Prebiotic selection and assembly of proteinogenic amino acids and natural nucleotides from complex mixtures

Article abstract of DOI:10.1038/nchem.2703

A central problem for the prebiotic synthesis of biological amino acids and nucleotides is to avoid the concomitant synthesis of undesired or irrelevant by-products. Additionally, multistep pathways require mechanisms that enable the sequential addition of reactants and purification of intermediates that are consistent with reasonable geochemical scenarios. Here, we show that 2-aminothiazole reacts selectively with two- and three-carbon sugars (glycolaldehyde and glyceraldehyde, respectively), which results in their accumulation and purification as stable crystalline aminals. This permits ribonucleotide synthesis, even from complex sugar mixtures. Remarkably, aminal formation also overcomes the thermodynamically favoured isomerization of glyceraldehyde into dihydroxyacetone because only the aminal of glyceraldehyde separates from the equilibrating mixture. Finally, we show that aminal formation provides a novel pathway to amino acids that avoids the synthesis of the non-proteinogenic α,α-disubstituted analogues. The common physicochemical mechanism that controls the proteinogenic amino acid and ribonucleotide assembly from prebiotic mixtures suggests that these essential classes of metabolite had a unified chemical origin.

Full text of DOI:10.1038/nchem.2703

ARTICLES
PUBLISHED ONLINE: 16 JANUARY 2017 | DOI: 10.1038/NCHEM.2703
Prebiotic selection and assembly of proteinogenic
amino acids and natural nucleotides from
complex mixtures
Saidul Islam, Dejan-Krešimir Bučar and Matthew W. Powner
*
A central problem for the prebiotic synthesis of biological amino acids and nucleotides is to avoid the concomitant
synthesis of undesired or irrelevant by-products. Additionally, multistep pathways require mechanisms that enable the
sequential addition of reactants and purification of intermediates that are consistent with reasonable geochemical
scenarios. Here, we show that 2-aminothiazole reacts selectively with two- and three-carbon sugars (glycolaldehyde and
glyceraldehyde, respectively), which results in their accumulation and purification as stable crystalline aminals. This
permits ribonucleotide synthesis, even from complex sugar mixtures. Remarkably, aminal formation also overcomes the
thermodynamically favoured isomerization of glyceraldehyde into dihydroxyacetone because only the aminal of
glyceraldehyde separates from the equilibrating mixture. Finally, we show that aminal formation provides a novel pathway
to amino acids that avoids the synthesis of the non-proteinogenic α,α-disubstituted analogues. The common
physicochemical mechanism that controls the proteinogenic amino acid and ribonucleotide assembly from prebiotic
mixtures suggests that these essential classes of metabolite had a unified chemical origin.
he conservation of the genetic code, amino acids and nucleo- prior to the reaction of triose sugars with 4, the predominant
tides in biology suggests a single origin of life on Earth1–13
.
product is the non-natural branched apiose 5 rather than the
T
Proteins are built from a highly restricted set of about desired product 1 (69% 5, 31% 1 (Supplementary Fig. 17)).
20 amino acids according to a universal triplet code of four ribonu- Moreover, the dominant ketose isomer problem is even more acute
cleotides. Therefore, it is essential to learn how this specific small in C2 and C3 sugar mixtures. Incubation of 2a and 2b (1:1) in phos-
constellation of molecules became irrevocably linked at the advent phate buffer yields 2a/2b/2q (1:0.11:0.89). Reaction of this mixture
of life1–21. In contrast to the narrow distribution of universal metab- with 4 (1 equiv.) returns 1 (11%) as a very minor component of a
olites observed in biology, typical prebiotic reactions are notorious complex product mixture that consists predominantly of the undesir-
for their complex product distributions. Accordingly, it has been able apiose aminooxazoline (5, 12%)) and tetrose aminooxazoline (6,
recognized that “the chief obstacle to understanding the origin of 70% (Supplementary Fig. 19)). These observations demonstrate the
RNA-based life is identifying a plausible mechanism for overcoming importance of identifying a process that could sequentially deliver
the clutter wrought by prebiotic chemistry”4. For example, the most- 2a and 2b in a pure form to support the recently proposed
efficient and specific proposed prebiotic pathway to the pyrimidine prebiotic synthesis of activated pyrimidine ribonucleotides3,24
ribonucleotides requires the synthesis of the key intermediate
.
