Comprehensive investigation of the energetics of pyrimidine nucleoside formation in a model prebiotic reaction

Article abstract of DOI:10.1021/ja900807m

The problem of β-nucleoside formation under prebiotic conditions represents one of the most significant challenges to the "RNA world" hypothesis. The possibility exists that alternative bases may have come before the contemporary bases (i.e., A, G, C, and

Full text of DOI:10.1021/ja900807m

Published on Web 10/19/2009
Comprehensive Investigation of the Energetics of Pyrimidine
Nucleoside Formation in a Model Prebiotic Reaction
,†
‡
‡
‡
Yinghong Sheng,* Heather D. Bean, Irena Mamajanov, Nicholas V. Hud, and
,
§
Jerzy Leszczynski*
Department of Chemistry and Mathematics & Whitaker Center, College of Arts & Sciences,
Florida Gulf Coast UniVersity, 10501 FGCU BouleVard South, Fort Myers, Florida 33965,
School of Chemistry and Biochemistry, Georgia Institute of Technology, 901 Atlantic DriVe,
Atlanta, Georgia 30332, and Interdisciplinary Nanotoxicity Center, Department of Chemistry,
Jackson State UniVersity, P.O. Box 17910, 1400 J. R. Lynch Street, Jackson, Mississippi 39217
Received February 2, 2009; E-mail: ysheng@fgcu.edu; jerzy@ccmsi.us
Abstract: The problem of ꢀ-nucleoside formation under prebiotic conditions represents one of the most
significant challenges to the “RNA world” hypothesis. The possibility exists that alternative bases may have
come before the contemporary bases (i.e., A, G, C, and U), including bases that more readily form
nucleosides. We previously reported the first successful synthesis of a pyrimidine nucleoside from a free
base and a nonactivated sugar in a plausible prebiotic reaction. Here we present a detailed computational
2+
study on the reaction at the density functional theory (DFT) level. The catalytic role of a Mg ion on the
reaction mechanism is also investigated. Our calculations demonstrate that a Mg2 ion, serving as a Lewis
acid, can afford the necessary stabilization to the base and leaving water molecule during glycoside bond
formation. The solvent effect is considered by the Onsager solvation model and also by an extended model
with the addition of explicit water molecules within the SCRF solvation model. In addition, predictions
regarding the formation of nucleosides from other pyrimidine bases are also addressed, providing valuable
insights into what chemical features of the bases facilitate glycoside formation in drying-heating reactions.
+
I. Introduction
yielding very little ꢀ-cytidine. The R-cytosine ribonucleoside
can be photoisomerized to the ꢀ-ribonucleoside, but in low yield.
Improvements in ꢀ-cytidine formation by this approach are
possible with phosphorylated ribose precursors or in a phosphate-
The formation of a glycosidic bond between ribose and free
pyrimidine bases under plausible prebiotic conditions has
represented a serious challenge to the “RNA world” hypothesis
and also to a number of its proposed precursors. Over 30 years
ago, Orgel and co-workers demonstrated that adenine and
hypoxanthine form glycosidic bonds with D-ribose when dried
5,6
rich reaction medium, but the prebiotic relevance of phos-
1
phorylated compounds presents another difficulty and it is
7
therefore important to seek alternative solutions to this problem.
2
For the model prebiotic reactions discussed above, none of
these approaches is able to produce all four nucleosides, or even
one purine and one pyrimidine. If the original bases differed
and heated together. However, this simple approach proved to
be less promising with the other bases: neither cytosine nor
uracil give rise to nucleosides under these conditions, nor does
guanine, which may be due to the low solubility of the free
8
from the canonical AUCG, then there is the possibility that
1
some of these bases might have formed glycosides more easily
than the current bases under prebiotic conditions. As part of
our investigation of this hypothesis, we recently reported the
first successful synthesis of a pyrimidine nucleoside from a free
base (2-pyrimidinone) and a nonactivated sugar (D-ribose) in a
guanine base.
Some research groups have elected to explore completely
different pathways to prebiotic pyrimidine nucleoside formation,
including the construction of the cytosine base on a sugar
phosphate. A stepwise synthesis of the base cytosine on a
9
3
plausible prebiotic reaction. However, it is still not clear what
preformed sugar was first accomplished by Sanchez and Orgel,
causes a kinetic barrier to the glycosidic bond formation for
the pyrimidines uracil and cytosine, and until a reasonably
efficient and plausibly prebiotic synthesis is found for uridine
and cytidine, it will be difficult to accept the proposal that RNA
was the first informational polymer of life.
and more recently explored in greater detail by the Sutherland
4
laboratory. This approach has proven to be regioselective,
†
Florida Gulf Coast University.
Georgia Institute of Technology.
Jackson State University.
‡
§
Given the fundamental importance of the glycosidic bond to
many aspects of chemistry and biology, the theoretical study
on glycosidic bond formation has received surprisingly little
attention. Most theoretical studies have focused on glycosidic
(
(
1) Zubay, G.; Mui, T. Origins Life EVol. B. 2001, 31, 87.
2) Fuller, W. D.; Sanchez, R. A.; Orgel, L. E. J. Mol. EVol. 1972, 1,
2
49.
(
(
3) Sanchez, R. A.; Orgel, L. E. J. Mol. Biol. 1970, 47, 531.
4) Ingar, A.-A.; Luke, R. W. A.; Hayter, B. R.; Sutherland, J. D.
ChemBioChem 2003, 4, 504.
(
5) Powner, M. W.; Anastasi, C.; Crowe, M. A.; Parkes, A. L.; Raftery,
J.; Sutherland, J. D. ChemBioChem 2007, 8, 1170.
(6) Powner, M. W.; Gerland, B.; Sutherland, J. D. Nature 2009, 459, 239.
(7) Keefe, A. D.; Miller, S. L. J. Mol. EVol. 1995, 41, 693.
1
6088 9 J. AM. CHEM. SOC. 2009, 131, 16088–16095
10.1021/ja900807m CCC: $40.75  2009 American Chemical Society

Energetics of Pyrimidine Nucleoside Formation
A R T I C L E S
1
0-13
bond solvolysis.
Stubbs and Marx presented a Car-Parrinello
The 6-31G(d) basis set is a good compromise between efficiency
1
8
and accuracy and has been proven to reproduce molecular
parameters well. The further expansion of the basis set has less
molecular dynamics study on the mechanism of acid-catalyzed
glycosidic bond formation between methanol and R-D-glucopy-
1
9
1
4
impact on the accuracy of the molecular geometries. We have
also used the 6-311+G(d,p) and 6-31++G(d,p) basis sets for some
of the studied species and obtained similar geometrical parameters
as predicted with the smaller basis set. In addition, the energy
ranoside in aqueous solution at 300 K. Due to the differences
in the reactants and reaction conditions, the reaction mechanism
for pyrimidine nucleoside formation is expected to be different.
