Molecular Biology, Vol. 34, No. 6, 2000, pp. 781–789. Translated from Molekulyarnaya Biologiya, Vol. 34, No. 6, 2000, pp. 921–930. Original Russian Text Copyright © 2000 by Kozlov, Orgel.
UDC 577.113
Over a period of many years, Professor Zoe A. Shabarova made major contributions to our understanding of the nonenzymatic template-directed ligation of short oligodeoxyribonucleotides [1–7]. The methods that she and her coworkers have developed are widely used and have still more potential applications in biotechnology and medicine. During this period, our laboratory at the Salk Institute followed a parallel course, using activated mononucleotides as substrates in place of oligomers. Past and present members of our group look back with pleasure on our many fruitful discussions with Professor Shabarova. In addition, Igor Kozlov, as one of Professor Shabarova’s former students, is indebted to her for mentorship and guidance. We would like to dedicate this paper to her memory.
Nonenzymatic Template-directed Synthesis of RNA from Monomers I. A. Kozlov1, 2 and L. E. Orgel1 1
2
The Salk Institute for Biological Studies, P.O. Box 85800, San Diego, CA, 92186 USA Currently at the Scripps Research Institute, Department of Chemistry, BCC 338, 10550 North Torrey Pines Road, La Jolla, CA, 92037 USA; E-mail:
[email protected] Received May 5, 2000
Abstract—Reviewed are the latest achievements in studying the information transfer mechanisms and the evolutionary significance of prebiotic RNA synthesis, the double helix structures most preferred in this respect, and the possible reasons for the prevalence of particular enantiomeric forms of nucleotides in template-directed synthesis. Key words: prebiotic RNA synthesis, activated mononucleotides, peptide nucleic acid, hexitol nucleic acid, altritol nucleic acid, wobble pairing, enantiomeric cross-inhibition
INTRODUCTION Nonenzymatic replication of RNA is of considerable interest with respect to the origins and development of the “RNA world,” an important intermediate stage in the origin of modern living organisms [8]. Substantial progress has been made over the last three decades in template-directed synthesis of RNA from activated mononucleotides. Work published before 1995 has been reviewed in detail [9, 10]. Here we review some of the progress in the field during the past five years. The most popular system for the nonenzymatic synthesis of RNA from mononucleotides uses nucleoside 5'-phosphoro-2-methyl imidazolides (Fig. 1) as substrates [11]. The standard reaction mixture that we have employed in most of our experiments contains 0.2 M 2,6-lutidine-HCl buffer (pH 7.9), 1.2 M NaCl,
and 0.2 M MgCl2. It has been reported that 2,6-lutidine-HCl can be replaced with Tris-HCl (pH 7.7) [12, 13] or HEPES (pH 8.0) [14]. In general, the preferred pH for the reaction is in the range 7–9 with an optimum of about 8. It has also been reported that NaCl can be omitted from the reaction mixture without significant loss in the oligomerization efficiency [12]. When G-rich templates are used, Na+ ions sometimes inhibit polymerization as they stabilize guanosine tetramers [12, 13]. The presence of Mg2+ ions is crucial, and the efficiency of reaction drops dramatically when Mg2+ ions are omitted. Two convenient experimental systems have been used extensively to study template-directed synthesis of oligonucleotides from monomers. The first system involves oligomerization of monomers on an oligomer, typically a decamer, template (Fig. 2a).
0026-8933/00/3406-0781$25.00 © 2000 MAIK “Nauka /Interperiodica”
782
KOZLOV AND ORGEL
Fig. 1. General structure of nucleoside 5'-phosphoro-2methyl imidazolides.
Again, the nature of most phosphodiester links was proved to be 3'–5' [15, 16]. The most informative experimental data are usually obtained after 5 days of incubation at 0°C. The reaction does not progress significantly after 14 days, because of substantial hydrolysis of the activated monomers. In some cases, a hairpin-based (Fig. 3a) oligomerization system [17–19] or an acridine-stabilized (Fig. 3b) system [12, 13] can be used in place of the simple system of an oligomer, a primer, and activated mononucleotides. In the last few years, similar methods have been used to study the template-directed kinetics of oligomerization of nucleoside 5'-phosphoryl-2-methyl imidazolides [14, 20–24].
