L-amino acids catalyze the formation of an excess of D-glyceraldehyde, and thus of other D sugars, under credible prebiotic conditions

Article abstract of DOI:10.1073/pnas.1001639107

Previous work by us, and others, has shown that the formation of amino acids on prebiotic earth with the geometric arrangement called the L configuration can be understood. Some meteorites of the carbonaceous chondritic type deliver unusual amino acids, with alpha-methyl groups, which have an excess of the L isomers. We previously showed that in decarboxylative transamination reactions under credible prebiotic conditions they produce normal amino acids that also have a preference for the L isomer, as is found in our proteins. We, and others, showed that as little as a 1%excess of the L isomers could be amplified up to a 95/5 ratio of L over D on simple evaporation of a solution, so life could start with such a solution in which the dominant L isomers would be selectively chosen. We now find that the geometry of sugars referred to D, as in D-ribose or D-glucose, is not an independent mystery. D-glyceraldehyde, the simplest sugar with a D center, is the basic unit on which other sugars are built. We find that the synthesis of glyceraldehyde by reaction of formaldehyde with glycolaldehyde is catalyzed under prebiotic conditions to D/L ratios greater than 1, to as much as 60/40, by a representative group of L-amino acids (with the exception of L-proline). The D/L glyceraldehyde ratio in water solution is amplified to 92/8 using simple selective solubilities of the D and the DL forms. This D center would then be carried into the prebiotic syntheses of larger sugars.

Full text of DOI:10.1073/pnas.1001639107

L-amino acids catalyze the formation of an excess
of D-glyceraldehyde, and thus of other D sugars,
under credible prebiotic conditions
Ronald Breslow1 and Zhan-Ling Cheng
Department of Chemistry, Columbia University, New York, NY 10027
Contributed by Ronald Breslow, February 9, 2010 (sent for review February 1, 2010)
Previous work by us, and others, has shown that the formation of
amino acids on prebiotic earth with the geometric arrangement
called the L configuration can be understood. Some meteorites
of the carbonaceous chondritic type deliver unusual amino acids,
with alpha-methyl groups, which have an excess of the L isomers.
We previously showed that in decarboxylative transamination re-
actions under credible prebiotic conditions they produce normal
amino acids that also have a preference for the L isomer, as is found
in our proteins. We, and others, showed that as little as a 1% excess
of the L isomers could be amplified up to a 95/5 ratio of L over D on
simple evaporation of a solution, so life could start with such a
solution in which the dominant L isomers would be selectively
chosen. We now find that the geometry of sugars referred to
D, as in D-ribose or D-glucose, is not an independent mystery.
D-glyceraldehyde, the simplest sugar with a D center, is the basic
unit on which other sugars are built. We find that the synthesis of
glyceraldehyde by reaction of formaldehyde with glycolaldehyde
is catalyzed under prebiotic conditions to D/L ratios greater than 1,
to as much as 60/40, by a representative group of L-amino acids
(with the exception of L-proline). The D/L glyceraldehyde ratio
in water solution is amplified to 92/8 using simple selective solu-
bilities of the D and the DL forms. This D center would then be car-
ried into the prebiotic syntheses of larger sugars.
(6). We were all preceded by Morowitz, who described the theory
of such processes and some experimental exploration of it (7).
With a dominance of the L-amino acids in solution, where life
might have started, it is reasonable that the life form that used the
dominant L-amino acids would win out over any life form based
on the minor amount of D amino acid still in solution.
Although there are still questions about why the special amino
acids that came in on meteorites had a small excess of the L form
over the D form, one idea is that they were formed in space as an
equal mixture, but then some of the D form was destroyed by
high-energy light that has right circular polarization and that is
known to be in this sector of the universe. Such selective destruc-
tion of D amino acids by right circularly polarized light has been
demonstrated experimentally (8). The excess of light with right,
rather than left, circular polarization is often ascribed to a syn-
chrotron or cyclotron process from a nearby neutron star. If so
it is an accident, and in other parts of the universe there could
be a preference for D amino acids if those regions have an excess
of left circularly polarized light. There should be no preference
for one or the other handedness of their amino acids if they do
not have any excess of one or the other polarized light. Such a
preference is important for life to start, because on prebiotic
earth equal mixtures of an amino acid and its mirror image would
not easily form biological molecules, such as proteins, with well-
defined structures.
