Search

Electrochemical proton gradients are the basis of energy transduction in modern cells, and may have played important roles in even the earliest cell-like structures. We have investigated the conditions under which pH gradients are maintained across the membranes of fatty acid vesicles, a model of early cell membranes. We show that pH gradients across such membranes decay rapidly in the presence of alkali-metal cations, but can be maintained in the absence of permeable cations. Under such conditions, when fatty acid vesicles grow through the incorporation of additional fatty acid, a transmembrane pH gradient is spontaneously generated. The formation of this pH gradient captures some of the energy released during membrane growth, but also opposes and limits further membrane area increase. The coupling of membrane growth to energy storage could have provided a growth advantage to early cells, once the membrane composition had evolved to allow the maintenance of stable pH gradients.

Very little is known about the distribution of functional DNA, RNA, and protein molecules in sequence space. The question of how the number and complexity of distinct solutions to a particular biochemical problem varies with activity is an important aspect of this general problem. Here we present a comparison of the structures and activities of eleven distinct GTP-binding RNAs (aptamers). By experimentally measuring the amount of information required to specify each optimal binding structure, we show that defining a structure capable of 10-fold tighter binding requires approximately 10 additional bits of information. This increase in information content is equivalent to specifying the identity of five additional nucleotide positions and corresponds to an approximately 1000-fold decrease in abundance in a sample of random sequences. We observe a similar relationship between structural complexity and activity in a comparison of two catalytic RNAs (ribozyme ligases), raising the possibility of a general relationship between the complexity of RNA structures and their functional activity. Describing how information varies with activity in other heteropolymers, both biological and synthetic, may lead to an objective means of comparing their functional properties. This approach could be useful in predicting the functional utility of novel heteropolymers.

Small changes in target specificity can sometimes be achieved, without changing aptamer structure, through mutation of a few bases. Larger changes in target geometry or chemistry may require more radical changes in an aptamer. In the latter case, it is unknown whether structural and functional solutions can still be found in the region of sequence space close to the original aptamer. To investigate these questions, we designed an in vitro selection experiment aimed at evolving specificity of an ATP aptamer. The ATP aptamer makes contacts with both the nucleobase and the sugar. We used an affinity matrix in which GTP was immobilized through the sugar, thus requiring extensive changes in or loss of sugar contact, as well as changes in recognition of the nucleobase. After just five rounds of selection, the pool was dominated by new aptamers falling into three major classes, each with secondary structures distinct from that of the ATP aptamer. The average sequence identity between the original aptamer and new aptamers is 76%. Most of the mutations appear to play roles either in disrupting the original secondary structure or in forming the new secondary structure or the new recognition loops. Our results show that there are novel structures that recognize a significantly different ligand in the region of sequence space close to the ATP aptamer. These examples of the emergence of novel functions and structures from an RNA molecule with a defined specificity and fold provide a new perspective on the evolutionary flexibility and adaptability of RNA.

The clay montmorillonite is known to catalyze the polymerization of RNA from activated ribonucleotides. Here we report that montmorillonite accelerates the spontaneous conversion of fatty acid micelles into vesicles. Clay particles often become encapsulated in these vesicles, thus providing a pathway for the prebiotic encapsulation of catalytically active surfaces within membrane vesicles. In addition, RNA adsorbed to clay can be encapsulated within vesicles. Once formed, such vesicles can grow by incorporating fatty acid supplied as micelles and can divide without dilution of their contents by extrusion through small pores. These processes mediate vesicle replication through cycles of growth and division. The formation, growth, and division of the earliest cells may have occurred in response to similar interactions with mineral particles and inputs of material and energy.

Threose nucleic acid (TNA), which has a repeat unit one atom shorter than that of DNA, is capable of Watson-Crick base pairing with DNA, RNA, and TNA. Because of its chemical simplicity, TNA is considered to be a possible progenitor of RNA. As an initial step toward developing the molecular tools necessary to investigate the functional capabilities of TNA by in vitro selection, we have screened a variety of DNA polymerases for TNA synthesis activity on a DNA template. We wish to report that several polymerases show surprisingly good ability to synthesize TNA using alpha-l-threofuranosyl thymidine-3'-triphosphate as a substrate.

An analogue of uridine triphosphate containing a cationic functional group was incorporated into a degenerate RNA library by enzymatic polymerization. In vitro selection experiments using this library yielded a novel receptor that binds ATP under physiological pH and salt conditions in a manner completely dependent on the presence of the cationic functionality. The consensus sequence and a secondary structure model for the ATP binding site were obtained by the analysis of functional sequences selected from a partially randomized pool based on the minimal parental sequence. Mutational studies of this receptor indicated that several of the modified uridines are critical for ATP binding. Analysis of the binding of ATP analogues revealed that the modified RNA receptor makes numerous contacts with ATP, including interactions with the triphosphate group. In contrast, the aptamer repeatedly isolated from natural RNA libraries does not interact with the triphosphate group of ATP. The incorporation of a cationic amine into nucleic acids clearly allows novel interactions to occur during the molecular recognition of ligands, which carries interesting implications for the RNA world hypothesis. In addition, new materials generated from such functionalized nucleic acids could be useful tools in research and diagnostics.

