Amino Acid PDF

Subscriber access provided by LIBRARY OF CHINESE ACAD SCI
Perspective
Evidence for the Likely Origin of Homochirality in Amino
Acids, Sugars, and Nucleosides on Prebiotic Earth
Ronald Breslow
J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/ja3012897 • Publication Date (Web): 25 Mar 2012
Downloaded from http://pubs.acs.org on April 12, 2012
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of the American Chemical Society is published by the American Chemical

Society. 1155 Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.
Page 1 of 15 Journal of the American Chemical Society
1
2
3
4
Evidence for the Likely Origin of Homochirality in Amino Acids, Sugars, and 
5 Nucleosides on Prebiotic Earth 
6  
7 Ronald Breslow 
8 Department of Chemistry, Columbia University 
9
10
New York NY 10024 
11 rb33@columbia.edu 
12  
13 Abstract.    
14 Over the past century the origin of terrestrial prebiotic homochirality has been the 
15
16
subject of many speculations.  For life to start on earth and elsewhere, it is critical 
17 that the building blocks of amino acids, sugars, and nucleosides be created in 
18 predominant homochiral form.  Recent findings of a modest excess L chirality of α‐
19 methylamino acids in some meteorites that landed on earth have furnished an 
20
important piece of evidence.  We have shown how these meteoritic components can 
21
22 furnish normal L amino acids, and therefrom D sugars and nucleosides, in high 
23 chiral excess under sensible prebiotic conditions.  Some important remaining goals 
24 are also described. 
25  
26
27
  Introduction. 
28 Currently  we  are  not  surprised  that  L  amino  acids  and  D  sugars  are  produced  in 
29 biological systems, since the enzymes that produce them are themselves homochiral 
30 (not  a  mixture  with  their  mirror  images.)    On  prebiotic  earth  no  such  chirally 
31 selective catalysts were there to make the first amino acids or sugars or nucleosides, 
32
33
so  many  scientists  have  speculated  on  how  such  selectivity  could  have  arisen  in  a 
34 previously  achiral  world.    Some  have  invoked  chiral  faces  of  quartz,  some  have 
35 suggested  an  accident  that  was  then  promulgated,  but  the  question  has  had  no 
36 important new impetus until recently. In 1969 a carbonaceous chondritic meteorite 
37 landed in Murchison Australia carrying many organic compounds (see an extensive 
38
39
recent review by Pizzarello and Groy1 including the earliest work by Kvenholden et 
40 al.2  and  by  Cronin  and  Pizzarello3  cited  in  the  review.)  These  compounds  were 
41 apparently able to survive the frictional heating as the meteorite passed through our 
42 atmosphere since they were initially at ca. 10K, and chondritic meteorites are pieces 
43 of  rock,  with  low  thermal  conductivity,  from  the  asteroid  belts  that  surround  the 
44
45 sun.  When the meteorite was split open the interior was still cold enough to freeze 
46 water.   
47    
48 Among the compounds identified were the amino acids alanine, valine, aspartic acid, 
49 glutamic acid, proline, and leucine, which were racemic, with equal mixtures of the L 
50
51 and D forms, along with achiral glycine. However, five amino acids were found that 
52 had  methyl  groups  instead  of  hydrogens  on  their  alpha  positions  (Figure  1),  and 
53 these had a range of small excesses of the enantiomers originally described as the L 
54 amino acids (in modern terminology they are the S enantiomers).  Since that time, 
55
these and other α‐methyl amino acids with small excesses of the S enantiomer have 
56
57 been found in the Murchison, Murray, and Orgueil meteorites.1 
58
59
60
  1 
ACS Paragon Plus Environment
Journal of the American Chemical Society Page 2 of 15
1
2
3
4
            