A decamer template is long enough to support oligomerization, and the products of the reaction can be analyzed by HPLC on an anion-exchange RPC5 column. A typical concentration of template is about 0.5 mM, while each activated monomer is present at a concentration of about 100 mM. The nature of phosphodiester linkages in the majority of the reaction products was shown to be 3'–5' using enzymatic degradation [15]. In the second experimental system, a primer is extended on an appropriate template (Fig. 2b). The primer carries a 32P label, and the reaction products are separated using polyacrylamide gel electrophoresis. The concentration of the template is typically 20 µM, and the concentration of each activated monomer is 50 mM. Reactions are carried out at 0°C for times ranging from several hours to 14 days.
Recently, several attempts have been made to infer the structure of the nucleic acid double helix that favors nonenzymatic RNA synthesis [12, 13, 25]. It is well known that right-handed nucleic acid double helices in dilute aqueous solution adopt two welldefined structures, A and B [26]. RNA double helices and RNA–DNA hybrid helices usually adopt the A structure with the ribose ring in its 3'-endo conformation, while DNA double helices generally adopt the B structure with the ribose ring in its 2'-endo conformation. Since template-directed synthesis with 2-methylimidazole derivatives of nucleotides is remarkably substrate-specific, it is almost certain that it takes place efficiently in only one of these two very different structures. Recent reports strongly support the view that the 3'-endo conformation of the ribose ring in the A structure of the double helix is the pro-
O N
N P O CH3
O–
CH2 O B
HO OH 2MeImpB 2MelmpB
(a) G C
G C
G C
G C
G C
G C
(b) G C
G C
G C
G C
G C
G4CG4
G C G4CG3
G C G C
5
G4CG2
X' X
G6
G C
G7 G8 G9 G10
G4C
G C G C 32
p
G4
G C G+C
G4 G
G
G3
Fig. 2. (a) Schematic representation of the template-directed oligomerization of 2-MeImpG on a decacytidylate template (upper panel) and a typical HPLC profile obtained after analysis on an RPC5 column of a reaction mixture after 5 days of incubation (lower panel). Peaks indicated as G3−G10 correspond to all 3'–5'-linked oligo(G) products. (b) Left panel, schematic representation of the 32 pG4 primer extension reaction with 2-MeImpG and 2-MeImpX' on C4XC4 templates, where X = G, T, A, and X' is the complement of X. Right panel, an example of the results of typical experiments after 20% PAGE analysis. The reaction mixture contained a 32pG primer, an RNA template C GC , and either 2-MeImpG or an equimolar mixture of 2-MeImpG and 2-MeImpC. The reaction 4 4 4 was terminated after 5 days. The fastest-moving band in the diagram corresponds to the primer. MOLECULAR BIOLOGY
Vol. 34
No. 6
2000
NONENZYMATIC TEMPLATE-DIRECTED RNA SYNTHESIS
T C
(a)
T
(b)
C T A G rG X' X' X' X' X' G A T C C X X X X X T
G
p32
Acridine
C
783
G C A C G X' X' X' X' C G T G C X X X X
Fig. 3. Scheme of (a) hairpin and (b) acridine-stabilized oligomerization systems.
(a)
, (b)
f
H H
O O
B H
O H
–
O P O
–
O
OH
H
O
O H3C N
N
O
H O
OH B
Matrix vfnhbwf
–
H O
O
H O OH H3C P O N O N O
H
OH
P O
O
vfnhbwf
HO O
P
O
Matrix
B
–
OH B
B
HO
H OH B
H
H OH A-type ldjqyfz cgbhfkm double A-nbgfhelix
OH B-type ldjqyfz cgbhfkm double helix B-nbgf
Fig. 4. Primer interaction with activated monomer.
