 2,
949–954 (2010).
37. Öberg, K. I. et al. The comet-like composition of a protoplanetary disk as
revealed by complex cyanides. Nature 520, 198–201 (2015).
38. Osterberg, R. Origins of metal ions in biology. Nature 249, 382–383 (1974).
39. Braterman, P. S., Cairns-Smith, A. G. & Sloper, R. W. Photo-oxidation of
hydrated Fe2+-significance for banded iron formations. Nature 303,
163–164 (1983).
40. Liu, R. & Orgel, L. E. Oxidative acylation using thioacids. Nature 389, 52–54
(1997).
41. Keefe, A. D. & Miller, S. L. Was ferrocyanide a prebiotic reagent? Orig. Life Evol.
Biosph. 26, 111–129 (1996).
Received 8 May 2016; accepted 23 August 2016;
published online 10 October 2016
References
1. Eschenmoser, A. & Loewenthal, E. Chemistry of potentially prebiological natural
products. Chem. Soc. Rev. 21, 1–16 (1992).
2. Jeong, H., Tombor, B., Albert, R., Oltvai, Z. N. & Barabási, A. L. The large-scale
organization of metabolic networks. Nature 407, 651–654 (2000).
3. Oparin, A. I. The Origin of Life (MacMillan, 1938).
4. Huber, C. & Wächterhäuser, G. Activated acetic acid by carbon fixation on (Fe,
Ni)S under primordial conditions. Science 276, 245–247 (1997).
5. Miller, S. L., Schopf, J. W. & Lazcano, A. Oparin’s ‘origin of life’: sixty years later.
J. Mol. Evol. 44, 351–353 (1997).
6. Morowitz, H. J., Kostelnik, J. D., Yang, J. & Cody, G. D. The origin of
intermediary metabolism. Proc. Natl Acad. Sci. USA 97, 7704–7708 (2000).
7. Zhang, X. V. & Martin, S. T. Driving parts of Krebs cycle in reverse through
mineral photochemistry. J. Am. Chem. Soc. 128, 16032–16033 (2006).
8. Eschenmoser, A. On a hypothetical generational relationship between HCN and
constituents of the reductive citric acid cycle. Chem. Biodiv. 4, 554–573 (2007).
9. Orgel, L. E. The implausibility of metabolic cycles on the prebiotic Earth. PLoS
Biol. 6, 5–13 (2008).
10. Cooper, G., Reed, C., Nguyen, D., Carter, M. & Wang, Y. Detection and
formation scenario of citric acid, pyruvic acid, and other possible metabolism
precursors in carbonaceous meteorites. Proc. Natl Acad. Sci. USA 108,
14015–14020 (2011).
42. Bowler, F. R. et al. Prebiotically plausible oligoribonucleotide ligation facilitated
by chemoselective acetylation. Nat. Chem. 5, 383–389 (2013).
43. Rudnick, R. L. & Fountain, D. M. Nature and composition of the continental
crust: a lower crustal perspective. Rev. Geophys. 33, 267–309 (1995).
44. Post, J. E. Manganese oxide minerals: crystal structures and economic and
environmental significance. Proc. Natl Acad. Sci. USA 96, 3447–3454 (1997).
45. Tebo, B. M. et al. Biogenic manganese oxides: properties and mechanisms of
formation. Annu. Rev. Earth Planet. Sci. 32, 287–328 (2004).
46. Hazen, R. M. et al. Mineral evolution. Am. Mineral. 93, 1693–1720 (2008).
47. Maynard, J. B. The chemistry of manganese ores through time: a signal
of increasing diversity of Earth-surface environments. Econ. Geol.
105, 535–552 (2010).
11. Sagi, V. N., Punna, V., Hu, F., Meher, G. & Krishnamurthy, R. Exploratory
experiments on the chemistry of the ‘glyoxylate scenario’: formation of
ketosugars from dihydroxyfumarate. J. Am. Chem. Soc. 134, 3577–3589 (2012).
12. Patel, B. H., Percivalle, C., Ritson, D. J., Duffy, C. D. & Sutherland, J. D. Common
origins of RNA, protein and lipid precursors in a cyanosulfidic protometabolism.
Nat. Chem. 7, 301–307 (2015).
13. De Duve, C. Singularites: Landmarks on the Pathway of Life (Cambridge Univ.
Press, 2005).
14. Powner, M. W. & Sutherland, J. D. Prebiotic chemistry: a new modus operandi.
Phil. Trans. R. Soc. B 366, 2870–2877 (2011).
15. Nelson, D. L. & Cox, M. M. in Lehninger Principles of Biochemistry 4th edn,
Ch. 13, 493–497 (W. H. Freeman 2004).
48. Trail, D., Watson, E. B. & Tailby, N. D. The oxidation state of Hadean magmas
and implications for early Earth’s atmosphere. Nature 480, 79–82 (2011).
