Formation of potentially prebiotic amphiphiles by reaction of β-hydroxy-n-alkylamines with cyclotriphosphate

Article abstract of DOI:10.1002/anie.200700394

(Chemical Equation Presented) The long and the short of it: In water, long-chain β-hydroxy-n-alkylamines are converted to their O-monophosphates by reaction with cyclotriphosphate. With short-chain pVhydroxy-n-alkylamines, N-triphosphates are formed instead. The difference results from the formation of surfactant assemblies from the long-chain compounds. This surfactant control of reactivity may be of relevance to the prebiotic formation of amphiphiles. P: phosphate unit.

Full text of DOI:10.1002/anie.200700394

Communications
DOI: 10.1002/anie.200700394
Amino Alcohol Phosphorylation
Formation of Potentially Prebiotic Amphiphiles by Reaction of
b-Hydroxy-n-alkylamines with Cyclotriphosphate**
Lee B. Mullen and John D. Sutherland*
Compartmentalization is thought to be crucial to the devel-
opment of evolvable genetic systems in the origins of life.[1–3]
Awide range of surfactant assemblies can form spontaneously
depending on conditions and the type of amphiphile,but only
bilayer structures such as vesicles are suitable for genetic
compartmentalization. Although double-chain amphiphiles
generally form vesicles more readily than single-chain amphi-
philes,the latter do form vesicles and other bilayer structures
in special cases.[3] A number of studies,for example,have
shown that long-chain carboxylic acid products of the
Fischer–Tropsch reaction form vesicles at a pH approximately
equal to the pKa of the acid in the bilayer.[4–7] However,the
nucleic acid compartmentalization properties of these assem-
blies are not ideal,and they are unstable to the ionic
conditions necessary for RNA folding and catalysis.[8]
Although these instability problems can be partly overcome
by the addition of glycerol esters,[8] we sought to find
alternative prebiotic amphiphiles that might have better
vesicle-forming properties.
ability,and because of a potentially predisposed phosphor-
ylation with cyclotriphosphate.
On the issue of prebiotic availability, b-hydroxy-n-alkyl-
amines are the potential reduction products of cyanohydrins,
which themselves could arise from the reaction of hydrogen
cyanide and alkanals. Medium- and long-chain alkanals are
found in the oxygenate fraction of Fischer–Tropsch reaction
products.[10]
As regards the phosphorylation of b-hydroxy-n-alkyl-
amines,we took our cue from the work of Quimby and
Flautt,[11] and Eschenmoser and co-workers (Scheme 1).[12]
On treatment with ammonia at pH 12,Quimby and Flautt
Because we envisage the formation of double-chain
amphiphiles,by a process of constitutional self-assembly,to
be more difficult than the formation of single-chain amphi-
philes,we first focused on the single-chain species. In
particular,we were attracted to a chemical scenario in
which cationic amphiphiles undergo partial conversion to
anionic amphiphiles since mixtures of oppositely charged
single-chain amphiphiles are known to spontaneously form
“catanionic” vesicles.[3,9] The presence of phosphate in the
lipids of contemporary biochemistry further steered our
search for alternative amphiphiles. We recognized that if a
singly charged cationic amphiphile could be phosphorylated,
then,depending on the chemistry and pH,the product might
be anionic due to the ability of a phosphate group to be
doubly negatively charged. b-Hydroxy-n-alkylammonium
salts seemed to us to be suitable cationic amphiphiles in this
regard,both with respect to their possible prebiotic avail-
Scheme 1. Two-step process for the phosphorylation of glycolaldehyde
hydrate 3 by cyclotriphosphate 1.[11,12] Initial ammonolysis of 1 to 2
allows the reversible tethering of a phosphorylating agent to 3 giving
5. Intramolecular phosphorylation of 5 assistedby divalent metal ions
followedby hydrolytic removal of the tethering attachment from 6 or 7
then gives the monophosphate product 4.
showed that cyclotriphosphate 1[13] is converted into amido-
triphosphate 2.[11] Eschenmoser and co-workers showed that
at near neutral pH in the presence of Mg2+ ions and 2,
glycolaldehyde hydrate 3 undergoes smooth conversion to its
monophosphate 4.[12] The reaction is thought to proceed by
reversible formation of the hemiaminal 5,intramolecular
phosphorylation to give the phosphoramidate 6,and subse-
quent hydrolysis of 6,or the open-chain form 7. A high pH is
necessary for the conversion of 1 to 2 in the first stage of this
two-step process in order that ammonia is predominantly in
its free base form. However,for the second stage a high pH
would be deleterious for two reasons. Firstly the Mg2+ ions—
needed to coordinate to the pyrophosphate moiety of 5
rendering it a good leaving group—would be removed from
solution by formation of the insoluble hydroxide salt.
Secondly,the hydrolysis of 6/7 requires that the nitrogen be
[*] L. B. Mullen, Prof. Dr. J. D. Sutherland
School of Chemistry
The University of Manchester
OxfordRoad, Manchester M139PL (UK)
Fax: (+44)161-275-4939
E-mail: john.sutherland@manchester.ac.uk
[**] This work was carriedout as part of EU COSTaction D27 “Prebiotic
Chemistry andEarly Evolution” andwas fundedby the Engineering
andPhysical Sciences Research Council through the provision of a
studentship to L.B.M. We thank Prof. A. Eschenmoser and Prof. R.
Krishnamurthy (The Scripps Research Institute, La Jolla, USA) for
bringing reference [15] to our attention.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
4166
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 4166 –4168

