Primitive Genetic Polymers - CSHL Pcshperspectives.cshlp.org/content/2/12/a002196.full.pdf · mer known as peptide nucleic acid (Nelson et al. 2000), and intermediate proposals, in

Primitive Genetic Polymers

Aaron E. Engelhart and Nicholas V. Hud

School of Chemistry and Biochemistry, Georgia Institute of Technology, Atlanta, Georgia 30332

Correspondence: [email protected]

Since the structure of DNA was elucidated more than 50 years ago, Watson-Crick basepairing has been widely speculated to be the likely mode of both information storage andtransfer in the earliest genetic polymers. The discovery of catalytic RNA molecules sub-sequently provided support for the hypothesis that RNA was perhaps even the first polymerof life. However, the de novo synthesis of RNA using only plausible prebiotic chemistryhas proven difficult, to say the least. Experimental investigations, made possible by the appli-cation of synthetic and physical organic chemistry, have now provided evidence that thenucleobases (A, G, C, and T/U), the trifunctional moiety ([deoxy]ribose), and the linkagechemistry (phosphate esters) of contemporary nucleic acids may be optimally suited fortheir present roles—a situation that suggests refinement by evolution. Here, we considerstudies of variations in these three distinct components of nucleic acids with regard to thequestion: Is RNA, as is generally acknowledged of DNA, the product of evolution? If so,what chemical and structural features might have been more likely and advantageous for aproto-RNA?

In contemporary life, nucleic acids provide theamino acid sequence information required

for protein synthesis, while protein enzymescarry out the catalysis required for nucleic acidsynthesis. This mutual dependence has beendescribed as a “chicken-or-the-egg” dilemmaconcerning which came first. However, requir-ing that these biopolymers appeared strictlysequentially may be an overly restrictive precon-ception—nucleic acids and noncoded peptidesmay have arisen independently and only laterbecome dependent on each other. Nevertheless,the requirements for the chemical emergence oflife would appear simplified if one polymer wasinitially able to store and transfer information as

well as perform selective chemical catalysis—two essential features of life.

The discovery of catalytic RNA molecules inthe early 1980s (Kruger et al. 1982; Guerrier-Takada et al. 1983) created widespread interestin an earlier proposal (Woese 1967; Crick 1968;Orgel 1968) that nucleic acids were the first bio-polymers of life, as nucleic acids transmitgenetic information and could have once beenresponsible for catalyzing a wide range of reac-tions. The ever-increasing list of processes thatinvolve RNA in contemporary life continuesto strengthen this view (Mandal and Breaker2004; Gesteland and Atkins 2006). Further-more, the rule-based one-to-one pairing of

Editors: David Deamer and Jack W. Szostak

Additional Perspectives on The Origins of Life available at www.cshperspectives.org

Copyright # 2010 Cold Spring Harbor Laboratory Press; all rights reserved; doi: 10.1101/cshperspect.a002196

Cite this article as Cold Spring Harb Perspect Biol 2010;2:a002196

1

on June 29, 2020 - Published by Cold Spring Harbor Laboratory Press http://cshperspectives.cshlp.org/Downloaded from

http://cshperspectives.cshlp.org/

complementary bases in a Watson-Crick duplex(Fig. 1) provides a robust mechanism for infor-mation transfer during replication that couldhave been operative from the advent of oligonu-cleotides. In contrast, there is no obvious andgeneral mechanism by which the amino acid se-quence of a polypeptide can be transferred to anew polypeptide as part of a replication process.

If we accept that nucleic acids must haveappeared without the aid of coded proteins,we are still faced with the question of how thefirst nucleic acid molecules came to be. Broadlydefined, there are two schools of thoughtregarding the origin of the earliest nucleic acids.In one school, it is proposed that abiotic chem-ical processes initially gave rise to nucleotides(i.e., phosphorylated nucleosides), which werethen coupled together to yield polymers identi-cal in chemical structure to contemporary RNA.In support of this model, Sutherland presents inhis article current progress toward discoveringpossible chemical pathways for the prebioticsynthesis of RNA mononucleotides, as well asmethods for their protein-free polymerization(Sutherland 2010).

A second school of thought, discussed inthis article, considers RNA to be a product ofevolution, and that a different RNA-like poly-mer (or proto-RNA) was used by the earliestforms of life. Just as the deoxyribose sugar of

DNA was likely the product of Darwinian evo-lution (selected for the hydrolytic stability itprovides this long-lived biopolymer), so, too,might the sugar, phosphate, and bases of RNAhave been refined by evolution. In this scenario,a proto-RNA is more likely to have spontane-ously formed than RNA, because a proto-RNAcould have had more favorable chemical charac-teristics (e.g., greater availability of precursorsand ease of assembly), but such a polymerwas eventually replaced, through evolution,by RNA (potentially after several incrementalchanges), based on functional characteristics(e.g. nucleoside stability, versatility in formingcatalytic structures). Thus, contemporary RNAmay possess chemical traits that, although op-timally suited for contemporary life, may havebeen ill-suited for the earliest biopolymers,with the converse being true for proto-RNA.

BACKGROUND

Challenges to “Reinventing” Proto-RNA

If proto-RNA came before RNA, possibly com-prised of different bases and a different back-bone, then the chemical space of proto-RNAcandidates seems almost limitless. Leslie Orgelonce called this scenario a “gloomy prospect”with regard to solving the origin of life (Orgel

O

OO

O

OO

OO

O

OOHO

O

H

H

HH

H

H

OHO

H

HGlycosidic

bond

Phosphoesterbond

Trifunctional moiety(ribose)

Nucleoside linker(phosphate)

Information unit(nucleobase)

OH

O

OOOP

O

–O

O

N

N

N

HO

NN

N

N

NA

U

N

N

NG

C

NN

NN

O

OP

P

P

O

O–

O–

–O

Figure 1. Two base-paired RNA dinucleotide steps with functional units discussed in the text annotated. Incontemporary life, the nucleoside linker is phosphate, and the information unit is one of the canonicalnucleobases (A, G, C, and U). The contemporary trifunctional moiety, ribose, is coupled via N,O-acetals tothe informational unit and via phosphoesters to the nucleoside linker.

A.E. Engelhart and N.V. Hud

2 Cite this article as Cold Spring Harb Perspect Biol 2010;2:a002196



1998). However, we do not see the prospect ofproto-RNAs as reason for pessimism. The pos-sibility that RNA had one or more predecessorsalso implies the possibility that there was a pol-ymer with more facile chemistry of assembly,which researchers could show to spontaneouslyassemble from simple precursors. Additionally,the chemical space within which proto-RNAwas formed is constrained by the molecules thatcould have been present on the prebiotic Earth.Although we will never know the precise con-tents of the prebiotic chemical inventory, modelprebiotic reactions continue to provide valuableinformation regarding plausible prebiotic mol-ecules (Miyakawa et al. 2002; Plankensteineret al. 2004; Saladino et al. 2004; Cleaves et al.2006). Additionally, studies of our neighboringplanets and their moons, meteorites, cometsand even interstellar space (Schwartz and Chang2002; Hollis et al. 2004; Pizzarello 2006; Martinset al. 2008) also provide clues to the types ofmolecules from which a proto-RNA mighthave emerged.

Sugars and nucleobases, two essential build-ing blocks of nucleic acids, have been identifiedas products in a rather wide range of model pre-biotic reactions. However, feasible prebioticreactions have not been shown for the efficientcoupling of ribose to all four nucleobases to cre-ate the canonical nucleosides, and only limitedevidence has been presented that preformednucleotides might be able to polymerize withoutresorting to nonprebiotic chemical activation.Thus, a conceptual gap still exists between ourideas for how the building blocks of RNA mighthave formed from the simple molecules thatare ubiquitous in the universe versus the chem-ical processes that gave rise to the first RNA-likepolymers.

All the present challenges to bridging thisconceptual gap between the “small moleculeworld” and the “polymer world” can be classi-fied as belonging to one of two major problems:molecular selection and polymerization reac-tions. There was likely a wide variety of mole-cules in the prebiotic chemical inventory, sothe first formidable challenge is to find an ab-iotic mechanism by which “useful” buildingblocks were selected from a complex mixture.

The second formidable challenge is to find reac-tions by which these building blocks could havebeen “correctly” joined to create “useful” (i.e.,self-replicating, catalytic or both) proto-biopoly-mers. In the following sections, we will elaborateon the challenge of RNA building block selectionand assembly, as well as possible solutions to“reinventing,” to use Albert Eschenmoser’s term(Eschenmoser 2007), a process for the de novosynthesis of an RNA-like polymer from plausibleprebiotic molecules and reactions.

Some Possible Constraints to Guide theSearch for Proto-RNA

Synthetic chemists have prepared numerousnucleic acid analogs, some as part of origin oflife studies (Schneider and Benner 1990; Pitschet al. 1993; Herdewijn 2001b; Benner 2004; Mit-tapalli et al. 2007a; Mittapalli et al. 2007b) andmany as part of a widespread search for thera-peutic agents (Kurreck 2003). These analogshave included changes to both the nucleobasesand to the polymer backbone. With regard toorigin of life investigations, some proposalsfor the difference between RNA and an earlierproto-RNA have been subtle, such as the pro-posal that hypoxanthine once functioned inthe place of guanine (these nucleobases differby only a single exocyclic amino group) (Crick1968). Other proposals are much more radical,such as the suggestion that the nucleobases wereonce connected by a completely different anduncharged backbone, as in the synthetic poly-mer known as peptide nucleic acid (Nelsonet al. 2000), and intermediate proposals, inwhich the backbone was a peptide, but stillnegatively charged (Mittapalli et al. 2007a;Mittapalli et al. 2007b) (Fig. 2). Physical andchemical studies of these and other nucleicacid analogs have provided invaluable informa-tion regarding which alternative polymersmight have predated RNA. An exhaustive reviewof all proposed predecessors of RNA is beyondthe scope of this article, and several excellentreviews have been published considering possi-ble variations in sugars and bases, to which thereader is referred (Benner 2004; Benner et al.2004; Eschenmoser 2007).