ductive conformation [12, 13, 25]. These data are supported by simple model building, which shows that the 3'-OH of the terminal nucleotides in a primer is positioned to attack an activated monomer in an A-form duplex. The 3'-OH in a B-form duplex is, however, directed toward the center of the duplex, away from the line of nucleophilic attack (Fig. 4). SOME RECENT RESULTS AND DISCUSSION (a) Copying of Oligocytidylate Templates Several studies describe the synthesis of RNA on templates related to nucleic acids but having modified backbones (Fig. 5). Peptide nucleic acids (PNAs) are nucleic acid analogs that instead of the normal phosphodiester backbone have an amide-bonded one [27] (Fig. 5). These oligomers bind tightly to complementary RNA and DNA sequences to form complexes similar to double-stranded nucleic acids [28, 29]. It is remarkable that synthesis of RNA on PNA templates takes place readily, although the reaction efficiency is lower than that on corresponding DNA templates [30]. These results provide support for the idea of the MOLECULAR BIOLOGY
Vol. 34
No. 6
2000
RNA world evolving from an earlier system that depended on a polymer containing the usual nucleic bases joined by a backbone simpler than that of RNA. We do not assert that PNA was necessarily that backbone. Hexitol nucleic acids (HNAs) are novel DNA analogs containing the standard nucleoside bases, but with a phosphorylated 1,5-anhydrohexitol backbone (Fig. 5). The six-membered hexitol ring mimics the furanose ring frozen in its 2'-exo, 3'-endo conformation [31]. Unlike previously described DNA or RNA analogs that have backbones based on a pyranose sugar ring [32, 33], HNA oligomers form duplexes with complementary DNA or RNA oligomers [34]. These duplexes have structures that closely resemble the structure of A-form DNA. Altritol nucleic acids (ANAs) are novel RNA analogs with a phosphorylated D-altritol backbone and a nucleobase at the 2-[S]-position of the carbohydrate residue (Fig. 5) [31]. They can be considered as HNA analogs that have an additional hydroxyl group intro-
784
KOZLOV AND ORGEL
O
O
NH B N
O P O CH2 O B O–
O P O CH2 O O–
B
O O
NH B N
O OH (H) O P O CH2 O B O–
O OH (H) O P O CH2 O O–
B
O O PNA GYR
O OH (H)
O OH (H)
RNA HYR(DNA) (LYR )
ANA FYR (HNA) (UYR )
Fig. 5. RNA and DNA analogs: (PNA) peptide nucleic acid, (HNA) hexitol nucleic acid, (ANA) altritol nucleic acid.
4 3
5 6 7 8 9
3
í 4
9 10 5 6
10
RNA
8 7
HNA
Fig. 6. Oligomerization of 2-MeImpG on RNA and HNA C10 templates after 14 days. The numbers above the peaks indicate the length of the all-3'–5'-linked oligo(G) products, T denotes the template. The C10 template was cleaved with RNase A before analysis. HPLC analyses of the reaction mixtures were performed on an RPC5 column.
duced into the six-membered hexitol ring. In a duplex, this group is directed into the minor groove and contributes to the duplex stability by increasing the hydration of the groove. The hydroxyl group may also help to preorganize a helical single-stranded structure that is optimal for the formation of the A-type double helix [35]. ANA–RNA and ANA–DNA duplexes are more stable than the corresponding HNA hybrids [35, 36]. The oligomerization of 2-MeImpG is slightly more efficient on the HNA and ANA templates than on corresponding RNA and DNA sequences (Fig. 6) and, after longer times, these templates give rise to significantly larger amounts of fully transcribed products [25, 37]. The slow and nonregiospecific addition of the last G residue on RNA and DNA templates has been attributed to the increasing instability of the template–substrate double helix as the 5' terminus of the template is approached [38]. Since HNA and ANA templates are preadapted to take the backbone conformation of A-DNA, they permit efficient chain elonga-
tion of the oligomeric products all the way to the 5' terminus of the template. Efficient oligonucleotide synthesis on PNA, HNA, and ANA oligomers, together with those reported in numerous other publications on nucleic acid analogs with modified backbones [32, 33, 39–42], emphasize an important question raised by Eschenmoser and his coworkers [39]. Why were RNA and DNA chosen as the genetic materials for all life on the Earth? Do the standard nucleic acids have some intrinsic advantage which we do not yet recognize, or is the choice of RNA a “frozen accident” that reflects the availability of β-ribonucleotides rather than their analogs at the dawn of the RNA world? (b) Information Transfer during Nonenzymatic Template-directed RNA Synthesis Information transfer in template-directed reactions is studied by copying oligomer templates containing MOLECULAR BIOLOGY
Vol. 34
No. 6
2000
d(C4TC4)
C4UC4
d(C4AC4)
C4AC4
NONENZYMATIC TEMPLATE-DIRECTED RNA SYNTHESIS
G9 G4AG4 G8 G4AG3 G7 G4AG2
G4UG4 G4UG3 G4UG2 G4UG G4U G4
G4 1 2 G+U (a)
3 4 G+A (b)
Fig. 7. Extension after 5 days of a 32pG4 primer on a C4AC4 RNA template (lane 1) or a d(C4AC4) DNA template (lane 2), a C4UC4 RNA template (lane 3), a d(C4TC4) DNA template (lane 4) in the presence of an equimolar mixture of 2-MeImpG and 2-MeImpU (lanes 1, 2) or an equimolar mixture of 2-MeImpG and 2-MeImpA (lanes 3, 4). The fastest-moving bands in the diagrams correspond to the primer.