5 + + +
H 2N H3N CO2- H3N CO2-
6 CO2-
7 Me Pr Me Me Et
8
9
10 e.e. 2.8% S e.e. 2.8% S e.e. 15.2% S
11
12 + +
H 3N H 3N CO2-
13 CO2-
14 Me Me Me
Me (3R)
15 (3S)
Et Et
16
17 e.e. 9.1% S
18
e.e. 7% S  
19  
20 Figure  1.    Some  α‐methyl  amino  acids  discovered  in  the  Murchison  and  other 
21 meteorites,  all  of  which  have  a  small  excess  of  the  S  configuration  that  was 
22 described  as  L.    The  enantioexcesses  showed  a  range  of  values  in  this  and  later 
23
24
work. 
25  
26 How do we know that they were not simply contaminants from earth, added after 
27 the  meteorites  landed?      There  are  three  arguments  against  this.    First  of  all,  the 
28 actual α‐methyl amino acids isolated from the Murchison meteorite are not found in 
29
30 our  biology  today,  but  the  meteorite  landed  here  only  41  years  ago.    Against  this, 
31 there are some α‐methyl amino acids produced by microorganisms4‐6 but not all the 
32 ones in the meteorites.    Secondly, enzymes do not produce compounds with only 
33 partial chiral selectivities; such partial chiralities are really symptomatic of species 
34
35
that  were  originally  racemic  but  have  then  been  partially  deracemized.    If 
36 microorganisms  had  invaded  the  meteorites  after  they  landed,  they  could  perhaps 
37 selectively  cause  destruction  of  the    D  enantiomers  of  the  amino  acids  if  they  had 
38 originally been racemic, but there is no evidence for such an unlikely process.  I will 
39 describe an alternative scenario for such partial deracemization below.  Finally, the 
40
41
α‐methyl  amino  acids  have  levels  of  13C  and  non‐exchangeable  deuterium  much 
42 higher  than  are  seen  in  molecules  formed  on  Earth.    These  high  levels  of  heavy 
43 isotopes  are  generally  seen  in  atoms  delivered  from  space,  where  isotopic 
44 fractionation is performed at very low temperatures.  They are generally accepted to 
45
be definitive proof that the species examined are not of terrestrial origin. 
46
47    
48 Microwave spectroscopy has detected hundreds of molecules in the interstellar gas 
49 clouds in space, and they include ammonia, hydrogen cyanide, hydronium ion, and 
50 various ketones and aldehydes and the acetylenes from which they can be derived.7 
51
Thus  the  meteoritic  unmethylated  and  methylated  amino  acids  were  probably 
52
53 formed  by  Strecker  reactions8  from  such  species.  The  reactants  are  aldehydes  for 
54 normal unmethylated amino acids and methyl ketones for the α‐methyl amino acids, 
55 along with HCN, NH3, and H3O+.  Methyl ketones are the products from reaction of 
56 terminal acetylenes with hydronium ion.  The multimolecular reactions would occur 
57
58 on  solid  particles  whose  larger  versions  are  asteroids,  and  they  would  be  initially 
59
60
  2 
1
2
3
4
formed  as  racemates.    They  could  develop  some  excess  of  the  L  forms  if  the 
5 racemates  absorbed  circularly  polarized  light  that  selectively  destroyed  one 
6 enantiomer, perhaps also using another process to amplify them.9  
7    
8 William  Bonner  had  shown  that  right  circularly  polarized  light  at  uv  wavelengths 
9
10
would  selectively  destroy  the  D  component  of  racemic  leucine,  leaving  a  small 
11 excess of the L amino acid.10  Astronomers have detected a small excess of circularly 
12 polarized  infrared  light  in  space,  whose  energy  is  too  low  to  deracemize  amino 
13 acids.11‐13 However, one author indicated that the same processes that produce the 
14 identified  circularly  polarized  light  in  the  infrared  might  also  produce  it  in  the 
15
16
ultraviolet.12  Very  high  energy  light  cannot  penetrate  our  atmosphere  for 
17 observation as infrared can, and could not penetrate it when our early atmosphere 
18 consisted  of  nitrogen  and  CO2.    As  one  possibility,  favored  by  many  astronomers, 
19 circularly polarized light could be formed in the universe by cyclotron processes in 
20 neutron  stars,  just  as  experimental  cyclotrons  produce  circularly  polarized  light 
21
22 with  opposite  chirality  above  and  below  the  circulation  plane.    Some  other 
23 astronomers  prefer  that  circularly  polarized  light  could  be  produced  by  magnetic 
24 dwarf stars. One of the important challenges for astronomy is to observe outside our 
25 atmosphere and detect circularly polarized short‐wavelength light, if it is there.  For 
26 instance,  observations  could  be  made  from  orbiting  satellites,  from  the  Hubble 
27
28 telescope,  or  from  the  new  Web  Space  Telescope.    I  am  trying  to  get  such 
29 observations made.   
30    
31 If there was also (yet undetected) right circularly polarized light with energy in the 
32
uv or higher irradiating the asteroid belt when the amino acids were present on a 
33
34 particle that later came to Earth, this could account for the small excesses of the L 
35 enantiomers  seen  in  the  α‐methyl  amino  acids.    If  this  also  happened  with  the 
36 unmethylated  amino  acids  that  are  found  in  our  proteins  they  could  racemize  by 
37 reversible  loss  of  their  alpha  protons  over  time,  explaining  why  they  are  found  in 
38
39
racemic form, but no such racemization is possible with the α‐methyl amino acids. 
40 (Isoleucine has two chiral centers, and it was not completely racemic.   The α‐center 
41 would  be  equilibrated  by  reversible  loss  of  its  somewhat  acidic  proton  but  the  β‐
42 center  could  not  be  equilibrated,  so  isoleucine  could  not  be  racemized  by  half 
43
44 conversion to its mirror image.)     
45    
46 Thus an attractive idea is that the L amino acids and the D sugars that are now the 
47 basis of life were first “seeded” by the arrival on Earth of α‐methyl amino acids in 
48
49
meteorites,  formed  on  asteroid  fragments  as  racemates  and  then  partially 
50 deracemized by circularly polarized light of short wavelength.14 Light of the needed 
51 wavelength  could  not  penetrate  Earth’s  atmosphere—now  or  in  the  past—so  the 
52 deracemization occurred in space and the result was then delivered to Earth.  Such 
53 homochirality  was  probably  important  since  mixture  of  enantiomers  would  not 
54
55
form  well‐defined  structures  on  random  incorporation  into  polypeptides  or 
56 polysaccharides  or  polynucleotides.    In  fact,  it  seems  likely  that  homochirality  is 
57 generally necessary for the origin of life anywhere, so without the kind of processes 
58
59
60
  3 
1
2
3
4
that  started  it  on  earth  other  planets  would  not  produce  life.    On  Earth  the  amino 
5 acid  chirality  was  what  we  call  L,  but  on  other  planets  the  mirror  image  systems 
6 could occur by the processes we describe.   
7    
8 We  have  recently  shown  how  such  α‐methyl  amino  acids  can  generate  normal 
9
10 unmethylated  biological  amino  acids  with  some  chirality  transfer,  and  how  that 
11 excess  chirality  can  be  amplified  under  credible  prebiotic  conditions,  producing 
12 aqueous solutions with very high enantioexcesses.  We have also shown how the D 
13 sugars and D ribonucleosides could have arisen on prebiotic Earth.  
14
 