Cite this article as Cold Spring Harb Perspect Biol 2010;2:a002196 3



In this article, we focus on a subset of poten-tial candidates for proto-RNA polymers that, inour opinion, would be closest to RNA in struc-ture and physical properties but are composedof alternative building blocks that would havemade prebiotic assembly more feasible. Onecould consider these structures as candidatesfor the later predecessors of RNA that resultedfrom proto-RNA evolution or, optimistically,as potential candidates for the first proto-RNA,if only a few changes were required to reach

RNA. Our selection of proposed proto-RNAcandidates on which to focus is guided by threehypotheses:

(1) The original molecular components of proto-RNA may have been different from, butstill similar to, those found in RNA. Three dis-tinct molecular components comprise RNA:the nucleobases, ribose and phosphate(Fig. 1). Each component has specific rolesfor which it appears to be very well (perhaps

O

O

OO

O

O

O

O

B

B

B

P–O

O

OO

P–O

OA

E

H I

F G

BC D

OO O

B

OO

O

O

O

OH

OH

OH

O

P–O

O

OO

P–O

OB

B

B

OOP–O

O

OOP–O

O

OO O

BP–O

–O

B

B

B

O OB

O

O O

O

B

OOHP

O

O

OO

P–O

–O

–O

–O

–O

O

OO

O O

OO

O

O

OO

O

O

OP

P

H H

H B

B

B

B

OH

HO

NHB

HH

H

H

NH

H

H

RR

R R–O

O

O

B

OOHP

O

OB

O

N

OH

O

O O

OO

OHN

NHNH

O

OO

H H

H B

B*

HRR

B

B

N O

O NH

N O

O NH

–O

NHO

HN

NH

O

OO

B*

NHO

HN

O

O

B*

Figure 2. Structures of RNA and selected analogs discussed in the text. (A) RNA; (B) FNA or “flexible nucleicacid,” originally named glycerol nucleic acid; (C) GNA, glycerol nucleic acid; (D) TNA,a-threofuranosyl nucleicacid; (E) pRNA, b-pyranosyl-RNA; (F) homo-DNA (R ¼ H) and b-allopyranosyl (R ¼ OH); (G) gaNA,glyoxylate-linked dinucleotide; (H ) PNA, peptide nucleic acid; (I) poly-(L-Asp-L-Glu) with the g-carboxylfunction of Glu conjugated to a nucleobase by an isopeptide linkage. For all structures shown, B representsone of the canonical nucleobases, and B�, in structure I, represents a noncanonical base.





optimally) suited. However, because of thelimited availability of some components,or the difficulty in forming covalent bondsbetween others, there is good reason toentertain the possibility that alternativecomponents were initially used by proto-RNA. Each of these three components couldhave been replaced on different timescalesover the course of evolution. Although thiscriterion may seem obvious regarding whatwould be recognized as RNA-like polymers,it does exclude from our discussion themore radical models of “genetic takeover”(Cairns-Smith 1982), in which an inorganicor decidedly non-RNA-like system gave riseto RNA, models which come with the addedburden of deciphering how a physicochem-ical transition could have transpired betweencompletely different molecular systems.

(2) The covalent bonds that connected the threemolecular components of proto-RNA were(periodically) reversible. In the early stagesof life, it would have been highly advanta-geous for chemical building blocks to beavailable for reuse. Mechanisms for recy-cling are easy to imagine, if the covalentbonds that joined the molecular compo-nents of proto-biopolymers could havebeen formed, broken and re-formed repeat-edly. Such a process would have allowedproto-biopolymers to be created that werethermodynamically favored structures (Liet al. 2002; Hud et al. 2007; Ura et al.2009) and for “errors” in synthesis to becorrected (e.g., replacement of non- ormispaired nucleobases with pairing ones).Proto-RNA monomers might have evenbeen repeatedly recycled into polymerswith different nucleotide sequences (andcorresponding functions) as survival pres-sures changed (Ura et al. 2009). Further-more, without some regular cycle bywhich RNA or a proto-RNA could havebeen depolymerized, any monomer incor-porated into a polymer that was not func-tionally useful would be wasted–acondition under which there simply may

have not been enough material to get lifestarted.1

(3) Condensation-dehydration reactions were,from the beginning, integral to the formationof proto-RNA. A wide range of biologicalbond-forming reactions involve the loss ofwater, including amide linkages in pep-tides, acetal linkages in polysaccharidesand ester linkages in phosphoglycerols. Per-tinent to the present discussion are thereactions required for nucleic acid forma-tion—glycosylation of a nucleobase toform a nucleoside, phosphorylation of anucleoside to form a nucleotide and con-densation of nucleotides to form oligonu-cleotides—all of which are dehydrationreactions. Such reactions are very appealingfrom a prebiotic standpoint, as their equili-bria can be modulated via water activity(i.e., through drying-wetting cycles), sug-gesting that many polymerization reactionsobserved in contemporary life may haveancient roots in drying reactions driven byperiodically fluctuating water activity—perhaps even dating back to the earlieststages of life.

We now will discuss the three molecularcomponents of nucleic acid polymers and con-sider the implications of the three hypothesespresented earlier in considering alternative can-didates for proto-RNA.

RECENT RESULTS

Selection of the Nucleobases: Was it C, U, A,and G from the Start?

Although nucleic acid chemists have provid-ed some insights regarding what might have

1If the earliest effective ribozyme was 50nt in length (similarto the hammerhead ribozyme), and the earliest genetic codecontained C, U, A, and G, this ribozyme would be one of 450

� 1030 possible sequences. A collection of one molecule ofevery possible 50mer would require a carbon mass of around107 kg. Assuming each polymer was produced in equal yield,and by an irreversible process, production of 1 picomole ofthe ribozyme would require 1019 kg of carbon (roughly thetotal weight of carbon in the Earth’s crust).





predated RNA, as we hope we have conveyedherein, we concede that it is still not obviousby what mechanism the nucleobases were orig-inally selected. The ability to form Watson-Crick base pairs should not be assumed tohave been the sole criterion. One observationthat runs counter to this assumption is thatthe free nucleobases and their nucleosides donot form Watson-Crick base pairs in aqueoussolution (Ts’o et al. 1963), but rather, theyform co-planar stacks, with their Watson-Crickedges interacting with the solvent. The hydro-gen bonds formed with water are of comparableenthalpy to those that would be formedbetween paired bases. The local preorganizationof nucleobases in oligonucleotides, on the otherhand, promotes intra- and interstrand stackingof base pairs, which contributes substantial freeenergy to duplex stability, whereas Watson-Crickhydrogen bonds have been reported to contrib-ute virtually no net free energy (De Voe andTinoco 1962; Yakovchuk et al. 2006).

The observation that free nucleobases donot form Watson-Crick base pairs in aqueoussolution inspired us to articulate a “paradox ofbase pairing” with regard to the emergence ofRNA (Hud and Anet 2000; Hud et al. 2007).Why would the nucleobases have been selectedfor inclusion in RNA or proto-RNA polymers forthe purpose of Watson-Crick base pairing if theydid not form base pairs before being linked by acommon backbone? To argue that the nucleo-bases formed nucleotides with ribose and phos-phate before being involved in base pairingseems to imply that nature either predictedthat the nucleobases would be useful for pairingpolymers in the future (which would violate thenonpredictive property of evolution); thatnature was simply lucky, and a series of chemicalreactions coincidentally created rather complexmolecules (i.e., nucleotides) that happened toform base pairs after polymerization; or thatsuch molecules existed in a metabolism-firstorigin of life scenario, and these moleculeswere later incorporated into a genetic polymer.Serious conceptual challenges have been raisedagainst metabolism-first scenarios (Anet 2004;Orgel 2008; Vasa et al. 2009). In any case, argu-ments for the prepolymer production and use

of nucleotides by an abiotic metabolic cycleremain largely philosophical, as we simply donot know at what point in evolution (of bothmetabolism and genetic polymers) that the spe-cific components of contemporary nucleotideswere selected. It is feasible, for example, thatphosphate was used first as a metabolic co-factor before being incorporated into geneticpolymers.

As one possible solution to the paradox ofbase pairing, we have proposed that there wasa template molecule, or a class of molecules,present in the prebiotic chemical inventory,that formed stacks with the nucleobases thatwere energetically favored over stacks withthemselves. Each template, or “molecular mid-wife,” would have been large enough to accom-modate two or more nucleobases, if the baseswere arranged in a particular hydrogen-bondedstructure (e.g., a Watson-Crick base pair) (Hudand Anet 2000; Hud et al. 2007). The stacking ofmultiple midwife molecules, interleaved withnucleobases, would have locally concentratedand organized the bases, thereby aiding in thepolymerization of the earliest proto-RNA.These midwife molecules would have no longerbeen necessary for polymer synthesis (and rep-lication) once evolution had produced a supe-rior means for carrying out these reactions(e.g., catalysis via protein or RNA enzymes).We envision that the midwife molecules wouldhave been similar to the small planar moleculesknown to intercalate the bases of contemporaryRNA and DNA (Ihmels and Otto 2005). Beyonda purely speculative model, dye molecules thatintercalate DNA and RNA base pairs havebeen shown to promote the template-directedsynthesis of nucleic acids. For example, interca-lators can provide a .1000-fold enhancementto the ligation of short oligonucleotides (i.e.,tri- and tetranucleotides) (Jain et al. 2004),and promote a specific base pairing (Hud, Jainet al. 2007; Horowitz et al. 2010).

Should we assume that Watson-Crick basepairs (i.e., adenine paired with uracil, guaninepaired with cytosine) were used by the earliestproto-RNA? The successful incorporation ofnonnatural base pairs into nucleic acid duplexeshas given ample reason to seriously consider





this question (Switzer et al. 1989; Piccirilli et al.1990; Geyer et al. 2003). Most of the nonnaturalbase pairs studied thus far have maintained theform of a base comprised of a six-memberedring (like a pyrimidine) paired with a base com-prised of fused five- and six-membered rings(like a purine), which can be accommodatedin a helix also containing canonical Watson-Crick base pairs. However, duplexes that con-tain all noncanonical base pairs have also beenshown, including duplexes with only purine–purine base pairs (Groebke et al. 1998; Battersbyet al. 2007; Heuberger and Switzer 2008b;Engelhart et al. 2009), pyrimidine–pyrimidinebase pairs (Mittapalli et al. 2007a; Mittapalliet al. 2007b), and even duplexes with tricyclicbases (Krueger et al. 2007).

If we restrict our consideration of potentialproto-RNA base pairs to those with a Watson-Crick pairing geometry, there are still twelvepossible types of nucleobases that can form sixtypes of base pairs with distinct hydrogenbond donor–acceptor arrangements (Fig. 3)(Benner et al. 2004), of which only two types

are commonly found in contemporary life.Some of these base pairs, at first glance, seemlike reasonable candidates as proto-RNA alter-natives to the canonical base pairs, such asisoC paired with isoG (Fig. 3, bottom left).However, even this close analog of the guanine-cytosine base pair comes with significant chem-ical challenges that could have prevented it frombeing a component of proto-RNA (Jaworskiet al. 1985). In particular, isoguanine exists toabout 10% as its enol tautomer in water (Sepiolet al. 1976; Voegel et al. 1993; Seela et al. 1995),which has a hydrogen-bonding pattern that iscomplementary to U/T. Moreover, some ofthe four noncanonical Watson-Crick-like basepairs could only be realized experimentally byusing a nonpyrimidine or a nonpurine base, orby the connection of the base to the sugarthrough a carbon-carbon bond (C-nucleosides)(Fig. 3) (Benner 2004). The use of non- purineor pyrimidine heterocycles may not be so prob-lematic from a prebiotic standpoint. Althoughpurines are abundant in model reactions of theprebiotic chemical inventory using, for instance,

pyDAA

DonorDonor

Donor

Donor

Donor

Donor

Donor

Donor

Donor

Donor

Donor

Donor

Donor

Donor

Donor

Donor

Donor

Donor

Acceptor

AcceptorAcceptor

Acceptor

Acceptor

Acceptor Acceptor

Acceptor

Acceptor

Acceptor

Acceptor

Acceptor

Acceptor

Acceptor

Acceptor

Acceptor

Acceptor

Acceptor

puADD

puADA

puDDA

pyADD

pyDDA

pyADA

puDAD

puAAD

puDAA

pyDAD

pyAAD

H H

H

HO

O

O

O

H

N

N

N

NN

NN

N

N

N

NH

H

H H

HH

R

R

R

R

N

NN

N

N

H H

H

HO

O

H

N

NN

R

R

N

N N

N

N

H

H

H

HO

N

H

O

N

NR

R

N

NN

N

N

H

H

H

HNH

N

O

N

NR

R

N

NN

O

N

H

H

H H

HNH

N

O

NN

R

R

N

NN

O

N

Figure 3. The six possible hydrogen-bonding patterns that can join 12 different nucleobases with Watson-Crickgeometry. Nomenclature: pu, a [6,5] fused ring system; py, a six-membered ring. The hydrogen-bonding patternof acceptor (A) and donor (D) groups from the major to the minor groove is indicated. For example, thestandard nucleobase cytosine is pyDAA. R indicates the point of attachment of the backbone. Reprinted withpermission from Benner et al., 2004. Copyright Elsevier Science Ltd.





formamide and HCN, other, helix-compatibleparent heterocycles were surely present in theprebiotic chemical inventory (Mittapalli et al.2007a; Mittapalli et al. 2007b), and these shouldbe considered. With respect to C-nucleosides,although they are well-known, they have beenless well-examined from a prebiotic standpointthan the canonical N,O-acetal linked nucleo-sides. Regardless of any potential difficulties inprebiotic synthesis, the demonstration of allsix possible base pairs with a Watson-Crickparing geometry, made possible through theapplication of synthetic and physical organicchemistry, has provided invaluable constraintsregarding what should be considered more orless likely as possible ancestral proto-RNA build-ing blocks.