pentose aminooxazoline (1) (Fig. 1a)2–6,22–24. However, the plausi- Results
bility of this proposed prebiotic synthesis of 1 has been questioned Selective extraction of glycolaldehyde (2a) and glyceraldehyde
because it is contingent on the strictly controlled sequential delivery (2b) from complex sugar mixtures. All the known prebiotic
of pure glycolaldehyde (2a) to cyanamide (3) to yield 2-aminooxa- syntheses of sugars, including the formose reaction and the
zole (4) and then of pure glyceraldehyde (2b) to 4 to yield the Kiliani–Fischer process, lead to mixtures of glycolaldehyde (2a)
desired product (1) (Fig. 1a). This is a serious problem because and glyceraldehyde (2b) together with other C4, C5 and C6
both these reactions lack the intrinsic selectivity required to yield sugars1–7,13,15–21. A separate synthesis of pure 2a and 2b is
exclusively their respective products (4 and 1) from mixtures of ostensibly implausible, and therefore
a means of physical
2a and 2b. The problem becomes increasingly worse in the presence separation is better suited to promote the desired selective and
of other sugars. Without a separate and sequential delivery of 2a and sequential delivery of these simple sugars6,14,28. Our attention was
2b, a complex mixture of undesirable by-products results en route to first drawn to the possible role of 2-aminothiazole (7) in
the canonical nucleotides4–7,19,23,25–27. To exacerbate matters further, mediating the sequestration, separation and accumulation of 2a
aldose glyceraldehyde (2b) is thermodynamically unstable with and 2b when we observed the precipitation of 2a and homochiral
respect to its ketose isomer dihydroxyacetone (2q). Triose equili- D-glyceraldehyde (D-2b) from water as aminals 8a and D-8b
bration is catalysed by specific base, general acid–base and metal- (ref. 14). As discussed below, 7 is expected to be abundant in a
ion (such as Zn2+) catalysis13,27, and equilibration yields more of 2q cyanosulfidic environment29. We now report that 7-induced
than the aldose isomer glyceraldehyde (2b) (2q/2b = 8.5:1; 25 °C, crystallization resolves several long-standing problems inherent to
pH 7 (Supplementary Figs 1–4)). Under the conditions required for ribonucleotide selection. We demonstrate the time-resolved
the formation of 4 (ref. 3), in aqueous solution 2b equilibrates very sequestration of 2a and 2b from complex sugar mixtures as
rapidly (<0.5 h) with its ketose isomer dihydroxyacetone (2q) (11% separate stable crystal reservoirs in the form of the corresponding
2b, 89% 2q (Supplementary Fig. 2)). If this equilibration occurs aminals 8a and 8b. Aminals 8a and 8b allow for sequential
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
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

NATURE CHEMISTRY DOI: 10.1038/NCHEM.2703
ARTICLES
a
b
Selective sequential
crystallization
N
O
O
2q
2b
NH2
C4 + C5 + C6
C2 + C3 + C4 + C5 + C6
Aldoses and ketoses
NH2
H2N
N
O
N
(ref. 26)
(ref. 25)
N
H2N
3
OH
O
N
OH
+
NH2
S
2a
O
HO
O
7
N
O
N
N
N
2a
2q
HO
HO
NH2
O
NH2
O
O
NH2
(refs 19,23)
HO
NH2
O
HO
5
6
4
1
8a
8b
HO
OH
OH
O
H2NCN
(3)
Selective
N
O
OH
N
N
S
crystallization
S
2q
(refs 5,27)
O
HN
R
NH2
2b
4
HN
HO
HO
rac-threo-5
O
O
B
N
8a R=CH2OH
rac-8b R=CHOHCH2OH
(ref. 3)
NH2
B = cyt
B = ura
O
HO
O
O
D-ribo-1
P
1
–
O
O
Figure 1 | Prebiotic ribonucleotide synthesis. a, Stepwise nucleotide synthesis from pure reagents (previous work). Pyrimidine ribonucleotide synthesis
(black arrows) requires the stepwise addition of glycolaldehyde (2a) and glyceraldehyde (2b) to avoid deleterious by-product formation (red arrows). Bottom
left: single-crystal X-ray structure of rac-apiose threo-furanosyl aminooxazoline (rac-threo-5). cyt, cytosine; ura, uracil. b, Nucleotide synthesis from complex
sugar mixtures (this work). Crystallization-controlled synthesis of ribose aminooxazoline (ribo-1) from a sugar mixture that contains an equilibrating mixture
of 2a, 2b, dihydroxyacetone (2q), L-erythrulose (2w), L-arabinose, D-lyxose, D-ribose, D-xylose, D-glucose, D-galactose, D-mannose, D-fructose and D-sorbose
mediated by sequential and separate 2-aminothiazole (7)-induced crystallization of glycolaldehyde aminal 8a and rac-glyceraldehyde aminal 8b. Centre:
single-crystal X-ray structures acquired directly from the sequential precipitates of C2-aminal 8a and racemic C3-aminal rac-8b from a complex sugar
mixture, and homochiral D-α-ribo-furanosyl aminooxazoline (D-ribo-1) isolated from bulk conglomerate crystals obtained from the reaction of C3-aminal
(rac-8b) with 2-aminooxazole (4), synthesized from the exposure of C2-aminal 8a to cyanamide (3).
formation of the key intermediates 2-aminooxazole (4) and pentose near-quantitative yield (>95%) from the cysteine precursor
aminooxazoline (1) from a highly complex mixture of C2, C3, C4, C5 β-mercaptoacetaldehyde (9a) and cyanamide (3) (Fig. 2). A very
and C6 sugars (Fig. 1b). Furthermore, the unprecedented sequestration high conversion was observed even in stoichiometric competition
of racemic aminal rac-8b removes 2b from the aldose/ketose with 26 other aldehyde, ketone and sugar species present in the
equilibrium, which overturns the inherent thermodynamic bias mixture (including 2a–2w and C2, C3, C4, C5 and C6 sugars
towards the ketose isomer dihydroxyacetone (2q)5,13,27
.