We report here a unique approach to selectively label and detect specific RNA in an RNA mixture (without separation or purification) using DNA polymerase, dNTP labels, and a short synthetic DNA template complementary to the 3(')-terminus of the RNA. The detection sensitivity is high, at attomole level (10-18 mole). The selective principle was demonstrated by individually labeling and detecting RNAs in a RNA mixture when different templates were provided. By taking advantage of the template-directed selectivity, poly(A) tail-containing mRNA in total RNA was detected and labeled at the 3(')-terminal on a poly(T) template. Nonradioactive labels, such as fluorophore and antigen labels, may also be used; this method can be applied in methodology for direct detection and quantification of viral RNAs.

TNA, or threose nucleic acid, is capable of Watson-Crick base pairing with DNA, RNA, and TNA; coupled with its chemical simplicity, this suggests that TNA is a possible progenitor of RNA. As an initial step toward developing the molecular tools necessary to investigate the functional capabilities of TNA by in vitro selection, we have screened a variety of DNA polymerases for activity on a TNA template. We report that despite having a repeating unit that is one atom shorter than that of DNA, several polymerases showed surprisingly good ability to copy limited stretches of TNA.

Aptamers, RNA sequences that bind to target ligands, are typically isolated by in vitro selection from RNA libraries containing completely random sequences. To see whether higher-affinity aptamers can be isolated from partially structured RNA libraries, we selected for aptamers that bind GTP, starting from a mixture of fully random and partially structured libraries. Because stem-loops are common motifs in previously characterized aptamers, we designed the partially structured library to contain a centrally located stable stem-loop. We used an off-rate selection protocol designed to maximize the enrichment of high-affinity aptamers. The selection produced a surprisingly large number of distinct sequence motifs and secondary structures, including seven different aptamers with K(d)s ranging from 500 to 25 nanomolar. The engineered stem-loop was present in the three highest affinity aptamers, and in 12 of 13 independent isolates with a single consensus sequence, suggesting that its inclusion increased the abundance of high-affinity aptamers in the starting pool.

The mRNA display approach to in vitro protein selection is based upon the puromycin-mediated formation of a covalent bond between an mRNA and its gene product. This technique can be used to identify peptide sequences involved in macromolecular recognition, including those identical or homologous to natural ligand epitopes. To demonstrate this approach, we determined the peptide sequences recognized by the trypsin active site, and by the anti-c-Myc antibody, 9E10. Here we describe the use of two peptide libraries of different diversities, one a constrained library based on the trypsin inhibitor EETI-II, where only the six residues in the first loop were randomized (6.4 x 10(7) possible sequences, 6.0 x 10(11) sequences in the library), the other a linear-peptide library with 27 randomized amino acids (1.3 x 10(35) possible sequences, 2 x 10(13) sequences in the library). The constrained library was screened against the natural target of wild-type EETI, bovine trypsin, and the linear library was screened against the anti-c-myc antibody, 9E10. The analysis of selected sequences revealed minimal consensus sequences of PR(I,L,V)L for the first loop of EETI-II and LISE for the 9E10 epitope. The wild-type sequences, PRILMR for the first loop of EETI-II and QKLISE for the 9E10 epitope, were selected with the highest frequency, and in each case the complete wild-type epitope was selected from the library., (Copyright 2002 John Wiley & Sons, Ltd.)

We describe the use of the SBP-tag, a new streptavidin-binding peptide, for both the one-step purification and the detection of recombinant proteins. The SBP-tag sequence is 38 amino acids long and binds to streptavidin with an equilibrium dissociation constant of 2.5 nM. We demonstrate that a single-step purification of SBP-tagged proteins from bacterial extract yields samples that are more pure than those purified using maltose-binding protein or the His-tag. The capacity of the immobilized streptavidin used to purify SBP-tagged proteins is about 0.5 mg per milliliter of matrix, which is high enough to isolate large quantities of proteins for further study. Also, the elution conditions from the streptavidin column are very mild and specific, consisting of the wash buffer plus biotin. This combination of high-affinity, high-yield, mild elution conditions, and simplicity of use makes the SBP-tag suitable for high-throughput protein expression/purification procedures, including robotically manipulated protocols with microtiter plates. Additionally, the SBP-tag can be used for detection since a wide variety of streptavidin-conjugated fluorescent and enzymatic systems are commercially available. We also present a new, rapid, method for the measurement of protein-protein, protein-peptide, or protein-small molecule equilibrium dissociation constants that require as little as 1 fmol of labeled protein. We call this method the spin-filter binding inhibition assay., (Copyright 2001 Elsevier Science.)

The hammerhead ribozyme was originally discovered in a group of RNAs associated with plant viruses, and has subsequently been identified in the genome of the newt (Notophthalamus viridescens), in schistosomes and in cave crickets (Dolichopoda species). The sporadic occurrence of this self-cleaving RNA motif in highly divergent organisms could be a consequence of the very early evolution of the hammerhead ribozyme, with all extant examples being descended from a single ancestral progenitor. Alternatively, the hammerhead ribozyme may have evolved independently many times. To better understand the observed distribution of hammerhead ribozymes, we used in vitro selection to search an unbiased sample of random sequences for comparably active self-cleaving motifs. Here we show that, under near-physiological conditions, the hammerhead ribozyme motif is the most common (and thus the simplest) RNA structure capable of self-cleavage at biologically observed rates. Our results suggest that the evolutionary process may have been channelled, in nature as in the laboratory, towards repeated selection of the simplest solution to a biochemical problem.