two or more different bases (Fig. 2b). In the context of the above evidence that efficient synthesis on C10 templates occurs only in A-type helices, it is interesting to compare synthesis on hetero RNA and DNA templates C4XC4. Copying is, indeed, more efficient on RNA [61]; the largest effects are observed when the heterobase is A, for which copying is least efficient (Fig. 7a). However, the most surprising difference between RNA and DNA templates is the degree to which RNA supports copying by wobble pairing (Fig. 8). Incorporation of G instead of A is appreciable when DNA is used as template and A is absent from the reaction mixture, but this wobble pairing can compete effectively with normal U–A pairing only on RNA
(Fig. 7b). If wobble pairing is a general feature of nonenzymatic synthesis of RNA and not a special problem associated with the use of phosphoroimidazolides as substrates, it is hard to see how a genetic system using four bases could have evolved de novo. Instead, one would have to consider an initial system involving only one purine and one pyrimidine base, or postulate the existence in a pre-RNA world of enzyme-like catalysts that suppressed wobble pairing. It will be important to determine whether the difference in the extent of wobble pairing on DNA and RNA is entirely due to the differences in the backbones, or whether the methyl group on T helps suppress wobble pairing. Since synthesis on RNA was found to be more efficient than on DNA, previous quantitative estimates of the rate of RNA copying based on deoxynucleotidecontaining oligomers [17, 19, 43] are too low. However, the qualitative conclusions [17, 19, 43] hold. RNA sequences involving A residues are copied less efficiently than other templates, and consecutive A residues are found to be a complete barrier to copying RNA, just as they are for copying DNA. RNA sequences very rich in C and G form secondary structures similar to those formed by DNA sequences, so G,C-rich sequences will not be efficient templates [44]. Thus, although copying may be more efficient on RNA than on DNA templates, exponential replication, that is, repeated rounds of copying, is still not possible. Nonenzymatic information transfer has also been studied on analogs of the nucleic acids. Peptide nucleic acid (PNA) heterosequences can be copied, but the reaction is less efficient than for corresponding DNA or RNA templates [16]. HNAs and ANAs support efficient information transfer in nonenzymatic template-directed reactions [15, 37]. All HNA templates C4XC4, where X = A, T, C, facilitate the synthesis of its 3'–5'-linked complement, and in each case RNA synthesis on an HNA oligomer is more efficient than on the corresponding DNA oligomer and as efficient as that on the corresponding RNA oligomer. The properties of ANA templates are qualitatively similar to those of HNA templates, but ANA templates are
O H
N H
N N R
A N N
O
O
N H N U N N O R R
H N U N N H O R G N N H H
Fig. 8. Wobble pairing in RNA copying. MOLECULAR BIOLOGY
Vol. 34
No. 6
2000
785
786
KOZLOV AND ORGEL
N N R N N R
O
H N
H
H N
N H
O
I N N
H N C H N U A N N N N N O O RR R H H N H H N O O N N C H N U G N H D N N N N N N R N H O N H O R R H H Fig. 9. Pairing with inosine (I) and 2,6-diaminopurine (D).
quantitatively superior whenever differences can be detected. The difference is most marked for the least efficient steps in template-directed elongation, the addition of the last nucleotide in the oligomeric product on C4XC4 templates and the incorporation of U opposite A on C4AC4 templates. In previously published works [15, 16, 19, 45] the incorporation of U opposite A on a template has proved to be significantly less efficient than the corresponding incorporations opposite U or G. It is striking that copying proceeds past an A residue in ANA as efficiently as past the other bases [37]. Nucleic acids occupy a central position in biochemistry, so studies of template-directed synthesis have been focused on the oligomerization of activated derivatives of the standard nucleotides, U, A, C, and G, although a few experiments with 2,6-diaminopurine nucleotide [46–48] and other nucleotide analogs [48, 49] have been reported. Oligomerization reactions involving inosine-5'-phosphate (I) or 2,6-diaminopurine nucleotide (D), although having limited application to biochemistry, are of considerable interest for prebiotic chemistry, since putatively prebiotic syntheses that yield adenine and guanine typically also lead to the formation of hypoxanthine and 2,6-diaminopurine [50]. It seemed possible that studies of template-directed chemistry might help explain why A and G were chosen as major components of RNA and DNA while I and D were excluded (Fig. 9). The replacement of the A–U base pair by the D–U pair improves the efficiency of nonenzymatic template-directed oligomerization reactions [48]. In some cases D–U pairs support reactions that are comparable in efficiency to those involving G–C pairs. The difference between A–U and D–U pairs is probably attributable to the presence of three hydrogen bonds in the D–U pair, compared with only two hydrogen bonds in the A–U pair. In this respect, the D–U pair resembles a G–C pair, both having three hydrogen bonds.