(Supplementary Figs 21 and 22)). Furthermore, because the
For 7 to play the role of ‘chemical chaperone’ for ribonucleotide addition of thiols to 3 is reversible, we also explored the synthesis
synthesis in a chemically complex environment it must have a plaus- of 7 from 9a in the presence of five other stoichiometric sulfides/
ible prebiotic synthesis of its own. Therefore, we examined the thiols (9b–9f (Fig. 2)) and observed a highly selective reaction; 7
synthesis of 7 in complex sugar mixtures. 7 was synthesized in was obtained in 86% yield. Moreover, isothiourea (10) undergoes
an efficient thiol exchange with 9a to furnish 7 in 72% yield
R
SH
under mild aqueous conditions (pH 7, room temperature (r.t.),
16 hours (Fig. 2)). Finally, we were drawn to investigate the synthesis
of 7 from prebiotically plausible disulfides29. It is known that
cyanide efficiently reduces disulfides30. Thus, the reaction of disul-
fide 2n with ammonium cyanide (11) furnishes 7 (14% yield),
whereas the reaction of disulfide 2n with cyanide and 3 gave 7 in
up to 75% yield (Supplementary Fig. 23). The facile and predisposed
assembly of 7 from a complex mixture of aldehydes, ketones, sul-
fides and thiols in water suggests that it is a highly apposite
reagent in the search for prebiotic selectivity (Fig. 2).
With 7 established as a plausible and highly robust component of
prebiotic cyanosulfidic chemical environments, we began to explore
C2 and C3 sugar separation, which is required for selective ribonu-
cleotide synthesis (Fig. 1a)3. We incubated 2a (0.25 M), 2b (0.25 M)
and 7 (0.5 M) in phosphate buffer (pH 7, 25 °C) and monitored the
reaction over 24 hours. Initially (0–6 hours), we observed a mixture
of aminals 8a and 8b. However, surprisingly, after 12 hours a com-
plete resolution of the C2 and C3 sugars was observed: 8a was the
only aminal isolated (77%, 24 hours (Fig. 3a and Supplementary
Table 3)) and all the C3 sugars remained in solution. To distinguish
between kinetic and thermodynamic effects in the selective for-
mation of the C2-aminal 8a, we incubated 2a with the C3-aminal
8b (1:1, 25 °C, three days, pH 7) in phosphate buffer and observed
an accumulation of the C2-aminal 8a in 56% yield. Conversely, incu-
bation of 2b with the C2-aminal 8a (1:1, 25 °C, three days, pH 7) in
phosphate buffer did not form the C3-aminal 8b—only the triose
H2NCN
N
S
9a R = CH2CHO
9b R = H
(3)
NH2
9c R = CH3
7
9d R = CH2CH3
9e R = (CH2)2CH3
9f R = (CH2)3CH3
H2NCN
H2N
(3)
9a
Me SH
NH
NC
9c
Me
S
10
NH4CN
O
O
O
(11)
RS CN
RSH
NH3
RS CN
S
S
SH
9a
RS
2n R = CH3
12 R = CH2CHO
Figure 2 | Multiple pathways to 2-aminothiazole (7) in aqueous
cyanosulfidic solution. Facile and selective synthesis of 7 at neutral pH from
9a (magenta) and cyanamide (3) (black arrow) and selective assembly of
7 is observed in competition with prebiotically plausible thiols (9b–9f).
9a reacts efficiently with methylisothiourea (10) to furnish 7 (red arrows),
which demonstrates a reversible thiol-exchange during the synthesis of 7.
Finally, disulfides 2n and 12 are observed to react with ammonium cyanide
(11) to afford 7 (blue arrows) under Strecker conditions. These pathways
collectively demonstrate the facile and selective assembly of 7 under mild,
aqueous cyanosulfidic conditions.
2
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

NATURE CHEMISTRY DOI: 10.1038/NCHEM.2703
ARTICLES
a
N
a
b
S
N
HN
2a
2b
OH
N
S
N
O
H
2q
2q
OH
2a
S
NH2
8a
(7)
2–
OH
O
HPO4
HO
OH
OH
O
2q
// //
2b
2a
2b
X
2a
2b
SH
N
O
9a
H2NCN
(3)
S
// //
N
HN
c
S
N
OH
2–
HO
OH
H
HPO4
OH
O
8b
2q
b
//
//
//
//
d
OH
OH
OH
O
O
2w/2c/2q
OH
2b
OH
2c
OH
OH
e
f
7
HO
OH
HO
2–
HPO4
O
2q
O
2w
// //
OH
8b
O
OH
2b
Figure 3 | Selective C2-aminal 8 sequestration from a mixture of C2 and
C3 sugars. a, Synthesis of glycolaldehyde aminal 8a from a mixture of 2a
(1 equiv.), 2b (1 equiv.) and 7 (2 equiv.) (red arrow). Selective multicomponent
glyceraldehyde aminal 8b synthesis from dihydroxyacetone (2q),
β-mecaptoacetaldehyde (9a) and cyanamide (3) (blue arrow), and selective
conversion of 8b into 8a on incubation with 2a (1 equiv.) in pH 7 phosphate
buffer (green arrows) demonstrate the thermodynamically controlled selection
of 8a and 2q from mixtures of C2 and C3 sugars. b, C3-specific aldose
sequestration by the crystallization of aminal 8b from an equilibrating C3- and
C4-sugar mixture (2b, 2c, 2q and 2w) on reaction with 7 in phosphate
buffer (blue arrows) provides a facile and highly selective separation of C3
and C4 sugars.
