A prebiotically plausible synthesis of pyrimidine β-ribonucleosides and their phosphate derivatives involving photoanomerization

Article abstract of DOI:10.1038/nchem.2664

Previous research has identified ribose aminooxazoline as a potential intermediate in the prebiotic synthesis of the pyrimidine nucleotides with remarkable properties. It crystallizes spontaneously from reaction mixtures, with an enhanced enantiomeric excess if initially enantioenriched, which suggests that reservoirs of this compound might have accumulated on the early Earth in an optically pure form. Ribose aminooxazoline can be converted efficiently into α-ribocytidine by way of 2,2'-anhydroribocytidine, although anomerization to β-ribocytidine by ultraviolet irradiation is extremely inefficient. Our previous work demonstrated the synthesis of pyrimidine β-ribonucleotides, but at the cost of ignoring ribose aminooxazoline, using arabinose aminooxazoline instead. Here we describe a long-sought route through ribose aminooxazoline to the pyrimidine β-ribonucleosides and their phosphate derivatives that involves an extraordinarily efficient photoanomerization of α-2-thioribocytidine. In addition to the canonical nucleosides, our synthesis accesses β-2-thioribouridine, a modified nucleoside found in transfer RNA that enables both faster and more-accurate nucleic acid template-copying chemistry.

Full text of DOI:10.1038/nchem.2664

ARTICLES
PUBLISHED ONLINE: 21 NOVEMBER 2016 | DOI: 10.1038/NCHEM.2664
A prebiotically plausible synthesis of pyrimidine
β-ribonucleosides and their phosphate derivatives
involving photoanomerization
Jianfeng Xu1, Maria Tsanakopoulou1, Christopher J. Magnani1, Rafał Szabla2 , Judit E. Šponer2,3,
*
Jiří Šponer2,3, Robert W. Góra4 and John D. Sutherland1
*
Previous research has identified ribose aminooxazoline as a potential intermediate in the prebiotic synthesis of the
pyrimidine nucleotides with remarkable properties. It crystallizes spontaneously from reaction mixtures, with an enhanced
enantiomeric excess if initially enantioenriched, which suggests that reservoirs of this compound might have accumulated on
the early Earth in an optically pure form. Ribose aminooxazoline can be converted efficiently into α-ribocytidine by way of
2,2′-anhydroribocytidine, although anomerization to β-ribocytidine by ultraviolet irradiation is extremely inefficient. Our
previous work demonstrated the synthesis of pyrimidine β-ribonucleotides, but at the cost of ignoring ribose aminooxazoline,
using arabinose aminooxazoline instead. Here we describe a long-sought route through ribose aminooxazoline to the
pyrimidine β-ribonucleosides and their phosphate derivatives that involves an extraordinarily efficient photoanomerization of
α-2-thioribocytidine. In addition to the canonical nucleosides, our synthesis accesses β-2-thioribouridine, a modified
nucleoside found in transfer RNA that enables both faster and more-accurate nucleic acid template-copying chemistry.
he first synthesis of pyrimidine ribonucleosides from ami- nucleobase destruction, for subtle chemical reasons. However,
nooxazoline and anhydronucleoside intermediates dates although our synthesis provided both canonical pyrimidine ribonu-
T
back nearly 50 years, but suffers from an extremely low-yielding cleoside-2′,3′-cyclic phosphates 9 and 10 in good yield, we had
step and the questionable prebiotic availability of ribose1. Sanchez concerns about our reliance on arabinose aminooxazoline 7 as
and Orgel found that treatment of the cyanamide derivative of an intermediate.
this sugar, ribose aminooxazoline 1, with cyanoacetylene 2 gave
We were motivated further to explore this area of chemistry in
the anhydronucleoside 3 and thence α-ribocytidine 4 in good more detail because we had also found that pentose cyanamide
yield, but ultraviolet irradiation of the latter gave β-ribocytidine 5 derivatives can be formed in an indirect way that avoids the
in only 4% yield1. Although the route had several attractive features, requirement for the free pentoses, which are unstable and difficult
the inefficiency of this photoanomerization compounded concerns to access. Thus, all four pentose aminooxazolines are formed by
about the availability of ribose and made the route seem somewhat the reaction of glyceraldehyde 11 with 2-aminooxazole 12,
implausible from a prebiotic standpoint (Fig. 1). Both the high crys- another compound that has desirable purification capabilities as
tallinity of ribose aminooxazoline 1 and its greater propensity to it sublimes readily4. Although ribose aminooxazoline 1 (44%
crystallize than its stereoisomers were noted by Sanchez and yield) and arabinose aminooxazoline 7 (30% yield) greatly predo-
Orgel1. In general, this behaviour is extremely attractive in prebiotic minate, and are thus apparently equally plausible as ribonucleotide
chemistry because it allows for the spontaneous purification and precursors, 1 selectively crystallizes from the product solution.
