C
D. J. Ritson et al.
Cluster
Synlett
that isocyanic acid requires a general acid to deliver a pro-
ton simultaneously as a nucleophile adds to it, nucleophilic
addition to a nitrile should also require a general acid.
Given these promising results we wondered if thiophos-
phate could add to a C≡N bond which was not part of a cy-
anohydrin yet may be of prebiotic significance (Table 1, en-
tries 5–8), and in the first instance we examined hydrogen
cyanide 1 using H13CN. After eight hours the 13C NMR spec-
trum clearly showed the thiocarbonyl signal of thioforma-
mide 21 at δ = 195 ppm. Using an increased recycle delay
and stopping any NOE build up we were able to obtain a
more accurate 13C NMR quantification, and integration of
H13CN and H13C(S)NH2 signals showed 47% conversion. Af-
ter 24 hours the yield of 21 only increased slightly as un-
avoidable side reactions began to compete, for example,
thioformamide and/or cyanide hydrolysis to formamide,
formation of thiocyanate, etc.7 Although thiophosphate 9
adds to HCN 1 the reactivity of the nitrile in 1 is greatly en-
hanced once it has added to a carbonyl, hence thiolysis of 1
does not interfere with cyanohydrin thiolysis – essential for
the described cyanosulfidic network.1 AICN 22 has been
documented in the literature as a pertinent prebiotic mole-
cule for some decades although tuning its reactivity has
proved problematic.8 We had previously attempted the thi-
olysis of 22 using NaSH with limited success (33%, 24 h) but
submitting 22 to reaction with thiophosphate 9 provided
the aromatic thioamide 23 in 85% yield after 24 hours (Ta-
ble 1, entry 6). Also, the putative importance of co-factors
in the chemical origin of life has been highlighted many
times previously,9 thus understanding the chemistry of the
co-factors and co-factor moieties would appear to be a
worthwhile endeavour. To this end 3-pyridinecarbonitrile
24 was reacted with thiophosphate 9 as before and after
eight hours the thioamide 25 could be seen crystallising out
of solution, the yield of 25 after 24 hours was 97% (Table 1,
entry 7). Finally, we attempted the thiolysis of glycine ni-
trile 26, which, under our standard conditions, did not give
a clean conversion and did not proceed to completion.
However, when we ran the same reaction with a lower ini-
tial pH of 5.0, the α-amino thioamide 27 was observed in
93% yield after only 0.5 hours and 97% yield after one hour
(Table 1, entry 8). This remarkable change in reactivity was
likely due to three factors all resulting from protonation at
the lower pH: firstly, when 26 is protonated a much larger
inductive effect increases the electrophilicity of the nitrile;
secondly, the electrostatic attraction between protonated
26 and 9 means the reaction is no longer diffusion con-
trolled; thirdly, the percentage of free amine, in aminoni-
trile or product, is reduced which diminishes their reaction
with the product thiocarbonyl (e.g. when glycine thioamide
(100mM) and glycine nitrile (100mM) were incubated with
NaH2PO4 (75 mM) at pH 7 new singlets could be seen form-
presumed to be thio and/or thio-imino variants of glycine
anydride).
Satisfied with the results of nitrile thiolysis we next
turned our attention to another aspect of thiophosphate
chemistry. In the protometabolic network we recently de-
scribed,1 copper(I)-promoted coupling of HCN 1 and acety-
lene gave acrylonitrile 28 which was then used for the pre-
biotic synthesis of arginine and proline, but if thiophos-
phate 9 was indeed prebiotically available, 28 could have
served an additional purpose. Nagyvary et al. used 28 to al-
kylate 9 allowing subsequent phosphorylation of 2′-deoxy-
thymidine giving a mixture of the 3′- and 5′-nucleotides.3b
Although the protocol used was questionable as a prebiotic
synthesis, we wondered if a related procedure could be
found which would still allow facile activation of 9 render-
ing it a prebiotically plausible phosphorylating agent.
Therefore we stirred adenosine (80 mM), dibasic 9 (200
mM) and 28 (300 mM) in formamide at 70 °C then inspected
1H NMR and 31P NMR spectra for evidence of phosphoryl-
ation. After 4 hours we found that ca. 36% of the adenosine
had been phosphorylated giving a mixture of 5′-, 3′- and 2′-
monophosphates in 16%, 8%, and 8% yield, respectively,
(confirmed by spiking with commercial standards) and
what were presumably the 5′,3′- and 5′,2′-bisphosphates in
a combined yield of 4% (yields determined by NMR spec-
troscopy). There was also 11% pyrophosphate based on to-
tal phosphorous content. Schoffstall reported the phos-
phorylation of nucleosides using orthophosphate 10 in for-
mamide some decades ago, which has now become a
standard prebiotic phosphorylation procedure.10 To ensure
our results were not simply due to in situ formation of 10
which then followed Schoffstall’s chemistry, we repeated
the reaction under otherwise identical conditions using
monobasic or dibasic 10 in place of 9. Using Na2HPO4 after 4
hours there was no phosphorylation, and in the case of the
more acidic NaH2PO4 we observed ca. 1% total phosphoryla-
tion of adenosine and ca. 2% pyrophosphate. Schoffstall’s
phosphorylation presumably proceeds via a mechanism
wherein phosphate 10 accesses the extremely rare (so ac-
counting for the sluggish rate of reaction) tautomer 29
which can then lose water to phosphorylate a nucleophile
via a metaphosphate-like transition state 30 (Scheme 3).
The nucleophile may be the nucleoside alcohol, giving a nu-
cleotide, or formamide, which would give a new phosphor-
ylating agent. During the thiolysis studies we had become
aware, as had Nagyvary,3c of how relatively easily 9 could be
hydrolysed.11 This must be due to the fact that as soon as
sulfur in 9 is protonated it can be lost as HS–, and although
the tautomer 31 is disfavoured it is far more accessible than
its oxygenous counterpart 29 (Scheme 3). Hence, we re-
peated the reaction of adenosine with 9 in formamide but
in the absence of acrylonitrile 28 and after 4 hours we ob-
served ca. 20% phosphorylation of adenosine in a similar ra-
tio of products to the analogous reaction with 28 present.
1
ing in the H NMR spectra and a black solid precipitated –
© Georg Thieme Verlag Stuttgart · New York — Synlett 2016, 27, A–D