Among the proposals for proto-RNA non-canonical base pairs, that of purine-purine basepairs is one of the oldest (Crick 1968). The abio-tic formation of proto-RNA containing only pu-rine bases is appealing for several reasons. First,an all purine system requires only one type ofheterocycle, and the variety of conditions underwhich purine production has been shown sug-gests that purines may be formed more easilyin model prebiotic reactions than pyrimidines(Oro 1961; Sanchez et al. 1966; Saladino et al.2004). Second, nucleoside formation, as dis-cussed later, might have occured more effi-ciently with purine bases. Third, the purinebases have much more favorable stacking inter-actions in aqueous solution than the pyrimi-dine bases (Ts’o, Melvin et al. 1963; Inoue andOrgel 1983). The potential prebiotic obstaclespresented by poor pyrimidine stacking in aque-ous solution is illustrated by the inhibitionof template-directed polymerization reactionsthat require the coupling of successive pyrimi-dine nucleotides (Joyce 1987). With respect tothe molecular midwife hypothesis discussedearlier, purine–purine base pairs would alsobe expected to form assemblies with midwifemolecules at lower concentrations and over awider temperature range than Watson-Crickbase pairs.

Crick was among the first to propose thatlife originally used purine-purine base pairs.In particular, he proposed that A.I base pairs

might have come first (Fig. 4) (Crick 1968).However, duplexes formed by homo-A andhomo-I polymers proved to be relatively unsta-ble, and these polymers were prone to formingtriplexes (Howard and Miles 1977). Eschen-moser and coworkers have subsequently shownthat homo-purine oligonucleotides (with G.isoGbase pairs) can form very stable duplexes whenthe standard ribofuranose nucleosides are re-placed by either a ribopyranose or a dideoxy-hexopyranose sugar (Krishnamurthy et al.1996; Groebke et al. 1998) (other sugar mod-ifications are discussed later). More recently,evidence has been presented that standard oli-godeoxynucleotides can form duplexes withmixed G.isoG and diaminopurine.xanthinebase pairs (Fig. 4) that are of comparable stabil-ity to Watson-Crick duplexes (Heuberger andSwitzer 2008a). In light of these recent advan-ces, the decades-old idea of purine-purine

H

H

H

H

H

N

N

N

N

N

N

N

NN

NHRR

iGG

O

O

H

H

H

H

H

N

N

N

N

N

N

N

NO

NHRR

XD

O

N

H

H

N

N

N

N

N

N

NO

NHRR

IA

N

Figure 4. Chemical structures of purine–purine basepairs. A . I, adenine with hypoxanthine (the base ofinosine); D . X, diaminopurine with xanthine; G . iG,guanine with isoguanine.





base pairs in proto-RNA now looks even moreattractive.

The Trifunctional Moiety: Was It Ribose fromthe Start?

In RNA, ribose serves as the branch pointbetween the polymer backbone and the sidechains (i.e., the nucleobases). This role requiresa chemical moiety that is, at least, trifunctional.Sugars with three or more carbons meet thisminimum criterion and have been studied;among those investigated thus far, ribose appearsoptimal (Eschenmoser 1999). The conforma-tional landscape of the b-furanose anomer ofribose is well accommodated in a helical arrange-ment when its nucleosides are linked by phos-phate. Although some nucleic acids synthesizedwith alternative sugars form more stable Watson-Crick duplexes than RNA (Eschenmoser 1999),ribose appears to provide an optimal balancebetween duplex stability and the ability to adoptthe more globular structures required for creat-ing a catalytically active ribozyme, as exemplifiedby the variety of RNA structures observed inthe ribosome (Ban et al. 2000; Wimberly et al.2000).

The optimal functionality of ribose suggeststhat its inclusion in RNA is more the result ofevolutionary refinement than prebiotic avail-ability. The yield of ribose in the formose reac-tion, which is often cited as a prebioticallyplausible synthetic route to sugars, is low withrespect to the numerous other sugars producedin this reaction (Decker et al. 1982). Possiblesolutions to these challenges have included pro-posals that prebiotic ribose was preferentiallystabilized by adduct formation with cyanamide(Springsteen and Joyce 2004), borate (Ricardoet al. 2004) or phosphate (Muller et al. 1990),each of these adducts being more stable thanthe unmodified sugar.

If ribose does not stand out for its ease ofabiotic synthesis or chemical stability, thenwhat sugars would have been more likely tohave come before ribose in proto-RNA? Twentyyears ago, Joyce et al. thoughtfully discussedthis question (Joyce et al. 1987). With the goalof identifying a predecessor to ribose with

enhanced prebiotic availability and tolerancefor incorporation of mixed stereoisomers (i.e.,D- and L-sugars), these authors proposed aproto-RNA with acyclic analogs of ribose,including an acyclic flexible nucleic acid (FNA)backbone that is tantamount to an RNA back-bone with “deletion” of the 20 carbon (Fig. 2).Subsequently, the Benner laboratory synthe-sized oligonucleotides containing these link-ages and found that each substitution within aDNA oligonucleotide resulted in significantlydepressed duplex stability, and complemen-tary thirteen-base oligonucleotides containingeleven FNA thymine residues failed to hybridizeto a DNA template (Schneider and Benner1990). Later work by Merle and colleaguesshowed weak interaction of FNA oligomerswith DNA—the strongest pairing being ob-served for a homo-FNA-adenine oligomerwith the homo-dT complement, although thispairing was still depressed relative to the analo-gous DNA-DNA duplex (Merle et al. 1995).Thus, in addition to being well-accommodatedin a Watson-Crick duplex, the conformationalpreorganization of ribose nucleosides is clearlyan important component of duplex stability.

The Eschenmoser laboratory has performedan extensive and systematic investigation of howchanges to the natural b-furanosyl ribonucleo-sides of RNA, to nucleosides with a sugar of dif-ferent stereochemical configuration, phosphateconnectivity, carbon chain length or ring form(i.e., pyranose or furanose), affect the abilityof nucleic acids to base pair (Eschenmoser2007). A few examples are shown as Figure 1D–F. Among their findings was the observationthat oligomers incorporating the pyranosylform of nucleosides with a pentose sugar(including ribose) can form more stableduplexes than their furanosyl forms. This obser-vation is intriguing, particularly taken in lightof the observation that pyranosyl nucleosidesare apparently formed along with furanosylnucleosides in heating-drying reactions (Beanet al. 2007), suggesting that both forms couldhave been available for proto-RNA. Eschen-moser has interpreted the intermediate helicalstability of the contemporary furanose nucleo-sides as an indication that the contemporary





form of RNA prevailed due, in part, to an opti-mal, intermediate helical stability, not maximalhelical stability. He suggests that the stronghybridization of pentopyranosyl backbonescould have come at a price in a proto-RNAworld—in the form of product inhibition anddepressed sequence fidelity in replication.

The four carbon sugar threose is of particu-lar interest as a potential ancestor of ribose(Schoning et al. 2000; Herdewijn 2001a; Wildset al. 2002). The nucleosides of this threosecan be connected by phosphate esters to createa backbone with a five-atom repeat (Fig. 2D),one atom shorter than the RNA backbone.Experiments with 30,20-linked threofuranosenucleic acid, or TNA, have generated significantexcitement in the origin of life community. Inaddition to the potential relative ease of forma-tion of threose in a prebiotic environment,duplexes of TNA oligonucleotides are of com-parable stability to RNA and DNA duplexes,and TNA forms hybrid duplexes with RNAand DNA (Schoning et al. 2000; Wilds et al.2002). It has often been speculated that cross-hybridization with RNA is an essential featurefor an alternative nucleic acid to be considereda possible ancestor of RNA (Orgel 2000), asthis ability would have been necessary for trans-fer of sequence information between the twopolymers over the course of evolution.

A backbone containing even fewer atoms,but of the same polymer repeat length as TNA,has been termed GNA (Fig. 2C). The structureand stability of duplexes formed by oligonucleo-tides with this backbone have been examined indetail by Meggers and coworkers (Zhang andMeggers 2005; Schlegel et al. 2009). These inves-tigators have shown that GNA oligonucleotideswith Watson-Crick complementary sequencesform antiparallel duplexes that are even morestable than DNA or RNA duplexes of the samenucleobase sequence (Zhang and Meggers2005). In contrast, little or no hybridization isobserved between complementary GNA andDNA oligonucleotides (relative to the DNAhomoduplex) and base pairing with RNA is ofroughly equal stability relative to homoduplexRNA. Meggers and coworkers have also pro-vided solution-state evidence (by CD and

NMR) for the preorganization of GNA (on thesingle-strand and single-residue levels) into ahelix-compatible conformer. These data illus-trate that even an acyclic backbone can beconformationally preorganized. A subsequentX-ray crystal structure revealed that the GNAduplex makes extensive interstrand stackingcontacts and minimal intrastrand contacts (incontrast to RNA and DNA, for which the con-verse is true).

The Szostak laboratory recently showed thata GNA region of a DNA-GNA chimeric strandcould act as a template for Bst polymerase(Tsai et al. 2007). Despite little or no hybridiza-tion between GNA and DNA, full-length poly-merization of the dodecamer GNA templatewas achieved. Although it is unlikely thatpolymerase-catalyzed replication was contem-poraneous with the first self-replicating poly-mers (which these authors do not suggest),their results show that, even if cross-hybridiza-tion is disfavored, the presence of a duplex-stabi-lizing molecule (a protein or RNA polymerase,or a small molecule) can relax this requirementfor information transfer. A parallel line of evi-dence for this principle comes from the Switzerlaboratory, who have shown that 20,50-linkedDNA and FNA can also act as substrates for nat-ural DNA polymerases, even though 20,50-linkedDNA and FNA hybridize weakly or not at all tonatural DNA (Sinha et al. 2004; Heuberger andSwitzer 2008b).