accumulation of an intermediate such that subsequent synthetic Augmenting earlier observations that this crystallization provides
steps in a sequence can occur without interference from the detritus a potential means whereby ribose aminooxazoline 1 might be
of earlier steps. Accordingly, we were prompted to revisit this chem- chemically purified1, we also found that it allows enantiomeric
istry and, having found that nucleobase destruction to give ribose enrichment if the input glyceraldehyde is non-racemic. This is
oxazolidinone 6 was the major photoreaction of α-ribocytidine 4 because 1 behaves as a true conglomerate (different enantiomers
(ref. 2), thus concluded that this route to the β-ribonucleosides that form separate crystals or crystalline domains such that solid-
from ribose aminooxazoline 1 was irredeemable. We then moved phase enantiomeric excesses are enhanced over those in the
on to explore routes that involve 2′-stereoinversion and the residual material in solution)4. Although we also found that 1
arabino-configured intermediates arabinose aminooxazoline 7 and can be partly epimerized to arabinose aminooxazoline 7 in phos-
2,2′-anhydro-arabinocytidine 8 (ref. 3). We found that phosphoryl- phate buffer—presumably on the basis of the mechanism,
ation of 8 occurred regioselectively on the 3′-OH group and was without loss of enantiomeric purity—the latter is the minor stereo-
followed by intramolecular 2′-stereoinversion to give β-ribocyti- isomer at equilibrium5. Accordingly, we felt that enantiomerically
dine-2′,3′-cyclic phosphate 9. Finally, we discovered that ultraviolet pure ribonucleotides would most plausibly derive from ribose
irradiation of 9 caused a partial conversion into the corresponding aminooxazoline 1 if a route that avoided arabinose aminooxazoline
ribouridine-2′,3′-cyclic phosphate 10, rather than predominant 7 could be found.
1MRC Laboratory of Molecular Biology, Francis Crick Avenue, Cambridge Biomedical Campus, Cambridge CB2 0QH, UK. 2Institute of Biophysics, Academy
of Sciences of the Czech Republic, Královopolská 135, 61265 Brno, Czech Republic. 3CEITEC – Central European Institute of Technology, Masaryk University,
Campus Bohunice, Kamenice 5, CZ-62500 Brno, Czech Republic. 4Department of Physical and Quantum Chemistry, Faculty of Chemistry, Wrocław
University of Technology, Wybrzeże Wyspiańskiego 27, 50-370 Wrocław, Poland. e-mail: johns@mrc-lmb.cam.ac.uk; rafal.szabla@gmail.com
*
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.2664
Results and discussion
ARTICLES
HO
O
OH
We recently discovered a synthetic route from hydrogen cyanide to
the sugars needed for the assembly of ribose aminooxazoline 1
based on reductive homologation with photochemically generated
hydrated electrons6,7. Hydrosulfide, the conjugate base of hydrogen
sulfide (pKa = 7.2), turned out to be the ideal source of these
electrons7. Hydrogen sulfide is a constituent of volcanic exhala-
tions and it can be produced by dissolution of metal sulfides in
cyanide solution7. Given their availability and their involvement
in the photochemical synthesis of the constituents of ribose ami-
nooxazoline 1, we wondered if these sulfur species and irradiation
might also play a role in the further conversion of 1 into ribonu-
cleotides. For the sake of clarity, we present our findings before we
discuss the mechanisms and implication of the reactions
we uncovered.
N
+
O
NH2
Ribose +
11
12
cyanamide
HO
HO
O
O
Pi
N
N
NH2
NH2
O
O
HO
HO
7
1
N
2
HO
HO
Formation, photoanomerization and hydrolysis of thionucleosides.
We first sought an inherently favoured reaction of ribose
aminooxazoline 1, or one of its derivatives, with hydrosulfide and,
in this way, discovered that the anhydronucleoside 3 undergoes
smooth thiolysis8 in aqueous formamide to give α-2-thioribocytidine
13 in 84% yield along with the hydrolysis product α-ribocytidine
4 in 16% yield (Figs 2 and 3a–c, and Supplementary Figs 30 and
31). We then investigated the ultraviolet photoanomerization of
α-2-thioribocytidine 13 and found that it took place extremely
efficiently to give β-2-thioribocytidine 14 in 76% yield (Figs 2 and
3c–e, and Supplementary Figs 32–34). Small amounts of the
starting material (11%) remained after irradiation at 254 nm for
2.5 days (in aqueous solution that contained hydrogen sulfide to
minimize photooxidation by adventitious oxygen) and there was
also a low level nucleobase loss (thiocytosine 15 detected in 4%
yield) and hydrolysis to α-ribocytidine 4 (9%).