Therefore, a comprehensive theoretical study is desirable. In
this work, we present a detailed computational study of
glycosidic bond formation during a pyrimidine nucleoside
formation reaction at the density functional theory (DFT) and
the MP2 level of theory. The goal of this study is (1) to
determine the mechanistic details of the glycosidic bond
formation in a pyrimidine nucleoside from a free base and a
nonactivated sugar in a plausible prebiotic reaction, (2) to
provide insights regarding optimal prebiotic reaction conditions
for nucleoside formation, (3) to investigate the role of metal
ions in the reaction mechanism, (4) to study the solvent effect
on the reaction mechanism, as well as (5) to explore different
pyrimidine base analogues that would be able to form the
glycosidic bond under reaction conditions that have been
demonstrated for purine and 2-pyrimidinone nucleoside synthesis.
2
+
evaluations for the Mg -catalyzed reaction at the B3LYP/6-
311+G(d,p), B3LYP/6-31++G(d,p), and MP2/6-311++G(d,p)
levels do not alter the reaction mechanism. Therefore, we will report
the results obtained from the 6-31G(d) basis set unless otherwise
noted.
The reactions, to which the theoretical models are compared,
were conducted in dH O solutions. The solvent effects on the
2
reaction mechanism were first studied by using the Onsager
20
model. In this model, the liquid is represented by a dielectric
continuum, characterized by its dielectric constant, ε. The solute is
placed in a cavity created in the continuum. The distribution of
electronic density of the solute polarizes the continuum and
generates an electric field inside the cavity, which in turn affects
the geometry and electronic structure. In order to mimic the
experimental conditions, the dielectric constant of ε ) 78.39
(
corresponding to water) was used.
However, the specific interactions, such as hydrogen bonding,
are not included in the Onsager model. In the literature, different
approaches have been adopted in order to study the effect of bulk
water on the reaction mechanisms. Car-Parrinello ab initio
II. Computational and Experimental Methodologies
A. Computational Methodologies. The choice of theoretical
level for a given chemical problem depends on the required accuracy
and the size of a molecule. DFT-B3LYP method has been
1
4,21
molecular dynamics
would be an excellent tool to study the
role of water molecules, especially when water molecules are
directly involved in the proton-transfer process. The combined
hybrid SCRF solvation models with explicit water molecules have
15
demonstrated to predict excellent molecular geometries. Therefore,
the possible reaction mechanisms for the glycosidic bond formation
between ribose and the free nucleoside bases (2-pyrimidinone, an
analogue of uracil and cytosine) were studied using the B3LYP
22
also been shown to be successful. In this work, we investigated
the role of specific water molecules using the Onsager solvation
model with explicit water molecules in the second coordination
1
6
nonlocal density functional approximation. In order to address
what causes the kinetic barrier to glycosidic bond formation for
pyrimidines, the reaction between ribose and uracil was also studied
for comparison. The geometry of reactants, products, transition
states, and intermediates were optimized by means of the Berny
23
sphere or shell of the complexes. All explicit molecule conforma-
tions were optimized. All calculations were carried out using the
2
4
Gaussian 03 program.
B. Materials and Experimental Methods. Zebularine and
2-pyrimidinone were obtained from Sigma-Aldrich. Magnesium
chloride was dried in vacuo prior to use. For solution-state studies
of zebularine glycosidic bond cleavage, 200 µL portions of solutions
1
7
approach, a modified Schlegel method. Vibrational frequency
calculations were performed to confirm whether the obtained
geometry represents a transition or minimum energy structure.
2
containing 48 mM zebularine with and without 3.6 M MgCl were
(
8) (a) Benner, S.; Burgstaller, P.; Battersby, T.; Jurczyk, S. Did the RNA
world exploit an expanded genetic alphabet? In The RNA World.
Second Edition: The Nature of Modern RNA Suggests a Prebiotic RNA
World; Gesteland, R. F., Atkins, J. F., Eds.; Cold Spring Harbor
Laboratory Press: Cold Spring Harbor, NY, 1999; pp 163-181. (b)
Hud, N. V.; Anet, F. A. L. J. Theor. Biol. 2000, 205, 543. (c) Kolb,
V. M.; Dworkin, J. P.; Miller, S. L. J. Mol. EVol. 1994, 38, 549. (d)
Mittapalli, G. K.; Osornio, Y. M.; Guerrero, M. A.; Reddy, K. R.;
Krishnamurthy, R.; Eschenmoser, A. Angew. Chem., Int. Ed. 2007,
(18) (a) Sheng, Y.; Leszczynski, J.; Garcia, A. A.; Rosario, R.; Gust, D.;
Springer, J. J. Phys. Chem. B 2004, 108, 16233. (b) Holm e´ n, A.; Broo,
A. Int. J. Quantum Chem. 1995, QBS22, 113. (c) Broo, A.; Holm e´ n,
A. Chem. Phys. 1996, 211, 147.
(19) (a) Hariharan, P. C.; Pople, J. A. Chem. Phys. Lett. 1972, 66, 217. (b)
Hehre, W. J.; Radom, L.; Schleyer, P. v. R.; Pople, J. A. Ab Initio
Molecular Orbital Theory; Wiley: New York, 1986.
(20) (a) Cossi, M.; Scalmani, G.; Rega, N.; Barone, V. J. Chem. Phys.
2002, 117, 43. (b) Cances, M. T.; Mennucci, V.; Tomasi, J. J. Chem.
Phys. 1997, 107, 3032. (c) Onsager, L. J. Am. Chem. Soc. 1936, 58,
1486. (d) Kirkwood, J. G. J. Chem. Phys. 1934, 2, 351. (e) Wong,
M. W.; Frisch, M. J.; Wiberg, K. B. J. Am. Chem. Soc. 1991, 113,
4776. (f) Wong, M. W.; Wiberg, K. B.; Frisch, M. J. J. Am. Chem.
Soc. 1992, 114, 523. (g) Wong, M. W.; Wiberg, K. B.; Frisch, M. J.
J. Chem. Phys. 1991, 95, 8991. (h) Wong, M. W.; Wiberg, K. B.;
Frisch, M. J. J. Am. Chem. Soc. 1992, 114, 1645.
(21) (a) Car, R.; Parrinello, M. Phys. ReV. Lett. 1985, 55, 2471. (b) Marx,
D.; Hutter, J. Ab-initio Molecular Dynamics: Theory and Implementa-
tion; Modern Methods and Algorithms in Quantum Chemistry Fors-
chungzentrum Juelich, NIC Series; 2000; Vol. 1, http://www.theochem.
ruhr-uni-bochum.de/research/marx/cprev.en.html.