Overall, synthetic organic chemists haveshown that numerous alternative sugars fromamong the tetroses, pentoses, and hexoses,and alternative conformations of ribose (e.g.,b-pyranosyl) can produce backbones that formdouble-stranded structures with Watson-Crickbase pairs, with some of these being evenmore stable than duplex RNA (Bolli et al.1997; Eschenmoser 1999; Schoning, Scholzet al. 2000; Eschenmoser 2004; Egli et al.2006). These results support the proposal thatproto-RNA backbones may have been con-structed from a sugar (or a mixture of sugars)other than ribose. The search through sugarspace is by no means complete. As one example,the ketose sugars have not been explored as pos-sible ancestors to ribose, but they should also be





considered as potentially simpler ancestors ofribose given the remarkably high yield withwhich the ketohexoses, such as fructose, areformed from DL-glyceraldehyde in a model pre-biotic reaction (Weber 1992).

Finally, from a reactivity perspective, aminosugars are also an attractive possibility. As dis-cussed later, the formation of phosphodiesterbonds in aqueous solution is problematic.Recent work by Szostak and coworkers (Mansyet al. 2008; Chen et al. 2009; Schrum et al.2009), as well as earlier work by Orgel and cow-orkers (Zielinski and Orgel 1985, 1987; Sieversand von Kiedrowski 1994), have shown the useof amino sugars in promoting polymerizationbecause of the greater nucleophicity of aminescompared with alcohols. The amino-substitutedphosphoimidazole-activated nucleosides usedin these studies were, obviously, prepared usingclassical synthetic organic techniques. Neverthe-less, given the promising polymerization resultsthat have been achieved with such nucleosides,as well as the presence of amino sugars in con-temporary life (e.g., glucosamine), potentialprebiotic reactions for the production of amino-substituted sugars are an intriguing area forfuture investigation.

Nucleoside Formation: How Did theNucleobase and Trifunctional Moiety FirstConnect?

The prebiotic origin of the glycosidic bond (anN,O-acetal) that connects the nucleobase toribose in a nucleoside has turned out to beone of the most vexing problems facing the ori-gin of RNA. Given that the synthesis of adenineby HCN polymerization was one of the earliestsuccesses of prebiotic chemistry (Oro 1960),and sugar formation from formaldehyde wasknown for a century prior (Butlerow 1861),many have speculated that prebiotic nucleosideformation resulted from the heating and dryingof preexisting nucleobases and ribose. Orgeland coworkers first reported the formation ofadenosine, inosine and guanosine when sam-ples of free adenine, hypoxanthine and guanine,respectively, were dried and heated with ribosein the presence of magnesium salts (Fuller

et al. 1972a, b). However, yields were low, andthe synthesis of guanosine was especially prob-lematic, because of the low solubility of guanine.More troublesome is the observation that thecanonical pyrimidine bases of RNA, uracil andcytosine, do not form nucleosides when driedand heated with ribose (Orgel 2004). It is worthnoting that even the previously reported forma-tion of adenosine in heating-drying reactions hasbeen described as difficult to reproduce (Zubayand Mui 2001).

Several alternative hypotheses for prebioticnucleoside formation have emerged as a resultof the difficulty of glycosidic bond formation.These hypotheses include the possibility thatthe nucleobases were formed on a preexistingsugar, or the converse, that the sugar was builtoff of the nucleobase. The earliest explorationof these alternative routes was again reported bythe Orgel laboratory (Sanchez and Orgel 1970).Starting with 50-phosphorylated ribose, followedby multiple steps, including chemical reagentaddition, UV anomerization, and hydrolysis,Sanchez and Orgel showed the synthesis of cyti-dine monophosphate. The Sutherland labora-tory has recently revisited this hypothesis withgreat vigor and has made impressive stridestoward the presentation of a comprehensivepathway for the stepwise abiotic formation ofb-DL-cytidine-20,30-cyclic phosphate, using cy-anamide, cyanoacetylene, glycolaldehyde, glyc-eraldehyde and inorganic phosphate as startingmolecules (Ingar et al. 2003; Anastasi et al.2006; Crowe and Sutherland 2006; Powneret al. 2009; Sutherland 2010).

The possibility that the nucleoside sugarwas originally synthesized on the nucleobases,the converse of the previously described nuc-leosidation hypothesis, has also received someexperimental support. Simply heating forma-mide in the presence of various mineral cata-lysts, as performed with a great variety ofminerals by the Saladino and DiMauro labora-tories (Saladino et al. 2007), have produced allfour canonical RNA bases, as well as purineacyclonucleosides (albeit with rudimentary“sugar” moieties) (Saladino et al. 2003). Giventhe plausibly prebiotic nature of formamide asa starting material and solvent (Benner et al.





2004), the one-pot synthesis of a proto-nucleo-side is a remarkable observation that certainlymerits serious consideration as an alternativeroute to nucleosides.

The concept of nucleic acid evolution alsoopens the door to that possibility that proto-RNA contained nucleobases and/or a trifunc-tional moiety that more easily formed nucleo-sides. The merit of this hypothesis was firstdemonstrated by Miller and co-workers, whoshowed that urazole, a five-membered hetero-cycle with the same Watson-Crick hydrogenbonding pattern as uracil, reacts spontaneouslywith ribose to give a- and b- pyrano- and fura-nosides (Kolb et al. 1994; Dworkin and Miller2000). More recently, Bean et al. investigatednucleoside formation with ribose and 2-pyrimidinone (Bean et al. 2007), a pyrimidinewith the same 2-oxo functional group as uraciland cytosine, but with a proton instead ofoxo- or amino-substitution at the 4 position.This base was shown to form nucleosides inheating-drying reactions, in which up to 12%of the remaining base was converted to theb-furanoside, possibly the highest conversionof a base to a nucleoside observed to date in asimple heating-drying reaction. Computationalstudies of this reaction have confirmed theimportance of the in-plane lone pair of 2-pyrimidinone in nucleoside formation, whichis also present in the imidazole ring of purines,and the importance of divalent cations in lower-ing the activation barrier to glycosidation(Bean, Sheng et al. 2007; Sheng et al. 2009).Together, the results obtained for urazole and 2-pyrimidinone illustrate that nucleoside forma-tion from a preexisting base and sugar maynot have been such a difficult step in proto-RNAformation, if the proto-nucleobases differedsomewhat from those of contemporary RNA.

Selection of the Nucleoside Linker: When DidNature Choose Phosphate?

The phosphate backbone is a chemical hallmarkof contemporary nucleic acids. As a triacid,phosphate provides nucleic acids with a meansto connect two alcohols, in the form of sugars,

via ester linkages, while maintaining a negativecharge (a key contributor to nucleic acid solu-bility in water). As Westheimer pointed out inhis insightful paper entitled “Why NatureChose Phosphates” (Westheimer 1987), it is dif-ficult to conceive of a chemical arrangementthat would provide better functional character-istics. The ester linkages provide connectivitythat is sufficiently labile so that enzymes canhydrolyze the polymer (for proofreading ormonomer salvage), whereas the negative chargeprovides sufficient screening of the phosphoruscenter from nucleophilic attack to stabilize thepolymer against hydrolysis. Additionally, theintrastrand electrostatic repulsion afforded bythe regularly-spaced negative charges may be cru-cial to an extended helical conformation and forinformation transfer that is rigorously dependenton Watson-Crick base pairing (contrast proteins,in which self-assembly via molecular recognitionoccurs, but not by the rule-based Watson-Cricksystem observed in nucleic acids). Westheimernoted that phosphate is unique here, as well: asa polyvalent acid with all acidic protons in closeproximity, its negative charge will afford greaterprotection from nucleophilic attack and intraresi-due repulsion than would a polyvalent acid withits ionizable moieties more distant from oneanother (e.g., citrate).

Benner and coworkers have experimentallyexplored the importance of backbone chargeon the ability for RNA-like molecules to formduplexes with Watson-Crick base pairs. As aneutral structural analog of the phosphodiester-linkage, they studied dimethylenesulfone-linked nucleosides (in which the O30 and O50

sugar atoms are replaced by methyelene groupsand phosphorus is replaced by sulfur) (Huanget al. 1991; Richert et al. 1996). The resultsfrom these studies were intriguing, but lessthan straightforward to interpret. Althoughshort sulfone-linked DNA analogs (sNAs)were shown to support Watson-Crick base pair-ing (Roughton et al. 1995), longer oligosulfonesappeared to have somewhat compromisedpairing abilities (Huang et al. 1991; Richertet al. 1996) and small changes in nucleobaseoligosulfone sequence resulted in appreciablechanges in oligomer solubility, folding and





aggregation (Eschgfaller et al. 2003; Schmidtet al. 2003). These authors interpreted theresults of these studies as evidence that (1)charged linkages are important for molecularrecognition by providing a repulsive energeticterm between the two backbones of a duplex,ensuring molecular recognition is largely afunction of the nucleobases, (2) that the regularrepeating charge limits intramolecular foldingof oligomers, thereby allowing these polymersto function well as linear duplexes and as tem-plates for the same during replication, and (3)that, given the dominance of charge in govern-ing the physical properties of these nucleic acidsin aqueous solution, changes in DNA/RNAnucleobase sequence have only a second-ordereffect on polymer properties (i.e., sequenceexerts more effect on the molecular recognitionof a complementary strand and less effect onhelical parameters), thereby allowing the useof virtually any possible nucleotide sequence(Benner et al. 2004).

Studies from the fields of biological andmedicinal chemistry have also provided someinsight regarding the importance of backbonecharge on nucleic acid base pairing. For exam-ple, methylphosphonates, in which one phos-phate oxygen is replaced by a methyl group,provide a neutral linkage that is among the clos-est possible structural analogs of the naturalbackbone (Miller et al. 1981). Oligonucleotideswith methylphosphonate substitutions supportWatson-Crick base pairing in water, as homo-duplexes and as hybrid duplexes with naturalnucleic acids (Miller et al. 1981; Kiblerherzoget al. 1991; Schweitzer and Engels 1999). Sur-prisingly, removal of negative charges does notresult in a substantial increase in duplex stabil-ity, as might be expected because of the reducedCoulombic repulsion between backbones com-pared with DNA and RNA. In contrast, twoneutral polymers with more radical structuralchanges, “morpholinos” and PNAs, have beenshown to form hybrid duplexes with RNA andDNA that are of greater stability than homo-DNA or homo-RNA duplexes (Summerton andWeller 1989; Egholm et al. 1993; Brown et al.1994; Summerton and Weller 1997), and PNAwill form homoduplexes with Watson-Crick

molecular recognition (Rasmussen et al. 1997;He et al. 2008). The hybridization of nucleicacids is the result of a complex interplay ofnumerous factors, including hydrogen bonds,nucleobase stacking interactions, backboneconformational landscapes, electrostatic inter-actions and hydration effects. Thus, the sensi-tivity of duplex stability to backbone charge isconsistent with the predicted importance ofthe phosphate backbone in facilitating basepairing, but the observation of several success-ful neutral polymers suggest we presently lacksufficient information to understand the preciseorigin of the observed charge-dependent differ-ential stabilities.