15
16   Results and Discussion 
17
18 We  addressed  two  questions.    1)  What  credible  prebiotic  chemistry  could  use  the 
19 small  excess  S  (L)  chirality  of  the α‐methyl  amino  acids  in  a  meteorite  to  seed  the 
20
21
formation  of  L  amino  acids  without  such  methyl  groups?  2)  How  could  credible 
22 prebiotic processes amplify small enantioexcesses to available high enantiopurity in 
23 solution?  The ideal is of course 100%, but enantioexcesses of 90% or better could 
24 probably be enough for primitive biology to select the dominant enantiomer. 
25    
26
27
  Formation of nonmethylated amino acids with some L excess.   
28 We  devised  a  process  of  decarboxylative  transamination:  α‐methyl  amino  acids 
29 reacted  with  α‐keto  acids  to  form  unmethylated  amino  acids  on  simple  heating  in 
30 the  presence  of  Cu(II)  (Figure  2).15    This  was  based  on  earlier  work  we  had  done 
31
32
showing  that  α‐methyl  amino  acids  could  perform  such  decarboxylative 
33 transaminations with pyridoxal.16 
34   
35 We  observed  some  chiral  transfer,  in  which  the  reaction  of  100%  e.e.  S‐α‐
36
37
methylvaline  with  sodium  phenylpyruvate  led  to  a  37%  enantioexcess  of  L‐
38 phenylalanine,  and  the  same  process  with  sodium  pyruvate  led  to  a  20% 
39 enantioexcess  of  L‐alanine.    The  reaction  of  sodium  dimethylpyruvate  with  S‐α‐
40 methylisoleucine  afforded  L‐valine  with  17%  transfer  of  chirality.    All  these 
41
processes  were  successful  only  in  the  presence  of  catalytic  Cu(II),  a  component  of 
42
43 some  meteorites  and  surely  present  on  prebiotic  earth.    Without  it,  or  with  Zn(II) 
44 instead, the reactions were much less successful, and the correct chiral transfer was 
45 not observed.  
46  
47
48
49
50
51
52
53
54
55
56
57
58
59
60
  4 
1
2
3 R
4 R CO2 -
+ - H2O
5 H 3N N - CO2
6 O- CO2- CO2-
O +
7 Me Me Cu(II)
O + H2O
8
9
10 R R H
CO2 - + H+ CO2 -
11 H2O R CO2 -
N N +
12
H NH3 +
13 Me Me O Me
14  
15  
16
17
18 Figure 2.  Transaminative decarboxylation of the imine from S‐α‐methylvaline and 
19 an α ketoacid.  In the presence of Cu(II), which catalyzes the reaction, the product 
20 amino acid has an excess of the L enantiomer, but not without the Cu(II).  This is 
21
22
expected if a square planar complex is formed between two imines and the Cu(II), so  
23 after decarboxylation of one ligand it is protonated selectively, guided by the 
24 stereochemistry of the other ligand.  
25  
26 The  process  was  credible  under  prebiotic  conditions,  involving  only  heating  the 
27
28
components  in  the  presence  of  some  catalytic  Cu(II)  ion.    As  we  described,15  DFT 
29 calculations indicated that a square planar Cu(II) complex of two imines that were 
30 formed  from  the  amino  acid  and  the  keto  acid  would  decarboxylate  one  of  them, 
31 which would then protonate to form the L amino acid steered by the chirality of the 
32 second  ligand,  as  we  observed.    However,  the  few  percent  of  S  excess  in  the 
33
34 meteoritic amino acids would produce only a 1% or so enantioexcess in the product 
35 amino acids.  Amplification of the resulting chiral excess would be critical to achieve 
36 a  dominance  of  the  L  amino  acids  large  enough—of  the  order  of  a  90/10  ratio  or 
37 better  of  L  to  D—that  new  organisms  or  pre‐organisms  would  select  them  for 
38 evolutionary success. 
39
40   Amplification of the Lamino acid excesses in solution.  
41 In  1969  Morowitz  had  proposed  a  process  by  which  a  small  enantioexcess  of  an 
42 amino  acid  could  be  amplified  to  form  solutions  with  large  enantioexcesses  under 
43 simple prebiotic conditions.17 This was an amplification of the concentration of the 
44 excess component in solution by concentrating it and removing the racemate, not by 
45
46 making  more  of  it.    We  conceived  the  same  idea  in  2006,18  not  aware  of  his  work, 
47 and Blackmond also published a version of the concept in 2006 slightly before our 
48 publication.19‐21   All our ideas were based on the general fact that most amino acids 
49 form racemic compound crystals that are less soluble, and higher melting, than the 
50
crystals of the L or the D enantiomers.  In such racemic crystals neighboring L and D 
51
52 molecules interact to lower the free energy—either by better crystal packing or by 
53 direct  stabilizing  interactions  in  the  crystals—compared  with  crystals  where  an  L 
54 molecule has only other L neighbors. 
55     
56
The  idea  is  that  a  mixture  of  D  and  L  amino  acids  with  some  excess  of  the  L 
57
58 component (or elsewhere in the universe perhaps the D component) would dissolve 
59
60
  5 
1
2
3
4
in  water,  and  as  the  water  evaporates  the  less  soluble  racemate  would  precipitate 
5 leaving  a  solution  with  increased  richness  in  the  L  component.    Because  the 
6 precipitation  of  the  racemate  depends  on  the  solubility  product  SP(DL)  =  [D][L], 
7 while that of the homochiral L compound depends only on the solubility [L], there 
8 would  be  feedback  to  increase  the  precipitation  of  the  racemate  as  the 
9
10
concentration of L increases.  Thus starting with a small excess of the L amino acid, 
11 the  final  L/D  ratio  would  become  very  large  with  even  a  modest  difference  in 
12 solubilities.  The Morowitz treatment17 is shown below. 
13  
14 S(L) = [L] 
15
16
SP(DL) = [D][L] 
17 [D] = SP(DL)/[L] 
18 [L]/[D] = S(L)2/SP(DL) 
19  
20 The  solubility  product  SP(DL)  is  equal  to  the  square  of  half  the  solubility  of  the 
21
22 racemate.  By this equation, if the homochiral crystals were only twice as soluble as 
23 were the racemates the final ratio [L]/[D] would be 16/1, a solution with 94% of the 
24 L  and  6%  of  the  D.  If  the  homochiral  crystals  were  three‐fold  as  soluble  as  the 
25 racemate the final L/D ratio would be 36/1. As we will describe later, this treatment 
26 is  not  quite  correct  and  can  overestimate  the  selectivity  for  reasons  we  have 
27
28 demonstrated. 
29   
30 The previous theories and experiments involved solutions at equilibria, but kinetics 
31 can often afford different results.  With equilibrium solubilities we saw L tryptophan 
32
amplified  to  94.5%  L  and  5.5%  D,  starting  with  any  arbitrarily  small  excess  (the 
33
34 treatment is true for any initial ratio, limited only by the practical need to achieve 
35 saturation in both the racemate and homochiral compound, so smaller amounts of 
36 water would be needed with smaller initial L excess).15  However, we found that this 
37 ratio  was  raised  to  99/1  in  solution  when  we  poured  a  small  amount  of  water 
38
through  the  final    dry  mixture,  imitating  the  effect  of  rainwater.15  Homochiral  and 
39
40 racemic crystals differ in their activation energies for subtracting (or adding) units, 
41 a form of dissociation (or binding), and in our case the homochiral crystals dissolve 
42 faster beyond the requirements of the equilibrium constants.   
43  
44
45
Forming D sugars under prebiotic conditions
46 We have investigated the formation of the simplest sugar belonging to the D series, 
47 D‐glyceraldehyde.  All other sugars in the D family—including D‐ribose, D‐glucose, 
48 D‐fructose,  D‐erythrose,  etc.—are  named  for  the  geometry  of  the  chiral  center 
49 furthest from the carbonyl group in the sugar, and this is the unique chiral center in 
50
51
D‐glyceraldehyde.    Sugars  were  probably  synthesized  on  prebiotic  Earth  by  a 
52 process  called  the  formose  reaction.22  D‐ribose  (C5H1005)  is  a  formal  pentamer  of 
53 formaldehyde  (CH2O),  whose  dimer  is  glycolaldehyde  and  whose  trimer  is 
54 glyceraldehyde.    In  the  formose reaction formaldehyde  is  treated with a  mild  base 
55 such as calcium hydroxide.  There is a period when nothing happens until there is 
56
57
the  sudden  conversion  of  the  formaldehyde  to  glycolaldehyde  and  glyceraldehyde 
58
59
60
  6 
1
2
3
4
and  even  larger  sugars.  We  proposed  and  demonstrated  the  mechanism  of  the 
5 formose reaction (Figure 3) by an autocatalytic cycle many years ago.23 
 