Angewandte
Chemie
Table 1: Yields [%] of the N-triphosphate 9, O-monophosphate 11, and
protonated,and related phosphoramidates are only very
slowly hydrolyzed at pH > 8.[14]
In the case of long-chain b-hydroxy-n-alkylamines 8,we
hoped to be able to demonstrate similar phosphorylation
chemistry using 1 (Scheme 2),but in a single reaction at one
residual starting material observed in the reactions of b-hydroxy-n-
alkylamines 8 with cyclotriphosphate 1 after 6 days.[a]
RCH(OH)CH2NH2 8
9
11
Residual 8
R=C2H5
R=C4H9
R=C6H13
R=C8H17
19
19
10
0
0
0
81
81
20
70
40 (55[b])
60 (45[b])
[a] A solution (or suspension) of 8 (100 mm) and 1 (250 mm) in D2O was
adjusted to pD 10 and maintained at 268C. NMR spectra were recorded
periodically (samples were filtered when necessary). The pD was
readjusted to 10 daily by addition of NaOD solution. For full
experimental details see the Supporting Information. [b] The yield of
11 (R=C8H17) increasedsignificantly after 4 weeks; (overall) yields of the
shorter-chain products were unchanged.
Scheme 2. Potential phosphorylation of b-hydroxy-n-alkylamines 8 by
cyclotriphosphate 1.
2+
pH,in the absence of Mg
ions,and with the key steps
controlled by chain-length effects. At a pH at which 8 is at
least partially available in free base form,reaction with 1 to
give an N-triphosphate 9 was anticipated.[15] In the absence of
effects due to the long alkyl chain,this species was not
expected to undergo intramolecular attack of the OH group
to give the phosphoramidate 10,because of the poor leaving-
group ability of pyrophosphate when it is unprotonated or not
coordinated to a divalent metal ion. However,if the alkyl
chain was sufficiently long for 9 to be incorporated into some
sort of surfactant assembly,we hoped that the unfavorability
of negative charge repulsion between hydrophobically asso-
ciated molecules of 9 would increase the degree of proto-
nation of the pyrophosphate moiety of 9,and consequently
increase its leaving-group ability. If cyclization to the phos-
phoramidate were then to occur,we again hoped that
incorporation into,or retention in,a surfactant assembly
would increase the degree of N-protonation of 10 such that
hydrolysis to the O-monophosphate 11 would occur.
Figure 1. 31P NMR analysis of the product from the reaction of 8
(R=C4H9) with 1. Selectedregions of the 1H-decoupled spectrum
(bottom) and 1H-coupledspectrum (top) are shown.
To test the foregoing predictions we used conventional
synthesis[16] to prepare a series of b-hydroxy-n-alkylamines 8
(R = C2H5, C4H9, C6H13, C8H17) by reduction of the corre-
sponding azidoalcohols which,in turn,were easily accessible
from the terminal epoxides (see the Supporting Information).
We then studied the reaction of 8 with cyclotriphosphate 1
under a variety of conditions. It was found that reaction with 1
was insignificant at pH < 10,but at pH 10 reaction occurred
over a number of days. For the short-chain compounds (R =
C2H5, C4H9), 8 was converted to the N-triphosphate 9
31P NMR spectrum (Figure 2). A key factor in the assignment
of the O-monophosphate 11 was the observation of a doublet
signal in the 1H-coupled 31P NMR spectrum at d = 4.0 (JH,P
=
8.5 Hz) which collapsed to a singlet in the 1H-decoupled
spectrum (Figure 2). To verify this assignment,we synthesized
an authentic sample of 11 (R = C8H17) and used it to spike the
1H NMR sample of the reaction products of 8 (R = C8H17) and
1 (see Figure S2 in the Supporting Information). The spiking
experiment confirmed the assignment. The second-longest-
chain compound tested 8 (R = C6H13) was converted to a
mixture of 9 and 11 (Table 1,see Figures S3 and S4 in the
Supporting Information),and over a prolonged period of
time,the amount of 11 was seen to increase while the amount
of 9 decreased.
(Table 1). The structure of
9
followed from 1H and
1
31P NMR analysis,[15] in particular the H-coupled 31P NMR
spectrum showed a doublet of apparent triplets (d = 0.0 ppm
J
P,P = 19.8 Hz, JH,P = 8.5 Hz, JH,P = 7.9 Hz for 9 (R = C4H9)) for
1
Pa which collapsed to a doublet (JP,P = 19.8 Hz) in the H-
decoupled spectrum (Figure 1). In the 1H NMR spectrum,
À
À
signals for H C1 and H C2 of 9 were shifted upfield relative
to the corresponding signals for 8 (see Figure S1 in the
Supporting Information). For the longest-chain compound
(R = C8H17), 8 was converted in moderate yield to a single
species which had NMR spectroscopic data consistent with
the O-monophosphate 11 (Table 1). No N-triphosphate 9 was
apparent as evidenced by the lack of a signal at d ꢀ 0 in the
The phosphorylation chemistry uncovered in this inves-
tigation is noteworthy for a number of reasons. Owing to the
long-chain effects,significantly different products form from
long-chain and short-chain b-hydroxy-n-alkylamines 8. The
Angew. Chem. Int. Ed. 2007, 46, 4166 –4168ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
4167

