SELEXION. Systematic evolution of ligands by exponential enrichment with integrated optimization by non-linear analysis

In vitro selection and evolution of functional proteins by using ribosome display

We report here a system with which a correctly folded complete protein and its encoding mRNA both remain attached to the ribosome and can be enriched for the ligand-binding properties of the native protein. We have selected a single-chain fragment (scFv) of an antibody 10(8)-fold by five cycles of transcription, translation, antigen-affinity selection, and PCR. The selected scFv fragments all mutated in vitro by acquiring up to four unrelated amino acid exchanges over the five generations, but they remained fully compatible with antigen binding. Libraries of native folded proteins can now be screened and made to evolve in a cell-free system without any transformation or constraints imposed by the host cell.

In vitro genetic selection analysis of alfalfa mosaic virus coat protein binding to 3'-terminal AUGC repeats in the viral RNAs

The coat proteins of alfalfa mosaic virus (AMV) and the related ilarviruses bind specifically to the 3' untranslated regions of the viral RNAs, which contain conserved repeats of the tetranucleotide sequence AUGC. The purpose of this study was to develop a more detailed understanding of RNA sequence and/or structural determinants required for coat protein binding by characterizing the role of the AUGC repeats. Starting with a complex pool of 39-nucleotide RNA molecules containing random substitutions in the AUGC repeats, in vitro genetic selection was used to identify RNAs that bound coat protein. After six iterative rounds of selection, amplification, and reselection, 25% of the RNAs selected from the randomized pool were wild type; that is, they contained all four AUGC sequences. Among the 31 clones analyzed, AUGC was clearly the preferred selected sequence at the four repeats, but some nucleotide sequence variability was observed at AUGC(865-868) if the other three AUGC repeats were present. Variant RNAs that bound coat protein with affinities equal to or greater than that of the wild-type molecule were not selected. To extend the in vitro selection results, RNAs containing specific nucleotide substitutions were transcribed in vitro and tested in coat protein and peptide binding assays. The data strongly suggest that the AUGC repeats provide sequence-specific determinants and contribute to a structural platform for specific coat protein binding. Coat protein may function in maintaining the 3' ends of the genomic RNAs during replication by stabilizing an RNA structure that defines the 3' terminus as the initiation site for minus-strand synthesis.

Nucleosome packaging and nucleosome positioning of genomic DNA

The goals of this study were to assess the extent to which bulk genomic DNA sequences contribute to their own packaging in nucleosomes and to reveal the relationship between nucleosome packaging and positioning. Using a competitive nucleosome reconstitution assay, we found that at least 95% of bulk DNA sequences have an affinity for histone octamer in nucleosomes that is similar to that of randomly synthesized DNA; they contribute little to their own packaging at the level of individual nucleosomes. An equation was developed that relates the measured free energy to the fractional occupancy of specific nucleosome positions. Evidently, the bulk of eukaryotic genomic DNA is also not evolved or constrained for significant sequence-directed nucleosome positioning at the level of individual nucleosomes. Implications for gene regulation in vivo are discussed.

Libraries for genomic SELEX

An increasing number of proteins are being identified that regulate gene expression by binding specific nucleic acidsin vivo. A method termed genomic SELEX facilitates the rapid identification of networks of protein-nucleic acid interactions by identifying within the genomic sequences of an organism the highest affinity sites for any protein of the organism. As with its progenitor, SELEX of random-sequence nucleic acids, genomic SELEX involves iterative binding, partitioning, and amplification of nucleic acids. The two methods differ in that the variable region of the nucleic acid library for genomic SELEX is derived from the genome of an organism. We have used a quick and simple method to construct Escherichia coli, Saccharomyces cerevisiae, and human genomic DNA PCR libraries that can be transcribed with T7 RNA polymerase. We present evidence that the libraries contain overlapping inserts starting at most of the positions within the genome, making these libraries suitable for genomic SELEX.