// //
7.2
6.66.0
5.5
5.0
4.5
4.0
3.5
3.0
Chemical shift/ppm
Figure 4 | Convergent crystallization-controlled synthesis of pure ribo-1
from a complex mixture of sugars. a–f, 1H NMR spectra (600 MHz) of the
reaction of one equivalent each of 2a (red), rac-2b (green), L-erythrulose,
L-arabinose, D-lyxose, D-ribose, D-xylose, D-glucose, D-galactose, D-mannose,
D-fructose and D-sorbose incubated in phosphate buffer (1 M, pH 7, 25 °C)
with 7 (6 equiv.). The initial sugar mixture is shown in a, and the
equilibrated sugar mixture after three days (b) shows the conversion of 2b
(green, 4.90 ppm, doublet, J = 5.4 Hz) into 2q (green, 4.30 ppm, singlet).
The subsequent four spectra are from the crystalline precipitate 8a that
accumulated 0–2 h after the addition of 7 (blue) (c), the crystalline
precipitate 8b that accumulated 48–480 h after the addition of 7 (d), the
quantitative transformation of aminal 8a and cyanamide (3) to give a 1:2
mixture of 4 (red) and 7 (blue) (e) and the crystalline conglomerate D-ribo-1
isolated on exposure of precipitate 8b to a 1:2 mixture of 4 and 7 (f).
equilibration (2b ↔ 2q) was observed, even with a twofold excess of 2b.
Therefore, clearly the equilibration of 2b with its more-stable ketose
isomer dihydroxyacetone (2q) drives the selective accumulation of
the C2-aminal 8a. However, remarkably, incubation of 2q with either
7 or 9a plus 3 (to effect an in situ synthesis of 7) delivered aminal
rac-8b in up to 87% yield (Fig. 3a and Supplementary Table 6),
which thereby deposited a reservoir of the crystalline aldose isomer Selective synthesis of pentose aminooxazoline (1) from a complex
2b and completely overturned the thermodynamic preference for the sugar mixture. We recognized that incubation of glycolaldehyde (2a)
C3 ketose isomer observed in aqueous solution. In contrast, phos- and glyceraldehyde (2b) in phosphate buffer prior to aminal
phate-induced isomerization of tetrose 2c to erythrulose (2w) com- precipitation would result in a very large differential accumulation of
pletely inhibited precipitation of the C4-aminal 8c. Consequently, C2-aminal 8a and C3-aminal 8b, which would provide the direct
only C3-aminal 8b (56%) was observed to precipitate from a stoichio- physical mechanism to separate 2a and 2b spatially, which is so
metric mixture of 2q and 2w after incubation (23 days) with 7 crucial to the synthesis of pentose aminooxazoline (1). To test our
(6 equiv.) in phosphate buffer (1 M, 25 °C, pH 7), which resulted in a hypothesis, a model prebiotic sugar mixture containing C2, C3, C4,
robust resolution of C3 and C4 sugars through the unique crystallization C5 and C6 sugars (Fig. 4a) was incubated in phosphate buffer (1 M,
of aminal 8b (Fig. 3b and Supplementary Table 8).
pH 7, r.t., three days (Fig. 4b)) to allow triose equilibration, and then
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
3
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

NATURE CHEMISTRY DOI: 10.1038/NCHEM.2703
ARTICLES
Table 1 | Prebiotic aminonitrile synthesis.
NH2
R1
O
NH2
O
KCN/NH4OH
KCN/NH4OH
R1
NC
R
H
NC
H
R
R
R
2a–2n
2o–2w
13o–13w
13a–13n
X 7
7
NH3
S
N
N
N
HN
NC
S
S
OH
N
HN
N
HN
KCN
R1
NC
H
N
R1
S
S
R
N
H
R
R
R
H
H
15
8o–8w
16a–16n
8a–8n
2
R
R1
8 (%)
13 (%)
Amino acid
a
b
c
d
e
f
g
h
i
j
k
l
m
n
o
p
q
r
CH2OH
CH(OH)CH2OH (rac)
R,R-CH(OH)CH(OH)CH2OH
H
CH3
CH2CH3
CH(OH)CH3
CH(CH3)CH3
CH(CH3)CH2CH3
CH2CH(CH3)CH3
CH2CH2CN
CH2CH2CH2NHAc
CH2CH2SCH3
CH2SSCH3
CH3
CH2OH
CH2OH
CH2CH3
CH2CH2CH3
CH(CH3)CH3
CH(CH3)CH2CH3
CH2CH2CH2CH2CH3
CHOHCH2OH
H
H
H
H
H
H
H
H
H
H
H
H
H
H
95
90*
32
96
99
85
70
94
92
89
89
90
75
79
0
0
0
0
0
0
0
0
0
>95
−
−
>95
>95
>95
91
>95
72
70
90†
>95
70
−
Ser
−
−
Gly
Ala
–
Thr
Val
Ile
Leu
Glu/Gln
(Pro/Arg)‡
Met
Cys§
−
−
−
−
−
−
−
CH3
CH3
CH2OH
CH3
CH3
CH3
CH3
CH3
CH2OH
83
>95
>95
88
78
s
t
u
v
w
87
84
90
−
−
−
Non-selective synthesis of aminonitriles 13a–13w from aldehydes and ketones by the traditional Strecker pathway (black arrows; previous work). Aldehyde aminal 8 selective synthesis of aminonitriles 13a and
13d–13m (blue arrows; this work); aminonitriles 15 were not observed. Isolated yields of aminals 8a–8w from the reaction of aldehydes/ketones 2a–2w (500 mM) and 2-aminothiazole (7 (1 M, pH 7, 24 h))
are reported. Aminonitriles 13a–13n were obtained by equilibration of aminals 8a–8n with potassium cyanide (1.5 equiv.) and ammonium hydroxide (5 equiv.) at pH 9.2; NMR yields are reported.