Communications
tant mixtures of 8 and 11 will form bilayer structures—
especially catanionic vesicles—remains to be determined. The
dynamic nature of the surfactant assemblies formed as the
reaction progresses also remains to be investigated. These
studies will likely be complex,but with the covalent chemistry
established,the elucidation of the noncovalent chemistry now
becomes a worthwhile target.
Received: January 29,2007
Published online: April 24,2007
Keywords: amino alcohols · amphiphiles · phosphorylation ·
.
prebiotic chemistry · surfactants
[1] J. W. Szostak,D. P. Bartel,P. L. Luisi, Nature 2001, 409,387.
[2] I. A. Chen,K. Salehi-Ashtiani,J. W. Szostak, J. Am. Chem. Soc.
2005, 127,13213.
[3] P. Walde, Origins Life Evol. Biosphere 2006, 36,109.
[4] J. M. Gebicki,M. Hicks, Nature 1973, 243,232.
[5] P. Walde,R. Wick,M. Fresta,A. Mangone,P. L. Luisi,
Chem. Soc. 1994, 116,11649.
J. Am.
[6] D. W. Deamer, Nature 1985, 317,792.
[7] T. M. McCollom,G. Ritter,B. R. T. Simoneit, Origins Life Evol.
Biosphere 1999, 29,153.
[8] P. A. Monnard,C. L. Apel,A. Kanavarioti,D. W. Deamer,
Astrobiology 2002, 2,139.
[9] E. W. Kaler,A. K. Murthy,B. E. Rodriguez,J. A. N. Zasadzin-
ski, Science 1989, 245,1371.
[10] R. B. Anderson, The Fischer-Tropsch Synthesis,Academic Press,
Orlando, 1984.
Figure 2. 31P NMR analysis of the product from the reaction of 8
(R=C8H17) with 1. Selectedregion of the 1H-decoupled spectrum
(bottom) and 1H-coupledspectrum (top) are shown.
[11] O. T. Quimby,T. J. Flautt, Z. Anorg. Allg. Chem. 1958, 296,220.
[12] R. Krishnamurthy,G. Arrhenius,A. Eschenmoser, Origins Life
Evol. Biosphere 1999, 29,333.
[13] For discussion of the prebiotic availability of phosphate and
polyphosphates including cyclotriphosphate see: A. W.
Schwartz, Philos. Trans. R. Soc. London Ser. B 2006, 361,1743;
Y. Yamagata,H. Watanabe,M. Saitoh,T. Namba, Nature 1991,
352,516.
formation of O-monophosphates 11 from the long-chain
compounds in a single reaction at one pH value and in the
absence of divalent metal ions is remarkable[17] especially
given the results of Quimby and Flautt,[11] and Eschenmoser
and co-workers.[12] Although we do not observe phosphor-
amidates 10,it would seem most reasonable to invoke them as
intermediates in the conversion of 8 to 11 especially given the
results of Eschenmoser and co-workers. It would thus appear
that they are rapidly hydrolyzed at pH 10 even though related
phosphoramidates are hydrolytically inert at pH > 8.[14] The
long-chain effects on the closure of 9 to 10 and the subsequent
hydrolysis of 10 to 11 mean that our goal of partially
converting a cationic amphiphile to an anionic amphiphile is
potentially possible depending on the pKa values of the
various species in surfactant assemblies. Whether the resul-
[14] R. A. Lazarus,P. A. Benkovic,S. J. Benkovic,
J. Chem. Soc.
Perkin Trans. 2 1980,373.
[15] M. Tsuhako,C. Sueyoshi,T. Miyajima,S. Ohashi,H. Nariai,I.
Motooka, Bull. Chem. Soc. Jpn. 1986, 59,3091.
[16] M. E. Dyen,D. Swern, J. Org. Chem. 1968, 33,379.
[17] In the case of the short-chain compounds,we even sought to
convert N-triphosphate 9 to O-monophosphate 11 by lowering
the pD from 10 to 7,but this instead caused reversion of 9 to b-
hydroxy-n-alkylamine 8.[15]
4168
www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 4166 –4168