Although phosphate esters are functionallysuperior linkages, they show several chemicalcharacteristics that may have made it difficultto incorporate them into the earliest biopoly-mers. The first is solubility. In the presence ofdivalent cations, free phosphate tends to forminsoluble minerals; outside of living organisms,the phosphate resources on Earth even today arefound as insoluble mineral deposits (Keefe andMiller 1995). Indeed, phosphate is a limitingreagent in much of contemporary life. Althoughrecent hypotheses have been proposed for theavailability of reduced forms of phosphoruson the earth which are more soluble (De Graafand Schwartz 2000; Bryant and Kee 2006;Schwartz 2006; Pasek 2008), the low aqueoussolubility of the contemporary (V) oxidationstate of phosphorus, in the presence of divalentcations, and the positive DH associated withphosphoester formation present problems forprebiotic phosphate chemistry.

Another chemical characteristic of phos-phate that would have made polymer formationchallenging is the kinetic barrier to spontaneousphosphate ester formation. The negative chargeof phosphate above pH � 1 presents a barrier toester formation, which, like hydrolysis, must pro-ceed via nucleophilic attack on the phosphorusatom. It is possible to activate phosphate forester formation (e.g., using methylimidazolides).However, such activated phosphates are neces-sarily high-energy compounds and, as wouldbe expected, hydrolysis competes with polymer-ization.





Phosphate activation poses problems forboth de novo production of oligomers frommonomers and the incorporation of oligomersinto higher polymers. In the case of monomerpolymerization, Ferris and coworkers haveshown that activated imidazolide derivativesof adenosine nucleotides will polymerize onmineral surfaces, but a large proportion of themonomer is incorporated into cyclic dimers(Miyakawa et al. 2006). In attempting to poly-merize tiled half-complementary hexanucleo-tides into higher oligo- and polymers bycarbodiimide activation, Kawamura and Oka-moto also observed substantial amounts ofcyclized starting material (Kawamura and Oka-moto 2001). In both of these examples, a signif-icant amount of starting material is irreversiblyincorporated into an undesired side product. Inaddition to being a nuisance in the lab, theseside products would have amounted to a fataland committed step in the synthesis of a nascentproto-RNA. This problem illustrates a difficultyin nonenzymatic polymerization that must betaken into account when considering how thenature of the synthetic routes to and structuralidentities of early genetic polymers: irreversiblelinkages are adaptive for an informationalpolymer only when mechanisms exist to makethem conditionally reversible (so as to allowproofreading).

The difficulties described earlier confoundthe efficient prebiotic chemical synthesis ofphosphate esters—ligation of activated sub-strates yields mixed regioisomers and poor rep-lication fidelity because of strand cyclization,base misincorporation, and premature productchain termination (Joyce and Orgel 1999; Kawa-mura and Okamoto 2001; Miyakawa al. 2006).Work by Usher and colleagues has providedencouraging evidence that the 20,50 linkageis more labile than the contemporary 30,50

regioisomer, suggesting that thermodynamicselection may be helpful in addressing a subsetof the aforementioned problems with phospho-diester linkages (Usher and McHale 1976).However, all these difficulties can be circum-vented by the use of a linker chemistry that ini-tially forms a low-energy, reversible bond,allowing for selection of the thermodynamic

product (i.e., in the case of RNA, the Watson-Crick base-paired product).

One remarkable example of the power ofreversible linkages was provided by Lynn andcoworkers, who showed that 50-deoxy,50-amino,30-deoxy, 30-formylmethyl-dT could polymer-ize via reductive amination in aqueous solution(Li et al. 2002). Polymerization occurred only inthe presence of a d(A8) template; without a tem-plate, no oligomer product was detected. Inter-estingly, only linear polymers were observed.Here, polymerization decreases the entropiccost of hybridization, driving base pairing(which is disfavored at the mononucleotidelevel in water) (Yakovchuk et al. 2006). In turn,hybridization increases the local concentrationof amines and aldehydes, driving the formationof imine linkages (which are otherwise disfa-vored in 55 M water). Since the irreversiblereduction step requires the imine linkage to bepresent, and imine formation was templatedependent, only linear polymers were formed.In this system, the subtle interplay betweenbase pairing and linkage formation drives thesimultaneous formation of reversible covalentand noncovalent interactions which are other-wise disfavored. Several other intriguing exam-ples of self-assembling reversible polymers havebeen shown, particularly by the Lehn laboratory(Sreenivasachary and Lehn 2005; Sreenivasach-ary et al. 2006; Sreenivasachary and Lehn 2008).

Acetal linkages have the potential to providereversible bonds, thereby providing the benefitsshowed by the reversible imine chemistry ofLynn. For example, the 30, 50-formacetal linkage(i.e., phosphate replaced by -O-CH2-O-), hasbeen examined by a number of laboratories(albeit not in fully modified oligonucleotides).Phosphorus (V) esters and acetals are bothtetrahedral and approximately isosteric, andit is known that aNAs can base pair with RNAs(Matteucci and Bischofberger 1991). Speci-fically, point substitutions of this linkageimpart a modest destabilization to DNA-DNAand DNA–RNA duplexes and show salt-dependent effects on RNA-RNA duplexes (sta-bilization at low salt, destabilization at highsalt) (Jones et al. 1993; Rice and Gao 1997;Rozners et al. 2007; Kolarovic et al. 2009).





Investigators have invoked compatibility withthe B-versus A-form helices, electrostatic, andhydration effects for these observed differentialeffects. Further, some evidence has been pre-sented that the destabilization associated withthis substitution in DNA–DNA duplexes issequence-dependent (Pitulescu et al. 2008).Finally, the 30-thioformacetal linkage has beenshown to slightly destabilize DNA-DNAduplexes and slightly stabilize DNA-RNAduplexes (Jones et al. 1993). The latter stabiliza-tion was explained by a combination of sugarpucker, bond lengths, and accessible torsionangles resulting from the 30-thio substitution.This sugar pucker effect was later confirmedby NMR spectroscopy (Rice and Gao 1997).

With the proper choice of aldehyde orketone, these linkages can provide a negativecharge that has proven critical to the phospho-diester backbone. For these reasons, acetal-linked nucleic acids (aNAs) are an intriguingpotential ancestral polymer (Hud and Anet2000). Acetal bond formation is favored overhydrolysis on water removal (Wiberg et al.1994). The concentration of substrates thatoccurs during dry-down, driving bond forma-tion, would also favor the hybridization ofshort, complementary sequences, providingan accessible means by which to drive polymer-ization.

The functionalization of an acetal with anegatively charged group, such as carboxylategroup in the case of the aldehyde glyoxylate,would provide a localized charge near the back-bone acetal linkage, thereby providing many ofthe favorable electrostatic properties affordedby phosphate. Glyoxylate has been shown toform in a prebiotically plausible reaction fromglycolaldehyde (Weber 2001). We have specu-lated that glyoxylate could have preceded phos-phate in an early proto-RNA, in the form ofglyoxylate-acetal nucleic acids (gaNAs, Fig. 2G)(Bean et al. 2006). Based on energy minimizedmolecular models, it appears that a gaNA du-plex would show helical properties very similarto RNA, suggesting that information transfercould occur between the two polymers. The for-mation of gaNA dinucleotides has been shownon heating and drying of a neutral solution

containing nucleosides, glyoxylate and mono-or divalent cations (Bean et al. 2006). These ace-tal linkages formed by glyoxylate and nucleo-sides are surprisingly stable to hydrolysis—nodetectable decomposition occurs after twoweeks at room temperature and neutral pH.Importantly, gaNAs can be hydrolyzed at mod-erately elevated temperature in the presence ofsalt (i.e., 60 8C, 1 M MgCl2). An accessiblemeans to modulate polymer stability (in thiscase, temperature and salt concentration),which provides an attractive means by whichto “turn on and off” thermodynamic controlof polymerization—a crucial feature in thechemical synthesis of an early informationalpolymer.

It is possible that other, similar, moleculescould form suitable linkages. For instance, pyr-uvate, like glyoxylate, provides a negative chargeand tetrahedral geometry, and ketal bond for-mation is still expected to be more facile thanphosphodiester bond formation. Further, pyru-vate is ubiquitous in life and formed in excellentyield in model prebiotic reactions (Weber2001).

It is important to note that additional pro-mising models have presented for the prebioticformation of oligonucleotides that begin witheither a reduced form of phosphorus or mono-nucleotides activated as cyclic monophos-phates. Reduced phosphorus species, suchas phosphite, in addition to showing the afore-mentioned enhanced solubility relative to phos-phates, enjoy enhanced reactivity. Schwartz andcoworkers reported high yields of uridine50-H-phosphonate (as high as 44% in the pres-ence of urea) in 60 8C drying reactions of aque-ous ammonium phosphite, uridine and,optionally, urea. Additionally, when dried at110 8C, these authors reported production ofsmall amounts of putative higher-order prod-ucts, the most abundant of which was producedin 4% yield, which they tentatively assigned asthe 50,50 H-phosphonate-linked dinucleotide.This species could be oxidized with iodine toa second species, assigned as the phosphate-linked dinucleotide (De Graaf and Schwartz2005). Like acetals, H-phosphonate diesters area conditionally reversible, low-energy linkage.





Although they are uncharged and hydrolyti-cally labile, upon oxidation, they become nega-tively charged, comparatively hydrolyticallyinert phosphate diesters, making phosphonatesanother attractive, conditionally reversible back-bone linkage. Finally, DiMauro and coworkershave recently reported that 30, 50-cyclic AMPand GMP will spontaneously form homo-Aand homo-G oligonucleotides, respectively, inaqueous solution (Costanzo et al. 2009). Theseauthors propose that nucleotides activated ascyclic monophosphates are potentially prebio-tic, based on their earlier demonstration of30, 50-cyclic AMP formation in formamide sol-utions containing phosphate minerals (Cos-tanzo et al. 2007).

As Westheimer, Benner, and others have ele-gantly discussed, phosphate affords RNA nu-merous positive characteristics (Westheimer1987; Benner et al. 2004). When consideringsome of the adaptive traits of RNA as a poly-mer—hydrolytic resistance, water solubility,extended conformation, and heritability of anarbitrary sequence because of rule-based mole-cular recognition—it is tempting to speculatethat phosphate might be the optimal linker forall possible RNA-like polymers. The roles ofphosphate in contemporary life, however, arepossible only because of a highly optimizedsuite of catalysts, which enable the transfer ofenergy between disparate substrates and the“phosphate economy” (i.e., via the agency ofATP and similar compounds). Absent such ca-talysis, the chemical synthesis (and subsequentproofreading) of phosphate esters and anhy-drides is problematic. Many of the same attributesof phosphate that are adaptive in contemporarylife could have been maladaptive in the proto-biopolymer world. We therefore suggest that otherchemistries, including those discussed here, beconsidered as possible linkers for the nucleosidesof proto-RNA. These linkers could have incor-porated some of the favorable traits of phosphate(e.g., electrostatic benefits, significant hydrolyticstability) while addressing some of the problem-atic ones (e.g., disfavored bond formation,lack of a proofreading mechanism in chemicalsynthesis) (Westheimer 1987; Bean, Anet et al.2006).