(22) (a) Privalo, T.; Samec, J. S. M.; Bckvall, J.-E. Organometallics 2007,
26, 2840. (b) Cimino, P.; Barone, V. THEOCHEM 2005, 729, 1. (c)
Amovilli, C.; Barone, V.; Commi, R.; Cances, E.; Cossi, M.; Mennucci,
B.; Pomelli, C. S.; Tomasi, J. AdV. Quantum Chem. 1999, 32, 227.
(d) Brancato, G.; Di Nola, A.; Barone, V.; Amadei, A. J. Chem. Phys.
2005, 122, 154109.
(23) Sheng, Y.; Roszak, S.; Leszczynski, J. J. Chem. Phys. 2004, 120, 4324.
(24) Frisch, M. J.; et al. Gaussian 03, Revision E.01; Gaussian, Inc.,
Wallingford CT, 2004.
4
6, 2478. (e) Mittapalli, G. K.; Reddy, K. R.; Xiong, H.; Munoz, O.;
Han, B.; De Riccardis, F.; Krishnamurthy, R.; Eschenmoser, A. Angew.
Chem., Int. Ed. 2007, 46, 2470.
(
9) Bean, H. D.; Sheng, Y.; Collins, J. P.; Leszczynski, J.; Hud, N. V.
J. Am. Chem. Soc. 2007, 129, 9556.
(
10) Fothergill, M.; Goodman, M. F.; Petruska, J.; Warshel, A. J. Am. Chem.
Soc. 1995, 117, 11619.
(
(
(
11) Dinner, A. R.; Blackburn, G. M.; Karplus, M. Nature 2001, 413, 752.
12) Gu, J.; Xie, Y.; Schaefer III, H. F. J. Am. Chem. Soc. 2005, 127, 1053.
13) Torres, R. A.; Himo, F.; Bruice, T. C.; Noodleman, L.; Lovell, T.
J. Am. Chem. Soc. 2003, 125, 9861.
(
(
(
14) (a) Stubbs, J. M.; Marx, D. Chem.sEur. J. 2005, 11, 2651. (b) Stubbs,
J. M.; Marx, D. J. Am. Chem. Soc. 2003, 125, 10960.
15) (a) Bauschlicher, C. W. Chem. Phys. Lett. 1995, 246, 40. (b) El-Azhary,
A. A.; Suter, H. U. J. Phys. Chem. 1996, 100, 15056.
16) (a) Lee, C.; Yang, W.; Parr, R. Phys. ReV. B. 1988, 37, 785. (b) Becke,
A. D. J. Chem. Phys. 1993, 98, 5648. (c) Miehlich, B.; Savin, A.;
Stoll, H.; Preuss, H. Chem. Phys. Lett. 1989, 90, 5622. (d) Stephens,
P. J.; Devlin, F. J.; Chabalowski, C. F.; Frisch, M. J. J. Phys. Chem.
1
994, 98, 11623.
(
17) Schlegel, H. B. J. Comput. Chem. 1982, 3, 214.
J. AM. CHEM. SOC. 9 VOL. 131, NO. 44, 2009 16089

A R T I C L E S
Sheng et al.
Table 1. B3LYP/6-31G(d)-Predicted Activation Energies of
Glycosidic Bond Formations between Ribose and 2-Pyrimidinone,
Uracil (in kcal/mol)
pathway A
pathway B
Gas Phase
a
ribose + pyrimidinone
ribose + uracil
ribose + protonated-pyrimidinone
52.7
50.8
61.1
59.2
34.3
a
-
Figure 1. A possible reaction pathway between ribose and 2-pyrimidinone
in the gas phase. Labeled distances are in angstroms. The structures are
depicted using standard CPK coloration.
Water
ribose + pyrimidinone
ribose + uracil
49.4
46.6
52.2
51.4
With Mg2
+
incubated in individual vials at 100 °C. Vials were removed at
designated times and stored at 4 °C until analyzed by HPLC. A
2
+
a,b,c
ribose + protonated pyrimidinone + Mg
20.2
2
+
a
saturated solution of MgCl
2
(calculated to be 4.8 M at 100 °C)
ribose + uracil + Mg
39.9
-
containing 48 mM zebularine was stirred at 100 °C, with 200 µL
aliquots removed every hour and stored at 4 °C until analyzed.
For drying-heating measurements of zebularine glycosidic bond
cleavage, 200 µL portions of solutions containing 5 mM zebularine
a
b
q
Data were originally reported in ref 9. ∆E ) 25.44 kcal/mol, the
single-point relative energies in water at the B3LYP/6-31G(d)-SCRF
Onsager solvation model) optimized geometries. ∆E ) 25.20 kcal/
mol, at extended model with the addition of explicit water molecules
within the SCRF solvation model; all explicit molecule conformations
were optimized.
c
q
(
and 200 mM MgCl
2
were dried for periods of 15-120 min at 100
°
C on 2 in. watch glasses. Each watch glass was then washed three
times with 200 µL of deionized water, and washings were combined
and lyophilized. The solid was resuspended in 200 µL of water
prior to HPLC analysis.
In this reaction pathway, 2-pyrimidinone first forms a complex
A1) with ribose. The complex is stabilized by three hydrogen-
(
For all samples, HPLC separation was conducted using a
bonding interactions: one between the ribose C1′ hydroxyl
oxygen atom (O1′) and the N1 proton (H1) of 2-pyrimidinone,
the second between the 2-pyrimidinone carboxyl oxygen atom
and the O1′ proton of ribose, and the third between the C5′
hydroxyl (O5′) proton of ribose and the carboxyl oxygen of
Phenomenex Luna NH
2
column (4.6 × 250 mm, 5 µ) and 80%
acetonitrile:20% water isocratic elution at 1 mL/min flow rate.
Zebularine degradation was determined by comparing the integrated
HPLC peak intensity of free 2-pyrimidinone with the intensity of
the remaining zebularine.
2
-pyrimidinone. The corresponding hydrogen-bond distances
III. Results and Discussion
amount to 1.938, 1.857, and 2.034 Å, respectively.
In the A1 complex, the orientation of 2-pyrimidinone with
respect to ribose, with H1 pointing toward O1′, facilitates the
proton transfer from N1 to O1′. As H1 approaches O1′, the
C1′-O1′ bond breaks. At the same time N1 of 2-pyrimidinone
approaches C1′ of ribose. A transition state corresponding to
the transfer of H1 from N1 to O1′ and the cleavage of the
C1′-O1′ bond, as well as the formation of C1′-N1 bond
A. Reaction of Ribose with 2-Pyrimidinone and Uracil
2
+
without Mg . Several possible mechanisms for the reaction
between ribose and 2-pyrimidinone have been investigated. At
the initiation of our theoretical studies, we had hypothesized
that the hydroxypyrimidine tautomer of 2-pyrimidinone might
be the tautomer active during glycosidic bond formation.