Aminonitriles 13o–13w were acquired by the equilibration of ketones 2o–2w (1 equiv.), potassium cyanide (1.5 equiv.) and ammonium hydroxide (5 equiv.) at pH 9.2; NMR yields are reported.
*Homochiral D-8b has previously been crystallized from an ethanol/water mixture14
nitrogen atoms of the proline/arginine precursors and can be elaborated to the canonical amino acids after amide hydrolysis13
.
†A mixture of α-aminonitrile 13k (62%) and α-aminoamide (28%) was obtained. ‡8l contains the four contiguous carbon/
.
§8n is converted into the Strecker precursor of cysteine 9a on disulfide reduction.
2-aminothiazole (7 (6 equiv.)) was added. Over the course of two hours sequestration from the solution, which results in a residual sugar
the selective crystallization of pure aminal 8a (>95%) was observed mixture in the supernatant devoid of 2a and 2b. Interestingly, we
(Fig. 4c). Further incubation of the supernatant, after the isolation of observed that the addition of 3 to this supernatant C4, C5 and C6
pure glycolaldehyde aminal 8a, resulted in the crystallization of pure sugar mixture directly induced further crystallization of ribo-1
glyceraldehyde aminal 8b (61%, 2–17 days (Fig. 4d)). No further (Supplementary Fig. 35). Therefore, we have demonstrated two
crystallization was observed. The onset of the crystallization of 8b is complementary pathways to accrue pure crystalline ribo-1 en
slow (9% after 24 hours), and the large observed time resolution for route to ribonucleotides from the most-complex sugar mixtures
aminal precipitation has significant implications for the spatial explored in prebiotic chemistry, and both pathways are only
resolution, accumulation, concentration and subsequent reactions of enabled by the selective formation of aminal 8. Of note, the
aminals 8a and 8b. Indeed, exposure of aminal 8a—sequestered from pathway described using aminals 8a and 8b to build 1 obviates
a mixture of 12 homologous sugars—to cyanamide (3 (0.65 M)) gave a the requirement to access the notoriously unstable sugar ribose1–7,15
quantitative yield of 2-aminoxazole (4 (Fig. 4e)). Incubation of 4 with
.
aminal 8b, which crystallized after 8a from the same complex sugar Prebiotic selection of proteinogenic amino acids. The prebiotic
solution, led to the selective formation of pentose aminooxazoline (1), origins of amino acids have been investigated for over 60 years.
followed by the direct crystallization of pure ribo-1 (Fig. 4f).
However, no reported prebiotic synthesis or meteoritic amino acid
Having demonstrated the crystallization-controlled synthesis of sample provides the restricted set of amino acids assigned to the
ribo-1 from the simplest C2 and C3 components of a complex genetic code2,8–13. For example, recently Sutherland and co-
sugar mixture, we next turned our attention to the residual (C4, workers demonstrated the stepwise prebiotic syntheses of 12
C5 and C6) sugars. It has been reported previously that, because aminonitrile (13) proteinogenic amino acid precursors, but,
of their rapid rate of reaction, the simplest sugars, 2a and 2b, paradoxically, essential ketones—such as acetone (2o),
thwarted previous attempts to crystallize ribo-1 by the action of 3 monohydroxyacetone (2p) and dihydroxyacetone (2q)—are
on complex sugar mixtures22. However, our sequential 7-induced required during the assembly of the branched carbon framework
crystallization from an aqueous solution of C2, C3, C4, C5 and C6 of valine and leucine13. Ideally, ketones would be excluded from a
sugars exploits the reactivity of 2a and 2b to facilitate their prebiotic aminonitrile 13 synthesis because prebiotic ketones
4
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

NATURE CHEMISTRY DOI: 10.1038/NCHEM.2703
ARTICLES
ketones would have been present in prebiotic mixtures8–13. Thus,
to investigate 7-induced selective amino acid synthesis, a mixture
of eight prebiotic aldehydes and ketones 2 (Fig. 5) was incubated
with 7 (10 equiv.). The quantitative crystallization of aminals 8a,
8d, 8e and 8h was observed after 24 hours, which allowed all the
unwanted ketones 2 (and excess 7) to be removed with the super-
natant to leave behind only pure proteinogenic amino acid precur-
sors. The precipitate, a mixture of aminals 8a, 8d, 8e and 8h, was
then converted smoothly into the corresponding aminonitriles
13a, 13d, 13e and 13h on the addition of ammonium cyanide
(11) (Supplementary Fig. 48). Therefore, this reaction is the first
selective synthesis of solely natural α-hydrogen amino acids from
prebiotic aldehyde/ketone mixtures.