O
O
N
N
+
+
NH2
NH2
N
N
O
O
HO
HO
8
3
Pi
H2O
N
Heat
N
NH2
NH2
HO
HO
O
N
O
N
O
O
O
O
–
HO
OH
P
O
O
4
9
hν
We then investigated the hydrolysis of the 2-thioribocytidine
nucleosides 13 and 14 and further chemistry of the hydrolysis
products. Mild acid hydrolysis of α-2-thioribocytidine 13 in dilute
formic acid (40 mM, pH = 3, 60 °C) gave α-2-thioribouridine 16
in 93% yield after 14 days (Supplementary Figs 36 and 37),
whereas the hydrolysis of β-2-thioribocytidine 14 under the same
conditions was slower and less selective, and gave β-2-thioribouridine
17 in 38% yield after 54 days, along with products of nucleobase loss
(2-thiocytosine 15, 34%; 2-thiouracil 18, 18% and ribose)
(Supplementary Figs 39 and 40). In contrast to what we found in
the 2-thiocytidine series, photoanomerization of α-2-thioribouridine
16 to β-2-thioribouridine 17 was not observed; rather, there was
extensive nucleobase destruction that resulted in the liberation of
ribose (Supplementary Fig. 35). Hydrolysis of β-2-thioribocytidine
14 at pH = 7 in the presence of phosphate buffer (100 mM, 60 °C)
was slow, but after 84 days cleanly furnished the canonical nucleo-
side β-ribocytidine 5 (41%) (Figs 2 and 3e–g, and Supplementary
Figs 41 and 42) along with small amounts of β-2-thioribouridine
17 (2%) and β-ribouridine 19 (3%) produced by the very slow
hydrolysis of 5 itself9. Hydrolysis at pH = 5 in the presence of phos-
phate buffer (100 mM, 60 °C) for a similar length of time gave more
β-2-thioribouridine 17 (8%) and less β-ribocytidine 5 (33%).
Although we only monitored the hydrolysis of β-2-thioribocytidine
14 for about one half-life at pH = 7 and pH = 5, more β-ribocytidine
5, β-ribouridine 19 and β-2-thioribouridine 17 would be produced
after longer periods.
O
HO
NH2
O
P
HO
N
NH
O
N
N
O
O
O
–
O
HO
OH
O
O
10
5
HO
O
NH
O
O
6 (»5)
HO
Figure 1 | Previous prebiotic syntheses of pyrimidine β-ribonucleotides via
2,2′-anhydronucleoside intermediates. The route initially described by
Sanchez and Orgel is indicated by the blue arrows1. Thus, treatment of
ribose with cyanamide gave a beautifully crystalline aminooxazoline 1, which
underwent reaction with cyanoacetylene 2 to give the anhydronucleoside 3
and thence, by hydrolysis, α-ribocytidine 4. However, anomerization to
β-ribocytidine 5 by irradiation of 4 with ultraviolet light was very inefficient
and this, along with doubts about the prebiotic provenance of ribose,
lessened the attractiveness of the route. We determined that one of the
reasons for the low photoanomerization yield was competing formation of
the oxazolidinone 6; we then found an alternative route to the pyrimidine
ribonucleotides via the arabinose aminooxazoline 7 and anhydronucleoside
8. The key steps were a regioselective phosphorylation and 2′-inversion
reaction of 8 to give β-ribocytidine-2′,3′-cyclic phosphate 9 and the partial
photochemical conversion of 9 into the corresponding uridine analogue 10.
We also found that pentose aminooxazolines, predominantly 1 and 7, can be
produced in an indirect way from glyceraldehyde 11 and 2-aminooxazole 12.
Although our route marked the first high-yielding synthesis of the pyrimidine
ribonucleotides under prebiotically plausible conditions, it did not exploit the
crystallinity of 1. Pi, inorganic phosphate.
Phosphorylation. Prebiotically plausible phosphorylation of
nucleosides has been demonstrated in hot formamide solution
and in urea melts10,11. Although the 5′-monophosphates or the
2′,3′-cyclic phosphates of the β-ribopyrimidine nucleosides 5 and
19 can be formed in this way, depending on the conditions11, we
also investigated the direct phosphorylation of β-2-thioribocytidine
14. Somewhat surprisingly, the 2′,3′-cyclic phosphate of the
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.2664
ARTICLES
HO
O
N
+
NH2
N
O
HO
3
HS– in
HCONH2/H2O
O
NH2
NH2
HO
HO
HO
O
O
O
N
NH
N
N
N
N
hν
X
Fast
»
17
S
S
O
HO
OH
16
HO
OH
13
HO
OH
4
+
H3O
hν
O
NH2
NH2
HO
HO
HO
NH2
+
O
O
O
O
N
NH
N
N
N
N
H2O
HN
N
HN
NH
17
+
+
S
S
O
Pi
Slow
HO
OH
HO
OH
HO
OH
S
S
Slow
15
18
17
14
5
H2O (ref. 9)
Very slow
Δ, Pi in
HCONH2
O
NH2
O
HO
HO
HO
NH2
+ 5
O
P
O
O
N
NH
N
N
N
NH
hν
(ref. 3)
+
N
N
O
10
O
O
HO
OH
19
O
O
O
O
NH2
20
P
–
9
–
O
O
O
O
Figure 2 | Synthesis of pyrimidine β-ribonucleosides and β-ribonucleotides involving photoanomerization of α-2-thioribocytidine 13. The reaction scheme
incorporates the new synthesis and the known hydrolysis of β-ribocytidine 5 to β-ribouridine 19 (ref. 9). Thiolysis of the ribose anhydronucleoside 3 gives
α-2-thioribocytidine 13 in high yield along with a small amount of α-ribocytidine 4. In marked contrast to 4, 13 undergoes a remarkably efficient
photoanomerization to give β-2-thioribocytidine 14 in excellent yield. The canonical pyrimidine ribonucleosides 5 and 19 can be produced from 14 by
hydrolysis. Phosphorylation of 14 is accompanied by conversion of the 2-thiocytosine nucleobase into cytosine, which provides a direct route to the canonical
pyrimidine ribonucleoside-2′,3′-cyclic phosphates 9 and thence 10 and an alternative route to the ribonucleoside 5. 2-Thioribouridines 16 and 17 can also be
produced by hydrolysis of 13 and 14, respectively, under acidic conditions.