However, our calculations revealed that the keto-enol tautom-
erisms of 2-pyrimidinone or protonated 2-pyrimidinone are
hampered by high activation energies (see Figure S1 of
Supporting Information). Therefore, it is unlikely that the
formation of ꢀ-nucleoside takes place through the hydroxypy-
rimidine tautomer, but rather through the pyrimidinone tautomer,
which is the preferred tautomer of 2-pyrimidnone in aqueous
(
TSA1) was identified and is shown in Figure 1. In the transition
structure TSA1, the distance between C1′ and O1′ elongates to
.554 Å, indicating that the C1′-O1′ bond is broken at this
3
point. The leaving OH group captures H1 and forms a water
molecule in TSA1. The distance between N1 and C1′ amounts
to 2.716 Å. The activation energy for this pathway is calculated
to be 52.7 kcal/mol.
1
4
solution. Stubbs and Marx also proposed that the formation
2
-Pyrimidinone can also approach ribose in a different
of methyl ꢀ-glucopyranoside from R-D-glucopyranose and
orientation where the carboxyl group is closer to the C3′
hydroxyl group (see Figure S2, Supporting Information), and
similarly, a high activation energy is predicted (61.13 kcal/mol).
The relative energies of the nucleoside product, zebularine, from
both pathways (PA1 and PB1) are -8.00 and -0.43 kcal/mol,
respectively, at the B3LYP/6-31G(d) level, indicating that both
pathways are exothermic and thermodynamically feasible.
However, the high activation energies for both pathways reveal
that glycosidic bond formation between ribose and 2-pyrimi-
dinone in the gas phase is kinetically inaccessible. In addition,
the activation energies for both pathways in water were also
evaluated by the Onsager solvation model and it was found that
the solvent effect only slightly lowers the activation barriers
N N
methanol proceeded through a D A mechanism, which in-
9
volves the formation of oxocarbenium ion. In our experiment,
only the ꢀ-conformation of the furanosyl ribonucleosides was
observed from the condensation of 2-pyrimidinone with D-
ribose; no R-furanosyl ribonucleoside was observed and, if
present, formed only in low yields. Therefore, the formation of
the C-N glycosidic bond in ꢀ-furanosyl ribonucleoside does
N N
not follow the D A mechanism. Under this circumstance, we
recognize that holding 2-pyrimidinone and the ribose C1′
hydroxyl in close proximity is crucial to form only the
-furanosyl ribonucleoside.
We first studied two possible reaction mechanisms of
ꢀ
glycosidic bond formation between ribose and 2-pyrimidinone
in the gas phase by holding 2-pyrimidinone and the ribose C1′
hydroxyl in close proximity. For comparison, the analogous
reaction between ribose and uracil was also investigated.
Reaction between Ribose and 2-Pyrimidinone without
(see Table 1). Therefore, the glycosidic bond formation between
ribose and a free 2-pyrimidinone base would not likely take
place in the gas phase or in pure water.
2+
Reaction between Ribose and Uracil without Mg . For
comparison, the possible reaction mechanisms of the glycosidic
bond formation between ribose and uracil were also studied.
2+
Mg . Figure 1 shows a possible reaction pathway for the
glycosidic bond formation between ribose and 2-pyrimidinone.
1
6090 J. AM. CHEM. SOC. 9 VOL. 131, NO. 44, 2009

Energetics of Pyrimidine Nucleoside Formation
A R T I C L E S
Figure 2. Acid-catalyzed glycosidic bond formation between ribose and
2+
2
-pyrimidinone in the absence of Mg . Labeled distances are in angstroms.
The structures are depicted using standard CPK coloration.
Two pathways similar to those of zebularine formation are
shown in Figure S3 (Supporting Information). Both pathways
possess high activation energies, 50.8 and 59.2 kcal/mol at the
B3LYP/6-31G(d) level. The inclusion of solvent effect only
slightly reduces the activation energies to 46.6 and 51.4 kcal/
mol, respectively (Table 1). This indicates that it is also unlikely
for uracil and ribose to form uridine in the gas phase or in a
pure water solution.
Figure 3. The structures and relative energies of reactants, complex,
intermediate, transition structures, and product for the reaction between
2+
ribose and protonated 2-pyrimidinone in the presence of Mg . Labeled
distances are in angstroms. Magnesium is drawn in yellow, and all other
atoms follow standard CPK coloration.
B. Brønsted Acid Catalyzed Reaction between Ribose and
-Pyrimidinone without Mg . In our previous work, ꢀ-ribo-
mental observations that zebularine formation is possible in the
presence of a Brønsted acid.
2+
9
2
9
nucleosides were synthesized from adenine (adenosine), hy-
poxanthine (inosine), and 2-pyrimidinone (zebularine) under two
reaction conditions: at pH 2.1 in the presence of a Brønsted
C. Lewis Acid Catalyzed Reaction of Ribose with 2-Py-
2+
rimidinone and Uracil in the Presence of Mg . We have shown
that zebularine formation by 2-pyrimidinone and ribose under
drying conditions can be promoted by divalent metal ion salts
2+
acid and at neutral pH in the presence of the Lewis acid, Mg .
Buffering the sample at pH 6.3 without the addition of
magnesium salts reduced nucleoside production to an unobserv-
able level, indicating that glycosidic bond formation is an acid-
9
at neutral pH, as was shown in earlier investigations of purine
2
nucleoside synthesis. Among several divalent metal ions tested,
2+
Mg is one of the most efficient catalysts for glycosidic bond
2
5
a
catalyzed reaction. As 2-pyrimidinone has a pK of 2.24, the
formation (Hud laboratory, unpublished). We therefore studied
glycosidic bond formation between protonated 2-pyrimidinone
majority of the base will be protonated at pH 2.1, the acidic
reaction condition created by the Brønsted acid; we therefore
modeled glycosidic bond formation between protonated 2-py-
rimidinone and ribose at the B3LYP/6-31G(d) level.
2+
and ribose in the presence of the Lewis acid Mg at the B3LYP/
6
-31G(d) level.
The mechanism for the reaction between protonated 2-pyri-
The acid-catalyzed reaction mechanism is shown in Figure
2+
midinone and ribose in the presence of Mg (Figure 3) is found
to be completely different from the mechanisms apparently
favored in the absence of metal ions.
2
. Only one pathway was located, similar to the uncatalyzed
zebularine formation in the gas phase. Initially, the protonated
-pyrimidinone forms a complex (C1) with ribose in which two
H-bonds are found: one is between H1 of 2-pyrimidinone and
O1′ of ribose, measuring 1.749 Å, and the other between the
O5′ of ribose and H6 from 2-pyrimidinone, measuring 2.052
Å. Next, the C1′ and N1 approach each other, and H1 migrates
from N1 of 2-pyrimidinone to the leaving C1′ hydroxyl group
to form a free water molecule, corresponding to the transition
structure (TSC1) shown in Figure 2. The distances between the
2
First, a distorted trigonal bipyramidal complex, Mg-COM,
is formed between ribose and protonated 2-pyrimidinone with
2+
the assistance of a Mg ion. The ribose is bidentately
2+
coordinated to the Mg ion at O5′ and the ring oxygen, O4′.