N
NH2
S
O
O
7
+
R
H
R
R1
2o R = R1 = Me
2a R = CH2OH
2d R = H
2e R = Me
2p R = CH2OH, R1 = Me
2q R = R1 = CH2OH
2t R = Me, R1 = CHMe2
2o
2p
2q
2t
2h R = CHMe2
NH2
Discussion
KCN/NH4OH
Quantitative
To avoid the concomitant synthesis of undesired or irrelevant
by-products alongside the desired biologically relevant molecules
is one of the central challenges to the development of plausible pre-
biotic chemistry1–7,12,13,15,19,22,32,33. Previous models have advocated
that kinetically controlled, segregated syntheses (under different
local geochemical conditions) are required to overcome the incom-
patibility of distinct reactions1–7. However, these models are necess-
arily highly contingent on the rapid exploitation of reagents as and
when they form. Accordingly, they are reliant on achieving a specific
and controlled order of synthetic steps under geochemical con-
straints and also they are incompatible with the accumulation or
purification of intermediates. Therefore, it is striking that we have
discovered a common physicochemical mechanism that controls
both α-hydrogen amino acid formation and ribonucleotide assem-
bly from prebiotically plausible mixtures. 2-Aminothiazole (7), in
conjunction with phosphate acting as a general acid–base catalyst,
+ 7
NC
H
R
8a (90%)
8d (99%)
8e (99%)
8h (89%)
13a R = CH2OH
13d R = H
13e R = Me
13h R = CHMe2
Figure 5 | High-yielding 2-aminothiazole (7)-controlled selective aldehyde
reactivity. Selective precipitation of four aldehyde aminals (8a, 8d, 8e and
8h) from a mixture of eight aldehydes and ketones (2a, 2d, 2e, 2h, 2o, 2p,
2q and 2t) on reaction with 7 followed by the directed aminonitrile 13
synthesis on equilibration of aminals 8a, 8d, 8e and 8h with ammonium
cyanide (1.5 equiv.) and ammonium hydroxide (5 equiv.) at pH 9.2
demonstrates a direct physicochemical mechanism for proteinogenic
amino acid selection.
undergo aminonitrile formation just as effectively as aldehydes, but facilitates an unprecedented sequential accumulation and purifi-
α,α-disubstituted amino acids are not genetically encoded (Table 1). cation of C2 and C3 aldoses (as stable crystalline aminals 8a and
To the best of our knowledge, there are no previously described 8b) from complex mixtures in the specific sequence required for
mechanisms to discriminate between aldehydes and ketones ribonucleotide assembly3. The time-resolved separation of 8a and
during aminonitrile synthesis. However, we recognized that 8b observed in our experiments suggests that simple flow systems,
aminal 8a is not only a C2 sugar, but also an amino acid (serine) such as a river, could provide a model prebiotic environment for
precursor, which prompted us to investigate whether the physical separation, accumulation and purification of 8a and
2-aminothiazole (7) could facilitate the separation of natural and 8b. The accumulation of 8b is achieved by dynamic resolution of
non-natural amino acid precursors, and thus unite the chemical the C3 sugars, such that the aldose isomer glyceraldehyde (2b) is
selection of amino acid and ribonucleotide precursors by a common sequestered from an unfavourable equilibrium with its thermodyna-
mechanism. Accordingly, we examined the reaction of 7 with the mically more-stable ketose isomer dihydroxyacetone (2q)27. This is
aldehydes and ketones 2d–2w, which are all prebiotic amino acid remarkable because the equilibration of 2b to 2q is not only highly
precursors2,8–13. All the aldehydes furnished their aminals (8d–8n) detrimental to stepwise ribonucleotide synthesis, but is also
in excellent yield (Table 1). In stark contrast, no ketone aminals especially rapid (<0.5 h (Supplementary Fig. 2)) under the con-
precipitated, even after extended incubation. This suggests that 7 ditions required for ribonucleotide assembly3, and there are no pre-
could be ideally suited to facilitate the facile separation of natural viously reported conditions that overcome the unfavourable
α-substituted amino acid precursors (aldehydes) and non-natural equilibrium that favours 2q (ref. 27). Moreover, the formation of
(α,α-disubstituted) amino acid precursors (ketones).
aminal 8b from 2q drives the recovery of the (first) chirogenic
A novel route to amino acids requires aminals 8 to participate centre (en route to sugars and ribonucleotides), and this recovery
directly in the synthesis of aminonitriles 13 without affording is coupled to crystallization. Therefore, we suggest that the crystal-
N-thiazolyl aminonitriles 15. Therefore, we were pleased to lizations outlined here may have further implications for the gener-
observe that the incubation of ammonium cyanide (11) with ation of homochiral ribonucleotides from achiral precursors, and
aminals 8a and 8d–8m gave a smooth conversion into aminonitriles investigations to exploit the crystallization of aminals 8 in chiral
13a and 13d–13m, respectively (Table 1). Even in the absence of amplification strategies are currently underway in our laboratory.