2-thiocytidine nucleoside was not detected when 14 was allowed to a solvent for subsequent phosphorylation chemistry, we envisioned
react with inorganic phosphate in hot formamide. Instead, using formamide—the hydration product of hydrogen cyanide.
β-ribocytidine 5 (48%) and its 2′,3′-cyclic-phosphate derivative 9 We therefore also evaluated aqueous formamide for the thiolysis
(24%) were detected along with 2,4-diaminopyrimidine 20 (21%) step and found it to allow smooth conversion of the anhydro-
(Supplementary Figs 43–46). As it has previously been shown that nucleoside 3 into α-2-thioribocytidine 13 with minimal competing
ultraviolet irradiation partly converts 9 into its uridine analogue hydrolysis. In the phosphorylation step, the formamide solvent
10 (ref. 3), the direct phosphorylation of β-2-thioribocytidine 14 has to be essentially anhydrous and this can easily be achieved
can therefore lead to both canonical β-ribopyrimidine nucleoside simply by heating.
2′,3′-cyclic phosphates 9 and 10.
The pH dependence of the hydrolysis of β-2-thioribocytidine 14
presumably results from the relative electrophilicities of C-2 and
Thiolysis mechanism. Mechanistically, thiolysis of the anhydro- C-4, dependent on the protonation state of 14 and relative leaving-
nucleoside 3 presumably involves repeated reversible nucleophilic group abilities. At low pH, ammonia can be lost from a tetrahedral
attack at C-2 by hydrosulfide followed by the occasional extrusion intermediate formed by nitrogen protonation and the attack of
of O-2′ from the tetrahedral intermediate thus formed to give water at C-4 of 14 (or the reversibly formed 5,6-double-bond
α-2-thioribocytidine 13. Although water or hydroxide are less hydration adduct thereof), whereas at neutral pH, hydrosulfide
nucleophilic than hydrosulfide, extrusion of O-2′ from the tetrahedral can be lost from a more rarely accessed tetrahedral intermediate
intermediate formed by the addition of these oxygenous formed by the attack of water or hydroxide at C-2. The more-
nucleophiles to the anhydronucleoside 3 must be a common event rapid dilute acid hydrolysis of α-2-thioribocytidine 13 is caused
and this is probably one of the reasons why thiolysis of 3 is by neighbouring-group participation by the 2′-OH, and a tricyclic
1
accompanied by hydrolysis to α-ribocytidine 4. The hydrolysis intermediate could be detected by H NMR spectroscopy during
might also be subject to general base catalysis by hydrosulfide. As the hydrolysis (Supplementary Figs 37 and 38).
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
© 2016 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
3

NATURE CHEMISTRY DOI: 10.1038/NCHEM.2664
ARTICLES
a
b
3 H-C5 3 H-C1ʹ
3 H-C6
13 H-C6
4 H-C6
13 H-C1ʹ
13 H-C5
4 H-C1ʹ 4 H-C5
c
13 H-C6
13 H-C1ʹ
13 H-C5
d
15 H-C6
4 H-C1ʹ
14 H-C6
14 H-C6
14 H-C6
13,14 H-C5
4 H-C6
14 H-C1ʹ
13 H-C1ʹ
13 H-C6
15
H-C5
4 H-C5
e
f
14 H-C1ʹ
14 H-C5
5 H-C5
5 H-C1ʹ
17 H-C5
5 H-C6
19 H-C6
14 H-C1ʹ
17 H-C1ʹ
14 H-C5
g
5 H-C5
5 H-C1ʹ
5 H-C6
8.0
7.5
7.0
δ(1H) (ppm)
6.5
6.0
Figure 3 | 1H NMR spectra that demonstrate the inherently favoured nature of the individual reactions of the newly discovered synthetic sequence.
1H NMR spectra were recorded at 400 MHz with samples of the reaction products (after removal of the solvents) and a synthetic standards dissolved in
H2O–D2O. a, Spectrum of the starting material, 2,2′-anhydroribocytidine 3. b, Spectrum of the products of the thiolysis of 3 by HS– in wet formamide
(singlet at δ ≃ 8.0 ppm results from residual formamide). Note the quantitative transformation of 3 and the predominance of α-2-thioribocytidine 13 over
α-ribocytidine 4 in the products. c, Spectrum of an authentic sample of α-2-thioribocytidine 13. d, Spectrum of the products of irradiation of 13 in aqueous
solution (singlet at δ ≃ 7.9 ppm results from residual dimethylformamide from the conventional synthesis of 13). The conversion of 13 into the anomer
β-2-thioribocytidine 14 is remarkably clean. e, Spectrum of an authentic sample of β-2-thioribocytidine 14. f, Spectrum of the products of hydrolysis of 14 in
pH = 7 phosphate buffer after about one half-life. Despite the conversion of 14 being only partial, note the clean production of β-ribocytidine 5 and the
appearance of its hydrolysis product, β-ribouridine 19, as well as β-2-thioribouridine 17, the product of the alternative hydrolysis of 14. g, Spectrum of an
authentic sample of β-ribocytidine 5.