Functional primordial proteins presumably originated from random sequences, but it is not known how frequently functional, or even folded, proteins occur in collections of random sequences. Here we have used in vitro selection of messenger RNA displayed proteins, in which each protein is covalently linked through its carboxy terminus to the 3' end of its encoding mRNA, to sample a large number of distinct random sequences. Starting from a library of 6 x 1012 proteins each containing 80 contiguous random amino acids, we selected functional proteins by enriching for those that bind to ATP. This selection yielded four new ATP-binding proteins that appear to be unrelated to each other or to anything found in the current databases of biological proteins. The frequency of occurrence of functional proteins in random-sequence libraries appears to be similar to that observed for equivalent RNA libraries.

We report the use of "mRNA display," an in vitro selection technique, to identify peptide aptamers to a protein target. mRNA display allows for the preparation of polypeptide libraries with far greater complexity than is possible with phage display. Starting with a library of approximately 10(13) random peptides, 20 different aptamers to streptavidin were obtained, with dissociation constants as low as 5 nM. These aptamers function without the aid of disulfide bridges or engineered scaffolds, yet possess affinities comparable to those for monoclonal antibody-antigen complexes. The aptamers bind streptavidin with three to four orders of magnitude higher affinity than those isolated previously by phage display from lower complexity libraries of shorter random peptides. Like previously isolated peptides, they contain an HPQ consensus motif. This study shows that, given sufficient length and diversity, high-affinity aptamers can be obtained even from random nonconstrained peptide libraries. By engineering structural constraints into these ultrahigh complexity peptide libraries, it may be possible to produce binding agents with subnanomolar binding constants.

Two analog uridine triphosphates tethering additional functionality, one a primary amino group and the second a mercapto group, were prepared and tested for their compatibility with in vitro RNA selection procedures. 5-(3-Aminopropyl)uridine triphosphate (UNH(2)) as a uridine substitute was a more effective substrate for T7 RNA polymerase than 5-(2-mercaptoethyl)uridine triphosphate (USH). However, both functioned in transcription assays of 100 nt templates to generate RNA transcripts in quantities sufficient to initiate RNA selection procedures. Transcription of RNA pools with T7 RNA polymerase and UNH(2) or USH occurred with efficiencies of 43 and 29%, respectively, of the values obtained for native UTP transcription. In addition, the transcribed RNA containing roughly 25% UNH(2) residues exhibited better substrate properties for SuperScript(TM) II RNase H reverse transcriptase than did RNA transcripts containing approximately 25% of the USH analog. With either analog, both transcription and reverse transcription proceeded with high fidelity for insertion of the analog residue.

We present a method that can significantly increase the complexity of protein libraries used for in vitro or in vivo protein selection experiments. Protein libraries are often encoded by chemically synthesized DNA, in which part of the open reading frame is randomized. There are, however, major obstacles associated with the chemical synthesis of long open reading frames, especially those containing random segments. Insertions and deletions that occur during chemical synthesis cause frameshifts, and stop codons in the random region will cause premature termination. These problems can together greatly reduce the number of full-length synthetic genes in the library. We describe a strategy in which smaller segments of the synthetic open reading frame are selected in vitro using mRNA display for the absence of frameshifts and stop codons. These smaller segments are then ligated together to form combinatorial libraries of long uninterrupted open reading frames. This process can increase the number of full-length open reading frames in libraries by up to two orders of magnitude, resulting in protein libraries with complexities of greater than 10(13). We have used this methodology to generate three types of displayed protein library: a completely random sequence library, a library of concatemerized oligopeptide cassettes with a propensity for forming amphipathic alpha-helical or beta-strand structures, and a library based on one of the most common enzymatic scaffolds, the alpha/beta (TIM) barrel., (Copyright 2000 Academic Press.)

The RNA world hypothesis implies that coded protein synthesis evolved from a set of ribozyme catalyzed acyl-transfer reactions, including those of aminoacyl-tRNA synthetase ribozymes. We report here that a bifunctional ribozyme generated by directed in vitro evolution can specifically recognize an activated glutaminyl ester and aminoacylate a targeted tRNA, via a covalent aminoacyl-ribozyme intermediate. The ribozyme consists of two distinct catalytic domains; one domain recognizes the glutamine substrate and self-aminoacylates its own 5'-hydroxyl group, and the other recognizes the tRNA and transfers the aminoacyl group to the 3'-end. The interaction of these domains results in a unique pseudoknotted structure, and the ribozyme requires a change in conformation to perform the sequential aminoacylation reactions. Our result supports the idea that aminoacyl-tRNA synthetase ribozymes could have played a key role in the evolution of the genetic code and RNA-directed translation.

In vitro selection allows rare functional RNA or DNA molecules to be isolated from pools of over 10(15) different sequences. This approach has been used to identify RNA and DNA ligands for numerous small molecules, and recent three-dimensional structure solutions have revealed the basis for ligand recognition in several cases. By selecting high-affinity and -specificity nucleic acid ligands for proteins, promising new therapeutic and diagnostic reagents have been identified. Selection experiments have also been carried out to identify ribozymes that catalyze a variety of chemical transformations, including RNA cleavage, ligation, and synthesis, as well as alkylation and acyl-transfer reactions and N-glycosidic and peptide bond formation. The existence of such RNA enzymes supports the notion that ribozymes could have directed a primitive metabolism before the evolution of protein synthesis. New in vitro protein selection techniques should allow for a direct comparison of the frequency of ligand binding and catalytic structures in pools of random sequence polynucleotides versus polypeptides.