aldol reaction ∣ formaldehyde ∣ formose reaction ∣ meteorites
Results and Discussion
or life to start, it was necessary that the amino acids that are
Fbuilding blocks for proteins, and the sugars that play many bio-
logical roles and are part of both RNA and DNA, have a single
handedness, called chirality. The amino acids could have been all
in the handedness referred to as L, analogous to the left hand, or
they could have been all D, analogous to the right hand, but on
our planet the L geometry was preferred. After many years of
speculation about why this is so, evidence has now accumulated
that the L handedness of amino acids was derived from a special
group of amino acids, with an extra methyl group that prevented
loss of their small excess of the L form by a process that can
scramble the handedness of ordinary amino acids on prolonged
heating. These special L-amino acids with the extra methyl group
were isolated from a meteorite that fell in Murchison Australia in
1969 (1, 2). We showed that they could generate normal amino
acids (without the methyl group) under credible prebiotic condi-
tions, and those also had an excess of the L isomer, along with a
smaller amount of the D amino acids (3).
Once there was a small excess of the L-amino acid, we showed
that it could be amplified in solution, because the DL 1∶1 mixture
was less soluble in water (4). By evaporating a water solution with
only a 1% excess of L-phenylalanine, for instance, we obtained a
solution with a 95/5 L/D ratio; equal amounts of the D and L
components precipitated from the solution as a less soluble crys-
tal in which the D and L components were bound to each other.
At the same time, Klussman et al. published the first of a series of
papers on such selective concentration of the L-amino acids in
water (5) and Hayashi et al. published the amplification of proline
Sugars can also have either L or D handedness, but the situation
here is a bit more complex. D sugars are named that way for the
handedness of the carbon next to the end of the sugar chain, but
there are other places where chirality (handedness) also occurs in
sugars.
In D-ribose (Fig. 1) all three chiral carbons have the handed-
ness defined as D, which in modern chemical terminology is R,
but in other sugars the other carbons can have R or S handedness
and they are still called D sugars if the next to bottom carbon has
the D (R) handedness. For example, in D-glucose (Fig. 1) the
next to bottom carbon still has this R configuration, but one other
carbon nearer the top of the molecule has the S configuration.
All the D sugars were probably initially derived from
D-glyceraldehyde, the simple three-carbon sugar (Fig. 1). This
has only one carbon atom at which the arrangement of the groups
can describe either D-glyceraldehyde (Fig. 1) or its mirror-image
L-glyceraldehyde (Fig. 1). In the D sugars such as D-ribose
or D-glucose or D-fructose other pieces are added to
D-glyceraldehyde: a two-carbon piece to make D-ribose (9), a
three-carbon piece to make D-glucose, etc. Those natural sugars
are all in the D family, suggesting that they were derived prebio-
tically from D-glyceraldehyde by adding such pieces.
Author contributions: R.B. and Z.-L.C. designed research; Z.-L.C. performed research;
R.B. and Z.-L.C. analyzed data; and R.B. wrote the paper; .
The authors declare no conflict of interest.
1To whom correspondence should be addressed. E-mail: Rb33@columbia.edu.
www.pnas.org/cgi/doi/10.1073/pnas.1001639107
PNAS ∣ March 30, 2010 ∣ vol. 107 ∣ no. 13 ∣ 5723–5725

O
H
CO2H
O
H
H
H
C
C
HO
OH
CH2OH
R
C
H
CH2OH
NH2
L-Glyceraldehyde
D-Glyceraldehyde
L-Amino Acids
alanine (R = CH3); serine (R = CH2OH);
valine (R = CH(CH3)2); leucine
(R = CH2-CH(CH3)2); phenylalanine
(R = CH2-C6H5); glutamic acid
(R = CH2-CH2-CO2H
O
H
O
H
C
C
H
HO
H
C
C
C
C
H
H
OH
H
C
C
C
OH
OH
OH
OH
H
OH
H
CH2OH
CH2OH
D-Ribose
CO2H
N
D-Glucose
Fig. 1. The two mirror-image optical isomers of glyceraldehyde, the simplest
sugar, and the structures of D-ribose and D-glucose.
H
L-Proline
On prebiotic earth, sugars are believed to have formed from
formaldehyde. In a process called the formose reaction, formal-
dehyde can indeed be induced to form a series of different sugars
by a process (10) whose early step was shown to involve the
reaction of formaldehyde with glycolaldehyde leading to glycer-
aldehyde (Fig. 2). Without any special catalysis this process forms
an equal mixture of D- and L-glyceraldehyde. Thus we have now
investigated whether adding L-amino acids (Fig. 3) to the reac-
tion mixture would catalyze the process and lead to a preference
in the product, and if so what the preferred product would be.