CHALLENGES AND FUTURE DIRECTIONS

Our primary objective in this article has been topresent some of the known challenges for theprebiotic synthesis of RNA, as well as potentialsolutions to these challenges that become avail-able if we accept the possibility that one or moreproto-RNAs predated RNA. More than a mereacademic exercise, the discovery (or “reinven-tion”) of a self-assembling RNA-like polymerwould have tremendous potential for applica-tions in biotechnology, medicine, and materialsscience.

Although our current examples of self-assembling polymers with reversible linkages arelimited, the properties of these polymers areimpressive and indicative of the promise heldby the development of this area of polymer sci-ence. For example, dynamic covalent assemblieshave been examined by the Lehn laboratory inthe form of guanosine hydrogels, and theyhave shown that these self-assembling materialscan act as controlled release media for bioactivesmall molecules (Sreenivasachary and Lehn2005, 2008). Similarly, these investigators, aswell as Leibler and coworkers, have also shownthat polymers with the capacity for constitu-tional reorganization show a number of intrigu-ing material properties, including self-healingand chemoresponsive characteristics (Sreeniva-sachary et al. 2006; Cordier et al. 2008).

Most recently, Ghadiri and coworkers haveshowed the ability for peptide nucleic acidpolymer with reversible linkages between a poly-cysteine backbone and thioester acyclonucleo-sides to undergo a form of evolution inresponse to the presence of different DNA tem-plates (Ura et al. 2009). Almost simultaneously,the Liu group reported a similar system, involv-ing amidation and reductive amination of aPNA backbone with acyclonucleosides contain-ing carboxylate and aldehyde functional groups(Heemstra and Liu 2009). Acyclonucleosidesformed in formamide mixtures recentlyreported by Di Mauro and coworkers share atleast some structural similarity with thoseused by the Ghadiri and Liu laboratories, sug-gesting that prebiotic routes may exist tonucleosides that multiple investigators have





now shown could couple to varied preformedbackbones. One can now imagine the develop-ment of RNA-like polymers following a similartheme, creating functional polymers that areenvironmentally responsive, perhaps evenevolving new functions in response to unantici-pated environmental signals, which is com-pletely feasible for polymers with reversiblelinkages that can assume the thermodynami-cally most favored state, as defined by their localenvironment.

Finally, the prospect of proto-RNA is reasonfor optimism that we will yet discover RNA-likepolymers with self-assembling properties thatare superior to any currently known. The chem-ical space that still must be searched is vast, butit is not infinite, and the studies discussed ear-lier have certainly provided clues to where wewill most likely find such polymers, and perhapseven the origin of RNA.

ACKNOWLEDGMENTS

We thank R. Krishnamurthy, P. Herdewijn,A. Schwartz and C. Switzer for helpful com-ments on this article. Support from the NASAExobiology Program and the National ScienceFoundation is gratefully acknowledged.

REFERENCES

Anastasi C, Crowe MA, Powner MW, Sutherland JD. 2006.Direct assembly of nucleoside precursors from two-and three-carbon units. Angew Chem Int Ed Engl 45:6176–6179.

Anet FAL. 2004. The place of metabolism in the origin oflife. Curr Opin Chem Biol 8: 654–659.

Ban N, Nissen P, Hansen J, Moore PB, Steitz TA. 2000. Thecomplete atomic structure of the large ribosomal subunitat 2.4 Angstrom resolution. Science 289: 905–920.

Battersby TR, Albalos M, Friesenhahn MJ. 2007. An unusualmode of DNA duplex association: Watson-Crick interac-tion of all-purine deoxyribonucleic acids. Chem Biol 14:525–531.

Bean HD, Anet FAL, Gould IR, Hud NV. 2006. Glyoxylate asa backbone linkage for a prebiotic ancestor of RNA. OrigLife Evol B 36: 39–63.

Bean HD, Sheng YH, Collins JP, Anet FAL, Leszczynski J,Hud NV. 2007. Formation of a b-pyrimidine nucleosideby a free pyrimidine base and ribose in a plausible prebi-otic reaction. J Am Chem Soc 129: 9556–9557.

Benner SA. 2004. Understanding nucleic acids using syn-thetic chemistry. Acc Chem Res 37: 784–797.

Benner SA, Ricardo A, Carrigan MA. 2004. Is there a com-mon chemical model for life in the universe? Curr OpinChem Biol 8: 672–689.

Bolli M, Micura R, Eschenmoser A. 1997. Pyranosyl-RNA:chiroselective self-assembly of base sequences by ligativeoligomerization of tetra nucleotide-20,30-cyclophos-phates (with a commentary concerning the origin of bio-molecular homochirality). Chem Biol 4: 309–320.

Brown S, Thomson S, Veal J, Davis D. 1994. NMR solutionstructure of a peptide nucleic acid complexed with RNAScience 265: 777–780.

Bryant DE, Kee TP. 2006. Direct evidence for the availabilityof reactive, water soluble phosphorus on the early Earth.H-Phosphinic acid from the Nantan meteorite. ChemCommun: 2344–2346.

Butlerow A. 1861. Bildung einer zuckerartigen substanzdurch synthese. Liebigs Ann Chem 120: 295–298.

Cairns-Smith AG. 1982. Genetic takeover and the mineralorigins of life. Cambridge University Press, Cambridge.

Chen JJ, Cai X, Szostak JW. 2009. N20!P30 phosphorami-date glycerol nucleic acid as a potential alternative geneticsystem. J Am Chem Soc 131: 2119–2121.

Cleaves HJ, Nelson KE, Miller SL. 2006. The prebiotic syn-thesis of pyrimidines in frozen solution. Naturwissen-schaften 93: 228–231.

Cordier P, Tournilhac F, Soulie-Ziakovic C, Leibler L. 2008.Self-healing and thermoreversible rubber from supramo-lecular assembly. Nature 451: 977–980.

Costanzo G, Saladino R, Crestini C, Ciciriello F, Di Mauro E.2007. Nucleoside phosphorylation by phosphate miner-als. J Biol Chem 282: 16729–16735.

Crick FHC. 1968. The origin of the genetic code. J Mol Biol38: 367–379.

Crowe MA, Sutherland JD. 2006. Reaction of cytidinenucleotides with cyanoacetylene: support for the inter-mediacy of nucleoside-20, 30-cyclic phosphates in the pre-biotic synthesis of RNA. ChemBioChem 7: 951–956.

De Graaf RM, Schwartz AW. 2000. Reduction and activationof phosphate on the primitive earth. Orig Life Evol B 30:405–410.

De Graaf RM, Schwartz AW. 2005. Thermal synthesis ofnucleoside H-phosphonates under mild conditions.Orig Life Evol B 35: 1–10.

De Voe H, Tinoco I. 1962. The stability of helical poly-nucleotides: base contributions. J Mol Biol 4: 500–517.

Decker P, Schweer P, Pohlmann R. 1982. Identification offormose sugars, presumable prebiotic metabolites, usingcapillary gas chromatography/gas chromatography-mass spectroscopy of n-butoxime trifluoroacetates onOV-225. J Chromatogr 244: 281–291.

Dworkin JP, Miller SL. 2000. A kinetic estimate of the freealdehyde content of aldoses. Carbohydr Res 39: 359–365.

Egholm M, Buchardt O, Christensen L, Behrens C, Freier S,Driver D, Berg R, Kim S, Norden B, Nielsen P. 1993. PNAhybridizes to complementary oligonucleotides obeyingthe Watson–Crick hydrogen-bonding rules. Nature365: 566–568.

Egli M, Pallan PS, Pattanayek R, Wilds CJ, Lubini P, MinasovG, Dobler M, Leumann CJ, Eschenmoser A. 2006. Crystalstructure of homo-DNA and Nature’s choice of pentose





over hexose in the genetic system. J Am Chem Soc 128:10847–10856.

Engelhart AE, Morton TH, Hud NV. 2009. Evidence ofstrong hydrogen bonding by 8-amino-guanine. ChemCommun: 647–649.

Eschenmoser A. 1999. Chemical etiology of nucleic acidstructure. Science 284: 2118–2124.

Eschenmoser A. 2004. The TNA-family of nucleic acid sys-tems: Properties and prospects. Orig Life Evol B 34:277–306.

Eschenmoser A. 2007. The search for the chemistry of life’sorigin. Tetrahedron 63: 12821–12844.

Eschgfaller B, Schmidt JG, Konig M, Benner SA. 2003. Syn-thesis and properties of oligodeoxynucleotide analogswith bis(methylene) sulfone bridges. Helv Chim Acta86: 2959–2997.

Fuller WD, Sanchez RA, Orgel LE. 1972a. Studies in prebi-otic synthesis. VI. Synthesis of purine nucleosides. JMol Biol 67: 25–33.

Fuller WD, Sanchez RA, Orgel LE. 1972b. Studies in prebi-otic synthesis: VII. Solid-state synthesis of purine nucleo-sides. J Mol Evol 1: 249–257.

Gesteland R, Atkins JF. 2006. The RNA world: The nature ofmodern RNA suggests a prebiotic RNA world (3rd ed.)Cold Spring Harbor Laboratory Press, Cold Spring Har-bor, NY.

Geyer CR, Battersby TR, Benner SA. 2003. Nucleobase pair-ing in Watson-Crick-like genetic expanded informationsystems. Structure 11: 1485–1498.

Groebke K, Hunziker J, Fraser W, Peng L, Diederichsen U,Zimmermann K, Holzner A, Leumann C, EschenmoserA. 1998. Why pentose- and not hexose-nucleic acids?Purine-purine pairing in homo-DNA: guanine, isogua-nine, 2,6-diaminopurine, and xanthine. Helv ChimActa 81: 375–474.

Guerrier-Takada C, Gardiner K, Marsh T, Pace N, Altman S.1983. The RNA moiety of ribonuclease P is the catalyticsubunit of the enzyme. Cell 35: 849–857.

He W, Hatcher E, Balaeff A, Beratan D, Gil R, Madrid M,Achim C. 2008. Solution structure of a peptide nucleicacid duplex from NMR data: Features and limitations. JAm Chem Soc 130: 13264–13273.

Heemstra JM, Liu DR. 2009. Templated synthesis of peptidenucleic acids via sequence-selective base-filling reactions.J Am Chem Soc 131: 11347–11349.

Herdewijn P. 2001a. TNA as a potential alternative to naturalnucleic acids. Angew Chem Int Ed Engl 40: 2249–2251.

Herdewijn P. 2001b. TNA as a potential alternative to naturalnucleic acids. Angew Chem Int Ed Eng 40: 2249–2251.

Heuberger BD, Switzer C. 2008a. An alternative nucleobasecode: Characterization of purine-purine DNA doublehelices bearing guanine-isoguanine and diamino-purine-7-deaza-xanthine base pairs. ChemBioChem 9:2779–2783.

Heuberger BD, Switzer C. 2008b. A nonRNA candidaterevisited: Both enantiomers of flexible nucleoside tri-phosphates are DNA polymerase substrates. J Am ChemSoc 130: 412–413.

Hollis JM, Jewell PR, Lovas FJ, Remijan A. 2004. Green banktelescope obserations of interstellar glycolaldehyde: Low-temperature sugar. Astrophys J 613: L45–L48.

Horowitz ED, Engelhart AE, Chen MC, Quarles KA, SmithMW, Lynn DG, Hud NV. 2010. Intercalation as a means tosuppress cyclization and promote polymerization ofbase-pairing oligonucleotides in a prebiotic world. ProcNatl Acad Sci 107: 5288–5293.