ammonia, N-thiazolyl aminonitriles 15 were not observed; only a Indeed, although rac-ribo-1 has only previously been observed to
quantitative formation of the relevant cyanohydrin 16 was noted form enantiomorphously twinned crystals (in which individual
(pH 3–10). Although the incubation of aminal 8b or 8c with crystals remain racemic but may contain homochiral domains)22,
cyanide yielded a smooth conversion to cyanohydrins 16b or 16c, we observed the crystallization of novel conglomerates of ribo-1
respectively (>95%, 25 °C, pD 7), aminonitriles 13b and 13c were from a synthesis mediated by aminal 8 (Fig. 1b), in which individual
not observed because of the rapid imidolactone formation crystals are purely homochiral. It is also striking that a previously
(Supplementary Fig. 85)31, which suggests their respective amino described mechanism for amplification and chirality transfer to
acids were not genetically encoded because of imidolactone-prohib- ribonucleotides—the phosphate-mediated conversion of ribo-1 in
ited aminonitrile synthesis.
to arabino-1 (ref. 24)—is mechanistically equivalent to the conver-
The breadth of amino acid precursors available on the early sion of 2b into 2q, which maximizes the resolution of C2 and C3
Earth remains unknown, but it is clear that both aldehydes and sugars by 7-induced crystallization.
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
5

NATURE CHEMISTRY DOI: 10.1038/NCHEM.2703
ARTICLES
13. 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).
14. Powner, M. W., Zheng, S. L. & Szostak, J. W. Multicomponent assembly of
proposed DNA precursors in water. J. Am. Chem. Soc. 134, 13889–13895 (2012).
15. Shapiro, R. Prebiotic ribose synthesis: a critical analysis. Origin Life Evol.
Biosphere 18, 71–85 (1988).
Given the close biological and generational relationship between
the proteinogenic amino acids and ribonucleotides, it is highly
significant that 7-induced crystallization provides an absolute chemical
selection for the natural amino acids (that is, those bearing an
α-hydrogen atom) from a complex mixture of Strecker aldehydes
and ketones 2 selected for their prevalence in abiotic chemistry.
Ketones such as dihydroxyacetone (2q), monohydroxyacetone (2p)
and acetone (2o) are required for the prebiotic synthesis of
branched-chain amino acids (such as valine and leucine) and lipid
precursors13. These ketones, and many others, all undergo effective
aminonitrile formation (Table 1), which leads to the prebiotic syn-
thesis of non-proteinogenic α,α-disubstituted amino acids alongside
the natural α-amino acids, and thus creates a puzzling dichotomy.
This has now been resolved by selective sequestration of the natural
α-amino acid precursors by accumulation of aminals 8. Restricting
the peptide sequence space is thought to have been advantageous
during the origins of life2,4, and 7 provides a simple chemical-selection
tool for α-hydrogen moieties that may underpin the later implemen-
tation of biological selection that would have refined (for example,
the exclusion of 13f) and expanded the amino acid repertoire to
16. Ricardo, A., Carrigan, M. A., Olcott, A. N. & Benner, S. A. Borate minerals
stabilize ribose. Science 303, 196 (2004).
17. Breslow, R. & Cheng, Z. L. L-Amino acids catalyze the formation of an excess of
D-glyceraldehyde, and thus of other D-sugars, under credible prebiotic
conditions. Proc. Natl Acad. Sci. USA 107, 5723–5725 (2010).
18. Sagi, V. S. et al. Exploratory experiments on the chemistry of the ‘glyoxylate
scenario’: formation of ketosugars from dihydroxyfumarate. J. Am. Chem. Soc.
134, 3577–3589 (2012).
19. Islam, S. et al. Detection of potential TNA and RNA nucleoside precursors in a
prebiotic mixture by pure shift diffusion-ordered NMR spectroscopy. Chem.
Eur. J. 19, 4586–4595 (2013).
20. Meinert, C. et al. Ribose and related sugars from ultraviolet irradiation of
interstellar ice analogs. Science 352, 208–212 (2016).
21. Becker, S. et al. A high-yielding, strictly regioselective prebiotic purine
nucleoside formation pathway. Science 352, 833–836 (2016).
22. Springsteen, G. & Joyce, G. F. Selective derivatization and sequestration of ribose
from a prebiotic mix. J. Am. Chem. Soc. 126, 9578–9583 (2004).
23. Anastasi, C., Crowe, M. A., Powner, M. W. & Sutherland, J. D. Direct assembly of
nucleoside precursors from two- and three-carbon units. Angew. Chem. Int. Ed.
45, 6176–6179 (2006).
24. Powner, M. W. & Sutherland, J. D. Phosphate-mediated interconversion of ribo- and
arabino-configured prebiotic nucleotide intermediates. Angew. Chem. Int. Ed. 49,
4641–4643 (2010).
25. Powner, M. W., Sutherland, J. D. & Szostak, J. W. Chemoselective
multicomponent one-pot assembly of purine precursors in water.
J. Am. Chem. Soc. 132, 16677–16688 (2010).
26. Saul, R. et al. Reaction of 1,3-disubstituted acetone derivatives with
pseudohalides: a simple approach to spiro[4.4]nonane-type bis-oxazolidines and
-imidazolidines (bicyclic carbamates, thiocarbamates, ureas, and thioureas).
Eur. J. Org. Chem. 1, 205–209 (2000).
provide further selective advantages4,34
.