Inhibitory DNA ligands to platelet-derived growth factor B-chain

We have identified a group of DNA molecules that bind to platelet-derived growth factor (PDGF)-AB with subnanomolar affinity from a randomized DNA library using in vitro selection. Individual ligands cloned from the affinity-enriched pool bind to PDGF-AB and PDGF-BB with comparably high affinity (Kd approximately 10(-10) M) and to PDGF-AA with lower affinity (> 10(-8) M), indicating specific recognition of the PDGF B-chain in the context of the hetero- or homodimer. The consensus secondary structure motif for most of the high-affinity ligands is a three-way helix junction with a three-nucleotide loop at the branch point. Photo-cross-linking experiments with 5-iodo-2'-deoxyuridine-substituted ligands establish a point contact between a thymidine nucleotide in the helix junction loop region and phenylalanine 84 of the PDGF-B chain. Representative minimal DNA ligands inhibit the binding of 125I-PDGF-BB but not of 125I-PDGF-AA to PDGF alpha- or beta-receptors expressed in porcine aortic endothelial (PAE) cells in a concentration-dependent manner with half-maximal effects of approximately 1 nM. The same ligands also exhibit a similar inhibitory effect on PDGF-BB-dependent [3H]thymidine incorporation in PAE cells expressing the PDGF beta-receptors. These DNA ligands represent a novel class of specific and potent antagonists of PDGF-BB and, by inference, PDGF-AB.

Representation of unique sequences in libraries of randomized nucleic acids

From a library of nucleic acid molecules, which are randomized in parts of their sequence, unique sequence variants can be selected for specific properties. The planning of such an in vitro selection experiment requires some consideration regarding how much DNA template or RNA transcript should be used initially. The amount applied depends on the number of randomized nucleotides and on the expectations of how often each conceivable and unique sequence combination should be represented in the experimental pool. We display graphs describing the probability for the representation of unique nucleic acid molecules in a randomized pool as a function of the mean representation k, defined by the ratio of sampled nucleic acid molecules to conceivable sequence combinations and we summarize the amounts required to represent unique sequences with 99% likelihood. The probability of representation, P = 1-e-k, can be applied also to 'sub-saturated' pools (k < 1) of nucleic acids with long randomized domains, where it is impossible to provide sufficient material for full sequence representation.

An enzyme-linked oligonucleotide assay

The recent development of in vitro methods to select high-affinity ligands by combinatorial chemistry methodologies promises unique and theoretically unlimited supplies of novel therapeutic and diagnostic reagents. One such combinatorial chemistry process, systematic evolution of ligands by exponential enrichment (SELEX), allows rapid identification, from large random sequence pools, of the few oligonucleotide sequences that bind to a desired target molecule with high affinity and specificity. We describe an enzyme-linked sandwich assay that uses a SELEX-derived oligonucleotide. This assay demonstrates that these oligonucleotides can be effective and useful analytical reagents.

Multiplex selection technique (MuST): an approach to clone transcription factor binding sites

We have used a multiplex selection approach to construct a library of DNA-protein interaction sites recognized by many of the DNA-binding proteins present in a cell type. An estimated minimum of two-thirds of the binding sites present in a library prepared from activated Jurkat T cells represent authentic transcription factor binding sites. We used the library for isolation of "optimal" binding site probes that facilitated cloning of a factor and to identify binding activities induced within 2 hr of activation of Jurkat cells. Since a large fraction of the oligonucleotides obtained appear to represent "optimal" binding sites for sequence-specific DNA-binding proteins, it is feasible to construct a catalog of consensus binding sites for DNA-binding proteins in a given cell type. Qualitative and quantitative comparisons of the catalogs of binding site sequences from various cell types could provide valuable insights into the process of differentiation acting at the level of transcriptional control.

The impact of oxidative stress on eukaryotic iron metabolism

The processes of iron uptake and distribution are highly regulated in mammalian cells. Expression of the transferrin receptor is increased when cells are iron-depleted, while expression of the iron sequestration protein ferritin is increased in cells that are iron-replete. Regulation of expression of proteins of iron uptake (transferrin receptor) and iron sequestration (ferritin) presumably ensures that levels of reactive free iron are not high in cells. Formation of reactive oxygen species occurs when free iron reacts with oxygen, and tight regulation of iron metabolism may enable cells to avoid engaging in destructive chemical reactions. Levels of intracellular iron are directly sensed by two iron sensing proteins. Iron regulatory protein 1 (IRP1) is a bifunctional protein; in cells that are iron-replete, IRP1 contains an iron-sulfur cluster and functions as cytosolic aconitase. In cells that are iron-depleted, IRP1 binds stem-loop structures in RNA transcripts known as iron responsive elements (IREs). Iron regulatory protein 2 (IRP2) binds similar stem-loop structures, but the mode of regulation of IRP2 is different in that IRP2 is rapidly degraded in iron-replete cells. The post-transcriptional regulation of genes of iron metabolism in mammalian cells ensures that cells have an adequate supply of iron, and also ensures that cells do not generate excess reactive oxygen species through the interaction of free iron and oxygen.