6
7 slow HO
8 2 H 2C O H2C CH=O
9 H2C O
10
11
12 HO H
13 2 H2C CH=O HO CH=O
14
15 CH2OH
16
17
18 HOH
19
HOHC CH=O
20
CH2OH O
21
22 HOH2C CH2OH
23
24
25 O
H2C O
26 HOHC CH2OH
27
CH2OH  
28
29  
30 Figure 3.  The formose reaction, producing sugars from formaldehyde. 
31  
32
33
A first molecule of glycolaldehyde is formed from formaldehyde by a slow unknown 
34 process,  possibly  involving  ionization  by  cosmic  rays,  and  this  then  reacts  with  an 
35 additional formaldehyde to form glyceraldehyde.   This is then in equilibrium with 
36 dihydroxyacetone by enolization and ketonization, and this then reacts with another 
37 formaldehyde  to  form  a  four‐carbon  ketosugar.    By  enolization  and  aldehyde 
38
39
formation  a  four‐carbon  aldehyde  is  formed  that  can  undergo  a  reverse  aldol 
40 reaction to form two molecules of glycolaldehyde.   After this, four glycolaldehydes 
41 result from another turn of the cycle, etc.   
42  
43 We  examined  the  formation  of  glyceraldehyde,  a  three‐carbon  sugar  that  is  the 
44
45
simplest  one  with  a  chiral  center.24  Glyceraldehyde  is  formed  in  the  formose  cycle 
46 by the reaction of formaldehyde with glycolaldehyde.   Without a chiral catalyst  this 
47 reaction would form both D‐glyceraldehyde and its enantiomer L‐glyceraldehyde in 
48 equal amounts, but we studied what would happen if the reaction were catalyzed by 
49 a chiral amino acid.  We can trace the formation of L amino acids back to the small 
50
51 excesses of S α‐methyl amino acids in the Murchison meteorite, so the question was 
52 simple.    Would  L  amino  acids  catalyze  the  formation  of  D‐glyceraldehyde 
53 preferentially, at least to a small extent, and if so could the preference be amplified 
54 to  a  sufficient  excess  that  the  dominant  D‐glyceraldehyde  would  be  selected  by 
55
56
incipient  life?    If  so,  all  of  the  D  sugars—built  on  the  D‐glyceraldehyde  by  adding 
57 pieces to its aldehyde group that do not change the geometry of the D center in the 
58
59
60
  7 
1
2
3
4
glyceraldehyde original unit—are part of the general origin of homochirality tracing 
5 back to the meteorites.  If not, a new source of homochirality will need to be found 
6 for the sugars.  As Table 1 shows, we examined various L‐amino acids from different 
7 classes. 
8  
9
10
 