The substitution of a G–C pair by an I–C pairs leads to a very large decrease in oligomerization efficiency [48]. The I–C pair is also much less efficient than A–U in facilitating oligomerization, although both are held together by two hydrogen bonds. The I−C pair must adopt a conformation that inhibits template-directed synthesis. The published data suggest that inhibition is strongest at the stage when a primer ending in a C residue opposite I in the template needs to be extended by the addition of a G residue opposite C in the template. The structural basis of this inhibition is unknown. Attempts to oligomerize 2-MeImpU and 2-MeImpC on D10 and I10 templates were unsuccessful, as were analogous attempts at primer extension with oligodeoxyribonucleotide hairpins [48]. The results described above confirm that the G–C base pair is exceptional in providing the necessary conformation for efficient nonenzymatic RNA synthesis using 2-methyl imidazolides of nucleoside 5'phosphates as substrates. Substitution of I for G in this reaction leads to poor incorporation of C and negligible extension of the resulting primer terminated by C. It is possible that a different activated derivative of C might polymerize efficiently, but present results suggest that the reason for the absence of I from replicating nucleic acids is the conformation of double helices containing I–C pairs. A similar explanation cannot account for the exclusion of D from nucleic acids. The replacement of A by D would lead to more efficient synthesis, so the choice of A rather than D is likely to reflect factors other than the efficiency of replication. The availability in the prebiotic environment is one possibility. Alternatively, optimization rather than maximization of the stability of double-helical RNA may have led to the selection of A rather than D. MOLECULAR BIOLOGY
Vol. 34
No. 6
2000
NONENZYMATIC TEMPLATE-DIRECTED RNA SYNTHESIS
3
DNA (RNA)
4
5 6
78
3
9
í HNA (ANA)
3
787
4 5
í
5
4
6
8 7
5
4
9
6 7
8 9
D-monomers
D- and L-monomers
-G D-G D-G D-G D-G D-G D-G D-G D-G D-G
D
C C C C C C C C C C
D-products D-template
-G
D
-G
L
-G
L
-G
L
-G
D
-G
L
-G
D
-G
D
-G
D
-G
L
C C C C C C C C C C
-G D-G D-G D-G D-G D-G D-G D-G D-G D-G
D
C C C C C C C C C C
Fig. 10. Enantiomeric cross-inhibition in template directed RNA synthesis. Oligomerization of D-2-MeImpG and D,L-2-MeImpG on different templates after 5 days. The arrangement of D and L residues on the template illustrated in this figure is a typical example chosen from 210 possibilities, all of which are in dynamic equilibrium. The numbers above the peaks indicate the length of the 3'– 5'-linked oligo(D-G) products, T indicates the template peak. HPLC analyses of the reaction mixtures were performed on an RPC5 column.
(c) Enantiomeric Cross-Inhibition in Template-directed RNA Synthesis All living organisms use the D-enantiomers of ribo- and deoxynucleotides as building blocks for their nucleic acids. Any abiotic synthesis of a nucleotide would yield equal amounts of the D- and L-isomers, so it is unclear why only D-nucleotides were chosen during evolution [51–55]. There are several publications that deal with the effect of L-nucleotides on the template-directed reactions of D-nucleotides [37, 56–59]. The effect of L-guanosine 5'-phosphoro-2-methylimidazole (L-2-MeImpG) on the nonenzymatic oligomerization of the D-enantiomer on a poly(D-C) template has been studied in some detail because of its relevance to prebiotic chemistry [56]. It was shown that the L-enantiomer is a potent inhibitor of the oligomerization, and is incorporated as a chain terminator in short oligo(D-G) products. A similar result was MOLECULAR BIOLOGY
Vol. 34
No. 6
2000