Background: In vitro selection experiments with pools of random-sequence nucleic acids have been used extensively to isolate molecules capable of binding specific ligands and catalyzing self-modification reactions., Results: In vitro selection from a random pool of single-stranded DNAs has been used to isolate molecules capable of recognizing the fluorophore sulforhodamine B with high affinity. When assayed for the ability to promote an oxidation reaction using the reduced form of a related fluorophore, dihydrotetramethylrosamine, a number of selected clones show low levels of catalytic activity. Chemical modification and site-directed mutagenesis experiments have been used to probe the structural requirements for fluorophore binding. The aptamer recognizes its ligand with relatively high affinity and is also capable of binding related molecules that share extended aromatic rings and negatively charged functional groups., Conclusions: A guanosine-rich single-stranded DNA is capable of binding fluorophores with relatively high affinity and of weakly promoting a multiple-turnover reaction. A simple motif consisting of a three-tiered G-quartet stacked upon a standard Watson-Crick duplex appears to be responsible for this activity. The corresponding sequence might provide a useful starting point for the evolution of novel, improved deoxyribozymes that generate fluorescent signals by promoting multiple-turnover reactions.

Most studies of the roles of catalytic metal ions in ribozymes have focused on inner-sphere coordination of the divalent metal ions to the substrate or ribozyme. However, divalent metal ions are strongly hydrated in water, and some proteinenzymes, such as Escherichia coli RNase H and exonuclease III, are known to use metal cofactors in their fully hydrated form [Duffy, T. H., and Nowak, T. (1985) Biochemistry 24, 1152-1160; Jou, R., and Cowan, J. A. (1991) J. Am. Chem. Soc. 113, 6685-6686]. It is therefore important to consider the possibility of outer-sphere coordination of catalytic metal ions in ribozymes. We have used an exchange-inert metal complex, cobalt hexaammine, to show that the catalytic metal ion in an acyl-transferase ribozyme acts through outer-sphere coordination. Our studies provide an example of a fully hydrated Mg2+ ion that plays an essential role in ribozyme catalysis. Kinetic studies of wild-type and mutant ribozymes suggest that a pair of tandem G:U wobble base pairs adjacent to the reactive center constitute the metal-binding site. This result is consistent with recent crystallographic studies [Cate, J. H., and Doudna, J. A. (1996) Structure 4, 1221-1229; Cate, J. H., Gooding, A. R., Podell, E., Zhou, K., Golden, B. L., Kundrot, C. E., Cech, T. R., and Doudna, J. A. (1996) Science 273, 1678-1685; Cate, J. H., Hanna, R. L., and Doudna, J. A. (1997) Nat. Struct. Biol. 4, 553-558] showing that tandem wobble base pairs are good binding sites for metal hexaammines. We propose a model in which the catalytic metal ion is bound in the major groove of the tandem wobble base pairs, is precisely positioned by the ribozyme within the active site, and stabilizes the developing oxyanion in the transition state. Our results may have significant implications for understanding the mechanism of protein synthesis [Noller, H. F., Hoffarth, V., and Zimniak, L. (1992) Science 256, 1416-1419].

We have previously isolated, by in vitro selection, an acyl-transferase ribozyme that is capable of transferring a biotinylated methionyl group from the 3' end of a hexanucleotide substrate to its own 5'-hydroxyl. Comparison of the sequences of a family of evolved derivatives of this ribozyme allowed us to generate a model of the secondary structure of the ribozyme. The predicted secondary structure was extensively tested and confirmed by single-mutant and compensatory double-mutant analyses. The role of the template domain in aligning the acyl-donor oligonucleotide and acyl-acceptor region of the ribozyme was confirmed in a similar manner. The significance of different domains of the ribozyme structure and the importance of two tandem G:U wobble base pairs in the template domain were studied by kinetic characterization of mutant ribozymes. The wobble base pairs contribute to the catalytic rate enhancement, but only in the context of the complete ribozyme; the ribozyme in turn alters the metal binding properties of this site. Competitive inhibition experiments with unacylated substrate oligonucleotide are consistent with the ribozyme acting to stabilize substrate binding to the template, while negative interactions with the aminoacyl portion of the substrate destabilize binding.

Background: In vitro selection has been shown previously to be a powerful method for isolating nucleic acids with specific ligand-binding functions ('aptamers'). Given this capacity, we have sought to isolate RNA motifs that can confer fluorescent labeling to tagged RNA transcripts, potentially allowing in vivo detection and in vitro spectroscopic analysis of RNAs., Results: Two aptamers that recognize the fluorophore sulforhodamine B were isolated by the in vitro selection process. An unusually large motif of approximately 60 nucleotides is responsible for binding in one RNA (SRB-2). This motif consists of a three-way helical junction with two large, highly conserved unpaired regions. Phosphorothioate mapping with an iodoacetamide-tagged form of the ligand shows that these two regions make close contacts with the fluorophore, suggesting that the two loops combine to form separate halves of a binding pocket. The aptamer binds the fluorophore with high affinity, recognizing both the planar aromatic ring system and a negatively charged sulfonate, a rare example of anion recognition by RNA. An aptamer (FB-1) that specifically binds fluorescein has also been isolated by mutagenesis of a sulforhodamine aptamer followed by re-selection. In a simple in vitro test, SRB-2 and FB-1 have been shown to discriminate between sulforhodamine and fluorescein, specifically localizing each fluorophore to beads tagged with the corresponding aptamer., Conclusions: In addition to serving as a model system for understanding the basis of RNA folding and function, these experiments demonstrate potential applications for the aptamers in transcript double labeling or fluorescence resonance energy transfer studies.