11 Table 1.  Ratio of D to L glyceraldehyde 
12  from glycolaldehyde and formaldehyde  
13 catalyzed by various L‐amino acids.   
14 The reaction conditions and analytical  
15
16
method are described in ref. 24. 
17 _________________________________ 
18 Serine (50.3/49.7)  Alanine (50.8/49.2) 
19 Phenylalanine (52.2/47.8) Valine (52.2/47.8) 
20 Glutamic acid (60.7/39.3) Proline (28.9/71.1)
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
  8 
1
2
3
4
 
5  
6 The  reaction  of  formaldehyde  with  glycolaldehyde  is  an  aldol  reaction,  and  many 
7 such  reactions  are  known  to  be  catalyzed  by  amines.    They  would  react  with  the 
8 glycolaldehyde to form an enamine, and this adds to the other carbonyl component.  
9
10
Thus  we  carried  out  the  reaction  of  glycolaldehyde  with  formaldehyde  in  water  in 
11 the presence of a variety of L amino acids.  We found (Table 1) that all the L amino 
12 acids  caused  the  formation  of  glycolaldehyde  with  a  small  excess  of  the  D 
13 enantiomer, with one exception:  L‐proline catalyzed the formation of an excess of L‐
14 glyceraldehyde.  (Some recent work by Blackmond25 indicates that in basic solution 
15
16
this proline preference is reversed when the carboxyl group is a carboxylate ion, so 
17 proline may not be an exception after all.) 
18  
19 The  preferences  were  modest.  There  was  a  60/40  ratio  of  D/L  glyceraldehyde 
20 catalyzed by L‐glutamic acid, for instance, and smaller ratios with the other L‐amino 
21
22 acids.    Proline  was  probably  not  present  in  large  amounts  among  the  early  amino 
23 acids, since it is formed in two secondary reactions from glutamic acid, so it would 
24 not  dominate  the  chiral  selectivity.    The  results  with  the  other  amino  acids  did 
25 support  the  idea  that  the  preference  for  D‐glyceraldehyde—and  then  the  other  D 
26 sugars derived from it—was simply the result of the formation of the other L amino 
27
28 acids on prebiotic Earth. 
29    
30 We then asked whether there was a likely process for amplifying the small excess of 
31 D‐glyceraldehyde  formed  in  this  way,  preferably  by  using  the  selective  solubilities 
32
that  had  worked  with  the  amino  acids.    This  turned  out  to  be  possible..    D‐
33
34 Glyceraldehyde  is  a  syrup  with  essentially  complete  water  solubility,  but  racemic 
35 DL‐glyceraldehde  is  a  solid  with  a  melting  point  of  145°C  and  a  limited  water 
36 solubility.  The striking difference has an explanation.  A previously published X‐ray 
37 structure determination showed that DL‐glyceraldehyde exists as a chair‐form six‐
38
membered ring dioxane dimer of one D and one L molecule, with all the substituents 
39
40 equatorial  as  in  chair  cyclohexane,  so  this  flat  dimer  molecule  packs  well  into  a 
41 crystal.26    With  the  same  structure  based  on  a  dimer  with  D‐glyceraldehyde  alone, 
42 one large group in the six‐membered DD ring would be axial, making the dimer less 
43 stable and also making crystal packing less favorable.  The result was that we could 
44
45
take the 60/40 ratio of D/L glyceraldehyde formed by catalysis with glutamic acid 
46 and  turn  it  into  a  92/8  D/L  ratio  in  water  solution  by  slow  evaporation  of  water, 
47 causing the racemic crystals to precipitate. 
48   
49 The formation and amplification of Dribonucleosides.   
50
51
An  aldol  condensation  of  D‐glyceraldehyde  wth  glycolaldehyde  would  lead  to  D‐
52 ribose.    We  are  studying  the  selective  catalysis  of  this  reaction,  in  which  two  new 
53 chiral centers are produced.  In the mean time, we examined the possibility that any 
54 excess  of  D  over  L  in  ribose  could  be  amplified  by  selective  solubilities,  as  in  the 
55 amino acid and glyceraldehyde cases. In contrast to the situation with amino acids, 
56
57
we  saw  that  D  ribose  and  DL  ribose  had  the  same  melting  point.27    Ribose 
58 apparently  forms  a  racemate  that  is  a  solid  solution,  with  essentially  identical 
59
60
  9 
1
2
3
4
properties, solubilities, and melting points at all D/L ratios.  In the solids a D‐ribose 
5 molecule  can  equally  well  have  either  a  D  or  an  L  as  its  neighbor.    We  did  not 
6 examine  their  solubility,  since  in  all  cases  that  we  know  of  a  lower  solubility 
7 correlated  with  a  higher  melting  point—both  melting  and  dissolving  break  up  the 
8 crystals.  For  this  reason,  D‐ribose  cannot  be  amplified  by  the  selective  solubility 
9
10
method.  Thus we examined the ribonucleosides (Figure 5).27 
11   
12 It  had  been  shown  previously  that  ribose  reacts  with  purines  under  prebiotic 
13 conditions to form ribonucleosides.28   In the first work pyrophosphate was present 
14 to activate the ribose, but in later work it was shown that the pyrophosphate could 
15
16
be replaced by the residue from dried seawater.  We are working on extending these 
17 results  to  pyrimidine  nucleosides  (see  another  proposal29  on  how  pyrimidine 
18 nucleosides  could  have  been  formed),  and  in  the  meantime  we  examined  whether 
19 ribonucleosides  with  small  excesses  of  the  D  form  could  be  amplified  by  the 
20 solubility method. 
21
22   
23 NH2 O
24 N NH
N
25
26 N N N O
27 HO HO
28 O O
29 H H H H
30 H H H H
OH OH OH OH
31
32
Adenosine Uracil
33
34
35
36 NH2 O
37
N N NH
38
39
N O N N NH2
40 HO
HO
41 O
O
42 H H H H
43 H H H H
OH OH OH OH
44
45
46 Cytidine Guanosine
47
48
49
 