obtained when the poly(D-C) template was replaced by an achiral PNA C10 template [59]. Enantiomeric cross-inhibition was attributed to the ability of L-guanosine to pair with a cytosine in a template with normal Watson–Crick geometry and then to adopt a syn conformation, which would bring the 5'-phosphate of the activated L-monomer close to the 3'-hydroxyl of the growing oligo(D-G) strand. However, the suggestion that a syn conformation is present [56] may need to be modified, since it has been shown that a (DC):(L-G) base pair, within an otherwise B-helical structure, forms a stable Watson–Crick-type basepaired structure in which L-G has an S-type sugar geometry and a low anti glycosyl conformation [60]. It was found recently that the enantiomeric crossinhibition is much less severe on HNA and ANA C10 templates than on DNA and RNA C10 templates [37, 58]. The most striking differences between the products are seen for G6 and longer oligomers. These
788
KOZLOV AND ORGEL
are formed in greater yield and with much higher regiospecificity when an HNA or an ANA template replaces a standard nucleic acid template (Fig. 10). The regiospecific formation of an n-mer requires n – 1 successive regiospecific reaction steps, so it is surprising all-D oligomers as long as the 6- and 7-mer are formed in substantial yield with only very small amount of side-products. Enantiomeric cross-inhibition must be due to the ability of L-2-MeImpG to compete with the D-enantiomer for the binding site on the C residue of the template adjacent to the 3' terminus of the growing oligo(G) chain. If L-2-MeImpG, after binding, is unable to form a covalent bond to the growing oligo(G) chain it will behave as a competitive inhibitor, but if it does form a covalent bond it will terminate the chain irreversibly. The increased enantio-selectivity of the oligomerization at –10°C as compared with 0°C [58] may indicate that L-2-MeImpG is excluded from the “active site” by competition with D-2-MeImpG at low temperature. Alternatively, L-2-MeImpG may occupy the “active site” but, at low temperature, may be confined to a conformation that prevents covalent bond formation. The observed strong temperature dependence supports the latter interpretation. The finding that HNA and ANA templates can select the D-isomer from an enantiomeric mixture is not directly relevant to prebiotic chemistry, since prebiotic synthesis of the hexitol or altritol nucleotides is hardly plausible. Nevertheless, it shows that efficient selection of one optical isomer from a racemic mixture of nucleotides during nonenzymatic oligonucleotide synthesis on an analog of a nucleic acid template is possible. These results suggest that there may be chiral, possibly prebiotic, nucleic acid-like polymers that can replicate without significant enantiomeric cross-inhibition. A study of nonenzymatic D-RNA synthesis on predominantly D-oligodeoxynucleotide templates containing one or several L-deoxynucleotides has been reported [57]. It was shown that when one L-dC or two consecutive L-dCs are introduced into a D-template, regiospecific synthesis of 3'-5' oligo(G)s proceeds to the end of the template, but three consecutive L-dCs block synthesis. Alternating D-,L-oligomers do not facilitate oligomerization of D-, L-, or racemic 2MeImpG. To explain the copying of L-residues, it was suggested that D-2-MeImpG interacts specifically with an L-dC residue in the template in such a way as to permit elongation of an all-D primer. It seems almost certain that base pairing between L-dC and D-2MeImpG occurs in the normal way with the formation of three hydrogen bonds. Then, since the syn conformation of pyrimidine nucleotides is unattainable, the most plausible way of bringing the 5'-phosphate of D2-MeImpG into contact with the 3'-OH of the primer