Howard FB, Miles HT. 1977. Interaction of poly(A) andpoly(I), a reinvestigation. Biochemistry 16: 4647–4650.

Huang Z, Schneider KC, Benner SA. 1991. Building-blocksfor oligonucleotide analogs with dimethylene sulfide,sulfoxide, and sulfone groups replacing phosphodiesterlinkages. J Org Chem 56: 3869–3882.

Hud NV, Anet FAL. 2000. Intercalation-mediated synthesisand replication: A new approach to the origin of life. JTheor Biol 205: 543–562.

Hud NV, Jain SS, Li X, Lynn DG. 2007. Addressing the prob-lems of base pairing and strand cyclization in template-directed synthesis – A case for the utility and necessityof ‘molecular midwives’ and reversible backbone link-ages for the origin of proto-RNA. Chem Biodiver 4:768–783.

Ihmels H, Otto D. 2005. Intercalation of organic dye mole-cules into double-stranded DNA: General principles andrecent developments. Top Curr Chem 258: 161–204.

Ingar A-A, Luke RWA, Hayter BR, Sutherland JD. 2003. Syn-thesis of cytidine ribonucleotides by stepwise assembly ofthe heterocycle on a sugar phosphate. ChemBioChem 4:504–507.

Inoue T, Orgel LE. 1983. A non-enzymatic RNA polymerasemodel. Science 219: 859–862.

Jain SS, Anet FAL, Stahle CJ, Hud NV. 2004. Enzymaticbehavior by intercalating molecules in a template-directed ligation reaction. Angew Chem Int Ed Engl 43:2004–2008.

Jaworski A, Kwiatkowski JS, Lesyng B. 1985. Why isoguanineand isocytosine are not the components of the geneticcode. Int J Quantum Chem 28, S12: 209–216.

Jones RJ, Lin K-Y, Milligan JF, Wadwani S, Matteucci MD.1993. Synthesis and binding properties of pyrimidine oli-gonucleotide analogs containing neutral phosphodiesterreplacements: The formacetal and 30-thioformacetalinternucleoside linkages. J Org Chem 58: 2983–2991.

Joyce GF. 1987. Nonenzymatic template-directed synthesisof informational macromolecules. Cold Spring HarborSymp Quant Biol 52: 41–51.

Joyce GF, Orgel LE. 1999. Prospects for understanding theorigin of the RNAworld. In The RNAWorld, Second Edi-tion: The Nature of Modern RNA Suggests a PrebioticRNA World (Eds. J.F. Atkins, R.F. Gesteland), 49–77.Cold Spring Harbor Laboratory Press, Cold SpringHarbor, NY.

Joyce GF, Schwartz AW, Miller SL, Orgel LE. 1987. The casefor an ancestral genetic system involving simple analogsof the nucleotides. Proc Natl Acad Sci 84: 4398–4402.

Kawamura K, Okamoto F. 2001. Cyclization and dimeriza-tion of hexanucleotides containing guanine and cytosinewith water-soluble carbodiimide. Viva Origino 29:162–167.

Keefe AD, Miller SL. 1995. Are polyphosphates or phosphateesters prebiotic reagents? J Mol Evol 41: 693–702.

Kiblerherzog L, Zon G, Uznanski B, Whittier G, WilsonWD. 1991. Duplex stabilities of phosphorothioate,





methylphosphonate, and RNA analogs of 2 DNA14-mers. Nucleic Acids Res 19: 2979–2986.

Kolarovic A, Schweizer E, Greene E, Gironda M, Pallan P,Egli M, Rozners E. 2009. Interplay of structure, hydrationand thermal stability in formacetal modified oligonu-cleotides: RNA may tolerate nonionic modifications bet-ter than DNA. J Am Chem Soc 131: 14932–14937.

Kolb VM, Dworkin JP, Miller SL. 1994. Alternative bases inthe RNA world: the prebiotic synthesis of urazole and itsribosides. J Mol Evol 38: 549–557.

Krishnamurthy R, Pitsch S, Minton M, Miculka C, Wind-hab N, Eschenmoser A. 1996. Pyranosyl-RNA: base pair-ing between homochiral oligonucleotide strands ofopposite sense of chirality. Angew Chem Int Ed Engl 35:1537–1541.

Krueger AT, Lu HG, Lee AHF, Kool ET. 2007. Synthesis andproperties of size-expanded DNAs: Toward designed,functional genetic systems. Acc Chem Res 40: 141–150.

Kruger K, Grabowski PJ, Zaug AJ, Sands J, Gottschling DE,Cech TR. 1982. Self-splicing RNA: Autoexcision andautocyclization of the ribosomal RNA interveningsequence of Tetrahymena. Cell 31: 147–157.

Kurreck J. 2003. Antisense technologies—Improvementthrough novel chemical modifications. Eur J Biochem270: 1628–1644.

Li X, Zhan Z-YJ, Knipe R, Lynn DG. 2002. DNA-catalyzedpolymerization. J Am Chem Soc 124: 746–747.

Mandal M, Breaker RR. 2004. Gene regulation by ribos-witches. Nat Rev Mol Cell Biol 5: 451–463.

Mansy SS, Schrum JP, Krishnamurthy M, Tobe S, Treco DA,Szostak JW. 2008. Template-directed synthesis of agenetic polymer in a model protocell. Nature 454:122–125.

Martins Z, Botta O, Fogel ML, Sephton MA, Glavin DP, Wat-son JS, Dworkin JP, Schwartz AW, Ehrenfreund P. 2008.Extraterrestrial nucleobases in the Murchison meteorite.Earth Planet Sci Lett 270: 130–136.

Matteucci MD, Bischofberger N. 1991. Sequence-definedoligonucleotides as potential therapeutics. Annu RepMed Chem 26: 287–296.

Merle Y, Bonneil E, Merle L, Sagi J, Szemzo A. 1995. Acyclicoligonucleotide analogues. Int J Biol Macromol 17:239–246.

Miller PS, McParland KB, Jayaraman K, Tso POP. 1981. Bio-chemical and biological effects of nonionic nucleic acidmethylphosphonates. Biochem Cell Biol 20: 1874–1880.

Mittapalli GK, Osornio YM, Guerrero MA, Reddy KR,Krishnamurthy R, Eschenmoser A. 2007a. Mapping thelandscape of potentially primordial informational olig-omers: oligodipeptides tagged with 2,4-disubstituted 5-aminopyrimidines as recognition elements. AngewChem Int Ed Engl 46: 2478–2484.

Mittapalli GK, Reddy KR, Xiong H, Munoz O, Han B, DeRiccardis F, Krishnamurthy R, Eschenmoser A. 2007b.Mapping the landscape of potentially primordial infor-mational oligomers: Oligodipeptides and oligodipep-toids tagged with triazines as recognition elements.Angew Chem Int Ed Engl 46: 2470–2477.

Miyakawa S, Joshi PC, Gaffey MJ, Gonzalez-Toril E, HylandC, Ross T, Rybij K, Ferris JP. 2006. Studies in the mineral

and salt-catalyzed formation of RNA oligomers. OriginsLife Evol B 36: 343–361.

Miyakawa S, Yamanashi H, Kobayashi K, Cleaves HJ, MillerSL. 2002. Prebiotic synthesis from CO atmospheres:Implications for the origins of life. Proc Natl Acad Sci99: 14628–14631.

Muller D, Pitsch S, Kittaka A, Wagner E, Wintner CE,Eschenmoser A. 1990. Chemie von a-Aminonitrilen.Aldomerisierung von Glycoladehyd-phosphat zu racemi-schen Hexose-2,4,6-triphosphaten und (in Gegenwartvon Formaldehyd) reacemischen Pentose-2,4-diphos-phaten: rac-Allose-2,4,6-triphosphat und rac-Ribose-2,4-diphosphat sind die Reaktionshauptprodukte. HelvChim Acta 73: 1410–1468.

Nelson KE, Levy M, Miller SL. 2000. Peptide nucleic acidsrather than RNA may have been the first genetic mole-cule. Proc Natl Acad Sci 97: 3868–3871.

Orgel L. 2000. A simpler nucleic acid. Science 290:1306–1307.

Orgel LE. 1968. Evolution of the genetic apparatus. J MolBiol 38: 381–393.

Orgel LE. 1998. The origin of life – a review of facts andspeculations. TIBS 23: 491–495.

Orgel LE. 2004. Prebiotic chemistry and the origin of theRNA world. Crit Rev Biochem Mol Biol 39: 99–123.

Orgel LE. 2008. The implausibility of metabolic cycles onthe prebiotic Earth. Plos Biology 6: 5–13.

Oro J. 1961. Mechanism of synthesis of adenine from hydro-gen cyanide under possible primitive Earth conditions.Nature 191: 1193–&.

Oro J. 1960. Synthesis of adenine from ammonium cyanide.Biochem Biophys Res Comm 2: 407–412.

Pasek MA. 2008. Rethinking early Earth phosphorus geo-chemistry. Proc Natl Acad Sci 105: 853–858.

Piccirilli JA, Krauch T, Moroney SE, Benner SA. 1990. Enzy-matic incorporation of a new base pair into DNA andRNA extends the genetic alphabet. Nature 343: 33–37.

Pitsch S, Wendeborn S, Jaun B, Eschenmoser A. 1993. Whypentose- and not hexose-nucleic acids? Part VII.Pyranosyl-RNA(‘p-RNA’). Helv Chim Acta 76: 2161–2183.

Pitulescu M, Grapp M, Kratzner R, Knepel W, DiederichsenU. 2008. Synthesis of formacetal-linked dinucleotides tofacilitate dsDNA bending and binding to the homeodo-main of PAX6. Eur J Org Chem: 2100–2106.

Pizzarello S. 2006. The chemistry of life’s origin: A carbona-ceous meteorite perspective. Acc Chem Res 39: 231–237.

Plankensteiner K, Reiner H, Schranz B, Rode BM. 2004. Pre-biotic formation of amino acids in a neutral atmosphereby electric discharge. Angew Chem Int Ed Engl 43:1886–1888.

Powner MW, Gerland B, Sutherland JD. 2009. Synthesis ofactivated pyrimidine ribonucleotides in prebioticallyplausible conditions. Nature 459: 239–242.

Rajamani S, Vlassov A, Benner S, Coombs A, Olasagasti F,Deamer D. 2008. Lipid-assisted synthesis of RNA-likepolymers from mononucleotides. Orig Life Evol B 38:57–74.





Rasmussen H, Kastrup J, Nielsen J, Nielsen J, Nielsen P. 1997.Crystal structure of a peptide nucleic acid (PNA) duplexat 1.7 A resolution. Nat Struct Biol 4: 98–101.

Ricardo A, Carrigan MA, Olcott AN, Benner SA. 2004.Borate minerals stabilize ribose. Science 303: 196–196.

Rice JS, Gao X. 1997. Conformation of formacetal and30-thioformacetal nucleotide linkers and stability of theirantisense RNA.DNA hybrid duplexes. Biochemistry 36:399–411.

Richert C, Roughton AL, Benner SA. 1996. Nonionic ana-logs of RNA with dimethylene sulfone bridges. J AmChem Soc 118: 4518–4531.