Covalent fusions between an mRNA and the peptide or protein that it encodes can be generated by in vitro translation of synthetic mRNAs that carry puromycin, a peptidyl acceptor antibiotic, at their 3' end. The stable linkage between the informational (nucleic acid) and functional (peptide) domains of the resulting joint molecules allows a specific mRNA to be enriched from a complex mixture of mRNAs based on the properties of its encoded peptide. Fusions between a synthetic mRNA and its encoded myc epitope peptide have been enriched from a pool of random sequence mRNA-peptide fusions by immunoprecipitation. Covalent RNA-peptide fusions should provide an additional route to the in vitro selection and directed evolution of proteins.

The solution structure of the ATP-binding RNA aptamer has recently been determined by NMR spectroscopy. The three-dimensional fold of the molecule is determined to a large extent by stacking and hydrogen bond interactions. In the course of the structure determination it was discovered that several highly conserved nucleotides in the binding pocket can be substituted while retaining binding under NMR conditions. These surprising findings allow a closer look at the interactions that determine stability and specificity of the aptamer as well as local structural features of the molecule. The binding properties of ATP binder mutants and modified ligand molecules are explored using NMR spectroscopy, column binding studies and molecular modeling. We present additional evidence and new insights regarding the network of hydrogen bonds that defines the structure and determines stability and specificity of the aptamer., (Copyright 1997 Academic Press Limited.)

Background: The protein enzymes RNA ligase and DNA ligase catalyze the ligation of nucleic acids via an adenosine-5'-5'-pyrophosphate 'capped' RNA or DNA intermediate. The activation of nucleic acid substrates by adenosine 5'-monophosphate (AMP) may be a vestige of 'RNA world' catalysis. AMP-activated ligation seems ideally suited for catalysis by ribozymes (RNA enzymes), because an RNA motif capable of tightly and specifically binding AMP has previously been isolated., Results: We used in vitro selection and directed evolution to explore the ability of ribozymes to catalyze the template-directed ligation of AMP-activated RNAs. We subjected a pool of 10(15) RNA molecules, each consisting of long random sequences flanking a mutagenized adenosine triphosphate (ATP) aptamer, to ten rounds of in vitro selection, including three rounds involving mutagenic polymerase chain reaction. Selection was for the ligation of an oligonucleotide to the 5'-capped active pool RNA species. Many different ligase ribozymes were isolated; these ribozymes had rates of reaction up to 0.4 ligations per hour, corresponding to rate accelerations of approximately 5 x10(5) over the templated, but otherwise uncatalyzed, background reaction rate. Three characterized ribozymes catalyzed the formation of 3'-5'-phosphodiester bonds and were highly specific for activation by AMP at the ligation site., Conclusions: The existence of a new class of ligase ribozymes is consistent with the hypothesis that the unusual mechanism of the biological ligases resulted from a conservation of mechanism during an evolutionary replacement of a primordial ribozyme ligase by a more modern protein enzyme. The newly isolated ligase ribozymes may also provide a starting point for the isolation of ribozymes that catalyze the polymerization of AMP-activated oligonucleotides or mononucleotides, which might have been the prebiotic analogs of nucleoside triphosphates.

We describe a simple method for 3'-end labeling RNAs of known sequence. A short DNA template is designed to anneal to the 3'-end of the RNA, with a two nucleotide 5' overhang of 3'-TA-5', 3'-TG-5' or 3'-TC-5'. The Klenow fragment of DNA polymerase I can then cleanly and efficiently extend the 3'-end of the RNA by the incorporation of a single alpha-32P-labeled dATP residue. This method can be used to label one RNA in a mixture of RNAs, or to label 5'-blocked RNAs such as mRNA.

The RNA world hypothesis is founded on the idea of an RNA replicase, or self-replicating RNA molecule, and presupposes the later emergence of ribozymes capable of catalyzing the synthesis of peptides. The recent demonstrations of ribozyme-catalyzed template-directed primer extension, and of ribozyme-catalyzed amide bond synthesis, confirm the plausibility of the RNA world, and highlight the steps that remain to be demonstrated in the laboratory.

The 'RNA world' hypothesis proposes an early stage in the evolution of life in which both genomic and catalytic functions were fulfilled by RNA. The evolution of RNA-catalysed protein synthesis would have been a necessary step in the transition from such an RNA world to modern protein-dominated biology. For this to have been possible, RNA must be capable of catalysing amide-bond formation using acylated carrier RNA substrates as amino-acid donors. We have used in vitro selection and evolution to isolate ribozymes with acyl transferase activity from a pool of random RNA sequences. One of these acyl transferases with a 5'-amino group transfers an amino acid to itself in a reaction that we propose to be analogous to peptidyl transfer on the ribosome.

We have found that nonenzymatic, template-directed ligation reactions of oligoribonucleotides display high selectivity for the formation of 3'-5' rather than 2'-5' phosphodiester bonds. Formation of the 3'-5'-linked product is favored regardless of the metal ion catalyst or the leaving group, and for several different ligation junction sequences. The degree of selectivity depends on the leaving group: the ratio of 3'-5'- to 2'-5'-linked products was 10-15:1 when the 5'-phosphate was activated as the imidazolide, and 60-80:1 when the 5'-phosphate was activated by the formation of a 5'-triphosphate. Comparison of oligonucleotide ligation reactions with previously characterized single nucleotide primer extension reactions suggests that the strong preference for 3'-5'-linkages in oligonucleotide ligation is primarily due to occurence of ligation within the context of an extended Watston-Crick duplex. The ability of RNA to correctly self-assemble by template-directed ligation is an intrinsic consequence of its chemical structure and need not be imposed by an external catalyst (i.e., an enzyme polymerase); RNA therefore provides a reasonable structural basis for a self-replicating system in a prebiological world.