These coordinations are facilitated through the interactions of
the lone pair electrons of the oxygen atoms and the d-orbitals
2+
of the Mg ion. The other three vertices of the trigonal
bipyramid are occupied by the protonated 2-pyrimidinone
carboxyl oxygen atom and two water molecules. The ribose and
the protonated 2-pyrimidinone are in cis-conformation so that
the N1 proton of the protonated 2-pyrimidinone is orientated
toward O1′, with a distance of only 1.737 Å. This orientation
facilitates the transfer of the N1 proton to O1′.
C1′-N1, N1-HH O, and C1′-OH2O amount to 2.662, 2.585, and
2
3
.726 Å, respectively.
It is interesting that the distances between N1-HH2O and
C1′-OH O are elongated from 1.718 and 2.704 Å, respectively,
2
in TSB1 to 2.585 and 3.726 Å in TSC1, while the C1′-N1
distance shortens to 2.662 Å in TSC1 from 2.915 Å in TSB1.
These changes are due to charge redistribution when a proton
is introduced into the system. The B3LYP/6-31G(d)-calculated
charge densities are +2.052e for ribose and -0.548e for
In the Mg-TS1 transition structure, the breaking of the
2+
C1′-O1′ bond is mediated by the Mg ion. One can see from
Figure 3 that a bond has actually formed between the leaving
OH group and the Mg atom. It is also apparent that the Mg-
bonded OH group forms a strong hydrogen bond with H1 of
2
-pyrimidinone in TSC1, while the charge densities on ribose
and 2-pyrimidinone in the uncatalyzed TSB1 are +1.227e and
0.687e, respectively. The larger positive charge on the ribose
fragment enhances the attraction between the ribose and the
-pyrimidinone in the TSC1, which leads to a shorter C1′-N1
distance, and a dramatic decrease of the activation energy. The
activation energy decreases from 61.1 kcal/mol for the uncata-
lyzed reaction in the gas phase to 34.3 kcal/mol for the acid-
catalyzed reaction. These calculations help explain our experi-
2
-pyrimidinone, with a bond length of only 1.580 Å. As a
-
consequence, the N1-H1 bond elongates from 1.019 Å in Mg-
2+
COM to 1.093 Å in Mg-TS1. The presence of the Mg ion
2
greatly lowers the activation barrier needed to break the
C1′-O1′ and N1-H1 bonds. The calculated activation energy
amounts to 20.15 kcal/mol, which is much lower than the
activation energies of the reaction mechanisms in the absence
2+
of a Mg ion.
In the Mg-TS1 transition structure the distance between C1′
and N1 measures 4.231 Å. It is expected that removing the
(
25) Brown, D. J. Nature 1950, 165, 1010.
J. AM. CHEM. SOC. 9 VOL. 131, NO. 44, 2009 16091

A R T I C L E S
Sheng et al.
Table 2. Relative Energies, Activation Energies of Complexes,
Transition Structures, Intermediates, and Product at the Different
Levels of Theory (in kcal/mol)
proton and hydroxyl group from 2-pyrimidinone and ribose,
respectively, would clear the way for C1′ and N1 to form a
glycosidic bond. However, instead of forming a zebularine
product, the transition structure Mg-TS1 leads to an intermediate
Mg-INT, in which the distance between C1′ and N1 becomes
quite long (6.979 Å at the B3LYP/6-31G(d) level). A larger
basis set was also used to confirm these structural parameters,
and the C1′-N1 distance only shortens to 6.929 Å at the
B3LYP/6-31++G(d,p) level. Considering that the reaction takes
place in water, the solvent effect was investigated by using the
Onsager model at the B3LYP/6-31G(d) level. These results
indicate that the solvent effect only slightly shortens the C1′-N1
distance (6.537 Å) in the intermediate Mg-INT.
B3LYP/
B3LYP/
6-31++G(d,p)
MP2/
6-31++G(d,p)
B3LYP/
b
6-31G(d)
a
a
a
6-31G(d)
Mg-C1
Mg-TS1
Mg-INT
Mg-TS2
Mg-P
0.00
20.15
6.59
14.04
2.32
0.00
19.00
4.08
12.13
1.41
0.00
22.89
10.00
18.97
-5.31
0.00
25.44
16.19
24.69
9.31
a
The single-point relative energies in the gas phase at the B3LYP/
b
6
-31G(d)-optimized geometries. The single-point relative energies in
water at the B3LYP/6-31G(d)-SCRF (Onsager solvation model) opti-
mized geometries.
In the transition structure Mg-TS1, the charge distributions
on N1, H1, OOH, and C1′ atoms are -0.634e, 0.447e, -0.873e,
and 0.564e, respectively. There is a competition of attractions
between both N1 · · ·H1 and H1 · · ·OOH pairs. The stronger
attraction between H1 · · ·OOH facilitates the migration of H1 to
the Mg-bonded hydroxyl group, forming a water molecule (Mg-
INT). The newly formed Mg-bonded water molecule forms a
hydrogen bond with the deprotonated nitrogen atom in the
intermediate Mg-INT (Figure 3), with charge distributions on
N1 and HH2O atoms of -0.567e and 0.312e, respectively. The
existence of this strong interaction temporarily stabilizes the
intermediate.
In order to form a glycosidic bond between the C1′ and N1
Figure 4. The optimized complex (Mg-COM-w) and transition state (Mg-
TS1-w) of the rate-determining step of glycosidic C-N bond formation, at
the B3LYP/6-31G(d) level using the SCRF solvation model with explicit
water molecules. Labeled distances are in angstroms. Magnesium is drawn
in yellow, and all other atoms follow standard CPK coloration.
atoms, the interaction between the N1 and HH O atoms must be
2
removed. This can be achieved by rotation of the Mg-bound
water molecule or the 2-pyrimidinone ring. In searching for the
corresponding transition state, structure Mg-TS2 was located
(
Figure 3). In this structure, the dihedral angle ∠C2-
levels of theory are similar. Thus, the use of a larger basis set
and the inclusion of solvation effects do not change the reaction
mechanism.