50  
51 Figure 4.  The nucleosides that were examined. 
52  
53 We  synthesized  L‐uridine  by  a  literature  method  and  made  a  1:1  mixture  with  D‐
54
55
uridine.  We  then  crystallized  the  racemate  to  purity  and  examined  its  properties.  
56 The crystals of uridine racemate had a melting point of 176°C, higher by 21°C than 
57 the  155°C  for  D‐uridine.    D‐Uridine  had  a  water  solubility  at  22°C  of  454  mg/mL, 
58
59
60
  10 
1
2
3
4
while the racemic uridine had a solubility of only 87 mg/mL, 5.2 times smaller.  By 
5 the Morowitz calculation this should afford a D/L ratio of 108/1 at saturation with 
6 both crystals.  We then dissolved both the D‐uridine and the DL‐uridine crystals to 
7 saturation together in water at 22°C, filtering away the excess, and saw 96% D and 
8 4% L in solution, a ratio of 24/1.  The large ratio makes the D‐uridine dominant, but 
9
10
it is less than predicted by the Morowitz theoretical treatment. 
11  
12 We  then  examined  the  situation  with  adenosine,  synthesizing  L‐adenosine  and 
13 mixing  it  with  D‐adenosine,  then  recrystallizing  the  racemic  crystals  to  purity.    In 
14 this case the melting point of the racemate was 243 ± 1°C, while for D‐adenosine the 
15
16
m.p. was 230 ± 1°C.  The solubility of D‐adenosine in water at 22°C was 5.2 mg/mL, 
17 6.3  times  as  large  as  the  0.8  mg/mL  for  the  DL  crystal.  There  should  have  been  a 
18 160/1  D/L  ratio  at  saturation  equilibrium  from  the  Morowitz  treatment,  but  we 
19 observed  a  99/1  ratio  of  D/L  adenosine.    Again  very  large,  but  again  less  than 
20 predicted  by  the  Morowitz  treatment.      With  cytidine  we  saw  decomposition  on 
21
22 melting of both the D and the DL crystals, so no melting points could be determined, 
23 but  the  solubilities  predicted  an  even  larger  amplification  than  for  the  other  two 
24 nucleosides.  D‐cytidine  had  a  solubility  of  192  mg/mL  in  22°C  water,  while  DL‐
25 cytidine  had  a  solubility  of  24.3  mg/mL.    The  Morowitz  equation  predicts  D/L  of 
26 250/1 at saturation, while we observed 199/1.  
27
28   
29 All three of these nucleosides can be amplified to very high D/L ratios by selective 
30 solubilities,  but  the  Morowitz  treatment  overestimated  the  ratios.    We  concluded 
31 that  this  could  reflect  the  fact  that  the  solubilities  being  measured  were  in  pure 
32
water,  but  the  solubilities  relevant  to  the  amplification  experiments  are  in  water 
33
34 along  with  a  second  component.    In  the  experiment,  the  solubility  of  the  racemate 
35 that is relevant is that in the presence of the homochiral component that is also in 
36 solution.    We  concluded  that  the  homochiral  dissolved  material  was  acting  as  a 
37 cosolvent,  increasing  the  solubility  of  the  racemate  over  that  in  pure  water.14    The 
38
dissolved homochiral component has two roles.  In one it helps drive the racemate 
39
40 out of solution by its role in the solubility product of the racemate, but in the other 
41 role  it  increases  the  solubility  of  the  racemate  by  acting  as  an  antihydrophobic 
42 cosolvent  like  ethanol,  with  which  relatively  hydrophobic  substrates  such  as 
43 nucleosides have greater solubility than in pure water.  If this is true the solubility of 
44
45
the homochiral crystals would also be increased by the presence of the racemate in 
46 solution,  but  at  final  saturation—with  little  dissolved  racemate—this  effect  would 
47 be smaller.  
48    
49 To  test  this  idea,  we  examined  the  solubility  of  racemic  uridine  in  water  with  D‐
50
51
cytidine added to saturation.14  This can mimic the cosolvent effect of D‐uridine, but 
52 does  not  play  a  role  in  the  solubility  product  of  DL‐uridine.    We  found  that  the  D‐
53 cytidine increased the solubility of DL‐uridine by 40% over that in pure water. Thus 
54 the Morowitz treatment overestimates the amplifications a bit, because it does not 
55 include  the  cosolvent  effect.    The  experimental  values  are  still  so  high  that  they 
56
57
could lead to the dominance of the D‐nucleosides on prebiotic Earth, even starting 
58 from a very small initial excess. 
59
60
  11 
1
2
3
4
 