If the L-amino acids had led to preferential formation of the
L-glyceraldehyde, it would be necessary to think of a completely
new reason for the dominance of D sugars in our biology. In fact,
all the L-amino acids we examined (except L-proline) catalyzed
the preferential formation of D-glyceraldehyde, as shown in
Table 1, although the enantiomeric excesses with L-alanine and
with L-serine were very small. The conditions of the reactions
were certainly credibly prebiotic: we simply held a solution of for-
maldehyde, glycolaldehyde, and various L-amino acids at 20 °C,
and then isolated and purified the glyceraldehyde formed as its
dinitrophenylhydrazone. The D/L ratio was then determined by
using chiral high-pressure liquid chromatography.
Fig. 3. The amino acids used as catalysts in the synthesis of glyceraldehyde.
structure determination shows (11), one L and one D enantiomer
form a six-membered ring, a substituted 1,4-dioxane in which the
substituents are equatorial and crystal packing is favorable. By
contrast, the dioxane structure from D-glyceraldehyde alone
would have an axial hydroxymethyl substituent and crystal pack-
ing would be less favorable. Thus selective solubility can amplify
the solution concentration of D-glyceraldehyde.
Starting with the 61/39 D/L ratio produced by catalysis with
glutamic acid, we find that the solution ratio is raised to 81/19
by equilibration with small amounts of water. When we started
with a solution that was 81/19 D/L in water, and slowly evapo-
rated the water, a small volume left after much of the DL com-
pound had precipitated was 92/8 D/L in solution. This is probably
not the upper limit of amplification. Also, we have shown with
amino acids that kinetic processes, such as would happen if rain
fell on the mixture and ran off before equilibrium was reached,
led to solution ratios that can be even higher than those reached
at equilibrium (3). We find that exposing the 61/39 D/L
glyceraldehyde mixture to water for only a few seconds raises
the solution ratio to 88/12.
The amino acids used are normal L-amino acids that can be
formed from the meteoritic alpha-methyl amino acids by heating
with simple components. Proline is not formed by transamination
of keto acids, as were used in our decarboxylative transaminations
using the meteoritic amino acids, and is unlikely to have been
present early on prebiotic earth. It can be formed by subsequent
nontrivial modifications of glutamic acid.
As the table shows, these reactions will lead to small or medi-
um excesses of the D sugars relative to the L sugars that are
formed from glyceraldehyde, but not to their dominance. For this
a process of amplification in water solution is needed, as was used
to amplify small excesses of L-amino acids to dominance in water.
We have now shown that this is possible.
We have also examined amplification recently (12, 13), focus-
ing on ribose, and found that ribose itself did not form a less so-
luble DL complex that would permit its amplification. However,
when the ribose was incorporated in the nucleosides uridine, ade-
nosine, or cytidine there were large differences in solubility be-
tween the D crystals and the less soluble DL crystals, so that very
small excesses of the D components were amplified to solutions
with as much as a 199/1 ratio of D to L. Thus again the solutions
would have allowed life to start with a preference for the domi-
nant D handedness in the components of RNA. After such am-
plification D-ribose itself could have been formed by hydrolysis of
the nucleosides.
Conclusions
D-Glyceraldehyde is a syrupy liquid at room temperature,
but DL-glyceraldehyde is a solid melting at 145 °C. As an x-ray
Of course, all such scenarios about how our prebiotic world could
have established dominant handedness in both amino acids and
sugars do not prove that this is the way it actually occurred.
H
catalyst
O
O
C
+
CH2
H2C OH
Formaldehyde
Table 1. Ratio of D to L glyceraldehydes synthesized from
glycolaldehyde and formaldehyde in the presence of amino acids
Glycolaldehyde
H
O
H
O
Amino acid
Ratio, D/L
C
C
L-serine
L-alanine
L-phenylalanine
L-valine
L-leucine
50.3/49.7
50.8/49.2
52.2/47.8
52.2/47.8
54.4/45.6
60.7/39.3
28.9/71.1
+
OH
H
HO
C
H
C
CH2OH
CH2OH
D-Glyceraldehyde
L-Glyceraldehyde
Fig. 2. The aldol reaction between glycolaldehyde and formaldehyde can
lead to both D-glyceraldehyde and L-glyceraldehyde, but with a preference
for the D isomer when an L-amino acid is the catalyst.
L-glutamic acid
L-proline
5724
∣
www.pnas.org/cgi/doi/10.1073/pnas.1001639107
Breslow and Cheng

dissolved in 17.0 mL deionized H2O. An amino acid (1.0 mmol) was added
to the resulting solution. The mixture was stirred at 20 °C for 3–5 days
(monitored by thin layer chromatography). Then 2,4-dinitrophenylhydrazine
(with 30% H2O, 11.6 g, 1.0 eq to aldehyde ) was added to the reaction
solution to form a mixture of dinitrophenyhydrazones. The glyceraldehyde
dinitrophenylhydrazone was isolated by flash chromatography on a silica gel
column, and shown to be pure by NMR spectroscopy (the spectrum was iden-
tical with that of the synthetically prepared pure compound). It was then sub-
jected to HPLC analysis and showed no significant impurity peaks.