The role of divalent cations in the mechanism of pyrophosphate-activated, template-directed oligoribonucleotide ligation has been investigated. The dependence of the reaction rate on Mg2+ concentration suggests a kinetic scheme in which a Mg2+ ion must bind before ligation can proceed. Mn2+, Ca2+, Sr2+, and Ba2+ can also catalyze the reaction. Although Pb2+ and Zn2+ do not catalyze the reaction in the absence of other divalent ions, they significantly modulate the reaction rate when added in the presence of Mg2+, with Pb2+ stimulating the reaction (up to 65-fold) and Zn2+ inhibiting the reaction. The logarithm of the ligation rate increases linearly, with slope of 0.95, as a function of pH, indicating that the reaction involves a single critical deprotonation step. The ligation rates observed with the different divalent metal ion catalysts (Mn2+ > Mg2+ > Ca2+ > Sr2+ = Ba2+) vary inversely with the pKa values of their bound water molecules. The pH profile and these relative ligation rates suggest a mechanism in which a metal-bound hydroxide ion located near the ligation junction promotes catalysis, most likely by deprotonation of the hydroxl nucleophile. The effects of changing either the leaving group or the attacking hydroxyl, together with the large delta S(++) value for oligonucleotide ligation (about -20 eu), are consistent with an associative transition state.

In Tom Stoppard's famous play [Rosencrantz and Guildenstern are Dead], the ill-fated heroes toss a coin 101 times. The first 100 times they do so the coin lands heads up. The chance of this happening is approximately 1 in 10(30), a sequence of events so rare that one might argue that it could only happen in such a delightful fiction. Similarly rare events, however, may underlie the origins of biological catalysis. What is the probability that an RNA, DNA, or protein molecule of a given random sequence will display a particular catalytic activity? The answer to this question determines whether a collection of such sequences, such as might result from prebiotic chemistry on the early earth, is extremely likely or unlikely to contain catalytically active molecules, and hence whether the origin of life itself is a virtually inevitable consequence of chemical laws or merely a bizarre fluke. The fact that a priori estimates of this probability, given by otherwise informed chemists and biologists, ranged from 10(-5) to 10(-50), inspired us to begin to address the question experimentally. As it turns out, the chance that a given random sequence RNA molecule will be able to catalyze an RNA polymerase-like phosphoryl transfer reaction is close to 1 in 10(13), rare enough, to be sure, but nevertheless in a range that is comfortably accessible by experiment. It is the purpose of this Account to describe the recent advances in combinatorial biochemistry that have made it possible for us to explore the abundance and diversity of catalysts existing in nucleic acid sequence space.

We have previously isolated a series of ribozymes with polynucleotide kinase activity [Lorsch, J.R., & Szostak, J.W. (1994) Nature 371, 31-36]. In order to learn how such newly evolved RNAs effect catalysis, we have determined a number of the kinetic and thermodynamic parameters for the reaction catalyzed by one of these ribozymes. This ribozyme, a class I polynucleotide kinase, catalyzes the transfer of the gamma-(thio)phosphate from ATP(-gamma S) to the 5'-hydroxyl of a 7-mer oligoribonucleotide. The kcat for the reaction with ATP-gamma S is 0.17 min-1 with a Km of approximately 3 mM. The Km for the oligoribonucleotide substrate 5'-HO-GGAACCU-3' is 2 microM, the same as the Kd for this substrate in the presence or absence of ATP-gamma S. Neither the binding of substrates nor the release of products is the rate-limiting step of the reaction. The binding of substrates and release of products appear to occur in a random fashion, with no synergy of binding between the ATP(-gamma S) and oligoribonucleotide substrates. The ribozyme binds the oligoribonucleotide substrate no more strongly than would be expected for the formation of a simple RNA-RNA duplex, suggesting that there are no tertiary contacts between the ribozyme and the RNA substrate. The oligoribonucleotide substrate binding site has been located, and the sequence specificity of the ribozyme could be altered by mutating this binding site. The ribozyme is specific for adenosine triphosphate substrates; GTP-gamma S reacts approximately 650-fold slower than ATP-gamma S. With ATP as the substrate, the Kms remain unchanged, but kcat decreases by a factor of 50, consistent with a rate-limiting chemical step occurring through a dissociative transition state. The pH independence (from pH 5.5 to 8.5) of kcat/Km and of the rate constant for the conversion of the ternary substrate complex into the ternary products complex is also consistent with a dissociative phosphoryl transfer mechanism. These results suggest that this newly evolved catalyst operates in a relatively simple manner, with independent substrate binding sites and without changing the mechanism of the underlying chemical reaction.

Avian myeloblastosis virus and Maloney murine leukemia virus RNase H-reverse transcriptases pause when they encounter a 2'-5' linkage or a 2'-thiophosphate in their template RNAs, but eventually read through these backbone modifications. Both reverse transcriptases pause after the 2'-5' linkage but before the 2'-thiophosphate. These results suggest that in the absence of precise information concerning the behavior of a given reverse transcriptase with respect to a particular lesion or modification, caution should be exercised in the interpretation of primer extension data that is being used to determine the existence of, or map the position of, a crosslink, site of chemical modification or non-standard linkage in an RNA template.