is to hold the D-2-MeImpG in the syn configuration. However, it is also possible that D-G forms a Watson– Crick pair with L-dC in the manner suggested by Urata et al. [60]. It is striking, and possibly relevant to the origin of the RNA world, that primer elongation with D2-MeImpG continues, albeit at a somewhat slower rate, past one or two L-dC residues in the template. If the RNA world had been the first organized biological world, as it has often been proposed, it must have arisen in an environment containing racemic nucleotides. These results suggest that once a “predominantly D-metabolism” was in place, a small proportion of L-monomers in the template or the substrate would not lead to termination of replication. Nonenzymatic replication may be more resistant to poisoning by the incorrect enantiomer than previously supposed. CONCLUSION Early work on sequence-specific oligomerization of activated mononucleotides was largely restricted to the use of oligodeoxynucleotide templates. It is now clear that synthesis on RNA is more efficient than on DNA, and that the stability of the A-form is a major factor in determining the efficiency of templatedirected synthesis on oligonucleotide analogs. The realization that RNA and DNA are members of a much larger class of polymers that support complementary, nonenzymatic oligomerization is an important finding in the context of prebiotic chemistry. REFERENCES 1. Shabarova, Z.A., Nucleosides Nucleotides, 1998, vol. 17, pp. 2063–2072. 2. Shabarova, Z.A., Biochimie, 1988, vol. 70, pp. 1323– 1334. 3. Dolinnaya, N.G., Sokolova, N.I., Gryaznova, O.I., and Shabarova, Z.A., Nucleic Acids Res., 1988, vol. 16, pp. 3721–3738. 4. Dolinnaya, N.G., Tsytovich, A.V., Sergeev, V.N., Oretskaya, T.S., and Shabarova, Z.A., Nucleic Acids Res., 1991, vol. 19, pp. 3073–3080. 5. Dolinnaya, N.G., Blumenfeld, M., Merenkova, I.N., Oretskaya, T.S., Krynetskaya, N.F., Ivanovskaya, M.G., Vasseur, M., and Shabarova, Z.A., Nucleic Acids Res., 1993, vol. 21, pp. 5403–5407. 6. Dolinnaya, N.G., Merenkova, I.N., and Shabarova, Z.A., Nucleosides Nucleotides, 1994, vol. 13, pp. 2169–2183. 7. Sokolova, N.I., Ashirbekova, D.T., Dolinnaya, N.G., and Shabarova, Z.A., FEBS Lett., 1988, vol. 232, pp. 153– 155. 8. Gesteland, R.F., RNA World, Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press, 1999. 9. Joyce, G.F., Cold Spring Harbor Symp. Quant. Biol., vol. LII, 1987 pp. 41–51. MOLECULAR BIOLOGY
Vol. 34
No. 6
2000
NONENZYMATIC TEMPLATE-DIRECTED RNA SYNTHESIS 10. Orgel, L.E., Accts. Chem. Res., 1995, vol. 28, pp. 109– 118. 11. Inoue, T. and Orgel, L.E., J. Am. Chem. Soc., 1981, vol. 103, pp. 7666–7667. 12. Kurz, M., Gobel, K., Hartel, C., and Gobel, M.W., Helv. Chim. Acta, 1998, vol. 81, pp. 1156–1180. 13. Kurz, M., Gobel, K., Hartel, C., and Gobel, M.W., Angew. Chem. Int. Ed. Engl., 1997, vol. 36, pp. 842–845. 14. Kanavarioti, A., J. Org. Chem., 1998, vol. 63, pp. 6830– 6838. 15. Kozlov, I.A., De Bouvere, B., Van Aerschot, A., Herdewijn, P., and Orgel, L.E., J. Am. Chem. Soc., 1999, vol. 121, pp. 5856–5859. 16. Schmidt, J.G., Nielsen, P.E., and Orgel, L.E., Nucleic Acids Res., 1997, vol. 25, pp. 4797–4802. 17. Wu,T. and Orgel, L.E., J. Am. Chem. Soc., 1992, vol. 114, pp. 317–322. 18. Wu, T. and Orgel, L.E., J. Am. Chem. Soc., 1992, vol. 114, pp. 5496–5501. 19. Wu, T. and Orgel, L.E., J. Am. Chem. Soc., 1992, vol. 114, pp. 7964–7969. 20. Kanavarioti, A. and White, D.H., Orig. Life Evol. Biosphere, 1987, vol. 17, pp. 333–349. 21. Kanavarioti, A. and Bernasconi, C.F., J. Mol. Evol., 1990, vol. 31, pp. 470–477. 22. Kanavarioti, A., Bernasconi, C.F., and Baird, E.E., J. Am. Chem. Soc., 1998, vol. 120, pp. 8575–8581. 23. Kanavarioti, A., Baird, E.E., Hurley, T.B., Carruthers, J.A., and Gangopadhyay, S., J. Org. Chem., 1999, vol. 64, pp. 8323–8333. 