However, the finding that the defining chiral center in D sugars
can be formed using L-amino acids as catalysts under really sim-
ple credible prebiotic conditions does simplify the requirements
for all proposals. It shows that the origin of the D sugars in life
today is not a mystery if indeed the small excesses of L alpha-
methyl amino acids were the seeds from space that started the
preference for the L isomers in normal amino acids. The L-amino
acids would then cause the preference for D sugars (see ref. 14 for
a different proposal and ref. 15 for a previous attempt at our gly-
ceraldehyde synthesis).
HPLC Analysis. HPLC was run on a Waters 600 HPLC system with a Waters 996
photodiode array spectrophotometer. The dinitrophenylhydrazones of D and
L glyceraldehyde were monitored by their absorbance at 350 nm. A Chiralpak
AD column was used for chiral analysis, with a mobile phase using gradient
hexane/EtOH. Under our conditions, the retention times for the dinitrophe-
nylhydrazones of L-glyceraldehyde and D-glyceraldehyde were 82 and
110 min, respectively.
Materials and Methods
Materials. Glycolaldehyde (dimer), glyceraldehyde, formaldehyde (37%
solution in water), dihydroxyacetone, 2,4-dinitrophenylhydrazine, and
amino acids were purchased from Aldrich. HPLC grade ethanol and hexanes
were obtained from Fisher Scientific.
Reaction and Derivatization Procedures. 60 mg of glycolaldehyde (1.0 mmol),
ACKNOWLEDGMENTS. This work was supported by the US National
and formaldehyde solution (3.0 mL, 37% H2O solution, 40.0 mmol) were
Science Foundation.
1. Cronin J, Pizzarello S (1997) Enantiomeric excesses in meteoritic amino acids. Science
275:951–955.
9. Pizzarello S, Weber A (2010) Stereoselective syntheses of pentose sugars under realistic
prebiotic conditions. Orig Life Evol B 40:3–10.
2. Pizzarello S (2006) The chemistry of life’s origin: A carbonaceous meteorite perspec-
tive. Acc Chem Res 39:231–235.
3. Levine M, Kenesky C, Mazori D, Breslow R (2008) Enantioselective synthesis and
enantiomeric amplification of amino acids under prebiotic conditions. Org Lett
10:2433–2436.
10. Breslow
21:22–26.
11. Senma M, Taira Z, Osaki K, Tagas
R (1959) On the mechanism of the formose reaction. Tetrahedron Lett
T
(1973) Glyceraldehyde: An x-ray study of
dl-glyceraldehyde dimer. J Chem Soc 880–881.
12. Breslow R, Cheng Z-L (2009) On the origin of terrestrial homochirality for nucleosides
and amino acids. Proc Natl Acad Sci USA 106:9144–9146.
13. Breslow R, Levine M, Cheng Z-L (2010) Imitating prebiotic homochirality on earth.
Origins Life Evol B 40:11–26.
4. Breslow R, Levine
M (2006) Amplification of enantiomeric concentrations under
credible prebiotic conditions. Proc Natl Acad Sci USA 103:12979–12980.
5. Klussman M, et al. (2006) Thermodynamic control of asymmetric amplification in
amino acid catalysis. Nature 441:621–623.
6. Hayashi Y, et al. (2006) Large nonlinear effect observed in the enantiomeric excess of
proline in solution and that in the solid state. Angew Chem 45:4593–4597.
14. Powner M, Gerland B, Sutherland
J (2009) Synthesis of activated pyrimidine
ribonucleotides in prebiotically plausible conditions. Nature 459:239–242.
7. Morowitz
mixtures. J Theor Biol 25:491–494.
8. Flores J, Bonner W, Massey (1977) Asymmetric photolysis of (RS)-leucine with
H (1969) Mechanism for the amplification of fluctuations in racemic
15. Weber A, Pizzarello S (2006) The peptide-catalyzed stereospecific synthesis of tetroses:
A
possible model for prebiotic molecular evolution. Proc Natl Acad Sci USA
G
103:12713–12717.
circularly polarized ultraviolet light. J Am Chem Soc 99:3622–3625.
Breslow and Cheng
PNAS
∣
March 30, 2010
∣
vol. 107
∣
no. 13
∣
5725