Phosphorylation mechanism. A mechanism for the photochemical the 2′,3′-cyclic phosphate ring of 22 constrains the sugar’s
conversion of α-2-thioribocytidine 13 into the β-anomer 14 was not conformational preferences12 such that intramolecular addition
immediately apparent, and further experimental and computational of the 5′-OH group to C-2 of the nucleobase with the loss of
investigations of this key reaction were carried out as described hydrogen sulfide also becomes possible. The 2,5′-anhydronucleotide
below. However, conventional organic reasoning did explain the 23 that results is susceptible to phosphorolytic cleavage to give
behaviour of 14 under phosphorylation conditions (Fig. 4a). β-ribocytidine-2′,3′-cyclic phosphate 9, and to aminolysis (by
Reversible phosphorylation in hot formamide is thought to ammonia produced by α-elimination from formamide) to give the
involve formidoylphosphate 21 formed by dissociative 2,4-diaminopyrimidine cyclic nucleotide 24. Glycosidic bond
nucleophilic displacement of water from a rare tautomer of the cleavage of 24 is then presumed to be facile because of the
phosphate monoanion (Fig. 4a, top). The ability of formamide to basicity of the nucleobase13 and the enhanced leaving group
displace an oxygenous leaving group from a phosphorus atom in ability of the protonated nucleobase. This cleavage would generate
such a molecular context means that if phosphate adds to an the free 2,4-diaminopyrimidine 20 that we observed, and an
electrophilic carbon and then undergoes attack by formamide, an unstable sugar moiety prone to decompose to a slew of products
oxygen atom is delivered to the electrophilic carbon. Thus, on that we did not observe, presumably because of the low yield of
heating with inorganic phosphate in formamide, β-2-thioribocytidine each individual product.
14 can either undergo phosphorolytic loss of hydrogen sulfide to
give β-ribocytidine 5, which can then undergo reversible Photoanomerization mechanism. We then turned our attention to
phosphorylation to give β-ribocytidine-2′,3′-cyclic phosphate 9 or the mechanism of photoanomerization of α-2-thioribocytidine 13,
be directly phosphorylated to the thermodynamically favoured and further sought an explanation as to why α-2-thioribouridine
β-2-thioribocytidine-2′,3′-cyclic phosphate 22. Intermolecular 16 did not undergo a similar epimerization on irradiation.
phosphorolytic loss of hydrogen sulfide from 22 to give Suspecting an interrupted Norrish type II reaction similar to that
β-ribocytidine-2′,3′-cyclic phosphate 9 is then possible. However, we observed for the photoanomerization of 2′-deoxycytidine14, we
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.2664
ARTICLES
a
–
–
+
H2N
O
P
O
H2N
+
–
H2O
+
O
P
OH2
–
O
O
O
O
O
21
NH
N
N
NH2
NH
O
HO
HO
O
P
O
O
P
N
N
N
N
14
20
NH3
–H2S
S
NH2
O
O
O
O
O
O
–H2S
Pi
P
22
23
24
O
O–
O
O–
O
O–
–H2S
Pi
5
9
13 H-C6
13 H-C1ʹ
13 H-C5
b
c
13,14 H-C5
14 H-C5
14 H-C6 13 H-C6
14 H-C6
14 H-C1'
d
e
13,14 H-C5
14 H-C6
13 H-C6
15 H-C5
15 H-C6
8.0
7.5
7.0
6.5
δ(1H) (ppm)
Figure 4 | Reaction mechanisms. a, Mechanisms of phosphorylation and the degradation of β-2-thioribocytidine-2′,3′-cyclic phosphate 22. The generation of
the active phosphorylating agent, formidoyl phosphate 21, in the phosphorylation of β-2-thioribocytidine 14 (top)—the phosphorylation of 14 itself involves a
hydroxyl group thereof attacking the imidoyl phosphate in a manner analogous to the reverse of the electron movements indicated by the arrows. The
mechanisms of the reactions that take place during the phosphorylation of β-2-thioribocytidine 14 are shown (bottom). b–e, 1H NMR spectra (400 MHz,
H2O–D2O) that demonstrate the incorporation of deuterium at C-1′ of α/β-2-thioribocytidine 13/14 during photoanomerization. Expansion of the spectrum of
α-2-thioribocytidine 13 that shows signals for H-6, H-1′ and H-5 (b). Spectrum of the products of irradiation of α-2-thioribocytidine 13 in D2O for 12 h that
shows signals for H-6 and H-5 of both 13 and β-2-thioribocytidine 14, but lacking signals for H-1′ of either anomer (c). Spectrum of 14 prior to irradiation (d).