49. Hazen, R. Paleomineralogy of the Hadean eon: a preliminary species list. Am. J.
Sci. 13, 807–843 (2013).
50. Anbar, A. D. & Holland, H. D. The photochemistry of manganese and the origin
of banded iron formations. Geochim. Cosmochim. Acta 56, 2595–2603 (1992).
51. Chatgilialoglu, C., Crich, D., Komatsu, M. & Ryu, I. Chemistry of acyl radicals.
Chem. Rev. 99, 1991–2069 (1999).
16. Walsh, C. T., Benson, T. E., Kim, D. H. & Lees, W. J. The versatility of
phosphoenolpyruvate and its vinyl ether products in biosynthesis. Chem. Biol.
3, 83–91 (1996).
17. Potter, S. & Fothergill-Gilmore, L. A. Molecular evolution: the origin of
glycolysis. Biochem. Mol. Biol. Educ. 21, 45–48 (1993).
18. Powner, M. W., Gerland, B. & Sutherland, J. D. Synthesis of activated
pyrimidine ribonucleotides in prebiotically plausible conditions. Nature 459,
239–242 (2009).
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
7
© 2016 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

NATURE CHEMISTRY DOI: 10.1038/NCHEM.2624
ARTICLES
52. Kasting, J. F., Holland, H. D. & Pinto, J. R. Oxidant abundances in rainwater and
the evolution of atmospheric oxygen. J. Geophys. Res. 90, 10497–10510 (1987).
53. Forsythe, J. G. et al. Ester-mediated amide bond formation driven by wet–dry
cycles: a possible path to polypeptides on the prebiotic Earth. Angew Chem. Int.
Ed. 54, 9871–9875 (2015).
54. Kim, Y. S., Wo, K. P., Maity, S., Atreya, S. K. & Kaiser, R. I. Radiation-induced
formation of chlorine oxides and their potential role in the origin of Martian
perchlorates. J. Am. Chem. Soc. 135, 4910–4913 (2013).
55. Leshin, L. A. et al. Volatile, isotope, and organic analysis of Martian fines with
the Mars Curiosity rover. Science 341, 6153–6162 (2013).
56. Solomon, S., Sanders, R. W., Garcia, R. R. & Keys, J. G. Increased chlorine
dioxide over Antarctica caused by volcanic aerosols from Mount Pinatubo.
Nature 363, 245–248 (1993).
62. Keefe, A. D., Newton, G. L. & Miller, S. L. A possible prebiotic synthesis of
pantetheine, a precursor to coenzyme A. Nature 373, 683–685 (1995).
63. Nargoski, R. W. & Richard, J. P. Mechanistic imperatives for aldose–ketose
isomerization in water: specific, general base- and metal ion-catalyzed
isomerization of glyceraldehyde with proton and hydride transfer. J. Am. Chem.
Soc. 123, 794–802 (2001).
Acknowledgements
In memoriam of Harry Lonsdale. This work was supported by the Simons Foundation
(31881), the Engineering and Physical Sciences Research Council (EPSRC (EP/K004980/1)),
the Leverhulme Trust (RGP-2013-189) and an award from the Origin of Life Challenge.
57. Jackson, W. A. et al. Widespread occurrence of (per)chlorate in the Solar System. Author contributions
Earth Planet Sci. Lett. 430, 470–476 (2015).
The research was conceived by M.W.P. Experiments were conducted by A.J.C. Both authors
58. Halperin, J. & Taube, H. The transfer of oxygen atoms in oxidation–reduction
reactions. III. The reaction of halogenates with sulfite in aqueous solution. J. Am.
Chem. Soc. 74, 375–380 (1952).
59. Bal, B. S., Childers, W. E. & Pinnick, H. W. Oxidation of α,β-unsaturated
aldehydes. Tetrahedron 37, 2091–2096 (1981).
60. Pascal, R., Taillades, J. & Commeyras, A. Strecker’s and related systems. IX.
Acetone as catalyst for the hydration of tertiary α-aminonitriles in aqueous basic
solution. Bull. Soc. Chim. Fr. 2, 177–184 (1978).
contributed to the design and analysis of experiments, and to writing the paper.
Additional information
Supplementary information and chemical compound information are available in the
online version of the paper. Reprints and permissions information is available online at
www.nature.com/reprints. Correspondence and requests for materials should be
addressed to M.W.P.
61. Nwaukwa, S. O. & Keehn, P. M. Oxidative cleavage of α-diols, α-diones, α-
hydroxyketones and α-hydroxy- and α-ketone-acids with calcium hypochlorite. Competing financial interests
Tetrahedron Lett. 23, 3135–3138 (1982).
The authors declare no competing financial interests.
8
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
© 2016 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.