Isolation of virus-neutralizing RNAs from a large pool of random sequences

RNA and ribonuclease-resistant RNA analogs that bound and neutralized Rous sarcoma virus (RSV) were isolated from a large pool of random sequences by multiple cycles of in vitro selection using infectious viral particles. The selected RNA pool of RSV-binding sequences at a concentration of 0.16 microM completely neutralized the virus. Of 19 sequences cloned from the selected pool, 5 inhibited RSV infection. The selected RNA and RNA analogs were shown to neutralize RSV by interacting with the virus, rather than by adversely affecting the host cells. The selection of the anti-RSV RNA and RNA analogs by intact virions immediately suggests the potential application of this approach to develop RNA and RNA analogs as inhibitors of other viruses such as human immunodeficiency virus.

In vitro selection methodologies to probe RNA function and structure

In vitro selection, or SELEX, has been used both to characterize the interaction of natural nucleic acids with proteins and to generate novel nucleic acid-binding species, or aptamers. Although numerous reports have demonstrated the power of the technique, they have not expanded on the methodologies that can be used for selection. This review focuses on the considerations and problems involved in selecting protein-binding aptamers from a random-sequence RNA pool. As an illustration, we describe two approaches to selecting aptamers to a particular target, the HTLV-I Rex protein. In the first, complete randomization is used to find an artificial, high-affinity RNA binding site. In the second, the contributions of individual nucleotides and/or base pairs to the natural Rex-binding element are determined by mutating the wild-type sequence and selecting active binding variants.

RNA ligands to human nerve growth factor

High affinity RNA ligands to human nerve growth factor (NGF) were selected from pools of random RNA using SELEX [Tuerk, C. and Gold, L. (1990) Science, 249, 505-510]. Nerve growth factor, which is a protein required for the development of neurons, is not known to bind nucleic acids as part of its natural function. We describe two of the selected RNA molecules in detail. One of them is highly structured, folding into a pseudoknot with an additional hairpin-loop; this structure provides salt-resistant binding to NGF. The other is unstructured and elevated salt concentrations inhibit its binding. These molecules compete with each other for NGF binding. Our RNAs may furnish useful diagnostic tools for the study of an important neurotrophic protein; additionally, they illustrate another example of the potential for nucleic acids to take part in novel binding interactions.

Recognition by viral and cellular DNA polymerases of nucleosides bearing bases with nonstandard hydrogen bonding patterns

The ability of DNA polymerases (pols) to catalyze the template-directed synthesis of duplex oligonucleotides containing a nonstandard Watson-Crick base pair between a nucleotide bearing a 5-(2,4-diaminopyrimidine) heterocycle (d kappa) and a nucleotide bearing either deoxyxanthosine (dX) or N1-methyloxoformycin B (pi) has been investigated. The kappa-X and kappa-pi base pairs are jointed by a hydrogen bonding pattern different from and exclusive of those joining the AT and GC base pairs. Reverse transcriptase from human immunodeficiency virus type 1 (HIV-1) incorporates dXTP into an oligonucleotide opposite d kappa in a template with good fidelity. With lower efficiency and fidelity, HIV-1 reverse transcriptase also incorporates d kappa TP opposite dX in the template. With d pi in the template, no incorporation of d kappa TP was observed with HIV reverse transcriptase. The Klenow fragment of DNA pol I from Escherichia coli does not incorporate d kappa TP opposite dX in a template but does incorporate dXTP opposite d kappa. Bovine DNA pols alpha, beta, and epsilon accept neither dXTP opposite d kappa nor d kappa TP opposite d pi. DNA pols alpha and epsilon (but not beta) incorporate d kappa TP opposite dX in a template but discontinue elongation after incorporating a single additional base. These results are discussed in light of the crystal structure for pol beta and general considerations of how polymerases must interact with an incoming base pair to faithfully copy genetic information.