5 Guanosine  behaved  differently.27    D‐guanosine  and  the  DL‐guanosine  crystals 
6 melted  with  decomposition,  but  we  could  still  examine  their  solubilities  in  22°C 
7 water.    DL‐Guanosine  had  a  solubility  of  0.84  mg/mL  while  D‐guanosine  had  a 
8 solubility of 0.46 mg/mL, essentially half that of the racemate.   This indicated that 
9
10
the racemate belongs to the third class of racemic crystals.  It is not a solid solution 
11 as  we  saw  with  ribose,  nor  a  racemic  compound  crystal  like  those  of  uridine, 
12 adenosine and cytidine.  It is a racemic conglomerate of D and L crystals, each with 
13 its own solubility.  Such a conglomerate is like that with racemic sodium ammonium 
14 tartrate, that allowed Pasteur to separate the D and L crystals by hand. 
15
16
   
17 This  inversion  in solubilities,  with  the  racemate more soluble  than  the homochiral 
18 compound, makes it impossible to amplify D‐guanosine by selective solubilities, but 
19 of  course  the  D‐ribose  from  hydrolysis  of  the  other  three  nucleosides  could  have 
20 been  used  to  synthesize  D‐guanosine  on  prebiotic  Earth.    The  D‐ribose  resulting 
21
22 from hydrolysis of the other three nucleosides could also play additional important 
23 biochemical roles, including conversion to D‐ribulose whose diphosphate is the key 
24 sugar in photosynthesis.
25
26
27  Conclusion.  This work30, 31 answers some of the questions in the general idea that 
28 the unusual amino acids delivered to Earth by the Murchison meteorite and related 
29 ones could have led to the dominance of L amino acids and D sugars on early Earth 
30 that would permit life to start.  Of course showing that it could have happened this 
31
32 way is not the same as showing that it did.  Proper theories need the possibility of 
33 falsification. This account would be in trouble if significant examples were found in 
34 which meteorites that landed on Earth contained R α‐methyl amino acids instead of 
35 the  S  examples  in  the  Murchison  meteorite  and  others  examined  so  far.      An 
36
alternative  scenario  to  ours  would  use  the  Murchison  α‐methyl  amino  acids  to 
37
38 catalyze the formation of some D sugars, as Pizzarello and Weber have shown32 (cf. 
39 refs.  (33)  and  (34)  for  related  work),  and  then  use  those  sugars  to  catalyze  the 
40 formation  of  the  normal  proteinogenic  L  amino  acids.    Good  credibly‐prebiotic 
41 examples of the latter process have not yet been produced.  
42
43
44 Further  work  is  needed  to  show  prebiotic  versions  of  the  conversion  of  D‐
45 glyceraldehyde  to  D‐ribose,  D‐glucose,  D‐fructose,  and  D‐2‐deoxyribose,  efforts 
46 underway  in  our  lab.    We  also  need  a  credible  way  in  which  nucleosides  and 
47 nucleotides  could  be  formed  from  these  sugars  and  pyrimidines.,  not  just  purines  
48 Finally,  of  course,  we  all  need  ways  in  which  these  and  other  sensible  building 
49
50 blocks could assemble into structures with the exciting properties of life. 
51   
52  An implication from this work is that elsewhere in the universe there could be life 
53 forms  based  on  D  amino  acids  and  L  sugars,  depending  on  the  chirality  of  circular 
54 polarized light in that sector of the universe or whatever other process operated to 
55
56 favor the L α‐methyl amino acids in the meteorites that have landed on Earth.  Such 
57 life  forms  could  well  be  advanced  versions  of  dinosaurs,  if  mammals  did  not  have 
58
59
60
  12 
1
2
3
4
the  good  fortune  to  have  the  dinosaurs  wiped  out  by  an  asteroidal  collision,  as  on 
5 Earth.  We would be better off not meeting them. 
6
7
Acknowledgements 
8
9
10 I thank my coworkers whose names appear in the references, and the U. S. National 
11 Science Foundation for support of this work.  
12
13  
14  
15
16
 