Finally, and importantly, in particular 7 does not inhibit the
formation of pentose aminooxazoline (1) by Mannich-type
reactivity14,25,35 or the formation of the canonical α-amino acids
through participation in a Strecker-type reactivity. Accordingly, it
is remarkable that 7 can react with ribonucleotide and amino acid
aldehyde precursors and facilitate their purification and accumu-
lation from prebiotic mixtures. These features make 7 an ideal
chemical chaperone for prebiotic multistep syntheses.
Data availability statement. The authors declare the data that
support the findings of this study are available within the paper
and its Supplementary Information files. X-ray crystallographic
data were also deposited at the Cambridge Crystallographic Data
Centre (CCDC) under the following CCDC deposition numbers:
D-ribo-1 (1477052, conglomerate), rac-threo-5 (1477054), aminals
8a (1477040), D-8b (1477041), L-8b (1477042), rac-8b (1477045),
D-8c (1477043), 8d (1477044), 8e (1477046), 8f (1477047), 8g
(1477051) and 8m (1477048). These can be obtained free of
charge from CCDC via www.ccdc.cam.ac.uk/data_request/cif.
27. Nagorski, 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).
28. Budin, I. & Szostak, J. W. Expanding roles for diverse physical phenomena
during the origin of life. Annu. Rev. Biophys. 39, 245–263 (2010).
29. Sagan, C. & Khare, B. N. Long-wavelength ultraviolet photoproduction of amino
acids on the primitive Earth. Science 173, 417–420 (1971).
30. Catsimpoolas, N. & Wood, J. L. Specific cleavage of cystine peptides by cyanide.
J. Biol. Chem. 241, 1790–1796 (1966).
31. Serianni, A. S., Nunez, H. A. & Barker, R. Cyanohydrin synthesis: studies with
[
13C]cyanide. J. Org. Chem. 45, 3329–3341 (1980).
32. Schwartz, A. W. Intractable mixtures and the origin of life. Chem. Biodiv. 4,
656–664 (2007).
33. Schwartz, A. W. Evaluating the plausibility of prebiotic multistage syntheses.
Astrobiology 13, 784–789 (2013).
34. Knight, R. D. & Landweber, L. F. The early evolution of the genetic code. Cell
101, 569–572 (2000).
35. Hein, J. E., Tse, E. & Blackmond, D. G. A route to enantiopure RNA precursors
from nearly racemic starting materials. Nat. Chem. 3, 704–706 (2011).
Received 18 September 2016; accepted 22 November 2016;
published online 16 January 2017
References
1. Eschenmoser, A. Etiology of potentially primordial biomolecular structures:
from vitamin B12 to the nucleic acids and an inquiry into the chemistry of life’s
origin: a retrospective. Angew. Chem. Int. Ed. 50, 12412–12472 (2011).
2. Sutherland, J. D. The origin of life—out of the blue. Angew. Chem. Int. Ed. 55,
104–121 (2016).
Acknowledgements
3. Powner, M. W., Gerland, B. & Sutherland, J. D. Synthesis of activated pyrimidine
ribonucleotides in prebiotically plausible conditions. Nature 459, 239–242 (2009).
4. Joyce, G. F. The antiquity of RNA-based evolution. Nature 418, 214–221 (2002).
5. Benner, S. A., Kim, H.-Y. & Yang, Z. Setting the stage: the history, chemistry and
geology behind RNA. Cold Spring Harb. Perspect. Biol. 4, a003541 (2012).
6. Robertson, M. P. & Joyce, G. F. The origins of the RNA world. Cold Spring Harb.
Perspect. Biol. 4, a003608 (2012).
7. Krishnamurthy, R. On the emergence of RNA. Isr. J. Chem. 55, 837–850 (2015).
8. Johnson, A. P. et al. The Miller volcanic spark discharge experiment. Science
322, 404 (2008).
9. Chyba, C. & Sagan, C. Endogenous production, exogenous delivery and impact-
shock synthesis of organic molecules: an inventory for the origins of life. Nature
355, 125–132 (1992).
10. Miller, S. L. Production of amino acids under possible primitive Earth
conditions. Science 117, 528–529 (1953).
This work was supported by the Simons Foundation (318881), the Engineering and Physical
Sciences Research Council (EP/K004980/1), the Leverhulme Trust (RGP-2013-189) and
through an award from the Origin of Life Challenge (M.W.P.) and a UCL Excellence
Fellowship (D.-K.B.). The authors thank K. Karu for assistance with mass spectrometry and
A. E. Aliev for assistance with NMR spectroscopy.
Author contributions
M.W.P. conceived the research. M.W.P. and S.I. designed and analysed the experiments. S.I.
conducted the experiments. D.-K.B. performed the crystallographic analyses. M.W.P. and
S.I. wrote 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.
11. Pizzarello, S., Cooper, G. & Flynn, G. in Meteorites and the Early Solar System II
(eds Lauretta, D. S. & McSween, H. Y. Jr) 625–651 (Univ. Arizona Press, 2006).
12. Wolman, Y., Haverland, W. J. & Miller, S. L. Nonprotein amino acids from spark
discharges and their comparison with the Murchison meteorite amino acids.
Proc. Natl Acad. Sci. USA 69, 809–811 (1972).
Competing financial interests
The authors declare no competing financial interests.
6
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.