O2-Mg-OH O between the pyrimidinone ring and the newly
formed water molecule with respect to the Mg atom measures
2
-
84.0°, and the dihedral angles between the pyrimidinone ring
and the other two water molecules with respect to the Mg atom
C2-O2-Mg-OH2O) are 175.5° and -93.5°. For comparison,
in Mg-INT the corresponding dihedral angles are -12.05°,
The role of water molecules on the reaction mechanism in
2+
the presence of a Mg ion was also investigated by an extended
(
model with the addition of explicit water molecules within the
SCRF solvation model. Only the first step of the reaction
between ribose and 2-pyrimidinone was studied because it is
the rate-determining step. The optimized structures of the
complex (Mg-COM-w) and the transition state (Mg-TS1-w)
are shown in Figure 4. From Figure 4 one can see that the
additional water molecules only slightly change the geometrical
parameters of the complex and the transition state, and the
activation energy amounts to 25.2 kcal/mol. Such a small change
in the activation energy is attributed to the fact that water
molecules are not directly involved in the bond breaking and
formation. Therefore, the results presented in our work, while
somewhat limited due to the continuum solvent model, do
provide a reasonable computational approach for comparing
alternative possible reaction pathways and provide results that
are consistent with experimental studies.
8
6.7°, and 175.2°, respectively. This difference indicates that
the pyrimidinone ring and newly formed water molecule have
rotated away from each other to form Mg-TS2. The frequency
analysis shows only one imaginary frequency (25.6i cm )
corresponding to the rotation of the pyrimidinone ring around
the Mg-O2 bond, associated with the rotation of the newly
formed water molecule in the opposite direction.
-
1
The existence of Mg-TS2 is quite reasonable because the
formation of the C1′-N1 bond is hampered by the hydrogen-
bond interaction revealed in Mg-INT, while the rotations of
the water molecule and the pyrimidinone ring remove this
interaction, clearing the way for N1 to approach the C1′ atom.
Since rotations usually require little energy, this step only needs
to overcome a small activation barrier of about 7.5 kcal/mol.
In Mg-TS2, the C1′-N1 distance amounts to 6.066 Å at the
B3LYP/6-31G(d) level. An IRC analysis confirmed that the
transition structure Mg-TS2 is the transition state connecting
the intermediate (Mg-INT) and the nucleoside product (Mg-
P), a distorted octahedral complex.
The solvent effect on the reaction mechanism of zebularine
formation was also investigated. The structures of the complex,
reactants, transitions structures, and intermediate were reopti-
mized at the B3LYP/6-31G(d) level using the Onsager model.
Table 2 summarizes the relative energies of all these species
All levels of calculations demonstrate that the formation of
the glycosidic bond between ribose and 2-pyrimidinone in the
presence of a Mg2 ion is a three-step reaction: First, 2-pyri-
midinone is protonated in the acid solution and forms a complex
with sugar with the assistance of a Mg2 ion. Second, the
C1′-O1′ bond breaks with the assistance of a metal atom, and
the departing hydroxide group abstracts a proton from 2-pyri-
midinone, yielding a water molecule. Finally, the rotations of
the newly formed water molecule and the pyrimidinone ring
clear the way for the creation of the glycosidic bond between
the C1′ and N1 atom of 2-pyrimidinone so that the zebularine
product is formed.
+
+
2
+
and the product of the Mg -assisted reaction at the different
levels of theory. We note that the energy profiles at the different
1
6092 J. AM. CHEM. SOC. 9 VOL. 131, NO. 44, 2009

Energetics of Pyrimidine Nucleoside Formation
A R T I C L E S
tially due to its lower pK
a
, which results in a reduction of the
protonated form of the base, relative to 2-pyrimidinone, at pH
.4.
The details of the B3LYP/6-31G(d)-optimized structures of
2
the pyrimidines are shown in Figure 5. Careful examination
of the geometrical parameters of the 2-pyrimidinone and
4
-pyrimidinone isomers reveals that the N-H bond length in
the protonated 2-pyrimidinone amounts to 1.019 Å, 0.006 Å
longer than in the unprotonated 2-pyrimidinone, which con-
tributes to the lower energy required for the abstraction of the
H1 proton during ribosylation. The N1-H1 and N3-H3 bond
lengths in the protonated 4-pyrimidinone are 1.020 and 1.017
Å, respectively. It seems reasonable that the protons connected
to either nitrogen atom of the pyrimidinone bases can be donated
in the reaction. However, according to the transition structure
Mg-TS1 presented in the section III.C, the coordination of the
carboxyl oxygen atom to the magnesium ion is crucial to
maintain the orientation of the pyrimidine ring so that the proton
can be transferred from the nitrogen atom to the hydroxyl group
in the reaction pathway. Once coordinated, H1 of 4-pyrimidi-
none is simply too far away to be able to adopt a proper
orientation with the C1′ hydroxyl group. We note, however,
that a proton could alternatively be delivered from solution.
Nevertheless, the N1-H1 bond would require more energy to
break, as the bond length of N1-H1 is slightly shorter than the
N3-H3 bond. H and H- C NMR data also indicate that N3
is more nucleophilic than N1, as the addition of 4-pyrimidinone
to acetaldehyde (an open-chain model of an aldose sugar) occurs
at N3 almost exclusively, i.e., no acetaldehyde addition at N1
was observed (see Figure S5 and Tables S1 and S2, Supporting
Information). Combining these results, it is predicted that the
nucleosides with 4-pyrimidinone connected through N3 to ribose
would be more favored.
Figure 5. Pyrimidines tested in the nucleoside-formation reaction with
ribose. (Bases are shown in their protonated forms.)
Reaction between Ribose and Uracil in the Presence of
2+
Mg . For comparison, the optimized structures of the complex,
transition structure, and product for the glycosidic bond forma-
tion between ribose and uracil in the presence of Mg2+ are
shown in Figure S4 (Supporting Information). Unlike the
reaction with 2-pyrimidinone, the reaction with uracil is a two-
2
+
step reaction in the presence of Mg : First the uracil forms a
complex (Mg-CU) with ribose and then deprotonation of N1
of uracil and the breaking of the C1′-O1′ bond are asynchro-
nically concerted with the glycosidic bond formation between
atoms C1′ and N1. The predicted activation energy amounts to
3
9.9 kcal/mol in the gas phase. When the solvation effects are
considered, the activation energy actually increases to 41.2 kcal/
mol. Thus, the glycosidic formation between ribose and uracil
is not likely to proceed under the reaction conditions investi-
gated. These theoretical results are in agreement with our
1
1
13
9
previous experimental observations that 2-pyrimidinone forms
ꢀ
-nucleosides in good yield (greater yields than both adenine
and hypoxanthine), while uracil does not form nucleosides in
detectable yields under any of the conditions tested.