17 References 
18 1. Pizzarello, S.; Groy, T.  Geochim. Cosmochim. 2011, 75, 645‐656. 
19 2. Kvenvolden, K.; Lawless, J.; Pering, K.; Peterson, E.; Flores, J.; Ponnamperuma, C; 
20 Kaplan, I.; Moore, C.  Nature 1970, 228, 923‐926.  
21
22
3. Cronin, J.; Pizzarello, S.  Science 1997, 275, 951‐955. 
23 4. Degenkolb, T.; Gams, W.; Bruckner, H.  Chem. & Biodiversity 2008, 5, 693‐706.  
24 5. Degenkolb, T.; Bruckner, H.  Chem. & Biodiversity 2008, 5, 1817‐43.  
25 6. Bruckner, H.; Becker, D.; Gams, W.; Degenkolb, T.  Chem. & Biodiversity 2009, 6, 
26 38‐56. 
27
28
7.   Herbst, E.; van Dishoek, E. Ann Rev. Phys. Astron. & Chem. 2009, 47, 427‐480. 
29 8. Strecker, A.  Annal. Chem. Pharm. 1850, 75, 27‐45. 
30 9. Glavin, D.; Dworkin, J.  Proc. Natl. Acad. Sci. USA 2009, 106, 5487–5492. 
31 10. Flores, J.; Bonner, W.; Massey, G.  J. Am. Chem. Soc. 1977, 99, 3622‐3625. 
32 11. Bailey, J.; Chrysostomou, A.; Hough, J.; Gledhill, T.; McCall, A.; Clark, S.; Menard, F.; 
33
34 Tamura, M.  Science 1998, 281, 672‐674. 
35 12. Bailey, J.  Orig. Life Evol. Biosph. 2001, 31, 167‐183. 
36 13. Buschermohle, M.; Whittet, D.; Chrysostomou, A.; Hough, J.; Adamson, A.; 
37 Whitney, B.; Wolff, M.  Astrophys. J. 2005, 624, 821‐826. 
38 14. Breslow, R.; Levine, M.; Cheng, Z.  Orig. Life Evol. Biosph. 2010, 40, 11‐26. 
39
40 15. Levine, M.; Kenesky, C.; Mazori, D.; Breslow, R.  Org. Lett. 2008, 10, 2433‐2436. 
41 16. Chruma, J.; Liu, L.; Zhou, W.; Breslow, R.  Bioorg. Med. Chem. 2005, 13, 5873‐
42 5883.  
43 17. Morowitz, H.  J. Theor. Biol. 1969, 25, 491. 
44
18. Breslow, R.; Levine, M.  Proc. Natl. Acad. Sci. USA 2006, 103, 12979‐12980. 
45
46 19. Klussmann, M.; Iwamura, H.; Mathew, S.; Wells, D.; Pandya, U.; Armstrong, A.; 
47 Blackmond, D.  Nature 2006, 441, 621. 
48 20. Klussmann, M.; Izumi, T.; White, A.; Armstrong, A.; Blackmond, D.  J. Am. Chem. 
49 Soc. 2007, 129, 7657‐7660. 
50
21. Noorduin, W.; Izumi, T.; Millemaggi, A.; Leeman, M.; Meekes, H.; Van Enckevort, 
51
52 W.; Kellogg, R.; Kaptein, B.; Vlieg, E.; Blackmond, D.  J. Am. Chem. Soc. 2008, 130, 
53 1158‐1159. 
54 22. Langenbeck, W.  Tetrahedron 1958, 3, 185‐196. 
55 23. Breslow, R.  Tetrahedron Lett. 1959, 1 (21), 22‐26. 
56
57
24. Breslow, R.; Cheng, Z.  Proc. Natl. Acad. Sci. USA 2010, 107, 5723‐5725.   
58 25. Hein, J.; Blackmond, D. G. Accts. Chem. Res. 2012, 30, xxxx.  
59
60
  13 
1
2
3
4
26. Senma, M.; Taira, Z.; Osaki, K.; Taga, T.  J. Chem. Soc. Chem. Commun. 1973, 880‐
5 881. 
6 27. Breslow, R.; Cheng, Z.  Proc. Natl. Acad. Sci. USA 2009, 106, 9144‐9146. 
7 28. Fuller, W. D.; Sanchez, R. A.; Orgel, L. E. J. Mol. Evoln. 1972, 1, 249‐257. W. D.  
8 29. Powner, M. W.; Gerland, B.; Sutherland, J. D. Nature 2009, 459, 239–242. 
9
10
30. Breslow, R.  Tetrahedron Lett. 2011, 52, 2028‐2032. 
11 31. Breslow, R.  Israel J. Chem., 2011, 51, 1‐7. 
12 32. Pizzarello, S.; Weber, A.  Orig. Life Evol. Biosph. 2010, 40, 3‐10. 
13 33. Burroughs, L.; Vale, M.; Gilks, J.; Hayes, H.; Christopher, J.; Clarke, P.  Chem.Comm. 
14 2010, 46, 4776‐4778. 
15
16
34. Fernandez‐Lopez, R.; Kofoed, J.; Machuqueiro, M.; Darbre, T.  Eur. J. Org. Chem. 
17 2005, 24, 5268‐5276. 
18  
19 Figure captions 
20  
21
22 Figure  1.    Some  α‐methyl  amino  acids  discovered  in  the  Murchison  and  other 
23 meteorites,  all  of  which  have  a  small  excess  of  the  S  configuration  that  was 
24 described  as  L.    The  enantioexcesses  showed  a  range  of  values  in  this  and  later 
25 work. 
26
27
 
28 Figure 2.  Transaminative decarboxylation of the imine from S‐α‐methylvaline and 
29 an α ketoacid.  In the presence of Cu(II), which catalyzes the reaction, the product 
30 amino acid has an excess of the L enantiomer, but not without the Cu(II).  This is 
31
32
expected if a square planar complex is formed between two imines and the Cu(II), so  
33 after decarboxylation of one ligand it is protonated selectively, guided by the 
34 stereochemistry of the other ligand.  
35  
36 Figure 3.  The formose reaction, producing sugars from formaldehyde. 
37
38
 
39  
40 Figure 4.  The nucleosides that were examined
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
  14 
1
2
3
4
Graphic Abstract 
5  
6  
7  
8  
9
10
 
11  
12  
13  
14  
15
16
 
17  
18  
19  
20  
21
22  
23  
24  
25  
26  
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
  15 

Amino Acid PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Amino Acid PDF

Uploaded by

Copyright:

Available Formats

Subscriber access provided by LIBRARY OF CHINESE ACAD SCI

Journal of the American Chemical Society is published by the American Chemical

You might also like