24. Kanavarioti, A. and Gangopadhyay, S., J. Org. Chem., 1999, vol. 64, pp. 7957–7964. 25. Kozlov, I.A., Politis, P.K., Van Aerschot, A., Busson, R., Herdewijn, P., and Orgel, L.E., J. Am. Chem. Soc., 1999, vol. 121, pp. 2033–2036. 26. Saenger, W., Principles of Nucleic Acid Structure, New York: Springer–Verlag, 1984. 27. Hyrup, B. and Nielsen, P.E., Bioorg. Med. Chem., 1996, vol. 4, pp. 5–23. 28. Egholm, M., Buchardt, O., Christensen, L., Behrens, C., Freier, S.M., Driver, D.A., Berg, R.H., Kirn, S.K., Norden, B., and Nielsen, P.E., Nature, 1993, vol. 365, pp. 566–568. 29. Eriksson, M. and Nielsen, P.E., Q. Rev. Biophys., 1996, vol. 29, pp. 369–394. 30. Böhler, C., Nielsen, P.E., and Orgel, L.E., Nature, 1995, vol. 376, pp. 578–581. 31. Winter, H.D., Lescrinier, E., Van Aerschot, A., and Herdewijn, P., J. Am. Chem. Soc., 1998, vol. 120, pp. 5381–5394. 32. Eschenmoser, A., Orig. Life Evol. Biosphere, 1994, vol. 24, pp. 389–423. 33. Eschenmoser, A., Orig. Life Evol. Biosphere, 1997, vol. 27, pp. 535–553. 34. Hendrix, C., Rosemeyer, H., Verheggen, I., Seela, F., Van Aerschot, A., and Herdewijn, P., Chemistry, Eur. J., 1997, vol. 3, pp. 110–120. 35. Allart, B., Busson, R., Rozenski, J., van Aerschot, A., and Herdewijn, P., Tetrahedron, 1999, vol. 55, pp. 6527– 6546. MOLECULAR BIOLOGY
Vol. 34
No. 6
2000
789
36. Allart, B. et al., Chemistry, Eur. J., 1999, vol. 5, pp. 2424–2431. 37. Kozlov, I.A., Zielinski, M., Allart, B., Kerremans, L., Van Aerschot, A., Busson, R., Herdewijn, P., and Orgel, L.E., Chemistry, Eur. J., 2000, vol. 6, pp. 151– 155. 38. Chen, C.-H.B., Inoue, T., and Orgel, L.E., J. Mol. Biol., 1985, vol. 181, pp. 271–279. 39. Beier, M., Reck, F., Wagner, T., Krishnamurthy, R., and Eschenmoser, A., Science, 1999, vol. 283, pp. 699–703. 40. De Mesmaeker, A., Altmann, K.H., Waldner, A., and Wendeborn, S., Curr. Opin. Struct. Biol., 1995, vol. 5, pp. 343–355. 41. Kuwahara, M., Arimitsu, M., and Sisido, M., J. Am. Chem. Soc., 1999, vol. 121, pp. 256–257. 42. Arya, D.P. and Bruice, T.C., J. Am. Chem. Soc., 1998, vol. 120, pp. 12419–12427. 43. Hill, A.R., Jr., Orgel, L.E., and Wu, T., Orig. Life Evol. Biosphere, 1993, vol. 23, pp. 285–290. 44. Joyce, G.F. and Orgel, L.E., J. Mol. Biol., 1986, vol. 188, pp. 433–441. 45. Joyce, G.F. and Orgel, L.E., J. Mol. Biol., 1988, vol. 202, pp. 677–681. 46. Grzeskowiak, K., Webb, T.R., and Orgel, L.E., J. Mol. Evol., 1984, vol. 21, pp. 81–83. 47. Webb, T.R. and Orgel, L.E., Nucleic Acids Res., 1982, vol. 10, pp. 4413–4422. 48. Kozlov, I.A. and Orgel, L.E., Helv. Chim. Acta, 1999, vol. 82, pp. 1799–1805. 49. Rembold, H., Robins, R.K., Seela, F., and Orgel, L.E., J. Mol. Evol., 1994, vol. 38. P. 211–214. 50. Miller, S.L. and Orgel, L.E., Englewood Cliffs, NJ: Prentice-Hall, 1974. 51. Mason, S.F., Nature, 1984, vol. 311, pp. 19–23. 52. Orgel, L.E., Trends Biochem. Sci., 1998, vol. 23, pp. 491–495. 53. Schwartz, A.W., Curr. Biol., 1997, vol. 7, pp. R477– R479. 54. Schwartz, A.W., J. Theor. Biol., 1997, vol. 187, pp. 523– 527. 55. Schwartz, A.W., Molecular Origins of Life, Brack, A., Ed., Cambridge, UK: Cambridge University Press, 1998, pp. 237–254. 56. Joyce, G.F., Visser, G.M., van Boeckel, C.A.A., van Boom. J.H., Orgel, L.E., and van Westrenen, J., Nature, 1984, vol. 310, pp. 602–604. 57. Kozlov, I.A., Pitsch, S., and Orgel, L.E., Proc. Natl. Acad. Sci. USA, 1998, vol. 95, pp. 13448–13452. 58. Kozlov, I.A., Politis, P.K., Pitsch, S., Herdewijn, P., and Orgel, L.E., J. Am. Chem. Soc., 1999, vol. 121, pp. 1108–1109. 59. Schmidt, J.G., Nielsen, P.E., and Orgel, L.E., J. Am. Chem. Soc., 1997, vol. 119, pp. 1494–1495. 60. Urata, H., Ueda, Y., Suhara, H., Nishioka, E., and Akagi, M., J. Am. Chem. Soc., 1993, vol. 115, pp. 9852– 9853. 61. Zielinski, J., Kozlov, I.A., and Orgel, L.E., Helv. Chem. Acta, 2000 (in press).