D. Will Other Pyrimidines Spontaneously Form Nucle-
osides with Ribose? The coupling of 2-pyrimidinone to ribose,
as described in this work, can produce ꢀ-nucleosides in relatively
good yield. However, the experimental conditions outlined
above might be somewhat specific for glycosidic bond formation
between 2-pyrimidinone and ribose. It remains to be demon-
strated whether similar conditions favor nucleoside formation
reactions with other pyrimidine bases and various sugars. Thus,
it is of great interest to determine which chemical features of
As discussed in the preceding section, the reaction of the
uracil with ribose is apparently not favorable because of the
localization of charge on N1, which enhances the N1-H1 bond
strength (N-H bond lengths in uracil are short, being 1.011
and 1.014 Å) and creates a substantial kinetic barrier to
pyrimidine glycosidic bond formation. The pK of N1 of uracil
a
is about 9.5, so that there is less than 1% of the anionic form
2
6
at neutral pH. However, Kimura et al. have shown that
placement of the uracil base near a cationic species can
2
-pyrimidinone are essential for this reaction to occur. As a
first step toward addressing this question, another pyrimidine
related to 2-pyrimidinone was also tested for its ability to form
nucleosides with ribose under similar reaction conditions. The
structures of the three pyrimidines tested in this study are shown
in Figure 5.
selectively reduce the pK
coordination of uracil within a supramolecular assembly, such
as in a coaxial stack with cationic intercalators, might
sufficiently lower the N1 pK to allow glycosylation.
E. Nucleoside DegradationsComparing Theory and Ex-
periment. By the principle of microscopic reversibility, the
transition state associated with glycosidic bond cleavage must be
the same, under identical reaction conditions, as the transition state
of glycosidic bond formation. The B3LYP/6-31G(d)-calculated
activation energy for zebularine glycosidic bond cleavage, in water
at neutral pH, is 26.1 kcal/mol. In the presence of Mg , this
activation energy is predicted to reduce to 13.6 kcal/mol. These
calculations are consistent with the observation that the glycosidic
a
of N1 by more than 2 pH units. Thus,
8
b
a
The pyrimidine most comparable to 2-pyrimidinone is 4-py-
rimidinone. This base has exactly the same chemical composi-
tion as 2-pyrimidinone, but the carboxyl group at C4 is adjacent
to only one nitrogen atom, rather than being in between the
two nitrogen atoms of the pyrimidine ring. This structural
2+
difference is associated with the lower 4-pyrimidinone pK
a
of
of 2-pyrimidinone. When free
-pyrimidinone is heated and dried under mildly acidic condi-
19
1.84, as compared the 2.24 pK
a
4
9
bond of zebularine is much less stable than that of uridine, which
has calculated uncatalyzed and Mg -catalyzed activation energies
of 45.2 and 39.0 kcal/mol, respectively.
tions (pH 2.4) in the presence of ribose, products are observed
with UV absorbance and HPLC retention times that are
consistent with 4-pyrimidinone nucleoside formation (data not
shown). However, these putative nucleosides are produced at
most in one-fourth the yield of those formed in the analogous
2+
(
(
26) Kimura, E.; Kitamura, H.; Koike, T.; Shiro, M. J. Am. Chem. Soc.
1
997, 119, 10909.
2-pyrimidinone reaction (i.e., under similar reaction conditions).
27) Burgess, J. Metal Ions in Solution; Ellis Horwood Ltd.: Chichester,
The decrease in the product yield in 4-pyrimidinone is poten-
England, 1978.
J. AM. CHEM. SOC. 9 VOL. 131, NO. 44, 2009 16093

A R T I C L E S
Sheng et al.
Thus, drying in the presence of Mg2 accelerates the rate of
+
3
zebularine hydrolysis by greater than 6.5 × 10 over that measured
2+
in the absence of Mg , corresponding to a decrease in transition
barrier height of more than 6.6 kcal/mol, which is in the very least
consistent with the predicted decrease of 12.5 kcal/mol. Given the
apparent additional reduction in the barrier height of glycosidic
bond cleavage upon drying from a solution containing MgCl
compared to that of a saturated MgCl solution at the same
temperature, our experimental results also support our computa-
2
,
2
2+
tional models for Mg -catalyzed glycosidic bond cleavage (and
formation), which include direct nucleoside coordination of a
partially dehydrated Mg2 ion.
+
IV. Conclusions
The present work was motivated by the long-standing problem
of how nucleosides were synthesized during the early stages of
life. Although it has been known for decades that purine bases
can form nucleosides with ribose when dried from an aqueous
solution, albeit in low yield, the first plausible prebiotic synthesis
of a pyrimidine nucleoside from a free base and a nonactivated
Figure 6. Hydrolysis of zebularine in the absence and presence of MgCl
2
.
2
Solution-state samples were 48 mM zebularine, in the absence of MgCl
filled circles) and in the presence of 3.6 M MgCl
M MgCl (filled triangles). Zebularine samples dried and heated (open
triangles) were initially 5 mM zebularine and 200 mM MgCl . All samples
(
2
(open circles) and 4.8
2
2
were neutral pH and maintained at 100 °C. Solid lines represent linear best
fits of the initial rates of zebularine degradation for the solution-state
samples. Error bars on dried and heated sample data represent the actual
values obtained for duplicate experiments, for which the average is plotted.
Calculated error bars for solution state data, based upon known sources of
error, are comparable to the size of data markers.
9
sugar was reported only recently. The lack of a mechanism for
the prebiotic synthesis for the pyrimidine bases has been considered
a crucial problem for most contemporary theories concerning the
origin of life. The discovery that 2-pyrimidinone and other
pyrimidine and pyrimidine-like bases can spontaneously form
nucleosides with ribose has significant implications regarding the
origin of the first RNA-like polymers, as it now seems feasible
that both purine and pyrimidine nucleosides could have been
formed on the prebiotic Earth without the aid of protein enzymes.
The reaction mechanisms of glycosidic bond formation between
ribose and free 2-pyrimidinone base have been investigated by
nonempirical theoretical calculations. The reaction potential energy
surfaces of the studied reaction mechanisms are charted in Figure
7. Upon the basis of the results obtained, we can draw the following
conclusions: (1) The theoretical study shows that the reaction
between ribose and free 2-pyrimidinone would not likely take place
in water at neutral pH due to the high activation energy. (2) The
study of the reaction between ribose and protonated 2-pyrimidinone,
which models an acidic experimental condition, shows that acid
does catalyze glycosidic bond formation, as the activation energy
decreases by 18.4 kcal/mol versus the reaction at neutral pH. (3)
The formation of the ꢀ-ribofuranosyl nucleoside product is greatly
facilitated by the presence of Mg2 ions in the acidic solution (pH
2.1). Our theoretical investigation shows that the reaction between
The zebularine degradation reaction presents an opportunity to
more directly compare theory and experiment, given that glycosidic
bond cleavage is a first-order reaction. Specifically, we sought to
compare experimental results for rates of zebularine glycosidic bond
2+
cleavage in the absence and presence of Mg with the calculated
change in transition barrier height corresponding to these reaction
conditions. In Figure 6, the fraction of zebularine remaining
(
2
normalized to the total concentration of zebularine and free
-pyrimidinone) is plotted as a function of time for four sets of
nucleoside degradation reactions. All reactions were at 100 °C and
2+
neutral pH but differed in Mg (and water) concentration/activity.