Spectrum of the products of irradiation of 14 in D2O for 36 h (e). The deuteration data shown indicate a mechanism of photoanomerization that involves the
exchange of H-1′ with the solvent.
irradiated α-2-thioribocytidine 13 in D2O because anomerization by preparative-scale ultraviolet reaction and characterized by NMR
this mechanism would (presumably) be associated with deuterium spectroscopy and liquid chromatography–mass spectroscopy
incorporation at C-1′. Gratifyingly, after irradiation for 12 hours (Supplementary Information).
we observed the complete deuterium incorporation at C-1′ of
both the product, β-2-thioribocytidine 14, and the residual Photoanomerization theory. Deuterium incorporation at the C-1′
starting material 13, which demonstrated that exchange is faster carbon atom is consistent with 2-thioribocytidine 13/14 undergoing
than anomerization (Fig. 4b,c). Furthermore, by 36 hours the photoanomerization via an excited-state C1′–H atom abstraction by
reaction appeared to reach equilibrium with the β- and α-isomers the C=S group (Fig. 5). The intermediate α-/β-25 formed after the
in a ratio of 85:15, as calculated using relative integrations of hydrogen abstraction possesses an SH group that can readily
signals caused by H-6 of both anomers in the1H NMR spectrum exchange the proton with a deuteron from the surrounding D2O
of the crude reaction products (Supplementary Fig. 47). To molecules. The SD group thus formed can subsequently deliver
confirm the equilibrium, the β-isomer was irradiated in D2O, this deuteron to the C-1′ position before or after anomerization
whereupon conversion into the α-isomer and C-1′ deuteration through a C1′ radical inversion. In the case of 2′-deoxycytidine,
were observed and, after 36 hours, equilibrium had again been the excited-state hydrogen-atom abstraction reaction occurs on
reached (β- and α-isomers in a ratio of 88:12) (Fig. 4d,e and the hypersurface of the lowest-lying nπ* singlet state (S1). The
Supplementary Fig. 48). In further support of the proposed minima of the S1 states of both 2-thioribocytidine 13/14 and
mechanism, the deuterated anomers were purified from
a
2-thioribouridine 16/17 are of nπ* character and, according to
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.2664
ARTICLES
a
NH2
NH2
HO
HO
O
O
N
N
H
N
H
N
S
S
HO
OH
HO
hν
OH
α-/β-25
Exchangeable
13/14
b
4
3
2
1
0
S1
S1/T2 ISC
T1/S0 ISC
T2
S1 min
T1
kISC ≈ 2.9×1010 s–1
T1 min
3
T1
2
1
0
S0
S0 min
S0
2.4 2.1 1.8
LIIC coordinate
2.4 2.2 2.0 1.8 1.6
S–H distance (Å)
Figure 5 | Mechanism of the key photoanomerization reaction. a, Conventional organic chemical depiction of the mechanism of the photoanomerization of
α/β-2-thioribocytidine 13/14 via the diradical 25. b, The same mechanism but analysed at a deeper level to explain the niceties of the reaction. The potential
energy (PE) profiles present the photoanomerization reaction of α-2-thioribocytidine. The PE cuts on the left-hand side were obtained by interpolations
between the Franck–Condon region, S1 minimum (min), S1/T2 minimum-energy crossing point and the T1 minimum. The PE profile on the right was obtained
from a relaxed scan along the S–H distance. The inset on the right corresponds to the relaxed scan along the S–H distance of α-2-thioribouridine, for which
the T1(ππ*) and S0 states do not cross. The energies were calculated using the ADC(2)/cc-pVTZ method and the computations of the SOC matrix elements
were performed at the CASPT2/CASSCF level (employing the cc-pVTZ-DK basis set). Details of the S1/T2 ISC rate-constant computations are given in the
Supplementary Information. LIIC, linearly interpolated internal coordinate.
our calculations, the S1/ground state (S0) crossing related to this between these two states. However, hydrogen abstraction is not
3
process is energetically accessible in both molecules. However, the possible on the surface of the ππ* state and, consequently, the S0
presence of sulfur significantly enhances spin–orbit couplings and T1 surfaces correlate adiabatically and do not cross (Fig. 5b,
(SOCs) in organic molecules and, in consequence, may enable inset box). The T1/S0 state crossing can be accessed along the
very efficient non-radiative transitions to the triplet manifold15–17
.
hydrogen-abstraction coordinate only after the change of the T1
3 3
In particular, it was shown that the thiothymidines 6-thioguanine character from ππ* to nπ*, which is related to an energy barrier
and 2-thiouracil undergo ultrafast intersystem crossing (ISC) that of 0.84 eV in the gas phase. This energy barrier is presumably
results in near-unity triplet yields17,18. Our estimates of the ISC even higher in bulk water because of the destabilization of nπ*
rates suggest that the photoanomerization of 2-thioribocytidine states in an aqueous environment (Supplementary Information)21.