Seven families of RNA ligases, previously isolated from random RNA sequences, fall into three classes on the basis of secondary structure and regiospecificity of ligation. Two of the three classes of ribozymes have been engineered to act as true enzymes, catalyzing the multiple-turnover transformation of substrates into products. The most complex of these ribozymes has a minimal catalytic domain of 93 nucleotides. An optimized version of this ribozyme has a kcat exceeding one per second, a value far greater than that of most natural RNA catalysts and approaching that of comparable protein enzymes. The fact that such a large and complex ligase emerged from a very limited sampling of sequence space implies the existence of a large number of distinct RNA structures of equivalent complexity and activity.

Single-stranded DNA can fold into well-defined sequence-dependent tertiary structures that specifically bind a variety of target molecules, raising the possibility that some folded single-stranded DNAs might exhibit catalytic activities similar to those of ribozymes and protein enzymes. Derivatives of the hammerhead ribozyme that contain a majority of deoxyribonucleotides retain the ability to cleave RNA, and a 'deoxyribozyme' was generated by leaving all essential ribonucleotides of the hammerhead on the RNA 'substrate'. Recently in vitro selection has been used to isolate a DNA sequence that shows Pb(2+)-dependent RNA-cleaving activity. Here we report the isolation by in vitro selection of a small single-stranded DNA that is a Zn2+/Cu(2+)-dependent metalloenzyme. The enzyme catalyses the formation of a new phosphodiester bond by the condensation of the 5'-hydroxyl of one oligodeoxynucleotide and a 3'-phosphorimidazolide on another oligodeoxynucleotide, and shows multiple turnover ligation.

Background: Many new ribozymes, including sequence-specific nucleases, ligases and kinases, have been isolated by in vitro selection from large pools of random-sequence RNAs. We are attempting to use in vitro selection to isolate new ribozymes that have, or can be evolved to have, RNA polymerase-like activities. As phosphorimidazolide-activated nucleosides are extensively used to study non-enzymatic RNA replication, we wished to select for a ribozyme that would accelerate the template-directed ligation of 5'-phosphorimidazolide-activated oligonucleotides., Results: Ribozymes selected to perform the desired template-directed ligation reaction instead ligated themselves to the activated substrate oligonucleotide via their 5'-triphosphate, generating a 5'-5' P1,P4-tetraphosphate linkage. Deletion analysis of one of the selected sequences revealed that a 54-nucleotide RNA retained activity; this small ribozyme folds into a pseudoknot secondary structure with an internal binding site for the substrate oligonucleotide. The ribozyme can also synthesize 5'-5' triphosphate and 5'-5' pyrophosphate linkages., Conclusions: The emergence of ribozymes that accelerate an unexpected 5'-5' ligation reaction from a selection designed to yield template-dependent 3'-5' ligases suggests that it may be much easier for RNA to catalyze the synthesis of 5'-5' linkages than 3'-5' linkages. 5'-5' linkages are found in a variety of contexts in present-day biology. The ribozyme-catalyzed synthesis of such linkages raises the possibility that these 5'-5' linkages originated in the biochemistry of the RNA world.

RNA enzymes are postulated to have catalysed all chemical reactions in the earliest living cells. This idea is now investigated in a search for alkyl transferases from a pool of random sequence RNAs. Selection for self-biotinylation yields a transfer RNA-like ribozyme that efficiently catalyses carbon-nitrogen bond formation. Ribozymes can thus promote reactions other than those involving the RNA sugar-phosphate backbone, suggesting that RNA may be capable of a broad range of catalytic activities.

We describe the complete chemical synthesis of a ribozyme that catalyzes template-directed oligonucleotide ligation. The specific activity of the synthetic ribozyme is nearly identical to that of the same enzyme generated by in vitro transcription with T7 RNA polymerase. The ribozyme is derived from a group I intron and consists of three RNA fragments of 36, 43, and 59 nt that self-assemble to form a catalytically active complex. We have site-specifically substituted ribonucleotide analogs into this enzyme and have identified two 2'-hydroxyl groups that are required for full catalytic activity. In contrast, neither the 2'-hydroxyl nor the exocyclic amino group of the conserved guanosine in the guanosine binding site is necessary for catalysis. By allowing the ribozyme to be modified as easily as its substrates, this synthetic ribozyme system should be useful for testing specific hypotheses concerning ribozyme-substrate interactions and tertiary interactions within the ribozyme.

RNA molecules that specifically bind riboflavin (Rb) and beta-nicotinamide mononucleotide (NMN) have been isolated by in vitro selection. A simple structural motif containing intramolecular G-quartets was found to bind tightly to oxidized riboflavin (Kd = 1-5 micromolar). DNA versions of the consensus sequence also bind, but with weaker affinity. DMS protection experiments show that the quartet structure of these aptamers is stabilized by interaction with the flavin. As a measure of their redox specificity, the aptamers do not show significant differential binding between oxidized and reduced forms of a 5-deazariboflavin derivative that is a close structural analog of riboflavin. In contrast to the lack of redox specificity of the riboflavin aptamers, RNAs selected for binding to the nicotinamide portion of NAD discriminate between NAD and NADH in solution by over an order of magnitude. A mutagenized pool based on one of the NMN aptamer sequences was used to reselect for NMN binding. Comparison of the reselected sequences led to the identification of the binding region of the aptamer. A complex secondary structure containing two interacting stem-loops is proposed for the minimal NMN-binding RNA. The same mutagenized pool was used to select for increased discrimination between NMN and NMNH. From these reselected sequences, a mutation within the binding region was identified that increases specificity for NMN. These experiments show that RNA can bind these cofactors with low micromolar affinity and, in the case of nicotinamide cofactors, can discriminate between the two redox states. These cofactor binding motifs may provide a framework for generating new ribozymes that catalyze redox reactions similar to those found in basic metabolic pathways.