In the absence of MgCl , the rate of zebularine degradation was
2
-3
-1
measured to be 1.2 × 10 s . In the presence of 3.6 M MgCl
2
,
-
2 -1
this rate increases to 1.8 × 10
s
and in the presence of 4.8 M
-
1
MgCl
2
to 0.23 s . The increased rate of zebularine hydrolysis with
2+
increasing Mg concentration and reduced water activity is
consistent with Mg2 acting as a catalyst in glycosidic bond
+
2+
cleavage. However, even at 4.8 M Mg , the apparent reduction
in the activation barrier to zebularine hydrolysis of 4 kcal/mol is
+
below the predicted reduction of 12.5 kcal/mol. We note that 4.8
M Mg2 corresponds to the saturation limit for MgCl
+
at 100 °C.
ribose and protonated 2-pyrimidinone in the presence of Mg ions
2+
2
2+
Furthermore, Mg in aqueous solution exists with a very favorable
first shell of hydration. Thus, to facilitate a more straightforward
comparison with the calculated transition state that involves a
partially dehydrated Mg directly coordinated to zebularine, we
also investigated the degradation of zebularine when dried from a
is a three-step reaction: First, 2-pyrimidinone is protonated in the
acid solution and forms a complex with a ribose sugar with the
assistance of a Mg2 ion. Second, the C1′-O1′ bond breaks with
the assistance of the metal ion, and the departing hydroxide group
abstracts a proton from 2-pyrimidinone, yielding a water molecule.
Finally, the rotations of the newly formed water molecule and the
pyrimidinone ring clear the way for the formation of the glycosidic
bond between the C1′ atom of ribose and N1 atom of 2-pyrimi-
dinone. With the assistance of a Mg2 ion at pH 2.1, the activation
energy dramatically drops to 20.2 kcal/mol.
27
+
2+
solution containing MgCl
5-120 min, similar results were obtained. That is, an average of
3% of the zebularine remained after drying and heating, regardless
2
at 100 °C. For drying-heating times of
1
1
+
of drying time. Thus, in 15 min or less, zebularine degradation
was essentially complete, with the remaining zebularine apparently
representing the equilibrium fraction of the nucleoside that exists
during a step in the drying-heating process, i.e., where nucleoside
formation and degradation are in equilibrium. Shorter drying times
were not attempted due to uncertainty in the completeness and
uniformity in sample drying, upon the basis of visual inspection
of samples at shorter times (i.e., 5 min). Nevertheless, if we consider
that 13% of the zebularine remains after 15 min, we are able to
The activation energy of the zebularine formation reaction, at
2+
20-25 kcal/mol in the presence of a Mg ion, is consistent with
the fact that the reaction requires only moderate heating at
atmospheric pressure. The role of water molecules in the reaction
mechanism was also investigated by the extended model with
addition of explicit water molecules within the SCRF solvation
model, but it was found to only slightly reduce the activation
energy.
-1
set a lower limit on the rate of zebularine degradation at 8.2 s .
1
6094 J. AM. CHEM. SOC. 9 VOL. 131, NO. 44, 2009

Energetics of Pyrimidine Nucleoside Formation
A R T I C L E S
Figure 7. Potential energy surfaces for the zebularine formation reaction between 2-pyrimidinone and ribose (red) and between protonated 2-pyrimidinone
2
+
2+
and ribose (black) in the absence of Mg and between protonated 2-pyrimidinone and ribose in the presence of Mg (blue).
Important contributions to the study of prebiotic nucleoside
formation have been made with phosphate/2′,3′-borate modifica-
tions of the sugar; therefore, it is of significant interest to understand
these nucleoside formation reactions in the context of our proposed
model for the origin of RNA, the proto-RNA nucleosides would
have been originally selected for their more favorable chemistry
of formation, but eventually replaced by uracil and cytosine through
evolutionary processes based upon the superior nucleoside stability
and functionality of these contemporary bases.
28
magnesium-mediated reaction mechanism. Phosphate and borate
are expected to affect the reactivity of ribose toward the bases due
to the negative charges they carry. These studies are currently
underway, and the comprehensive investigation on reaction mech-
anisms between 5′-phosphorylated ribose or 2′,3′-borate ribose and
Acknowledgment. We thank A.E. Englehart for helpful dis-
cussions. This research was supported by the College of Arts &
Sciences, Florida Gulf Coast University and National Science
Foundation (Grant No. CHE-0739189).
2-pyrimidinone will be reported in a separate work.
Although the pyrimidine bases presently found in RNA do not
Supporting Information Available: The keto-enol tautom-
erism transition structure and enol form of protonated 2-pyri-
midinone (Figure S1); an alternative reaction pathway between
ribose and 2-pyrimidinone in the gas phase (Figure S2); two
possible reaction pathways between ribose and uracil in the gas
phase (Figure S3); the structures of reactants, complex, inter-
mediate, transition structures, and product for the reaction
spontaneously form nucleosides in a simple heating and drying
9
reaction, our experimental observations suggest that alternative
pyrimidine (or pyrimidine-like) bases could have formed the
nucleosides required for an ancestral RNA-like polymer. The
protopyrimidine bases 2-pyrimidinone and 4-pyrimidinone studied
in this work have more favorable glycosylation reaction kinetics
than cytosine and uracil, suggesting that their nucleosides, or the
nucleosides of other bases that more readily undergo glycosylation,
could have been more abundant in the prebiotic environment and
therefore might have been present in proto-RNAs. Within such a
2+
between ribose and uracil in the presence of Mg (Figure S4);
1
1
13
the H NMR spectrum and the HSQC and HMBC H- C
correlational spectroscopy data of nucleophilic addition of
4-pyrimidinone to acetaldehyde (Figure S5, Tables S1 and S2);
(
28) (a) Muller, D.; Pitsch, S.; Kittaka, A.; Wagner, E.; Wintner, C. E.;
Eschenmoser, A. HelV. Chim. Acta 1990, 73, 1410–1468. (b) Ricardo,
A.; Carrigan, M. A.; Olcott, A. N.; Benner, S. A. Science 2004, 303,
the full list of authors of ref 24; and Cartesian coordinates for
all the stationary points optimized at the B3LYP/6-31G(d) level
used in the paper (Table S3). This material is available free of
charge via the Internet at http://pubs.acs.org.
1
96. (c) Li, Q.; Ricardo, A.; Benner, S. A.; Winefordner, J. D.; Powell,
D. H. Anal. Chem. 2005, 77, 4503. (d) Benner, S. A. Acc. Chem. Res.
004, 37, 784. (e) Semmelhack, M. F.; Campagna, S. R.; Federle,
M. J.; Bassler, B. L. Org. Lett. 2005, 7, 569.
2
JA900807M
J. AM. CHEM. SOC. 9 VOL. 131, NO. 44, 2009 16095