Roughton AL, Portmann S, Benner SA, Egli M. 1995.Crystal-structure of a dimethylene sulfone-linked ribodi-nucleotide analog. J Am Chem Soc 117: 7249–7250.

Rozners E, Katkevica D, Stromberg R. 2007. Oligoribonu-cleotide analogues containing a mixed backbone of phos-phodiester and formacetal internucleoside linkages,together with vicinal 20-O-methyl groups. Chembiochem8: 537–545.

Saladino R, Ciambecchini U, Crestini C, Costanzo G, NegriR, Di Mauro E. 2003. One-pot TiO2-catalyzed synthesisof nucleic bases and acyclonucleosides from formamide:Implications for the origin of life. ChemBioChem 4:514–521.

Saladino R, Crestini C, Cicirielloc F, Costanzo G, Di MauroE. 2007. Formamide chemistry and the origin of informa-tional polymers. Chem Biodiv 4: 694–720.

Saladino R, Crestini C, Costanzo G, DiMauro E. 2004.Advances in the prebiotic synthesis of nucleic acids bases:implications for the origin of life. Curr Org Chem 8:1425–1443.

Sanchez R, Ferris J, Orgel LE. 1966. Conditions for purinesynthesis—Did prebiotic synthesis occur at low temper-atures? Science 153: 72–73.

Sanchez RA, Orgel LE. 1970. Studies in prebiotic synthe-sis. V. Synthesis and photoanomerization of pyrimidinenucleosides. J Mol Biol 47: 531–543.

Schlegel MK, Xie XL, Zhang LL, Meggers E. 2009. Insightinto the high duplex stability of the simplified nucleicacid GNA. Angew Chem Int Ed Engl 48: 960–963.

Schmidt J, Eschgfaller B, Benner SA. 2003. A direct synthesisof nucleoside analogs homologated at the 30- and50-positions. Helv Chim Acta 86: 2937–2958.

Schneider KC, Benner SA. 1990. Oligonucleotides contain-ing flexible nucleoside analogs. J Am Chem Soc 112:453–455.

Schoning KU, Scholz P, Guntha S, Wu X, Krishnamurthy R,Eschenmoser A. 2000. Chemical etiology of nucleic acidstructure: The a-threofuranosyl-(30!20) oligonucleo-tide system. Science 290: 1347–1351.

Schrum JP, Ricardo A, Krishnamurthy M, Blain JC, SzostakJW. 2009. Efficient and rapid template-directednucleic acid copying using 20-amino-20,30-dideoxyribo-nucleoside250-phosphorimidazolide monomers. J AmChem Soc 131: 14560–14570.

Schwartz AW. 2006. Phosphorus in prebiotic chemistry.Philos Trans R Soc London B 361: 1743–1749.

Schwartz AW, Chang S. 2002. From Big Bang to primordialplanet: Setting the stage for the origin of life. In Life’s

Origin: The Beginnings of Biological Evolution (Eds.)W. Schopf, 46–77. University of California, Berkeley.

Schweitzer M, Engels JW. 1999. Sequence specific hybridiza-tion properties of methylphosphonate oligodeoxynu-cleotides. J Biomol Struct Dyn 16: 1177–1188.

Seela F, Wei C, Kazmierczuk Z. 1995. Substituent reactivityand tautomerism of isoguanosine and related nucleo-sides. Helv Chim Acta 78: 1843–1854.

Sepiol J, Kazmierczuk Z, Shugar D. 1976. Tautomer-ism of iso-guanosine and solvent-induced keto-enolequilibrium. Zeitschrift fur Naturforschung C-A. J Biosci31: 361–370.

Sheng Y, Bean HD, Mamajanova I, Hud NV, Leszczynski J.2009. A comprehensive investigation of the energeticsof pyrimidine nucleoside formation in a model prebioticreaction. J Am Chem Soc: in press.

Sievers D, von Kiedrowski G. 1994. Self replication of com-plementary nucleotide-based oligomers. Nature 369:221–224.

Sinha S, Kim PH, Switzer C. 2004. 20,50-linked DNA is atemplate for polymerase-directed DNA synthesis. J AmChem Soc 126: 40–41.

Springsteen G, Joyce GF. 2004. Selective derivatization andsequestration of ribose from a prebiotic mix. J AmChem Soc 126: 9578–9583.

Sreenivasachary N, Hickman DT, Sarazin D, Lehn JM. 2006.DyNAs: Constitutional dynamic nucleic acid analogues.Chem-Eur J 12: 8581–8588.

Sreenivasachary N, Lehn JM. 2005. Gelation-driven compo-nent selection in the generation of constitutional dy-namic hydrogels based on guanine-quartet formation.Proc Natl Acad Sci 102: 5938–5943.

Sreenivasachary N, Lehn JM. 2008. Structural selection inG-quartet-based hydrogels and controlled release of bio-active molecules. Chem-Asian J 3: 134–139.

Summerton J, Weller D. 1989. Uncharged morpholino-based polymers having achiral intersubunit linkages.US Patent #5,034,506

Summerton J, Weller D. 1997. Morpholino antisenseoligomers: design, preparation, and properties. AntisenseNucleic Acid Drug Dev 7: 187–195.

Sutherland JD. 2010. Ribonucleotides. Cold Spring HarbPerspect Biol 2: a005439.

Switzer C, Moroney SE, Benner SA. 1989. Enzymatic incor-poration of a new base pair into DNA and RNA. J AmChem Soc 111: 8322–8323.

Ts’o P, Melvin I, Olson A. 1963. Interaction and associationof bases and nucleosides in aqueous solutions. J AmChem Soc 85: 1289–1296.

Tsai CH, Chen JY, Szostak JW. 2007. Enzymatic synthesis ofDNA on glycerol nucleic acid templates without stableduplex formation between product and template. ProcNatl Acad Sci 104: 14598–14603.

Ura Y, Beierle JM, Leman LJ, Orgel LE, Ghadiri MR. 2009.Self-assembling sequence-adaptive peptide nucleic acids.Science 325: 73–77.

Usher DA, McHale AH. 1976. Hydrolytic stability of helicalRNA—Selective advantage for natural 30,50-bond. ProcNatl Acad Sci 73: 1149–1153.





Vasas V, Szathmary E, Santos M. 2010. Lack of evolvability inself-sustaining autocatalytic networks constraints meta-bolism-first scenarios for the origin of life. Proc NatlAcad Sci 107: 1470–1475.

Voegel JJ, Altorfer MM, Benner SA. 1993. The donor-acceptor-acceptor purine analog-transformation of 5-aza-7-deaza-1H-isoguanine (¼ 4-aminoimidazo-[1,2-a]-1,3,5-triazin-2(1H)-one) to 20-deoxy-5-aza-7-deaza-isoguanosine using purine nucleoside phosphorylase.Helv Chim Acta 76: 2061–2069.

Weber AL. 1992. Prebiotic sugar synthesis: hexose andhydroxy acid synthesis from glyceraldehyde catalyzed byiron(III) hydroxide oxide. J Mol Evol 35: 1–6.

Weber AL. 2001. The sugar model: Catalysis by amines andamino acid products. Orig Life Evol B 31: 71–86.

Westheimer FH. 1987. Why Nature chose phosphates.Science 235: 1173–1178.

Wiberg KB, Morgan KM, Maltz H. 1994. Thermochemistryof carbonyl reactions. 6. A study of hydration equilibria. JAm Chem Soc 116: 11067–11077.

Wilds CJ, Wawrzak Z, Krishnamurthy R, Eschenmoser A,Egli M. 2002. Crystal structure of a B-form DNA duplexcontaining (L)-a- threofuranosyl (30! 20) nucleosides:A four-carbon sugar is easily accommodated into thebackbone of DNA. J Am Chem Soc 124: 13716–13721.

Wimberly BT, Brodersen DE, Clemons WM, Morgan-Warren RJ, Carter AP, Vonrhein C, Hartsch T, Ramak-rishnan V. 2000. Structure of the 30S ribosomal subunit.Nature 407: 327–339.

Woese C. 1967. The evolution of the genetic code. InThe genetic code (Eds.), 179–195. Harper & Row,New York.

Yakovchuk P, Protozanova E, Frank-Kamenetskii MD. 2006.Base-stacking and base-pairing contributions into ther-mal stability of the DNA double helix. Nucleic Acids Res34: 564–574.

Zhang LL, Meggers E. 2005. An extremely stable and orthog-onal DNA base pair with a simplified three-carbon back-bone. J Am Chem Soc 127: 74–75.

Zielinski WS, Orgel LE. 1985. Oligomerization of activatedderivatives of 30-amino-30-deoxyguanosine on poly(C)and poly(dG) templates. Nucleic Acids Res 13: 2469–2484.

Zielinski WS, Orgel LE. 1987. Oligoaminonucleoside phos-phoramidates. Oligomerization of dimers of 30-amino-30-deoxy-nucleotides (GC and CG) in aqueous solution.Nucleic Acids Res 15: 1699–1715.

Zubay G, Mui T. 2001. Prebiotic synthesis of nucleosides.Origins Life Evol B 31: 87–102.





12, 20102010; doi: 10.1101/cshperspect.a002196 originally published online MayCold Spring Harb Perspect Biol

Aaron E. Engelhart and Nicholas V. Hud Primitive Genetic Polymers

Subject Collection The Origins of Life

Constructing Partial Models of Cells

et al.Norikazu Ichihashi, Tomoaki Matsuura, Hiroshi Kita,

The Hadean-Archaean EnvironmentNorman H. Sleep

RibonucleotidesJohn D. Sutherland

An Origin of Life on MarsChristopher P. McKay

Characterize Early Life on EarthHow a Tree Can Help−−Deep Phylogeny

RandallEric A. Gaucher, James T. Kratzer and Ryan N.

Primitive Genetic PolymersAaron E. Engelhart and Nicholas V. Hud

Medium to the Early EarthCosmic Carbon Chemistry: From the Interstellar

Pascale Ehrenfreund and Jan Cami

Membrane Transport in Primitive CellsSheref S. Mansy

Origin and Evolution of the RibosomeGeorge E. Fox

The Origins of Cellular LifeJason P. Schrum, Ting F. Zhu and Jack W. Szostak

BiomoleculesPlanetary Organic Chemistry and the Origins of

Kim, et al.Steven A. Benner, Hyo-Joong Kim, Myung-Jung

From Self-Assembled Vesicles to ProtocellsIrene A. Chen and Peter Walde

the Origins of LifeMineral Surfaces, Geochemical Complexities, and

Robert M. Hazen and Dimitri A. Sverjensky

The Origin of Biological HomochiralityDonna G. Blackmond

Historical Development of Origins ResearchAntonio Lazcano

Earth's Earliest AtmospheresKevin Zahnle, Laura Schaefer and Bruce Fegley

http://cshperspectives.cshlp.org/cgi/collection/ For additional articles in this collection, see

Copyright © 2010 Cold Spring Harbor Laboratory Press; all rights reserved


http://cshperspectives.cshlp.org/cgi/collection/

http://cshperspectives.cshlp.org/cgi/collection/


Documents

Primitive Genetic Polymers - CSHL Pcshperspectives.cshlp.org/content/2/12/a002196.full.pdf · mer known as peptide nucleic acid (Nelson et al. 2000), and intermediate proposals, in