13/14 may occur predominantly on the triplet hypersurface.
Furthermore, the C=S⋯H–C1′ distance is significantly larger in
The initially populated excited state of 2-thioribocytidine 13/14 may the T1 minimum of 2-thioribouridine 16/17 (2.65 Å) than in the
undergo barrierless vibrational relaxation towards the minimum on the respective minimum of 2-thioribocytidine 13/14 (2.43 Å), which
corresponding S1 potential energy surface (Fig. 5b). As the minimum- should significantly affect the accessibility of this reaction
energy structure of the T1 state of 2-thioribocytidine is also of nπ* pathway. Therefore, we conclude that photoanomerization via an
character, the corresponding 1nπ*(S1) → 3nπ*(T1) ISC operates interrupted Norrish type II reaction in the triplet manifold is
without molecular orbital change and should be rather slow according clearly hindered in the case of 2-thiouridine. Unable to lose
to the El-Sayed rules and the results of our calculations. Therefore, energy by T1/S0 state crossing along the hydrogen-abstraction coor-
we suggest (based on reasoning given in the Supplementary dinate and subsequent vibrational relaxation, the 3ππ* state under-
Information) that the primary ISC channel is most probably related goes other, as yet undefined, photochemistry that results in
3
to the 1nπ*(S1) → ππ*(T2) transition. The T1 state can then be accessed nucleobase destruction and liberation of ribose.
almost immediately after the S1 → T2 ISC because at the S1/T2
minimum-energy crossing point the T2 and T1 states are almost degen- Conclusions
erate (yielding a S1/T2/T1 three-state crossing). The hydrogen-atom The route to the canonical pyrimidine β-ribonucleosides 5 and 19
abstraction can then take place in the T1 state and when the S–H that we have uncovered in this work is short and high yielding.
distance approaches 1.55 Å, the T1 and S0 electronic states cross.
The 5′-phosphates of 5 and 19 can be produced easily under plaus-
Similarly, as found for 2-thiouracil17,19,20, the T1 state of ible conditions, as can their 2′,3′-cyclic phosphates 9 and 10
2-thioribouridine 16/17 has a ππ* character and it can be more (ref. 11). Alternatively, the latter can be produced by irradiation of
efficiently accessed from the S1 state because of the large SOCs the phosphorylation products of β-2-thioribocytidine 14. The
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.2664
ARTICLES
11. Lohrmann, R. &. Orgel, L. E. Urea–inorganic phosphate mixtures as prebiotic
phosphorylating agents. Science 171, 490–494 (1971).
12. Saenger, W. Principles of Nucleic Acid Structure (Springer, 1984).
13. Brown, D. J. & Jacobsen, N. W. 612. Pyrimidine reactions. Part IV. The
methylation of 2,4- and 4,5-diaminopyrimidine and related compounds.
J. Chem. Soc. 3172–3179 (1962).
14. Szabla, R. et al. Excited-state hydrogen atom abstraction initiates the
photochemistry of β-2′-deoxycytidine. Chem. Sci. 6, 2035–2043 (2015).
15. Mai, S., Marquetand, P. & González, L. Intersystem crossing pathways in the
noncanonical nucleobase 2-thiouracil: a time-dependent picture. J. Phys. Chem.
Lett. 7, 1978–1983 (2016).
β-ribonucleosides and nucleotides all derive from ribose aminooxa-
zoline 1, crystallization of which provides an opportunity for
enantioenrichment because of its conglomerate behaviour.
Enantiomerically pure pyrimidine β-ribonucleotides can thus be
considered prebiotically plausible if a means of producing glyceral-
dehyde 11 with sufficient enantiomeric excess for the subsequent
crystallization of 1 to give a single enantiomer is demonstrated.
Having enantiomerically pure β-ribonucleotides is crucial to producing
RNA by oligomerization because racemic (or even scalemic) mixtures
can either be expected to produce vast numbers of diastereo- 16. Martínez-Fernández, L., Corral, I., Granucci, G. & Persico, M. Competing
ultrafast intersystem crossing and internal conversion: a time resolved picture for
the deactivation of 6-thioguanine. Chem. Sci. 5, 1336–1347 (2014).
17. Pollum, M. & Crespo-Hernández, C. E. Communication: the dark singlet state as
isomerically isomeric oligonucleotides or result in the inhibition of oli-
gomerization through enantiomeric cross-inhibition22. As different
phosphorylated β-ribonucleosides can be produced, oligomerization
could result in 5′- or 2′- and/or 3′-phosphorylated RNA.
a doorway state in the ultrafast and efficient intersystem crossing dynamics in
2-thiothymine and 2 thiouracil. J. Chem. Phys. 140, 071101 (2014).
18. Taras-Goślińska, K., Burdziński, G. & Wenska, G. Relaxation of the T1 excited
state of 2-thiothymine, its riboside and deoxyriboside-enhanced nonradiative
decay rate induced by sugar substituent. J. Photochem. Photobiol. Chem. 275,
89–95 (2014).
19. Gobbo, J. P. & Borin, A. C. 2-Thiouracil deactivation pathways and triplet states
population. Comput. Theor. Chem. 1040, 195–201 (2014).