We have used in vitro selection to isolate adenosine/ATP-binding DNA sequences from a pool of approximately 2 x 10(14) different random-sequence single-stranded DNA molecules. One of these aptamers has been characterized and binds adenosine in solution with a dissociation constant of 6 +/- 3 microM. Experiments with ATP analogs indicate that functional groups on both the base and the sugar of ATP are involved in the ligand/aptamer interaction. The binding domain of this aptamer was localized to a 42 base sequence by deletion analysis. A pool of mutagenized versions of this sequence was then synthesized and screened for functional adenosine binding sequences; comparison of the selected variants revealed two highly conserved guanosine-rich regions, two invariant adenosine residues, and two regions of predominantly Watson--Crick covariation. This data led us to propose a model of the ATP-binding DNA structure which is based on a stable framework composed of two stacked G-quartets. The two highly conserved adenosine residues may stack between the top G-quartet and the two short stems, forming a pocket in which the adenosine or ATP ligand binds. Site-directed mutagenesis, base analog substitution studies, and the design of highly divergent but functional sequences provide support for this model.

We have previously described a collection of mutations in conserved residues of the core of the Tetrahymena self-splicing intron. Most of these single base substitutions have less than 10% of the activity of their parental intron derivative [Couture, S., et al., (1990) J. Mol. Biol., 215, 345-358]. We examined the effect of two agents known to stabilize RNA structure, spermidine and ethanol, on the activity of many of these mutant RNAs. In the presence of either 5 mM spermidine or 20% ethanol most substitution mutations were partially or completely suppressed. These conditions also increased the temperature optima of both wild-type and mutant ribozymes. In addition, we find that mutations are also suppressed by a high concentration of GTP, a substrate in the reaction which is bound specifically by the intron. Thus we observe a general suppression of mutations in an RNA enzyme (ribozyme) by spermidine, ethanol and by substrate stabilization. These results are consistent with the idea that most mutations destabilize the folded structure of the ribozyme and can be suppressed by any of a variety of stabilizing influences.

We have isolated a large number of polynucleotide kinase ribozymes from a pool of RNA molecules consisting of an ATP-binding domain flanked by regions of random sequence. Different classes of kinases catalyse the transfer of the gamma-thiophosphate of ATP-gamma S to the 5'-hydroxyl or to internal 2'-hydroxyls. An engineered version of one class is able to catalyse the transfer of thiophosphate from ATP-gamma S to the 5'-hydroxyl of an exogenous oligoribonucleotide substrate with multiple turnover, thus acting as a true enzyme.

1H NMR studies of a 30-nucleotide RNA oligonucleotide (RBE3), which contains a high-affinity binding site for Rev of the HIV-1 Rev responsive element (RRE), two derivatives of RBE3 (RBE3AA and RBE3-A), and the complex of RBE3 with peptides derived from the RNA binding domain of HIV-1 Rev, are presented. The high-affinity binding site of the RRE consists of an asymmetric internal loop and surrounding Watson-Crick base pairs. In the wild-type RRE, one of the stems is closed by a loop; this is replaced in REB3 by the stable UUCG tetraloop. NOE data suggest that the internal loop of the free RNA contains structural features that have been predicted on the basis of in vitro selection experiments [Bartel, D.P., et al. (1991) Cell 67, 529-536]. The structural features include a Gsyn.Ganti base pair, a Ganti.Aanti base pair, and a looped out U. When the Rev peptide is bound to the RNA, the base pairs in the internal loop appear to be stabilized, although the RNA chemical shifts indicate that the RNA conformation undergoes some changes when bound by Rev peptide.

The group I self-splicing introns are RNA enzymes that catalyze phosphodiester-exchange reactions. These ribozymes have a highly specific binding site for guanosine, a substrate for the first self-splicing reaction (Bass & Cech, 1984). The binding site for guanosine has been localized to a specific region of the ribozyme (Michel et al., 1989), but the conformation of the bound guanosine substrate remains unknown. Most analogs of guanosine with substituents at C8 have a preference for the syn conformation; however, some C8-substituted analogs have the potential to form a hydrogen bond between the C8 substituent and the 5'-hydroxyl that would stabilize the anti conformation; we have found that analogs with the potential to form such a hydrogen bond are more active substrates than those that cannot form such a hydrogen bond. These observations led us to test 8-5'-O-cycloguanosine, which is locked in the anti conformation, and 8-(alpha-hydroxyisopropyl)guanosine, which is locked in the syn conformation; the former is active as a substrate, while the latter is inactive. These results strongly suggest that guanosine is bound to the ribozyme in the anti conformation and provide an additional constraint on structural models of this RNA enzyme. We have also examined a series of N7-substituted guanosine analogs; this position had previously been assumed to be unimportant for substrate binding since 7-methylguanosine is an excellent substrate. However, we have found that 7-deazaguanosine and 7-methyl-7-deazaguanosine are less active substrates than guanosine. We discuss several models for the role of N7 in guanosine binding.