Distinct binding specificities and functions of higher eukaryotic polypyrimidine tract-binding proteins

In higher eukaryotes, the polypyrimidine-tract (Py-tract) adjacent to the 3' splice site is recognized by several proteins, including the essential splicing factor U2AF65, the splicing regulator Sex-lethal (Sxl), and polypyrimidine tract-binding protein (PTB), whose function is unknown. Iterative in vitro genetic selection was used to show that these proteins have distinct sequence preferences. The uridine-rich degenerate sequences selected by U2AF65 are similar to those present in the diverse array of natural metazoan Py-tracts. In contrast, the Sxl-consensus is a highly specific sequence, which can help explain the ability of Sxl to regulate splicing of transformer pre-mRNA and autoregulate splicing of its own pre-mRNA. The PTB-consensus is not a typical Py-tract; it can be found in certain alternatively spliced pre-mRNAs that undergo negative regulation. Here it is shown that PTB can regulate alternative splicing by selectively repressing 3' splice sites that contain a PTB-binding site.

Modified RNA sequence pools for in vitro selection

We report the use of modified RNA, in which the 2'-OH group of pyrimidines is replaced by a 2'-amino (2'-NH2) group to identify high affinity ligands specific for human neutrophil elastase (HNE) by in vitro selection. Compared to unmodified RNA the 2'-NH2-modified RNA ligands show enhanced stability in human serum and urine. Use of RNase T1 cleavage data in the presence of K+ and Li+ ions suggests that the modified RNA ligands selected for HNE form an intermolecular G-quartet structure.

Inhibition of receptor binding by high-affinity RNA ligands to vascular endothelial growth factor

The proliferation of new blood vessels (angiogenesis) is a process that accompanies many pathological conditions including rheumatoid arthritis and solid tumor growth. Among angiogenic cytokines that have been identified to date, vascular endothelial growth factor (VEGF) is one of the most potent. We used SELEX [systematic evolution of ligands by exponential enrichment; Tuerk, C., & Gold, L. (1990) Science 249, 505-510] to identify RNA ligands that bind to VEGF in a specific manner with affinities in the low nanomolar range. Ligands were selected from a starting pool of about 10(14) RNA molecules containing 30 randomized positions. Isolates from the affinity-enriched pool were grouped into six distinct families on the basis of primary and secondary structure similarities. Minimal sequence information required for high-affinity binding to VEGF is contained in 29-36-nucleotide motifs. Binding of truncated (minimal) high-affinity ligands to VEGF is competitive with that of other truncated ligands and heparin. Furthermore, truncated ligands from the six ligand families inhibit binding of [125I]VEGF to its cell-surface receptors. Oligonucleotide ligands described here represent an initial set of lead compounds in our ongoing effort toward the development of potent and specific VEGF antagonists.

RNA binding specificity of hnRNP A1: significance of hnRNP A1 high-affinity binding sites in pre-mRNA splicing

Pre-mRNA is processed as a large complex of pre-mRNA, snRNPs and pre-mRNA binding proteins (hnRNP proteins). The significance of hnRNP proteins in mRNA biogenesis is likely to be reflected in their RNA binding properties. We have determined the RNA binding specificity of hnRNP A1 and of each of its two RNA binding domains (RBDs), by selection/amplification from pools of random sequence RNA. Unique RNA molecules were selected by hnRNP A1 and each individual RBD, suggesting that the RNA binding specificity of hnRNP A1 is the result of both RBDs acting as a single RNA binding composite. Interestingly, the consensus high-affinity hnRNP A1 binding site, UAGGGA/U, resembles the consensus sequences of vertebrate 5' and 3' splice sites. The highest affinity 'winner' sequence for hnRNP A1 contained a duplication of this sequence separated by two nucleotides, and was bound by hnRNP A1 with an apparent dissociation constant of 1 x 10(-9) M. hnRNP A1 also bound other RNA sequences, including pre-mRNA splice sites and an intron-derived sequence, but with reduced affinities, demonstrating that hnRNP A1 binds different RNA sequences with a > 100-fold range of affinities. These experiments demonstrate that hnRNP A1 is a sequence-specific RNA binding protein. UV light-induced protein-RNA crosslinking in nuclear extracts demonstrated that an oligoribonucleotide containing the A1 winner sequence can be used as a specific affinity reagent for hnRNP A1 and an unidentified 50 kDa protein. We also show that this oligoribonucleotide, as well as two others containing 5' and 3' pre-mRNA splice sites, are potent inhibitors of in vitro pre-mRNA splicing.