20. Mai, S., Marquetand, P. & González, L. A static picture of the relaxation and
intersystem crossing mechanisms of photoexcited 2-thiouracil. J. Phys. Chem. A
119, 9524–9533 (2015).
21. Besley, N. A. & Hirst, J. D. Ab initio study of the electronic spectrum of
formamide with explicit solvent. J. Am. Chem. Soc. 121, 8559–8566 (1999).
22. Joyce, G. F., Schwartz, A. W., Miller, S. L. & Orgel, L. E. The case for an ancestral
genetic system involving simple analogues of the nucleotides. Proc. Natl Acad.
Sci. USA 84, 4398–4402 (1987).
23. Grosjean, H., de Crécy-Lagard, V. & Marck, C. Deciphering synonymous codons
in the three domains of life: co-evolution with specific tRNA modification
enzymes. FEBS Lett. 584, 252–264 (2010).
24. Heuberger, B. D., Pal, A., Del Frate, F., Topkar, V. V. & Szostak, J. W. Replacing
uridine with 2-thiouridine enhances the rate and fidelity of nonenzymatic RNA
primer extension. J. Am. Chem. Soc. 137, 2769–2775 (2015).
Finally, a side product of the chemistry documented in this
paper, β-2-thioribouridine 17, and its derivatives are found in the
anticodon stem loop of certain transfer RNAs where they are
involved in split-box decoding23. Furthermore, replacing β-ribouridine
19 with β-2-thioribouridine 17 leads to both faster and more-
accurate non-enzymatic template copying chemistry, which suggests
that 17 might have played a role in chemical RNA replication at an
early stage in the origin of life24. Taken with the foregoing, the
formation of β-2-thioribocytidine 14 reported herein suggests that
its role in RNA replication chemistry should also be investigated.
Note added in proof: After acceptance of this manuscript, a paper from
the groups of González, L. and Crespo-Hernández, C. E. describing a
similar origin of efficient triplet state population in 2-thiocytosine was
published online (Mai et al. Nat. Commun. 7, 13077; 2016).
Received 10 June 2016; accepted 30 September 2016;
published online 21 November 2016
Acknowledgements
References
This work was supported by the Medical Research Council (no. MC_UP_A024_1009), a
grant from the Simons Foundation (no. 290362 to J.D.S.), grant 14-12010S from the Grant
Agency of the Czech Republic and by the project CEITEC 2020 (LQ1601) with financial
support from the Ministry of Education, Youth and Sports of the Czech Republic under the
National Sustainability Programme II. Support from a statutory activity subsidy from the
Polish Ministry of Science and Higher Education for the Faculty of Chemistry of Wrocław
University of Technology is gratefully acknowledged. Theoretical calculations were partly
performed at the Wrocław Center for Networking and Supercomputing and
Interdisciplinary Centre for Mathematical and Computational Modelling in Warsaw.
1. Sanchez, R. A. & Orgel, L. E. Studies in prebiotic synthesis. V. Synthesis and
photoanomerization of pyrimidine nucleosides. J. Mol. Biol. 47, 531–543 (1970).
2. Powner, M. W. et al. On the prebiotic synthesis of ribonucleotides:
photoanomerisation of cytosine nucleosides and nucleotides revisited.
ChemBioChem 8, 1170–1179 (2007).
3. Powner, M. W., Gerland, B. & Sutherland, J. D. Synthesis of activated
pyrimidine ribonucleotides in prebiotically plausible conditions. Nature
459, 239–242 (2009).
4. 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).
5. 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).
6. Ritson, D. & Sutherland, J. D. Prebiotic synthesis of simple sugars by photoredox
systems chemistry. Nat. Chem. 4, 895–899 (2012).
7. Ritson, D. J. & Sutherland, J. D. Synthesis of aldehydic ribonucleotide and
amino acid precursors by photoredox chemistry. Angew. Chem. Int. Ed. 52,
5845–5847 (2013).
8. Kawaguchi, T. et al. Enzymatic reactivity and anti-tumor activity of
1-(β-D-arabinofuranosyl)-2-thiocytosine derivatives. Chem. Pharm. Bull. 48,
454–457 (2000).
9. Frick, L., Mac Neela, J. P. & Wolfenden, R. Transition state stabilization by
deaminases: rates of nonenzymatic hydrolysis of adenosine and cytidine.
Bioorg. Chem. 15, 100–108 (1987).
Author contributions
J.D.S. supervised the experimental research, and J.X., M.T. and C.J.M. performed the
experiments. J.E.S., J.S. and R.W.G. oversaw the theoretical work, which was carried out by
R.S. All the authors contributed intellectually as the project unfolded. J.D.S. wrote most of
the paper and J.X., M.T., C.J.M. and R.S. further contributed and assembled the
Supplementary Information.
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
J.D.S. and R.S.
10. Schoffstall, A. M. Prebiotic phosphorylation of nucleosides in formamide.
Orig. Life 7, 399–412 (1976).
Competing financial interests
The authors declare no competing financial interests.
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
7
© 2016 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.