Characterization of an in vitro-selected RNA ligand to the HIV-1 Rev protein

A small RNA ligand with high affinity for the HIV-1 Rev protein, generated by the SELEX in vitro evolution method, was used in a series of chemical modification studies to aid in determining the secondary structure of the ligand, to detect which modifications interfere with the binding of the ligand to Rev, and to find those modifiable groups that are protected from attack when bound to the Rev protein. This SELEX RNA ligand, like the high-affinity binding site of the Rev-responsive element, seems to bind the Rev protein within or along the major groove. There are two major regions of the RNA that interact with the Rev protein, and these regions appear to be close in space. Additionally, this high-affinity ligand has been used as the basis for an additional "biased randomization" SELEX procedure, in an effort to gain comprehensive information on the RNA sequences and structural elements necessary for efficient binding to the Rev protein. This complementary experimental approach supports the structural conclusions of our chemical modification data.

High-affinity RNA ligands to basic fibroblast growth factor inhibit receptor binding

We have isolated RNA ligands with low-nanomolar affinity and high specificity to basic fibroblast growth factor from a pool of 10(14) molecules containing 30 randomized positions by the systematic evolution of ligands by exponential enrichment (SELEX) procedure. High-affinity ligands could be classified into two families based on sequence and secondary structure similarities. Representative RNA ligands from the two families compete with one another as well as with heparin for binding to the protein. Furthermore, we show that these ligands inhibit the first step in the signaling pathway of basic fibroblast growth factor: binding of the growth factor to its cell-surface receptors. These findings emphasize the general usefulness of SELEX as a tool for discovering potent, specific oligonucleotide antagonists of target proteins.

Thrombin-binding DNA aptamer forms a unimolecular quadruplex structure in solution

We have used two-dimensional 1H NMR spectroscopy to study the conformation of the thrombin-binding aptamer d(GGTTGGTGTGGTTGG) in solution. This is one of a series of thrombin-binding DNA aptamers with a consensus 15-base sequence that was recently isolated and shown to inhibit thrombin-catalyzed fibrin clot formation in vitro [Bock, L. C., Griffin, L. C., Latham, J. A., Vermaas, E. H. & Toole, J. J. (1992) Nature (London) 355, 564-566]. The oligonucleotide forms a unimolecular DNA quadruplex consisting of two G-quartets connected by two TT loops and one TGT loop. A potential T.T bp is formed between the two TT loops across the diagonal of the top G-quartet. Thus, all of the invariant bases in the consensus sequence are base-paired. This aptamer structure was determined by NMR and illustrates that this molecule forms a specific folded structure. Knowledge of this structure may be used in the further development of oligonucleotide-based thrombin inhibitors.

CASTing for multicomponent DNA-binding complexes

Degenerate oligonucleotides and polymerase chain reaction-based reiterative selection techniques have been used to define the consensus binding sites for an increasing number of transcription factors. The use of crude nuclear extracts rather than purified proteins permits multicomponent complexes to form, and allows the technique to generate information about the combinatorial interactions involved in gene regulation.

Selection of high affinity RNA ligands to the bacteriophage R17 coat protein

RNA ligands with high affinity for the bacteriophage R17 coat protein were isolated from a pool of random RNA molecules using SELEX. Of the 38 ligands isolated, 36 were found to contain a hairpin very similar to the naturally occurring coat protein binding site in the R17 genome. The common features of these 36 sequences provide a consensus binding site and predict components of a hairpin that promote favorable interaction with the coat protein. These include a tetraloop of primary sequence AUCA and a variable-length stem with a bulged adenosine residue at a specific stem position. The predicted consensus agrees well with the highest-affinity RNA binding site of the R17 coat protein, identified through classical but laborious techniques. These results demonstrate the value of SELEX as a tool for isolating high affinity RNA ligands to a specific target protein, and the further value of those ligands to point the researcher toward natural sequences for that target protein.

RNA pseudoknots that inhibit human immunodeficiency virus type 1 reverse transcriptase

High-affinity ligands of the reverse transcriptase of human immunodeficiency virus type 1 (HIV-1) were isolated by the SELEX procedure (systematic evolution of ligands by exponential enrichment) from RNA populations randomized at 32 positions. Analysis of these ligands revealed a pseudoknot consensus with primary sequence bias at some positions. We demonstrated that at least one of the ligands inhibits cDNA synthesis by HIV reverse transcriptase but fails to inhibit other reverse transcriptases. These experiments highlight the power of SELEX to yield highly specific ligands that reduce the activity of target proteins. Such ligands may provide therapeutic reagents for viral and other diseases.